MATHEMATICS
1
S. L. PARSONSON
Senior Mathematics Master
Harrow School
Preface page v
1 Numbers and inequalities 1
2 Vectors and vector geometry 24
3 Coordinates 45
4 Polynomials 68
5 Functions and inequalities 78
6 The trigonometric functions 87
7 Probability in finite outcome spaces 103
8 Finite series and the binomial theorem 130
Revision exercise A 151
9 Mathematical induction 155
10 Expectation 169
11 Further vectors 189
12 Further trigonometry 215
13 Matrices 1 238
14 Matrices 2 258
15 Linear equations 281
16 Discrete probability distributions 301
Revision exercise B 328
Bibliography 339
Answers 343
Index 375
iii
Preface
The present book is the first of a two volume course covering those parts
of modern ' A' level pure mathematics syllabuses not normally included
in standard calculus texts. The reason for omitting calculus is simply to
keep the book within manageable proportions and it is felt that there are a
number of excellent modern calculus books available. Although no formal
calculus is done here, it is expected that a student will be studying calculus
concurrently with this text; thus, for example, it is assumed in Chapter 5
that the reader can differentiate simple algebraic expressions and sketch
the graphs of rational functions. In Chapter 16, the exponential function
is used.
It is hoped that the order of presentation of topics in this book will offer
an effective teaching programme, but variations can be made at the dis-
cretion of the teacher. Thus, for example, the chapters on probability can
be deferred while the two chapters on trigonometry could be taken in con-
junction. Again certain chapters contain work which might be deemed
suitable for a second reading; for example, Sections 4 and 5 of Chapter 10.
In some ways Chapter 1 offers the most difficult problem of presentation
in the whole book: it is necessary later to be able to refer to rational and
irrational numbers, and yet to devote too much time in the early stages
to such topics may not be desirable. The author hopes that he has found
an acceptable compromise but some might yet find parts of Chapter 1 too
formal, in which case they are strongly advised to leave the detailed study
of Sections 2, 3 and 4 for a second reading.
To learn mathematics, constant practice is necessary, some of it repeti-
tive. The book is therefore liberally supplied with exercises for the student.
Most of the questions represent fairly straightforward applications of the
bookwork, although in the Miscellaneous Exercises, included at the end
of all but two of the chapters, will be found some rather more searching
questions. Furthermore, questions marked Ex. occur in the text; it is
hoped that most, if not all, of these will be attempted by the pupil as he
covers the associated bookwork—certainly those marked with an asterisk
should be regarded as obligatory.
The syllabuses for M.E.I. ' A' level, the Joint Matriculation Board
syllabus in Mathematics (Advanced) and Further Mathematics (Advanced)
and the University of London Revised Syllabus in Mathematics (Advanced)
PREFACE
have been particularly kept in mind in the writing of this book and the
planning of the next. Other boards are in the process of devising new
syllabuses and draft copies published by them indicate that this book will
probably cover the necessary work. The School Mathematics Project have
produced their own text-books but the present book may be used to supple-
ment these if desired.
It is a pleasure to record my thanks to Mr M. J. Rawlinson who read
part of the text and made a number of valuable suggestions; to Mr A. J.
Moakes who read the entire book and whose detailed criticisms have done
much to remove obscurities and improve the presentation; and to my wife,
who also read the entire book and who lent invaluable assistance in check-
ing answers. I am grateful for permission to reprint examination questions
from the following Boards: Oxford and Cambridge (0 &C, M.E.I., S.M.P.),
the Joint Matriculation Board (J.M.B.), the Cambridge Local Examination
Syndicate (Cambridge), University of London (London), and the Oxford
Delegacy of Local Examinations (Oxford); also to the Clarendon Press
for permission to use Oxford Scholarship questions (0.S.) and to the
Cambridge University Press and University Registry for the use of Cam-
bridge Scholarship (C.S.) and Mathematical Tripos (M.T.) questions.
S.L.P.
vi
1. Numbers and inequalities
Ex. 1. Explain why the method of conversion from one scale to another by con-
tinued division works.
Ex. 2. Evaluate 11011 x 1011 in the binary scale. Perform several other multi-
plications of this type and check your answers by conversion to the denary scale.
Ex. 3. Evaluate 1110011 - 11011 in the binary scale. Perform several other
divisions of this type and check your answers by conversion to the denary scale.
Exercise 1(a)
1. Solve the following equation for x, working in the binary scale throughout:
(i) 101x+11 = 1101;
(ii) 11x-11111 = 1011;
... 1 x
0 + 11
= 111;
(iii) 101
(iv) 11(10x — 101) = 10x + 101.
2. Express the denary number 275 in the scales of 2, 3 and 12. (For the last
part, you will need to supplement the digits 0, 1, ..., 9 by t = ten and e = eleven.
Why ?)
3. With the notation of Question 2, evaluate 4te x 19t in the scale of 12.
4. With the notation of Question 2, evaluate e89t + 2ee in the scale of 12.
5. Explain the following method of conversion from the binary form 1100110111
to denary form 823
1 100 110 111 Binary
1 4 6 7 Octal
512 + 256 + 48 + 7 = 823 Denary
6. Show that 1331 is a perfect cube whatever the base b, provided that b is greater
than 3.
2
1] THE INTEGERS
7. The number x lies between 100 and 999 (denary scale) and the number y is
formed by writing the digits of x in the reverse order. Prove that x y is divisible
by 99, where x y means the difference between x and y.
8. Show that the difference between any number with four digits (in the denary
scale) and the number with these digits reversed is divisible by 9, and that, if the
two middle digits are the same, the difference is also divisible by 37.
Example 1. Prove, using only the laws (i)—(vi) above, that (— a) (— b) = ab.
First observe that
aa+ a0 = a(a+ 0) by (iv)
= as by (v)
= aa+ 0 by (v),
a0 = 0
on adding — aa to both sides and using (ii) and (i). Thus, any integer multi-
plied by zero gives the answer zero.
Next, consider
ab + [a(— b)+ (— a) (—b)] = a[b + (— b)]+ (— a) (—b) by (iii), (iv)
= a0 +(— a) (—b) by (vi);
= 0 + (— a) (—b) by result
proved above;
=(— a)(—b) by (v).
3
NUMBERS AND INEQUALITIES 1]
But
ab + [a(— b)+ (— a) (—b)] = ab + [a + (— a)] (—b) by (ii) and
(iv);
= ab + 0(— b) by (vi);
= ab +0 by result
proved above;
= ab by (v).
Thus ab = (— a) (— b).
Laws (i)—(vi) above supply almost all the apparatus required for the
manipulation of integers. However, as Example 2 below shows, they need
to be supplemented by one more law:
(vii) The cancellation law: ab = a c b = c, provided a 0. (The
sign =- is read as ' implies '; -= means `is implied by' and means `implies
and is implied by'.)
In this section we have given only a brief survey of the logical structure
of the integers. The reader interested in acquiring a deeper understanding
of this topic and, indeed, of the other topics mentioned in this chapter
should consult one of the books mentioned in the Bibliography at the end
of the book.
3. RATIONAL NUMBERS
As soon as numbers began to be applied to problems more complex than
the mere counting of objects, the value of subdividing the interval between
two integers must have become apparent. For example, when a unit of
length was defined, lengths must have been met that were not an exact
integral number of units. Thus the concept of a fraction or, as we shall
prefer to call it, a rational number, was evolved.
In order to emphasize the fact that we are now dealing with a new type
of number, we shall avoid the familiar fractional form for rational numbers
at first, and instead we shall define a rational number as a pair of integers
written in a definite order thus : (p, q), where q + 0. (We may call this an
ordered pair of integers: (2, 3) and (3, 2) represent different rational
numbers. See Chapter 4.)
(In what follows it may help the reader to see what is happening if he
bears in mind that our aim is to demonstrate that the rational number
(p, q) is what he would call p/q.)
Two rational numbers (pi, q1) and (p,, q2) are said to be equal if
p2q1 = 0; if pia,— p,q, + 0, they are said to be unequal.
Ex. 8. Prove that (1, 2), (2, 4), (13, 26) are all equal. Prove more generally that
(p, q) = (kp, kq), provided k * 0.
Since rational numbers are newly defined objects, rules for adding,
subtracting, multiplying and dividing them must be given, for these opera-
tions have so far only been applied to integers. The rules are
(Pr, q1)+ (P2, q2) = (P1q2+ P2qi, q1q2); (Pi, — (P2, q2) = (Piq2 —P2qi, q1q2);
(p1, q2) x (pa, q2) = (P1P2, qiq,); (Pi, 0 (P2, q2) = (P1q2, P20,
Example 3. Verify that the distributive law holds for rational numbers.
We have to show that
(p,, q1) x [(P2, q2)± (pa, q3)] = q1) x (Pa, q2)± (pi, q1) x (ma),
= (P1P293+P1Paq2, 91q2q3))
5
NUMBERS AND INEQUALITIES [1
Ex. 10 shows us that we may identify the integers with rational num-
bers of the form (p, 1); Ex. 11 then shows us that the rational number
(b, a), when multiplied by the integer (a, 1) gives the integer (b, 1). In
more familiar language (b, a) has just the properties we associate with
the 'number' b/a. We may thus regard a rational number as `the quotient
of two integers'.
It thus follows that the statement ' x is a rational number' is equivalent
to the statement 'integers p and q may be found such that x = p/q'.
The reader will be familiar with the process of expressing a rational
number as a decimal. For example,
18 1 1 1
= 0.024 = 0 x fo +2+Tiy2 +4 x
625
Conversely, a terminating decimal can always be expressed as a rational
number in the form p/q; for example
0.175 = =
However, the terms 'rational number' and `terminating decimal'
are not synonymous. Thus, .* is certainly a rational number, but
= 1.142857142857142857...
and the process of division cannot be brought to an exact conclusion.
However, this expression does recur; the reader will probably be familiar
with the notation.
1.142857.
In fact, any rational number may be expressed either as a terminating or
a recurring decimal, and conversely, any terminating or recurring decimal
represents a rational number. The proof of the italicized part of this state-
6
3] RATIONAL NUMBERS
88
- 37.
4. IRRATIONAL NUMBERS
It is useful to depict the integers and rational numbers as points on a
straight line. In Figure 1.1, an origin 0 is taken, representing the number
zero, and equal intervals are measured to the right, the end-points repre-
senting 1, 2, 3, ... and to the left representing —1, —2, — 3, .... The posi-
tion of a point representing a rational number may be defined quite simply ;
I I I I I I I
—3 —2 —1 0 1 2 3
Fig. 1.1
•
2
Fig. 1.2
Numbers such as V2 which are not rational are called irrational. Since
rational numbers are represented by terminating or recurring decimals,
the decimal representation (if it exists) of an irrational number is non-
terminating and non-recurring.
Irrational numbers obey the same laws of combination as rational
numbers. The proof of this lies beyond the scope of this book, as it requires
a systematic definition of irrational numbers in terms of the rationals ;
8
4] IRRATIONAL NUMBERS
Exercise 1(b)
1. Where does the proof of Example 5 break down if you try to prove ,/4 is
irrational ?
2. By assuming that ,;/2 = plq, where p, q are integers cancelled down into their
lowest form, show that ./2 is irrational.
3. Prove that V3 is irrational.
4. Express as (i) a decimal, (ii) a bicimal, the fraction 3(denary notation). If a
decimal recurs, does it bicimal representation necessarily recur ?
5. Prove that, if a rational number of the form pl q (cancelled down into its
lowest form) is expanded as a recurring decimal, the length of the block of digits
that recurs is less than q.
6. Which of the following statements are always true and which are not? If a
statement is true, prove it; if it may be false, give a counter-example (that is, an
example that illustrates its falsity):
(i) rational + rational = rational;
(ii) rational + irrational = irrational;
(iii) irrational + irrational = irrational;
(iv) rational x rational = rational;
(v) rational x irrational = irrational;
(vi) irrational x irrational = irrational.
7. If a and b are integers and Vb is irrational, prove that (a+ Jb)3is irrational.
(The results of Question 6 may be used in this question.)
8. Prove that it is always possible to find an irrational number that lies between
two given rational numbers a and b.
9. If pi, p„Paare three unequal prime numbers, prove that
1 1 1
— +— +-
P1 P2 P3
is a rational number, but not an integer.
Can you generalize this result?
10. You are given a ruler with only integral units of length marked on it. Show
how to construct, geometrically, the following lengths:
(i) A/3 ; ; (iii) \/(2 + A/2).
11. Prove that, if k is an integer and Vk is rational, then Vk is an integer. Show
further that, for k > 1, V(k2—1) is irrational and deduce that, for k 1,
V(k — 1) + ,/(k+ 1) is irrational.
9
NUMBERS AND INEQUALITIES [I
5. SURDS
Although the existence of irrational numbers is mathematically significant,
from the point of view of practical arithmetic all numbers may be regarded
as rational, since any irrational number may be approximated to by a
terminating decimal. Indeed, wherever measurements are concerned, the
answers obtained must necessarily be in the form of rational numbers,
since every measuring device must eventually reach the limits of its possible
accuracy. However, just as is simpler to handle than its decimal approxi-
mation 0.3333, so J2 is often easier to deal with than its approximation
1-414.
An expression involving only rational numbers and their roots (not
necessarily square roots) is called a surd; thus surds form a class of
irrational numbers, though there are irrational numbers, such as 7r, which
cannot be expressed as surds.
Examples of surds are:
1 2
4J2; :/;; (V2 + v2 11); ,?,/3 + V2.
Example 8. Simplify:
= 1 — V3 + V2
1 —(5-2A/6)
= 1+A/2—V3
2(V6— 2)
_ (1+1/2-1/3) (A/6+2)
2(6-4)
— 1/6+2+1/12+2,/2-1/18-24/3
4
_ 2 + V6 — V2
4
(ii) Suppose 14-41/6 = (Va 21/b)2.
-
11
NUMBERS AND INEQUALITIES
Exercise 1(c)
1. Express in the form aV b, where b has no perfect squares as factors
(i) V50; (ii) V363; (iii) V2400; (iv) V192; (v) V1452.
2. Simplify the following surds:
(i) (V5 1)2; (ii) (20—V2)2; (iii) (V3+1)2; (iv) (V3 V2)3;
- —
V117. V147— V3
(iv) (v)
3 ' 4
12
51 SURDS
6. SETS OF NUMBERS
Suppose we have a collection of numbers, the whole collection being
denoted by the letter U. Suppose, too, we ask a question which may be
answered unambiguously 'Yes' or 'No' for each number (or element)
of 11. Then the collection of those elements of U for which the answer is
`Yes' is said to form a set S, which is a subset of the universal set U,
13
NUMBERS AND INEQUALITIES (1
written S s U. The collection of those elements of U for which the answer
is `No' is called the complement of S, written S'.
It may happen that the answer to the question is 'No ' for each
element of U; in this case we say S is empty and write this as S = 0.
(0 is usually referred to as the null set, or empty set.)
As an example, suppose that U is the set of integers from 1 to 10. We
may write this as
U = {1, 2, 3, ..., 10}.
The question may be posed: 'Is the element x of U divisible by 3 ?'
Those elements for which the answer is 'Yes' form a set A where
A = {3, 6, 9).
The complement of A is given by
A' = {1, 2, 4, 5, 7, 8, 10).
Again, the alternative question 'Is the element x of U irrational ?' may be
asked and unambiguously answered for each element of U. The answer is
`No' in every case; the set defined is thus 0 and its complement is U.
In this example, the universal set U contains only a finite number of
elements. It may well happen, however, that U contains an infinite number
of elements; for example, U might be the set of all positive integers. The
proper subset E defined by the question 'Is the element x of U even ?'
also contains an infinite number of elements. (P is a proper subset of Q
if P is contained in Q but is not the whole of Q, written P = Q. P s Q
means that P is a subset of Q, but may be the whole of Q.)
It is convenient to develop a shorthand notation for the somewhat
cumbersome method we have used so far in defining our sets. The method
adopted is to use braces {:} with a colon (:) or vertical line (I) (read as
`such that') separating the two necessary pieces of information:
(i) within the bracket and to the left of the colon is stated the universal
set from which the elements are drawn;
(ii) within the bracket and to the right of the colon is the statement
defining which particular elements of the universal set are to be chosen.
For instance, for the sets U and A mentioned above we may write
Ex. 20. If U is the set of positive integers {1, 2, 3, ..., 9} write down the elements of
the following sets:
(i) {x e 21 : x + 4 E IT} ; (ii) {x a 1T : Jx is rational).
14
6] SETS OF NUMBERS
Ex. 21. If U is the set of all positive integers {1, 2, 3, ...} describe in words the
following sets:
(i) {x + 1) e 1.4; (ii) {x 11: x-1 e 11};
(iii) {x e -Ayx ell}; (iv) {x e : x2 — 5x 6 = 0} ;
—
Ex. 22. Two sets A and B which are both subsets of the same universal set U are
said to be equal (A = B) if they contain precisely the same elements.
Show that A = B -.44- A B and B S A.
11
Fig. 1.3
An B A'
Fig. 1.5 Fig. 1.6
Ex. 23. Verify, using Venn diagrams, that the operations of union and intersec-
tion are associative; that is
A u (B u C) = (A U B)U C; A n (B = (A n B) n c.
15
NUMBERS AND INEQUALITIES
Ex. 24. Verify, using Venn diagrams, that the operation of union is distributive
over intersection, and that intersection is distributive over union; that is
A u (B n C) = (A U B) n (A U C); A n (B U C) = (A n B) U (A n C).
Ex. 25. Verify de Morgan's Laws, using a Venn diagram:
(i) (A u = A' n B'; (ii) (A n = A' U B'.
(The reader may be able to identify these laws with the logic of AND (n),
OR (U) and NOT 0.)
Ex. 26. If U is the set of all positive integers and
A= {x E e 111, B= {x e e IX}
describe A n B, using the {:} notation.
Ex. 27. If 11 = {a, b, c, d, e} and A = {a, c, e} write down the subsets:
(i) A U A'; (ii) A n A'; (iii) A U 11; (iv) A n 11; (v) A U 0; (vi) A n ;
(vii) 11'; (viii) 0'.
Ex. 28. If, with the notation of Ex. 27, B = {a, b, e}, write down the subsets
(i) A n B; (ii) A' u B; (iii) A n B'.
Verify de Morgan's Laws (see Ex. 25) in this particular case.
Three particular sets, the set of all integers, the set of all rational numbers
and the set of all real numbers occur throughout mathematics with such
unfailing regularity that it is convenient to introduce a notation by which
to refer to them.
The set of all integers is denoted by Z.
The set of all rational numbers is denoted by Q.
The set of all real numbers (that is, the union of the set of all rational
numbers and the set of all irrational numbers) is denoted by R.
With a suitable definition of each of the terms integer, rational and real
we have the following relation between them
Z Q R.
The sets of all positive integers is denoted by Z+, of all positive rational
numbers by Q+ and of all positive real numbers by R+.
7. INEQUALITIES
We have already asserted that the integers and the rational numbers may
be ordered; that is, given two integers (or rational numbers) x and y,
we may answer the question 'Is x greater than y ?' The same is true for the
set of all irrational numbers, or, indeed, for the union of the sets of rational
and irrational numbers (the set of real numbers).
If x, y are two real numbers, x > y (read x is greater than y') means
that the point representing the number x lies to the right of the point
representing the number y. Similarly for x<y(`x is less than y'). When
comparing two positive or negative numbers x and y in this way, we see
that x> y x—y > 0 (i.e. x—y is positive),
x< y a x—y < 0 (i.e. x—y is negative),
x= y p x—y = 0 (i.e. x—y is zero).
The number zero separates the positive and negative numbers; it is itself
neither positive nor negative.
By continued use of these results we see that we may add or subtract
any number from both sides of an inequality without altering the validity
of the inequality sign.
Inequalities are of value in defining sets. Thus, for example, to take a
rather trivial case,
{x E Z: 0 < x < 3) = {1, 2}.
A more substantial example would be
{x e R: x2 < 4}.
Solving an inequality means redefining the set of elements which satisfy
the given inequality in as simple a manner as possible.
The reader should have observed that in, for example, the result
x > y and z > 0 xz > yz
the two-way implication was not used: it is not valid to deduce from
xz > yz that x > y and z > 0. However, inequalities involving products
may be solved by observing that, if the product is positive, then the factors
are either both positive or both negative. Another useful observation is
that a squared number is always positive or zero.
Example 12. Prove that the expression y = x2 + x +1 is positive for all real
values of x, and find its minimum value.
It will be useful in the solution of this example to employ the notation
: x y means ' x is greater than, or equal to, y'.
Now
y = x2 +x+1
= (x++)2 +4-,
= (x+i)2,
3
y — 1 0 or y 4.
Theorem 1.1. The A.M. of two positive numbers is at least as great as their
G.M.
Proof. Since the two numbers are positive, we may write them as
u2 and v2.
Let A = (u2 +v2)/2, G = uv; we have to prove that
A — G 0,
19
NUMBERS AND INEQUALITIES [1
that is u 2+ v2
uv 0,
2
L.H.S. = 1(12± V2 — 2uv)
= +(u v)2
> 0, unless u = v.
We have thus proved that A > G unless u2 = v2in which case A = G.
Example 13. 200 m of wattle fencing are to be bent to form three sides of a
rectangular enclosure, the fourth side being a straight hedge. Find the length
of the rectangle if the area to be enclosed is to be a maximum.
Let x m be the length, y m the breadth, A m2the area of the enclosure.
Then we have
(i) x + 2y = 200, (ii) xy = A.
Both x and 2y are positive; thus, by the theorem just proved, their A.M.
is at least as great as their G.M. That is
1-(x + 2y) V(2xy)•
Using (i) and (ii), this gives
100 V(2A),
i.e. A 5 5000.
Equality occurs only if x = 2y, i.e. x = 100, y = 50. Thus the maximum
area occurs when the length is 100 m.
Exercise 1(d)
1. Verify, using a Venn diagram, the results:
(i) A U (A' U B)' = A n (B' U A),
(ii) (A U B) n (A' n = A U (B n C'),
(iii) [A' u (B' n C')]' = (A n B) U (A n C).
2. If R is the set of real numbers and
A = {x e R: x > 3}, B = {x e R; x < 4}
describe, using the {:} notation
(i) A U B; (ii) A n B; (iii) A' U B.
3. Taking the universal set, U as the set of all integers, describe using the {:}
notation
(i) the set of all negative integers;
(ii) the set of all positive integers divisible by 6;
(iii) the set of all integers, excluding 0, ± 1, ± 2.
4. A = {x e R: — 1 x < 3}, B = {x e R: 2 -4 x < 4}, C = {x e R:x > 3}.
Find expressions for the following sets, using the {:} notation:
(i) A n B;(ii) A' n C; (iii) (A U B) n C'; (iv) A U (B n C); (v) (A' U B') n C;
(vi) A' U (B' n C); (vii) A' n B n C.
20
INEQUALITIES
positive?
12. If xy = 25, find the least possible value for x + y, if both x and y are positive.
13. If xy = 18, find the least possible value for 2x + y, if both x and y are positive.
14. If x + y = 2, find the maximum value of xy.
15. If 2x + 3y = 120, find the maximum value of xy.
16. Some netting is required to make three sides of A
a rectangular chicken-run, the fourth side being an
existing wall. Find the least length of netting needed
for an area of 50 m2.
17. Figure 1.7 shows part of the framework of a kite.
ABCD is a rectangle; P and Q are the mid-points of Q C
AB, CD respectively. Find the maximum area of
ABCD that can be made if 4 m of wood are available. Fig. 1.7
What are the dimensions of the kite necessary to attain this area?
Miscellaneous Exercise I
1. Express 120 as a binary number.
Find the least number of stamps required so that any value from p to 60p
(in steps of -1-p) may be selected, and give their values. (You may assume stamps
of any value are available if required.)
21
NUMBERS AND INEQUALITIES [1
12. Show that the expression x2+ 8xy — 5y2— k(x2 +y2) can be put in the form
a(x+by)2when k has either one or other of two values. Find these values and the
values of a and b corresponding to each value of k.
Prove that when the variables x and y are restricted by the relation
x2 + y2 = 1,
but are otherwise free, then
—7 < x2 + 8xy — 5y2 < 3. (0 & C)
13. a, b, c, d are four unequal positive numbers. By using Theorem 1.1 and
considering first the pair of numbers -Ka +b) and +(c+ d), and then, separately,
the pairs a, b and c, d, prove that
1(a b + c+ d)> (abcd)1.
Deduce, by considering the four unequal numbers a, b, c and +(a+ b + c),
that
+(a + b + c) > (abc)1.
What happens to the last inequality (i) if a = b = c; (ii) a = b c?
Suggest a generalization of the results proved above.
14. If x+ 2y+ 3z = 1, where x, y, z are positive numbers, find the maximum
value of xyz.
If u+v = 1, where u, v are positive numbers, find the maximum value of u2v.
15. A cylindrical vessel, with one end open, is made from a given piece of
material. Show that its volume is greatest if the height and radius are equal.
16. a, b, c are non-zero rational numbers. Show that, if
al-Fg+c* = 0,
and if none of al, bl, c* is rational, then al, bl, cl are each rational multiples
of the same irrational number.
17. Show that, if pig is a good approximation to V2, then (p2 +2q2)/(2pq) is a
better one. Starting with p = q = 1, show that V2 Ps,' 577/408, and estimate the
accuracy of this approximation.
23
2. Vectors and vector geometry
Ex. 1. Can you suggest any addition to the list of vector and scalar quantities
given in Section 1?
Ex. 2. Can three displacements be added together? Does it matter in what
order it is done?
Ex. 3. A rotation may be given a magnitude (size of angle turned through),
direction (axis of rotation) and sense (positive being direction for which rotation
is anticlockwise). Show that rotations are NOT vector quantities. (This example
shows that physical quantities exist that are neither scalar nor vector quantities.)
Fig. 2.3
We may thus drop the brackets when adding three vectors together, and
write a +b +c without ambiguity.
If AB represents the vector a then BA will be said to represent the vector
26
3] OPERATIONS WITH VECTORS
(Note that the zero vector, unlike all others, has no direction.)
If AB = a and AC = b then, since CA + AB = CB, the line segment
CB represents a — b (see Figure 2.5). (Notice that we have used the corn-
mutativity of vector addition here.)
To develop the algebra of vectors further, we need a new definition, the
multiplication of a vector by a number (scalar).
If k is any positive number, ka is the vector of magnitude k la I in the same
direction as a and with the same sense.
b
Fig. 2.5 Fig. 2.6
*Ex. 4. Prove that 0.a = 0. What does this tell us about the magnitude of the
zero vector?
2 PPM
27
VECTORS AND VECTOR GEOMETRY [2
Ex. 5. If a is a vector of magnitude 4 units due north and b is a vector of magni-
tude 1 unit due east, describe the vectors:
(i) 2a; (ii) -3b; (iii) a -4b; (iv) 2(a +b); (v) 2a - 6b.
Ex. 6. If a = 3x + y, b = x - 2y, find x, y in terms of a, b, justifying your argu-
ment in terms of Rules 1-5.
Ex. 7. If a is a vector of magnitude 1 unit due north and b is a vector of magni-
tude 1 unit in the direction N 60° E, describe the following vectors (using a scale
drawing or trigonometry, if necessary):
(i) 2a + b; (ii) - (2a - b); (iii) b -a; (iv) - a - 2b.
Ex. 8. If a is a vector of 2 units due north in a horizontal plane, b is a vector of
3 units due east in a horizontal plane and c is a vector of 1 unit vertically up out
of the plane, describe the vectors:
(i) a - 2c; (ii) - b- 3c; (iii) 3a + 2b + 6c; (iv) 3a - 2b- 6c.
Ex. 9. ABCD is a parallelogram with AB 2 units and BC 1 unit. If a is a vector of
1 unit in the direction AB and b is a vector of 1 unit in the direction of BC, and
if E is the mid-point of CD, find, in terms of a and b, the vectors AC, DB, AE,
EB.
4. COMPONENTS OF A VECTOR
It is often useful to express a given vector r as the sum of a number of
vectors in specified directions
r
If this is done, a1, a2, anare called components of r.
Given a vector r and specified directions two questions immediately
arise :
(i) Can r be split into components in these directions?
(ii) If the answer to (i) is ' Yes', in how many alternative ways can this
be done ?
We shall answer these questions for a vector r and three non-parallel,
non-coplanar directions in three dimensions by proving two theorems of
types that constantly recur in mathematics. The first theorem is an example
of an Existence Theorem: it will answer question (i) above, telling us that a
solution does exist; the second theorem is a Uniqueness Theorem: it will
answer question (ii) above, telling us that the solution which we know
exists is the only possible one, that is, that it is unique. The significance
of these two theorems will become apparent as we proceed.
Theorem 2.1 (The Existence Theorem). Given a non-zero vector r and three
non-coplanar t, non-parallel vectors a, b, c, there exist numbers A, it, v, such
that
r = Aa-Ficb+vc.
f That is, displacements represented by a, b, c, cannot all be chosen to lie in one plane.
28
4] COMPONENTS OF A VECTOR
Fig. 2.7
We have thus proved our Existence Theorem by showing that, with the
given data, a set of components may be constructed. To demonstrate the
truth of the Uniqueness Theorem we employ a common device in mathe-
matical arguments: we assume that it is false and show that this leads to a
contradiction (see Chapter 9).
and any non-zero multiple of a cannot lie in this plane, by the data. Thus
= A2, which contradicts our initial assumption.
Theorems 2.1 and 2.2 have been proved for three dimensions; they hold
equally well in two dimensions, for a vector r lying in the plane deter-
mined by a and b, where a kb.
An expression such as Aa +#b + vc is called a linear combination of the
vectors a, b, c. The theorems we have just proved may be restated in the
form 'Given three non-parallel, non-coplanar vectors a, b, c any non-zero
vector r may be expressed uniquely as a linear combination of a, b, c'.
If numbers A, #, v (not all zero) exist such that
Aa+,ub+vc = 0
the vectors a, b, c are said to be linearly dependent; if no such numbers
exist, a, b, c are said to be linearly independent.
*Ex. 10. Prove that, if 0, A, B, C are coplanar, then the position vectors a, b,
c are linearly dependent and, conversely, if a, b, c are linearly dependent, then
0, A, B, C are coplanar.
*Ex. 11. Prove the Existence and Uniqueness Theorems for components in two
dimensions.
Ex. 12. ABCD is a parallelogram, and E is the mid-point of CD. AB = a,
AD = b, AE = x, BE = y. Express x and y in terms of components in the direc-
tions a and b and also express a and b in terms of components in the directions x
and y.
Ex. 13. Do either the Existence or the Uniqueness Theorems hold in three di-
mensions if four directions a, b, c, d are given?
Exercise 2 a
1. If v = a +2b, w = 2a—b, express the following vectors in the form
A.a
(i) v+w; (ii) 2v +3w; (iii) v — 3w; (iv) 2(v—w).
2. If u = a +3b, v = 2a— b, express a and b in terms of u and v.
3. If u = a + b + c, v = a +2b +c, w = a— b— 2c, find, in terms of a, b, c:
(i) u+ v+w; (ii) u—v—w; (iii) 2u+ v—w; (iv) u+ 2v + 3w.
4. If u = a+b—c, v = 2a— b+c, w = 3a +2b+c, find a, b, c in terms of
u, v, w.
5. If i and j are perpendicular vectors of unit magnitude, i pointing due east and
j due north, find the magnitudes and directions of the following vectors:
(i) 2i; (ii) —3j; (iii) 3i + 4j; (iv) i— j; (v) —3i +4j.
30
4] COMPONENTS OF A VECTOR
Fig. 2.8
and do the same for the vectors OQ and OR. Express PQ in terms of a, b, and
RQ in terms of b, c. Find two different expressions of the form aa + /3b +7c for
the vector OQ, with none of a, fl, y zero. Show that, if one of these expressions
is subtracted from the other, the known relation b = c+ a is obtained.
12. In Figure 2.8 OC = u, OD = v, QP = w. Obtain expressions for OP, OQ,
OR in terms of u, v and hence obtain w in terms of u and v.
Why is it not possible to express CD in terms of the two vectors PQ, AR?
13. ABCDEF is a regular hexagon, and AB = a, BC = b. Find CD, DE, EF,
FA in terms of a and b.
14. ABCDA'B'C'D' is a cuboid whose base ABCD is a square of side 2 units.
The sides AA', etc., are vertical and of magnitude 4 units. E is the mid-point
31
VECTORS AND VECTOR GEOMETRY [2
5. APPLICATIONS TO GEOMETRY
So far we have developed the algebra of vectors using geometrical argu-
ments; we now reverse the process and show that the algebra of vectors
may usefully be employed to deduce geometrical results. The vector treat-
ment of geometrical problems has the great advantage that it is equally
applicable to two or three dimensions.
To describe a geometrical configuration consisting of a number of points
A, B, C, ... we must be able to locate each point. This may be done by
taking a fixed point 0 (called the origin) and referring to a point A by
the line segment OA. If a is the vector representative of all the line seg-
ments equal and parallel to OA, then a is called the position vector of the
point A with respect to the origin 0. Thus, given an origin 0, we may
32
5] APPLICATIONS TO GEOMETRY
refer to all points in the plane by their position vectors relative to this
origin 0.
First, we establish a result known as the Section Formula. This theorem
enables us to write down the position vector of any point on a given line
and so is of importance in setting up a vectorial description of a geometrical
configuration.
Example 1. Show that the medians of a triangle ABC are concurrent and
find the position vector of the centroid (meet of the medians) in terms of the
position vectors a, b, c, referred to some origin 0.
Let the mid-points of BC, CA, AB be L, M, N respectively and call their
position vectors 1, m, n (Figure 2.10).
33
VECTORS AND VECTOR GEOMETRY [2
4A(b+c)+ Aa
i.e.
A+ A
For varying values of A and we obtain different points of the line AL;
in particular, if we choose A = 2, is = 1, we obtain a point whose position
vector is symmetrical in a, b, c and so lies equally well on BM or CN.
Thus, the medians of a triangle are concurrent at the centroid, G, whose
position vector is given by
a+b+c
g— 3 •
Notice, incidentally, that we have shown that AG/GL = 2, etc.
and s— r = c)
Thus QP = RS and so PQRS is a parallelogram (one pair of sides equal
and parallel).
Ex. 14. Show that the joins of mid-points of opposite edges of the tetrahedron
ABCD are concurrent at a point G.
Ex. 15. Show that, in a tetrahedron ABCD, the joins of vertices to the centroids
of opposite faces are concurrent at the same point G as that obtained in Ex. 14.
*Ex. 16. What condition must the numbers a, /3 possess for the points with
position vectors a, b and aa +fib to be collinear?
34
APPLICATIONS TO GEOMETRY
The Section Formula may be restated in a form that gives a test for the
collinearity of three points whose position vectors are known.
cla+ fib+yc = 0
and fld-y = 0
a+/3+y
then A, B, C are collinear.
Proof. (i) Given that A, B, C are collinear, suppose AB/BC = Alit. Then
by the Section Formula
pa + Ac
b=
1H-A
i.e. pa—(,1+A) b+Ac = 0.
Take a = — + A), y = A and the result follows.
(ii) Given that numbers a, y exist such that
aa+fib+7c = 0
and a+/3+y = 0
we have as +fib = (a +ft) c, by substitution.
Now we know that a, f3 y are not all zero and we may assume, without any
,
Denoting the position vectors of the various points in the usual way,
we have, using Theorem 2.4:
fv+aa+a'a' = 0,
jl 1+a+a' = 0;
fv+fib+rb' = 0,
1. 1+fl+fi' = 0;
_fv+7e+7'e' = 0,
1 1+7+7' = O.
Fig. 2.11
Fig. 2.12
Theorem 2.5 (condition for four points to lie on a plane). Four points A, B,
C, D have position vectors a, b, c, d.
Then
(i) If A, B, C, D lie on a plane, there exist numbers a, ft, y, 8, not all
zero, such that
aad - igb+yc+8d = 0,
a+fl+y+8 = 0.
Conversely,
(ii) If there exist numbers a, fl, y, 8, not all zero, such that
aa-Fflb+yc+ad = 0,
a+fi-Fy+8 = 0
then A, B, C, D are coplanar.
Proof. (i) Given that A, B, C, D are coplanar, d satisfies an equation of
the form
r = —(A+11-1)a+Ab+ite,
i.e. —(A+,a-1)a+Ab+icc—d = 0
and the first result follows if we set
cc = — (A -Fit -1), A' = A, Y = It, a = —1.
(ii) Given that numbers a, ft, y, 8 exist such that
aa+flb+yc+ad = 0,
a-1-13+y+8 = 0.
We have —(13±y±8)a+ Ab+yc+ad = 0,
i.e. fi(b— a) + y(c — a) = 8(a — d),
or fiAB+yAC = &DA.
But AB, AC both lie in the plane ABC and so therefore does (MA. It
follows that D lies in the plane ABC and the proof is complete.
*Ex. 19. Show that the equation of the line through A parallel to the direction
defined by the vector u is r = a+ Au.
38
6] EQUATIONS OF LINES AND PLANES
*Ex. 20. Show that the equation of the plane through A parallel to the directions
defined by the two vectors m, n is r = a + Am +mi.
Ex. 21. Show that the mid-points of two pairs of opposite edges of a tetrahedron
ABCD are coplanar.
Example 4. ABCD is the base of a cube whose vertical edges are AA', BB',
CC', DD'. X is the point of trisection of BB' nearer B, Y is the point of
trisection of CC' nearer C' and M is the mid-point of BC.
If D'M cuts the plane AXY at Z, find the ratio in which Z divides D'M.
Take side of cube as 6 units and call vectors of unit magnitude in the
directions AB, AD, AA' respectively i, j, k.
y = 6i+6j+4k
and so the position vector of any point on the plane AXY is given by
r = Ax+ity
= 6(A+µ) i +6,4 +2(A +2,u) k.
Again, m = 6i + 3j,
d' = 6j + 6k
and so D'M = 6i — 3j— 6k.
39
VECTORS AND VECTOR GEOMETRY [2
Exercise 2 (b)
1. ABC is a triangle. U lies on AB produced so that AB = }AU, V lies on AC
so that AV = 2VC. Find, in terms of the position vectors a, b, c of A, B, C,
the position vectors of U, V and the mid-point of UV.
2. Draw a diagram showing the relative positions of the points whose position
vectors are a, b and 3a — 2b.
Prove that these three points are collinear.
3. If G is the centroid of the triangle ABC, prove that
GA + GB + GC = 0.
Suggest a generalization of this result.
4. ABC is an equilateral triangle of side 3 cm. P, Q lie on BC, CA respectively
and are such that AQ = CP = 2 cm. R lies on AB produced so that BR = 1 cm.
Prove that PQR is a straight line.
5. OABC is a parallelogram and the position vectors of A, B, C relative to 0
are respectively a, b, c. M is the mid-point of BC. Write down, in terms of b and
c, and hence in terms of a and c, the position vector of a general point X on the
line OM. Deduce that, if X lies on AC, then X is the point of trisection of AC
nearer C and also the point of trisection of OM nearer M.
6. OAB is a triangle; M is the mid-point of AB and T is the point of trisection
of OB nearer B. TM produced meets OA at X. If OA = a and OB = b, write
down the position vectors of T and M, and hence of a general point P of TM.
Deduce the position vector of X and find a relation between the points 0, A
and X.
40
6] EQUATIONS OF LINES AND PLANES
10. Prove that, if ABC is a plane through the origin 0 from which position
vectors are measured, and if A, B, C have position vectors a, b, c, then there
necessarily exits a relation of the form
pa +qb+rc = 0 (p, q, r * 0)
where 0 is the zero vector.
The lines AO, BO, CO meet BC, CA, AB in L, M, N respectively. Prove that
the position vector of L is
b+ c.
q+r q+r
BL.CM . AN
Deduce that
LC .MA .NB — +1,
where the magnitude and sense of each line segment is taken into account. (M.E.I.)
(The result proved in Question 10 is known as Ceva's Theorem: it and its
converse are very useful for proving concurrency theorems.)
13. P, Q are variable points on two skew lines. Find the locus of the mid-point
of PQ. Can you generalize this result?
15. Two vectors a and b, such that b is not a multiple of a, are given in a plane.
Two other vectors c and d are defined by the equations
c = yia+y2b,
d = ai a+82b.
Prove that any vector ma +fib can be expressed in terms of c and d provided
0. Find the coefficients of c and d.
Y3.82 — Y2.61. *
Explain the geometrical significance of the condition y182 — y2Si *0. (0 & C)
43
VECTORS AND VECTOR GEOMETRY [2
44
3. Coordinates
1. UNIT VECTORS
A unit vector is a vector of unit magnitude. Thus, a unit vector is a sort
of ' signpost ' giving a direction and sense; any vector may be split into
the product of a number (its magnitude) and a unit vector with the required
sense and direction. Thus we may write
r = rP,
where r = In and t is a unit vector in the direction of r. (The notation
employed here is useful and should be adopted by the reader: given any
vector x, its magnitude may be written simply as x while a unit vector
with the same sense and direction as x is written x.) Unit vectors are useful
in that they enable us to deal separately with the magnitude and direction
of a given vector.
Now suppose we choose two directions in a plane and specify them by
unit vectors i and j. By the Existence and Uniqueness Theorem for a
plane (see Chapter 2) any displacement AB in the plane, represented by
d may be split into components in the i and j directions
d = pi+ qj
where p and q are uniquely determined. The vector d, which is representa-
tive of a whole class of displacements, of which AB is a typical member,
may be written in the alternative form
d= .
1
Thus, ( —1) represents the vector 2i — j, ( ) the vector i and (_ 10) the
0
vector — j. Of course, the representation of a vector in this new notation
depends upon our original choice of base vectors i and j. Again, there is
no necessity for the base vectors to be unit vectors, though they will almost
invariably be so. To illustrate these last two remarks, consider two unit
vectors i and j and the vector a where
a = 3i — j.
Now a = (i + j) + 2(i — j)
and so we have
a =— 3) with i, j as base vectors,
45
COORDINATES [3
while
a = ( ) with (i + j), (1— j) as base vectors.
2
Notice that (i + j) and (i — j) are not unit vectors. (For example, if i and j
are perpendicular, neither (i+j) nor (i—j) is a unit vector.)
Similarly, in three dimensions, if non-coplanar base vectors i, j, k
are chosen, any vector d may be split into three components in the i, j, k
directions
d = pi+qj+rk
and d may be represented in the alternative notation as
d = (q)
11
a= (
—i .
a3
Ex. 1. Rewrite 3i +4j— k and (i + j) — 2(j — 2k) in the column vector notation,
with i, j, k as base vectors.
Ex. 2. Express 3i— j—k as a column vector with (j+k), (k +1), (i+ j) as base
vectors. Where do you use the Uniqueness Theorem in your solution?
Ex. 3. Evaluate a+ b, 2a— 3b and 3(a— 2b) where
( 1 2)
a= 0), b= 1.
—1 —1
Ex. 4. If
2 ( 3)
x = (— y)
4
—1 —2
what are the values of x and y? Explain how you make your deductions.
2. COORDINATES
Suppose we now have some geometrical configuration in a plane which we
want to describe algebraically. Let us choose any point 0 of the plane as
origin; then all the points of the plane may be specified by their position
vectors with respect to 0. To give more detailed information, these position
vectors may be split into their components in the direction defined by the
right-handed pair of orthogonal unit vectors i, j.
Thus, if
r = OP
we may write r = xi +yj.
x and y are called, respectively, the x and y
coordinates of the point P and are usually
written as (x, y). The lines through 0 in the i
and j directions are called the x and y axes
respectively, and may be denoted by Ox and Oy. 0 •B
In Figure 3.1 the following points have been C .
plotted :
A(1, 3); B(3, — 1); C( 1, —2); D( 3, 0).
— —
Fig. 3.1
>- x
Fig. 3.2
From Figure 3.2 we have
P2P1 = P2Q +
= (x
1 —x2) i + 0,1—Y2) j.
Thus, applying Pythagoras's Theorem to the right-angled triangle P2P1Q,
P1 P2 = [(Xi X2)2+ (Y1 Y2)21
49
COORDINATES [3
For three-dimensional coordinates the same argument applies but
Pythagoras's Theorem must be applied twice. In Figure 3.3, the feet of the
perpendicular from P1, P2to the plane Oxy are Q1, Q2, and RQ1, Q2R
are respectively parallel to Oy and Ox. Thus we have
P2P1 = P2 Q
+
=+R(21+ QPI
= - + (yi
y2 - ) j+(z1 -z2) k
and so P1P2 = AiRxi - x2)2 + (Yi +(z1-z2)2]
on applying Pythagoras's Theorem to triangles P1P2 Q and Q1Q2R.
Fig. 3.3
Exercise 3(a)
1. Evaluate as single column vectors:
1 1
1
(i) (_3)— _3
(1); (ii) 2 ( 12) + 3 ( 10,•
—1 —2 )
{
(Y —Yi = A(Y2 —Y1).
Eliminating a between these two equations we obtain the following result.
The Cartesian equation of the line joining the points P1(x1, Yi)P2(x2,Y2) is
x—x1= y— yl
x2 —x1 Y2—Y1
(The reader must note carefully the distinction in this equation between
x, y on the one hand and x1, yl, x2, y2on the other: x, y are the coordinates
of a general point P on the line; x1, yi, x2, Y2 are the coordinates of two
specified fixed points of the line. If we take U as the set of all points
(x, y) in the plane, the equation (1) is the defining condition for the set of
points comprising the line; that is, the set
{(x, u: x — xl = x2 — xl
Y — Yi Y2 —Y1) .
The same remarks hold for the vector equation r = r, + A.(r, —r1), in which
r1, r2are the position vectors of two given points.)
Equation (1) may be rewritten in the form
Y2 Y1 (x
Y— = -
The number m is called the gradient of the line. The sign of the gradient
tells us which way the line is sloping (see Figures 3.4 and 3.5). In case (i),
x1 — x2and y,—y, have the same sign and m > 0; in case (ii) x, and
Y1 — Y2 have opposite signs and m < 0. In both cases, the positive value of
m is equal to tan a, where a is the acute angle made by the line with the
x axis. Equation (2) gives the form for the straight line when a point on the
line and its gradient are known.t
t The reader who has already met trigonometric ratios of obtuse angles will realize
that if the straight line AB meets the x-axis at P, then the gradient of AB is tan LxPA in
all cases, where LxP A is measured in the positive (anticlockwise) sense from the x axis
(see Chapter 6).
53
COORDINATES [3
ol
y
>x
0 X
Example 3. Find the equation of the line L, through (2, —1) with gradient
—1 and the equation of the line L2joining the points (0, 7) and (-1, 4).
What are the coordinates of the intersection of L, and L2?
Draw a sketch showing the relative positions of L, and L2and the coordi-
nate axes.
L, has equation y+ 1 = — -1-(x— 2),
i.e. x+2y = 0,
54
41 COORDINATE GEOMETRY IN A PLANE
x —0 _ y —7
L2has equation —1-0— 4 — 7'
i.e. 3x—y+7 = 0.
Since all points lying on L1satisfy the x + 2y = 0 and all points lying on
L2satisfy 3x —y + 7 = 0, the point of intersection of the two lines must
satisfy both equations simultaneously.
x+2y = 0,
Solving
l ax—y = —7
we have x = —2, y= 1
and so the point of intersection of L, and
L2is ( -2, 1).
(In this case the sketching of the lines L1
and L2is straightforward: we know two points
on L2, and L1clearly passes through the origin.
In general, to sketch the position of a straight
line whose equation is given, first find the
coordinates of the points where it cuts through
the coordinate axes.) Fig. 3.6
Ex. 17. Draw a sketch to show the positions of the straight lines:
(i) 3x-2y-6 = 0; (ii) 2x+y + 3 = 0; (iii) 2x-5y = 0.
Ex. 18. Find the equations of the lines through (2, —3) with gradients:
CO 2; (ii) — i.
Ex. 19. Find the equations of the lines joining the pairs of points:
(i) (2, — 3), (3, 2); (ii) (0, 1), (2, —1);
(iii) (-3, —1), (-1, 2); (iv) (1, 2), (-3, 2).
Ex. 20. What are the gradients of the following lines:
(i) 2x—y-3 = 0; (ii) x+y+1 = 0; (iii) 3x+4y+2 = 0?
Suppose now we have two lines, L1and L2, the gradients of which are
respectively m1and m2.
(i) If L1and L2are parallel, the angles that they make with the x axis
are equal and
m1= m2.
(ii) If L1and L2 are perpendicular then either they are parallel to the
coordinate axes or they have gradients of opposite sign: in the second case,
55
COORDINATES [3
Fig. 3.7
*Ex 21. With the notation above, prove that, if mi. = m2, L1and L2 are parallel.
*Ex. 22. Again with the notation above, prove that, if m1m2= —1, Lland L2 are
perpendicular.
Example 4. Find the equations of the lines L1, through (1, —1) parallel to
3x— y +7 = 0 and L2, through (2, —3) perpendicular to 2x+5y +1 = 0.
Find also the equation of the line L3joining the origin to the intersection
of L1and L2.
The gradient of 3x —y + 7 = 0 is +3 and so L1has equations
y+ 1 = 3(x-1),
i.e. 3x — y — 4 = 0.
The gradient of 2x + 5y + 1 = 0 is --I and so the gradient of a perpen-
dicular line is +I. Thus, the equation of L2 is
y +3 = i(x — 2),
i.e. 5x-2y-16 = 0.
(With practice, the reader will soon be able to derive such equations
more rapidly than is done here. For example, in the second case, any line
perpendicular to 2x + 5y +1 = 0 is clearly of the form 5x-2y = k and,
since the required line passes through (2, —3), k = 5 (2)— 2 (-3) = 16.)
Now consider the equation
(3x — y — 4) + A(5x — 2y — 16) = 0,
Ex. 23. Find the equations of the lines L1, through (4, —10) parallel to 3x — y = 0
and L2, through (1, 2) perpendicular to 2x —y — 1 = 0. What are the coordinates
of the intersection of L1and L2?
Ex. 24. Find the line joining the point ( — 1, 2) to the meet of the two lines
x+3y-1 = 0,x-4y+2 = 0.
Ex. 25. Find the equation of the line through the meet of the two lines
3x—y+1 = 0, 4x-3y+2 = 0
perpendicular to the line 3x —y— 1 = 0.
Ex. 26. Find the intercept cut off on the transversal 5x-2y+ 3 = 0 by the two
parallel lines 3x—y + 1 = 0 and 3x —y + 4 = 0.
Exercise 3 (b)
1. Find the equations of the lines through the stated points with the given
gradients:
(i) (0, — 2), 2; (ii) (-1, — 1), —2; (iii) (3, —1), — 1; (iv) (a, b), —a/b.
2. Find the equations of the lines joining the following pairs of points:
(i) (1, — 2), (2, 1); (ii) ( 2, 3), (— 1, 4); (iii) (2, — 1), (— 1, — 1);
- —
7. Find the orthocentre of the triangle whose vertices are (0, 1), (1, 2), (4, 3). (The
orthocentre is the meet of the altitudes of a triangle.)
8. The mid-points L, M, N of the sides, BC, CA, AB of a triangle ABC have
coordinates (2, 1), (3, — 3), (4, — 5). Find the coordinates of A, B, C.
9. The coordinates of the vertices of a triangle are (6, 0), (— 1, 1) and (5, —7).
Find the coordinates of the centre of the circumcircle.
10. The coordinates of the vertices of a triangle are (1, —4), (3, —2) and
(- 11, 12). Find the coordinates of the centre of the circumcircle and deter-
mine the coordinates of the points where this circle cuts the axes.
11. Show that, for all values of A, the line whose equation is Ax +y = 1-2A
always passes through the point (-2, 1). What is the equation (in terms of A)
of the perpendicular line through (1, 2)? Show that, whatever the value of A,
the intersection of these two lines always satisfies the equation
xa ±y2 +x-3y = 0.
What locus does this last equation represent?
12. OABC is a rectangle in which OA = 30C. M is the mid-point of OC and L
is the point of trisection of CB nearer C. OL, AM meet at X. By setting up axes
along OA and OC and assigning suitable coordinates to the various points,
determine the ratio in which X divides OL.
13. Prove that, for all values of A, the line
(1— A) x+ Ay = 3 — 7A
passes through a fixed point. What are the coordinates of the point?
14. Find the coordinates of the centre of the circumcircle of the triangle ABC,
where A, B, C have coordinates (— 1, — 3), (-2, —2), (5, 5) respectively.
Prove that the point P(6, 4) lies on the circumcircle and prove further that the
feet of the perpendiculars from P on to BC, CA, AB are collinear.
15. Find the equation of the line joining A(a, 0) and B(0, b). A', B' are the feet
of the perpendiculars from A, B to a variable line through the origin. If A'P, B'P
are respectively parallel to the y and x axes, what can be said about the position
of P?
16. What are the equations of the reflections of the line 2x— 3y+ 6 = 0 in
(i) the x axis; (ii) the y axis;
(iii) the line 2x — 3y = 0; (iv) the line 2x— 3y = 3 ?
We shall content ourselves for the moment with demonstrating the general
form of the Cartesian equations of lines and planes in space.
We may define a plane as a set of points in space which
(i) contains at least three non-collinear points;
(ii) has the property that, given any two points R, and R2 of the set,
all points of the line R1R2 belong to the set.
From this definition, three non-collinear points
P1(x1, 311., z].), P2 (x2, Y2, z2), P3 (x3, Y3, z3)
clearly define a plane. If r1, r2, r3are respectively the position vectors of
P1, P2, /33, we have seen that the plane P1P2P3is (see Chapter 2.6)
r = r, + A(r2 —r„)+p(r3 —r1).
In column vector form this may be written
(X1 X2 -XI X3 -
= Y1) A (Y2 Y1) + Y3 — Yl
Z
X) Z1 Z1 Z3 - Z1
The vector u defines the direction of the line and may be termed a
direction vector. Since u = ii+mj +nk, 1, m, n are proportional to the
cosines of the angles that u makes with the directions i, j, k, that is, with the
directions of the coordinate axes. 1, m, n are usually called direction
ratios for the line; if u is a unit vector, the constant of proportionality is
one and 4 m, n are called direction cosines (see Figure 3.8, where 1 = cos 01,
m = cos 02, n = cos 03).
Fig. 3.8
Example 5. Find the equation of the plane through A (1, 1, 1), B(0, 1, — 2),
C (0, 0, — 1) and the point of intersection of this plane with the line
x— 3 _ y — 2 _ z— 2
1 — —2 — 3 '
Let the equation of the required plane be
ax+by+cz+d = 0.
Since the coordinates of A, B, C all satisfy this equation,
a+b+c =—d,
b-2c = —d,
—c = —d
which gives a = —3d, b = d, c = d.
The equation of the plane ABC is thus
3x — y — z —1 = 0.
Any point of the given line has coordinates
(3 +A, 2-2A, 2 + 3A).
This lies on the plane if
3(3 + A) — (2— 2A)— (2 + 3A)— 1 = 0
giving A = —2
and so the point of intersection has coordinates
(1, 6, —4).
3-2 61
COORDINATES [3
tion (— 1, 1,
Eliminating A (a) between the x and y coordinates of a general point of
L1, and (b) between the y and z coordinates, we see that the two planes
x+4y = 5, y—z = 0
both contain L1. It follows that, for all values of k, the equation
(x + 4y — 5)+k(y—z) = 0 (1)
represents a plane containing L1.
By setting # = 0, we see that the point (0, 1, 2) lies on L2(but not on
L1). This lies on the plane (1) if
(0 +4 — 5) + k(1 — 2) = 0,
i.e. k = —1.
The equation of the plane containing L1and L2 is thus seen to be
x + 4y — 5 — 1(y — z) = 0,
i.e. x+3y+z-5 = 0.
Exercise 3 (c)
1. Find (a) direction ratios, (b) the direction cosines, for the lines joining the
following pairs of points:
(i) (1, 2, 3), (2, 3, 4); (ii) (2, -1, 3), (1, 1, 2);
(iii) (3, 1, 2), (5, -1, 1); (iv) (1, 3, 5), (5, 3, 1);
(v) (2, -3, - 4), ( -3, 2, 1);
(vi) (a+ A, 2a + 2A, 3a + 3A), (a- it, 2a - 211, 3a - 3,a);
(vii) (a2, ab, b2), (a2, ac, c2); (viii) (1, a, a2), (a2, a, 1).
2. Find the equations of the planes through the following sets of points:
(i) (1, 0, 0), (0, 1, 0), (0, 0, 1);
(ii) (3, 2, 0), (0, 3, -1), (1, 0, -2);
(iii) (2, 0, 0), (1, 1, -1), (6, -5, 3);
(iv) (2, 2, 0), (-1, 1, -4), (1, 1, -1);
(v) (-1, 3, 1), (1, -3, -3), (3, -1, 5);
(vi) (1, 2, 3), (2, -2, 8), (-1, 2, -7).
3. Find the equations of the lines joining the following pairs of points:
(i) (0, 0, 0), (2, 1, -3); (ii) (1, 2, -1), (2, 3, -3);
(iii) (1, 1, -1), ( -1, 4, 2); (iv) (-2, 1, 3), (2, 1, -3);
(v) (1, 4, 2), (1, -1, 2); (vi) (a, 2a, 3a), ( -a, a, 2a).
4. Find the equation of the plane through the line of intersection of the planes
x+y -3z = 2, 2x-y-z =1
(i) containing the origin;
(ii) containing the point (1, 1, -1);
(iii) parallel to the x axis.
5. Find the equation of the plane through the line of intersection of the planes
2x+2y+3z-1 = 0, 3x-y-z+2 = 0
(i) containing the origin;
(ii) containing the point (1, 1, 1);
(iii) parallel to the z axis.
63
COORDINATES [3
6. Write down the equation of the plane parallel to the x axis which contains
the line
x-1 y-3 =z+1
=
—1 2 —3 •
Find the equation of the plane containing this line and the origin.
7. Find the equation of the plane through the point (1, 1, — 5) containing the
line
x+1 _ y+5 _ z
3 — 4 — —5'
8. Write down, in terms of a parameter A, the coordinates of a general point of
the line
x-8 = y 1 = z-1
—
3 1 0•
Where does this line cut the plane x -2y — 3z— 2 = 0?
9. Find the coordinates of the point where the line joining (2, 3, 1) and (4, 7, 3)
cuts the plane 2x +y—z— 3 = 0.
10. Find the equations of the line joining (2, 1, 1) and (1, — 1, 2). Where does this
line meet the plane x — y +az = 0?
Is your result true for all a? Explain.
11. Determine the direction cosines of the line of intersection of the planes
2x—y+z— 9 = 0, 4x+y+2z — 6 = 0.
Where does this line meet the plane x+y+z = 0?
12. Prove that the two lines
x-1y+1 z+2
= =
1 2 3 '
x+1 y z-1
= =
1 1 1 '
are skew.
13. Prove that the two lines
x-2 y+4 z -1
= =
3 1 —1
x+1 y+4 z+2
and = =
2 1 —2
meet. What are the coordinates of their point of intersection? Find the equation
of the plane containing them.
14. Prove that the two lines
x -3 y— 2 z-4 4
= =
2 1 1
x +1 y -1 z
and = =
2 0 3
meet. What are the coordinates of their point of intersection? Find the equation
of the plane containing them.
64
5] COORDINATE GEOMETRY IN SPACE
15. Find the coordinates of the point common to the three planes:
x— 3y+ z— 6 = 0,
2x— y + 2z— 2 = 0,
3x+2y+z+2 = 0.
16. Find the coordinates of the point common to the three planes:
x — 2y + z — 7 = 0,
2x + 3y — 4z + 26 = 0,
3x+y+3z+7 = 0.
17. Find the coordinates of the point where the line determined by the two planes
x — 3y— z + 8 = 0,
x — y+ z— 2 = 0
cuts the plane containing the x and y axes.
Miscellaneous Exercise 3
1. Find the area of the trapezium whose vertices are:
(x2, Y2), (x3, Y3), (x2, 0), (x3, 0).
Deduce that the area of the triangle P1(x1, P2(x2, Y2), P3(x3, Y3) is
Mx2Y3 —x3Y2+ x3Y1 —x1Y3+ x1Y2 —x2Y1)•
What is the condition that the three points P1P2P3should be collinear?
2. ABCD is a rectangle with AB = 2AD. E is the mid-point of AB and F is
the mid-point of BC. CE, AF meet at X. By setting up a suitable coordinate
system, prove that DXB is a straight line and determine the ratio DX/XB.
What is the area of the quadrilateral AXCD?
3. ABCD is a rectangle and points P, Q are taken on AB, AD respectively.
The rectangle APRQ is completed. If the lines BQ, DP meet at X, prove that
X, R, C lie on a straight line.
4. Find, in terms of a parameter m, representing the gradient, the equation of
a variable straight line through the fixed point (a, b).
If this line cuts the x and y axes at A and B respectively, and the parallelogram
OABP is completed, prove that, whatever the value of m, the point P lies on the
curve
xy = xb — ay.
Sketch this curve for the case a = b = 1 for values of x > — 1.
5. Show that, by a suitable choice of coordinate system, the equations of two
coplanar non-parallel lines may be taken in the form
Y = ax, y = — ax.
65
COORDINATES [3
Two fixed lines OA, OB are drawn and a variable line, passing through the
fixed point C, cuts OA, OB at P and Q respectively. If the feet of the per-
pendiculars from P, Q to OB, OA are U, V, prove that UV passes through a fixed
point.
6. Two fixed straight lines, L1and L2, meet at 0. Through a fixed point A two
lines AP1P2and AQ1Q2are drawn to cut L1at P1and Q1, and L2 at P2 and Q2.
Prove that whatever the position of the two lines drawn through A, the point of
intersection of P1Q2 and P&Llies on a fixed straight line through 0.
7. Points P1, P2, P3 are taken on the x axis, and Q1, Q2, Q3 on the y axis. L1
is the point of intersection of P2 Q3 and P3 Q2, L2 of P3 Q1and P1 Q3, L3 of P1 Q2
and P2 Q1. Prove that L1, L2, L3 lie on a straight line.
(This is a particular case of Pappus's Theorem, which holds more generally
for two sets of three collinear points on any two straight lines.)
8. Prove that the lines x+1 y-1 z-3
—2 2 1
x—1 y— 3 z— 2
and
3 1 6
are skew.
By considering the family of planes through one of the lines, prove that
there is just one common transversal to the two lines which passes through the
origin and find its equation.
9. Prove that no three of the points (1, 1, 2), (-2, — 6, 3), (— 1, 1, 5), (2, 4, 2) are
collinear but that all four points lie on a plane.
10. Explain why the three planes
x— z +1 = 0,
x — y —z+5 = 0,
x+2y—z+2 = 0
have no common point.
What can you say about the intersections of the planes:
x—z+1 =0,
x—y—z+5 = 0,
x+2y— z-7 = 0?
11. ABCDA'B'C'D' is a cube, with square faces ABCD, A'B'C'D' and vertical
edges AA' etc. M is the mid-point of C'D'. The plane AB'M cuts BD' at X
and CC' produced at Y. Find the ratios BX: XD' and CC': C' Y.
12. Prove that the three lines . x _ y _ z
——
.x-1 _y _z
. ——
+1 y z
L3. 3 1—2
are skew.
66
5] MISCELLANEOUS EXERCISE 3
67
4. Polynomials
1. POLYNOMIALS
Expression of the form
xs - 3x2+ 4x + 2
and x5- x + 1
are called polynomials. More generally, a polynomial of degree n is an
expression of the form
ao xn + ai xn-+ a2 xn-2+ an (ac, + 0)
involving only multiples of positive integral powers of x and a constant
term, an. A polynomial of degree n is completely determined if its (n +1) co-
efficients (including the constant term or coefficient of x°) are given and
two polynomials of degree n are said to be identically equal if the coefficient
of xr in each polynomial is the same for all values of r, and conversely.
For example, writing = to mean 'identically equal to'
ax3 -x -2 -E 2x8 -bx2 -x + c
-4=> a = 2, b = 0, c = - 2.
Ex. 2. If P(x) x4 -3x3+ 4x2 - x-1, show that P(1) = 0 and find P(- 1), P(0)
and P(2).
Ex. 3. Show that, if P(x) Q(x), then P(k) = Q(k) for all values of k. The
converse of this proposition is also true: if P(k) = Q(k) for all values of k,
then the coefficients of xr in P(x) and Q(x) are the same for all r (see Ex. 7).
Either statement may be taken as the definition of 'identically equal to'.
68
1] POLYNOMIALS
Exercise 4(a)
1. If P(x) x3+5x2+ 3x + 1; find P(1), P(— 1), P(1), P(3).
2. If (x- 2) (x— 3) (x-4) -.7- ax3+ bx2+ cx+ d, find a, b, c, d.
- x3— 3ax2+ 3a2x — a3. For what real values of x does
3. Prove that (x — a)3 a
(x — a)3= — a3?
4. If a(x— 2)2 + b(x— 2)+ c 3x2 -8x-1, find a, b, c.
5. Express 7x3 — x2 + 3x-4 in the form a(x — 1)3+ b(x — 1)2+ c(x-1)+ d.
6. Express 4x3 +12x2 + 6x in the form a(x +1)3+ b(x +1)2+ c(x +1)+ d.
7. Express 3x3 +2x2— 11x— 10 in the form a(x +1)3+ b(x +1)2+ c(x + 1)+d and
also in the form cc(x-2)3 + fi(x— 2)2+y(x— 2) + a.
Find the three roots of the cubic equation 3x3 +2x2 — 11x— 10 = 0.
8. Express x3+ 4kx2+ 3k2x— k3in the form a(x+ k)3+ b(x+ k)2+c(x+ k)+ d.
9. If (ax2+ bx+ c) (x +1) E- 0, prove that ax2+bx+c -m 0.
f We shall not associate a degree with the zero polynomial. The reader will appreciate
that, by elementary algebra, the product of a polynomial of degree m by a polynomial
of degree n is a polynomial of degree m+n. This result holds good even if one or both
of the polynomials is a constant polynomial, but fails if one of the polynomials is the
zero polynomial if a finite degree is associated with this polynomial.
69
POLYNOMIALS [4
{a, b, c} =
and so the cubic polynomial 2x3+ 5x2 — 3x — 10 has a linear factor (x + 2).
(It also has a quadratic factor (2x2 +x — 5))
It is important to realise at the outset that possession of a factor is not
an absolute property of a polynomial but depends on the restrictions we
place upon the coefficients. For example, if we restrict our polynomials
to have only rational numbers as coefficients, the polynomial x2— 2
has no linear factors; on the other hand, if we allow our coefficients to be
real numbers, x2 -2 = (x— V2) (x + V2). (This is often expressed by the
statement: 'x2 -2 is irreducible over the rational field but reducible over
the real field'. The word 'field' has a technical meaning and will be defined
in Volume 2; all that is necessary at the moment is to realize that over
the rational field' means that all polynomials under discussion must have
rational numbers as coefficients.)
The theorems we are about to prove do not depend upon whether we
choose the rational field or the real field for our coefficients provided, of
course, that we are consistent. We shall not, therefore, allude to the field
from which the coefficients are drawn, but the reader may, if he wishes
interpret all the coefficients as, say, rational numbers.
It is easy to see that, if P(x) has a factor (x a), then P(a) = 0. For
P(x) has a factor (x — a)
P(x) (x a) Q(x)
=- P(a) = 0 Q(a)
P(a) = O.
A converse of this result is known as the Factor Theorem.
70
2] FACTOR AND REMAINDER THEOREMS
Example 2. Find the remainder when x3— 5x2— x + 2 is divided by (i) x-1;
(ii) x+2; (iii) 2x+ 1.
Write
P(x) _= x3— 5)0— x + 2.
(i) P(x) (x —1) Qi(x)+ R1;
put x = 1: R1= P(1) = 1-5-1+2 —3;
Exercise 4(b)
1. Find the remainder when:
(i) x3— x + 2 is divided by x-4;
(ii) 3x3— 5x2+ x + 2 is divided by x — 2;
(iii) 4x4—2x2 + x — 3 is divided by x + 1;
(iv) x5— x-1 is divided by x+ 3 ;
(v) 4x3— 5x2+2 is divided by 2x — 1;
(vi) 4x3— 8x + 1 is divided by 2x + 1.
2. Show that x — 2 is a factor of the polynomial x3— 4x2 + x+ 6 and hence factor-
ize the expression completely.
72
2] FACTOR AND REMAINDER THEOREMS
3. Factorize the following polynomials over the rational field as far as possible:
(i) 2x3 + 7x2 — 5x— 4; (ii) 12x3 + 5x2— 19x — 12;
(iii) 2x3 +7x2 — 17x — 10; (iv) x3 + 3x2— 2x — 6;
(v) x3 — x2 — x— 2; (vi) x3 — 7x2 +7x +15;
(vii) x3 + 2x2 — 7x— 2; (viii) 8x3 +12x2 — 2x — 3.
4. Factorize the polynomials given in Question 3 over the real field as far as
possible.
5. If x3 — 5x2+7x— a has a factor x — 2, find a.
6. If 2x3 +ax2 — 5x-1 is divisible by 2x + 1, find a.
7. If x3 + 3x2 + ax— 1 leaves a remainder of 3 on division by (2x +1), find a.
8. If 3x3 + ax2+ bx— 2 is divisible by both x+2 and 3x +1, find a and b.
9. If ax3+ 3x2+ bx — 3 is divisible by both x-1 and 2x+ 3, find a and b.
10. By expressing x4 + 1 in the form x4 + 2x2 + 1 — 2x2, show that x4 + 1 is re-
ducible over the real field.
11. Factorize x4 + 3x3 — 1 5x2 +9x+ 2: (i) over the rational field; (ii) over the real
field.
12. ax4+ 2x3— 4x2—2x+ b has factors (x— 1) and (x— 2), find a and b and fac-
torize the expression completely.
Theorem 4.3. If P(x) is a cubic polynomial with leading term ao x3, and
if P(a) = = P(y) = 0 for 3 unequal numbers a, 13, y, then
P(x) = ao(x ct) (x-13) —
Proof. Since P(a) = 0, P(x) (x — a) Q1(x), where Q1(x) is a polynomial
of degree 2.
Thus, since PO = 0, 0 = (fi— a) Q,(13). But 161 — + 0 and so Q103) = 0.
Q1(x) therefore has a factor (x—fl).
Q1(x) (x — ie) Q2(x), where Q2(x) is a polynomial of degree 1.
By the same argument,
Q2(x) (x—Y) Q3(x), where Q3(x) is a polynomial of degree 0, that is,
Q3(x) = k, where k is some number. Substituting,
P(x) = k(x — a) (x— fi)(x— y).
73
POLYNOMIALS [4
can be found such that P(x) = Q(c<), ..., P(8) = Q(8), prove that P(x) = Q(x).
Exercise 4(c)
1. Prove that
(x — b) (x — c) (x — c) (x — a) (x — a) (x — b)
I O.
(a — b) (a — + (b — c) (b — a) (c — a) (c — b)
74
31 THE FACTOR THEOREM
2. Prove that
(x— b) (x — c) (x— d)+(x— c) (x— d)(x— a) +(x — d)(x — a) (x — b)
(a — b) (a — c) (a — d) (b— c) (b — d) (b — a) (c — d)(c — a) (c — b)
(x— a) (x— b) (x— c)
+ O.
(d— a) (d — b) (d— c) 1 .=
3. Prove that
a(x — b) (x— c) b(x— c) (x— a) c(x— a) (x— b)
(a — b) (a — c) (b — c) (b — a) (c — a) (c — b) x O.
4. Prove that
x(x — a) .4_ x(x — b) x(x— c) x
=
bc(c — a) (a — b) ca(a — b) (b — c) ab(b — c) (c — a) abc 0.
5. By considering the quadratic polynomial axe + bx + c, show that, if n * 0,
n * 1 and
a+b+c = 0,
an2+bn+c = 0,
a+bn+cn2 = 0,
then a = b = c = 0.
6. Prove that the cubic polynomial P(x) defined by
a(x— fl) (x—y) (x -8) b(x— a) (x— y) (x-8) c(x— a) (x— fl) (x— (1)
P(x) =
(f3)()() (fi — a) (fi — (fl — — ce) — /3) (Y —
d(x— a) (x—/3) (x — y)
(a — ce) fl) (8— 1')
is the unique cubic polynomial such that P(a) = a, P(/?) = b, P(y) = c, P(8) = d.
Find the cubic polynomial P(x) such that
P(— 1) = — 3, P(0) = 2, P(1) = 5, P(2) = 12.
7. Find the cubic polynomial P(x) such that
P(0) = 1, P(1) = —1, P(2) = —1, P(3) = 13.
8. Find the cubic polynomial P(x) which takes the value — 5 when x has the
values —1, 0, 1 and is such that P(2) = 1.
9. State and prove the identity theorem for quartic polynomials.
Miscellaneous Exercise 4
1. We shall say that a polynomial P(x) 'has the property R' if the remainder on
division of P(x) by the fixed factor x— a is R. Prove that
(i) if P1(x) and P2(x) both have the property R, so also has the polynomial
P(x) where P(x) = P1(x) + P2(a);
(ii) if P3(x) has the property R and P4(x) has the property S, show that the
polynomial Q(x) E P3(x) P4(x) has the property RS.
What can be said about the polynomial P(x) if it has the property 0?
75
POLYNOMIALS [4
2. Prove that, if P(x) is any polynomial of degree n 2, then P(x) may be ex-
pressed in the form
P(x) (x – a) (x – b) Q(x) + Rx + S,
where R, S are numbers and Q(x) is a polynomial of degree n- 2 (all numbers
and coefficients taken from the same field—say the rationals).
Express 5x3+ 9x2 – 8x-11 in the form
(x –1) (x + 2) Q(x)+ Rx + S
and show you can check your values for R and S by putting x = 1 and x = –2.
3. Find the remainder on dividing the following polynomials by x2 – x– 6:
(i) 7x3– x2 – 30x– 30; (ii) x8 – 1.
4. The remainder when the polynomial P(x) is divided by (x – a) (x – b), where
a * b, is expressed in the form A(x – a) + B(x – b). Find explicit values for A
and B.
Prove that, if identically equal remainders are obtained on division by
(x– a) (x– b) and (x– a) (x– c),
then (b – c) P(a)+ (c – a) P(b)+ (a – b) P(c) = 0.
What can you deduce about the remainder when P(x) is divided by
(x – b) (x – c)?
5. By using the Remainder Theorem or otherwise, find the factors of
a2(b + c) + b2(c + a) + c2(a + b) + 2abc.
Hence or otherwise show that, if a+ b + c = 0, then
a3+ b3 + c3 = 3abc.
Find (i) the factors of
(p + 2q –303+ (– 3p + q + 203+ (2p – 3q +
(ii) the prime factors of
733– 473 – 263. (Cambridge)
6. If axe + 2bx + c can be written in the form
A(x – a)2 +B(x–fi)2,
prove that aafl+b(a+fl) +c = 0.
Is it true that, if numbers a, /3 can be found such that
aaft+b(a+fi)+c = 0
then ax2+ 2bx + c can be written in the form A(x – a)2 +B(x–fi)2 ?
7. Prove that, if lx2+ mx + n is equal to zero for three distinct values of x,
then / m = n = 0.
Hence, or otherwise, prove the identity
a2(x – b) (x – c) +b2(x– c) (x – a) + c2(x – a) (x – b) _ 2
(a– b) (a– c) (b – c) (b – a) (c – a) (c – b) x•
76
3] MISCELLANEOUS EXERCISE 4
Prove also that
(x+a)3 (x+ b)3 (x + c)3
+ + 3x+a+b+c.
(a — b) (a — c) (b — c) (b — a) (c — a) (c — b)
(0 & C)
8. Factorize:
(i) bc(b — c)+ ca(c — a)+ ab(a— b);
(ii) a(b — c)2+ b(c— a)2+ c(a— b)2+8abc;
(iii) a(b — c)3+ b(c — a)3+ c(a — b)3.
9. Given that p, q, r are distinct values of x which satisfy the equation
a
+ + =1
x—a x—/3 x-7
prove that, for all values of x other than x = a, x = x=y
a b c (x — p) (x — q) (x — r)
1
x—a+x—/3+x - y (x— a) (x—/3) (x-7)•
By considering the value of P(x) e (x — p) (x — q) (x — r) for different values of x,
prove that, if a, b, c are all positive and a < fl < y, then p, q, r lie one in each of
the sets
{x e R: a < x < 13}, {x e R; fl < x < y}, {x e R: x > y}.
Prove also that
(p — a) (q — a) (r — a)
a=
— /6) (a —
and that
a (p — q)(p — r)
+ +
— c02 — fir /3) (P — Y)
(0 & C adapted)
10. When divided by (x2 + x+ 1), a cubic polynomial F(x) leaves a remainder
(2x+3). When F(x) is divided by x(x + 3), the remainder is 5(x + 1). Find F(x).
(Cambridge)
11. f(x), g(x) and G(x) are cubic polynomials in x. The polynomials h(x) and
H(x) are respectively the remainders left when f(x) is divided by g(x) and by
G(x). Prove that, if ax3 and Ax3are respectively the terms of highest degree in
g(x) and G(x), then
[aG(x) — Ag(x)] f(x) E aG(x) h(x) — Ag(x) H(x). (Cambridge)
77
5, Functions and inequalities
1. FUNCTIONS
The concept of an association between the elements of two sets is of prime
importance in mathematics. For example, the table of sines in a book of
mathematical tables exhibits an association between the set A of angles
between 0° and 90° and the set B of real numbers between 0 and 1. We
could exhibit the fact that sin 30° = 0.5 by writing the ordered pair
(30°, 0.5). The word `ordered' used here reminds us that the order in
which we write the elements down within each pair must be taken into
account; that is, the ordered pair (a, b) is not the same as the ordered pair
(b, a).
We can put this concept of an association between the elements of two
sets in a more precise way if we introduce the idea of the Cartesian product
of two sets A and B. The Cartesian product of A and B, written A x B is
the set of all ordered pairs (a, b) where a E A and b E B. For example, if
A = {1, 2, 3}, B = {3, 4, 5},
then
A x B = {(1, 3), (1, 4), (1, 5), (2, 3), (2, 4), (2, 5), (3, 3), (3, 4), (3, 5)1.
Ex. 1. If the set A has m elements and the set B has n elements, how many
elements has the set A x B?
Fig. 5.1
If we have a function f: B the set A is called the domain off and the
set of image points (a subset of B in general, but possibly the whole of B)
is called the range of f. The set B is called the codomain of f. The image
of the element a e A is written f(a), thus, f(a) is an element of B or, in
other words, (a, f(a)) is an element of A x B.
Thus, since 1/(1+ x) can take any positive value less than 1 by a suitable
choice of positive x, the range off is the set 0/ e R+: 0 < y < 11.
(ii) f(2) = 1/(1+ 2) = 1;f(N/2) = 1/(1+ V2) = 1. (Notice that both
2 and N/2 belong to R+ and so f(2), f(V2) are defined.)
(iii) If 1/(1+x) = 1, x = 3.
number. Even so, the phraseology is loose: the notation f(x) (or y) stands
for the image of x under f, not for the function itself.
by g(x) = x3 3x2 5x + 2.
- -
(0,4)
Ex. 5. Explain why the relation depicted in Figure 5.3 is not a function.
Ex. 6. Show that no inverse function exists for the function f: R R where
f(x) = x2, but that an inverse function does exist for the function g: R+ R+,
whereg(x) = x2.
Exercise 5(a)
1. The function f:R—> R is defined by f(x) —x-1. Find f(0), f(1), f(2),
f(3),f(-1).
2. The function g: R+ ---> R is defined by
x2 +x+1
g(x)
x.i I
Find g(1), g(2), g(3).
3. The function f: R R is defined by
(— x if x < 0,
f(x) = 0 if x = 0,
x if x > 0.
(This is normally written f(x) = IxI; read f(x) equals the modulus of x'.) What is
the range of f? Sketch the graph of f.
A second function g: R --> R is defined by g(x) = x— Ix 1. What is the range
of g? Sketch the graph of g.
81
FUNCTIONS AND INEQUALITIES [5
4. The function f: R -> R is defined by
x
f(x) = trx/ , x * 0,
0, x = O.
Write down the values of f(- 1) and f(1). Sketch the graph of f.
5. A g R and B g R. The set f of all ordered pairs (x, y) is formed, where
x e A, y e B and y = 4,/x. Show that f is not a function if
(i) A = R+, B = Q; (ii) A = Q, B = R.
Suggest sets A and B for which f is a function.
6. A = {x e R: 1 --5 x ...
. 2} and f: A -> R is defined by f(x) = 2/x. The function
g: A -)- R is defined by g(y) = 1 +y2. A new function h: A -> R is defined by
h(x) = g{f(x)] (this is usually written h = g o f). Find
(i) h(); (ii) the range of h.
7. Sketch the graph of the function!: R -+ R where f(x) = lx + 11.
8. Sketch the graph of the function g: R -÷ R where
g(x) = lx+11+1x+21+1x+31.
9. A = {xe R: 0 < x < 3} and f: A -* R is defined by f(x) = x2 - 3x+2. Find
(i) !(1),.f(2);
(ii) the range of f;
(iii) the subset of A whose elements have the image 1 under f;
(iv) the subset of A whose elements have the image 2 under f.
10. A = {x e R: - V6 -.<... x ‘. V6} and f: A -> R is defined by f(x) = x3- 6x.
Find the range and sketch the graph of f.
11. If f: R R is defined by f(x) = (x- 1)s, show that f-1: R -> R exists and
determine its form explicitly.
12. Determine the numbers which are invariantt under the mapping f: R -+ R
where f is defined by
{(x2 +x+1) (x+3)-1 (x * -3),
f(x) =
0 (x = - 3).
Show how to illustrate invariance graphically for a general function g: R -+ R.
13. f: R R is a given function, A a Rand the set of images of all the elements
in A is denoted by B. If C is the set of numbers whose images belong to B,
prove that A c C and that the inclusion may not be strict.
14. f: R -)- R is a given function, X c R and the set of images of all the elements
in X is denoted by f(X). If A c R, B c R is it necessarily true that
(i) f(A n B) = f(A) n f(B); (ii) f(A U B) = f(A) U f(B)?
15. Sketch the graph of the function f: R+ R defined by f(x) = x2. Show that
the function g: R+ -> R, where g(x) = xi-, is the inverse off. Sketch the graph of g.
What connection exists between the graphs off and g?
t An element a is invariant under the mapping f if f (a) = a.
82
11 FUNCTIONS
16. If x is a real number, [x] denotes the greatest integer less than or equal to x.
(For example, [- 2-1-] = - 3, [4k] = 4, [- 1] - 1.) The function f: R -› Z is
defined by f(x) = [x]; sketch its graph.
A functionfis called periodic if there exists a number k such that f(x+ k) = f(x),
for all x in the domain; k is then called the period of f. Show that the function
defined by f(x) = x-[x] is periodic and find its period.
What feature does the graph of any periodic function possess?
2. INEQUALITIES
The function, f, defined by
1
fix) 3 - 2x
has for its domain the set R with the single point x = i deleted; if we call
this set D, we may write D = R-{3}. The inequality
1
<1 (1)
3 -2x
defines a set A c D, where
1
A = {x D: < 1};
3 2x
that is, A is the subset of elements of D which map into elements less
than 1. By `the solution' of the inequality (1) is meant the determination
of the simplest expression for the set A.
It is always good practice, in solving an inequality, to express one side
of the inequality as zero. Thus
1 2(x- 1)
A = {x E D: 1 <0} {x e D: 3-2x <
3 -2x
To determine the values of x satisfying the inequality
2(x-1) < 0
3 -2x
it is best to express the working in tabular form (see Chapter 1.7). The
critical values of x E R are x = 1 and x = Z.
83
FUNCTIONS AND INEQUALITIES [5
Fig. 5.5
(x 3) (x+ 1)
-
(x-3) (x+ 1)
has as its domain the set D where
D=R { — - 1, 3};
that is, the set of real numbers, less the two numbers —1 and 3. (D may
be written alternatively as the complement of the set
{x E R: x = — 1 or x = 3}.)
84
2] INEQUALITIES
2 5x +2x2— 4x— 6
Now
(x-3) (x+1)± 5 — 5(x-3) (x+1)
2x2 +x— 6
—5(x-3)(x+1)
(2x — 3) (x + 2)
- 5(x-3) (x+1)'
We thus seek a simpler description of the set A, where
(2x-3) (x +2)
A = c D: 0).
5(x-3) (x+1)
The critical values of x e R are x = —2, x = —1, x = 2, x = 3.
As above, we set out the work in tabular farm.
3
D x<-2 x=-2 —2<x<-1 —1<x<1 x=-2. 2- <x<3 x> 3
— 0 + + + + +
— — — — 0 + +
— — — — — — +
x+1 — — — + + + +
(x+2)(2x-3)
+ 0 — + 0 — +
5(x-3)(x±1)
Fig. 5.6
85
FUNCTIONS AND INEQUALITIES [5
Ex. 7. Verify the solution of Example 2 by evaluating f(- 3), f(- 11), f(0), f(2),
f(4). Explain why these are appropriate values to use for checking.
*Ex. 8. Show that the two expressions
(2x-3) (x +2)
and (x-3) (x- (x +1) (x+ 2)
(x - 3) (x + 1)
have the same sign for all values of x -1 or 3. Hence solve the inequality of
Example 2.
Exercise 5(b)
Solve the following inequalities, illustrating your solutions by means of a sketch.
1 3- x 1
1. - < 2; 2. < 2; 3. < 1.
x x 1-x '
x 2
4. < 2- 5. ..>.- 1; 6.2 3x> -1;
1+x ' 3- x '
2
7. > x; 8. -
-.C. x; 9. < •
x-1 x- 3 x+1 3'
2 x- 5x+7
10. < 1. 4 < 2; 12. 1,
x2 - 5x+ 6 ' 11. x2-1 x2+5x+6 >
x-4 6(3x + 1)
13. < 1. 14. <1•
x2 4
- ' (x-1) (x- 2) (x+ 3) '
x- 3 1
15. < 1;• 16. < - 1;
x2+ x-3 x2 -3
1 1
17. -1 <2- x< 2. 18. 0 < < 3.
2x- 5
3 1
19. < 1; 20. < 1;
2-3x x2 -1
x2 -3x+3 I x2 -21
21. 1; 22. < 1.
x2 - 4x + 3 I x2 - 7x+ 6
86
6. The trigonometric functions
1. RADIANS
We shall assume in this chapter that the reader is familiar with the usual
measurement of angles in degrees and also with the definitions of the sines,
tangents, etc., of acute angles in terms of the sides of a right-angled tri-
angle. We shall also assume that arc-lengths in circles may be satisfactorily
defined and measured and, in particular, that the circumference and area of
a circle of radius r are respectively 27-rr and nr 2.
An alternative measure of angle may be defined as
follows. Figure 6.1 shows a circle of radius r, centre 0.
The magnitude of the angle LAOP may be defined as
the ratio of the arc-length, s, of AP to the radius, r.
In the particular case in which s = r, the angle LAOP
is called a radian. Thus, an arc-length of 2r subtends an Fig. 6.1
angle of 2 radians at the centre of the circle. (We may
write this LAOP = 2 rad', or, since a ratio of two lengths is a pure
number, simply LAOP = 2' provided no confusion is likely to arise.)
*Ex. 1. Show that a right-angle is Pr rad. Convert 180° to radians and use your
result to express in radians the following angles: 30°, 60°, 45°, 75°, 135°, 225°,
330°.
*Ex. 2. Express in degrees the following angles: i rad, in rad, in rad, 1r rad,
lig rad, in rad, Air rad.
(i) P (ii)
Fig. 6.2
We shall adopt the usual convention that angles are measured as positive
in the anticlockwise sense, the sense being denoted, when necessary, by an
arrow. Figure 6.2(i) shows an angle of — Pr, Figure 6.2 (ii) an angle of
31_ both measured from OA.
"
*Ex. 3. Prove that the range of both the tangent and cotangent function is the
set R.
*Ex. 4. Prove that the range of the secant and cosecant function is the complement
of the set {x E R: —1 < x < 1}.
The following table of the images of the sine, cosine and tangent func-
tions for some commonly occurring angles is worth memorizing:
Also the acute angle which defines the same numerical values of x and y
is 715-77" (see Figure 6.4(i)).
Thus, sin irr = + sin =
Fig. 6.4
sines (and cosecants) are positive; in the T quadrant, only tangents (and
cotangents) are positive; in the C quadrant, only cosines (and secants) are
positive. Notice, as an aid to remembering the diagram, that the letters
(read anticlockwise) spell the word 'CAST'.
Fig. 6.6
Suppose OP is the unit vector defined by the angle 0 and OQ the unit
vector defined by the angle -17r — 0 (see Figure 6.6). If OP = xi +yj, then
OQ = yi+xj. It follows that
sin (17r — 0) a- cos 0, cos (-1-7r — 0) E sin 0, tan (-177.— 0) cot 0,
cosec — 0) sec 0, sec (-3:77-- 19) cosec 0, cot (Pr — 0) tan 0
92
2] THE TRIGONOMETRIC FUNCTIONS
(provided, of course, that 0 lies in domains for which both sides are de-
fined).
*Ex. 5. Find similar simplifications for the values of the trigonometric functions
for the angles
(i) -1-7/- + 0; (ii) Tr 0; (iii) n +0.
—
Exercise 6(b)
1. Write down the values of cos 47r, tan lir, cosec lir, cot PT, sin P
' T, sec iv,
cos fa, tan in, cosec (—in), sec (-177).
2. Write down the values of cos 315°, tan 135°, cosec 330°, sin (— 135°),
cot (-120°), sec 240°, sin 480°, sec (— 210°), cot (— 60°), sin 1020°.
3. Use your tables to find the values of sin 215°, cos 128°, tan (— 40°), cosec 161°,
sin (-200°).
4. Solve the following equations for 0, giving all values lying in the interval
0 27r:
(i) sin 0 = —1; (ii) tan 0 +1 = 0; (iii) sec 0 +2 = 0;
(iv) cot 0 = V3; (v) 4 sin2 0 = 3; (vi) 2 cos (0 — 170+ ,/3 = 0;
(vii) sin 0+ cos 0 = 0; (viii) 2 sin 30 = 1; (ix) sec (0 + Pr) = 1;
(x) 3 sect + = 4; (xi) cosec2(0 + Pr) = 1; (xii) tang 20 = 3.
5. Solve the following equations for 0, giving all values lying in the interval
— 7T < 0 <
(i) 2 sin2 0+sin 0 = 0; (ii) 2 cos2 0+3 sin 0 = 0;
(iii) cos 20 = sin 0; (iv) 2 tan 0 + sin 0 = 0;
(v) sin (iir— 0)+ cos 0 = 0; (vi) cos 0 = 2 cot 0;
(vii) tan 0 + cot air— = 0; (viii) tan 0+ cot 0+sec 0 = 0;
(ix) cote 0+ cosec 0 +1 = 0; (x) sec 30+ cosec 0 = 0.
6. Find the maximum and minimum values of:
(i) 3/(2 + sin 0); (ii) (1 + cos 20)2 ; (iii) sin2 0+ 2 sin 0 + 2.
7. No tables to be used in this question.
(i) If lir < x < it and sin x = 4, find cos x and tan x.
(ii) If lir < x < it and tan x = —1, find sin x and cos x.
(iii) If it < x < irr and sec x = —3, find cos x and tan x.
(iv) If 4ir < x < it and sin x = I, find tan x and sec x;
(v) If zit < x < 2ir and sec x = 4, find cosec x and cot x.
8. Prove that, provided sin 0 0.
1 — cos 0 _ 1
sin 0 cosec 0+ cot 0'
94
2] THE TRIGONOMETRIC FUNCTIONS
Is the following converse result true? Values of 0 and ¢ can be found such that
any point on the surface has a position vector of the form
r = a cos 0 cos g5i+ a sin 0 cos cbj+ a sin qlk.
18. Discuss the possibility of solving the following equations for 0:
(i) sin 0 = a+ 1/a; (ii) 2ab cos 0 = a2+b2;
(iii) cost 0 +4 = sin 0 +4 cos 0.
19. What are the maximum and minimum values of the expression 4/(2 + sin x)?
Without attempting to solve the equations exactly, state how many values of x
in the range 0 < x < 27 satisfy
(i) the equation cos x(2+ sin x) = 4;
(ii) the equation tan x(2 + sin x) = 4.
Fig. 6.7
Ex. 6. How do the graphs of sin x and cos x illustrate the identities
sin (-Pr— x) E cos x, sin an + x) a" cos x,
sin (Zi -— x) —cos x, sin an + x) —cos x?
Ex. 7. Sketch roughly the graphs of 2 sin x, sin 2x, sin xj.
Figure 6.8 shows the graph of the tangent function (continuous line)
and cotangent function (dotted line) for values of the domain
— 27r < x < 27r. Recall that the tangent function is undefined at values
of x which are odd multiples of 17r but, by taking x sufficiently close to such
96
3] GRAPHS OF TRIGONOMETRIC FUNCTIONS
values, we make 'tan x I arbitrarily large. A similar remark holds for the
cotangent function at values of x which are even multiples of 1-Tr. (A line
such as x = in is called an asymptote for the curve y = tan x.)
il i 11 I .Y I 11
1%. I I 1I I
ilt II
I\ I It I % I I1
I \ I ■I.
I \ I I \
I \I 1 \ \,■ I X
...
—27r 3 iN —IT V ■0 /1
"\ ir
I — 2 ; I — 2. \ *. \I
IC I-
I I \I I \% I
I I I I 1 I II
I I,I
I I II I I
Fig. 6.8
Ex. 8. How does Figure 6.8 illustrate the identities tan air x) — cot x,
tan (PT+ x) —cot x, tan (171. —x) = cot x, tan air +x) —cot x?
Ex. 9. Sketch roughly the graphs of tan 3x and !tan xi.
Figure 6.9 shows the secant function (continuous line) and cosecant
function (dotted line).
It i YI
II 1 I
t I
Ii t
I t
s
t
II
I % t t
I I
— --I— —
—27737r •-•71 ir it
—2 2
1 tt
1
II I
Fig. 6.9
Ex. 12. Suggest appropriate domains and ranges for arccotangent, arcsecant and
arccosecant.
Ex. 13. Write down the values of arcsin 2iarccos 1/,/2 and arctan (-1).
98
4] INVERSE TRIGONOMETRIC FUNCTIONS
Fig. 6.12
Exercise 6(c)
1. Show that the function f defined by
f(x) = cos x+ sin x
Estimate from your graphs the required angle, giving your answer in degrees
to the nearest degree. (0 & C '0')
5. Solve graphically the equation tan 2x° = 2 cot x°, giving all solutions x such
that 0° < x° < 180°.
6. Draw the graph of y = 1— 2 sin 60x° for values of x from 0 to 6, using 2 cm
as unit on both axes. By drawing a suitable straight line on the same diagram,
read the solutions, within the given range of values of x, of the equation
2(1 — 2 sin 60x) = x. (0 & C ` O')
7. Solve graphically the equation
sin x = cos 2x +1
for values of x for which 0 < x < IT.
8. Using the same axes sketch the graphs of: (i) sin x, (ii) 2 sin x, (iii) 1 +2 sin x.
9. Using the same axes, sketch the graphs of (i) sin x, (ii) — sin x, (iii) 2 — sin x.
10. Sketch the graph of 2(1 — cos x).
11. Sketch the graph of 1— tan x.
12. Sketch the graph of 1+ 'sin x I.
13. Sketch the graph of sin (x+ kir).
14. Sketch the graph of 1 + 2 cos (x
15. Find the values of:
(i) sin (arccos -1); (ii) cos (arcsin —1);
(iii) tan (arccot V3); (iv) sin (arctan — 1);
(v) cos (arccot — 1 /V3); (vi) sec (arccos I);
(vii) cosec (arcsin I); (viii) cot (arcsin —10).
16. Simplify the expressions (i) tan (arcsin x); (ii) sin (arctan x).
17. Sove for x the equation tan (arccot x) = 2 sin (arccos x).
18. Simplify the expression sec {arccos [tan (arccot x)]}.
What restriction must be placed upon the value of x?
19. If arcsin x = 7r— arccos y, find a relation between x and y.
20. Prove that:
(i) arcsin x + arccos x = -17r;
(ii) arcsin [cos (arcsin x)] + arccos [sin (arccos x)] = -fir.
(0 & C modified)
Miscellaneous Exercise 6
1. A chord PQ of a circle of radius r subtends an angle 20 rad at the centre of
the circle and 20 < 7r. Find an expression for the area of the smaller of the two
portions into which PQ divides the circle.
100
4] MISCELLANEOUS EXERCISE 6
If this area is a fraction 1/27r of the area of the circle, prove that sin 20 = 20 —1.
Obtain graphically an approximate solution in radians of this equation.
(0 & C '0')
2. If a circle of radius a is drawn with its centre on the circumference of another
circle of radius a, find the area common to both circles.
3. Find the maximum and minimum values of the expression
sin2 0+6 sin 0+4.
4. Solve the equation
12 cost 0 = 6 + sin 0
giving all values of 0, in radians, lying between 0 and 27r.
5. A globe representing the earth consists of a sphere of radius 10 cm. Find the
length of the circumference of the small circle representing the 30° parallel of
latitude.
Two points A, B on the 30° parallel of latitude differ in longitude by 90°.
Calculate:
(i) the length of the straight line joining A and B;
(ii) the angle subtended by AB at the centre of the globe;
(iii) the great circle distance between A and B, in cm, correct to 3 S.F.
(0 & C O')
`
12. Show by means of a rough graph that the equation cos x = x12,77r has
2n +1 positive roots when n is a positive integer. Show also that, if r is a positive
integer less than n, the root nearest to 2rir is 2rn + a where a is the smallest
positive root of the equation
r x
cos x = -+.
n 2177r
Draw a careful graph of y = cos x for --fn < x < -in and from it find the five
positive roots of the equation
4n cos x = x.
Give your answers in the form kn., where k is correct to two decimal places.
(0 & C)
13. Show that 0 = 2nn+ a and 0 = (2n + 1) it-a, where n is an integer, both
satisfy the equation sin 0 = sin a. Prove, furthermore, that any solution of the
equation must have one or other of these forms.
0 = 2rm +a or 0 = (2n +1) - a is called the general solution of the equation
sin 0 = sin a. Find forms for the general solution of
cos 0 = cos a and tan 0 = tan a.
14. Find the general solutions of the equations
(i) tan x = cot (x+ Pr); (ii) 2 cost x- cos x-1 = 0.
15. AB is a chord of a given circle, centre 0 and radius a, subtending an angle
0 at 0 (0 < 7r). Prove (i) that the area of the triangle OAB is -ia2 sin 0, and (ii)
that the area of the minor segment of the circle of which AB is the chord is
-1-a2(0- sin 0).
Prove also that the area of the part of the circumcircle of the triangle AOB
lying outside the given circle is
(sin 0-0 cos \
a2 (0 & C)
1 + cos 0 •
16. In a circle centre 0, two radii OP, OQ contain an angle 0 radians (0 < 7r).
If the area of the sector OPQ is A and the length of the chord PQ is 2c show that
1 - cos 0 = c20/A. Draw on the same diagram on squared paper the graphs of
the functions y = 1- cos 0 and y = c20/A in the particular case where
A = it cm2, c = 1.3 cm
for values of 0 between 0 and 7T. (Take 1 cm to represent 0 1 on the x axis and
4 cm to represent one unit on the y axis.) Read off from your graphs the value
of 0 between 0 and it which satisfies the equation 1- cos 0 = c20/A in this par-
ticular case. Use your value of 0 to find the radius of the circle in this case.
(Cambridge)
17. The function f: R+ R is defined byf(x) = x sin 1/x. Sketch the graph of the
function.
102
7. Probability in finite outcome spaces
Ex. 1. Verify that 2! = 2, 3! = 6, 4! = 24. Write down the values of 5!, 6!, 7!.
ni
n x (n — 1) x (n — 2) x x (n — r + 1) =
(n— r)!*
103
PROBABILITY IN FINITE OUTCOME SPACES [7
We may modify our question again and ask: 'how many different hands
of four cards may be dealt from a standard pack?' The difference here is
that we are not concerned with the order in which the cards are dealt:
we only want to know the number of selections (sometimes called combina-
tions) of four cards that can be made from fifty-two cards. Suppose for the
moment that this number is N. Then each of these N different hands may
be rearranged in 4! = 24 ways. There are then 24N different arrangements
of four cards. But we already know that the number of different arrange-
ments is 52 x 51 x 50 x 49. Thus
52 x 51 x 50 x 49
N— = 270725.
24
Once more, our argument may be generalized: since there are nP,
possible arrangements of n objects taken r at a time, the number of selec-
tions, or combinations, of n unlike objects, taken r at a time, is
„P, n!
r! r! (n —
The number of combinations of n unlike objects, taken r at a time, is
written either as nC,. or (n) . The latter notation is used almost universally
now, although the former has its merits: we can read as 'n choose r',
reminding us that we are choosing, or selecting, rather than arranging.
We shall retain „Cr in this chapter, but revert to ( ) in later chapters,
r
when the need arises. — n
)e n r — n!
r r!(n—r)!•
(
(Notice that again, since 0! has been defined as 1, 9,C,„ = 1 by the formula,
which accords with the commonsense result that there is just one way of
selecting n objects from n.)
104
1] ARRANGEMENTS AND SELECTIONS
arrangements, the letter a1, a2, a3, armay be rearranged amongst them-
selves in r! ways; similarly the letter b1, b2, b3, . bsmay be arranged in s!
ways and so on. Thus, if we drop the suffixes, the number of distinguish-
able arrangements becomes
n!
r! s! ...•
Hence we have the following important rule.
Given n objects, r alike of one kind, s alike of another kind and so on, the
number of arrangements of the n objects in a row is
n!
r! s!
For example, the number of different arrangements of the letters of the
word SELECTIONS iS
10!
= 907200.
2! 2!
Ex. 8. In how many ways may four letters P and six letters Q be arranged in a
row ?
Exercise 7(a)
1. Five boats are entered for a race. Assuming that they all finish and that there
are no dead heats, in how many possible orders can they pass the finishing line?
2. In a fonn of thirty boys, a first and a second prize are to be awarded; in how
many ways can this be done?
106
1] ARRANGEMENTS AND SELECTIONS
3. You have a form on which you have to give your first six choices for univer-
sity, in order of preference. If you have a list of twelve universities offering the
course you want, in how many ways can you complete the form?
4. A fruit machine has three windows, in each of which appears independently
one of six pictures. How many different arrangements (taking the order into
account) are possible?
5. How many five-digit numbers can be formed from the digits 1, 2, 3, 4, 5,
using each digit once? How many of them are even? How many are even and
greater than 30000?
6. A committee of four is to be selected from six Labour and six Conservative
M.P.s. How many possible committees are there? In how many will the mem-
bers of the Labour Party have a majority?
7. In how many ways may a tennis team of six members be selected from fifteen
available players? In how many ways may a first and second team be chosen?
8. Twelve people are to divide up into three sets of four players for a whist drive.
In how many ways can this be done?
9. Of ten electric light bulbs, three are faulty but it is not known which. In how
many ways may three bulbs be selected ? How many of these selections will
include at least one faulty bulb?
10. How many different bridge hands (thirteen cards) are there which contain
(i) all four aces, (ii) three aces and one king? (Leave your answer in factorial
form.)
11. Criticize the following attempted solution of Example 1:
`A man must sit on the committee, and he can be chosen in four ways. The
remaining two members may now be chosen arbitrarily from among the six
remaining people: this may be done in 6 C2 = 15 ways. Thus, the total number of
possible committees is 4 x 15 = 60.'
12. Prove that the numbers of distinguishable arrangements of n objects in a
row, if two are alike and the rest different, is in!.
What is the number of distinguishable arrangements if three are alike and the
rest different?
How many different telephone numbers can be made using all the digits of
the number 4225267?
13. How many different arrangements are there of the letters of the word
QUEUE? In how many of these arrangements do the letters QU appear together,
in that order?
14. Find in how many ways a batting order (eleven men) may be made if Smith
is to bat before Brown.
15. A diagonal of an n-sided polygon is a line joining two non-adjacent vertices.
How many diagonals does an n-sided polygon possess?
16. In how many ways can two ls, two 2s and two 3s be thrown with six dice?
107
PROBABILITY IN FINITE OUTCOME SPACES [7
17. Given n unlike objects, find the number of ways of dividing them into three
unequal groups of sizes p, q, r where p + q+ r = n.
In how many ways can they be divided into two groups of size p and one of
size n— 2p, if no attention is paid to the order of the groups?
18. A pair of integers is selected from the set of positive integers 1, 2, 3, ..., n.
In how many ways may this be done? [In each pair the order of the integers is
immaterial, e.g. (2, 3) and (3, 2) count as one pair only.]
If the integers in each pair are multiplied together show that, in the case when
n is odd, the number of products which will be odd integers is - k (n2—1). If
n is large show that this number is approximately one-quarter of the total number
of products. (Cambridge)
the outcome space: any other set whose elements represent distinct pos-
sible outcomes and which exhausts all the possibilities will do, e.g.
S2 = {G, F, P},
where G denotes a good result (no defectives);
F denotes a fair result (one or two defectives);
P denotes a poor result (more than two defectives).
Generally speaking, it is best to choose as an outcome space one that
gives as much detail as possible about the result obtained, but there are
exceptions to this and other criteria may be adopted.
Ex. 10. Two coins are tossed and the result (in terms of heads and tails) noted.
Suggest three possible outcome spaces.
Ex. 11. A card is drawn from a pack and its value and suit noted. Criticize the
following outcome space:
S = {card is an ace, card is a heart, card is neither a heart nor an ace}.
Ex. 12. Suggest an outcome space for Ex. 11 if the experiment is concerned only
with the drawing of a heart or an ace from a pack of cards.
Ex. 13. A count is made of the numbers of girls and boys in a family. Under
what circumstances would the set
S = {there are more girls than boys, there are more boys than girls}
constitute an outcome space?
3. PROBABILITY DISTRIBUTIONS
We now assign to each elementary event si of our outcome space S a
positive fraction pi (i.e. 0 < pi < 1) called the probability that the outcome
of our experiment will be si. Furthermore, for consistency, we shall make
the sum of all the probabilities piover the entire outcome space 1. Thus,
the probabilities associated with an outcome space S = {si, s2, sn } are
real numbers such that
0 < pi < 1, (1)
Pi+P2+133+ •+Pn = 1. (2)
Such a set of probabilities is said to constitute a probability distribution
for the given outcome space.
It will be observed that so far we have attached no meaning to the
numbers pi —we have only placed certain restrictions upon their possible
values. Provided we observe these restrictions we have a mathematically
meaningful system. However, it is desirable that a mathematical system
should have some relevance to the physical world. If a mathematical
system describes some physical situation, we are said to have created a
109
PROBABILITY IN FINITE OUTCOME SPACES [7
Ex. 14. Of the outcome spaces you suggested for Ex. 10 which do you think may
reasonably be supposed to be uniform?
Ex. 15. Two dice are thrown and the total score is noted. Criticize the assigning
of a uniform distribution to the outcome space
S = {2, 3, 4, ..., 12}.
110
3] PROBABILITY DISTRIBUTIONS
Ex. 16. Suggest an outcome space for the experiment of throwing two dice
that may be given a uniform distribution.
Ex. 17. From the weather records of a certain town taken over the past thirty
years, on the average five days in November have been recorded as foggy. Is
it justifiable to assume that the probability of a November day in that town
being foggy is i?
Ex. 18. The number of boys and girls in a family of five children are noted.
Justify the assumption that the outcome space
S = {there are more girls than boys, there are more boys than girls}
has a uniform distribution (i.e. p = I for each of the two elementary events).
4. PROBABILITIES OF EVENTS
Any subset E of the outcome space S of an experiment is called an event.
(Notice that the use of the word in `elementary events' is consistent with
this definition, since {s} c S.) If we have assigned a probability distribu-
tion to our outcome space it becomes meaningful to consider the prob-
ability of the event E (by which we would mean, in a practical example, a
measure of our degree of belief that the experiment will result in one of the
elementary events belonging to E). A moment's consideration should
show that the following definition is plausible, at least in simple cases.
Given a sample space S with associated probability distribution, and an
event E c S, then the probability of the event E, written Pr (El S) (read: `the
probability of event E given S'—more explicitly, `the probability of E
given the probability distribution of the outcome space S') is defined as
the sum of the probabilities of all the elementary events belonging to E.
More formally, if S = {s1, s2, sn} and piis the probability associated
with si, and if E {s„, s„ s,} then
Pr (EIS) = 1911 ±Prz+Pr3+ • • • +Pr..
*Ex. 19. How would you interpret the probabilities obtained in Examples 3 and
4?
Ex. 20. What is the probability of obtaining one head and two tails if three coins
are tossed ?
Ex. 21. What is the probability of obtaining a total of more than 10 from the
throw of two dice?
Ex. 22. There are one thousand tickets issued in a lottery and prizes are awarded
for twenty of them. What is the probability of any specified ticket securing a
prize?
Ex. 23. What is the probability that a bridge hand contains (i) just one heart
(a singleton heart); (ii) a singleton in just one suit? (Leave your answers in
factorial form.)
Probability is occasionally formulated in terms of odds rather than as
a fraction. If the probability that a horse will win a race is estimated as I,
then the probability that it will not win is aand the odds against its winning
are 3 to 1 (sometimes written as 3 to 1 against'). Again, if the probability
that another horse will win a race is 3, then the odds in favour of its winning
are 2 to 1 (or 2 to 1 on').
Ex. 24. The odds on horse A to win a race are 3 to 1 on and on horse B 4 to 1
against. Write down the probabilities
(i) that A wins;
(ii) that B wins;
(iii) that A does not win.
Exercise 7(b)
1. Consider families consisting of six children, all of different ages. The 'type'
of family is defined by an ordered set of the form {B, B, G, B, G, B} the elements
of which represent the sex of each child, starting with the eldest. How many
different types are there? In how many different types are there three boys
and three girls? What is the probability that a family of six children will consist
of three boys and three girls?
112
4] PROBABILITIES OF EVENTS
Example 5. Two dice are thrown. What is the probability of scoring either a
double, or a sum greater than 9 ?
We take as our outcome space the set of 36 pairs (i, j), where i, j run
independently from 1 to 6. We make the assumption that the events are
equiprobable and so attach a probability to each elementary event of -316.
Write
E1= {(1, 1), (2, 2) (3, 3) (4, 4) (5, 5) (6, 6)},
E2 = {(4, 6) (5, 5) (6, 4) (5, 6) (6, 5) (6, 6)},
t The notation n means the union of the n sets E1, E2, E,.; that is
i-1
E1 UE2 U...U E,,.
114
51 THE ADDITION LAWS
Example 6. Two dice are thrown. What is the probability of not getting a
double?
Making the same assumptions as in Example 5 and using the same nota-
tion,
Pr (E.; = 1—Pr (E,IS), by Theorem 7.3,
1-*
5
-
E.Y. 25. In a class of boys, one-third have black hair and one-quarter have brown
eyes. What deductions can you make?
Ex. 26. If a card is drawn from a well-shuffled pack what is the probability that
it is either an ace or a king?
Ex. 27. If a card is drawn from a well-shuffled pack, what is the probability that
it is an ace or a heart?
Ex. 28. If a number is selected at random from the integers I, 2, ..., 30 what is the
probability that it is
(i) divisible by 2;
(ii) divisible by 3;
(iii) divisible by 6;
(iv) not a multiple of 2 or 3 ?
Ex. 29. If one thousand tickets are issued in a lottery in which there are two
first prizes, eight second prizes and ten third prizes, what is the probability of not
securing a prize with one ticket?
Now since F is the new outcome space, the new probability distribution
for F must have the property that the probabilities of the elementary
events contained in F sum to 1. Making the plausible assumption that the
relative proportion of the weights attached to the elementary events of F
remain unaltered, this is equivalent to scaling up the probabilities in F by
a factor 1/Pr (FM S). The probability that E will occur is thus Pr (E n FMS),
scaled up by this factor 1/p(FIS). We are thus led to make the following
definition:
Pr (E n FMS)
Pr (ELF) =
Pr (FMS) •
Example 7. A bag contains twenty balls, ten of which are red, eight white
and two blue. The balls are indistinguishable apart from the colour. Two balls
are drawn in succession, without replacement. What is the probability that
they will both be red?
We take as our outcome space S the set of 20 x 19 = 380 possible
selections of two balls. Since the balls are indistinguishable apart from
colour, we may impose a uniform probability distribution upon S.
Now define R1as the event 'the first ball picked is red', R2as the event
`the second ball picked is red' and similarly W1, W2, B1, B2.
We require to find Pr (R1 n R2 15).
Now Pr (R1 S)=14) and Pr (R,IRD—the probability that a red ball is
selected from a bag now containing nine red, eight white and two blue
balls is -A,-.
Thus,
Pr (R, n R2 I S) = Pr (R2 IR1) . Pr (R,IS)
_
-
9 10
19.20
_ 9
- 38•
Example 8. With the data of Example 7, what is the probability that we obtain
a blue and a white ball (in either order)?
Adopting the notation of Example 7, we are required to calculate the
probability of obtaining one or other of the two mutually exclusive events
n W2and W„ n B2; that is Pr {(B, n wo u n Bois).
Now
Pr {(B1n W2) U n BO = Pr (B1n S)+ Pr (W1n B2 S),
since the events are mutually exclusive using Theorem 2;
= Pr (W21B1). Pr (.1311 S)+ Pr (B2 I . Pr (W, IS)
= 1 9. A+-A-.-A
116
6] THE MULTIPLICATION LAWS
The arguments employed for the solution of the preceding examples may
be represented diagrammatically by a probability tree (see Figure 7.1).
Starting at 0 on the extreme left, the end-points of the first three branches
represent all the possible (mutually exclusive) results of the first draw.
The end-points of the next three sets of branches represent all the possible
.,......„...- R2
9
8
R1 19 W2
2 ___
-----""*"
10 /1 0 —."'"'s B2
E
R2
10
8 7
20 Wi 19 W2
2 ,
13 B2
20 10 ....,./.. R2
8
.1°
Bi - 19 W2
.'.'"''''1
D ''''''' /32
Fig. 7.1
outcomes of the second draw, the various sets corresponding to the dif-
ferent draws on the first round. The various probabilities are then attached
to the branches as shown and the probability of any desired path from
0 to some end-point on the right may be found by multiplying the suc-
cessive branch probabilities together.
For example
Pr (W,r) R, IS) = io x9 =9
Pr (B2 IS) = Pr (R1 n B2 IS) + Pr (W1n B2 IS)+Pr (Bin B2 IS)
(events mutually exclusive)
= Pr (R1 IS). Pr (B2 IR1) + Pr ( W, IS). Pr (B2I W1)
+ Pr (Bi lS). Pr (B21131)
_ 10 126 286 .126 +226 .11si
1
— 10 •
(Can you see a simpler method of arriving at this last probability ?)
117
PROBABILITY IN FINITE OUTCOME SPACES [7
Probability trees are particularly helpful if a number of probabilities
have to be read off. If only one probability is required, a simplified tree may
often help. For example, in Example 7 we are essentially concerned with
two alternatives: red and not red. Figure 2 shows a suitable tree for this
example.
Fig. 7.2
Ex. 30. Prove that, if Pr (A n Bls) = Pr (AIS) Pr (B S), then Pr (A1S) = Pr (A IB).
Ex. 31. A bag contains two white balls and six red balls; a second bag contains
four white balls and four red balls. The balls are indistinguishable apart from
colour. What is the probability of obtaining two white balls if
(i) one ball is drawn from each bag;
(ii) two balls are drawn from the first bag;
(iii) two balls are drawn from the second bag?
Ex. 32. Criticize the following argument:
`A ring of anti-ballistic missiles is estimated to give a probability of 02 of
destroying any incoming missile. A set of five such rings will therefore render
a city immune to missile attack.'
Ex. 33. 'In a form of twenty-four boys, six failed their English examination and
four failed their Mathematics. The probability that a boy, selected at random
from the form, failed in both English and Mathematics is thus r64 x 14 = ,
Exercise 7(c)
1. What is the probability of not throwing a six with four throws of an un-
biased die? What is the probability of throwing at least one six?
2. On average, 2 % of the electric light bulbs of a certain type prove to be
faulty: what is the probability that a batch of twelve such bulbs will be free
of defectives ?
3. If four people are chosen at random, find the probability that no two of them
were born on the same day of the week. (M.E.I. '0')
4. A bag contains a dozen apples, of which three are bad. If two apples are
withdrawn at random, find the probability that
(i) both are good ;
(ii) both are bad;
(iii) one is good and one bad. (M.E.I. '0')
5. Show that there is a better chance of throwing at least one 6 with four throws
of a single die than there is of throwing at least one double 6 with twenty-five
throws of a pair of dice. (This is a modification of a famous problem in the hiS-
tory of probability, first proposed by a gambler, the Chevalier de Mere and
transmitted by him to the French mathematician Fermat, who managed to solve
it.)
6. A bag contains two red, three white and four blue balls. If two balls are drawn
in succession without replacement, what is the probability of drawing a red and a
blue ball, in either order? What is the corresponding probability if the first ball
is replaced before the second is drawn ?
7. A bridge hand consisting of thirteen playing cards contains two aces. If
five cards are drawn at random from the hand, find the probability that the five
will contain neither of the aces. (M.E.I. '0')
119
PROBABILITY IN FINITE OUTCOME SPACES
8. A teacher has twelve pupils in his form, eight boys and four girls. On a school
outing he can provide transport in his car for three pupils and decides to draw
lots for the seats. Construct a tree to show on its branches the probabilities as to
the sex of the winning pupils in the different stages of the draw. Hence, or other-
wise, find the probability that those winning will contain
(i) two boys and one girl,
(ii) at least two girls. (M.E.I. `O')
9. Given five different letters and their respective envelopes, in how many ways
can one letter be placed in each envelope if this is done at random?
Find in how many cases only two of the letters will be in their correct en-
velopes.
What is the probability that at least three of the letters are in their correct
envelopes? (M.E.I. `O')
10. What is the probability that, after a pack of cards has been well shuffled,
two aces will lie at the top ?
11. A bag contains three red, four white and five blue balls, indistinguishable
apart from their colour. If two balls are drawn successively, without replace-
ment, estimate the probability of obtaining two balls of different colours.
12. A machine has N components, each of which has a probability of 2/3 of
breaking down. The machine will function provided at least one of its compon-
ents is functioning. What value should be chosen for N if the machine is to be
99 % efficient ?
13. A pack of fifty-two ordinary playing cards is shuffled and a card is with-
drawn. D denotes the event that the card is a diamond, K that it is a king, R
that it is red. Prove by calculating the appropriate probabilities that D and K
are independent, that K and R are independent, but that D and R are not in-
dependent.
What is the value of Pr [(D' n K) U (D n K')]? (M.E.I.)
14. Three cards are drawn from a pack of fifty-two cards and, when they have
been replaced and the pack shuffled, a second set of three cards is drawn. Find
the chance that the six cards drawn should include at least one ace. (M.E.I.)
15. A bag contains five white and three red balls. Balls are drawn in succession
and are not replaced. Show that the chance that the first red ball will appear
at the fifth draw is 3/56. (M.E.I.)
16. Three bags, A, B, C, contain respectively three white and two red balls, four
white and four red balls, five white and two red balls. A ball is drawn unseen
from A and placed in B; then a ball is drawn from B and placed in C. Find the
chance that if a ball is now drawn from C it will be red. (M.E.I.)
17. How many times should an unbiased die be thrown if the probability that
a six should appear at least once is to be greater than 9/10? (M.E.I.)
18. A batch of fifty articles contains three which are defective. The particles are
drawn in succession (without replacement) from the batch and tested. Show that
the chance that the first defective met will be the rth article drawn is
(50 — r) (49 — r)/39200. (M.E.I.)
120
6] THE MULTIPLICATION LAWS
19. Two men A and B play a game in which A should win eight games to every
seven won by B. If they play three games, show that the probability that A will
win at least two games is approximately 0-55. (Cambridge)
20. The chance of any one engine of a four-engined aeroplane failing on a long
journey is 5 %. If only one engine fails the chance of the aeroplane completing
the journey is 80 %; if two engines on opposite wings fail, its chance of com-
pleting the journey is 50 %. It cannot fly with two engines out of action on the
same wing. Find the chance that the aeroplane will complete the journey.
(0 & C)
21. (i) From a bag containing five red, four white and three green balls, three
are drawn together at random. Find the chance of their being
(a) all of different colours,
(b) all of the same colour.
(ii) Two six-faced dice whose faces bear the numbers 1 to 6 respectively are
thrown together. Find the chance of
(a) the total score being exactly 8,
(b) the total score being greater than 8. (0 & C)
Example 10. Two bags contain coloured balls as shown in the table below.
Pr (Wn I)
Pr (W)
Pr (WI1).Pr (I)
Pr (W)
=
Example 11. Three machines produce the same type of electrical component,
20 % of the total output coming from machine A, 50 % from machine B and
30 % from machine C. Tests conducted in the past show that 5% of the com-
ponents from A and 1% from each of B and C prove faulty. A component
selected at random from the total output is proved to be faulty. What is the
probability that it came from machine A?
Call A the event 'component comes from machine A',
B the event 'component comes from machine B',
C the event 'component comes from machine C',
F the event 'component found faulty'.
We have to determine Pr (AIF).
Now, from the data, we may assign the following probabilities:
Pr (FIA) = ao; Pr (FIB) = 100; Pr (FIC) = 1,130;
Pr (A) = ; Pr (B) = -M,; - Pr (C) =
Now Pr (AIF) = Pr (A n F)/Pr (F) and so the problem reduces to deter-
mining Pr (A n F) and Pr (F), given the numerical values above. But
Pr (A fl F) = Pr (F fl A)
= Pr (FIA). Pr (A)
_
— loo•
Again F = (F n A) U (F n B) U (F n C) and these three events are
mutually exclusive (no component can come from more than one machine).
Thus
Pr (F) = Pr (F n A) + Pr (F n B)+ Pr (F fl C)
= Pr (FI A) . Pr (A) +Pr (FIB) . Pr (B) +Pr (FI C). Pr (C)
= 11 j_ 3 _
100 200 • 1000
9_
500•
Pr A n F)
(
Ex. 35. Of two pennies, one is double-headed, the other normal. If one of the
pennies is selected at random and tossed twice, what is the probability of obtain-
ing two heads? If one of the coins is selected at random, tossed twice and gives
two heads, what is the probability that it is double-headed ?
Ex. 36. Find the probabilities that the faulty article of Example 11 came from
(i) machine B, (ii) machine C.
5 PPM 123
PROBABILITY IN FINITE OUTCOME SPACES [7
Ex. 37. With the data of Example 11, find the probability that a component
selected at random and shown to be not faulty comes from machine A.
Ex. 38. Two articles come from one of the machines of Example 11, but it is
not known which one. On testing, both are found to be faulty. Find the respec-
tive probabilities that they come from machines A, B and C.
Exercise 7(d)
1. One of four dice is known to be biased, showing a 6, on average, three times
as often as each of the other scores. A die is chosen at random and thrown three
times and three 6s appear. What is the probability that it is one of the unbiased
dice?
2. A factory manufactures three different qualities of light bulb, A, B and C in
the ratio 1 :2 : 3. The bulbs are indistinguishable in external appearance but
extensive tests indicate that, on the average, 1 % of type A, 4 % of type B and
5 % of type C are below the advertised standard. A batch of six, all of the same
type is sent in error without a distinguishing label. If none of these turns out to be
defective, estimate the probability that they were of type A.
3. Of a large sample of items, 40 % were produced by machine A and 30 % by
each of machines B and C. The machinery was very unreliable, machine A pro-
ducing 10 % defective items, machine B 20 % defective items and machine C
30 %. If an item, selected at random from the sample, proves to be defective,
what is the probability that it came from machine A?
4. A card is missing from a pack of 52 cards. If this is the only information you
have, what is the probability that the missing card is a spade? The pack is well
shuffled and the first card is removed and proves to be a spade. What would your
assessment of the probability that the missing card is a spade be now? The card
removed is now replaced and the pack shuffled. The top card again proves to be
a spade. What is your assessment of the probability now?
5. Four bags, I, II, III, IV, contain white and black balls as shown in the follow-
ing table.
Bag I II III IV
Number of white balls 1 2 3 4
Number of black balls 9 8 7 6
A die is thrown; if a one appears, bag I is chosen, if a two or three, bag II,
if a four or five, bag III and if a six, bag IV. A ball is then drawn at random from
the bag selected.
5-2 125
PROBABILITY IN FINITE OUTCOME SPACES
If you are told that a black ball has been drawn, what should your assess-
ment be of the probability that it came from bag I?
6. A certain rare disease from which one in ten thousand of the population suffers
is diagnosed by a test which reveals the presence of the disease in 95 % of the
cases of those tested who actually have the disease. However, it also incorrectly
yields a positive reaction in 1 % of the cases of those who are not suffering from
the disease. If a person selected at random from the population shows a positive
reaction, what is the probability that he is actually suffering from the disease?
7. One of four pennies is known to be double-headed, the other three being
normal. One of the pennies is selected at random and tossed three times. If the
result is three heads, what is the probability that the coin tossed is the double-
headed penny ?
8. An engineer is responsible for servicing three small computer installations
A, B, C. The probability that A, B, C will break down on any day are 44, T PT)
respectively. The probabilities that the three computers will not be in operation
a, i
after 7 p.m. are 1, respectively. On one day the engineer receives a message
that one of the computer installations has broken down at 8 p.m. but the
message does not specify which computer. Calculate the probabilities that the
breakdown is at A, B, C and so decide the first installation which should be
checked by the engineer. (M.E.I. adapted)
9. The probability that a radio message sent from a ship is not picked up by a
land-station is p. The land-station replies to all messages received and its trans-
mission back has probability q of not being picked up by the ship. Find the
probability that the land-station failed to receive the message if the ship gets
no reply.
Miscellaneous Exercise 7
1. A number is drawn at random from the set of thirty-seven numbers
0, 1, 2, ..., 36,
and is then replaced.
(i) What is the chance that the number drawn will be odd?
(ii) What is the chance that in two consecutive draws the numbers drawn will
be the same?
(iii) What is the chance that in one hundred consecutive draws the number 0
will not be drawn?
(iv) What is the chance that in five consecutive draws none of the numbers
will be odd?
Give the answers to (i) and (ii) as vulgar fractions, and evaluate the answers
to (iii) and (iv) approximately as decimals. (Cambridge)
2. An encyclopaedia consisting of n (? 4) similar volumes is kept on a shelf
with the volumes in correct numerical order; that is, with volume one on the
left, volume two next and so on. The volumes are taken down for cleaning and re-
placed on the shelf in random order. What are the probabilities of finding
exactly n, (n-1), (n-2), (n— 3) volumes in their correct order on the shelf?
(Cambridge)
126
71 MISCELLANEOUS EXERCISE 7
3. You have available two pairs of dice. The dice of one pair are both true but,
in the other pair, one is true and the other is known to be biased in some way.
Which pair would you choose if you were playing a game in which to win you
had to throw (a) a score of 7; (b) a double?
4. Two persons X and Y play a game in which X deals from a standard pack of
fifty-two cards. Each player receives three cards and the game is won by X if he
holds more red cards than there are black cards held by Y.
By considering the number of black cards in play or otherwise, calculate the
probability that the dealer wins a game. What odds should be fixed by the dealer
in order that he may expect just to show a profit? (M.E.I. ' A')
5. Four cards, the aces of hearts, diamonds, spades and clubs are well shuffled
and then dealt two to player A, the other two to player B. A is then asked whether
at least one of his two cards is red. He replies in the affirmative. In the light of this
information we wish to calculate the probability that he holds both the red aces.
Consider the argument:
`We know that he has one red ace; without loss of generality we may suppose
that it is the heart ace. Among the other three cards there is no reason why one
more than another should be the diamond ace: one only out of three equally
likely possibilities gives A both the red cards: the required chance is thus 4.'
Criticize this argument and produce a correct argument and answer. (C.S.)
6. A bag contains twelve coloured balls: 4 yellow, 3 white, 2 black, 2 red and
1 green. A prize is offered for obtaining 2 red balls in two draws, there being an
option of either replacing or not replacing the first ball. Which would you choose ?
Generalize your result to the case of a bag containing N balls of s different
colours where there are m, balls of colour i, and prove your conjecture.
7. Raffles of two types, A and B, are held. In raffles of type A there are 2N
tickets, of which 2n carry prizes; in those of type B there are N tickets of which
n carry prizes. Which would you prefer: two tickets in a raffle of type A, or two
tickets in separate raffles of type B?
8. A six-digit telephone number (i.e. any number between 100000 and 999999)
is dialled at random. Show that the probability that one or more digits are dialled
in succession (as in 322916 or 322216 or 322911 but not as in 329216) is about
41 %. (Cambridge)
10. Articles are distributed among three boxes in such a way that each article
has the same chance of being placed in any one of the boxes. Find the probability
that at least one of the boxes remains empty when r articles have been dis-
tributed.
Deduce that, if r articles are similarly distributed among five boxes, the prob-
ability that exactly two boxes remain empty is 6(3r-1 — 2r + 1)/5r-1. (Oxford)
127
PROBABILITY IN FINITE OUTCOME SPACES [7
drives at 30 m.p.h. but if forced to stop at the lights, he accelerates uniformly from
rest at 4 ft/s2until he attains that speed. Estimate the probabilities of his finding
(i) both sets of lights in his favour;
(ii) both sets of lights against him;
(iii) only one set of lights against him.
Suggest two ways in which the relative timing of the two sets of lights could
be adjusted to ensure that the probabilities would be the same if travelling in
either direction and calculate the probabilities.
(Assume throughout that his progress is not interfered with by other vehicles,
and also, that he drives at 30 m.p.h. right up to the lights and then stops
instantaneously if they are red at that instant.) (SMP)
18. Three players, X, Y, Z, play a game and at each round of the game they have
an equal probability of winning that round. The player who wins each round
scores one point and the game is won by the first person to score a total of
3 points. X wins the first round; calculate the probability that Y wins the game.
(M.E.I.)
19. The basic premium under a certain motor insurance policy is EA p.a. The
scale of No Claims Bonus provides 25 % discount after the first year without
claim, 331 % after two consecutive years free of claim and 50 % after three or
more consecutive years free-of claim. In addition a preferred policy holder's dis-
count of 20 % is given on the basic premium (which may already be discounted
by the No Claims Bonus) for four consecutive years without claim. If a preferred
policy holder's discount has been granted it is not lost by reason of a claim.
Suppose that for a given driver there is a constant probability q of no claim in
a given year. Evaluate the probabilities for the possible premiums in the sixth
year of the policy, in terms of q. (M.E.I.)
20. n2balls, of which n are black and the rest white, are distributed at random
into n bags, so that each bag contains n balls. Determine the probability that at
least one bag contains no black ball.
129
8, Finite series and the Binomial Theorem
1 357 2r -1
Fig. 8.1
of the sequence {r 3-1} are 0, 7, 26, 80, 124, .... Again, the sequence
1, 3, 5, 7, 9, ... could be written {2r -1}-but see Ex. 5, which illustrates
the danger of assuming that the form of the rth term is determined by the
first few terms of a sequence.
Sequences are sometimes defined by giving, say, the first term and a
relation between the (r - 1)th and rth terms. For example, the sequence
defined by the relation
Ur = +1 (1)
*Ex. 1. What are the first four terms of the sequence given by
ur= ru„_,+ 1,
(i) if It, = 0; (ii) if u1 = 2?
Ex. 3. Write down the first six terms of the following sequences:
(i) {3r}; (ii) {3r-5}; (iii) {2r-1}; (iv) {r(r2 1)};
-
Ex. 7. Evaluate:
4 3 4
(i) E (r + 2); E r2; E r3;
r=1 r=1 r=1
3 2 1
(iv) E (r- 1)2; (V) E (vi) E (- 1)r r2;
r=0 r =0 r+2' r=1
4 (1 1 \
(vii) E 2r;
r=0 r=2 kr r+11
Ex. 8. Write in E notation:
(i) 1+2+3+4; (ii) 1-2+3-4;
(iii) 2+4+6+8; (iv) 1+2+4+8+16;
(v) 1+++++++1+*; (vi) 1.2+2.3+3.4+4.5+5.6+6.7;
(vii) 1 - 1 + 1 - 1 + 1 - 1 + 1 - 1; (viii) 1+2+3+4+...+n.
*Ex. 9. Prove that, if {u7 } and {v,} are two sequences, then
(ar + v7) = E ur +
r=m r=172 r=m
*Ex. 10. Prove that, if {ur} is any sequence and a a fixed number, then
n n
aur = a .
r=m r=m
*Ex. 11. Interpret and evaluate E c, where c is fixed. (That is, find the sum of the
r=1
first n terms of the constant sequence defined by Ur = c, all r.)
2. ARITHMETIC SEQUENCES
Sequences such as (1, 3, 5, 7, _I, {11, 21, 3i, 5, ...}, {4, 1, -2, -5, _} are
called arithmetic sequences, their distinguishing feature being that the
difference between any pair of successive terms is constant. If we call this
common difference d, and the first term a, then the arithmetic sequence
takes the form {a +(r-1) d}.
Ex. 12. Write down the rth terms of each of the following arithmetic sequences:
{1, 4, 7, ...}, {1, 2+, 4, ...}, {2, 0, -2, ...}.
Ex. 13. Find to which integer r the following terms correspond in the given
arithmetic sequences:
(i) 91 in {1, 3, 5, ...}; (ii) 139 in {3, 7, 11, ...);
(iii) -253 in {2, -1, -4, ...}; (iv) 61.7 in {3.2, 4.7, 62, ...};
(v) -67 in {71, 61, 5i, ...}.
133
FINITE SERIES AND THE BINOMIAL THEOREM [8
Ex. 14. Taking what you consider to be the simplest continuation in each case,
write down, in E form, the sum of the first n terms of each of the following
series:
(i) 1.4.7+4.7.10+7.10.13+10.13.16+...;
(ii) 2.3.5+4.6.9+6.9.13+8.12.17+...;
(iii) 24 +44 + 67 +8"+...;
1.3 3.7 5.11 7.15
(iv) 5 + 8 + 11 +...;
1 2 3 4
(v) 1.2.3 3.5.10+ 5.8.17 7.11.24'
= na++dn(n+1)—nd
= -in[2a + (n —1) d].
This method is easily applicable to any arithmetic series and depends only
upon memorizing the single important result
E r = in(n + 1).
r=1
Example 1. The tenth term of a certain arithmetic series is —29 and the
twentieth term is — 69. Find the sum of the.first twenty terms.
Let the first term be a and the common difference d. Then
a+9d = — 29, a+ 19d = — 69,
giving a = 7, d = — 4. The rth term of the arithmetic series is thus
134
21 ARITHMETIC SEQUENCES
7 +(r —1) (-4) = 11-4r and the sum of twenty terms is given by
20
E (11 - 4r) = 220-2.20.21
r=1
= — 620.
Exercise 8(a)
1. Evaluate:
6 4 2
(i) E (r2+ r); -; (iii) E (1 +r+r2).
r=1 2.=.1 r -2
2. Write in E notation :
(i) 1.3 + 4.6 + 7.9 + (to n terms) ;
(ii) 1.2.3 — 2.3.4 + 3.4.5 — 4.5.6 + (to n terms) ;
(iii) 12_32+52_72 + (to n terms).
3. Find the first four terms of each of the sequences defined by the following
recurrence relations and initial terms:
(i) ur = 3/4_1+1, u1 = 1;
(ii) = 0, u1 = 1, u2 = 1;
(iii) u, = ui = 2;
(iv) 2ur a,-1 -4,_1 = 3, u1 = 1.
4. If {ur } is an arithmetic sequence with common difference d and first term a,
show that ursatisfies the recurrence relation
llr = ur_i+ d.
If s„ = E u,., determine a recurrence relation involving only s,, sn-1, sn-2
r=1
and d. What is si?
5. Find the rth term of the following arithmetic sequences:
(i) {3, 12, 21, ...}; (ii) ...}; (iii) {12, 4, —4, ...}.
6. Find the number of terms in the following arithmetic series:
(i) 1+3+5+...+101; (ii) 2+9+16+...+485;
(iii) 5+2+(-1)+...+(-82).
7. Sum each of the series in Question 6.
8. The fourth term of an arithmetic series is — 14 and the tenth term is — 44.
Find an expression for the rth term and hence obtain the sum of the first twenty
terms.
9. The first term of an arithmetic series is 9 and the sum of the first ten terms
is 450. Find an expression for the rth term and deduce an expression for the sum
to n terms. Check your final result by setting (i) n = 1; (ii) n = 10.
10. The ninth term of an arithmetic series is 24 and the sum of the first nine terms
is 126. Find an expression for the rth term and deduce an expression for the sum
to n terms.
135
FINITE SERIES AND THE BINOMIAL THEOREM [8
11. How many terms of the series 1 + 8+15+ ... are required to give a sum of
5500?
12. How many terms of the series 12+ 6+ 0+ ... are required to give a sum of
- 3348?
13. Show that
(r+1)2 = E r2 +2 E r + n
r=1 r=1 r=1
n
and deduce that E r = in(n+ 1).
r=1
14. If E sr = n2 + n, prove that {s7} is an arithmetic sequence, and find the first
r=
three terms.
15. Prove that the sum of n terms of any arithmetic series may be written in
the form an2+ bn + c. Prove the converse of this result: that if
s,. = an2+bn+ c,
r=i
then {sr} is an arithmetic sequence.
16. If a2, b2, c2are the first three terms of an arithmetic sequence, prove that
(b+ c)-1,(c+ a)-1,(a+b)--1are the first three terms of another arithmetic
sequence (provided b+ c * 0, c+ a * 0, a+b * 0).
17. If 3 + x + y + 29 is an arithmetic series, find x and y.
If a + xi+ x2+ ...+ xn+ b are the (n+ 2) terms of an arithmetic series, find xr
in terms of a, b, r and n.
18. An arithmetic sequence whose first term is 8 has a common difference of 1.
Find N, given that the sum of the first 2N terms is equal to the sum of the next
N terms.
3. GEOMETRIC SEQUENCES
Sequences such as {1, 2, 4, 8, ...}, {3, -6, 12, -24, ...}, {1.5, 0.75, 0.375...},
in which each term is a fixed multiple of the preceding term are called
geometric sequences. If we call the first term of the sequence a, and the fixed
multiple, or common ratio, p then the geometric sequence takes the form
{apr-1}.
Ex. 15. Which of the following sets of numbers could be the first three terms of
geometric sequences?
(i) {1, -5, 25, ...}; (ii) {2, 24, 2F1, ...} (iii) {l, 4, 9, ...};
(iv) {1, - Fl, 121, ...}; (v) {7, -1, -4, ...}; (vi) {1, -1, 1, ...).
Ex. 16. Find to which value of r the following terms correspond in the given
geometric sequences:
(i) 256 in {2, 4, 8, ...}; (ii) 1458 in {2, 6, 18, ...};
(iii) 10-" in {10, 1, 0.1, ...}.
136
3] GEOMETRIC SEQUENCES
Since, for Ip I < 1, lim pn = 0 (consult a calculus text-book for the de-
ryb—>C0
E apr 1 — a/(1— p)
r =1
less than any positive number, however small, by taking a sufficiently
large n, provided Ipl < 1. Thus, if we write
n o0
we have
r=1
apr-' = 1 %, provided IPI< 1.
Ex. 17. Comment upon the number of decimal points used in the calculations in
Example 2 and on the accuracy achieved.
Exercise 8(b)
1. Find the sum of the following geometric series, simplifying your answers as
far as possible, but without evaluating terms of the form p":
(i) 1 + 4 + 16 + ... (12 terms); (ii) 1 — 2+4 — 8+ ... (20 terms);
(iii) A/2+ 2+24 + ... (16 terms); (iv) 1+1+ ... (50 terms)
2. The third term of a geometric sequence is 5 and the eighth is 160. Find the
first term and the common ratio.
3. Find two possible geometric sequences whose fourth term is 9 and whose
eighth term is 144.
4. Which is the first term of the geometric sequence 1, +, ... which is less than
10-8 ?
5. How many terms of the geometric series 1 +1++ + ... must be taken to give
a sum greater than 2-10-6 ?
6. How many terms of the geometric series 1 +i+A+... must be taken to
give a sum greater than 4-10-3 ?
7. If x 4, x, x + 6 are the first three terms of a geometric sequence, find x and
-
If all the terms on the left-hand side are added together we have precisely
E r(r +1); if all the terms on the right-hand side are added together, we are
r=1
left with —4-0.1.2 at the beginning and in(n + 1) (n+2) at the end, all
the others having cancelled out in pairs. Thus
r(r +1) = in(n + 1) (n + 2).
r=1
1 1 1
Ex. 18. Show that
r(r+1) = r r+1
In the example given above, sr = r(r+ 1) and ur= 4-(r — 1) r(r+ 1);
in Ex. 18,
1
s,. and u,. = — 1/r.
r(r +1)
This technique for summing series is called the method of differences; its
use depends, of course, upon our ability to discover the new sequence {ur}.
*Ex. 19. Show that
r(r + 1) (r + 2) = i[r (r +1) (r + 2) (r + 3)— (r — 1) r(r + 1) (r + 2)]
and deduce that
n
E r(r +1) (r + 2) = in(n +1) (n + 2) (n +3).
r=1
140
4] FURTHER SUMMATION
Proof of (ii):
r(r +1) (r +2) = r3+3r2 +2r
and so r3 = r(r + 1) (r +2)— (3r 2+ 2r)
= r(r + 1) (r +2)-3r(r+1)+r.
n n n n
Thus r3 = r(r + 1) (r+2)— 3 r(r+1)+ r
r=1 r=1 r=1 r=1
Example 4. Evaluate
(r +1) (n+ r).
r=1
(r+1)(n+r) = E (r2+r(n+1)+n)
r=1 r=1
Exercise 8(c)
1. Sum the following series:
(i) 12 +22 +32 +...+122 ; (ii) 13 +23 +32 + ...+ 122 ;
(iii) 32 +42 +52 + ... +202; (iv) 12 +32 + 52 +72 + + (2n — 1)2.
2. By expressing r(r + 1) (r +3) in the form r(r +1) (r + 2) + r(r + 1), sum the series
r(r + 1) (r +3).
r=1
4. Find (rs+r+2).
r=1
142
4] FURTHER SUMMATION
n
6. Find (r2 + 2r)
r=1
n n
(ii) by expressing it as E (r + 1)2 — E 1.
r=1 r=1
Hence evaluate
r -1(r + 1) + (r+ 3).
r —1
8. Evaluate
r=1 r(r +1) (r+2)(r+3).
1
9. Evaluate E
r=1 r(r+ 1) (r+ (r + 4)'
10. Find the sums to n terms of the following series:
(i) 1.4+4.7+7.10+10.13+...;
(ii) 12 .2+22 .3+32 .4+42 .5+...;
(iii) 1.2.3+4.5.6+7.8.9+10.11.12+...;
(iv) 12 .2.3+22 .3.4+32 .4.5+42 .5.6+....
11. Show that
x(x+ 1) (2x + 1) A x(x + 1) (x + 2) (x+ 3) + B(x —2) (x — 1) x(x + 1)
for certain constant values of A and B, and find these values. Hence find the sum
of the first n terms of the series
1.2.3+2.3.5+3.4.7+4.5.9+.... (0 & C)
12. Prove that
(2n+ 1) (2n + 3) (2n-1) (2n+ 1)_ 2(2n + 1)
(n+1)(n+2) n(n + 1) — n(n + 1) (n + 2) .
Find the sum of the first n terms of the series
3 5 7
+ + + (0 & C)
1.2.3 2.3.4 3.4.5
n
13. Deduce r2from the identity (2r + 1)3 — (2r — 24r2 +2.
r=1
n 1
14. Evaluate
r=2 1'(r2 1)•
15. Prove that the sum of the squares of the first n even numbers exceeds the
sum of the squares of the first n odd numbers by n(2n + 1). Hence, or otherwise,
find the sum of the squares of the first n odd numbers. (0 & C)
143
FINITE SERIES AND THE BINOMIAL THEOREM [8
1 2 1
1 3 3 1
1 4 6 4 1
1 5 10 10 5 1
Fig. 8.2
Each line of Pascal's triangle starts and finishes with a 1; terms in be-
tween are obtained by adding together the two terms on either side of it in
the row above.
*Ex. 23. Supply the sixth, seventh and eighth lines of Pascal's triangle.
Ex. 24. Check that the expansion of (x+a)6satisfies conditions (i), (ii) and (iii)
above.
Ex. 25. Show that, if we add the 'zeroth' line, 1, to Pascal's triangle, the expan-
sion of (x+ a)° satisfies conditions (i), (ii) and (iii).
By direct verification as above, we may show that conditions (i), (ii) and
(iii) hold for all suitably small values of n; the Binomial Theorem can be used
to show that they are true for all positive integers n. However, before
proceeding to the general theorem, we shall solve an example by means
of Pascal's triangle.
144
THE BINOMIAL THEOREM
and the last three terms will not affect the first six significant figures.
Direct calculation of the line above shows us that
(x + a)n = xn xn-2a2+ (
n
) x n-rar __Fan
2
(iii) Since
in-1\ + in-1\ (n-1)! 4. (n-1)!
kr - 1 ) k r ) (r -1)! (n- r)! r! (n- r -1)!
(n-1)! [1 11
= (r-1)!(n-r-1)! n-r +r
n!
20 . 19
= 1 + 20( - 2x) + (4x2) ...
1.2
= 1- 40x +760x2 ....
Exercise 8(d)
1. Write down the expansions of
(i) (x -31)6 ; (ii) (2x+4y)6 ;
(iii) (1 +2x)7; (iv) (2x+6y)6.
2. Find the coefficient of x2in the expansions of each of the following expres-
sions:
(i) (3x+1)'; (ii) (2-x)7; (iii) (1 +2x)'2; (iv) (3 -4x)6.
3. Write down the first three terms in the expansions of each of the following
expressions in ascending powers of x:
(i) (1 +2x)8; (ii) (1+-1x)16 ; (iii) (2-x)10 ; (iv) (3x-y)7.
4. Write down the general term in the expansion of
2x-1)8.
(
x
146
5] THE BINOMIAL THEOREM
Hence determine the constant term in this expansion (that is, the term not
involving x).
5. Determine the constant term in the expansion of
(x2 _- 1 )6
3x
6. Find correct to five significant figures, the values of
(i) (1.01)8; (ii) (0.998)8; (iii) (9.99)5 ; (iv) (2.98)5.
In each case, justify the accuracy of your result.
7. Find, without using tables, the value of
(2 + V5)4 + (2 — V5)4.
8. Write down the coefficient of a3x7 in the expansion of (a+ x)". Find the co-
efficient of a3b4c3in the expansion of (a+ b + c)".
9. Expand in ascending powers of x, as far as terms in x2, each of the following
expressions:
(i) (1 +x+ x2)4; (ii) (1 — x + x2)8; (iii) (1 + 2x — x2)5; (iv) (2— x— 2x2)6.
10. If a device measures linear dimensions with an error of less than 4%, find
the possible percentage margin of error if it is used to measure the volume of a
cube.
11. Find the sum of the coefficients of the powers of x (including x°) in the ex-
pansion of (1 +x)n.
If (1 + x + x2)4 = ao + x + a2 x2+ ...+ a8 x8, evaluate
8 8
(i) E ar; (ii) E (— 1)*
r =0
Miscellaneous Exercise 8
1. Find (i) the sum, (ii) the product, of the first n terms of the geometric sequence
{l, 3, 9, ••.}.
2. The first term of a geometric series is 18 and the sum to infinity is 20. Find
the common ratio and the sum of the first six terms. Find also in its simplest form
the ratio of the nth term to the sum of all the subsequent terms of the infinite
series. (J.M.B.)
147
FINITE SERIES AND THE BINOMIAL THEOREM [8
r=p
6. Numbers C1, C2, C3, ... are defined as follows:
C1= 2, = 2(C1+ C2
for n = 1, 2, .... Find C2, C3 and C4 and prove that the numbers C2, C3, C4, ...
form a geometric progression.
Find the sum of n terms of the series
C1+C2 +C3 +.... (0 & C)
7. The numbers ao, ai, ..., an; co, cn_1are the coefficients in the two ex-
pansions
(1 + x)" = ao + ai x+ aox2 + + an x",
(1+ x)n-1= co + ci x+ co x2 + +
Prove that:
(a) ao = co and an = cn-1;
(b) c cr = ar for 1 < r < n — 1;
(c) rar = ncr_i for 1 r < n.
Use these results to show that
r+ 1
ar — ar-i+ ao =
for0<.r<n— 1. (0 & C)
8. Integers a, b, d are connected by the relation a = b+ d. By using the binomial
expansion of (b + d)n where n is a positive integer, show that a' — bn-1(b + nd) is
exactly divisible by d2.
Replace b by a — d in this result, and hence show that if a is the first term,
d the common difference and 1 the nth term of an arithmetic progression, then
an— (a— d)'i-11is exactly divisible by d2.
Show that 518 -214is exactly divisible by 9. (Cambridge)
148
5] MISCELLANEOUS EXERCISE 8
9. Prove that
1 +2+3 +...+n = -1-n(n+1).
Prove also that
12 — 22 + 32 — 42 + + (2n + 1)2 = (n + 1) (2n+ 1).
Show that, when m is any positive integer, odd or even, the sum of the first m
terms of the series 12_22 + 32_42 +...
is ( lm(m + 1). (O & C)
10. Prove that (V2+ 1)2a = AA/2 + B, where A and B are integers. Prove further
that B2 — 2A2 = 1.
1+(l+a)+(l+a+a2)+...+(1+a+a2+...+an-1) = a(a
(a-1)2 a 1 U —
13. Find, in binary form, the square of the number whose expression in binary
notation consists of the r digits 1.
By putting m = n, deduce that the sum of the product in pairs of the first n
integers is
n(n2—1) (3n+ 2).
15. Verify that r2(r + 1)2 — (r — 1)2 r2= 4r3
and deduce that 13 +23 + ... +n3= -102(n+ 1)2.
Prove that 13 + 33 + 53 + + (2n — 1)3 = n2(2n2 — 1)
and find the sum of the cubes of all odd numbers less than 50 that are not multi-
ples of 5.
149
FINITE SERIES AND THE BINOMIAL THEOREM [8
150
Revision exercise A
1. If the expressions
x3 — (a+ 2) x + 2b and 2x3 + ax2 — 4x — b
have a common factor x+ 3, find the values of a and b, and find a second com-
mon factor. (0 & C '0')
2. Find the mid-point of the line joining the points (-4, 2) and (2, — 6), and
show that the equation of the perpendicular bisector of this line is 3x— 4y = 5.
Prove that the point (3, 1) is one vertex of a square of which the points (-4, 2)
and (2, — 6) are opposite vertices, and find the coordinates of the fourth vertex.
(0 & C '0')
3. You are given that the equation
3x5 -9x+5 = 0
has a root approximately equal to 1. By substituting x = 1 + h, and neglecting
h2and higher powers show that x = 11 is a closer approximation to this root.
4. (i) The fourth term of an arithmetic series is 55, and the tenth term is 45.
Calculate the sum of all the positive terms.
(ii) Find, correct to one decimal point, the sum of the first fifteen terms of the
geometric series
1 +(1.02)+ (1.02)2 + (1.02)2 + (0 & C '0')
5. In how many distinct ways may the letters of the word SYZYGY be arranged?
In how many of these do the three Ys appear together ?
(What does the word syzygy mean ?)
6. Without using tables, prove that tan 60° =
A rectangle ABCD lies in a horizontal plane and DE is a line of length h
drawn vertically upwards from D; EA, EB, EC are joined. If the lines AE, CE
make angles of 60°, 45° respectively with the horizontal, express AD and CD in
terms of h, and hence calculate the angle which BE makes with the horizontal.
(0 & C '0')
7. Find the equation of the line of gradient —1 which passes through the point
of intersection of the lines x + 3y— 5 = 0 and 2x —y— 7 = 0.
8. If x2+2x + 2 is a factor of x4 + axe + b, find a and b. Express x4 + 16 as the
product of two quadratic factors whose coefficients are real numbers.
9. ABC is a triangle: points Q, R are taken on AC, AB respectively so that
AQ = 2QC, 2AR = RB.
Express AB, AC in terms of RQ, BC.
10. If in. < x < 71 and sin x = I, find (i) cos x, (ii) tan x, (iii) cosec x.
151
REVISION EXERCISE A
22. The set S consists of all points in a plane whose coordinates relative to a
given pair of perpendicular axes are both integers (in the units used). If A, B, C
are members of S, prove that the area of the triangle ABC is a rational number
of square units.
23. Eliminate 0 between the pair of equations
a sec 0+ b cosec 0 = c,
a' sec 0 — b' cosec 8 = c'.
24. Solve the inequality
x —1
< 2
x2-3x-4
and illustrate your answer by sketching the graph of
x- 1
Y = x2 — 3x — 4
n
25. Evaluate E r(r +1) (2r +3).
r =1
Find the position vector of the point of intersection of this plane with the line
r = 3i+ 5j 5k+ v(2i + 3j k).
— —
The errors in the three cases are respectively E1, E2, E3. Find a relation connecting
E2, E3.
153
REVISION EXERCISE A
31. A bag contains three white and four red balls, another bag contains four
white and three red balls. One of the bags is selected at random and two balls
are drawn from it. If both balls are red, what is the probability that, of the five
balls remaining in the selected bag, two are red?
32. A and B are two points with position vectors a, b relative to a given origin 0.
Prove that the equation of the internal bisector of the angle AOB is
r = A(A + 6).
a
(Recall that is a unit vector in the direction of a.)
What is the equation of the external angle bisector?
Find the position vector of the incentre of the triangle ABO in terms of
a = lad, b = obi, c = IABI and the unit vectors a, b. (The incentre of a triangle
is the point of intersection of the bisectors of the angles of the triangle.)
33. Evaluate the expressions
(i) ,/[1 +.,/{1 +,/(1 + ...)}]; (ii) .,/[2 + ./{2 +.,/(2 + ...)}].
If n is an integer, when is an expression of the form
.,/[n + ,l{n+ V(n+ ...)}1
a rational number ?
34. Two cards are missing from a pack. If three cards are drawn at random from
the remaining fifty (without replacement) and all are found to be aces, what is the
probability that, if a fourth card is drawn, it will also be an ace?
154
9. Mathematical induction
Example 1. Prove that x2+ y2+ z2— yz— zx— xy 0 for real x, y, z.
E = x2+ y2+ z2— yz— zx— xy
2E = (y — z)2+ (z — x) + (x — y)2
2E 0, for real x, y, z,
E 0, for real x, y, z.
The reader must always be on his guard, when writing p q, that this
is indeed what he means: a genuine deduction from the immediately pre-
ceding statement is implied. One of the principle sources of error among
6 PPM 155
MATHEMATICAL INDUCTION [9
Ex. 1. `If —1 > 0, then (-1)2> 0, which is true, and so —1 > 0.' Comment.
Ex. 2. A class is asked to prove that, if a/b =cid, then
(a + b)I(a— b) = (c + d)I(c— d) (a, b, c, d unequal and non-zero).
Criticize the following proof:
`a+bc+d
b=
c—d (a + b) (c — d) = (a — b) (c + d)
a—
ac+bc— ad— bd = ac+ ad— bc— bd
2bc = 2ad
a/b = c/d
which is true and so the original proposition is true.'
Ex. 3. Criticize the proof of the identity cosec2 A tan2 A-1 E tan2 A given below :
`If cosec2 A tan2 A-1 E tan2 A, then
1 sine A
= 1 + tan2A -i- sec2 A
sine A cost A
which is manifestly true, since
1
= sec A.'
cos A
Ex. 4. In Example 2, how do you know that x = 4 is the only root of the equation
V(2x + 1) + Vx = 5?
Ex. 5. Solve Question 6, Exercise 1 c, using the implication signs, and explaining
carefully which steps are not reversible.
Exercise 9(a)
1. Disprove the converse of the result 'if a quadrilateral is a rectangle then its
diagonals are equal'.
2. All angles 0 of the form 0 = kir (k an integer) have the property that
sin 0 = tan 0. Is the converse result true?
3. If a, b e Z+ and if the operation o is defined by
ab— a— b
aob
a+1 '
prove that aob = bo a = a = b.
159
MATHEMATICAL INDUCTION
4. Solve the equation V(x— 4)+ 2 = V(x + 12), stating which steps in your
solution are reversible and which are not.
5. Solve the equation Al(2x +1)— Aix = 1, stating which steps in your solution
are reversible and which are not.
6. Solve the equation V(2x — 1) + V(x + 3) = 9, stating which steps in your solu-
tion are reversible and which are not.
7. Explain the fallacy in the following argument:
`n2 +1 has no real factors. Put n = 3: thus 10 is a prime number.'
8. Prove that the difference of the squares of two odd numbers is divisible by 8.
State and prove a converse theorem.
9. Prove that (x+h) is a factor of axe + bx+ c if and only if ah2— bh+ c = 0.
10. Prove that, if a * —1, then the simultaneous equations in x, y:
ax+y = b,
lx+ay = b
can be solved for x and y.
What happens if a = 1?
Is it true that the equations cannot be solved if a = —1?
11. Two well-known theorems of number theory are:
(i) any odd prime of the form 4n + 1 is expressible as the sum of the squares
of two integers;
(ii) any odd prime of the form 4n + 3 is not expressible as the sum of the
squares of two integers.
Combine these two theorems into a single theorem by using the phrase `if and
only if'.
12. a, b, c, d are four unequal, positive numbers. Prove that
a—c (a2 +c2 a c
= if and only if =
b— d b2+d2)
13. If P(x) is a polynomial with a repeated linear factor, that is, if
P(x) = (x— h)2Q(x),
where Q(x) is a polynomial, prove that (x— h) is also a factor of P'(x). (P'(x) is
the derivative of P(x).)
14. Prove that, if 0 < x < 1,
4 — 3x
then (1 — x)1 < .
4—x
15. ABC is any triangle and M is the mid-point of BC. Prove that
AB2+ AC2= 2AM2+ BM2+ CM2(Apollonius's Theorem).
Is it true that, if M is a point on BC such that
AB2+ AC2= 2AM2+ BM2+ CM2
then M is the mid-point of BC?
160
1] A NOTE ON MATHEMATICAL PROOFS
2. MATHEMATICAL INDUCTION;
AN INTRODUCTORY EXAMPLE
Consider the series
1 1 1
1.2+ 2.3+3.4 +—
Suppose we define the sequence {sn} in such a way that snrepresents the
sum of the first n terms of this series. Then by direct computation,
S1 = S3 = 4, S4 =
The form taken by the first few values of s7, suggests that
n
sn (n c Z +) .
n+1
Let us define a new sequence tfrj in such a way that
n
fn_ n +1'
Then we wish to prove that the two sequences {sn} and {fn} are the same,
in the sense that sn = fnfor all n E Z.
As a first step, we observe that
n+1
fn+i= by the definition off„
n+2'
= 1
n+1 (n+1) (n+2)
1
= (1)
Also,
1
Sn+i = Sn n+2)' by the definition of sn. (2)
±(n+1)(
161
MATHEMATICAL INDUCTION
(1) and (2) define recurrence relations (see chapter 8.1); with the initial
terms they enable us to determine uniquely any term of either of the
sequences {sn}, {M. But we further observe, in this case, that
(i) the two recurrence relations are the same;
(ii) sl = z =
Since the recurrence relations (1) and (2) develop successive terms of the
sequences {s„} and {fn} in the same way, and since the two sequences
have the same starting point, we deduce that sn = fnfor all n.
The observation that (1) and (2) represent the same recurrence relation
enables us to deduce that if sn = fn, then sn+, = or, expressed
alternatively,
sn = fn Sn+1 =fn+
1.•
The fact that s1 = flthen enables us to infer the equality of s„ and fn for
all n E Z± ; for
s1 =f s2 = s, =f3 •
Example 5. Prove that 13n— 6n-2is divisible by 7 for all positive integers
greater than m, where m is a certain integer, to be specified.
Write un = 13" — 6n-2; we have to prove that unis divisible by 7 for all
n m, where m is some integer, to be found.
(i) u1 is a fraction, and the concept of divisibility does not apply.
yr, = 132 -6° = 168 = 7.24 and u2is therefore divisible by 7: the induction
starts.
Let pnbe the statement `unis divisible by 7'.
163
MATHEMATICAL INDUCTION [9
+ = xn + a+ 2 xn-2a2+ + xn-rar + + an
1
for positive integral n.
1 1
(i) (x + a)1= x + a = ( ) xl + x°a1and the induction starts.
1
(ii) Consider the product
\ n-r a? + ... +ani .
(x+a) [xn+ (1) xn-i a+ ( ) xn-2 a2 + ... + ( ) x
2 r
Each term in the product is of degree (n+ 1), the first is xn+1, the last is
an+1and the term containing xn+l-rar is
x xn-rar + a ( n ) xn±'-rar-1
r —1
F
= / + (r —n 1)1 xn+l-rar
= (n + 1) xn+l-rar, by Vandermondes
r Theorem (see Chapter 8).
Thus
(x+a)[xn + (7) xn-ia+ ( ) xn-2a2+ ... + (n) xn-rar + ...+an]
2 r
= xn+i+in+ 1\ xna+ in+ 1)x
n-1a2 + ... + I n+ ) xn+l-rar + ...+ an+1
k 1 ) k 2 k rl
and hence
(x+a)n = xn + (11) xn-ia+ ( ) xn-2a2+ ...+ ( ) xn-rar + ... + a"
1 2 r
1) xna+(11+1\ xn_1a2 4....
(x+ a)n+1= xn+1+ ( fl +
1
k 2J
+ in + 1) xn-a-rar + ...± an+1
k r
and the proof is complete, by mathematical induction.
164
3] METHOD OF MATHEMATICAL INDUCTION
Pr U Eils) = E Pr(E,IS).
(i=1 i=i
165
MATHEMATICAL INDUCTION [9
(i) The result is certainly true if n = 2, by Theorem 7.2, and the induc-
tion starts.
k-1 k-1
(ii) Write F = U Eiand suppose that Pr (FIS) = E Pr (EilS).
i=1 i=1
Now F n Ek = rb, since the events are mutually exclusive. Thus
Pr (F U EkIS) = Pr (FI S)+ Pr (Ek IS), by Theorem 7.2
k-1
= E Pr (Ei lS)+ Pr (Ek i S), by the assumption made above
i=1
k
= E Pr (Ei lS)
i=i
and the theorem is proved, by mathematical induction.
Exercise 9(b)
Throughout this exercise, It represents a positive integer. Prove the results 1-14.
using the principle of mathematical induction.
1. 1 +2+3+4+...+n = in(n+1).
2. 1+3+5+...+(2n-1) = n2.
3. 12 +22 +32 +42 +...+n2 = in(n +1) (2n +1).
4. 13 +23 +33 + ... +n3= 4n2(n +1)2.
5. 1.1+ 3.2+ 5.4+ + (2n - 1).2n--3= 3 +2'(2n- 3).
6. 3.1!+7.2!+13.3!+...+(n2 +n+1) n! = (n +1)2n! 1. -
1 1 1 1 1 1
+ +
7' 1.2.3 2.3.4 3.4.5 + + n(n +1) (n + 2) 4 2(n+ 1) (n + 2).
1 1 1 1 3 2n+3
8.
13+ 24 + 35+-+n(n+2) = 4 2(n+1)(n+2).
9. n6 -n is divisible by 30.
10. 12n +2.5n-4is divisible by 7.
11. 26n + 32n-2is divisible by 5.
12. 52n +22n-2.3n-3is divisible by 13.
16. Prove that 5n < rz! for all sufficiently large n. Add such precision as you can
to the phrase 'sufficiently large n'.
17. Prove by mathematical induction that the sum of the angles of a convex
n-sided polygon is (2n — 4) right-angles.
18. Prove that an n-sided convex polygon has -in(zz— 3) diagonals.
19. A straight line separates a plane into 2 regions; two straight lines separate
the plane into 4 regions and so on. If n straight lines (no three concurrent, no two
parallel) separate the plane into u„ regions, prove that
un= 1-(na + n + 2).
20. If n is a positive integer, prove that
52n-1-2_ 24n — 25
is divisible by 576. (Cambridge)
1 1 1
21. If f(n) = 1+ — +— + ...+ —
22 32 n2
prove that E [(3r2 + 3r + 1) f(r)] = (n +1)3f(n)— -in(n + 1). (Cambridge)
22. Prove that n
(r+k-1) = n(n + 1) ... (n+ k)
r(r+1)(r+2)
r=1 k+ 1
23. If al, a2, a3, ..., anare all positive, prove that
11 1
(ai+ a2 + + a„) (-+-+ ...+ —) n2.
a1 a2 an
24. Prove that, for any positive integer n,
2n 2n(2n — 2) 2n(2n — 2) (2n — 4)
+ + + ... (to n terms) = 2n.
2n —1 (2n —1) (2n— 3) (2n —1) (2n — 3) (2n — 5)
(0 & C)
25. A motorist estimates that, by travelling along a main road at a certain steady
speed, the probability that the next set of traffic lights will be green if the last
set was green is aand that the probability that the next set of lights will be
green if the last set was red is 4. He sets out one day to test his theory. Prove that,
if the first lights he meets are green and if
pn = Pr (the nth set of lights is green when he reaches them)
then Pn 3(a)n-1 (n % 1).
26. A man repeatedly tosses an unbiased coin, scoring 1 for each head and 2
for each tail. If pnis the probability that his score will ever be n, prove that
Pn = + Dn• —
27. n bags, numbered 1 to n, each contain one white and one black ball. A
ball is taken at random from bag one and placed in bag two; a ball is then taken
167
MATHEMATICAL INDUCTION [9
from bag two and placed in bag three and so on, until finally a ball is taken from
bag n. Prove that the probability that this ball is white is
( )
1
2 1 + 3n-1
168
10. Expectation
1. RANDOM VARIABLES
In Chapter 7 we considered random experiments described by outcome
spaces. With each elementary event we associated a number, its proba-
bility, which denoted our degree of belief that the experiment would result
in that particular events occurring. The elementary events themselves
may be described in various ways: in some cases, it is natural to denote
them by a number, e.g. for the fall of a die we could denote our elementary
events by the numbers 1, 2, 3, ..., 6; but sometimes no such natural numeri-
cal description exists. For example, for a single toss of a coin, the two ele-
ment set {heads, tails) is the natural choice of outcome space, the elements
being labelled by the descriptions ' heads' and `tails'. In order to make a
mathematical analysis of random experiments it is helpful to describe the
possible outcomes numerically even in those cases where no such 'natural'
description exists. Thus, in the case of coin spinning, the event ' tails' could
be denoted by the number 0, the event ' heads ' by the number 1. Such
values are called values of a random variable.
Let us pause here to recapitulate. Suppose we have a random experi-
ment whose possible outcomes are the n elementary events sb s2, sn.
Then we may take as our outcome space the set
S sz, • • •, sn}•
I I 1
.P2=6 A=6 P6=6
1 1 n --
1 1 -1 1
P1 = 3 I '3-6 I P 6-6
P1 P2 PIT
I I I I I I
f t
s 4I 4I I S
I
1
I
_L
I
I
I
1
I 1
1
1 1
stssb
1 s l 530 540 s50 ScEl
Saal
S , 1 1 1 1 1 1
"o 1 1 1 1
11 I 1
I I ■ 1 1
I I 1 I I I I 1
= 1 x3 = 3 I x4 = 5 I
XL X2 Xn x2 = 2 x4 = 4 xa = 6
Fig. 10.1 Fig. 10.2
r, 1 5
ri= -6 P2 = 6
I I
I I
—Not a
Double double
x1= 1 X2 = 0
Fig. 10.3
1 1
P2 = 4 FL =
1 1
f1 = 74 F3 = 74
I
ItI tI
1
s1 1 $2 I 3 3 I 34 1
Th :1=1 r =1
1 1 I
xi. =-I-5 x3 = 0
x2 =0 x4 = 5 -
Fig. 10.4
170
1] RANDOM VARIABLES
cerned with the financial outcome and our choice of random variable
reflects this concern. If the experiment is repeated a number of times,
the sum of the values of the random variable obtained denotes A's net
gain in pence.
With these examples behind us we now make the following definition:
a variable whose value is a number determined by the outcome of a random
experiment is called a random variable.
The discerning reader will have observed that a random variable is a
function from a given sample space into a set of numbers, the values of
the random variable being the images under this function. The words
`random variable', however, are often loosely used to denote the image—
a vice to which we ourselves shall succumb from time to time for the sake
of brevity.
In the same way that a function is denoted by f and the image of z
under./ by f(z), so a random variable may be denoted by X and its associ-
ated value for the ith element of the outcome space by xi.
Ex. 1. A man pays 10 pence to throw two dice. For any double he receives his
stake momey back, together with a prize: one pound for a double six and 20
pence for any other double. Suggest a suitable outcome space, probability dis-
tribution and random variable for this experiment.
Ex. 2. A man insures his life for £1000, paying a premium of EX. Suggest a
suitable choice of random variable to describe the situation.
2. EXPECTATION
A mass of numerical data often has an indigestible appearance and conveys
very little, unless subjected to considerable analysis. A device commonly
used to give an overall impression of the data is to determine their average
value. For example, if the heights of a hundred boys are measured, the
complete set of results may usefully be characterized by giving their average
height.
We shall now investigate what meaning can be attached to the phrase
`the average value of the random variable X associated with a given
probability distribution'. If an experiment is repeated N times, where N
is a large number, our interpretation of the probability p, as a measure
of our confidence in securing the outcome s1leads us to expect roughly
pi N occurrences of si, with similar results for s2, s3, etc. Thus we should
anticipate a score of x1 on IAN occasions, x2 on IAN occasions and so on,
giving an approximate average value of our random variable X for all N
171
EXPECTATION [10
repetitions as
Nxi +p2Nx2+ +p„Nx,,
N = Pixi+P2x2+ • • •+Pnxn
n
= 1=1
E Pi xi.
The larger the value of N, the more confidence we should place in the value
of E pi xias an estimate of the average. We are thus led to formulate the
i=
following definition:
The expectation e(X) of the random variable X is defined by the equa-
tion
e(X) = pi xi•
S(X) = ‘
r
1-1 u
= 21
= 3.5.
It is perhaps surperfluous to remark that no one sufficiently familiar
with an unbiased die would expect a score of 3.5 on any one throw. 3.5
simply represents the best estimate available, prior to the actual experiment,
that we can make of the final average score if the die is thrown a number of
times; the larger the number of throws, the better we expect our estimate
to be. (The reader is strongly recommended to test the accuracy of this
forecast if ever he finds time lying heavily on his hands by throwing a die,
say a hundred times, and computing his average score.)
Example 2. Two players, A and B, play the following game with three coins:
A pays a stake of 10 pence and tosses the three coins in turn. If he obtains
three heads, his stake is returned, together with a prize of 30 pence; for two
consecutive heads, his stake money is returned, together with a prize of
10 pence. In all other cases, B wins the stake money. Is the game fair?
We must first consider what is meant by the question `is the game fair ?'
Intuitively it seems reasonable to label a game as ' fair ' if, in the long-run
neither side anticipates any considerable financial gain. Mathematically,
a game between two players is fair if the expectation of gain for either
player is zero.
172
2] EXPECTATION
On the assumption that the coins are unbiased and that the results of the
tosses are independent, Pr (HHH) = (Z)3= 1; similarly,
As our random variable we take the net gain, in pence, for A on each
particular elementary event: x, = +30, x2 = + 10, x3= + 10, x4= —10.
Summarizing in tabular form, we have:
S Anything
HHH 'HHT THH
else
P 1 1t. 1 i
X 30 10 10 —10
Ex. 3. An experiment with three possible outcomes sl, s2, s3, has probability
distribution {4, -4-, 4} and associated random variables 3, 2, 1. Determine the
expectation.
Ex. 4. A man pays 1 penny to throw three unbiased dice. If at least one six
appears he receives back his stake money together with a prize consisting of the
number of pennies equal to the number of sixes thrown. Does he expect to win
or lose?
173
EXPECTATION [10
Exercise 10(a)
1. The values of a random variable, together with their associated probabilities
for four different experiments, are given in the tables below. Calculate AX) in
the four cases.
(i) xi 0 1 2 3 4 5
p, A- A- 1 IAA
xi —2 —1 0 I 2
pi Ai AAA
x; 1 2 3 4 5 6
(iv) xi 1 2 3 4 5 6 7 8 9 10
pi 0 a 0 1 0 1 0 1 0 1
4. Two dice are thrown; find the expectation of the higher score showing (or
the score of one of them, if they fall alike).
5. If the probability that a man aged sixty will survive another year is 0.9,
what premium should he be charged for a life insurance policy of £1000? (If he
survives the year, he receives no money back.)
6. X1and X2 are two random variables, each with values 0, 1, 2, 3, ..., 9, and
each possessing a uniform probability distribution. Evaluate
(i) g(X,— X2); (ii) I Ari- x2I).
7. Two bags each contain ten coloured discs as shown.
174
2] EXPECTATION
A player stakes a certain sum of money for the privilege of drawing two discs,
one from each bag. For two discs of the same colour his stake is returned and, in
addition, he is awarded a prize of 10p for two reds, 20p for two greens and 25p
for two blues. For two discs of different colours he loses his stake.
Show that, if the stake money is 8p, he can anticipate gaining in the long run,
but that with the stake at 9p he should expect to lose.
8. The game of Question 7 is repeated, but the player now tosses a coin to decide
which bag he must choose from: if he tosses a head, he chooses bag I, if a tail,
bag II; he then draws a disc at random from the chosen bag, notes its colour and
replaces the disc. He repeats the process again and is paid prizes as in the pre-
vious question. Determine the minimum stake (to the nearest penny) required to
ensure that the player will show a loss in the long run.
9. The game of Question 8 is repeated but the discs are not replaced between
draws. Determine the minimum stake (to the nearest penny) required to ensure
that the player will show a loss in the long run.
10. The game of Question 7 is repeated, but the player is now required to place
his stake after the result of the first draw from bag I is known. If he draws a red
disc first time he pays 2p, if a green, 9p, if a blue, 12p. Show that, in the long run,
the player expects to win.
Show further that the above stakes are the fairest available that still give an
advantage to the player, in the sense that, if any one of the stakes were increased
by 1p, the bank would then expect to win.
11. A man pays a stake to throw two dice. If he scores a total of 3 or 11, he
receives 40p. For a total of 5 or 9 he receives 20p, and for a total of 7, 10p (in
each case the stake money being returned too). He loses his stake money for any
even score. Show that he expects to win if the stake money is 21p, but that, if the
prizes for scoring 5 or 9 and 7 are reversed, he would then expect to lose.
12. Two identical bags contain respectively (i) four fivepenny pieces and twelve
tenpenny pieces, (ii) nine fivepenny and seven tenpenny pieces. You are allowed to
select a bag and draw a coin at random from it. If the coin you draw is a five-
penny piece, what would be a fair price for you to offer for the bag you did not
select ?
13. The path in Figure 10.5 represents a simple maze along which a rat is made
to run. It starts at S and has to finish at F. If it makes a mistake at A by turning
along AA' it will return to A and be forced by the construction of the maze, to
turn towards F, and similarly at each of the other junctions. The probability of
taking either of the two paths available at each junction is -. Find the expected
number of mistakes the rat will make in running from S to F.
A' B' C' D'
S F
0 A BCD
Start Finish
Fig. 10.5
14. What is the expected number of moves that can be made by (i) a bishop,
(ii) a knight, placed at random on an empty chess board?
175
EXPECTATION [10
15. Two dice are thrown in 'one turn', each turn costing 5p. If a prize of 40p is
given for a double six and a prize of 20p for any other double (together, in both
cases, with the stake money), determine the loss to a person playing the game
one hundred times.
16. A man puts three coins into a bag, deciding at random for each coin sepa-
rately whether it is to be a fivepenny or a tenpenny piece. Calculate the expected
total value of the coins he puts in his pocket.
17. A man puts three £5 notes into one envelope and three E1 notes into a simi-
lar envelope. Each year at Christimas, he chooses one envelope at random and
gives his nephew a note from it. As soon as either envelope is emptied by his
taking the last note from it, the process ends.
(i) State the different totals which the nephew may have received when the
process ends;
(ii) for each of these totals calculate the chance of its occurrence;
(iii) deduce that the nephew's expectation of gain is £12.375.
(Cambridge adapted)
= e(Z)
= Pi(Xi Max)2.
i=1
The positive square root of the variance, cr„ is called the standard
deviation of X; it has the advantage of being measured in the same units
as X. The standard deviation (or, equivalently, the variance) has been
shown, in an intuitive sense, to give a measure of spread; its deeper signifi-
cance will be appreciated by the reader only as he gains a fuller grasp of the
subject, when it will be seen that standard deviation plays a central part in
the development of probability theory.
There is no difficulty in extending the concept of expectation to other
functions of a random variable. Indeed, given a function f that maps the
random variable X into another random variable Y, where yi = f(x,),
177
EXPECTATION [10
6 .fiX)] = E Ptflxi);
i=1
with this notation cr = eq(x—it.)2].
Ex. 6. Find the mean and variance of the random variable X whose probability
distribution is shown in the table below:
X 0 1 2 3 4
Pi III I
Ex. 7. Find the variance for the number of heads showing if three unbiased coins
are tossed.
*Ex. 8. A random variable has mean # and variance a-2. If a constant c is added
to each value of the random variable, describe (without embarking upon any
detailed working) what difference this will make to (i) the mean, (ii) the variance.
Example 3. Find the mean and variance of the numbers 1, 2, ..., 20, assuming
that they are uniformly distributed.
We use Theorem 10.1, with a = 10,
e[(y-10)] = -216{-(9+8+...+1)+(1+2+...+9)+10]
= 0.5,
p = 10.5;
e[(x.- 10)2] = 26 [2(12+22+ ... +92) + Hp]
= -216[1.9.10.19+100]
= 33.5,
o.2 = 33.5 — (0-5)2
= 3325.
The next theorem determines the mean and standard deviation of a
linear function of the random variable. Intuitively, the results are easy to
understand: multiplication of every value of X by A clearly results in a
proportionate increase in the measures of the average (mean) and spread
(standard deviation), while a shift of origin shifts the average by the same
amount but has no effect upon the spread.
Theorem 10.2. If A, a are any numbers:
(i) the mean of the random variable Y = AX+ a is Ap,+ a,
(ii) the standard deviation of the random variable Y = AX +a is Ao-.
Proof (i) ‘[AX + a)] = E pi(Axi+ a)
i-----1
=A E pi xi+a E pi
1:=1 1:=1
= Alt+a.
179
EXPECTATION [10
= A2 EPi(xi -/L)2
i=i
= A2o2
Our final theorem of this section is a result due to the great Russian
mathematician P. L. Chebyshev (1821-94). Its importance lies in the fact
that it gives us an interpretation for standard deviation in terms of prob-
abilities; that is, it gives us, in terms of A, an upper limit for the probability
that a value of the random variable lies further away from the mean than
A standard deviations, whatever the associated probability distribution.
(Of course, if we are given some information about the probability dis-
tribution we can usefully refine the inequality, but this does not detract
from Chebyshev's result viewed as a general theorem.)
180
4] SOME THEOREMS
Ex. 9. A probability distribution has mean 3 and variance 0-2 = 1-2. Give an
approximate limit for the probability
Pr (lx-31 > 2u).
The probability distribution for the random variable X (X = 1, 2, ..., 5) is
given in the following table:
X 1 2 3 ' 4 5
P A 1 4 1 A
Find the mean and standard deviation. What is Pr (Ix-31 > 2a)? Comment
upon the results obtained and any discrepancy you observe.
Ex. 10. Find an expression for g[(X—,u)3] in terms of e(A-3), p and u.
Theorem 10.5. If the random variable X, whose p.g.f. is G(t), has mean ,u
and standard deviation o, then,
(i) G(1) = 1;
(ii) G'(1) = it;
(iii) G"(1) = a2 + ft2 -
= E r 2 pr— r=1
E rpr
r=i
= 1(X2)- 1(X)
= (Cr2 /42) -it, by Theorem 10.1, with a = 0.
Example 4. Two unbiased dice are thrown. Find the expectation of the total
score, and its variance.
Let the random variable X denote the score from the first die; the ran-
dom variable Y the score from the second die. Then we have to calculate
the expectations and variance of the random variable Z = X+ Y. Now
Gz(1) = *(ti+t2+ + te),
GAO = 6 (t1+0+ ... +16),
.*. Gz(t) = -P6(t1+ t2+ t6)2,
Ex. 11. Three coins are tossed and the number of heads noted. Find the p.g.f.
for the resulting distribution and determine pand 0-2.
Pr (X = r) =
Notice that E Pr (X = r) = = 1,
r=0 1 11
2
which is as it should be. Indeed, all the results of Chapter 7 hold for such
outcome spaces; detailed proofs will be omitted, because they depend upon
the manipulation of convergent series. However, assuming the obvious
generalizations of the definitions and theorems, the calculations involved
are not difficult. For instance, in the example above
6r(X) = r.2
1 .
1
2r+
But r = (1 — X)-1(IXI < 1)
r=0 X
Since the game can go on indefinitely, X may take any positive integral
value 1, 2, 3, ... 12 1)r-1 1
Pr (X = rIA wins) = 33' 3
9 (1y-1
and 6'(X) = E
r=1
7 1
= 9• — D2
— 9
Ex. 12. A man tosses a coin until a head appears and is paid a number of pence
equal to the number of tosses he makes. What is his expectation?
Exercise 10(b)
1. X is a random variable with mean it and standard deviation a. Write down
the means and standard deviations of the following random variables:
(i) — X; (ii) X+1; (iii) 3X-1; (iv) (X—#)/(r.
2. Calculate the mean and variance for each of the following distributions:
(i) X 0 1 2 3 4 5
P i 1 15 25 1 t
X 1 2 3 4 5 6
X —3 —2 —1 0 1 2 3 4
1
P As 235 4 230 215 110 1-1,3 1
3. Calculate the means and variances for X2in each part of Question 2.
4. A cubical die is so weighted that the probability of obtaining any face is
proportional to the score showing on that face. Find the mean and variance of
the score obtained.
5. If the random variable X can take the values 1, 2, 3, ..., n, all values being
equally likely, calculate the mean and variance for X.
185
EXPECTATION 110
Miscellaneous Exercise 10
1. Two bags contain red and white discs as shown in the table below:
Red White
Bag I 5 15
Bag II 10 10
One of the bags is selected at random and a disc drawn from it proves to be red.
If the red discs are now valued at £1 each and the white discs are valueless, what
would be a fair price to pay for the remaining discs in the selected bag?
2. (The St Petersburg Paradox.) A coin is spun. If a head is obtained first time
you are paid 1p ; if you get a tail followed by a head you receive 2p; for two tails
followed by a head 4p, the next prize being 8p and so on. Show that, however
much you are prepared to pay to play the game, your expected profit will be
positive.
186
6J MISCELLANEOUS EXERCISE 10
Criticise any assumptions you have made and indicate what further know-
ledge you would require before offering a more realistic 'fair price' for the game.
If the banker against whom you are playing starts with a capital of 100p, what
would be a fair price for you to offer him before playing the game?
3. A and B alternately throw an unbiased die, A having the first throw. The game
terminates when a six is thrown. B agrees to pay A £1 if A throws a six on his
first throw, £3 for a six on his second throw, £5 for a six on his third throw, and
so on. What would be a fair price for A to offer to play the game?
4. Two men play a game with two dice. A has a true die, whereas B has a die
which is biased so that each of the even faces is twice as likely to occur as each
of the odd faces. The two players throw their own die and A wins from B the
sum of the numbers thrown when the sum is even and the numbers are unequal.
B wins the sum from A when it is odd; and the game is drawn when the two num-
bers are equal.
Calculate the expectation of the game to A.
5. X and Y are two random variables defined over the same probability dis-
tribution. A new random variable Z is constructed, where Z = X+ Y. Prove
that the mean of Z is the sum of the means of X and Y.
Is it also true that the variance of Z is the sum of the variances of X and Y?
6. If X is a random variable with mean uand standard deviation a prove that
the standard deviation of the random variable (X--#)2 is
{6[(x -F)4] - a4}1.
7. n points are marked on a line at 1 cm intervals. If two of the points are chosen
at random, find the probability that they will be r cm apart. Deduce that the
expected distance apart of the two points is +(n +1).
8. A and B play a game in which A's chance of winning is p, while B's is q, where
p+q = 1. They have a contest, the winner being the first to score two consecutive
successes. Prove that the expected number of games is
(2 +pq) (1- pq)-1 .
9. Xis a random variable such that Pr (X < 0) = 0 and the mean and standard
deviation of X are respectively g and a. Prove that, for any k 1,
k 1
Pr (X < kit)
Deduce Chebyshev's result, that
Pr (it - Aa < X < it+Acr) 1 - j .
10. The random variable Yhas a distribution such that Pr (Y = r) is given by the
coefficient of Br-1in pt(1 - 0q)--4 for r t and zero otherwise, where 0 < p < 1,
q = 1-p. Find the expectation of Y. (M.E.I.)
11. Xis a random variable that can take all positive integral values 0, 1, 2, 3, ....
If CO
pr = Pr (X = r) and qr = E p,
s=r+1
7 PPM 187
EXPECTATION [10
b
Fig. 11.1
7-2 189
FURTHER VECTORS [11
(between 0° and 180°) through which one of the vectors must be rotated to
bring it into coincidence in the same sense as the other vector (see Figure
11.1). It follows that the sign of the dot product is positive if the angle thus
defined between the two vectors is acute and negative if the angle is obtuse.
We have said that the word ' multiplication ' is used in this context
because of the similarity shown between dot products and ordinary pro-
ducts. This similarity we now endeavour to display, but to give added force
to our warning against pressing the analogy too far, we start by observing
two laws of multiplication of ordinary numbers that do NOT hold for
dot products.
Dot multiplication does NOT give closure; that is, a.b is not a vector.
Dot multiplication is NOT associative; indeed, (a .b).c is not even defined,
for (a . b) is a number, not a vector. (On the other hand, we can talk about
(a .b) c, which represents a vector in the direction of c and whose magni-
tude is lad lb I cos 0 times that of c.)
However, there is a sort of associative rule that holds: if a, b are two
vectors and k is a number,
k (a .b) = (k a) .b = a. (kb) .
The proof of this follows immediately from the definitions and may be
left as an exercise for the diligent reader.
Another reassuring note may be struck by observing that the commuta-
tive law holds for dot products
a.b = b . a.
Again, the proof of this is immediate from the definition and may, with
even more confidence, be left as an exercise for the reader.
The final rule for dot products, that of distributivity over vector addi-
tion, i.e.
a.(b+c) = a.b+a.c
is more subtle and the proof is not immediately obvious. However, the
effort expended in demonstrating this result is amply rewarded, for with
it we shall have set up all the apparatus necessary for an effective algebra
of vectors.
We begin by defining the projection of a vector a in a direction specified
by a unit vector u. Suppose the displacement OA represents a; draw the
plane n through 0 perpendicular to u. Then, if D is the foot of the per-
pendicular from A to 7r,
a = OD+DA
= OD +pu
(see Figure 11.2). Further, the expression of a as the sum of two vec-
190
1] THE DOT PRODUCT
tors, one in the plane it and one in the direction u is unique. For suppose
that
a = OD +pu
= OD' +piu,
Fig. 11.2
Now
(b — a) . (b — a) + — . (d — = (c — b). (c — b) + (d — a) . (d — a)
b.b-2a.b+a.a+d.d-2c.d+c.c=c.c-2b.c+b.b+d.d-2d.a+a.a
= a.b+c.d
(c — a) . (b — d) = 0.
But c — a * 0, b — d * 0,
.*. AC _LBD.
Exercise 11(a)
1. Show that a.a—b.b = (a — b). (a + b).
Interpret this equation geometrically and discuss the particular case la I = lb I.
2. If a, b are non-zero vectors such that la + b I = b I, show that a and b
are perpendicular.
3. A is a fixed point with position vector a and c is a constant vector. If the
variable point P has position vector r where r satisfies the equation
(r — a) . (r — a) = c . c
what can be said about the position of P?
4. Prove that b+ c = 0 and lal = Ibl (a — b). (a— c) = 0.
Interpret this result as a geometrical theorem.
5. 0 XYZ is a rhombus of side a; P is any point of 0 Y (or 0 Y produced). Prove
that OP .YP = XP2 — a2.
6. ACBD is a straight line and 0 is a point not on the line. LAOB, LCOD are
both right-angles. If OA = a, OB = b, AC = x, CB = y, BD = z, find the
position vectors (relative to 0) of C and D.
Deduce that
CB. BD _ OB2
AC .AD 0A2*
7. ABCD is a tetrahedron in which AB1CD and AC1 BD. Prove that AD IBC
and that
AB2+ CD2 = AC2+ BD' = AD2+ BC2.
8. If each edge of a tetrahedron is equal to the opposite edge, prove that the
line joining the mid-points of any two opposite edges is at right-angles to each
of these edges.
9. Prove that, if two of the joins of mid-points of opposite edges of a tetra-
hedron are at right-angles, the remaining edges are equal.
10. ABCD is a skew quadrilateral (that is, the vertices A, B, C, D do not all lie
in the same plane). The mid-points of AC, BD are P, Q respectively. Prove that
AB2+BC2+ CD2+ DA2 = AC2+ BD2+ 4PQ2.
193
FURTHER VECTORS [11
Ex. 4. If a = i— 3j + k, b = 2i + 2j — k, find a. b.
Ex. 5. Show that the two skew lines
(x-1 y+1z+1
=
2 3 2 '
x +1 y-2 z
1 2 —4
are perpendicular.
Ex. 6. Find the unit vector perpendicular to the plane A (3, 1, 1), B (2, 5, 0),
C (-1, —1, 15).
Ex. 7. Deduce the two distance formulae of Chapter 3 using dot products.
Ex. 8. Prove vectorially that the angle between two diagonals of a cube is
arccos
A line is specified given a point P1lying upon it and its direction; let
us suppose its direction is defined by the unit vector u = /i +rnj. Here
/ = cos 0, and m = cos 02, where 01and 02are the angles made by the
vector u with the axes Ox and Oy respectively (senses being taken into
account) (see Figure 11.5).
The line through P1(x1, y„) in the direction u thus has the equation
x — x, _y — y, _
r m —
(see Figure 11.5, where 01+ in). Since cos 02 = sin 01, this equation may
be rearranged in the form
sin 0
Y — = cos 011 (x x1)
u2 = Ull ludcos ft
4
= V65
Notice in particular that, from equation (3) the two lines are perpendicular
<r> 412+ = 0.
Provided m1* 0, m2* 0 this second condition may be written
4 /2 = —1.
m1 m2
But, as was noted above, 4/m1and 12/m2are the gradients of the two lines;
we have thus regained, by an alternative method, the result of Chapter 3.4.
Fig. 11.6
197
FURTHER VECTORS [11
Theorem 11.3. The perpendicular distance from the point P2(x2, y2) to the line
ax+by+c =0 is
ax2+by2+c
(a2+ b2)
Proof. The given line may be rewritten in the form
x+ c/a y
b
—a
or, making the denominators components of a unit vector, in the equiva-
lent form
x+ c/a
b (az b2) — aIV(a2+ b2)
Thus the unit vector defining the direction of the line is
b. a
u = v(412 4.b2) 4J(a2
b2)
Fig. 11.7
Ex. 9. Find the perpendicular distances of the point (2, — 1) from the lines
3x + 4y + 3 = 0 and 4x — 3y — 6 = 0. What can you deduce about the position of
(2, —1) relative to the two given lines?
Ex. 10. Sketch the two lines 3x +y— 3 = 0, 3x +y — 6 = 0 and shade in the region
determined by the four inequalities x > 0, y > 0, 3x + y — 3 > 0, 3x + y — 6 < 0.
Ex. 11. Find the equation of the bisector of the acute angle between the lines
x = 0 and 5x + 12y — 60 = 0. Find the incentre of the triangle formed by the
lines x = 0, y = 0, 5x+12y — 60 = 0.
Ex. 12. a, b, c are constants. Show that, as 0 varies, the straight line
x cos 0+y sin 0+acos0+b sin 0 + c = 0
touches a certain circle. Find the centre and radius of the circle.
Ex. 13. Find the equations of the lines parallel to 8x+ 6y+3 = 0, and distant
1 unit from it.
Exercise 11(b)
1. Determine the projection upon (a) the x axis, (b) the line 3x — 4y +1 = 0 of
the line segments joining the following pairs of points:
(i) (2, 3), (4, 4); (ii) (— 1, 2), (1, 5);
(iii) (-1, 2), (2, — 3); (iv) (a, b), (a+ b, a— b).
199
FURTHER VECTORS [11
2. Find the cosines of the acute angles between the following pairs of lines:
(i) x—y+ 1 = 0, 3x—y+2 = 0; (ii) 3x+2y+1 = 0, x-2y-3 = 0;
(iii) 4x — y + 1 = 0, 3x +4y = 0; (iv) ax+ by = 0, ax—by = 0.
3. Determine the perpendicular distances between the following points and lines.
(Leave your answers in surd form.)
(i) (0, 0), x + 4y— 1 = 0; (ii) (-1, 2), 2x— 3y— 4 = 0;
(iii) (-2, —3), x— 3y+ 5 = 0; (iv) (h, —k), hx—ky+2hk = 0.
4. Find the distances between the following pairs of parallel lines:
(i) x —2y +1 = 0, x-2y+3 = 0; (ii) 2x+y— 3 = 0, 4x+2y+ 3 = 0;
(iii) 5x—y-1 = 0, 5x— y+ 3 = 0; (iv) ax+ by + a = 0, ax+ by— b = 0.
5. Find the equations of the pairs of lines, parallel to the following lines, and
distant three units from them:
(i) 3x — 4y + 1 = 0; (ii) 8x + 6y +5 = 0;
(iii) 5x+ 12y + 2 = 0; (iv) 3x + 2y + 1 = 0.
6. Draw rough sketches and shade in the areas determined by the following
inequalities:
(i) x > 0, y > 0, y < 2, 2x + 3y — 8 <0;
(ii) x > 0, y > 0, x—y— 1 < 0, 3x+y-7 < 0;
(iii) 2x— y > 0, x — 2y < 0, 2x— y + 2 > 0, x— 2y + 4 > 0;
(iv) 1x+y1 < 1; lx—Y1 < 1;
(v) + y — 11 < 1; 12x — 3y — 11 < 1.
7. Find the equations of the bisectors of the acute angles between the following
pairs of lines:
(i) x = 0, 3x — 4y = 0; (ii) 3x-4y-1 = 0, 4x-3y-2 = 0;
(iii) 5x + 12y +3 = 0, 6x— 8y + 3 = 0.
8. Find the equation of the image by reflection of the line x—y = 4 in the line
2x+y = 1. (London)
9. ABC is a triangle; A is the point (0, 3) and B is the point (— 5, —2). The
orthocentre (meet of the altitudes) of the triangle is the point (— 1, 1). Find
(i) the coordinates of C; (ii) the area of the triangle;
(iii) the tangent of the angle ACB.
10. Show that the reflection of the point (a, ft) in the line y = mx is the point
((1 —m2)a +24 2ma— (1 — m2) fi\
1+m2 l+m2 f•
A is the point (p, q), B is the reflection of A in the line y = x and C is the
reflection of B in the line y = — x. Find the coordinates of C.
Show that C is the reflection of A in the line px +qy = 0. (London)
200
31 FURTHER TWO-DIMENSIONAL GEOMETRY
11. A triangle ABC lies wholly within the first quadrant and has an area of
4-1- sq. units. The equation of one side is 2x— 5y + 23 = 0 and the vertices A and
B are (1, 5) and (3, 4) respectively. Find the equations of the other two sides, the
angle ABC, and the coordinates of the orthocentre of the triangle. (London)
12. If f(x, y) a--ax + by + c, discuss the changes in value and sign of f(X, Y),
when the point P with coordinates (X, Y) moves in the x, y plane, in relation to
the line whose equation is f(x, y) = 0.
Determine the smallest value attained by the expression
22x+ lly-21,
when x and y vary subject to the simultaneous restrictions
3x+4y 12,
2x — y 2. (Cambridge)
and it follows from equation (6), that the perpendicular distance from the
origin to this plane is the numerical value of
d
4/(a2 +b2+0).
Ex. 14. Write down the unit vectors perpendicular to
(i) x-2y+2z = 0; (ii) 3x—y+z = 6; (iii) 3x+2y+z = 2.
Ex. 15. Write down the equation of the planes through the given points and
perpendicular to the given vectors:
(i) (1, 2, 0), — 3j + k; (ii) (2, —1, — 1), j+k;
—
Theorem 11.4. The perpendicular distance from the point P1(x1, yl, z1) to the
plane ax + by+ cz + d = 0 is
ax1+ cz1+ d
,l(a2+62+ c2)
Fig. 11.9
Proof Let 1'2(x, y2, z2) be any point of the plane ax + by + cz +d = 0 and
let N be the foot of the perpendicular from P1to the plane (Figure 11.9).
If n is the unit vector perpendicular to the plane, then
P,N = P2 P1. n
= [(x,— x2) i+ (y1—y2) j+(z1—z2) k]
Example 3. Find the complete locus of a point which moves so that its per-
pendicular distance from 8x—y+ 18 = 0 is twice its perpendicular distance
from 7x+ 4y 6+= 0. (0 & C)
Let P(x, y) be any point of the plane. If d1, d2represent the perpendicular
distances from P to the two lines
8x—y+ 18 7x + 4y + 6
—t ; d2 — ±
V65 V65
The required locus is the totality of points P which satisfy d1 = 2d2; that is
8x—y+18 _ 14x+ 8y+ 121
{(x, y) : ±
V65 V65 f•
This is equivalent to
8x—y+ 18 14x+ 8y+ 121
(0c, y) V65 V65 I
8x—y+18 14x+ 8y+ 121
U {(x, y':
J65 V65 f
= {(x, y):2x +3y —2 = 0) U {(x, y) : 22x+ 7y+ 30 = 0}.
The locus is thus the pair of straight lines
2x + 3y — 2 = 0, 22x + 7y + 30 = 0.
1 0 1
and so a general point P of the first line has coordinates (A+3, A+2, 3A+4)
and a general point Q of the second line has coordinates (,u, — 1, +1).
The two lines are clearly not parallel, since their direction cosines are
different. Thus they are either skew or else they meet in a point. If they
meet in a point, we must have
. tA+3=
A +2 = —1,
3A+4 = ,u+1,
for some A, A. But solution of the first two equations gives A = —3, A = 0
and these values clearly do not satisfy the third equation. The lines are
therefore skew. Now
OP = (A+3) i+(A+2) j+(3A+4)k,
OQ = pi—j+(,u+l)k,
QP = (A—ic+3)i+(A+3) j+(3A—,u+3)k.
But the directions of the two given lines are given by the vectors
i+j+ 3k
and i+k.f
Thus QP is perpendicular to both lines
-4=> —#+3)+(a+3)+3(3A—,a+3) = 0
and (A ,u+ 3)
— + (3a-12+3) = 0.
Solving these two equations gives A = - = 1 and so the coordinates of
the end-points of the common perpendicular to the two lines are (2, 1, 1)
and (1, —1, 2) and the length of the common perpendicular
= V[(2 — 1)2 + (1 + 1)2 + (1 — 2)2]
= V6.
t Since we are going to use the condition for perpendicularity, in which the right-
hand side of the dot product is zero, there is no need to make these vectors unit vectors
since this would only introduce a factor that would divide out.
204
4] FURTHER THREE-DIMENSIONAL GEOMETRY
Note: A more elegant method for finding the direction of the common
perpendicular to two skew lines will be described when the vector product
is introduced.
Example 5. Find the image by reflection of the point (11, — 13, 8) in the plane
2x-3y+z+1 = 0.
The vector 2i— 3j +k is perpendicular to the given plane and so the line
through (11, —13, 8) perpendicular to the plane has equations
x-11 _ y+13 _ z —8 _ a
2 —3 — 1 •
The coordinates of a general point P of this line are
(2A + 11, — 3A — 13, A + 8)
and P lies in the plane
<r> 2(2A + 11)— 3( — 3A — 13) + + 8) + 1 = 0
<=> 14A+70 = 0
A = —5.
Thus, the reflection of (11, —13, 8) in the plane is given by A = —10, i.e.
(-9, 17, —2).
Exercise 11(c)
1. Prove that the vectors:
a = i+2j+4k, b = 2i-3j+k
are perpendicular.
Find the angle between the vectors
c = 3i+ j + 2k, d = i— 2j —3k.
2. Find the equations of the planes through the given points which are normal
to the given vectors:
(i) (0, 0, 0), j + k; (ii) (1, 2, —3), 2i+ j —3k;
(iii) (-1, 2, —4), 3i—k; (iv) (2,4, —1), 6i-3j — 2k;
(v) (2, 3, 4), 4i+ j —3k; (vi) (a, b, c), ai+ bj + ck.
3. Find the cosines of the angles between the following pairs of lines. (Leave
your answers in surd form.)
x y-1 z+1 x 1 y+2 - z-1.
2 —3 — —1' 1 —0 —2 '
205
FURTHER VECTORS Ell
5. Find the cosines of the acute angles between the following lines and planes.
(Leave your answers in surd form.)
x-1y-2 z- 3
= x+y+z = 0;
1 = 2 3,
x-3y+2 -
z+1
(n) -1 = 0 2x-y- z = 0;
2 '
x-2 y-1 z
(iii) = = -1' 3x-2y+2z+4 = 0;
2 2
x-3 y-4 =z+3
(iv) = 4x-3y-4z = 0;
4 -3 4
x-1 y+1 z-2
(v) = = y -z = 0;
1 0 -1'
(vi)
x- a
= y-b=z-c px+qy+rz = 0.
1 m n'
6. Find the unit vectors normal to the planes through the following sets of points,
and deduce the equations of the planes:
(i) (1, 1, - 1), (2, 3, 1), (5, -1, -13);
(ii) (2, 6, 1), (0, 3, 1), (4, 0, -2);
(iii) (1, 4, 1), (2, 7, 2), (-3, 0, -1);
(iv) (1, 0, -1), (0, -4, - 1), (3, 2, 2);
(v) (1, 1, 2), (2, 2, 3), (-2, 1, 11);
(vi) (0, c/b, b/c), (c/a, 0, - a/c), (b/a, alb, 0).
206
4] FURTHER THREE-DIMENSIONAL GEOMETRY
208
4] FURTHER THREE-DIMENSIONAL GEOMETRY
Miscellaneous Exercise 11
1. The perpendicular distance from the origin to a straight line 1 is of length
p and makes an angle a with the positive x axis. Prove that the equation of the
line may be taken in the form
x cos a +y sin a = p.
What is the equation of the parallel line through the point P1(x1, yi)? Deduce
the perpendicular distance from P1to the given line.
2. Find the equations of the straight lines which bisect the angles between the
straight lines
3x- 4y- 11 = 0, 12x + Sy - 2 = 0. (0 & C)
209
FURTHER VECTORS [11
3. Find the equation of the bisector of the acute angle between the lines
4x + 3y — 12 = 0, 12x + 5y — 60 = 0. (0 & C)
4. Sketch the triangle formed by the lines
3x— 4y — 4 = 0, 12x— 5y + 6 = 0, 7x+ 24y — 56 = 0,
and verify by calculation and reference to your sketch that the point (1, 1) is the
centre of the inscribed circle. (0 & C)
5. A straight line is drawn through A(h, k) and P(x, y) so that AP makes an
angle 0 with the positive direction of the x axis, and AP = r. Prove that
x—h _y—k
cos e — sin 0 — r.
Three vertices of a square are E(2, 2), F(— 2, 2) and G(— 2, — 2); a straight
line of gradient ais drawn from A(— 3, — 1) to meet FG at P and EF at Q. Use
the formulae in the first part of the equation to find the length of PQ.
Find also the radius of the circle with centre at the origin to which APQ is a
tangent. (0 & C)
6. Points A and B lie on the same side of a line 1, C is the optical image of A in
1, and BC meets 1 at P. Prove that 1 is a bisector of LAPB. Given that 1 has
the equation x +3y = 5, and that the coordinates of A and B are (1, —2) and
(— 11, 2) respectively, verify that these points lie on the same side of 1. Prove
that the coordinates of C are (3, 4), and find the coordinates of P. (0 & C)
7. Prove that the equation of the circle, centre (12, 13), radius 7, is
x2 + y2 — 24x— 26y + 264 = 0.
Two sides of a triangle are
5x-12y+5 = 0, 12x— 5y+ 12 = 0
and the incentre is the point (12, 13). Prove that the third side touches the circle
x2 + y2 — 24x— 26y + 264 = 0. (0 & C)
8. Prove that the perpendicular bisectors of the sides of a triangle are concur-
rent (at the circumcentre) and that the altitudes of a triangle are concurrent (at
the orthocentre).
9. Prove that the circumcentre S, the centroid G and the orthocentre H of a
triangle ABC have position vectors (referred to any origin 0) which satisfy the
relation
h+2s-3g = 0.
Deduce that SGH is a straight line and determine the ratio SG: GH.
Prove further that the mid-point, N, of the line SH is the centre of the circle
through the mid-points of the triangle ABC. What is the radius of this circle?
10. OABC is a tetrahredron in which OA is perpendicular to BC and OB is
perpendicular to AC. PQR is a triangle such that A is the mid-point of QR, B
is the mid-point of RP and C is the mid-point of PQ.
210
4] MISCELLANEOUS EXERCISE 11
Prove that the length of the perpendicular to the plane from the point S with
position vector s is In .s – pi; and find the position vector t of the mirror image
T of S in the plane—that is, of the point T such that TS is perpendicular to the
plane and bisected by it. (M.E.I.)
18. Referred to a given system of rectangular coordinates in space of three-
dimensions, the points A, B have coordinates (1, 0, 0), ( – 1, 0, 0) respectively.
A variable point P has coordinates (x, y, z). Write down the direction-cosines
of PA, PB and prove that, if 0 is the angle between them, then
(x2 +y2+ z2– 1)2
cost 0 =
[(x2+ y2+ z2+ 1) – 2x] [(x2+ y2 z2+ 1) + 2.x]•
Deduce that, if P is restricted to lie in the plane z = 0, then
(x2+ y2– 1)2 = [(x2+ y2– 1)2+ 4y2] cost 0
and find the equations of the two circles on which P must lie if the angle 0 is kept
constant. (M.E.I.)
19. Two planes x– 3y + 2z = 2,
2x – y – z = 9,
meet in the line 1. Find the equations of
(i) the plane through the origin which contains 1,
(ii) the plane through the origin which is perpendicular to 1.
Find also the coordinates of the reflection of the origin in 1. (C.S.)
20. Find the value of k such that the line joining the points (– 2, k, – 9), (2, 1, 7)
intersects the line joining (-2, –4, 4), (7, 2, 1).
What are the coordinates of the point of intersection?
21. A regular tetrahedron ABCD has the face ABC in the xy plane, the origin
is the centre of that face, the vertex A is at the point (1, 0, 0) and the vertex D
is on the positive half of the z axis. Find the coordinates of B, C, D and of the
centre of the tetrahedron, and the direction ratios of the normals to the four
faces.
Hence or otherwise show that, if any line makes angles a, y, 8 with the faces
of a regular tetrahedron, then
sine a + sin2/3 + sin2y + sin2 8
22. Prove that the equation of the straight line through the given point A,
position vector a, and in the direction of the unit vector b is
r = a+tb,
where t is the distance from A of the variable point P of the line whose position
vector is r.
Prove also that the equation of the plane through the point C (position
vector c) whose normal is in the direction given by the unit vector d can be
expressed in the form
r .d = c.d.
212
4] MISCELLANEOUS EXERCISE 11
The plane through the point C(1, 2, 4) has normal in the direction
d=
Find the length of the shortest distance from the point A (3, 4, 5) to the plane.
(M.E.I.)
23. To find the position of an underground rock layer a number of vertical
borings are made at points on horizontal ground which form a coordinate grid.
Results are as follows, the unit of distance both horizontally and vertically
being 300 m:
Grid-point (0, 0) (2, 0) (0, 2) (2, 2)
Depth 0.270 0.225 0.162 0.117
Show that these results are consistent with this part of the rock layer being a
plane. Find the (x, y) equation of the line in which this plane when produced
would meet the ground, and show that the plane would be inclined at about 31°
to the ground.
What would you conclude if a boring at (1, 1) gave a depth of about 0.18?
(M.E.I.)
24. Calculate the shortest distance between the line of intersection of the planes
x-8y+2z+9 = 0, x-2y—z+6 = 0
and the line of intersection of the planes
2x+y+8z-12 = 0, x—y+z-6 = 0;
and show that the line which cuts both these lines at right-angles passes through
the point (-4, 12, 8).
25. Find the coordinates of the mirror image of the point (p, q, r) in the plane
ax+by+cz+d = O.
A ray from the origin is reflected successively in the planes
x+y—z+l = 0
and x —y + 2z— 1 = 0,
and then passes again through the origin. Find the points at which it meets the
two planes.
26. Prove that the lines
x+1 y+1 z+2 x-1 y+3 = z
2 3 1 3 1 —2
do not intersect.
Find the equation of the plane through the origin which contains the first
line, and find also the direction cosines of the line through the origin which meets
both lines. (Oxford Mod.)
27. The line x// =ylm = zln is reflected in the plane ax+ by+ cz + d = 0. Show
that the equation of the resulting line is
(a2 + b2 + c2) x+ 2ad b2 c 2) y+2bd
b+c2) 1— 2a(al + bm+ cn) =(a2+ b2 c2) m-2b(a1+ bm + cn)
(a2+ b2+ c) z +2cd
02+1,2 +C2•
n — 2c(a1+ bm+ cn)•
213
FURTHER VECTORS [11
Hence, or otherwise, find the equation of the plane such that the angle between
it and the plane /x + my+ nz = 0 is bisected by the plane ax+ by+ cz = 0.
(Oxford Mod.)
28. Write down the equations of the axes Ox, Oy, Oz. What is the equation of the
plane containing Ox and the point (a, b, c)?
The roof of a rectangular house consists of four inclined planes, each sloping
upwards at an angle of 45° to the horizontal. What is the angle between two
adjacent faces of the roof?
29. ABCDA' B'C' D' is a cubical box, with faces ABCD, A' B'C' D' and edges AA',
BB', CC', DD'. E is the point on BB' such that BE = lEB'; F is the mid-point
of CC'. Find the angles between
(i) the line A' B and the plane AEF;
(ii) the plane A' BC and the plane AEF.
30. Two lines through the origin have unit direction vectors 11i + mi j + nik and
121+ m2 j + n2 k. Prove that the locus of points equidistant from the two lines is a
pair of planes and deduce the equations of the straight lines which bisect the
angles between the given lines.
31. Two straight paths on a plane hillside are at right-angles, and make angles 0
and 0 respectively with the horizontal. If the hillside itself makes an angle a with
the horizontal, prove that sine
0 + sin2 ¢ = sine a.
Prove also that the acute angle between the projections of the paths on a
horizontal plane is
arccos (tan 0 tan cb).
32. A line of slope, in an easterly direction, of a plane hillside is inclined at an
angle a to the horizontal. A line of slope of the hillside in a southerly direction
is inclined at an angle fi to the horizontal. Prove that the actual inclination, 0,
of the hillside to the horizontal is
0 = arctan (tang a + tan2 fi)1.
A vertical pole of height h is placed on top of the hill. Show that the angle 0
subtended by it at a point distant a down the line of greatest slope through the
foot of the pole is given by
h
tan 0 = a sec
0 + h tan 0.
Find 0, if h = 16, a = 36, a = 30° and ft = 45°. (London)
33. Let G be the centroid of the acute-angled triangle ABC of circumradius R.
Show that
AG2+BG2+ CG2 2R2.
34. If 0 is the point on AB such that AO = 20B, and if /3is any point, prove that
AP2+2BP2 — 30P2
is independent of the position of P.
What is the locus of a point which moves so that the sum of the squares of its
distances from the vertices of an equilateral triangle is constant?
214
12. Further trigonometry
The proof of (i) and (ii) is effected by expressing the unit vector r (a) in
terms of i and j directly and (b) in terms of p and q and hence in terms of i
and j. QB
Thus (a) r = cos (0 + 0) i+ sin (0+ 0) j
and (b) r = cos Op + sin 0q.
But p = cos Oi + sin Oj
and q = — sin &i+ cos Oj
Fig. 12.2
(see remarks preceding this theorem).
Substituting for p and q in (b) this gives
(c) r = cos 0(cos Oi+ sin 0j) + sin 0(— sin Oi+ cos OD
= (cos 0 cos 0 — sin 0 sin 95) i+ (sin 0 cos 95+ cos 0 sin 95) j.
But, in a plane, the expressions for the components of a vector r in the
two directions i and j are unique (see Chapter 2) and so, comparing (a)
and (c),
(i) cos (0 + 0) = cos 0 cos 95— sin 0 sin 0;
(ii) sin (0 + 0) = sin 0 cos 95+ cos 0 sin 95.
Formulae (iii) and (iv) now follow immediately, on writing — 0 for 0 and
recalling that sin (— 0) — sin 0, cos (— 0) = cos 95,
(iii) cos (0- 0) = cos 0 cos 0 + sin 0 sin 0;
(iv) sin (o— 0) = sin 0 cos 0 — cos 0 sin 0.
Ex. 2. Express sin 15°and cos 15°in surd form by writing 15° = 45° — 30°.
Ex. 3. Express cos 165° and sin 105° in surd form.
Ex. 4. If sin x =A- and sin y = s find sin (x +y) and cos (x +y)
(i) when x and y are both acute angles;
(ii) when x is acute and y is obtuse;
(iii) when x and y are both obtuse angles.
Ex. 5. Express cos (A+ B) cos (A B) in terms of sin A and sin B.
— —
Theorem 12.2. For any angles 0, 0 for which all of the expressions appearing
are defined
tan 0+ tan 0
(v) tan (o+0) =
1 — tan 0 tan 0 '
_ tan 0 — tan 0
(vi) tan (0— 0) =
1 +tan 0 tan 0'
216
1] COMPOUND AND MULTIPLE ANGLES
*Ex. 6. What values of 0 and cb must be excluded in (a) formula (vi); and (b)
formula (ix)?
cos 20
1— sin 20
cos 20
The formulae we have proved are also useful in dealing with combina-
tions of inverse trigonometric functions, as is shown in the next example.
V61'
where tan a = 5, i.e. a = 50° 12'. Also, using tables,
3
= cos 67° 25'.
V61
218
1] COMPOUND AND MULTIPLE ANGLES
No. Log.
3 0.4771
61 1.7853
,,/61 0 8926
Thus, cos 50° 12' cos 0 — sin 50° 12' sin 0 = cos 67° 25',
i.e. cos (50°12' + 0) = cos 67° 25',
.*. 50° 12' + 0 = 67° 25' or 292° 35'
and 0 = 17° 13' or 242° 23'.
y = V17 G17
1
sin
20+V17
cos 20)
Ex. 7. Find the maximum and minimum values of the following expressions:
(i) sin x - 3 cos x; (ii) 2 cos x+ sin x; (iii) 5+3 sin x+ 4 cos x;
(iv) 1 sin x cos x;
— — (v) 1/(2— sin x+ cos x).
Ex. 8. Solve the following equations, giving all solutions lying between 0° and
360° inclusive:
(i) cos x— sin x = 1; (ii) 3 cos x +4 sin x = 5;
(iii) sin x +2 cos x = 1; (iv) 12 sin x-5 cos x = —4.
Ex. 9. If tan A = 2, and tan B = 1, find tan (A + B) and cot (A — B).
Ex. 10. If x+y = in and tan x = 4 find tan y. ,
8 PPM 219
FURTHER TRIGONOMETRY [12
Ex. 11. If cos x = 4, find cos 2x. If, furthermore, 0 < x < 77, find sin 2x.
Ex. 12. Find the maximum and minimum values of 1 + sin x cos x.
Ex. 13. Find the values of x in the interval 0 < x < 27r which satisfy the equa-
tion cos 2x = 3 sin x 1.
-
Ex. 14. Find the value of tan irr without using your tables.
Ex. 15. Find sin (2 arcsin x) and cos (2 arccos y).
Ex. 16. By writing x = arctan 4 and y = arctan 4 and considering tan (x+y),
show that arctan 4+ arctan i = arctan i.
Ex. 17. Express arcsin 4 + arcsin "A in the form arcsin x.
The multiple angle formulae can be used to derive three very useful
expressions, for sin 0, cos 0 and tan 0, in terms of tan 10. Their value
derives from the fact that, with their help, a trigonometric expression may
be written as a rational functiont of a single variable t = tan 40. The reader
will find that they have applications to problems in calculus as well as
ordinary trigonometry.
= 1 — tan2 210
sec2-1-0
= 1 — t2
1+t2 '
2
(iii) tan 0 = by dividing (i) by (ii).
12t t2'
t A rational function is defined as the ratio of two polynomials.
220
1] COMPOUND AND MULTIPLE ANGLES
1 +t
1 + t * 0 (10 + (2n + 1) /7/),
1 — t'
R.H.S. — 4 + tan 1
tan-In 20
1—tan ill' tan 10
1 +t
since tan in * 1.
1— t'
Theorem 12.4.
(i) sin A sin B Ecos (A — B) — cos (A + B)] ;
(ii) cos A cos B +[cos (A + B)+ cos (A — B)];
(iii) sin A cos B = Esin (A + B)+ sin (A — B)];
(iv) cos A sin B = l[sin (A + B) — sin (A—B)].
Proof. (i) R.H.S. = +[(cos A cos B+ sin A sin B)
— (cos A cos B — sin A sin B)]
= 1.2 sin A sin B
= L.H.S.;
(ii), (iii), (iv) are all proved in a similar fashion.
Theorem 12.5.
A— B
(i) sin A+ sin B= 2 sin A+
2 cos
B 2
A— B
(iii) cos A+ cos B 2 cos A +
2 cos
B 2 ;
B— A
(iv) cos A—cos B 2 sin A 2 B sin
• 2
Proof.
FA+BA—Bl
+ +sin FA+B A—B1
R.H.S. = sin
L 2 2 j L 2 2 t
by Theorem 12.4 (iii)
= sin A+ sin B
= L.H.S.
(iv) are all proved in similar fashion.
222
2] SUM AND PRODUCT FORMULAE
B+C = 90°— A
= 1-2 sin2 —
A +2 sin A B—C since
2 2 cos 2 ' 2 2
= 1— 2 sin A
— [sin A cos13— Cl
2 2 2
Ex. 20. Prove the identity sin x+ sin 2x + sin 3x E sin 2x(2 cos x+ 1). Hence
find all values of x in the interval 0 x 27r satisfying the equation
sin x+sin 2x+ sin 3x = 0.
*Ex. 21. Solve the equation sin x+ sin 2x = 0 by expressing the left-hand side in
factor form, giving all solutions in the interval 0 x < 271. Solve the equation
again in the following alternative ways:
(i) by rewriting it as sin 2x = sin (— x), etc.;
(ii) by rewriting it as sin x + 2 sin x cos x = 0.
Exercise 12(a)
1. Evaluate in surd form:
(i) sin 75°; (ii) cos 105°; (iii) tan 105°; (iv) cot 75°.
2. Prove that sin 3A = 3 sin A —4 sin3 A and that cos 3A = 4 coss A -3 cos A.
By considering the equation sin 2A = cos 3A, express in surd form:
(i) sin 18°; (ii) cos 18°; (iii) sin 54°; (iv) tan 108°.
3. Prove that the following identities hold for all angles A, B for which the
expressions appearing are defined:
(i) sin (A+ B) sin (A— B) E cos2B— cost A;
(ii) cos (A — B)— sin (A + B) E (cos A — sin A) (cos B — sin B);
cos (A+ B) + cos (A— B) _
( ii ) = cot B;i
sin (A+ B)— sin (A— B)
(iv) sin (A+ B) sin B al cos A—cos (A+ B) cos B;
sec Asec B cosec A cosec B
(v) sec (A+ B)
cosec A cosec B — sec A sec B'
(vi) tan 2A(cot A — tan A) E. 2;
(vii) cos 4A 4(cos4 A+sin4 A) — 3;
(viii) cot A — cot 2A cosec 2A;
sin A+ sin B
(ix) cos A + cos B = tan 1(A + B);
and hence, using the formula sin 30 -a-. 3 sin 0-4 sin' 0, show that cos 36° is
a root of the equation
8x4 -8x2 +x+1 = 0. (0 & C)
1—tan2 0
12. (i) Prove that cos 20 7._
1+tan2 0
and, without using tables, deduce that
tan 221° = —1
13. (i) If a = sin 0+ cos 93 and b = cos 0 + sin 95, prove that
cos (0 — q5) = 2ab/(a2+ b2)
and hence, or otherwise, find tan (0 — 93) in terms of a and b.
(ii) Express 11 sin2 x+12 sin x cos x+6 cos° x in the form a+ b sin (2x- 0)
where a, b and 0 are constants to be determined.
Hence prove that
2 < 11 sin2 x+ 12 sin x cos x+6 cost x < 15. (Cambridge)
14. Prove that sec x+ tan x = tan air +-1-x)
and express in a similar way (i.e. as the tangent of an angle)
(i) sec x— tan x; (ii) cosec x— cot x. (Cambridge)
15. Prove
2 sin 0(cos 20+ cos 40 + cos 60) = sin 70 — sin 0.
Deduce that cos 47/ +cos VT + COS trr = —4.
Show also that cos jr77 + COS + cos 4r = (Cambridge)
16. Express the function cos x+2 sin x in the form R sin (x +a) where R is
positive and 0° < a < 360°. State the values of R and a. Hence, or otherwise,
find the values of x in the range 0° to 360° inclusive which satisfy the equations
(i) 2 cos x + 4 sin x = 1; (ii) cos x(cos x + 2 sin x) = 1. (Cambridge)
17. If u = cos 0+ sin 0, v = cos 0 — sin 0, prove that
cos 20 = uv, sin 20 = u2 — 1, u2 + v2 = 2.
Prove that, if 0 is a root lying between —180° and 180° of the equation
cos 20+ a sin 20 = cos 0 — sin 0,
then it is either 0° or 90° or it is a root of the equation
(a+ 1) cos 0 + (a— 1) sin 0+a = 0.
Solve the equation
cos 20+7 sin 20 = cos 0 — sin 0
for values of 0 between —180° and 180°. (0 & C)
226
SUM AND PRODUCT FORMULAE
18. Find all the pairs of values (8, 93) lying between 0 and 27-r that satisfy the
equations cos 0+ cos 95 = cos -4-7r,
sin 0+ sin ¢ = sin 17r. (0 & C)
19. Prove, by induction or otherwise, that
cos a + cos (a + fl)+ cos (a + 2fi)+ + cos [a + (n-1) fl]
E cos [a +1(n— 1) fl] sin -infl cosec
r2
Fig. 12.5 Fig. 12.6
(i)
Fig. 12.7
Ex. 23. Show that the area, A, of the AABC is given by the formula
A = ibc sin A.
Write down the two similar expressions for A and deduce the Sine Rule in the
form
a
sin A sin B sin C.
Example 10. In AABC, BC = 6 cm, CA = 4 cm, AB = 5 cm. Find the
angles of the triangle.
228
3] SOLUTION OF TRIANGLES
Clearly we must use the Cosine Rule to begin with. We select the angle B,
since this is the smallest angle, and the tables of cosines are marginally
more accurate for smaller than larger angles.
16 = 25 + 36 — 60 cos LB,
cos LB = 0.75,
LB = 41° 24'.
sin A sin 46° 24'
By the Sine Rule
6 2 '
sin LA = 0.9919(5),
LA = 82° 42'.
By the angle sum property of a triangle
LC = 55° 54'.
Example 11. Discuss the solution of a triangle in which A, a, c are given.
(This is often called the ambiguous case, since it is sometimes possible
to draw two triangles.)
*Ex. 24. Interpret the four congruency conditions for triangles in terms of the
solution of triangles by trigonometry.
229
FURTHER TRIGONOMETRY [12
Example 12. A hillside faces due north and is inclined at 20° to the horizontal.
A path up the hill has a bearing of 120°. Find the angle the path makes with
the horizontal.
Fig. 12.9
Example 13. Prove the theorem of Apollonius that, in any triangle ABC with
median AM, AB2+ AC2= 2AM2+2BM2.
An aircraft flying on a constant course and at constant height with speed
V is observed from a station on the ground at times 0, t, 2t to have elevations
a, fl, y respectively. Prove that
V = kl(cot2 a-2 cote ft+ cote -y)/tV2.
If the bearings of the aircraft from the station at times 0 and 2t are 0,
and 02respectively, determine the course of the aircraft.
0
Fig. 12.13
Exercise 12 (b)
1. Two points B and C on the bank of a straight river are 120 metres apart.
It is observed that a point A on the opposite bank is such that the angle ABC
is 72° 15' and the angle ACB is 38° 30'. Find the width of the river, correct to the
nearest metre. (0 &C '0')
2. In LABC, s = 1-(a + b + c); deduce from the Cosine Rule that
cos 4 IFscs-
2 V L be
If the area of the triangle is A, use this result to prove
A [(s - b) (s -c)-1
(i) si
n-= (ii) A = V[s(s- a) (s - b) (s - c)].
bc
3. In the AABC, a = 14, b = 15,c = 13. Calculate
(i) the area of the triangle;
(ii) sin A;
(iii) the radius of the circumcircle. (0 & C '0')
4. Prove that, in any triangle ABC,
a - b sin A-sin B
(i) a+ b sin A + sin B'
A- B a-b C
(ii) tan 2 = a+bcot 2 .
232
4] SOLUTION OF PROBLEMS
In a triangle ABC, a = 13.41 cm, b = 9.63 cm and LC = 34°. Find the size
of the angles LA and LB.
If in a triangle a = 5, b = 4 and cos (A — B) = prove that cos C = *
and that c = 6. (0 & C)
5. A straight river is 80 m wide. A man on one bank observed that the angle of
elevation of the top of a tree directly opposite him is 18° 20'. He walks 60 m
along the bank. Find the angle of elevation of the top of the tree from his new
position. (0 & C '0')
6. In a triangle ABC, a = 20 m, b = 28 m, c = 32 m. Prove that LB = 60°.
From the points A, B, C which are on level ground, the top of a flagstaff has
the same angle of elevation, 30°. Calculate the height of the flagstaff.
(0 & C '0')
10. ABCD is a regular tetrahedron and M is the mid-point of the edge CD. Find
the angle between the plane ABC and
(i) the plane ABM; (ii) the line AM.
11. From the top of a cliff the angle of depression of a ship, steaming on a con-
stant course at 12 km per hour, is 15° and its true bearing is 75° north of west.
Two minutes later, the angle of depression is 12° and the true bearing is due west.
Find
(i) the height of the cliff (to the nearest metre);
(ii) the ship's course (to the nearest half-degree).
12. A and B are two points on one bank of a straight stretch of a river, P is a
chimney exactly opposite A and 20 m from the other bank. The angles of eleva-
tion of the top of the chimney from A and B are a and fi respectively. Calculate
the width of the river to the nearest metre, given that
a = 45°, ft = 30° and AB = 200 m.
233
FURTHER TRIGONOMETRY [12
Prove that, if the angle of elevation of the top of the chimney from a point C,
midway between A and B, is y, then
4 cote y= 3 cot' a + cot2 ft
whatever the height of the chimney, the width of the river, and the distance
AB may be. (0 & C)
Miscellaneous Exercise 12
1. Prove that arctan i + arctan 4 = 41r and that
arctan i + arctan i + arctan 4 = in.
2. Solve the equations:
(i) arctan 1+ arctan x = arctan 4;
(ii) 2 arcsin (x — y) = 3 arccos (x+y) = Tr.
3. Two circles with centres 0 and C meet in P and Q. The radii of the circles are
a and b and the angle CPO is a. Prove that the angle between the common tan-
gents of the circles is 0, where
(a— b)2 cot2 40 = 4ab sine 4a. (0 & C)
4. Find the set of values of x which satisfy the inequality
2 sin x+sin 2x > 0. (SMP)
5. Under what circumstances is it true that
x +y
arctan x+ arctan y = arctan ?
1 — xy
Find values of x, y for which the relationship above is (i) true, (ii) untrue.
6. Prove that, in any triangle ABC,
a b c
sin A sin B sin C
In the triangle ABC the angle B is a right-angle and 0 is a point inside the tri-
angle at which all the sides subtend the angle 120°. If 0 is the angle CBO, prove
that
c+a,13
tan 0 =
a+ cA/3
If the angle C = 30°, show that CO = 2A0. (0 & C)
7. Show that 4 arctan 4— arctan th = 4T (a result commonly known as
Machin's Formula).
235
FURTHER TRIGONOMETRY [12
8. Find all pairs of angles x, y such that 0 < x < ir 0 < y < 77-which satisfy
,
16. The wave-train well away from a ship is modelled by the equation
z = a sin [(x+y— ct)Ip]
where z is the height of the sea's surface above the mean horizontal plane in
which (x, y) are Cartesian coordinates, a is the maximum height of the waves, c
is the fixed speed, and p a fixed length. Sketch a diagram showing an airman's
view of the waves, indicating the lines of the crests and troughs and the direction
in which the waves are travelling; and prove that the wave velocity and distance
between successive waves are c/V2 and irpV2 respectively.
The wave train of another ship, given by
z = a sin [(x— y — ct)Ip]
is superimposed on the other one. Prove that on the lines y = (N + I-) rip, where
N is an integer, the sea is undisturbed. (SMP)
17. Find the range or ranges of values of c such that the simultaneous equa-
tions
cos 0+ sin q5 = 1,
sec 0+ cosec 95 = c
are satisfied by real values of 0 and 0.
Obtain the general solutions of these equations when c = 61.
237
13. Matrices 1
1. INTRODUCTION
Any pair of simultaneous equations in two unknowns, x and y,
f ax+ by = c,
(dx+ey = f
is completely specified if we know
(i) the coefficients on the left-hand side, which we may write as the
rectangular array
(ad be)
and (ii) the two numbers on the right-hand side, which we may write as
the rectangular array
(Cf)
Ex. 1. Solve the simultaneous equations whose coefficients are given by the
rectangular array
P— 1)
2
and whose right-hand sides are given by the rectangular array
( 112)
The answer is the pair of values 4 (for x) and — 3 (for y), which can be expressed
as an array
_ 34) ,
as we saw in Chapter 2.
(read `three by five matrix'; that is, a matrix with three rows and five
columns):
ABCDE
M 2 5 ( 7 6 3
E 1 6 8 6 2 ).
F 3 5 8 7 0
(878)
(ii) the 3 x 2 matrix (2 5
1 6);
35
(iii) the 1 x 5 matrix (or row-vector)
(3 5 8 7 0).
More generally, any rectangular array of m rows and n columns of ele-
ments is called an m x n matrix:
an a12 a13 ... al.
a21 a22 a23
2. LINEAR TRANSFORMATIONS
If we are given a pair of axes in a plane, any point P may be located by
its coordinates (x, y), so that OP = xi +yj. Suppose we now associate
with each point P(x, y) a unique point P'(x', y') such that
x' = x +2y,
y' = x — y.
For example, the point P(1, 2) gives rise to P'(5, — 1); the point Q( —2, 3)
gives rise to Q'(4, — 5). Such an association of points is a mapping, or func-
239
MATRICES 1 [13
tion, of the set of points in the plane into itself and is often referred to as a
transformation of the plane into itself. We may specify our mapping by the
2 x 2 matrix
which we call the matrix of the given transformation. For the present,
such a matrix is to be regarded simply as an inert array of coefficients
defining a transformation. In Section 3 we shall consider combinations of
transformations and the resulting matrices; in Section 4, rules for combining
matrices will be developed and the inert arrays which we have at present
will come to life. Finally, in Section 5 matrices, with their new-found
vitality, will be used to illuminate the concept of a geometrical transforma-
tion.
We may also regard the transformation as mapping the vector OP into
the vector OP' and the vector OQ into the vector OQ'. If we call the
transformation T, then we write
T(OP) = OP'
and T(OQ) = OQ'.
Notice carefully that T(OP) is a vector: the position vector of P'. T(OP) is
called the image of OP; P' is the image of P and may be written T(P).
The transformation defined above has an important property: it maps
the vector OP + OQ into the vector sum of T(OP) and T(OQ). Let us
demonstrate this property first with the points P and Q above:
OP = i+2j, OQ = —2i+ 3j.
Let OR = OP + OQ
so that R is the point (— 1, 5) and OR = — i + 5j. By the definition of T, R'
has coordinates (-1+2.5, —1— 5), i.e. (9, — 6) and so
OR' = 9i — 6j.
But OP' + OQ' ----- (51— j)+ (4i— 5j)
= (91— 6j)
= OR'.
Now let P be any point (h1, k1), Q any point (112, k2) and let R(h3, k3) be
the point such that
OR = OP + OQ,
or h3i+k3 j = (111-Fh2)i+(k+k2)j,
240
2] LINEAR TRANSFORMATIONS
*Ex. 3. By writing x in the form x+0, show that, for any linear transformation
T, T(0) = 0.
An example of a mapping, T, that is particu- P(x, y)
larly easy to visualize geometrically is that in
which the image P' of P is obtained by a half- ax
turn about the origin (see Figure 13.1). The co-
ordinates of P'(x' ,y') in terms of P(x,y) are given P' (x' , y')
by
x' = —x, Fig. 13.1
= —Y
and thus the matrix of this transformation is
k 0 -1
241
MATRICES 1 [13
3. LINEAR TRANSFORMATIONS
AND THEIR MATRICES
Suppose we have a linear transformation T whose matrix A is given by
(a b\
A=
c
Since i is the position vector of the point /(1, 0) and j is the position vector
of the point AO, 1), we have
T(i) = ai+ cj, T(j) = bi+ dj.
Recalling that vectors may be written in column form, this result shows us
that the first column of A, regarded as a vector, is the position vector of the
image of I and the second column of A, regarded as a vector, is the position
vector of J.
A= , w = (2 1\
Ex. 9. If
‘2 3/ kl —1)
and P is the point (1, 1) sketch in a diagram the effects of S, T and U upon the
point P.
*Ex. 10. The transformation L is defined by
L(x) = cT(x).
(L is sometimes written as cT.) Prove that L is a linear transformation and that
its matrix is
cw ell
(
cy czl
*Ex. 11. The transformation M is defined by
M(x) = S[T(x)],
that is, x is first transformed under T into T(x), and T(x) is then transformed
under S into M(x). (M is called the product of S and T and may be written as
ST.) Prove that M is a linear transformation and that its matrix is
taw+ by ax+ bz\
‘cw + dy cx + dzI '
Exercise 13(a)
1. The transformation T is defined by the equations
x' = 2x—y,
y' = x-3y.
245
MATRICES 1 [13
What is the image of the point P(— 1, 3) ? What point Q gives rise to the point
Q'(3, 5) ?
2. A transformation is represented by the matrix
(
— 2 — 3) •
Find the image of the point (-1, 2) and also the point whose image is (14, 9).
What points are invariant under this transformation? (A point is invariant
under a transformation if it maps into itself.)
3. What points are invariant under the transformation whose matrix is
(4 — 6)
1 —1 •
4. A transformation whose matrix is
ki)
is called a shear. Consider its effect upon the square with vertices (0, 0), (1, 0),
(1, 1), (0, 1) and suggest a reason for the name.
5. Prove that the shear transformation defined in Question 4 is linear.
6. Describe the following transformation in geometrical terms:
x' = y,
y' = 0.
Draw a sketch to illustrate the transformation; mark in a number of points and
their images.
Prove that the transformation is linear.
7. Answer the same questions as in Question 6 for the transformation
x' = 2y,
y' = 2x.
8. Sketch a diagram to illustrate the effect of the transformation which has matrix
(ac b
d)
on the four points P(0, 0), Q(1, 0), R(1, 1), S(0, 1).
Determine the area of the image quadrilateral Pf Q' R' S'.
What can you say about the particular transformation in which ad— bc = 0?
9. Prove that the translation transformation defined by
x' = x,
= y+ 1
is not a linear transformation.
246
3] MATRICES OF LINEAR TRANSFORMATIONS
10. Find the matrix of the linear transformation which maps the points U, J
whose position vectors are given by
u= (1), ]= (?)
into the points U', J' whose position vectors are given by
Example 2. If/2 a 3\ c —2 3\
kl 0 bf e '
then a = —2, b = 4, c = 2, d = 1, e = 0. The two matrices
52
0 1)
(
/0 2 0
and
k5 1 0
are not equal.
(ii) Addition of matrices.
Two matrices are said to be conformable for addition if they have the
same number of rows and the same number of columns.
By analogy with the corresponding rule for column vectors, we make the
following definition: if two matrices are conformable for addition, A +B
is the matrix whose i, jth element is the sum of the i, jth elements of A and B.
If A and B are not conformable for addition, A +B is not defined.
Example 3. If 1 —3 2) 2 2 —3
A= 4 0 3 , B = (1 —3 4) ,
1 —1
( 2 3 —3 1
then (1+2 —3+2 2-3 3 —1 —1
A+B = 4+1 0-3 3+4 = 5 —3 7.
1+3 —1-3 2+1 ) 4 —4 3
If 1 3
C = (1 1),
2 4
then A + C is not defined.
(iii) Multiplication of matrices by numbers.
Again working by analogy with column vectors, if A is any matrix and k
is any number, kA is a matrix, conformable for addition with A, whose i, jth
elements is k times the i, jth element of A.
248
4] THE ALGEBRA OF MATRICES
Example 4. If
1 2 3 6 —2 —4
A = 3-1, 3A= 9-3, —2A = —6 2.
1 —1 3 —3 —2 2
We may now define the subtraction of two matrices A and B which are
conformable for addition by the equation A—B = A + (— 1) B.
Example 5. With the matrices A and B of Example 3
—1 —5 5
A—B = ( 3 3 —1).
—2 2 1
(iv) Zero matrices.
A zero matrix is a matrix all of whose elements are zero. Provided no
confusion is likely to arise, a zero matrix may be written 0 but note that
there are many different zero matrices; for example,
(0\ \ / 0\
\O) k0 Of k0 0 0)
are all zero matrices, and are all different.
(v) Multiplication of matrices.
Two matrices A and B are said to be conformable for the product AB if the
number of columns in A is the same as the number of rows in B. We now
define the product AB of two matrices A and B which are conformable for
the product AB as follows: the i, jth element of AB is obtained from the ith
row of A and the jth column of B by multiplying together the corresponding
elements and adding. If A, B are not conformable for the product AB, then
AB is not defined.
Example 6. If
(1 3 2 —1 1 —1
A= 5 —1 3 and B= 2 0 2,
1 1 2 0 3 1
then
((1) (— 1)+ (3) (2)+(2) (0) (1) (1)+ (3) (0) + (2) (3)
AB = (5) (— 1)+ (— 1) (2)+(3) (0) (5) (1) + ( 1) (0) + (3) (3)
-
(1) ( — 1) + (1) (2) + (2) (0) (1) (1) + (1) (0) + (2) (3)
(1) ( — 1)+ (3) (2)+(2) (1)
(5) ( — 1)+( — 1) (2)+(3) (1)
(1) ( — 1) + (1) (2)+ (2) (1)
(5 7 7
= -7 14-4.
1 7 3
249
MATRICES 1 [13
If 1 2) 2 8
C= —1 0 then AC = 12 19
2 3 4 8
but CA is not defined.
Ex. 13. If 3 2 —1 6 0 2)
A =( B= 1 1 — 1
2 2 —1/ 2 3 0
find
(i) A+B; (ii) 2A-3B; (iii) AB; (iv) BA.
Ex. 14.1f 1 —1
A = (2 1), B = (23 3 1)
0 2 \ —2 0
find
(i) AB; (ii) BA.
*Ex. 15. If
A = (1 —1) B = (0 2) e = (1 ,
\2 1/' 21' 0 0
verify that
(i) A+B = B+A; (ii) (A+B)+C = A+(B+C);
(iii) A(BC) = (AB)C; (iv) A(B+ C) = AB+ AC.
Show also that AB = BA.
In Ex. 15 above, certain of the basic laws of algebra have been shown to
hold good for the particular matrices A, B, C. With the definitions of
equality, addition and multiplication we have given, the following laws
may be shown to hold good for all matrices A, B, C, provided all the sums
and products are defined:
(i) Matrix addition is
(a) Commutative: A+B = B+A;
(b) Associative: A + (B + C) = (A + B) + C.
(ii) Matrix multiplication is
(a) Associative: A(BC) = (AB) C;
(A(B+C) = AB + AC,
(b) Distributive over addition:
(A +B)C = AC + BC.
However, matrix multiplication is non-commutative; indeed, two
matrices A, B may be conformable for the product AB and yet BA is
not defined. This is not to say that matrix multiplication is never com-
mutative.
Ex. 16. If A = (4
1 2) 1 4
and B = (— 8
3 3)
verify that AB = BA.
250
41 THE ALGEBRA OF MATRICES
Ex. 17. Show that, if A is any 2 x 2 matrix, B is any 3 x 2 matrix and I is the 2 x 2
identity matrix, (i) AI = IA = A; (ii) BI = B.
Ex. 18. Show that, if I is the 3 x 3 identity matrix and A is any matrix with three
columns, AI = A.
5. LINEAR TRANSFORMATIONS
AND THEIR MATRICES (CONTINUED)
The rules for matrix addition and multiplication given above enable us
to obtain a deeper insight into the machinery of linear transformations.
If P has coordinates (x, y) and its image, P', has coordinates (x', y') then
the position vectors (in column vector form) of P and P' are respectively
(iii) If the image of the point P(x, y, z) is the origin, then x, y, z satisfy
the equations
(a) 2y+ z = 0,
(b) 3x+ y+2z = 0,
(c) 3x— y+ z = 0.
Planes (a) and (b) meet in the line x = y = --1z which clearly lies also in
plane (c). Thus, all points with coordinates of the form (A, A, —2A) map
into the origin.
Exercise 13(b)
1. If = (—1
3 234 0
2) and B = (-1
2 —2 4)'
evaluate
(i) A+B; (ii) A—B; (iii) 2A—B; (iv) 2A+ 3B.
2. If
A = 3 0 3 and B = 1
1 2 —1 2 —1 2
evaluate
(i) A+B; (ii) 2A-3B; (iii) 3A—B;
(iv) 4A-2B.
3. If 3 0
\ ' B= /\ c = (011/\
A = (-1 2
2 3/ ‘I 1)' 2
evaluate
(i) A-3B+2C; (ii) 2A(13—C); (iii) A2 — AB+AC.
Verify that (A+B)2 = A2 +AB+BA+B2; can this be put in the simpler form
A2 +2AB+B2 ?
4. A, B, C are all 3 x 3 matrices. Remove brackets in the following expressions:
(i) A(B-2C); (ii) (A+B) (A— C); (iii) (A—B—C) (A+B+C).
5. If
1 2
A= —31), B= (
2
3 —1 0)'
I 2
find AB and BA.
6. If
A = 0 —1 —1 , B = 1 —1 —2 ,
2 1 1/ 1 1 1
find
(i) AB; (ii) BA; (iii) A2;
(iv) B2; (v) (2A—B)2; (vi) (A+2B) (A—B).
7. If 2 1 —1 0 2 1
A= 1 1 2), B = 1 3 C2),
—1 2 1/ —2 —1
9-2 253
MATRICES 1 [13
find
(i) AB; (ii) BA; (iii) A2;
(iv) B2 ; (v) A2 -4B2 ; (vi) (A — 2B) (A + 2B).
8. If 1 3 (
A= —2 1 —2, B= 2 1 0, C= 2 1 ,
1 2 3 0 2 3 —2 1
evaluate
(i) AC; (ii) BC; (iii) A + 2B.
Verify that (A + 2B) C = AC + 2BC.
9. If 1 2)
A= —3 0,
B=—120
1 0 1)'
2 —1
find
(i) AB; (ii) BA; (iii) ABA; (iv) BAB.
10. A matrix M is said to be transposed into the matrix M' if the first row of M
becomes the first column of M', the second row of M becomes the second column
of M', and so on. Write down the transposes of the matrices:
0 b 0)
M =(x) y, T = (0 0 c .
a 0 0
Calculate the matrix products M'M and TM; show also that (TM)' = M'T'.
If the elements of M are the Cartesian coordinates of a point P, what informa-
tion is provided by the element of M'M?
If the matrix T describes a transformation of the point P of three-dimensional
space, interpret geometrically the equation:
(TM)' (TM) = M'M,
and find all appropriate values of a, b and c. (SMP)
11. Show that under the linear transformation defined by the matrix
3 —1 — 1)
A= (2 2 1
8 0 —1
any point P in space is mapped into a certain plane, and find the equation of this
plane.
Show further that all points of the line
x-1 = y —1 = z — 3
1 —5 8
are mapped into a certain point, and find the coordinates of this point.
12. Show that, under the linear transformation defined by the matrix
2 1 1)
A= (3 2 1 ,
1 1 0
any point P in space is mapped into a certain plane, and find the equation of this
plane.
254
5] LINEAR TRANSFORMATIONS (CONTINUED)
1 1 — —1
map into a certain point and find the coordinates of this point.
13. The linear transformation T of three-dimensional space into itself maps
the point P into the point P' and the point Q into the point Q'. The points P and
Q are distinct. Show that
(i) if P' and Q' are distinct, then all points of the line PQ map into points
of the line P'Q'
(ii) if P' and Q' coincide, then all points of the line PQ map into P'.
14. The linear transformation T of the plane into itself has matrix
1 —2
A=(
1 — 1) •
Show that T maps the interior of the triangle 0(0, 0), P(2, 0), Q(1, 1) into the
interior of another triangle O'P'Q' and find the coordinates of the vertices
0' P', Q'.
Prove also that the image of the centroid of the triangle OPQ is the centroid
of the triangle O'P'Q' .
Miscellaneous Exercise 13
1. If (a b
A = c d) does (I = A (1
map a set of parallel lines into another set of parallel lines when ad * bc?
What happens to the transformation defined by A when ad = bc? Is it still
one-one? (A transformation is said to be one one if each image P' arises from a
-
unique P.)
2. What are the position vectors of the images of the points whose position
vectors are i, j, k under the linear transformation whose matrix, A, is given by
a2 a,
A= 1)2 1•3) ?
C1 C2 C3
4. A is the matrix of the transformation that rotates all points in the plane
through an angle a; B is the matrix of the transformation that rotates all points
in the plane through an angle Calculate BA and simplify the resulting matrix.
What result does this illustrate?
5. Can you find a point P with position vector x which is mapped into itself by
the tranformation whose matrix is A, where
( 8 13 \
A=
— 2 —2f •
Can you determine any real values of A which would enable you to find non-zero
vectors x such that Ax = Ax? If you can, find both A and the corresponding vectors.
6. Answer the same questions as in Question 5 for the linear transformations
with matrices
2 —1
(i) (3 5) —4 5)•
7 5)
7. The transpose of the matrix A, denoted by A', is defined as the matrix whose
ith row is the same as the ith column of A. For example, if
2), B = (42
1. —3 5
1
2
3
1),
1
A= ( 4
3 —5 — 2
then 1 1 2 4 3
A' = ( —3 4 and B' = 1 2 —5 .
25 3 1—2
If C and D are both 2 x 2 matrices, prove that (CD)' = D'C'.
8. A symmetric matrix A is a matrix such that A' = A. If A and B are two
symmetric matrices, find a further condition that A and B must satisfy to ensure
that AB is a symmetric matrix.
9. A square matrix is said to be diagonal if its only non-zero elements lie on the
leading diagonal (that is, top left to bottom right). For example, the following
two matrices are diagonal:
(0
20 0
(2 0)
0 —3 '
0 0 I
Prove that, if D1, D2are two 3 x 3 diagonal matrices, then
(i) D1D2is a diagonal matrix; (ii) D1D2= D2D1;
(iii) Di = D1.
If A is a 3 x 3 matrix and D any diagonal 3 x 3 matrix, what can you say
about A if AD = DA?
10. A linear transformation is given by the matrix
(1 3 2)
2 1 1 .
0 5 3
256
51 MISCELLANEOUS EXERCISES 1
Show that
(i) all points are mapped into points on a certain plane;
(ii) all points on the line
x -1 =
y— 1 z -1
1 3 —5
are mapped into the same point.
11. Let A, B, C be real 2 x 2 matrices and write
[A, B] = AB—BA, etc.
Prove that
(i) [A, A] = 0;
(ii) [[A, B], C]+ [[B, C], A] + [[C, A], B] = 0;
(iii) [A, B] = I [A, B1 = mIlm-1for all positive integers m.
At each step you should state clearly any properties of matrices which you use.
The trace Tr(A), of a matrix
A = (an a12)
a fi a22
is defined by Tr(A) = an+a22.
Prove that:
(iv) Tr(A+B) = Tr(A)+ Tr(B);
(v) Tr (AB) = Tr(BA);
(vi) Tr(I) = 2.
Deduce that there are no matrices satisfying [A, B] = I. Does this in any way
invalidate the statement in (iii)? (M.E.I.)
12. If M denotes the matrix /1 0
k1 2
and I denotes the matrix (10 0
1)
257
14. Matrices 2
Ex. 1. The linear transformation T maps points of the x-axis according to the
rule
T(x) = kx.
Interpret T in the case where k = 1.
Ex. 2. The linear transformation T maps points in a plane into points in the same
plane according to the rule
T(X) = Ax where A is a 2 x 2 matrix.
Interpret T in the case where
A = (1 0\
k0 11 •
Ex. 3. Interpret the matrix 1 0 0
I = (0 1 0
0 0 1
as the matrix of a linear transformation.
The question thus naturally arises: 'Is there a matrix analogue to the
reciprocal of a number ?' ; that is, `given a square matrix A, does there exist
a matrix B such that BA = I?'
We begin our investigation by actually constructing such a matrix B
for a given 2 x 2 matrix A. Suppose
(1 3
A=
2 7/ •
We effect the construction in two stages : we first form a matrix Clsuch that
258
1] THE INVERSE OF A MATRIX
1
the product C,A has as its first column ( ); we then form a matrix C2 such
0
that C2(C,A) = I. By associativity, it follows that C2 C, = B.
Since the element all of A is already 1, we have only to assume that C,A
makes the element of the second row and first column of the product zero.
Choose
1 1 3
Cl = 2_ ()) • then CI A = (0 /) .
1
Next choose
C2 = ; then C2(C,A) = (1 0
0 1)
Thus if we take
1) we have BA = I.
B = C2 C1 = (_ 27 —3
and the effect of pre-multiplication by C1has been to multiply the first row
of A by c (r; = cr1, r2 = r2, r3 = r3). Notice again that C1has been ob-
tained by performing the required elementary row operation on I. (The
reader may wonder why the restriction c 0 has been imposed; so far as
the result is concerned, this is of course unnecessary, but multiplication of a
row by zero must be specifically excluded when we use elementary row
operations to determine inverses.)
Ex. 6. Interpret C1as the matrix of a linear transformation.
(iii) The third elementary row operation: addition to any one row a con-
stant multiple c of another row.
Consider 1 0 0
M1 = (0 1 0) ,
0 c 1
( all a12 a13
MI. A = a21 a22 a23
a31+ ea, a32+ ca22 a33 ca2,
and we have a matrix whose first two rows are those of A, but whose third
row is obtained by adding to each element of the third row of A, c times
the corresponding element of the second row (r; = r1, r2 = r2, r,' = r3+ cr2).
260
THE INVERSE OF A -MATRIX
A=
G 7)
to a unit matrix in terms of pre-multiplications by elementary matrices.
We wish to effect the transformations
1 7
(2 3) (0
1 3
1) (01 0
1).
Example 1. If A= 3 4\
2/'
find B such that BA = I.
261
MATRICES 2 [14
(-1 2
4) ol 0
1)
—1 4) ri = r2
3 2 0
r,' = 1 1
0)
1 ri = r1 0 13)
(— 12
C) rz = r2+3r1 1
r1=5K
(— "
0 10
0) r2--= ra
= r,— r, 1 2
3 )
(-05 10
0) ra = r, 1
1 0\ ri = —r,/5 (
0 10/ r2' = ra 1 3
( 0
1 01) = r1 (
5 -5
= r2/10 1_ 3_
10 10
1
Thus B = I
A" A
Notice that, in the third transformation, we multiplied the first row by 5
to avoid introducing fractions for a little longer—an artifice worth re-
membering. With practice, the reader will learn to contract some of the
working, though it is best to write it out in full at first. It is always advisible
to check one's working by computing BA.
(12 —5\ /2 5\
(iii) A = (iv) A=
5 —2/ ' k3 9 .
Example 2. If
1 2 5
A 2 3 4),
= (1 1 2
find B such that BA = I.
262
11 THE INVERSE OF A MATRIX
1 2 5) 1 0 0
2 3 4 ( 0 1 0
( 1 1 2 0 0 1
(1 2 5) 1 0 0
0 —1 —6 r2 = r2 -2r1 1 0
1 1 2 (— 0
2 0 1
(1 2 5) ( 1 0 0)
0 —1 —6 —2 1 0
0 —1 —3 r; = r3 r, — — 1 0 1
1 2 5 ( 1 0 0)
0 —1 —6 —2 1 0
(0 0 3 r,,-' = r3 r2
— 1 —1 1
1 0 7 — I.,' = r1+2r2 —3 2 0
0 —1 —6 —2 1 0
(0 0 3 1 —1 1
(1 0 —7) (-3 2 0)
0 —1 0 r2 = r2 +2r3 0 —1 2
0 0 3 1 —1 1
1 0 0 K = ri +3r, — I — 1 1
0 — 1 0 0—1 2
(0 0 3 1 1 — 1
(1 0 0) 1
I -3 3)
0 1 0 rz = —r2 0 1 —2
0 0 3 1 —1 1
(1 0 0) 7)
- 3
0 1 0 0 1 —2
0 0 1 rg = ir3 1
3
2 1
3)
Thus B= 0 1 —2
B .
Ex. 10. Find, using elementary row operations, matrices B such that BA = I in
each of the following cases:
6 8 5 4 7 3
(i) A= (3 5 3); (ii) A= (2 5 2);
2 3 2 5 13 5
263
MATRICES 2 [14
4 3 5 4 1 1
(iii) A = (4 2 3); (iv) A = (3 2 3).
6 3 5 4 3 4
Having determined the inverses of some specific matrices, we next
consider whether all square matrices A have an inverse. The perceptive
reader may have observed that the process outlined above breaks down if
a complete row of zeros is obtained. In fact, if this occurs at any stage, A has
no inverse. Matrices for which no inverse can be found are called singular
matrices; matrices which have an inverse are called non-singular matrices.
The proof of a necessary and sufficient condition for singularity is deferred
to the next section.
Before leaving the consideration of elementary row operations it should
be pointed out that entirely analogous elementary column operations exist
and post-inverses (that is, matrices C such that AC = I) may be found by
post-multiplication by elementary matrices.
Exercise 14(a)
1. Find matrices B such that BA = I in each of the following cases:
(i) A = /2 1\ .
(ii) A =
/2 3\ .
(iii) A =
i 0 —1\
kP
3 2) ' k5 4 k -2 3)
(3 \. 1h h+1\
(iv) A = (v) A =
k2 7/ ' kh —1 h f•
2. Find matrices C such that AC = I for each of the matrices A of Question 1.
3. Show that the matrix /3 2\
k6
is singular.
4. Find matrices B such that BA = I in each of the following cases:
1 0 2) 4 8 3 2 2 —1
(i) 3 1 1 ; (ii) 3 5 1 ; (iii) (3 7 2);
(4 1 4 1 4 3 2 5 2
3 4 1 3 4 5 3 6 2
(iv) 2 3 1 ; (v) 4 3 11 ; (vi) (1 4 2.
(3 7 2 1 0 3 2 4 2
Verify in each case that AB = I.
5. Show that the matrix (1 2 3
2 3 1
1 1 —2
is singular.
264
1] THE INVERSE OF A MATRIX
1
may be written in the matrix form
3x+5y+4z = 3,
2x+4y+3z = 2
Av = w.
Pre-multiply both sides of this equation by B and hence solve the equations.
3x+2y+3z = 17,
1
8. Solve the equations
2x-3y +4z = —7,
2x— y+3z = 1
by the method of Question 7.
9. For what value of a is the matrix
(4 —3 —1
2 4 a
3 —5 —4
singular?
10. If AX = B where
(3 4 4 1 1 3
A = 1 3 2, B = 2 0 1 ,
2 3 3 3 1 —5
find X.
265
MATRICES 2 [14
2. DETERMINANTS
Recall that a square matrix A is non-singular if a (pre- or post-) inverse
may be found; otherwise it is singular. We now establish a simple criterion
for singularity for 2 x 2 matrices.
a31 a32 a3
then the determinant of A written
all an a13
a21 a22 a23 or, more briefly, det A (or IA l)
a31 a32 a33
is defined by
a22 a23 a21 a23 + a13 a21 a22
det A = - a12
a32 a33 a31 a33 a31 a32
a33 — all a23 a32 - a12 a21 a33 +a12 a23 a31 +a13 a21 a32 -
= an aa22 a a13 a22 a31
It is simply verified by direct calculation that det A may be expressed in
the two alternative forms
a12 a13 au a13 - a23 au a12
det A = -1121 + a22 a31 a
a32 a33 33 a31 a32
267
MATRICES 2 [14
or
a12 a13 an a13 an a12
det A = a31 — a32 + a33
a22 a23 a21 a23 a21 a22
These three forms give the expansion of the determinant by rows. The
determinant may also be expanded by columns; for example, expanding
by the first column
a22 a23 a12 a13 a12 a13
det A = all — a21 + a31
a32 a33 a32 a33 a22 a23
The term a22 a23
a32 a33
is called the minor of a„; similarly, the minor of a12 is
a21 a23
a31 a 33
(see expansion by first row) while the minor of a32 is
als
all
azl a23
(from the expansion by the third row). To find the minor of any term,
write down the determinant and cross-out the row and column containing
the term under consideration; the required minor is the 2 x 2 determinant
that remains. See Figure 14.1, where the minor of a2, is being sought: it is
au a 12
or (a a32 — a31).
a31 a32
an 012 an
a32 /233
the second row, or by the third column, we should include the term
all a 12
— a23
a31 a32
A minor, together with its correct sign as given in Figure 14.2, is called a
cofactor, the cofactor of the term ad being written A. Thus
all a13
A22 = = a33 — an;
a31 a33
an a13
A32 = — = a13a21— a11a23.
a21 a2.3
In terms of cofactors, the expansion of the determinant may be written
det A = all A11 +a12 Al2 a13 A13 (first row),
or det A = Al2 a22 A22 ± a32 A32(second column), etc.
Determinants of orders 4, 5, 6, ... may be defined similarly. Thus, for
the 4 x 4 determinant
all alt a13
a21 a22 a23 a24
1 1 0 0
a h g
2 1 3 4
(iv) (v) h b f
1 — 1 3 2
1 2 1
g f
3
269
MATRICES 2 [14
3. PROPERTIES OF DETERMINANTS
We shall now develop some simple properties of determinants, restricting
ourselves to the 3 x 3 case for simplicity, although the results and proofs
all extend to n x n determinants.
First observe that the interchange of rows and columns does not affect
the value of a determinant: for we may expand in the first case by the first
row and in the second case by the first column and, in either case, the
corresponding cofactors remain the same. Thus, any property possessed
by the rows of a determinant is possessed equally well by the columns.
For brevity, we shall denote the determinant
an a12 a13
a22 a23 by A.
a31 a32 a33
Property 1.
an +x +y ±z
a13 au. a13 x y z
a21 a22 a, a21 a22 a23 a21 a22 a23
a31 a32. a33 a31 a32 a33 a31 a32 a33
For
L.H.S. = (a„+ x) An+ (an+ y) Al2 ± (an + Z ) A13 = R.H.S.
One simple but useful result that follows immediately from property 2
can now be derived. It will be recalled that a„A„+a„A„+a„A„ = det A
and similarly for other rows and columns. Now suppose we multiply each
element of some row (or column) by the cofactors of the corresponding
elements of another row (or column); for example
all A21 + a12 A22 + a13A23.
Now
an a12 a13
all A21 a12 A22 +a13A23 = a31 a32 a33 = 0 (r„ = r3)•
an a12 a13
A similar result holds for any other combination of rows (or columns).
Such an expansion is called an expansion by alien cofactors and we have
demonstrated that expansion by alien cofactors gives the value zero.
A13 An A33
that is, the adjugate is obtained by substituting for each element its co-
factor and transposing (see Miscellaneous Exercise 13, Question 7).
Proof (3 x 3 matrices)
an a12 ai3 (All A21 a)
A31 A 0 0
A (adj A) = (21 a22 a23 Al2 A22 A32 = 0 A 0I,
a3, 32 a3, ) A13 A23 A3 0 0 A
since each term in the product is either an expansion of the determinant,
or an expansion by alien cofactors;
= AI, where A = det A.
A precisely similar result holds for (adj A) A.
Corollary. If det A + 0, det (adj A) = (det A)2 (3 x 3 matrix).
Proof Take determinants of both sides of A (adj A) = (det A) I and
divide both sides by det A.
It can be shown that, in fact,
det A = 0 det (adj A) = 0
and so the restriction that det A must be non-zero may be dropped.
If det A + 0 we have now obtained an inverse for A, namely
adj A
since A. adj A= I
det A' det A
Theorem 14.4.
_ adj A
det A
is the unique inverse of the non singular matrix A.
-
Proof. Suppose L is a left inverse other than A-1; that is, suppose that
LA = I, L A. Then
L = LI =- L(AA-1) = (LA) A-1 = = A-1,
which contradicts the assumption that L A-1. Similarly for right inverses.
Exercise 14(b)
1. Evaluate the following determinants:
1 3 2 -7
(i) (ii)
2 4 ; -1 3
6 5 -2 3
(iii) (iv)
5 4 ; -4 -5
4. Find the adjugate of each of the following matrices and, in each case, evaluate
the product of the matrix with its adjugate. Write down the determinant of each
matrix.
1 2 1 5 1 —2
3 7 5); (ii) — 2 5 4 ;
(i) (5 13 9 3 6 2
1 3 1 4 11 5
(iii) (3 10 5); (iv) (1 4 2) .
1 4 5 1 2 1
5. Find the inverses of the following matrices by the adjugate-determinant
method:
( 1 —2 1 2 5 1
(i) 3 —1 5 ; (ii) (1 0 3);
—1 4 0 3 1 2
3 7 2 2 4 1
(iii) (2 6 3); (iv) (1 4 2) .
1 4 2 1 2 1
8. 1 2 7 2 0 0
A= (0 1 2), B = (1 1 0.
1 2 8 0 2 —1
Solve for X the matrix equation
A2 +AX = B.
3 4 4 0 2 1
9. A= (2 1 5) , B = (1 1 0,
1 8 8 3 —2 1
1 2 1 3 1
C= (2
1 —2 2, D= 2 2 1 .
—1 -- 1 2 3 0 1
276
3] PROPERTIES OF DETERMINANTS
Miscellaneous Exercise 14
1. Prove that
11 (-1 0) (0 1 0) (a b) c d\
11 ) 01 1) (1 1 c d b)•
Deduce that the rows of a 2 x 2 matrix can be interchanged by operations
which add multiples of one row of the matrix to the other, together with opera-
tions changing the sign of a row.
Find a 2 x 2 matrix X such that
ba
(a
c =
2. Do 2 x 2 matrices A, B with integer entries, exist such that
(a) AB = 0, BA * 0?
(b) AB = BA = 0, A * 0, B * 0, A * B?
(c) AB = BA * 0, A, B * I, A * B?
In each case, if your answer is 'yes', justify it by giving examples of suitable
A, B. (M.E.I.)
3. If
a= (b)
a
11. If (a 0 0 d
0 b 0 e
A=
00 c f
d e f x
where a, b, ..., fare real and a > b > c, prove that the equation
det (A— xI) = 0
has real roots. (0 & C adapted)
12. If A and B are square matrices such that AB = 0, prove that either A = 0
or B = 0 or det A = det B = 0.
13. Two matrices A and B are said to commute if AB = BA. Let
A= (2 1o)\ •
prove that every matrix B that commutes with A can be expressed in the form
B = AA where A and it are scalars and I is the unit matrix. Obtain ex-
pressions of this form for A2 and A. (M.T.)
14. Prove that, if a is a 3 x 3 matrix, with adj a = A and det a = a, that
aA = Aa = aI.
Assuming that a * 0, prove that:
(i) det A = a2;
(ii) X = A is the only solution of the matrix equation aX =
(iii) adj A = aa.
Which, if any, of the above results, are still valid if the restriction on a is
removed?
15. If A is a 3 x 3 matrix with an* 0, and if the cofactor of every element of
A is zero (that is, if adj A = 0), prove that A has the form
an a12 a13
kale ka12 ka13 •
( /an 1 a12 Ian
16. What is meant by the statement that multiplication of real numbers is
associative and is distributive over addition?
Prove that, if crib bo co(i, j = 1, 2, 3) are real numbers, then
3 ( 3 3 3
E E aiabafl) = aja 1 E bafi c fli).
fl=1 cc = 1 a=1 fi = 1
279
MATRICES 2 [14
280
15. Linear equations
3x —y =
Fig. 15.1
Writing
1 a21x+ a22Y = b2.
(1)
and we have
(32 - 25) (;) = (1)
(x
y) = 119 (-2
523) CO
ix\ = 1 22 \
\ y) 19 k
The solution may be stated in the form x = y = — 9 or, in terms of the
linear transformation T whose matrix is A, the unique point --h) maps
into (4, 1).
284
2] TWO UNKNOWNS; CONTINUED
Fig. 15.2
M
J
Fig. 15.3
*Ex. 3. If det A = 0, show that every point P of the plane maps into a point of
the line OF.
Since, in this latter case, T maps every point P into a point of the line
01', a point B not lying on the line cannot be the image of any point under
T: the equations Ax = b are inconsistent.
However, if B does lie on 0/', we may write the position vectors of J',
I', B as r, mr, cr respectively; that is
Aj = r, Al = mr, b = cr.
Then, P(x, y) is a point which maps into B
cr = A(xi +yj)
cr = x(Ai) + y(Aj)
cr = (rnx+ y) r
ra P lies in the straight line y + mx = c,
and infinitely many points map into B, by the reversibility of the argument.
10-2 285
LINEAR EQUATIONS [15
Exercise 15(a)
1. Solve, using the inverse matrix, the following pairs of simultaneous equations:
(i) 3x— y = 1, 13x+2y = 1, I 70x + 105y = 12,
(ii) (iii)
2x+3y = 19; 2x+7y = 29; A 5x— 7y = 5.
2. Solve for x the following equations, commenting upon any special cases that
arise in the two cases:
(i) ax + a2 = b2— bx; (ii) ax+ b = bx+ c.
3. Solve for x, y the following pairs of equations, commenting upon any special
cases that arise for particular values of the coefficients a:
3x+ ay = —2, (ii\ f x+ 2y = 4, f 3x+ ay = 1,
{3x +2y = 3; 2x+ay = 2a; ax+ 12y = 2.
4. Solve for x, y the following pairs of equations, commenting upon any special
cases that arise for particular values of the coefficients a, b:
f ax+ by = 2, f ax+ 3y = 6; (iii) ax+by = 1,
t12x-4y = a+ b; (11) x — by = 2; tbx+ ay = — 1.
5. Discuss the solution of the following simultaneous equations for various
values of a:
x+ y = 1,
2x+ay = 5,
[ax+ y = —3.
6. Discuss the solution of the following simultaneous equations for various
values of the coefficient a:
fax +3y = 4,
3x— ay = 6,
x-2y = a.
286
2] TWO UNKNOWNS; CONTINUED
7. Discuss fully the solution of the following simultaneous equations for the
unknowns x, y:
f ax+by=c,
la2x+b2y = c2.
8. Under what conditions do the homogeneous equations
fai x+biy = 0,
ia2 x+b2y = 0
have a non-trivial solution (that is, a solution other than x, y = 0) ? Solve the
equations fully in the case where non-trivial solutions exist.
9. Solve for x, y the equations:
f x cos 0 +y cos cb = cos a,
x sin 0+y sin q5 = sin a.
10. Sketch the set, E, of points in the plane whose coordinates satisfy the
inequality x+y > 0. The linear transformation, T, has matrix A where
A= (1 2)
5 0
Sketch the image set T(E).
11. T is a linear transformation whose matrix is A, where
A = (2
1) •
If E is the set of points whose coordinates satisfy the inequality y > 0, sketch the
image set T(E).
have in common the single point (1, 1, —2)—a result readily derived by
successively eliminating two of the variables.
287
LINEAR EQUATIONS [15
2x— y+4z = 2.
If we write
L1 x+y+3z- 1, L2 E. x-2y+z-1, L3 2x—y+4z-2,
then it is readily seen that L3 = 4+4; thus, these three equations
represent three planes through a common line (see page 62) and so pos-
sess infinitely many solutions (one degree of
freedom); see Figure 15.5.
Again, the system
2x+y+3z = 6,
x+y+4z = 1,
Fig. 15.5
x— z= 1,
has no solution, for by subtracting the first two equations we obtain
x — z = 5, which is inconsistent with the third equation. Geometrically,
x — z = 5, which is a plane through the line of intersection of the two
planes 2x +y+ 3z = 6 and x+y+ 4z = 1,
is parallel to the third plane x—z = 1 and thus the line of intersection of
the first two planes is parallel to the third plane. A similar result would
hold whatever line of intersection was chosen and so the three planes form
a triangular prism of uniform cross-section; see Figure 15.6.
+ ai2Y+ an z =
f aux
anx+ a22 Y +a23z = b2, l (3)
1a31x+a32y+a33z = b3,1
b3
Again, the problem of solving a set of simultaneous equations may be
interpreted as the problem of finding the point P whose image, under a
given linear transformation, is a given point B.
If det A + 0 we have, as in section 2,
Ax = b
-4.> A-1AX = A-lb
x = A-lb;
the unique point P, whose position vector is A-1b, has image B under the
linear transformation T.
Ex. 7. Show that A-1is the matrix of a linear transformation S—the inverse
of the transformation T.
As an example of the solution of equations whose matrix, A, is non-
singular consider
x+ y+z = 4,
x— y—z = —2,
x+2y—z = 1,
{
1 1 1 i 0(by theusual
es inversion pro-
A 1 —1 — ; A-1 =
= (1 2 —1 + H-
and we have yx) = (4.
0 A- 1
0) (_ 24
(z —+ 1)
/1
=
The solution is x = 1, y = 1, z = 2 or, in terms of the linear transforma-
tion whose matrix is A, the point (1, 1, 2) maps into the point (4, —2, 1).
291
LINEAR EQUATIONS [15
The analysis of the case in which det A = 0 follows the same pattern as
the two-dimensional case and is pursued in the following series of exercises.
Ex. 8. Show that, under the linear transformation T whose matrix is given by
a, a2 a 3)
1)1 b2 b3
A = (c1 C2 Cs
the images of the points I, J, K, whose position vectors are i, j, k, are I', J', K',
whose position vectors are respectively the first, second and third columns of A.
Ex. 9. Show that, under T, the unit cube, three of whose edges are 04 OJ, OK,
transforms into a parallelepiped (a figure all of whose faces are parallelograms).
Ex. 10. Show that the volume of the parallelepiped obtained in Ex. 8. is det A.
(This result is necessary for the remaining exercises in this section, but its proof
is rather hard and may be omitted at a first reading. The best proof depends
upon the triple scalar product, which will be met in Book 2.)
Ex. 11. If det A = 0, show that 0, I', J', K' are coplanar. What does this tell us
about the columns of A?
Ex. 12. If the point D, with position vector d, is the image of a point P under
a linear transformation whose matrix is singular, what can be said about
(i) the point D; (ii) the vector d?
Ex. 13. Show that, if det A = 0, and if D lies in the plane OI'J'K', then all the
points of a certain line map into D.
Ex. 14. In the extreme case in which 0, I', J', K' are collinear, what can be said
about the columns of A? If Ax = d has a solution in this case, what can be said
about (i) the point D; (ii) the vector d?
In an actual example, det A should be calculated and the various cases in
which det A = 0 should be treated by reverting to the original equation.
Matrix methods are admirably suited to studying the structure of systems
of equations but rarely constitute an efficient procedure for their solution
in individual cases.
Thus det A = 0 if a = 2 or 3.
Case (i). a * 2, a * 3.
The equations have a unique solution for all values of b.
Case (ii). a = 2.
The equations reduce to
x+2y+ z = b, (1)
2x+3y+ z = 1, (2)
5x + 8y + 3z = 1. (3)
Eliminating z between (2) and (3), and between (1) and (2),
x+ y = 2, (4)
. x+ y = 1—b. (5)
Thus, the equations are consistent if and only if b = — 1, in which case we
have a one-parameter solution x = A, y = 2 — A, z = A-5 (on using equa-
tions (4) and (1)).
Case (iii). a = 3.
The equations reduce to
3x + 7y + 3z = 3b, (6)
7x + 9y + 3z = 3, (7)
{ 5x + 8y + 3z = 1. (8)
(6) + (7)— 2(8) gives 0 = 3b+3-2.
Thus, the equations are consistent if and only if b = --I, in which case we
have a one-parameter solution x = A, y = 2(1 — A), z = 3(11A — 15) (from
equations (7) and (8)).
5. HOMOGENEOUS EQUATIONS
We now consider the system of linear equations
an x+ any + anz = 0,
a21x+a22y+a23 z = 0,
.[
an x+ any+ a33z = 0,
or Ax = O.
Such a system is said to be homogeneous. If A is regarded as the matrix of a
linear transformation T we seek these points which are mapped into the
origin under T. It is easy to see that the origin maps into the origin under
any linear transformation; that is, that x = 0 is always a solution for a set
of homogeneous equations—we call this the trivial solution—but it is of
more interest to look for non-trivial solutions x + 0.
293
LINEAR EQUATIONS [15
1 3 1
x _y_ z
maps all points of the line
2 —1 1
into the origin.
What does this tell us about the system of homogeneous equations
1
3x+7y+ z = 0,
x- y - 3z = 0,
x+3y+ z = 0?
We conclude this chapter by proving a necessary and sufficient condition
for the existence of non-trivial solutions of the homogeneous equations
Ax = 0.
Theorem 15.1
Ax = 0 has a non-trivial solution a det A = 0.
Proof. (i) If det A + 0, A-1exists and x = 0 is the only solution.
Thus, if Ax = 0 has a non-trivial solution, then det A = 0, and we
have proved that a necessary condition for the existence of a non trivial
-
Exercise 15 (b)
1. Solve the equations {3x+ y+ z = -2,
2x+2y+3z = 8,
x+3y+2z = 6
(i) by successive elimination; (ii) by using the inverse matrix.
294
5] HOMOGENEOUS EQUATIONS
7.
A = (2 1 —1), B = (3 —1 2 .
1 —4 —1/ 5 —1 —2)
A maps the point P into the point Q and B maps the point Q into the point R,
whose coordinates are ( -5, 0, 0). Find the coordinates of P.
295
LINEAR EQUATIONS [15
Miscellaneous Exercise 15
1. Find the inverse of the matrix
( 1 0 3
—1 2 —1 .
2 —4 1
Given the equations 2x1-4x2 + x, = a,
xi +3x3 = b,
— xi+ 2x3 -- x3 = c,
find the solutions
(i) when a = 1, b = 1, c = —1; (ii) when a = — 1, b = 2, c = 0.
(M.E.I.)
2. Let A be a 3 x 3 real matrix and b a three-rowed column vector. It is proposed
to solve the equation
Ax = b
and a particular solution x = xois noted. Prove that any other solution may be
written in the form
x = xo +u,
where u is a solution of the equation Ax = 0. Prove conversely that any vector
xo+u is a solution of the equation, where xois a fixed particular solution of the
equation and u is any vector such that Au = 0.
Interpret geometrically the equation and its solutions if A is singular (i) when
xoexists, and (ii) when there is no such particular solution xo. (M.E.I.)
297
LINEAR EQUATIONS [15
3. If 2 3 6
M= 6 2 —3),
(3 —6 2
form the product MM' and show that M' = 49M-1.
Without multiplying out, state the product M'M, giving reasons for your
answer. Hence, or otherwise, find the solution of the equations:
2x1+3x2 + 6.; = 1,
6x1 +2x2 -3x3 = 1,
3x1-6x2 +2x3 = 2. (M.E.I.)
4. The simultaneous equations
x+ 2y = 4,
2x— y = 0,
3x+ y = 5
may be written in matrix form as
/2
1—2
1\ ix\ /4\
or AX = B.
li 5)
\y)= '21
300
16. Discrete probability distributions
1. INTRODUCTION:
THE UNIFORM DISTRIBUTION
In Chapter 7 we discussed the concept of an outcome space for a random
experiment. To each elementary event we ascribed a probability and the
entire set of probabilities was described as a probability distribution for the
outcome space. In this chapter we shall enumerate various possible prob-
ability distributions that find frequent use as mathematical models for
random experiments. We shall generally define our outcome space in terms
of an associated random variable X (see Chapter 10), and we shall thus be
able to talk about the mean and variance of the distribution; however, it
must be remembered that, for example, the mean is defined as e(X) and
so, if a new random variable, Y, is chosen, the mean naturally changes, too.
Unless there is any possibility of confusion, we shall refer to the mean and
variance of X as ,tt and cr2respectively.
Suppose we have an outcome space consisting of n elementary events,
with associated random variable, X, where
X = {1, 2, 3, ..., n}.
Perhaps the simplest assumption we can make about the n possible out-
comes is that each one is as likely to occur as any other. Thus we take
Pr (X = r) = (1)
Equation (1) defines the uniform distribution for n possible outcomes.
The uniform distribution is, in a way, fundamental, for we may often
subdivide the elements of an outcome space in such a way that the new
outcome space may reasonably be given a uniform distribution. This is not
always possible, however; a simple counter-example might be the fall of a
biased coin.
The mean, ,u, of the random variable X with a uniform distribution is
given by
= (X)
n 1
=Er .-
r=1 n
= +(n+1). (2)
301
DISCRETE PROBABILITY DISTRIBUTIONS [16
r2.1— 1)2
n
= +1) (2n + 1)— i(n + 1)2
= 1)• (3)
The reader will find a large number of examples on the uniform dis-
tribution at the end of Chapter 7.
Ex. 1. A man spins a coin and throws a die. For a head he scores —1, for a
tail +1, and this he adds to the score showing on the uppermost face of the die.
The values of the random variable, X, where
X = {0, 1, 2, ..., 7}
represent his total score. Why do you think a uniform distribution would be
an unsuitable mathematical model for this eight-point space? Suggest a way of
subdividing the sample space so that a uniform distribution would be suitable.
Example 1. Six unbiased dice are thrown. What is the probability of securing
three or more sixes?
The dice may be assumed to fall independently and so the random vari-
able X (= number of sixes showing) has a binomial distribution with
probability * of success on each of the six Bernoulli trials.
Thus
Pr (X?.... 3) = 1— p(X < 3)
= 1- NY + CO
0.06.
V 0 + (2) (:)402)
303
DISCRETE PROBABILITY DISTRIBUTIONS [16
We next derive expressions for the mean, ,u, and variance, 0-2, of the
number of successes in a binomial distribution. We shall give two methods,
the first one using the definition (4) directly, the second employing the
probability generating function (and using Theorem 10.5).
Method I. By definition
on writing r = (r-1)+1,
= n(n — 1) p2np — n2p2, since p + q = 1,
= npq.
Thus we arrive at the important result: the mean and variance of the
number of successes in a binomial distribution B(n, p) are given by
,u = np, Cr2 = npq. (5)
Method II. Since the p.g.f. for the binomial distribution is given by
G(t) = (pt +q)n
we have
G'(t) = np(pt G"(t) = n(n — 1) p2(pt +q)n---2(n ?„-. 2).
Putting t = 1, A = np, 0.2 ,u2— A = n(n— 1)p2and the results follow.
304
2] THE BINOMIAL DISTRIBUTION
Example 2. A very large number of balls are in a bag, one-eighth being black
and the rest white. Twelve balls are drawn at random. Find
(i) the probability of drawing three black balls and nine white balls;
(ii) the probability of drawing at least 3 black balls;
(iii) the expected number of black balls in the sample;
(iv) the most likely number of black balls in the sample.
Drawing a ball from the bag and noting its colour may be regarded as a
Bernoulli trial (a black ball drawn being a success, p = 1). Since a ball is
not replaced before its successor is drawn, the twelve Bernoulli trials
constituting our sample are not strictly independent; however, since the
number of balls in the bag is very large, we may assume that Pr (black ball
drawn) remains sensibly constant at *; that is, each trial is independent of
what has previously occurred. Thus the random variable
X = number of black balls contained in a sample of 12
has a binomial distribution B(12, *).
0182;
(iii) e(X) = np
= 1.5.
(12) nrn12-r .
(iv) Write p(r) = Pr (X = r)
r
12 r+1 /7\11-r
12 r
—
7(r +1)•
Thus P( +1)
r)1) < 1 if r 1
p(r)
12-0 12
and so p(1) = p(0) = — p(0)
7(0 +1) 7
is the maximum of the p(r); that is, the most likely number of black balls
is 1.
305
DISCRETE PROBABILITY DISTRIBUTIONS [16
Exercise 16(a)
1. A coin is tossed four times. Find the probability that heads appear
(i) at the first two tosses, followed by two tails;
(ii) just twice in the four throws;
(iii) at least twice.
2. 10 % of the very large number of articles produced by a machine are faulty.
What is the probability that a random sample of ten articles will
(i) be free of faulty articles;
(ii) contain more than two faulty articles?
3. From a packet containing a large number of seeds, 40 % of which are adver-
tised to give red flowers and the others white, 10 plants are produced. What is the
probability
(i) that all the plants have red flowers;
(ii) that all the plants have white flowers;
(iii) that half the plants have red flowers and half white?
4. 10 % of the very large number of articles produced by a machine are faulty.
If articles are taken at random and tested, how many articles will be tested, on
average, before the first faulty article is found? What is the probability that the
testing procedure will have to go on longer than this before the first faulty article
is found?
5. (i) In a trial, eight coins are tossed together. In one hundred such trials how
many times should one expect to obtain three heads and five tails?
306
2] THE BINOMIAL DISTRIBUTION
have been obtained, we reverse the process by asking how long we must go
on repeating the trial until a stipulated number of successes is obtained.
First, suppose we wish to calculate the probability that the first success
occurs at the rth repetition of a Bernoulli trial, the probability of success
in any one trial being p. (As in Section 2, we assume that each trial is
independent of what has previously occurred.) We take as our random
variable X, where
X = number of trials up to and including the trial which results in the
first success.
There is no upper limit to the value X may take: we can theoretically
continue obtaining failures for ever, although the probability of doing so
steadily decreases.
Now Pr (X = r) = Pr (initially (r -1) failures, followed by a success),
i.e.
Pr (X = =qr 1p. (6)
Equation (6) defines the geometric distribution for the random variable
X = {1, 2, 3, ...}. It is easily seen that this is a valid probability distribution,
for co
E qr-1
p =__ P
r=1 1—q
= 1.
The mean, ,u, and variance, o-2, of the random variable X for a geometric
distribution may be derived without undue difficulty: we give two possible
methods.
CO
Pq
= P+1—q
= 1, since p+q = 1,
,u = S(X) = llp.
qe(x 2) = E r2p—r
y = E (r-1)2 pqr-1.
r=1 r= 2
308
3] BINOMIAL DISTRIBUTIONS
P + 2(14 P)-177
= P + 2{(1/P) — P} — (1— P)
= (2/p)— 1,
o.2 = e(X2)—#2
2 1 1
= P —P2
= qip2•
It = 1 =p2 q
-
(7)
Method II. The p.g.f. for the geometric distribution is given by
G(t) = or)
= E pqr-i tr
r=1
= pt/(1—qt).
Thus G'(t) = 111(1 — qt)2, G°(t) = 2pq/(1—qt)3
and results (7) follow on substituting t = 1 and using Theorem 10.5.
The geometric distribution is a suitable probability model in a number
of practical examples; for example, in inverse sampling, articles are tested
until a faulty article is found.
observe that, if X = r, the rth outcome must be the kth success, probability
p, and the previous (k 1) successes are binomially distributed B(r 1, p).
— —
Thus
r —1) (p, r k)
Pr (X = r) =
P K-1 P q
= (k —1) ple qr—k k). (8)
E Pr (X = r) ak+.
r=k s=0
= G(1)
= 1.
Furthermore, G'(t) = kpktk-11(1—qt)k±1,
k(k —1+2qt) Pk ik-2
G"kt —/
1 qtyc-1-2
kg'
and it follows that = p, = (9)
310
3] BINOMIAL DISTRIBUTIONS
Ex. 6. Devise an experiment for which you feel the negative binomial distribu-
tion would constitute a suitable mathematical model.
Ex. 7. Dice are thrown in succession until two sixes have appeared; find
(i) the probability that two dice in all are thrown;
(ii) the probability that more than four dice are needed;
(iii) the expected number of throws required.
= e—aea
=1
and Pr (X = r) > 0, for all r, equation (10) does constitute a valid prob-
ability distribution.
We shall show that the Poisson distribution arises as the limit of a
binomial distribution in which n-->- cc and p--->0, in such a way that np = a
remains constant. Thus the Poisson distribution will form a suitable
probability model, not only to situations where the binomial distribution
applies directly and n is very large and p very small but also to situations
where the binomial distribution is not directly applicable. For example,
consider requests for trunk call connections made at a telephone exchange.
Suppose that, for the period 7.00 a.m. to 8.00 a.m. they average out at
three and suppose further that the request takes a negligible time to make.
Dividing the hour up into a large number of short time intervals (say 720
of 5 seconds each), the probability of there being a call in any one of the
intervals may be taken as 7T1T), on the assumption that the calls arrive
independently throughout the hour, and the probability of two calls in
the interval is negligible. Thus we have 720 independent repetitions of a
Bernoulli trial (`Is there a call or not in a given 5 second interval ?') with
small probability of success p = rzt-c), and the Poisson distribution may
be regarded as a good mathematical model to employ.
We must now verify our original assertion that a binomial distribution
311
DISCRETE PROBABILITY DISTRIBUTIONS [16
Pr (X = r) = (nr) pr qa-r
n(n— 1) (n — 2) ... (n —r + 1)
r! P 19)n (1 P)-r
1 \ 11 \ ... (1
n1 k
- 1 - 2 — r 1
1 ( n1 n )
ar pyr,
r! k
since np = a.
Since r is a fixed number, the product of the r factors in the numerator of
the first fraction tends to 1 as n -> oo, and (1 — p)-r tends to 1 as p 0.
Furthermore,
(1 — clI n—> e-a t
n
e-aar
and we have Pr (X = r)_>
r!
as required.
The Poisson distribution is completely described if we are given the
value of a; thus, a Poisson distribution is a one-parameter distribution and
a Poisson distribution with parameter a may be referred to as P(a).
The mean, it, and the variance, a.2, of the Poisson distribution P(a) may
be found by the direct computation of €(X) and ‘(2(2). Thus
= g(X)
ar
=fi r.
r=0
ao
ar-1
= 2.E
=1 (r — 1)! . a e-a
= a.
Similarly, 0.2 = 60(A72)__ 1,2
= a l as above.
For the Poisson distribution P(a),
p, = a, o-2 = a,
i.e. the mean and variance of the Poisson distribution P(a) are both equal
to a.
t For the proof of this and other results concerning the exponential function, see one
of the books on calculus mentioned in the bibliography.
312
41 THE POISSON DISTRIBUTION
Ex. 8. Prove that, if the random variable X = {0, 1, 2, ...} has equal mean
and variance, this does not necessarily mean that X has a Poisson distribution.
(Hint: consider a suitable uniform distribution.)
0.433;
(iii) Pr (X > 4) = 1 — e-4(1 + 4 +41 -Eci ,f
44
. 1)
0-371.
Exercise 16(b)
1. The random variable X can take values 0, 1, 2, 3, .... Given that X has a
Poisson distribution, mean 2, calculate the probabilities
Pr (X = 0), Pr (X = 1), Pr (X = 2), Pr (X = 3),
Pr (X = 4), Pr (X = 5), Pr (X > 5).
2. Samples of forty articles at a time are taken periodically from the continuous
production of a machine and the number of samples containing 0, 1, 2, ... de-
fective articles are recorded in the following table:
Assuming that this is the mean of the population and that the Poisson distribu-
tion applies, find the chance of:
(a) a sample containing four or more defectives;
(b) two successive samples containing between them four or more defectives.
(0 & C)
3. The following table shows the results of recording the telephone calls handled
at a village telephone exchange between 1.00 p.m. and 2.00 p.m. on each
of a hundred weekdays (e.g. on thirty-six days no such calls were made):
Calls 0 1 2 3 4 or more
Days 36 35 22 7 0
9. Mass-produced articles are taken at random from a batch and tested until a
faulty article is found. If the twenty-first article proves to be the first defective,
is this at variance with the assumption that 10 % of all the articles are faulty?
10. If X is distributed according to a Poisson distribution with mean A, write
down the probability that X = r.
How is this probability modified if the values X = 0 are unobservable? Prove
that the mean is now A/(1 — e-A), and find the second moment].about X = 0.
(Oxford)
11. Evaluate Pr (X = r+ 1)/Pr (X = r) for the Poisson distribution P(a). If
the numbers of misprints on pages of an uncorrected proof have a Poisson
distribution with mean 2.7, what is the most likely number of misprints to be
found on a given page?
12. The average proportion of bad eggs in an egg packing station is one in
2000. The eggs are packed in boxes containing six eggs each.
(i) Evaluate to two significant figures the probability that a box contains
one bad egg.
(ii) A housewife complains if she obtains two or more boxes, with one bad
egg each per hundred boxes. What is the probability that she complains.?
(M.E.I.)
13. During World War II, 537 flying bombs fell on south London. The dis-
tribution of the number of hits in 576 areas, each of 0.25 km2, is given in the
table. Compare the actual frequency with the theoretical frequency obtained by
assuming that the aim was effectively random and followed a Poisson distribution
with the same average number of hits.
No. of hits 0 1 2 3 4 5 6 or more
Frequency 229 211 93 35 7 1 0
(M.E.I. adapted)
however, here assume that such a sample has been drawn, and concentrate
upon the second part of the problem: inferring the quality of the population.
A simple approach would be to take a sample of size n, test each article
and reject the batch as sub-standard if, say, more than m of the articles
prove faulty. We begin by assuming that we have some percentage, 100p
say, as an upper limit to the number of faulty articles which may be allowed
before the manufacturing process is stopped and corrected for any fault.
Example 5. Samples of size 20 are taken from a large batch of articles pro-
duced by a machine; if more than two faulty articles are discovered the
batch is withdrawn, otherwise it is accepted. What is the probability that, if
the machine is producing 5 % of faulty articles, the batch will be accepted?
If the proportion of faulty articles rises to 10 % of the total output, what is the
probability the batch will now be rejected?
Since the batch is large, we may assume that the removal of twenty
articles does not sensibly alter the proportion of defectives and so the
binomial distribution B(20, 0.05) may be taken as a suitable mathematical
model for the experiment, where our random variable is the number, X,
of defectives in a sample of size 20.
(i) If Pr (article faulty) = -2-10-, then
Pr (not more than 2 faulty articles in 20)
/19\20+20 /19\19( 1 0 .1 (19\ 18 12
— k2o)20) k 20) +21.29 k20/ (20) (20)
0.925.
(ii) If Pr (article faulty) = -A-, then
Pr (more than 2 faulty articles in 20)
= 1(190)20_ ( 9 )19 ( 1 ) 20.19 ( 9 )" ( 1 )2
20 10 10 1.2 10 10
0.649.
Thus, if only 5 % of the articles produced are faulty, the chance of ac-
cepting the batch is about 921 %, whereas, if the number of faulty articles
increases to 10 %, there is a 65 % chance of rejecting the batch.
Ex. 9. Comment upon the efficiency of the test given in Example 5. Suggest
ways in which you feel it could be improved.
The testing procedure exhibited above is an example of a single sampling
inspection scheme. Any sampling inspection scheme is open to two types
of error:
(I) A batch with an acceptable number of faulty articles may be
rejected.
(II) A batch with an unacceptable number of faulty articles may be
accepted.
316
5] SAMPLING INSPECTION SCHEMES
Ex. 10. What objections are there to increasing the sample size in order to reduce
a and ft?
1.0
0.8
0.6
0.4
0.2
6. HYPOTHESIS TESTING
Probability theory finds a further application in the testing of statistical
hypotheses. Consider the following problem.
Example 7. A coin is tossed ten times and shows nine heads. Is this sufficient
evidence to support the claim that the coin is biased?
In order to attack this problem, we adopt the approach made in testing
scientific theories: we make a hypothesis and then consider whether or not
observational evidence supports this hypothesis. In problems of statistical
inference, our hypothesis is generally called the null hypothesis (because it
often takes the form of an assumption about the absence of bias in the
population): in this example, we shall take as our null hypothesis the state-
ment the coin is unbiased; that is, Pr (heads on any toss) = 2.
Having set up our null hypothesis we may ask the question 'Is the prob-
ability of obtaining a result as bad as or worse than, the observed result so
small that we are reluctant to ascribe it to chance ?' If it is, we agree to
reject our null hypothesis in favour of some alternative hypothesis.
To complete the framework of our solution we must further decide
what numerical value to prescribe to the indefinite phrase so small' ;
following common practice, we shall define p = -216as a small probability
(the so-called 5 % significance level'); in other cases it might be more
reasonable to take p = o o (` 1% significance level ').
Returning to the given example, our null hypothesis, H,, is given by
11,:p = 1 .
Our question now takes the form ' What is the probability of obtaining
a result as bad as, or worse than, nine heads with an unbiased coin ?' Since
we are comparing the hypothesis 1-10:p = 2 with the alternative hypothesis
111:p -I, we are simply testing for absence of bias and thus we have no
reason to distinguish between an apparently abnormal number of heads
or tails; it follows that a result `as bad as, or worse than, nine heads' is to
be interpreted as the result 'nine or ten heads or tails '.
We adopt as our probability model for the situation the binomial dis-
tribution. Thus we have
Pr (9 or 10 heads or tails/Ho) = 2[(2)1°+warm
47.
Since this is less than -A, we infer that the result is significant at the 5 %
level: we have reason to believe that the coin is biased.
Sometimes a test of a statistical hypothesis is constructed before the
observational evidence is obtained. For instance, in our last example, we
might have compared the null hypothesis:
Ho :P =
with the alternative hypothesis:
ability of any particular battery being faulty is small), with random variable
X, where
X = r when there are r faulty articles in a batch.
(2.4)r
Thus we have Pr (X = rIH0) = e-2.4 x
r!
and we wish to find the least integer n such that
Pr (X > nIH0) < 0.05.
Equivalently we have to find the least integer n such that
Pr (X < nIH0) 0.95.
By direct computation
Pr (X < 4114) = e-2.4 (1+2.4+ 2.2412 + 2.3413 + 2.414)
Exercise 16(c)
1. A machine is believed, on average, to produce 0.1% of faulty articles. Esti-
mate, to 2 D.P., the probability of finding a batch of five hundred articles free of
defectives
(i) by using a binomial distribution;
(ii) by using a suitable Poisson distribution.
A batch of five hundred is tested and found to contain two faulty articles.
Comment.
2. A large batch of manufactured articles is accepted if either of the following
conditions is satisfied:
(i) a random sample of 10 articles contains no defective article;
(ii) a random sample of 10 articles contains one defective article and a second
random sample of ten is then drawn which contains no defective articles. Other-
wise the batch is rejected.
If, in fact, 5 % of the articles in a batch to be examined are defective, find the
chance of the batch being accepted [(0.95)10 = 0.5987]. (0 &
3. From a batch of manufactured articles a sample of ten is taken and each
article is examined. If two or more articles are found to be defective the batch is
321
DISCRETE PROBABILITY DISTRIBUTIONS [16
P = (1 —P)9(1+9P).
Find an expression for P if it is now decided to modify the scheme so that when
one defective is found in the sample a second sample of ten is taken and the batch
rejected if this second sample contains any defectives.
In the second case, what will be the average number sampled per batch over a
large number of batches when p = 0.05? (M.E.I.)
5. State the formula for the probability that a variable following a Poisson
distribution of mean m takes the value r. Prove that the variance of r is m.
Past experience has shown that the number of defective items produced in a
shift by a certain machine is a Poisson variable of mean 4. A new employee in his
first shift produced six defectives. Is this clear evidence that he is operating the
machine inefficiently? (Cambridge)
6. The probability P(r) that there will be r damaged tomatoes in a crate can be
taken as
mr
P(r) =
r!
where m is the expectation of r. Over a large number of crates the value of in
has been found to be 10. In the first crate from a new supplier the value of r
was 4. Test whether this is significant evidence that the value of m for this
supplier is less than 10; explain carefully the logic of your argument.
(Cambridge)
7. A die is thrown six times. If a score of six is made on three of these occasions,
have you any reason for believing that the die is biased?
8. A cubical die with faces marked 1 to 6 is thrown n times. Show that on the
hypothesis that the die is unbiased the chance that the face marked 4 will appear
uppermost not more than once is p, where
1,1+5\ 15\n
P =
k 5 k6f
322
6] HYPOTHESIS TESTING
If n = 40 and the face marked 4 comes uppermost exactly once, test whether
the hypothesis that the die is unbiased is contradicted
(a) at the 1 % significance level;
(b) at the 0.1 % significance level. (Cambridge)
9. A bag contains ten balls, each of which is either black or white, but otherwise
the balls are indistinguishable one from another. Three balls are drawn without
replacement, and all are found to be white. Test the hypothesis
Ho : there is an equal number of black and white balls in the bag.
The three balls are now replaced and three more are drawn; these are found to
be two white and one black. Test the hypothesis Ho again.
10. A sampling inspection scheme is operated by taking a random sample of size
10 from each large batch of a product. The batch is rejected if more than one
defective is found and otherwise the batch is accepted. Plot the operating charac-
teristic of the given plan.
Explain the applications of operating characteristics when choosing a suitable
plan. (M.E.I.)
[You are advised to use a hand calculating machine for the first part, if you
have one available.]
11. A double sampling inspection scheme is devised as follows: a sample of size
15 is drawn from a batch of articles; if the sample contains none or one faulty
article the batch is accepted, if more than three, it is rejected. If it contains two
or three faulty articles a second sample is taken; the batch is now rejected if this
second sample contains more than one faulty article. If the batch contains
100p % faulty articles, find
(i) the probability that the batch will be accepted;
(ii) the value of p which gives the largest expected sample size;
(iii) the size of this largest expected sample.
12. An assembled instrument contains two critical components A and B. Sample
tests show that we may expect one in ten of A and one in eighteen of B to be
defective. Estimate and compare the costs of the following inspection plans,
per hundred fully tested instruments:
(i) to test every component before assembly at a cost of 2p for each A tested
and 3p for each B tested; or
(ii) to test every instrument after assembly, if this test adds nothing to the cost
but making good a defective instrument costs on average 24p. (M.E.I.)
Miscellaneous Exercise 16
1. A pack of cards is cut and the suit of the exposed card noted; the pack is then
well shuffled and the process repeated. Find
(i) the probability that a spade will appear for the first time on the fourth cut;
(ii) the average number of cuts required before the first spade appears;
(iii) the average number of cuts required to expose cards of all four suits.
2. Show that the probability of r successes in n independent trials is the coefficient
of tr in the expansion of (q + pt)n , where p is the chance of success in a single
trial and q = 1—p. Prove that the mean number of successes is np.
323
DISCRETE PROBABILITY DISTRIBUTIONS [16
Samples, each of eight articles, are taken at random from a large consignment
in which 20 % of the articles are defective. Find the most likely number of de-
fective articles in a single sample and the chance of obtaining precisely this
number.
If a hundred samples of eight are to be examined, calculate the number of
samples in which you expect to find three or more defective articles. (0 & C)
3. Two coins are identical in appearance, but one is unbiased while the other
gives, on the average, heads three times as often as it does tails. One of the coins
is taken at random and tossed four times. If two heads and two tails appear,
what is the probability that it is the unbiased coin?
4. Explain briefly what is meant by a Poisson distribution of rare events and its
relation to the binomial distribution. Prove that the mean of the distribution is
equal to its variance.
A shopkeeper's sales of washing machines are four per month on the average.
Assuming that the monthly sales fit a Poisson distribution, find to what number
he should make up his stock at the beginning of each month so that his chance
of running out of machines during the month will be less than 4 %. (0 & C)
5. A machine depending for its energy upon four complexes of solar cells, each
complex functioning independently of the others, will work provided one of the
four complexes is working. Each complex has probability p of failing. The
machine is redesigned to have six complexes and will function if two of the six
complexes are working. Is the new design an improvement on the old?
6. A and B play N games, each of which must result in a definite win to one or
other player. A's chance of success in any one game is p. For his rth win, A
receives from B £r, for his sth loss, he gives B £s. Find A's expected gain.
7. A coin is tossed repeatedly an even number, 2n, times. Show that, whatever
the bias (provided Pr (H) = 0 or 1) the probability of obtaining the same
number of heads as tails decreases as n increases.
8. In lawn tennis a set is won by the first player to win six games, except that
if the score reaches 5-5 the set is won by the first to lead by two games. Two
players have chances respectively p and q of winning in any game (p+q = 1);
games may be treated as independent. Find the chance that a set lasts exactly
2n + 2 games (n 5). (Cambridge)
9. The number of eggs laid by an insect has Poisson distribution with mean it.
If the probability that an individual egg survives is p, show that the number
of eggs surviving has Poisson distribution, and determine its mean value. (You
may assume that the survival of an egg is independent of the fate of the other
eggs laid.)
10. An experiment has probability p of success. In n independent trials, pois the
probability of an even number of successes (pa = 1). Prove that
Pn-(1-2P) P.-1 = P.
If f(t) = i po,r, prove that
n=0
f(t) = [1- (1- p) t]l{(1- t) [1 - (1-2p) t]}.
Deduce that Pn = i0 + (1 -2P)ni.
324
61 MISCELLANEOUS EXERCISE 16
14. A bag contains a large number of red, white and blue dice in equal numbers.
If n are drawn at random, show that the probability P(n, r) of drawing exactly
r red dice is equal to the term containing (-Dr(4)n-rin the expansion of ef + ip.
If r dice are thrown, find the probability Q(r, s) of throwing exactly s sixes.
If n dice are drawn from the bag and the red dice drawn are thrown, show
that the probability of throwing exactly s sixes is
n-s
I P(n, s+ t) Q(s+ t, s)
t=o
and prove that this is equal to a term in a binomial expansion. Explain why a
binomial distribution is obtained. (C.S.)
17. The number of a certain type of organism in a given volume of water has
Poisson distribution with mean 2. A test, applied to indicate absence of the
organism has a 90 % chance of success if the organism is in fact absent, but also
indicates absence in 10 % of those cases in which they are present. If the test is
applied and indicates absence, what is the probability that the water is free of
the organism?
18. The probability that a source emits r a-particles in a given time is propor-
tional to ArIr!, where is a constant. Obtain the constant of proportionality
and calculate the mean of r.
The probability of the same source emitting s fl-particles in the same time is
proportional to vs/s!. Assuming that the two types of particle are emitted in-
dependently, write down the probability that r a-particles and s fl-particles are
given off in this time and show that the probability that a total of n particles of
the two types are emitted is
,-(1.44-v) + 07,1n !. (Cambridge)
19. The probability that any randomly chosen rat from the colony used in a
certain laboratory will show a certain undesirable characteristic is p; it is known
that the value of p for the colony is either 0.4 or 0.6 and it is desired to set up a
sampling scheme to decide which value is correct. The procedure is to take rats
one at a time at random and test for the presence of the characteristic; after n
rats have been tested, let r be the number with the characteristic.
Show how to determine limits L„ and U„ such that for known n
5 7 9 11 13 15 17
3 317 580 768 881 942 973 988
4 87 290 517 704 831 909 954
5 10 96 267 467 647 783 874
6 19 99 247 426 597 736
7 2 25 99 229 390 552
8 4 29 98 213 359
9 0 6 32 95 199
10 1 8 34 92
11 0 1 9 35
20. At a certain seed testing station it is found that a proportion 0.4 of a certain
type of seed is fertile. By accident the remaining stock of this seed (whose total
amount is very large) is completely mixed with an equal quantity of a second
type of seed which is believed to be completely infertile. If this latter assumption
is true, what is the probability that a seed taken at random from the mixture
will germinate?
Each of seven pots is planted with two seeds taken at random from the mixture.
Six pots eventually produce one or more plants each. Is this result consistent, at
the 5 % level of significance, with the infertility postulate for the second type of
seed? (Cambridge)
21. Derive the probability p(r) of obtaining exactly r successes in n independent
trials, the probability of a success being p at each trial. Determine the mean and
variance of this distribution.
Find an expression for [p(r)]/[p(r I)] and hence or otherwise find the con-
—
ditions that rff, must satisfy if p(r„) is such that no other value of p(r) is greater
than p(rm).
Show that the mode of a binomial distribution differs from the mean by less
than unity. (M.E.I.)
327
Revision exercise B
1. Find the equation of the line L1through the origin with gradient 1, and
also the equation of the line L2 perpendicular to L1and passing through the point
(4, -1).
2. In a large university, one-third of the men and one-quarter of the women
read science. If four men are selected at random, what is the probability that not
more than one reads science? If two men and two women are selected at random
what is the probability that not more than one of the four reads science?
What is the probability that, in a mixed group, one man reads science and the
other man and the two women read something else?
3. A square lamina ABCD of side a is held with the corner A on a horizontal
plane. The feet of the perpendiculars from B, C, D on to the plane are B', C', D'
and the angles B' AB, D' AD are a, /3. The angle B'AD' is O. Prove that
(i) cos 0 = - tan cc. tan )3;
(ii) the area of the triangle B'AD' is -1a2(cos2 a cost sine a sin2 fl)1;
(iii) the inclination of the lamina to the horizontal can be expressed in the
form
arccos [cos (cc+ fi) cos (a-/3)]t. (0 & C)
4. If a + -14(4n +1)7r, find x if sin (x+ a) = cos (x- a).
What can you say about x if a = 1(4n +1)70
5. Find the sum to infinity, S, of the geometric series
1 1 1
1 + -+ +
A/10 10 1000
If SN denotes the sum of the first N terms of this series, find the least value of
N such that S and SN are the same correct to 3 D.P.
6. A tennis match usually consists of either three or five sets, and ends when one
side has won a majority of the sets. If the probability of a side winning a set is p,
and if the result of each set is independent of any previous results, show that the
probability of a match going its full legth is 2pq in the case of a three-set match
and 6p2q2in the case of a five-set match (q = 1-p).
Show that the first probability is always greater than the second, if p 0 or 1.
7. The function f: R -> R+ is defined by
f(x) = Ix+1I+ I2x-11+
Sketch the graph of f and determine the least value of f(x).
8. Call a matrix of the form cx xy)
328
REVISION EXERCISE B
9. Factorize into linear factors the expression 2x2 + 5xy— 3y2— 3x+ 5y — 2 and
describe geometrically the set of points {(x, y) : 2x2 + 5xy — 3y2— 3x + 5y — 2 = 0}.
Describe geometrically the sets of points
(i) {(x, y): x2 — 2y2 = 0); (ii) {(x, y) : x2 + 2y2 = 0).
10. i, j, k are unequal positive integers and x, y, z are real numbers. Prove that
x(j— k) + y(k — i) + z(i—j) = 0 <r> x, y, z are respectively the ith, jth, kth terms of
an arithmetic sequence.
13. What are the first three terms in the expansion of (1 — 2x)24in ascending
powers of x?
Find the value of 0.9824, correct to 4 D.P.
16. If A is the point (a cos a, b sin a) and B is the point (a cos /3, b sin /3), show
that the equation of the line AB is
x a+ fi y a+13 fi
Q cos sin = cos
2 b 2 2
What is the connection between a and /3 if AB passes through the origin?
17. Prove that the number of spheres that can be drawn to pass through three
given points and touch a given plane is 2 or 1 or 0, explaining how the three cases
arise.
18. OA, OB, OC are three concurrent straight lines lying in one plane. P is a
point outside the plane such that the angles PO A, POB, POC are equal. Prove
that PO is perpendicular to the plane.
19. Find the matrix of the linear transformation which reflects all points in the
plane x+y+z = 0.
20. ABCDA'B'C'D' is a parallelepiped, with opposite faces ABCD, A'B'C'D'
and edges AA', etc., AB = BB' = B'C' , and the mid points of BB', A' D', A'B'
-
are respectively F, G, H. Find the ratio in which the plane AD' H divides the line
FG.
21. Two boxes each contain one hundred cards, numbered 1 to 100. A card is
taken from each box, the numbers on the two cards noted, and they are then
returned to their respective boxes. If the process is repeated fifty times, find the
probability that at no stage a pair of cards bearing the same number will be
drawn. How many draws are needed for the probability to be >1?
22. Prove that lg 15 is irrational. (lg 15 means log„ 15.) Is it true that lg x is
irrational a x is not a power of 10?
23. Evaluate the following determinants:
b— c c— a a— b (x-p)2 (y-p)2 (z-p)2
(i) c— a a— b b — c ; I 0,
(11)
a—b b— c c— a (X r)2 (y— r)2 (z— r)2
24. Define the projection of a vector in a given direction. ABCD is a quadri-
lateral in which LA = LC = 90°. The feet of the perpendiculars from B, D
to AC are X, Y respectively. Prove that AX = CY.
25. In a large crate of oranges, 100p % are bad. A random sample of ten oranges
produces two bad ones. On the assumption that this is the most likely number
of bad oranges to find in the sample, what can you say about the possible values
of p?
26. ABCD is a square lying in the plane 2x— y — 2z = 5. If A has coordinates
(1, 1, —2) and C has coordinates (5, 5, 0), find the coordinates of B and D.
27. The function, f, defined by
f(x) = sin x+ cos x
has for its domain the set of real numbers {x e R: 0 < x < zr}. Find the range
off.
330
REVISION EXERCISE B
28. Factorize the expression ax2+(ac+ b — c) x+ c(b— c) into two linear factors.
Hence factorize the expressions
(i) a+b—c+bc+ca—c2;
(ii) a — b + c+bc— ca— c2;
(iii) a3b2+ a2bc + ab2— abc + bc — c2.
29. Describe geometrically the transformation of three-dimensional space repre-
sented by premultiplying the position vector of the point (x, y, z) by the matrix
0 1 — 12)).
—0
Which points are left unaltered by this transformation? (O.S.)
30. When A and B play chess the chance of either winning a game is always
+ and the chance of the game being drawn is always 4. Find the chance of A
winning at least three games out of five.
If A and B play a match to be decided as soon as either has won two games,
not necessarily consecutive games, find the chance of the match being finished
in ten games or less. (J.M.B)
31. Solve the equations
(i) 2 sine x+ sin 2x = 0;
(ii) sin x = cos (-Pr— x);
giving all solutions in the interval 0 x ‘. 2ir in each case.
32. Solve the inequality x2 -1 1
x2 -4 5
and illustrate your solutions by means of a sketch of the curve
x2—1
Y — x2-4 '
33. The base AB of a triangle ABC is fixed and K is a fixed point on AB. The
vertex C of the triangle moves so that the perpendicular distances of K from CA
and CB are always equal in length. Prove that, in general, the locus of C is a
circle through K.
What is the exceptional case? (Cambridge)
34. A pitcher is taken daily to a well and back. Its chance of being broken on
an outward trip is pi. If it survives the outward journey it has a further chance p2
of being broken on the return. Show that on any day its change of being broken
is P, where
P = Pi+P2 —PiP2.
Show that the chance that the pitcher will survive for at least n days is Q",
where Q = (1—p1) (1 —p2), and find the chance it will survive (n-1) days but
be broken on the nth day.
If two such pitchers are each taken independently to the well and back daily,
prove that the chance that they survive for exactly the same number of days is
P1(1+ (C.S.)
331
REVISION EXERCISE B
35. Find the image by reflection of the point (4, 3, 1) in the plane
3x+2y+3z+ 1 = 0.
36. A and B are two towers, B being four miles due east of A. The true bearings
of a flagpole C from A and B are a° east of north and a° west of north respec-
tively; the true bearings of a second flagpole D from A and B are (cc + fi)° east
of north and (oc— j3)° west of north respectively. Draw a sketch-plane to indicate
the positions of A, B, C, D.
Assuming that A, B, C and D are on level ground, prove that D is 4 sine f3 cosec 2a
miles south of C and 2 sin 213 cosec 2a miles east of C. (0 & C)
37. If ax4+ 3x3+ bx2+ cx+ 2 is exactly divisible by (x+2) (x2 -1), find the
values of a, b, c.
40. Using the formula for tan (A+ B) in terms of tan A and tan B, show that,
if A, B, C are the angles of a triangle, then
tan A + tan B+ tan C = tan A tan B tan C.
Calculate, in degrees and minutes, the angles of a triangle ABC if
tan A: tan B: tan C = 1:2: — 6. (Cambridge)
41. Find the equation of the plane containing the origin, the point (1, 1, 1)
and the point (3, 1, — 1). Find also the direction ratios of the line of intersection
of this plane with the plane x+y—z— 4 = 0.
332
REVISION EXERCISE B
50. Find the image by reflection of the plane x = 0 in the plane x+y+z = 4.
51. Using the same axes draw the graphs of y = sin x° and y = 1-270 for
values of x from 0 to 360. Estimate from your diagram the roots of the equation
x
sin x° + — = 1.
270
The equation sin x° = mx+ c, where m and c are constants, has roots x = 60,
x = 330. With the aid of your diagram estimate
(i) the other root of the equation;
(ii) the values of in and c. (Cambridge)
52. Prove that, when any polynomial f(x) is divided by (x— a), the remainder
is f(a).
The remainder, when f(x) is divided by (x— a) (x— b) is written in the form
A(x— a)+ B(x— b); prove that
f(b) f(a)
A= B =
b— a ' a— b
When ax3+ fix2+ yx+ a is divided by x2 — 1, the remainder is Kx where K is
a constant ; when it is divided by x2— 4, the remainder is K. Prove that
a = —fi = —1,7 = 6 = —iK. (0 & C)
53. Let A(0) and B(0) denote the matrices
cos 0 sin 0 cos 0 — sin 0)
sin 0 — cos 0)' (sin 0 cos 0
respectively.
(i) If
(c,) = A(0) ( y)
show that the point (X', Y') is the mirror image of (X, Y) in the line y = x tan -I-0.
(ii) Prove that
A(0,) A(82) = B(95),
where 95 is some angle (to be determined), and hence, or otherwise, explain the
relation between the points (X', Y') and (X, Y) when
p\ = we)
Y'f y) •
54. One side of a hill has the form of an inclined plane, the line of greatest
slope, AB, being due north. A man starts from A and walks to B, the distance
being d1km. At B he walks for a distance d2km in a straight line, inclined at
an acute angle 0 to AB, reaching a point C. The vertical heights of B and C
above the level of A are h1km and h2km respectively. Prove that
di(h2 — hi)
cos 0o = -
h1d2 •
The bearing of C from A is g5 east or west of north. Prove that
(.-4 sin 0
tan 95 = (0 & C)
h2A1(42— h'f)•
55. Show that, for all but two specific values of k, the equations
x— ky+2z = 3,
3x+ y— 2z = 1,
k x — 9y+2z = 7,
represent three planes with a s'ngle common point.
Interpret geometrically the two special cases.
56. In how many different ways may a red balls, b white balls and c black balls
(indistinguishable apart from colour) be arranged in a straight line?
Find the coefficient of x4y4Z4in the expansion of (x+ y+2z)12.
57. Solve the inequality 5x + 1
< 1
x2 — x-6
and illustrate your solution by drawing a sketch of the curve
5x +1
Y = x2—x-6'
58. Three spheres of radii 1, 2, 3 units, touch each other externally. Prove
that a plane which touches each sphere makes an angle arcsin V13) with the
plane containing the centres of the spheres. (0 & C)
59. Find all values of x, y and z which satisfy the equations
—y+z = u,
x —z = v,
—x+y = w,
where v-2w = a,
—u +3w = b,
2u-3v = c. (C.S.)
60. If I deal cards one by one from a well-shuffled pack, find the expectation of
the number of cards I shall have to deal
(i) to secure a spade; (ii) to secure all the spades.
335
REVISION EXERCISE B
61. ABCDA'B'C'D' is a rectangular box with faces ABCD, A' B'C' D' and edges
AA', etc. AA' D' D is a square of side 1 unit, AB = 2 units and M is the mid-
point of C' D' . By setting up a suitable coordinate system, or otherwise, find the
acute angle between the planes BMA', AB' D' , correct to the nearest degree.
62. Prove that (a— x)2 (a— y)2 (a— z)2 (a — w)2
(b — x)2 (b — y)2 (b — z)2 (b — w)2
(c (c (e, (c —w)2 = 0. (C.S.)
67. The points P1and P2on the surface of the earth (assumed to be a sphere of
radius r and centre 0) are at the same (northern) latitude A, and their respective
longitudes are L1and L2. Show that the length C of the route from P1to P2 via a
circle centre 0 (called a great circle) satisfies the equation
elements with probabilities po,pi,p2, ... show that the probability PT of r progeny
is given by the coefficient of i r in G(t), where
G(t)= E p„
rty
n=0
When
Pn" = 2n+1
find the probability that the number of progeny is greater than or equal to N.
By considering G(t) and its derivatives, or otherwise, find T., the expected value
of r, the number of progeny, and the expected value of (r— r)2. (O.S.)
72. A, B and C are the three angles of a triangle. Show that
sin A sin B sin C
cos A cos B cos C = 0. (C.S.)
sin3 A sin3 B sin3 C
73. If 11 is a positive integer and p a prime number, ccp(n) denotes the greatest
integer k such that plc divides n. If n is written in the form
N
n = E ary (0•--5 a, p-1),
r=0
N
show that n— a,
a(n!) = r=0 (C.S.)
p —1 •
74. ABCD is a parallelogram and E a point not necessarily in the plane ABCD.
Show that a2 +c2 +g2 +h2 = b2+ d 2 +e2, these being the lengths shown in the
figure, and find a relation involving only a, b, c, e, f. (You may use vector
geometry.) (C.S.)
338
Bibliography
GENERAL
Allen, R. G. D. Basic Mathematics (Macmillan). A broad survey of mathematical
ideas—useful as a reference book.
Backhouse, J. K. and Houldsworth, S. P. T. Pure Mathematics (Longmans,
2 volumes). Written some years ago, with traditional 'A' levels in mind—
consequently rather old-fashioned, but workmanlike in its approach and
contains a great number of examples for the student.
Kemeny, J. G., Murkill, H., Snell, J. L. and Thompson, G. L. Finite Mathe-
matical Structures (Prentice-Hall). An introductory survey of some modern
topics, containing a much more thorough account of the ideas sketched in
Ch. 9.1 and also including chapters on linear algebra and probability.
S.M.P. Advanced Mathematics. (Cambridge, 4 volumes). Specifically written
for the new S.M.P. 'A' level, these books are full of new ideas, presented in
a lively and stimulating manner.
ALGEBRA
Beckenbach, E. F. and Bellman, R. An Introduction to Inequalities (Random
House.) A readable introduction to the subject.
Birkhoff, G. D. and Maclane, S. A Survey of Modern Algebra (Macmillan,
New York). Really a university text book, but beautifully written and con-
taining much that can be read with profit by sixth-formers. Gives a much more
thorough account of the ideas outlined somewhat sketchily in Ch. I.
Cohn, P. M. Linear Equations (Routledge and Kegan Paul). A sound introduc-
tion at low cost.
Durell, C. V. and Robson, A. Advanced Algebra (Bell, 3 volumes). Written a
long time ago, but these books wear well and still contain a great deal of valu-
able reading; they also contain an excellent selection of problems, many of
which are stimulating and challenging.
Ferrar, W. L. Higher Algebra for Schools (Oxford). A good introductory course
in traditional sixth-form algebra.
Maxwell, E. A. Algebraic Structures and Matrices (Cambridge). Carries the work
beyond 'A' level standard; contains some good exercises.
Neill, H. and Moakes, A. J. Vectors, Matrices and Linear Equations (Oliver and
Boyd). Covers the course of linear algebra required for the M.E.I. 'A' level
examinations. Approaches the subject from a rather different standpoint to
that adopted in this book.
Niven, I. Numbers: Rational and Irrational (Random House). A comprehensive
account of the work of Ch. 1.
339
BIBLIOGRAPHY
CALCULUS
Hardy, G. H. Pure Mathematics (Cambridge). A very famous book. Not easy
reading, but repays careful study. A work for the specialist.
Matthews, G. Calculus (Murray). An excellent modern text.
Maxwell, E. H. Calculus (Cambridge, 4 volumes). A comprehensive text, useful
as a work of reference.
Quadling, D. A. Mathematical Analysis (Oxford). An excellent introduction to
university work: presupposes some acquaintance with the calculus.
Scott, D. B. and Timms, S. R. Mathematical Analysis (Cambridge). A university
text, but contains much good reading for the serious sixth-form student.
Siddons, A. W., Snell, K. S. and Morgan, J. B. A New Calculus (Cambridge,
3 volumes). A standard text-book; volume 3 is a particularly good introduc-
tion to advanced work.
Toeplitz, 0. The Calculus, A Genetic Approach (Chicago). A stimulating ap-
proach to calculus by tracing its historical development.
PROBABILITY
Feller, W. An Introduction to Probability Theory and Its Applications, Volume 1
(Wiley). An advanced text, but highly recommended for the serious post-
'A' level student.
Mosteller, F. Fifty Challenging Problems in Probability (Addison-Wesley). Some
good material here for the problem addict.
Mosteller, F., Rourke, R. E. K. and Thomas, G. B. Probability and Statistics
(Addison-Wesley). A very elementary introduction to probability—wordy at
times, but thorough.
Parzen, E. Modern Probability Theory and its Applications (Wiley). A very
thorough introduction, carrying the subject well beyond 'A' level.
340
BIBLIOGRAPHY
but can be read by the specialist with profit. Strongly recommended to all
sixth-formers.
Smith, D. E. A Source Book in Mathematics (Dover, 2 volumes). Contains
excerpts from the writings of great mathematicians of the past—all within the
understanding of the sixth-former.
Weston, J. D. and Godwin, H. J. Some Exercises in Pure Mathematics (Cam-
bridge). A fine collection of problems, based on the Welsh 'A' level (alterna-
tive syllabus). Contains detailed solutions.
342
Answers
CHAPTER 1
PAGE
Exercise 1(a)
1. (i) 10; (ii) 1110; (iii) 10000; (iv) 101.
2. 100010011, 101012, lte.
3. 8e242. 4. 3e, remainder 19.
5 Ex. 7. Yes, unless a2 = b2 0.
7 Ex. 12. (i) 7/9; (ii) 134/333; (iii) 283/45.
Ex. 13. (i) 11.11; (ii) 0101101; (iii) 1001001;
(iv) 0.11010001111010111000.
Ex. 15. (i) 0.3; (ii) 1-714285.
Exercise 1 (b)
9 4. (i) 0.3; (ii)
6. (i) True; (ii) True; (iii) False; (iv) True; (v) True (provided rational
number is not zero); (vi) False.
11 Ex. 17. (i) 242; (ii) 3A/2; (iii) 3A/6; (iv) 500; (v) 42 \/3.
Ex. 18. (i) 11+4.N/6; (ii) 3J6; (iii) 2A/3; (iv) 8.
Ex. 19. (i) 1A/3; (ii) (2+ V2); (iii) M2+ V2); (iv) 3+ A/6; (v) 7-4V3.
Exercise 1(c)
12 1. (i) 5 A/2; (ii) 11A/3; (iii) 20A/6; (iv) 8A/3; (v) 22A/3.
2. (i) 6-2V5; (ii) 14-4A/6; (iii) 10+6.0; (iv) 9A/3 -11V2;
(v) 28-16V3.
3. (i) 2,8/2; (ii) 3A/5; (iii) 5 \/3; (iv) A/13; (v) (3A/3)/2.
13 4. (i) V5/5; (ii) V21/3; (iii) 2(A/7+2)/3; (iv) (13+4V3)/11;
(v) V6+ V3- A/2 2. -
343
ANSWERS
PAGE
Exercise 1(d)
20 2. (i) R; (ii) {x e R: 3 < x < 4); (iii) {x e R: x < 4}.
3. (i) {x e Z: x < 0); (ii) {x e x/6 e Z+};
(iii) {x e Z: x< - 3 or x 3).
4. (i) {x a R: 2 < x < 3}; (ii) {x E R:x > 3};
{x e R: -1 < x< 3}; (iv) {x R: -1 < x< 3);
(v) {x R: x > 3}; (vi) {x e R: x < -1 or x 3};
(vii) {x R: 3 < x < 4}.
21 5. (i) {x R: 0 < x < 1}; (ii) 0 ; {x e R: 1 < x < 2};
(iv) {x R: -2< x< -1}.
6. (i) {x R: x < -5); (ii) {xe R: x > 2}; (iii) {x e R: x < 7};
(iv) {x e R: x > 0}; (v) {x e R: x < 0 or x > 1).
7. (i) {x E R: x < -3 or x> 1}; (ii) {x e R: x< -3 or x >
{x E R: < x < 3); (iv) 0 ; (v) {x e R: -1 < x < 5}.
344
ANSWERS
PAGE
21 8. {x e R: -3 < x < -1 or x > 2}.
9. (i) 1; (ii) 3; (iii) 4. 10. (i) Negative; (ii) positive.
11. x < -4; -2 < x < 1; x > 3.
12. 10. 13. 12. 14. 1. 15. 600.
16. 20 m. 17. 3 m2; AB = 1 m, AD = I m.
Miscellaneous Exercise 1
1. 1111000; Seven: -fp, 1p, 2p, 4p, 8p, 16p, 32p.
22 2. a = 2, b = 3. 6. a = 1, b = 4, c = 3.
8. 1-1/J7. 9. (i) 10; (ii) 1. 10. Unaltered.
11. (Expressions not unique) (i) A n C; (ii) A' n B'; (iii) C' n B;
(iv) B n (A' U C').
23 12. k = -7, a = 8, b = k = 3, a = - 2, b = -2.
13. (i) Expressions equal; (ii) strict inequality.
14. 1/162, 4/27.
CHAPTER 2
28 Ex. 5. (i) 8 units N; (ii) 3 units W; (iii) 4V2 units NW;
(iv) 2V17 units N 14° 2' E; (v) 10 units N 36° 52' W.
Ex. 6. x = (2a + b)/7, y = (a- 3b)/7.
Ex. 7. (i) V7 units N 19° 6' E; (ii) V3 units S 30° E;
(iii) 1 unit S 60° E; (iv) V7 units S 40° 54 W.
Ex. 8. (i) 2J2 units at 45° with downward vertical in plane through SN
line;
(ii) 3V2 units at 45° with downward vertical in plane through EW line;
(iii) 6V3 units at 45° to N, E and upward vertical;
(iv) 6V3 units at 45°to N, W and downward vertical.
Ex. 9. (i) 2a + b; (ii) 2a-b; (iii) a + b; (iv) a- b.
30 Ex. 12. x= b + y = b - la, a = x - y, b = -f(x + y).
30 Ex. 13. Yes, no.
Exercise 2(a)
1. (i) 3a+ b; (ii) 8a + b; (iii) - 5a + 5b, (iv) - 2a+ 6b.
2. a = (u + 3v)/7, b = (2u - 0/7.
345
ANSWERS
PAGE
30 3. (i) 3a + 2b; (ii) -a+ 2c; (iii) 2a + 5b + 5c; (iv) 6a + 2b - 3c.
4. a = ;(u+ v), b = ( - u- 4v + 3w)/9, c = (- 7u- v+ 3w)/9.
5. (i) 2 units E; (ii) 3 units S; (iii) 5 units N 36° 52' E;
(iv) /2 units SE; (v) 5 units N 36° 52' W.
31 6. (i) 13 units S 67° 23' E; (ii) 13 units N 67° 23' E;
(iii) N/29 units N 21° 48' W.
7. (i) 1.53 units N 22° 30' E; (ii) 2.8 units S 59° 38' E;
(iii) 1.47 units S 16° 20' E; (iv) 7.4 units N 67° 30' W.
8. -20x -10,,l2y.
9. A = 2 sec 70°, u = -2 tan 70°; A' = 1.02.
10. 9 units: (i) 12° 45'; (ii) 125° 28'.
11. OP: (i) a+b; (ii) 2a+c; (iii) 2b-c.
OQ: (i) 2a+ 3b; (ii) 5a + 3c; (iii) 5b - 2c.
OR: (i) 3a + 4b; (ii) 7a+ 4c; (iii) 7b - 3c.
PQ = a+2b; RQ = -2b+c.
12. 3u + + u + v, w = iu -
13. b-a, -a, -b, a-b.
14. a-2b, 2a+2b+2c, a+2b+ 2c, b-2c, -a-b-4c.
32 15. -a+2b+4c, a+b+4c, a+2b+2c;
(-3ED'+2EG+2EF)/7, (ED'-3EG+4EF)/7, (ED'+4EG-3EF)/14.
16. (i) j+2k; (ii) i+j+ 2k; (iii) i - j 2k; (iv) 2i+j+2k;
-
18. (i) (4u - 2v)/3; (ii) (4v - 2u)/3; (iii) 2(v - u).
19. (i) (ii) w); (iii) 1(u+ w), (iv) 4(v-u).
21. AC'-2AD', 2AC-AC'.
22. bV2 - a, b-a/2, -a, - b, a- b.,/2, a,/2- b.
35 Ex. 16. a+fi = 1.
Exercise 2(b)
40 1. 2b-a,4(2c+a),-4(-a+ 3b+c). 5. A(b+c), A(2c + a).
6. 4b, (a + b), 2a, OA = AX.
41 7. la+ic, -a+11); x = 1(2a+4b+c).
346
ANSWERS
PAGE
Miscellaneous Exercise 2
1. b+2 cos 72° a, a+2 cos 72° b.
3. (i) Parallelogram; (ii) equal diagonals; (iii) rectangle.
43 14. 3:1:2.
15. W2 -001(7182 — 72 al), ( 6.71 CtY2)/(71 82 72 81) ; c, not parallel.
CHAPTER 3
Exercise 3(a)
0 5 4 5 0
51 1. (i) (2); (ii) ( 4 (iii) (— 3); (iv) ( 0); (v) (— 3).
2 —8 —1 \-1
/ 8
3. (i) (-1
—+); (ii) (2).
2
4 0 2 2
4. (i) x = ( ;7°) , y = (— 171); (ii) x = ( 4), y = (0), z= —64).
_ .§7 -5 1
12 PPM 347
ANSWERS
PAGE
51
5. CO' (-11)' (-31)' CO'
Exercise 3(b)
1. (i) 2x— y — 2 = 0; (ii) 2x + y + 3 = 0; (iii) x + 2y-1 = 0;
(iv) ax + by — (a2+ b2) = 0.
2. (i) 3x — y — 5 = 0; (ii) 7 x — y+ 11 = 0; (iii) y +1 = 0;
(iv) 3bx + 2ay — 5ab = 0.
3. (i) i; (ii) —1; (iii) 5; (iv) (b+ a)/(b — a).
4. (i) 2x — y — 7 = 0, x+ 2y-1 = 0;
(ii) 5x+ 4y— 11 = 0, 4x— 5y— 17 = 0; (iii) x — 3 = 0, y+ 1 = 0. 2V5.
5. (i) (3, — 2); (ii) (— (a3 + b3)I(a2+ b2), — ab(a — NI(a2+ b2)).
6. (i) 43x + 3y = 0; (ii) 7 x — y — 2 = 0; (iii) 10x— 6y — 9 = 0.
58 7. ( — 3, 10). 8. (5, — 9), (3, — 1), (1, 3). 9. (2, — 3).
10. ( 5, 4), (— 5 ± 2V21, 0), (0, 4 ± 5V3).
—
PAGE
58 12. 3:4. 13. (3, —4). 14. (2, 1). 15. Lies on AB.
16. (i) 2x + 3y + 6 = 0; (ii) 2x+ 3y— 6 = 0; (iii) 2x-3y-6 = 0;
(iv) 2x-3y-12 = 0.
62 Ex. 27. (i) z = 0; (ii) z = 2.
63 Ex. 31. 3y- 5z = 0. Ex. 32. (2A + 3, — A 2,
- — 3A 5), (1,
- — 1, —2).
Ex. 33. 4(x+ 3) = y-1 = — 4(z 1). - -
Exercise 3(c)
1. (i) [1, 1, 1], [1/J3, 1A13, 1/V3]. The direction ratios only are given in
the remaining seven parts; (ii) [1, —2, 1]; (iii) [-2, 2, 1]; (iv) [1, 0, —1];
(v) [— 1, 1, 1]; (vi) [1, 2, 3], (vii) [0, a, b+ c]; (viii) [1, 0, —1].
2. (i) x+y+z = 1; (ii) x+y-2z = 5; (iii) 2x+y —z = 4;
(iv) 3x — y — 2z = 4; (v) 5x + 3y 2z = 2; (vi) 5x z = 2.
— —
64 6. 3y + 2z— 7 = 0; 7x 4y 5z = 0.
- -
7. 2x+y+ 2z+ 7 = 0.
8. (3A+8, A+1, 1), (5, 01). 9. (1, 1, 0).
10. x 2 = 4(y— 1) = — (z 1), (3, 3, 0).
- —
Miscellaneous Exercise 3
2. 2:1, 51. 4. mx — y +(b— am) = 0.
66 8. x = —y =
10. (i) Planes form a prism;
(ii) planes all pass through the line x — z + 1 = 0, y = 4.
11. 2:1; 2:1. 12. 1844+2A-5F = 0.
67 14. a(fl+ y)/(a a).— 15. x± z.N/3— 2 = 0.
12-2 349
ANSWERS
CHAPTER 4
PAGE
68 Ex. 1. 3, 6, 6. Ex. 2. 8, —1, 5.
69 Ex. 4. 1, —1, 0.
Exercise 4(a)
1. 10, 2, 318, 82. 2. 1, —9, 26, —24. 3. 0.
4. 3, 4, —5. 5. 7, 20, 22, 5. 6. 4, 0, —6, 2. 7. 3, —7, —6, 0;
3, 20, 33, 0; —1*, —1,2. 8. 1, k, —2k2, — k3.
70 11. —9, 27, 3.
Exercise 4 (b)
72 1. (i) 62; (ii) 8; (iii) —2; (iv) —241; (v) 1i; (vi)
2. (x — 2) (x — 3) (x + 1).
73 3. (i) (2x+ 1) (x + 4) (x-1); (ii) (x+ 1) (3x-4) (4x+ 3);
(iii) (x + 5) (x — 2) (2x+ 1); (iv) (x+ 3) (x2 -2); (v) (x-2) (x2 + x+1);
(vi) (x+1) (x-3) (x — 5); (vii) (x— 2) (x2 + 4x + 1);
(viii) (2x— 1) (2x+ 1) (2x+ 3).
4. As for Qu. 3 except (iv) (x + 3) (x — (x + V2);
(vii) (x —2)(x + 2 — V3) (x+ 2+ 0).
5. 2. 6. —5.
7. —61. 8. 4, —5. 9. 2, —2.
11. (i) (x —1)(x-2) (x2 + 6x+ 1);
(ii) (x-1) (x— 2) (x+ 3+ 2,/2) (x+ 3 — 2.„/2).
12. 0, 4; 2(x+ 1) (x-1) (x —2).
Exercise 4(c)
75 6. x3 x2 +3x+2.
— 7. 2x3-5x2+x+1.
8. x3 x 5.
— —
Miscellaneous Exercise 4
76 2. R = —2, S = — 3.
3. (i) 18x + 6; (ii) 1261x + 2777. 4. P(b)/(b — a), P(a)I(a— b).
5. (b + c) (c+ a) (a + b). (i) 3(p +29 — 3r) (— 3p + q + 2r) (2p — 39 + r);
(ii) 2, 3, 13, 47, 73.
350
ANSWERS
PAGE
CHAPTER 5
78 Ex. 1. mn.
80 Ex. 3. 2, 3, -12, -13; 0, (3 ± V29)/2.
Ex. 4. 2, 3, -12, -13; 0.
Exercise 5(a)
81 1. -1, -1, 5, 23, -1. 2. 11, 21, 3i.
3. {yeR:y ?..--0},{yeR:y -. 5 0}.
82 4. -1, 1. 6. 24, {zeR: 2< z< 5}.
9. (i) 0, 0; (ii) {y e R: -1< y < 2}; (iii) {(3 - A/5)/2, (3 +8/5)/2};
(iv) 0.
10. {y e R: - 4V2 < y ---. 412}. 11. f-1(Y) = 1+.0.
12. 1. 14. (i) No; (ii) no. 15. Reflection in x = y.
Exercise 5 (b)
86 1. x < 0 or x > 1. 2. x < 0 or x > 1.
3. x < 0 or x > 1. 4. x < -2 or x> - 1.
5. 1 < x < 3. 6. x < i or x > 1.
7. x < -1 or 1 < x < 2. 8. x > 3 or x = 2.
9. -1 < x < 1. 10. x < 1 or 2 < x < 3 or x > 4.
11. -2 < x < -1 or 1 < x < 11.
12. -3 < x < -2 or -1 < x < 1.
13. x < -2 or 0 < x < 1 or x > 2.
14. x < -5 or -3 < x < 0 or 1 < x < 2 or x > 5.
15. x < - 1(,/13 + 1) or x > -1-,(V13 - 1).
16. - 8/3 < x < -V2 or V2 < x < 813.
17. x < 11 or x ,.. 3. 18. x > 2f.
19. x < -1 or x > 11. 20. x < -8/2 or x > V2.
21. x < 0 or 1 i < x < 2. 22. x < -11 or 31- < x < 5.
351
ANSWERS
CHAPTER 6
PAGE
IT IT n 577 377 577 1177
87 Ex. 1. - - -
6'3'4' 12' 4'4' 6 •
Ex. 2. 45°, 150°, 120°, 40°, 105°, 36°, 48°.
Exercise 6(a)
777 577 777 377 777 117r 577- 197r
88 1. - -
Exercise 6(b)
94 1. -1, -1, 2, -V3/3, -V2/2, 2, 0, V3/3, -V2, -2V3/3.
2. A/2/2, -1, -2, -V2/2, V3/3, -2, V3/2, -2V3/3, -V3/3, -V3/2.
3. -0.5736, -0.6157, -0.8391, 3.0716, 0.3420.
A (
.1777 117r t..‘37r 777 27r 477. 7T 777
(iii) T, (iv) 6 , 6;
`11 6 ' 6 ; °I 4 4
77 27r 477 577 137r 177 . 37 77
(v) s , 3 , T ,3; (vi) , j, ,(vii) 7, 4 ,
2 , 2'
(vu)
.. 777 7T 5n 117r
- — - , , —-
12' 12 12 12
7T 37r 577 7T 77 377 777
(viii) no solution; (ix) 2; (xxj\)
-
.717
- - 8 ' -8' 4 ' 8, i•
352
ANSWERS
PAGE
Exercise 6(c)
99 2. 0, in, 7T, Ir 2n.., 3. 35-2. 4. 65°.
100 5. 35.3, 144.7. 6. 0.4, 3.32, 5.14. 7. 0.90, 2.25.
15. (i) A/3/2; (ii) V3/2; (iii) A/3/3; (iv) - A/2/2; (v) 1; (vi) 3;
(vii) 4; (viii) - A/3/3.
16. (i) x/A/(1 - x2); (ii) x/A/(1+ x2). 17. V2/2.
18. x, Ix' 1. 19. x2+ y2 = 1.
Miscellaneous Exercise 6
1. 0.967.
101 2. 2a2(ka-V3/4). 3. 11, -1. 4. 0.73, 2.41, 3.99, 5.43.
5. 54.4 cm. (i) 12.25 cm; (ii) 75.5'; (iii) 13.2 cm.
6. 0. 7. sin x # ± sin y
8. ar1 ;7570 or ( g, AlT).
5 T, 2 9. (x2 - y2)2 = 16xy.
102 11. 0.88n, 0.78n, 0.69n, 0.60n, 0.50n.
12. 0.46n, 1.63n, 2.31n, 3.95n, 4n.
13. 27777 + n7T + cc,
14. (i) i(3n+ 1) n.; (ii) 3(3n± 1) 77 or 2kn. 16. 0.39n, 2.3 cm.
353
ANSWERS
CHAPTER 7
PAGE
Exercise 7(a)
1. 120. 2. 870.
107 3. 665280. 4. 216. 5. 120, 48, 30. 6. 495, 135.
7. 5005, 420420. 8. 34650. 9. 120, 85.
10. (i) 48!/(9!39!); (ii) 48! 16/(9! 39!). 12. in!, 840.
13. 30, 12. 14. 19958400. 15. ln(n -3). 16. 90.
108 17. n!/(p! q! r!), n!/{3(n-2p)! (p!)2}. 18. ln(n- 1).
109 Ex. 10. {HH, HT, TH, TT}, {0H, 1H, 2H}, {coins fall the same, coins fall
differently}.
112 Ex. 20. Ex. 21. 112-. Ex. 22. A.
Ex. 23. (i) 132(391)2/(27! 52!); (ii) 13 (39!)2/(27! 51!).
Ex. 24. (i) (ii) (iii) 4.
Exercise 7(b)
1. 64, 20, 6•
113 2 ,614, 64 , 39 1.1 uy, 1
664, 614,
354
ANSWERS
Exercise 7(c)
PAGE
Exercise 7(d)
125 1. 36047. 2. 0.1997. 3. A• 4. +, gr. 5.
126 6. 1459 4 ; 7• A-. 8. 41, 0, A•
9. pl(p + q - pq).
Miscellaneous Exercise 7
1. (i) P; (ii) (iii) 0.065; (iv) 0.036.
,., 1 1 1
n!' 0' 2(n 2)!' 3(n -3)!•
-
14. (i) 0.09; (ii) 0.111; (iii) 0.336; (iv) small; (v) no.
16. (pi+ (P2+ A) (p3+ A) = A2.
355
ANSWERS
PAGE
128 19. A, (1 -q)(1-q4); (1 -9) 94; -q) q; -9) R2;
4A,(1 93;
q4.
CHAPTER 8
Ex. 13. (i) 46; (ii) 35; (iii) 86; (iv) 40; (v) 100.
134 Ex. 14. (i) E (3r- 2) (3r + 1) (3r+ 4); (ii) E 6r2(4r+ 1); (iii) E (202r-2;
r=1 r=1 r=1
Exercise 8(a)
135 1. (i) 112; (ii) (iii) 15.
356
ANSWERS
PAGE
135 3. (i) 1, 4, 13, 40; (ii) 1, 1, 1, 1; (iii) 2, 22, 24, 28; (iv) 1, 2, 1a, 114.
4. s 2s + 5,2= d.
5. (i) 9r -6; (ii) 1(2r-1); (iii) 20-8r.
6. (i) 51; (ii) 70; (iii) 30. 7. (i) 2601; (ii) 17045; (iii) -1155.
8. 6- 5r, - 930. 9. 8r +1, n(4n + 5). 10. 1(5 r + 3), 4n(5n + 11).
136 11. 40. 12. 36. 14. 2, 4, 6. 17. 111, 20+.
18. 63.
Ex. 15. (i), (ii), (iv) and (vi). Ex. 16. (i) 8; (ii) 7; (iii) 32.
Exercise 8(b)
138 1. (i) ;(412-1); (ii) -1(220 -1); (iii) (2+ ,/2) (28 -1);
(iv) 1[3 -3-48].
2. 1i, 2.
3. 18, 2i, 41, ... or -18, 2 , - 4}, .... 4. 18.
5. 21. 6. 29. 7. 27. 8. 16.
9. 1.126x 108km. 10. 1, 1, 3.
139 11. £316.1. 12. £313.07. 13. (i) 19; (ii) 6; (iii) 3.
14. NI . 15. 23/ years. 16. {pa(pn -1)- an(p -1)} (p -1)-2.
17. ur = 18. 4(1- r) (1- en) (1+ r)-1.
140 Ex. 21. in(n + 1) (n+ 2) (n+ 3) (n + 4) (n+ 5).
141 Ex. 22. 1{(n + 1) (n + 2) (n+ 3) (n+
Exercise 8(c)
142 1. (i) 650; (ii) 6084; (iii) 2865; (iv) -}n(4n2 - 1).
2. A-n(n + 1) (n+ 2) (3n + 13), -&n(n + 1) (n+ 2) (3n +17).
3. &n(n+ 1)(n+ 2) (3n +1).
-
PAGE
143 11. +, -+; in(n +1)2(n+2).
-
Exercise 8(d)
146 1. (i) x6 - 6x2y+15x4y2 - 20x3y3 +15x2y4 - 6xy6 + y6;
(ii) 32x6 +40x4y+20x3y2 + 5x2y3+ ixY4 +
(iii) 1 + 14x + 84x2 + 280x3 + 560x4 + 672x6 + 448x6 + 128x';
(iv) 64(x4 +18x6y+135x4y2 +540x3y3 +1215x2y4 +1458xy2 +729y4).
2. (i) 54; (ii) 672; (iii) 264; (iv) 19440.
3. (i) 1 + 16x+ 112x2 ; (ii) 1 + 121x + 1x2 ; (iii) 256(4 - 20x + 45x2);
(iv) - y7 + 21y4x- 189y4x2.
8
4. ( ) 2nx2r-8,1120.
r
147 5. 27. 6. (i) 1.0615; (ii) 0.98411; (iii) 0.99501; (iv) 235.01.
7. 322. 8. 120, 4200.
9. (i) 1 + 4x + 10x2 ; (ii) 1- 6x+ 21x2; (iii) 1 +10x+ 35x2;
(iv) 16(4-12x- 9x2).
10. ±1-1%. 11. 2n. (i) 81; (ii) 1.
Miscellaneous Exercise 8
1. (i) 1-(3n - 1); (ii) 3jn(n-1) 2. A, 19.99998, 9.
148 3. e. 4. {nxn+i (n+ 1) xn +1)(x 1)-2,
- -
5. in + 1\
6. 4, 12, 36; 2.3n-1.
kp + 11
149 12. {(1-bn) nYn(1 - b)} (1- b)-2.
—
PAGE
Revision Exercise A
151 1. 5, 3; x - 1. 2. (-5, - 5). 4. (i) 1110; (ii) 17.3.
5. 120, 24. 6. 40.9°. 7. 7 x +7y 29 = 0.
-
CHAPTER 9
Exercise 9(a)
160 4. 13. 5. 0 or 4. 6. 13.
10. No solution if a = - 1 unless b = 0.
15. No: e.g. right-angle at B and M coincident with B.
359
ANSWERS
CHAPTER 10
PAGE
171 Ex. 2. 1000- X, - X.
173 Ex. 3. 14. Ex. 4. Lose. Ex. 5. Yes, if x = 1.
Exercise 10(a)
174 1. (i) 2+; (ii) - -,÷); (iii) 311; (iv) 51. 2. (i) 41; (ii)- 1+;
(iii) 6A; (iv) 10.
3. 3p. 4. 13661. 5. £100. 6. 0, 3.3.
175 8. 9p. 9. 7p. 12. E1.32. 13. 2.
14. (i) 8.75; (ii) 5.25.
3I 8 13 15 16 17
176 15. 27-Ip. 16. 22Ip. 17. X
16p 2 3 3 2 3 3 •
178 Ex. 6. 2, 11. Ex. 7. 1. Ex. 8. (i) #+c; (ii) 0.2.
181 Ex. 9. 4, 0. Ex. 10. g(X)3-3/2cr2-#3.
183 Ex. 11. 1(1+ 03, 11-, 1.
185 Ex. 12. 2p.
Exercise 10 (b)
1. (i) (ii) ft+ 1, a; (iii) 3#- 1, 30.; (iv) 0, 1.
2. (i) 2.81, 4.41; (ii) 3.7, 2.01; (iii) 0.15, 3.53; (iv) 150, 3000.
3. (i) 12.31, 1211; (ii) 15.7, 102.8; (iii) 3.55, 16.75;
(iv) 2.55 x 104, 2.83 x 108.
4. 41, 2+. 5. (n+ 1), -i1
2-(n2- 1).
186 6. 7. 9. (t5 + ts + t + 3)n/6n, in.
10. 15
8 7, 2.415, 1.71.
, 11. il, 1166890, 11659•
12. 6.5, 2.36. (i) (ii) 0.
Miscellaneous Exercise 10
1. £7.33. 2. 4p favours player, 4-Ip favours banker.
187 3. Just over £3. 4. -1*. 5. Yes, if X, Yindependent.
7. 2(n - r)l{n(n-1)}. 10. 0p-1. 11. 3(cr2 +#2 -#).
(\ 113\ + 18\ 113\ .4_ (7\ 113\
188 14. 3-141113\ +
lk 4 / kl/ k 5 J k2/ k 6 ) \3/ k 7
6) (183) + (193)).
+ (4
16. 1, 71.
360
ANSWERS
CHAPTER 11
PAGE
Exercise 11(a)
193 3. Sphere, centre A, radius lel.
195 Ex. 4. -5. Ex. 6. (3i + j + k)/411.
199 Ex. 9. 1, 1; lies on angle bisector.
Ex. 11. 3x+ 2y - 10 = 0, (2, 2). Ex. 12. ( a,
- - b), lc'.
Ex. 13. 8x+ 6y + 13 = 0, 8x+ 6y - 7 = 0.
Exercise 11(b)
1. (i) 2, 1 ; (ii) 2, 1;
1; (iii) 3, I; (iv) b, (3a- 20/5.
200 2. (i) 245/5; (ii) 465/65; (iii) 807/85; (iv) (a2 62)1(a2+ b2)1. -
3. (i) 417/17; (ii) 12413/13; (iii) 6V10/5; (iv) (h+ k)2(h2+ k2)-i.
4. (i) 245/5; (ii) 945/10; (iii) 2426/13; (iv) (a + b) (a2+ b2)-I.
5. (i) 3x - 4y + 16 = 0, 3x - 4y - 14 = 0;
(ii) 8x+ 6y + 35 = 0, 8x+ 6y - 25 = 0; (iii) 5x + 12y + 41 = 0,
5x+12y-37 = 0; (iv) 3x+ 2y+ 1 ± 3,/13 = 0.
7. (i) 2x - y = 0; (ii) 7 x 7 y 3 = 0; (iii) 128x+ 16y + 69 = 0.
- -
8. x 7y-18 = 0.
- 9. (i) (9, -9); (ii) 521; (iii) +1.
10. ( p, -q).
-
12. 37.
202 Ex. 14. (i) - 2j+ 2k); (ii) (31 j + k)N11 ; (iii) (3i+ 2j+ k)/,/14.
-
Ex. 15. (i) x- 3y+ z = -5; (ii) x - y+ z = 2; (iii) 2x+ 3y- 5z = - 16.
Exercise 11(c)
205 1. 110.9°. 2. (i) x y+z = 0; (ii) 2x+y-3z = 13;
-
PAGE
206 4. (i) 34/11/11; (ii) V15/15; (iii) V66/11; (iv) 174/58/174;
(v) ab(a2+ b2)'; (vi) lab bc (a2+62+ c2)-1.
- -
207 7. (i) (2, 3. -1); (ii) (4, 7, -5); (iii) (1, 1, 1); (iv) (0, -1, 3).
8. (i) (9, 5, 3); (ii) (7, -10, -3); (iii) (-6, 16, 6); (iv) (-9, 9, 12);
(v) (15, 13, 2); (vi) ( 2ad(a2+10 + c2)-1, -26d(a2+62+ c2)-',
- - -
Miscellaneous Exercise 11
1. 1p x, cos a - yisin al.
-
PAGE
212 18. x2 +y2= 1 ± 2y cot 0.
19. (i) x- 5y + 4z = 0; (ii) x+y+z = 0, (6, -2, -4).
20. -11, (1, -2, 3).
21. B(-1-, V312,0), C(-1, - 012, 0), D(0, 0, V2), (0, 0, V2/4); -4:0:V2,
2:±2V3:V2, 0:0:1.
22. 2.
213 23. 5x+12y- 60 = 0. 24. 2V6.
25. x = [p(62+ c2 - a2)-2a(bq + cr + d)] (a2+ 62+c2)-1, etc.;
- (-i, -1, D.
26. 15x-9y-3z = 0, 1:2:- 1.
27. E[(a2+ 62+ c2) l 2a(a1+ bm+ cn) x] = 0.
-
CHAPTER 12
215 Ex. 1. Yes, e.g. A = B = 60°.
216 Ex. 2. 4(V6- V2), 4(,/6+V2). Ex. 3. - i(V6+ V2), i(V6+ V2).
EX. 4. (1) a g; (111) -14, H.
Ex. 5. -2 sin A sin B.
217 Ex. 6. (a) 0 0 * (2k + 1) ;
- (b) 0 *inn ±
219 Ex. 7. (i) ± V10; (ii) ± V5; (iii) 10, 0; (iv) 1 ± V2; (v) 1(2 ± V2).
Ex. 8. (i) 0, 270, 360; (ii) 53.1; (iii) 90, 323.1; (iv) 4.9, 220.5.
Ex. 9. 1, 7. Ex. 10. 1.
220 Ex. 11. -4, 9V2. Ex. 12. 14, 4. Ex. 13. in, in.
Ex. 14. V2-1. Ex. 15. 24(1 - x2), 2y2 -1. Ex. 17. arcsin
221 Ex. 18. (1+ 2t- t2) (1 + t2)-1.
223 Ex. 19. (i) 2 sin 2x cos x; (ii) 2 cos 3x cos x; (iii) - 2 sin 2x sin x;
(iv) V2 sin (4n+ x); (v) V2 sin x.
224 Ex. 20. 0, fir, in-, 7T, '47T, in-, 277. Ex. 21. 0, in, 77, 177, 27T.
363
ANSWERS
Exercise 12(a)
PAGE
224 1. (i) a(/6+ V2); (ii) - f(V6- V2); (iii) - (2 + V3); (iv) 2- V3.
2. (i) 71(,/5 1); (ii) *A/(10+ 2V5); (iii) 1(0+1);
-
Exercise 12(b)
232 1. 76 m. 3. (i) 84; (ii) (iii)
233 4. 1012°, 44.8°. 5. 14.8°. 6. 9.3 m. 7. 36° 8. 127.8°.
9. (i) h(h2 + 2a2)-4; (ii) h(h2+a2)-1;
(iii) 2ah(112 +a9-1; (iv) Al(112 2a2) N(2172+2a2).
10. (i) 35.3°; (ii) 64°. 11. (i) 77 m; (ii) S 46° W. 12. 121 m.
234 13. 529 cm2. 14. (i) 64.9°; (ii) 69.4°; (iii) 31.6°.
16. 1.12r. 17. 1.96, 56.8°.
235 19. arccos {(0-a4-b4 +2a2b2)1(2b2c2+ 2c2a2+ 2a2b2 a4 - -
Miscellaneous Exercise 12
2. -15-; - a. 4. 2mr < x < (2n+1)7r, n e Z.
236 8. art, r!-1- 10, (1±, 1Th), ag, 1410. 13. 0, fir, 7T.
14. -1-(1 ± J2). 15. tan la = alb.
237 17. c > 4.
(i) 8 = 2nrr ± arccos 5, 0 = 2m7r+arcsin 1;
(ii) 8 = 2nrr ± arccos = (2m+ 1) arcsin -
364
ANSWERS
CHAPTER 13
PAGE
245 Ex. 9. S(P) = (3, 5), T(P) = (3, 0), U(P) = (6, 5).
Ex. 12. Reflection in Oxy plane.
Exercise 13(a)
1. (-5, —10), (t, —
246 2. (-5, —4), (3, —5); (0, 0). 3. The line x — 2y = 0.
6. Reflect in x = y and then project on to x axis.
7. Reflect in x = y and stretch by factor 2.
8. lad — bc1; all points map into line cx— ay = 0.
(1- r rs\ . ii. (1
0 k
i) , (1
0 _ 10).
247 10.
q s )
—
(9 2 1 —12 4—8 18 —1 4
250 Ex. 13. (i) 2 1 2); (ii) (— 1 —3 9); (iii) (12 9 2);
4 5 —1 — 2 —5 —2 12 —1 2
22 16 —8 —1 51
(iv) ( 2 0 3). Ex. 14. (i) ( 7 4 2); (ii) (__ 3\
9 4 7 6 —4 0 — 5/ '
Exercise 13(b)
1. (0 (2
1 5
1 4\ . (i 4\
253
6/i)
' k k -3 5 —2/ '
1 8\ /3 13 8
(iii) ( 4
7
8 Of ' (iv)1 ‘4 0 16)
0 4 2 — 5 —2 —1 —4 4 2
2. (i) (4 1 6; (ii) 3 —3-3 ; (iii) 8 —1 6;
3 1 1 —4 7 —8 1 7 —5
—6 4 2
(iv) ( 10 —2 6) .
0 10 —8
i —104 5
3. (i) . (iii)
1 4p (ii) (12612
10) ; (-218)
8 '
4. (i) AB — 2AC; (ii) A2 — AC+ BA—BC;
(iii) A2 + AB— BA + AC— CA— (B+ C)2.
365
ANSWERS
PAGE
5 0 2
6 3
253 5. ( — 5
_ 15 1)' (1 6) •
8
5 3 0 —1 —3 3 5 2 4
6. (i) — 2 0 1); (ii) (-3 —1 1); (iii) (-2 0 0);
8 2 —5 3 0 2/ 4 0 4
/8 0 —10 20 8 0 —18 —7 30
(iv) (0 0 — (v) ( 2 2 —6); (vi) (— 6 —2 5).
5 1 — 3/ —1 —3 19 —8 —4 19
3 8 0 1 4 5 6 1 —1
7. (i) — 3 3 5); (ii) ( 5 4 5); (iii) ( 1 6 3);
0 3 1 —7 1 2 —1 3 6
0 5 2 6 —19 —9 10 —11 —19
(iv) 3 11 1 ; (v) —11 —38 —1 ; (vi) —27 —40 —1
( —5 —9 2 19 39 —2 33 43 —4
—1 6 —3 2 —1 3 6
254 8. (i) 4 1 ; (ii) 4 —1 ; (iii) 2 3 — 2.
( —1 4 —2 5 1 6 9
1 2 2 .. ( —7 —\
9. (i) ( 3 —6 0); (ii) 3 11;
—3 4 —1
—1 0
5 —14 — 2\
(iii) ( 21 6); (iv)
(-2 6 1) •
—17 — 5
0 0 a
10. (x y z), (b 0 0 ; T preserves lengths; each ± 1.
0 c 0
11. 2x+y z= 0, (-1, 7, 5).
—
Miscellaneous Exercise 13
2. Shear; x, z coordinates fixed.
3. (i) (0, 0) only; (ii) line 2x —y = 0; (iii) all points in the plane.
256 5. (0, 0) only; no.
6. (i) (0, 0) only; A= 10: line 7x 5y = 0, A= —2: line x+y = 0;
-
366
ANSWERS
PAGE
256 8. AB = BA.
9. A diagonal, provided non-zero elements of D distinct.
257 11. No.
CHAPTER 14
— 2\ I 2
262 Ex. 9. (i) 3
—7 51' \-1 —lf'
(iii)
(—
15
2); (iv) (_
Exercise 14 (a)
—1 5 —4 —5
265 7. ( —1 2 8. ( 7).
2 —6 0
3), v = (-160
5). 6
—9 —1 29
9. 6. 10. (-3 —1 14).
10 2 —35
367
ANSWERS
PAGE
267 Ex. 11. (i) 3; (ii) 5; (iii) 7.
Ex. 12. (i) 2; (ii) —3; (iii) 0;
(iv) —3; (v) abc+2fgh—aft bg2 —ch2.
273 Ex. 13. (i) 12; (ii) —9; (iii) 0, (iv) 1.
Exercise 14(b)
275 1. (i) —2; (ii) —1; (iii) —1; (iv) 22.
2. (i) —30; (ii) 0; (iii) 0; (iv) 1; (v) 24; (vi) —9; (vii) 3;
(viii) —12; (ix) 8; (x) 6.
3. (iii) Singular.
—2 —5 3 —14 —14 14
276 4. (i) ( — 2 4 —2 , —21, —2; (ii) 16 16 —16), 0, 0;
4 —3 1 —27 —27 27
30 —11 5 0 —1 2
(iii) (-10 4 —2), 21, 2; (iv) ( 1 —1 —3), I, 1.
2 —1 1 —2 3 5
—2
10 15
5. (i) (— 5 1 —2); (ii) A( 7 1 —5);
1 —2 5 1 13 —5
( 0 —6 9 0 —1 2
(iii) —4 —1 4 —5); (iv) ( -4 4 —4).
2 —5 4 —1 0 2
6. (i) 12; (ii) 1; (iii) —10. 7. (i) (b— c) (c— a) (a— b);
(ii) (b — c)(c— a) (a — b) (a + b + c); (iii) (b— c) (c— a) (a — b)(bc+ ca+ ab).
5 —10 —4 —3 33 4
8. 5— 4 0. 9. 0— 8 0).
(—3 0 —9 ( 2 — 2 —1
66 36
277 10. ( 100 —173 —95 .
—39
147 249 136
Miscellaneous Exercise 14
1. ( 2. (a) Yes; (b) yes; (c) yes.
1 10).
1
a2 ab ac
() a2 ab ac
3. (ba b2 bc ; ba b2 +1 bc+2 .
ca cb c2 ca cb + 2 c2+4
278 6. Yes. 9. (b — c)2(c — a)2(a — b)2.
368
ANSWERS
PAGE
279 13. A2 = 2A +I, A-1= -A+ 2I.
2 -2 1
280 17. 1 - 2 , 91, 2 1 -2,
\2
1 2 2 1 2 2
1 -7 8
-7 - 4).
8 4 1
CHAPTER 15
285 Ex. 2. (ff, -A).
286 Ex. 5. True for any x provided b = 0.
Exercise 15(a)
1. (i) (2, 5); (ii) (- 3, 5); (iii)
2. (i) If a * - b, x = b- a; if a = - b, true all x.
(ii) If at b, x = (c- b) (a - b)-1; if a = b * c, no solution;
if a = b = c, true all x.
3. (i) x = (4+ 3a) (3a- 6)-1, y = 5(2-a)-1, provided a * 2.
If a = 2, equations inconsistent.
(ii) If a * 4, x 0, y = 2; if a = 4, x = 2(2 -A), y = A, all A.
(iii) If a * ± 6, x = 2(6 +a)-1, y = (6 +a)-1
If a = 6, x = A, y = 6(1- 3A); if a = - 6, equations inconsistent.
4. (i)If a * - 3b,x= (b2+ ab + 8)(4a+ 12b)-1,y = (24 - ab- a2)(4a + 126)-1 .
If a = -3h , 1,2 *4, equations inconsistent; if b = 2, x = A, y = 3A+ 1;
if b = -2, x = A, y = 3A-1.
(ii) If ab * -3, x = 6(1 + b) (ab + 3)-1, y = 2(3-a) (ab+ 3)-1.
If a = 3, b = -1,x = A, y = 2 - A ; if ab = - 3, a * 3 equations inconsistent.
(iii) If a2 * 62, x = (a - b)-1, y = (b - a)-1. If a = b, equations inconsistent;
if a+b = 0, x = A, y = (aA-1) a-1.
5. Inconsistent unless a = 3 or -1;
if a = 3, x = - 2, y = 3; if a = -1, x = 2, y = -1.
6. Inconsistent unless a = 1, 2 or -3. If a = 1, x =lsl,y = I;
if a = 2, x = 2, y = 0; if a = -3, x = y =
287 7. (i) a * 0, b * 0, a * b, x = c(b- c) a-1(b- a)-1, y = c(c- a) b-1(b - a)-1;
(ii) a = 0, b * 0, c * 0, equations inconsistent unless
b = c, when x = A, y = 1; (iii) a = 0, b * 0, c = 0, consistent,
x = A, y = 0; (iv) a * 0, b = 0, c * 0, equations inconsistent unless
369
ANSWERS
PAGE
287 a = c, when solution is x = 1, y = A; (v) a * 0, b = 0, c = 0, con-
sistent, x = 0, y = A; (vi) a = b * 0, c * 0, inconsistent unless
a = b = c, when solution is x = A, y = 1—A; (vii) a = b = 0, c * 0,
inconsistent; (viii) a = b = c = 0, consistent, x = A, y = ,u.
8. al bs —a2bi = 0, x = Aar', y =
9. If 0 * kff +0 (k integral), x = sin (q5— a) cosec (0— 0),
y = sin (0 — a) cosec (0— 0); if 6 = +0, equations inconsistent unless
0 and q5 both differ from a by an integral multiple of 7T. If 0 = a+ 2mn,
q5 = a+ 2nn, x = A, y = 1—A; if 0 = a+ 2nvr, = a + (2n+ 1) n,
x = A, y = A-1; if 0 = a+ (2m+1)7r, q5 = cz+2mr, x = A, y = A +1;
if 0 = a+(2m+1)7r, q5 = a+ 2(n+ 1)n, x = A, y = —1—A.
292 Ex. 12. (i) Lies in plane 01:1'K'; (ii) can be expressed in terms of any
two of i', j', k'.
Ex. 14. Columns of A proportional; (i) lies on 01'; (ii) Ai'.
Exercise 15(b)
294 1. (-2, 0, 4).
(1 6 5
295 2. 1 1 2I. (i) (0, 1, —1); (ii) (2, 0, 1); (iii) (-35, —5, —7).
—1 1 —1
3. Any point of line 1-(x-1) = -}(y + 1) = +(z+ 2).
4. (i) Line A, 1(6— 7A), 1(13-11A); (ii) inconsistent: planes form
prism in direction 5i — 7j -11k.
5. b = 10, line; b * 10, inconsistent.
6. a = 2, inconsistent; a * 2, consistent only if b = a +2, giving line.
7. (2, 1, —3).
296 8. a = 5, inconsistent; a = —1, line; unique point otherwise.
9. 4, (0, A, 2A). 10. (A, A, — 2A), (1+A, A, — 2A).
11. 4-(x+ 6) = y = — A-(z— 12), parallel.
297 14. Tetrahedron 7x-6y-5z > 0, x — 2y — z < 0, 9x-8y-7z < 0,
3x-4y-3z > — 2.
Miscellaneous Exercise 15
1 6 3
1. 5 I 1), (-2, —1, 1), (4,1, /).
0 —2 —1
(
2. (i) Line or plane; (ii) inconsistent.
370
ANSWERS
PAGE
298 3. - 4).
5. A = b - a, B= c- a, C (b - c) (c- a);
x = (b2 + c2) (a - b)-' (c - a)-1, y = (c2 + a2) (a - 6)-1(b - c)-',
z = (a2 + b2)(b- c)--1 (c -
CHAPTER 16
—X•. 2. n
306 Ex n09 - n
1,, 5,01 6 5-,nO. Ex. 3. (i) 2.5; (ii) 2.
Ex. 4. F8,1.44. Ex. 5. I, 116, 7 to 1.
Exercise 16(a)
1 . 16 s i,
, , Il 2. 0.348, 0-071. 3. 0.000105, 0.00605, 0.201.
Exercise 16(b)
313 1. 0.135, 0.271, 0.271, 0.180, 0.090, 0.036, 0.017.
2. 1.45; (a) 0.06; (b) 0.33.
314 3. (i) 1; (ii) 0.846; (iii) 0.013. 4. 0.167, 0.594.
5. 0.030, 0.106, 0.185, 0.216, 0.188, 0132, 0.077; 3, 4-25.
6. 3; 60, 222. 7. 20, 0.029. 8. 0.302, 0.060.
315 9. No: Pr (X 21) 0121. 10. (A2 +.1)/(1 -e-A).
11. a(r + 1)-1, 2. 12. 0.0030, 0.0379.
13. 226.8, 211.4, 98.6, 30.6, 71, 1.5.
371
ANSWERS
Exercise 16(c)
PAGE
Miscellaneous Exercise 16
1. (i) 76; 00 4; (iii) 2. 1, 0.335; 20.
324 3. H. 4. 8. 5. Yes, if p < 1; no otherwise.
6. IN(N+ 1) (2p-1).
- 8. s(2pq)n (p2+q2).
9. FP.
325 11. 0.3.
12. 20, 60, 90, 90, 67, 40, 33. (i) £157.50;
(ii) £167.50; (iii) £160.75.
(s+s lc)(s±n k )
p"-0q'(1- p)n-s-k (1 - q)k,
(rs)
14. Q(r, s) = ( )
s
ars.
15. (i) 99.1%; (ii) 0.03; (iii) 0.1.
326 16. 0.011, 0.870, 0.516; 23.7. 17. 0.585.
18. e-F, u, e-(ii+v)prV3{115!)-1.
327 20. 02; no. 21. p(n- r +1)(r - rp)-1, rm= [p(n+1)].
372
ANSWERS
Revision Exercise B
PAGE
328 1. x 2y = 0, 4x + 2y — 15 = 0.
— 4. 3(4k+1) r.
5. 9(10+V10), 7. 7. 3. 8. No.
329 9. (2x —y + 1) (x+ 3y-2), two lines; (i) two lines; (ii) origin.
—8 11 — 10
12. 1( — 9 12 — 9). (i) (20, 20, —23); (_ 137, _5, i
32).
10 —13 11
13. 1 — 48x + 1104x2, 0.6158. 14. (i) 37x-27y = 0;
(ii) 31x+27 = 0.
15. (i) 3x + 4; (ii) 5x+7.
1 —2 — 2
330 16. a ft = (2k +1) 7. 19.3- 2 1 —2).
—2 —2 1
20. 3:1. 21. About I, 69.
23. (i) 0; (ii) 2(q r) (r p) (p q) (y z) (z
— — — — — x) (x y).—
25. -A- < p < A-. 26. (2, 5, — 3), (4, 1, 1).
27. {y e R: — 1 < y J2}.
331 28. (x+ c) (ax + b — c). (i) (1 + c) (a + b — c); (ii) (c — 1)(b— c — a);
(iii) (ab + c) (a2bc + b — c).
29. Plane x+z = 0.
30. 55132, 2048. 31. (i) 0, an., rr, 17, 2n; 00 AT1T, 1477
32. —2 < x --I or -1, x < 2. 34. Q"--1.1).
332 35. (-2, —1, —5). 37. 2, —4, —3. 38. 4.
—18 26 —19 (1 0 0 ( 1 0
39. ( — ) 0 1 , 0 1 0
23 —33 24/ 0 1 0 —1 —10 1.
—18 —19 26 —18 26 —19
(i) (— 2 — 2 3); (ii) (— 2 3 — 2).
23 24 —33 43 —62 45
40. 26° 34', 45°, 108° 26'. 41. x —2y+z = 0, 1:2:3.
333 42. (i) 4n; 9; W. 43. a = 2" 1, fl = 1.
-
PAGE
334 49. (i) -4(b-c)(c- a) (a b); -
50. x- 2y- 2z + 8 = 0.
51. 53°, 155°, 344°. (i) 158'; (ii) in = -0.005, c = 1.7.
53. 0 = 01 -02.
335 55. k = 5, line; k = -7, prism. 56. 554400.
57. x < -5, -1 < x < 1, x > 7.
59. u+v+ w = 0 and 3a+ 2b+ c = 0 for consistency;
line (A, A + i(b- a), A- -1(2a+ b)).
60..'144, 419-2-47.
336 61. 47°. 63. x-1 = -1(y+ 2) = z.
64. (i) (ii) 1(a - b)° (a + b)-'; b(b +2a) (a + 6)-2in both cases.
65. 5567. 66. Yes, provided A = 0.
337 68. A = - 2, line; A = 1, inconsistent.
69. (Al' + Am') (Al + 70. 23 min.
338 71. 3-N, 1. 74. 2a2+ 2c2+ f 2 = 2b2 2d2 e2.
374
Index
Preface page v
17 Complex numbers (1) 379
18 Polynomials and partial fractions 401
19 Complex numbers (2) 423
20 Mappings in the Argand diagram 438
21 Quadratic equations and quadratic functions 450
22 The parabola and rectangular hyperbola 461
23 Polynomial equations 483
24 Vector products and their applications 506
25 Continuous probability distributions 520
26 Numerical solution of equations 552
27 The ellipse and hyperbola 573
28 Further matrices 609
29 Further coordinate geometry 633
Revision exercise C 657
Bibliography 680
Answers 683
Index 703
iii
Preface
This book completes the course designed to cover the work required for
modern 'A' level pure mathematics syllabuses (including probability),
particularly the M.E.I. syllabus. As explained in the Preface to Volume 1,
the calculus has been excluded, since it is already adequately covered by
existing texts, but a knowledge of the subject is demanded throughout the
book, and particularly in Chapter 25, on continuous probability.
In places the subject is carried beyond the bare limits of ' A' level
requirements: most chapters contain a small proportion of work which
would probably best be left to a second reading, while Chapters 20, 24, 28,
29 contain a substantial proportion of ' S ' level work.
As in Volume 1, the book is liberally supplied with exercises for the
student. Most questions are straightforward applications of the bookwork,
though a few harder questions may be found in the Miscellaneous Exer-
cises. The questions marked Ex. occurring in the text illustrate the associ-
ated bookwork: those marked with an asterisk should be regarded as
obligatory.
The same nomenclature as in Volume 1 has been used to indicate the
source of examination questions and I am grateful to the Examination
Board of the University of London and the Oxford and Cambridge Schools
Examination Board for permission to reproduce their questions. I should
also like to record my thanks to Dr N. A. Routledge and Mr A. J. Moakes
who read the book and made many valuable suggestions and to my wife
who once again lent invaluable assistance in checking the answers.
S.L.P.
v
17. Complex numbers (1)
The extension of the number system to the set Q vastly increases the
number of equations with a solution, but it is not difficult to formulate equa-
tions in terms of the elements of Q which have no rational solution:
consider, for example, the equations
x2 — 2 = 0 or 2 sin x = 1 or 10° = 5.
Many such equations acquire solutions if we augment the set Q into
the set R of all real numbers. Although it is possible to define R in terms
of the elements of Q, we have contented ourselves with the intuitive geo-
metrical concept of R as `completing the number line' but, since real num-
bers may be approximated arbitrarily closely by rational numbers, it
seems reasonable to assume that laws (i)—(viii) above hold in R as well.
Since we have `filled in the gaps' of the number line it might seem
reasonable to suppose that our work is done and that any equation in-
volving elements of R as coefficients should have elements of R as solutions.
Such is not the case however; indeed, if we modify the three equations of
the previous paragraph only slightly to
x2 + 2 = 0 or sin x = 2 or 10° = — 5,
none of these equations has a solution in R. However, these equations have
solutions provided we extend the number systems still further. It will be
the purpose of the remainder of this chapter to introduce such an extension
in a reasonably informal manner. The resulting augmented set we shall call
the set of complex numbers and denote it by C. Since C will contain R as
a subset, it will be necessary to verify that the laws of combination (i)—(viii),
suitably redefined, hold in C also.
Up to this point, each extension of the number system has led inexorably
to a further extension. However, it has been shown (by C. F. Gauss in the
Fundamental Theorem Algebra) that no further such extensions are needed,
in the sense that any polynomial equation with its coefficients in C will
have all its roots lying in C as well. But the real significance of the set C
goes deeper than simply enabling solutions to be found for any polynomial
380
2] MANIPULATION OF COMPLEX NUMBERS
(2—j)-1 = 1 2 = 2+
Thus
2- j 2+j 4+1 5 5j
more generally,
1 b
+ = bo (a, b, not both zero).
a+ bj a bj
— a2+b2 +
Ex. 4. Show that the multiplicative inverse of (2—j) obtained above is unique;
that is, show that
(2—j)-1= x+yj x = f, y =
Ex. 8. Write down the real and imaginary parts of the following complex
numbers:
(i) 3 —2j, (ii) (3 —2j)2, (iii) (3 —2j)--1.
Ex. 9. Show that Re, Im may be interpreted as functions C R. Find
Re [Im (a + bj)] and Im [Re (a + bj)],
R being regarded as a subset of C.
Ex. 10. If z1= 2—j, z2 = 3+2j find:
(0 (ii) za , (iii) (z1+ zz)*, (iv) (ziz 2)*, (v) (1/Z1)*, (vi) (z1+ 3 z 2) *, (vii) (z + zD* .
Ex. 11. The conjugacy function!: C —> C is defined by f(z) = z*. For what subset
of C is f the identity function?
Ex. 12. Find:
(i) 13 + (ii) (iii) 11/(1 — DI, (iv) 1cos + j sin B1, (v) 11—cos 0+j sin O.
Ex. 13. The word 'modulus' has, prior to its use in this chapter, been used in the
context of real numbers only, to mean Jae. Show that the two uses coincide if
R is regarded as a subset of C.
*Ex. 14. If z1, z2are two complex numbers, and z2* 0, prove that:
(i) (zi+z2)* (ziz2)* = 44, (iii) zi z;' = 14 2,
1)* 1 z2
(iv) 12 , (v) z1+4 = 2 Re (z1),
Z2 Z8 i z2i
(vi) zl - zi = 2j Im (z,).
The results proved in Ex. 14 are of great importance and should be
committed to memory.
The concept of modulus enables us, in a sense, to order complex num-
bers, but this is not entirely analogous to the ordering of real numbers,
since different complex numbers can have the same modulus. It is not
possible to order the complex numbers in the same way as we order the
reals.
Exercise 17(a)
1. If z1 = 1 +2j, z2 = 2—j, z3= 4+ 5j, express as complex numbers in the stand-
ard form a+ bj:
(i) z,+ z2+ z3; (ii) 3z1— z2+ 2z3 ; (iii) z1i2 ; (iv) +2z3 ; (v) z2z3+z3 z1+z1z2;
(vi) z2 zs; (vii) (z1— jz2)/(z2 + 21z3) ; (viii) + 1--; (ix) (z2 + Z3)3 ;
-
z
(x) (Z1— Z2) (Z2 iZ3) ; (xi) AZ? +:1Z2) ; (xii) — Z2 Z3)/(Z2 — z3 z1.
2. If z1= cos 01+j sin 01, z2 = cos 02 +j sin 02, show that
z1z2 = cos (01+ 02)+ j sin (01+ 02),
and find z1jz2. Evaluate zi2 and (z1z2)2.
3. Solve the following equations:
(i) x2— 4x + 5 = 0; (ii) 2x2— 2x + 1 = 0; (iii) x2 — 5x+ 7 = 0;
(iv) jx2 — 2x-2j = 0.
383
COMPLEX NUMBERS (1) [17
r=0
Ex. 15. Plot the points corresponding to the complex numbers 1 +2j, 1-2j,
—j, +j. What is the geometrical relationship that exists between the points
representing z and z*?
Demonstrate geometrically the result
z z* <=> z e R.
*Ex. 16. If P represents the complex number a + bj (P is the affix of the complex
number a+ bj) show that la+ bj1 =
and thus the position vector representing the sum of two complex numbers
is the sum of the position vectors which separately represent the two numbers
(see Figure 17.1).
*Ex. 17. Describe the vector representing the complex number z1 — z2.
Ex. 18. Mark in the Argand diagram the affixes of the complex numbers 2—j,
2(2—j) and 1(2—j). Interpret multiplication of a complex number z by the
real number (i) a > 0, (ii) b < 0, in terms of an operation upon the vector OP
representing z in the Argand diagram.
*Ex. 19. Mark in the Argand diagram the affixes of the complex numbers 3 + 2j,
j(3 +2j), — (3 + 2j), — j(3 + 2j). Interpret multiplication of a complex number z
by the pure imaginary number j in terms of an operation upon the vector OP
representing z in the Argand diagram. Show that multiplication of z by the pure
imaginary number bj is represented in the Argand diagram by an enlargement of
magnitude I bl followed by (i) an anticlockwise rotation through -flr if b > 0, or
(ii) a clockwise rotation through ig if b < 0.
Ex. 20. Interpret geometrically the statement j2 = —1.
Fig. 17.3
— 4/3+j = 2 (-1+2-j)
2 — j = ,15(75 —
Ex. 21. Find the modulus and principal argument of each of the following com-
plex numbers, and write each in modulus-argument form
(i) 1 + A/3j; (ii) 2+2j; (iii) —1; (iv) j; (v) —j; (vi) — A/3 — j; (vii) A/3 — j;
(viii) 1 + j tan a; (ix) tan a+ j; (x) 1 — cos a— j sin a (use half angle formulae).
*Ex. 22. Show that arg z = — arg z* and hence find arg z-1in terms of arg z.
Ex. 22. By writing z in the modulus-argument form r(cos 0+j sin 0), discuss the
geometrical interpretations for the multiplication of the complex number z
by: (i) the real number a > 0; (ii) the real number a < 0; (iii) the pure imaginary
number bj, b > 0; (iv) the pure imaginary number bj, b < 0.
Then
wz = sr (cos 0 +j sin 0) (cos 0+j sin 0)
= sr [(cos 0 cos 0— sin 0 sin 0)+j (cos 0 sin 0 + sin 0 cos 0)]
= sr [cos (0+ 0)+ j sin (0 + 0)],
and thus the product, wz, is a complex number with modulus Mk! and
an argument arg w + arg z. In words, the modulus of a product is the
product of the moduli and an argument of the product is the sum of the
arguments. Geometrically, if the affix of z is P, multiplication by w
(i) enlarges OP by a factor I wl, and (ii) rotates OP anticlockwise through
an angle arg w.
Ex. 24. If z is any complex number and w = u+vj (u, v real and positive), plot
in the Argand diagram, the affixes of z, jz, uz, jvz, uz+ jvz. Deduce that
Iwzi = and arg (wz) = arg w+ arg z.
Is it necessary to modify your demonstration in any way if either u or v is
negative?
Fig. 17.4
Ex. 25. Form the product (1 +j) (J3 + j) and interpret your result geometrically.
What are the values of cos 75° and sin 75°?
Now suppose that w, z1and z2 are any three complex numbers. Let
P, Q, be the affixes of z1and z1+ z2respectively (Figure 17.5). Then OP
represents z1and PQ represents z2. Now enlarge triangle OPQ linearly
by a factor I wl and rotate through an angle arg w to bring it into coinci-
dence with the triangle ORS. Then OR represents wz1, RS represents wz2
and OS represents w(z1+z2). But OS = OR+RS and we have thus veri-
fied the distributive law
w(z1+ z2) = wz,+ wz2.
Fig. 17.5
w 11
44' w
*Ex. 27. Draw diagrams similar to that of Figure 17.4 to illustrate the construc-
tion from the affix of the complex number z, of the affixes of the complex num-
bers (i) z2; (ii) z3; (iii) z-1.
390
3] THE ARGAND DIAGRAM
Fig. 17.6
For the second part of the question, we know that 1z— wl gives the
distance between the affixes of z, w in the Argand diagram, and the set S
is thus represented by the circumference and interior of the circle, centre
(1+j) and radius 1.
We demonstrate two methods of deducing the final part of the question.
Method (i) (Analytical)
From the inequalities
lz/ z21 ?: lizil — lz,11, 14+ z2! < 14+14,
—
we have
lz-21 = 1(z-1 —j)—(1 —j)1 ?: Hz— 1 —j1 —11 —j1 ?; V2-1,
lz-21 = 1(z-1---i)+(-1+DI lz-1—j1 + I —1-Fil V2+1.
Method (ii) (Geometrical)
Let C be the centre of the circle 1z— 1 —j1 = 1 (which touches the real
axis at A, say) and let B be the affix of the number 2. Let BC cut the circle
at P, Q (see Figure 17.7).
391
COMPLEX NUMBERS (1) [17
Fig. 17.7
Ex. 28. The triangle inequalities stated in the Example above may be deduced
without recourse to geometry. Show that
z1 Z2 + z1 z2 = 2 Re (z1 4)
and deduce that Izi +z21 2 = 1z11 2+ 2 Re (z1z:)+1z21 2.
Show also that Re (z14) k1I1z21
and deduce that Izi +z21 Izil +
Prove the second triangle inequality in a similar fashion.
Fig. 17.8
392
3] THE ARGAND DIAGRAM
Ex. 29. Confirm, by direct computation of (a+ bj)-1and the inverse of the matrix
ta —
that the structure of division is also preserved.
kb a )
*Ex. 30. Interpret the correspondence between the complex number cos 0+ j sin 0
(cos 0 — sin 0)
and the matrix in the light of their geometrical properties.
sin e cos 0
Ex. 31. If C, R have their usual meanings and M is the set of all matrices of the
form (b
a a
b), a, b e R, determine the nature of the function f: M R which
corresponds to the modulus function C R.
P
Fig. 17.9 Fig. 17.10
Similarly q = b + ( — 1 — 2j)
= 1+2j-1-2j
=0
and = 1+2j+1+2j
= 2+4j.
To rotate AB through an angle ± 177.we seek complex numbers z such
that z I = 1 and arg z = ± 3r; thus
z = cos Pr + j sin ill
= (1 ± j1/3).
Exercise 17(b)
1. Determine the modulus and principal argument of each of the following
complex numbers, where necessary leaving the modulus in surd form and giving
the argument in degrees, correct to the nearest degree: (i) 3-4j; (ii) 2+3j;
(iii) — 1 —2j; (iv) 1+j; (v) —2+5j; (vi) —2—j; (vii) 12-5j; (viii) —9-40j.
2. Simplify the following complex numbers:
(i) (cos 0+j sin 0)3; (ii) (cos 8+j sin 0)/(cos 95—j sin 95);
(iii) (cos 0+j sin 8) (sin 8+j cos 0); (iv) (cos 87r +j sin 71-)2;
(v) (cos + j sin - 2-7r) (cos *IT +j sin .-277);
(vi) (cos -7-r + j sin *77-) (cos *7 —j sin *7);
(vii) (cos *IT +j sin-A-7r)/(cos *Tr —j sin ,. .27r);
(viii) (1 + cos 8 + j sin 0)(1— cos 95—j sin q).
3. Determine the principal argument of the complex number
(sin 0+j cos 0) (cos 95+j sin 0)
(i) when —17-r < — 6 s ; (ii) when frr < — 0 s
4. If P is the affix of the complex number z, show how to construct geometrically
the affixes of: (i) 1+z; (ii) 1-2z; (iii) 2j+ 3z; (iv) 1 — jz; (v) (1 — j) z; (vi) j+ z2;
(vii) z+z2; (viii) (1 +j) (1 — z).
5. If P is the affix of the complex number z, show how to construct the affixes
of: (i) z*; (ii) z—z*; (iii) 1+2/z; (iv) (1+z)* ; (v) 1/(1+z); (vi) j/(j—z);
(vii) z/(1 — z); (viii) (1— z)-2.
6. If P, Q are respectively the affixes of the complex numbers z, w show how to
construct the affixes of: (i) 2z+ w; (ii) 1(z+ 2w); (iii) 2(3z— w); (iv) z+jw;
(v) (1+j) (z+w); (vi) (z +1) (w +1); (vii) w/z; (viii) j(z— w).
7. Describe geometrically the set S = {z e C: zI — 2 —j1 < 1} and prove that, for
all z e S, V5-1 < izi 0+1.
8. Describe geometrically the set S = {z e C:lz— 21 < 2} and prove that, for
all z e S, 3 < jz+2+3j1 < 7.
9. Show that jz-6j1 < 3= 10 I z -5+6j1 < 16.
10. Show that I z —3—j1 < 1= 00-1 < 1z1 < N/10+1.
11. Prove, using mathematical induction, the following results:
(0 Izi+z2+...+znI < Iz1 j+jz21+...+ jz„I;
(ii) zni =
(iii) arg (z1z2z3...zn) = E arg zi;
i=1
(iv) (21+ z2 + Z3 + + Zn)* = ;
1=1
(v) (Zi Z2 Z3. . * = z1 Z: z3...z,,.
12. The function!: C -›- C is defined by f(z) = 2z +j. Describe the effect of this
mapping geometrically and find the value of z which is invariant under the
mapping, again interpreting your result geometrically.
396
3] THE ARGAND DIAGRAM
(z C: Im (z—ja) = 0}
z—b
is represented by a straight line in the Argand diagram.
Miscellaneous Exercises 17
1. By writing 1 = — j2, express z = (V3 +1)+ j(V3 —1) as the product of two
complex numbers and hence write down the values of 1z1 and arg z.
2. If lz — 2 — < 2 and 1w — 5 — 5j1 < 1, find the maximum and minimum values
of Iz— w(.
3. Describe geometrically the relation between the complex numbers 6j and
2 + 4j and the complex numbers (i) 1 + 5j; (ii) 3 + 3j.
If A is an arbitrary real number, what can you say about the complex number
(1 + A) + j(5 + A) ? (over)
397
COMPLEX NUMBERS (1) [17
If B, C are the affixes of the complex numbers 6j, 2+ 4j, find the two complex
numbers whose affixes, A, A' are such that the triangles ABC, A' BC are equi-
lateral.
4. z, z2,w,,w2 are four complex numbers and
,
b—c c— a a—b'
prove that the triangles BCX, CA Y, ABZ are similar.
Prove also that the triangles ABC and XYZ have the same centroid.
(0 & C)
7. The cross-ratio of four complex numbers z3, z4, written (ziz2, z3z4), is
defined by
Z3) (Z2 Z4)
(Z4 Z2, Z3 Z4)
= (Z1 .Z4) (Z2 Z2) •
15. In the Argand diagram A, B, C and D are four points representing the com-
plex numbers z1, z2, z3 and z4respectively. If 0 is the origin, prove that the
triangles OAB and OCD are similar if
Z1 Z3
Z2 Z4
400
18. Polynomials and partial fractions
Ex. 1. Explain how the method above may be modified for division by ax+b,
a * 1, by first dividing P(x) by a. Obtain Q(x) and R when
P(x) E 2x4 — 15x3 — 33x2 — x+ 14
is divided by 2x + 3. (In such cases Q(x) will not, of course, generally have integral
coefficients.)
P(x) .= A0(x a)n + Ai(x a)n-l +A2(x — a)'2" + ... + An_Ax a)+ A.
- — — —
2. DIFFERENCES
Consider the quartic polynomial
f(x) .- x4 — 5x3 + x+ 2.
Tabulated overleaf are the values of f(x) for x = —4(1)5 (that is, for all
values of x in steps of 1 from x = — 4 to x = 5). The first and second
columns give respectively the values of x and f(x), while each subsequent
column gives the successive differences between entries in the preceding
column.
It is seen from this table that, although the first and second difference
columns follow no readily discernible pattern, the third difference column
contains the terms of an arithmetic sequence and the fourth differences are
constant. This is no coincidence: for any polynomial of degree n, the nth
differences are constant (see Ex. 11).
403
POLYNOMIALS AND PARTIAL FRACTIONS [18
x f Af A2f AY AY mf
—4 574
— 359
—3 215 200
—159 — 90
—2 56 110 24
—49 —66 0
—1 7 44 24
—5 —42 0
0 2 2 24
—3 —18 0
1 —1 —16 24
—19 6 0
2 — 20 —10 24
—29 30 0
3 — 49 20 24
—9 54
4 — 58 74
65
5 7
If we are content to assume for the moment that the fourth differences
continue constant, it is an easy matter to extend the table working back-
wards from the constant fourth difference, to supply further values for
the polynomial :
x I of Lvf Ay A4f
2 —20
— 29
3 —49 20
—9
4 —58 74 24
65
5 >152 24
217
6 224 254
471
7 695
404
2] DIFFERENCES
Ex. 7. The cubic polynomial f(x) has the values f(0) = 2, f(1) = — 8, f(2) = — 20,
f(3) = —28. Form a difference table and deduce the values of f(4), f(5).
Ex. 8. Explain the following check on a difference table : the sum of any column
of differences is equal to the -difference between the first and last entries in the
preceding difference column.
Of A2f A2f
0 0
99
0.1 0.099 —6
93 —6
0.2 0.192 —12
81 —6
0.3 0273 —18
63 —6
0.4 0.336 —24
39 —6
0.5 0.375 —30
9
0.6 0.384
Ex. 9. Extend the table above, back from the constant difference in the 03f
column, to obtain f(- 0.1) and f(0.7).
Air = frd-i
and A2f, = A(4) = — Af., etc.
Similarly, if we work back from x, through values x_1, x_2, x_3, ... to obtain
2 PPMII 405
POLYNOMIALS AND PARTIAL FRACTIONS [18
f Af 6,2f A3f
x_2 f-2
AT-2
x -1 f1 AY-2
Af-t AY-2
X0 fp 2f 1
Afo L 3f-1
xl fl fo
Ofi
x2 fl
where h is the difference between successive values of x for which the poly-
nomial f(x) is being tabulated.
*Ex. 10. If f(x) = x(x -1) (x-2) ... (x- n+ 1) (and thus f(x) is a polynomial
of degree n) prove that, if f is tabulated for integral values of x,
(i) Af(x) = nx(x -1) (x- 2)...(x-n+ 2),
(ii) A2f(x) = n(n-1) x(x-1)(x-2)...(x n +3).
-
Suggest a form for Arf(x) and prove your conjecture by mathematical induction.
*Ex. 11. Prove that any polynomial f(x) of degree n may be expressed in the
form
f(x) Aix(x- 1) (x - 2) ... (x - n + 1)+ A2 x(x 1)...(x n+ 2) + + An x+ A.+3.,
- -
where A1, A2, An+iare numbers, and deduce, using Ex. 10, that the nth
differences of an nth degree polynomial are constant. Prove furthermore, that
Av.= Ain!.
*Ex. 12. Prove, by mathematical induction, the Gregory-Newton formula
Example 3. The quadratic polynomial f(x) has values f(2) = 15, f(3) = 41,
f(4) = 81. Find an explicit form for f(x).
Since f(x) is a quadratic polynomial, the second differences will be con-
406
2] DIFFERENCES
stant. Thus, we are able to build up the table of differences below to infer
the values of f(0) and Af(0).
zi2f
0 5
—2
1 3 14
12
2 15 14
26
3 41 14
40
4 81
Thus f(0) = 5, Af(0) = —2, AY(0) = 14, and by the Gregory—Newton
formula (Ex. 12)
f(n) = f(0)+ (7) Af(0)+ (3) 6.2f(0)
= 5-2n+ 7n(n--1)
= 7n2— 9n + 5.
The Gregory—Newton formula was only proved for integral n, but
f(x) = 7x2— 9x + 5
is the unique quadratic polynomial defined by the three values
f(2) = 15, f(3) = 41, f(4) = 81.
(We may thus be permitted to write formally
(x
1) (2
1
lx(x — 1), 6x(x-1) (x-2), ...
Exercise 18 (a)
1. Use the method of synthetic division to find the quotient and remainder:
(i) when x3 — 5x2 +x+16 is divided by x-2;
(ii) when x4 + 3x3 — 9x2 -23x+14 is divided by x+3;
(iii) when 3x4 + x3 -12x2 — 11x— 24 is divided by x+2;
(iv) when x4 — 5x— 5 is divided by x-4;
(v) when 7x5 -3x-5 is divided by x + 5.
2-2 407
POLYNOMIALS AND PARTIAL FRACTIONS [18
2. Use the method of synthetic division to find the quotient and remainder:
(i) when 2x3 — 5x2 — 17x is divided by 2x+ 1;
(ii) when 3x4 +2x3 -3x2 + 7x+ 6 is divided by 3x+2;
(iii) when 2x6 -9x4 +2x3 + 7 x2+7 x— 3 is divided by 2x-1;
(iv) when 4x4 — 9x3 — 9x'+x+ 2 is divided by 2x + 1;
(v) when 3x6 — x6 — 21x3+ 9x2 — 14x+ 12 is divided by 3x-1.
3. Use a hand calculating machine to evaluate the following polynomials for
the given values of x, giving your answers to 1 decimal place:
(i) 3.5x3 +7.1x2 +2.8x+5.9, x = 2.9;
(ii) 7.7x3 -9.6x2 — 3.8x +17.2, x = F97;
(iii) 8.6x4 — 9.2x3 + 7.3x2 — 5.8x+ 13.2, x = 2.61.
4. Express the following polynomials in the form
ao(x — h)"+ ai(x — h)"-i+ + a„,_1(x—h)+
for the given values of h
(i) x3+x2+2x+1, h = 1;
(ii) 2x3 — 17x2 + 30x, h = 2;
(iii) xa + 7x2 — 6x— 4, h = —3;
(iv) x4-1, h = —2.
5. Use the method of synthetic division to find the quotient and remainder:
(i) when x4 + (1 + 4j) x2 — 2x + (1 + j) is divided by (x+j);
(ii) when x4 + (1 —j) xa — jx2 + 3(1 + 2j) x+ (4-3j) is divided by (x— 2j).
6. By expressing P(x) x 3—11x2+35x— 30 in the form
ao(x 3)3 + ai(x — 3)2+ a2(x— 3) + a3,
—
3. PARTIAL FRACTIONS
The process of adding together fractions with polynomials as denomina-
tors will be familiar to the reader. For example
2 3
x -3 x —2
2(x — 2) — 3(x — 3)
(x — 3) (x -2)
5— x
(x — 3) (x — 2)•
The reverse process is called 'resolving a rational function into partial
fractions'. Such a resolution is of importance in differentiation and integra-
tion and also in obtaining power series expansions of rational functions,
as we shall see later in this chapter.
Two questions arise—can the resolution always be effected, and, if it can,
are the partial fractions obtained unique? Detailed analysis of these
questions lies outside the scope of this book.t
9 a b
Ex. 14. Express in the form +1, a, b integers, and show that such a resolu-
tion into arithmetical partial fractions is not unique.
73 a b c
Ex. 15. Express 105in the form 1,+ 3 + a, b, c integers.
is greater than, or the same as, the degree of Q(x), we may divide out until
this is no longer the case. For example, we should express
x4 — 2)0— 3x2 + x+ 4
x2 — 1
x— 2
as x2 -2x 2
x2 — 1
We now suppose that this has been done; that is, that we have a rational
function of the form
YA
*Ex. 20. The fact that (x — 3) cancelled at the final stage in the example above was
no coincidence. Suppose that f(x), F(x) are polynomials and that F(x) has a
linear factor x — a, that is,
F(x) E (x — a) g(x), g(a) = 0.
Show that, near x = a, the rational function f(x)/F(x) is approximately
f(a) 1
g(a) x — a'
f(x)g(a)— g(x)f(a)
with deviation
(x— a) g(x) g(a) •
Prove that the numerator of this fraction is a polynomial with a factor (x — a) and
deduce that cancelling is always possible.
Example 5. Express
4x2 — 3x+ 5
f(x) =
(x-1)2 (x+2)
in partial fractions.
f(x) has singularities at x = 1 and x = —2. The results proved in Ex. 20
suggest that we consider the singularity x = —2. Near x = —2, f(x) is
approximately
4( — 2)2 -3( —2)+ 5\ 1 \ = 3
(
(-2-1y ) kx+21 x+2.
The deviation is
4x2 — 3x + 5 3 _ 4x2 — 3x+ 5 — 3(x2 — 2x + 1)
(x-1)2 (x+2) x+2 (x-1)2 (x+2)
x2 + 3x+2
(x-1)2(x + 2)
(x+ 1) (x+2)
(x— 1)2(x + 2) '
3 + x+1
and thus we have f(x) =
x+ 2
The reduction can, if desired, be carried one stage further by writing
x+1 (x-1)+2:
3 1 2
f(x) = + +
x+ 2 x— 1 (x — 1)2 '
Example 6. Express
x2 + 12
f(x) —
(x-2) (x2 +4)
in partial fractions.
f(x) has just one singularity, at x = 2. Near x = 2, f(x) approximates
to
(22 +12\ 1 \_ 2
22+4 ) kx-2) x-2*
The deviation is
x2 +12 2 _ x2 + 12-2(x2 + 4)
(x-2) (x2 +4) x-2 (x-2) (x2 +4)
(x-2) (x + 2)
(x — 2) (x2 + 4)
x+2
x2 +4'
2x+2
and thus we have f(x) =
x-2 x2 +4.
414
3] PARTIAL FRACTIONS
Exercise 18(b)
1. Express in partial fractions
11-7x 4—x 9x — 22
(0 (ii) • (iii)
3— 7x+ 2x2' x2 +2x' 2x2 +5x-3'
2x-1. 2x2 +3 13x-19
(iv) (v) • (vi)
1—x2 ' x2+x (x — 1) (x +2) (x-3)'
x2 -8x+1 . 20—x3 x4
(vii) x — x3 f(viii) x2 -4 (ix)
x2 — 5x + 6'
1 + 3x . 1 + 6x2 x-2
(x) (xi) (xii) •
x2 + 2x3' (2 + x) (1 — 2x)2 ' (2+x2) (1+x)'
2x (x-1)2 x3-Fx2 +x
(xiv) • (xv)
1 + 8x3; (x-3)2(x— 2)' (x+ 1) (x2 +2)'
2x2+ 3x + 2
(xvi)
x4(x— 2) •
2. Find numbers A, B, C, D such that
2x2 + 8x + 3 A(x —2)3+ B(x — 2)2(x — 1) + C(x — 2) (x — 1) + D(x —1)
and deduce the partial fraction expansion of
2x2+ 8x + 3
(x-2)3 1) •
4. Express
1
x+1)(x+2)
in partial fractions.
Hence evaluate
1
r=1 (r +1) (r+2).
1
Evaluate similarly
r=1 (r+1) (r + 4)•
1
5. Express
1+x3
(i) as the sum of two partial fractions with real coefficients;
(ii) as the sum of three partial fractions with complex coefficients.
6. Express 1/(x2 — 3x + 2) in partial fractions and deduce the partial fraction
expansion of y2/(2y2 — 3y +1).
Express x2/(8x2 — 6x + 1) in partial fractions.
4. POLYNOMIAL APPROXIMATIONS:
THE BINOMIAL SERIES
We saw, in Chapter 8, that, for any number x and any positive integer n
If x is numerically much less than one, so that, say, its rth and higher
powers are negligible, the identity above may be used to give a polynomial
approximation of degree (r 1) for (1 + x)n. Thus, for example, with n = 10,
—
Ex. 25. Use the quadratic approximation to (1+x)1° obtained above to find
(0.998)10 correct to 4 decimal places.
1 1.3 1. 3 5
(iv) (1 + x)1/2 = 1 + ax)- (-1-x)2 + (1-x)3- ( 4 (Ix)4 --;
1.3 1.3.5 1.3.5.7
(v) (1+ X)-1I 2= 1— (-IX) + —
2!
(-1-x)2 (x)3+ (x)4
3! 4!
Ex. 27. Suppose, when x is very small, that (1+ x)I can be approximated by
1 + ax + bx2. Then 1+ x and (1 +ax+ bx2)2are nearly equal. Deduce that, for a
`best fit', 2a = 1 and a2 + 2b = 0, giving the first three terms of expansion (iii) of
Ex. 26.
Deduce similarly a quadratic approximation to (1 + x)- if x is very small.
= 3 (1 - 2x)-i - (1 - xj)
1 ti+x+ x2 ± n
j(1 + 2x + 4.x2 + 8x3) -
6 k 2 4
++Ix+8lx2 +16 x3.
t The decision as to what constitutes a 'sufficiently small' x depends, of course, upon
the accuracy demanded of the answer. In numerical approximations, answers are usually
required to a certain number of significant figures or places of decimals and the point
at which to truncate the binomial •series can be found by inspection.
It can be shown that
N (n-1) (n— 2)...(n— r+ 1)
lim xr = (1+x)'-1
N—>co r=1 r!
for any real n, provided lx1 < 1. The result is also true for x = 1, provided n > —1 and
for x = —1 provided n > 0. It is not true for any Ix! > 1, unless n is a positive integer
or zero.
417
POLYNOMIALS AND PARTIAL FRACTIONS [18
1
(ii) = (1 — x)--1(1 +x)-1
(1 — x) V(1 + x)
(1 + x + x2 + x3) (1 —ix+ ix2 --hx3)
1 + x(1 --1-)+ x2(1 -1+ + x3(1 -1+ 8-16)
1 +-ix+8x2+126x3.
1+0.1+0.015+0.0025+0.00044+0.00008.
Thus 21/5 1.11802 and V5 2.23604.
To estimate the accuracy of our answer we have to obtain an upper limit
for the sum of the remaining terms in the series. Now the coefficient, u„
of (A)' in the expansion above is given by
1.3.5 . (2r — 1)
= r!
ur 2r-1 1
and we have — 2— - < 2.
ur_1
Thus < 2u,_„
418
4] POLYNOMIAL APPROXIMATIONS
and the sum of the remaining terms in the binomial series above is certainly
less than
1.3.5.7.9.11
6! ( A)6 [1+ 126+ Ito+ dciv+...]
= 1.3.5.7.9.11
6! (m) Lr11 1
- (sum of infinite geometric series)
1.805 x 10-5.
Since 4.4 x 10-4and 8 x 10-5are overestimates of the fifth and sixth terms
of the series above we may safely conclude that the error in taking 1.11802
as 10 is less than F805 x 10-5and so our answer of 2.23604 for has an
error of less than 3.61 x 10-5.
Ex. 29. In fact, V5 = 2.236068 (6 d.p.), and our limit for the error in Example 8
appears to have been rather crude. Explain why this is so. How could a more
accurate estimate of the error be obtained?
Ex. 30. To 4 d.p., J5 = 2.2361, while in Example 8 above, we obtain V5 = 2.2360
(to 4 d.p.) although we subsequently show our error to be less than 5 x 10-5.
Explain.
7 7 j '
,z 1
7 2y+ 44y92)
(1— since y is approximately zero
= 7 491x 3) + 3 : 3 (X 3)2*
419
POLYNOMIALS AND PARTIAL FRACTIONS [18
Exercise 18(c)
1. Obtain cubic polynomial approximations for the following expressions in the
neighbourhood of x = 0:
1 1 1 1
(i) 1 (ii) • (iii) (iv) (1 _ .02; (v) — x)3.
3 —x' (1 — 2x)2;
2. Obtain cubic polynomial approximations for the following expressions in the
neighbourhood of x = 0.
1 1
(i) ; (ii) V(1 4x); (iii) V(4 x); (iv) ✓ (v) V(21
V(1-2x) (4 x)' x)•
3. Obtain cubic polynomial approximations for the following expressions, in
the neighbourhood of x = 0:
(i) (1 + x)*; (ii) (1 +2x)-4; (iii) (1— x)-1; (iv) (2 — x)-5; (v) (8— x)}.
4. Obtain cubic polynomial approximations for the following expressions, in
the neighbourhood of x = 0:
1 1 1
(i) • (iii)
(1 — x) (1 — 2x) ; (x — 2) (x — 3) (1 + x)2 (2 + x)'
x(x+ 3) 1+4x
(iv) ; (v)
(1 + x2) (1+2x) (1 — 2x)2(1 + 2x2) •
5. Obtain quadratic polynomial approximations for the following expressions
in the neighbourhood of x = 0:
1 1 1
(i) (ii) • (iii) •
(1+ x) (1 + x)' (1 + x2) V(1— 2x)' V(1 5x + 6x2)'
—
. vr 1(1— x) ' 1
(v)
1+x (1 + x)2V (1 + 2x) •
6. Obtain quadratic polynomial approximations for:
(i) (1+ x)-1in the neighbourhood of x = 1;
(ii) (2— x)--1in the neighbourhood of x = — 1;
(iii) V(3 +2x) in the neighbourhood of x = —1;
(iv) (2+3x)* in the neighbourhood of x = 2;
V(3— x)
(v) in the neighbourhood of x = —1.
7. Prove that
1 Al(1+x2)-1
V (1 + x2) + 1 x2
and deduce that, if x is very small,
1
1X2.
V(1 ± X2)± 1 8
8. Find, correct to 5 significant figures, the values of (i) (0.998)1/3; (ii) (1.02)112 ;
(iii) (4.01)1/2; (iv) (0799)0.
420
4] POLYNOMIAL APPROXIMATIONS
12. If x is so small that x3and higher powers of x may be neglected, express the
function
A/(4+x)
1— 2x+ (1 + 3x)2/3
in the form, a+ bx+ cx2. (0 & C)
Miscellaneous Exercise 18
1. The polynomial x6 + Ax6+Ex4 — Axs + Ax2+ Bx+ C is exactly :divisible by
xs +1; find A, B, C.
2. If (1 + x+ x2)n E a0+ ai x+ a2 x2+ ...+a2n X2n, write down the values of ao
and a2„ and prove that a,. = a2n-r.
2n 2n
Show that E a,. = 3" and find E (— 1)rar.
r =1
4. Prove that
x2+2bx+1 (x—a)(x+2b+a)+a2+2ab+1.
1
Put
(x-2) (x2 + 2bx + 1)
into partial fractions
(i) when b =
(ii) when b = 1. (0 & C)
5. Find A, B such that, for all values of x other than 1, — —1,
x+1 Ax+B A(x+1)+B
(2x— 1) (2x + 1) (2x + 3) (2x-1) (2x + 1) (2x + 1) (2x+ 3).
Find the sum to n terms of the series whose rth term is
r +1
(0 & C)
(2r —1) (2r +1) (2r+3)'
6. Prove that, if —1 < x
1.3 1.3.5 1.3.5.7
1— x + — x2 x3 + x4 + — (1 + 2x)-112.
1.2 1.2.3 1.2.3.4
Deduce that
1 1.3 1.3.5 1.3.5.7
1+ + + + + V2.
4 4.8 4.8.12 4.8.12.16
7. Sum to infinity the series:
1 1.3 1.3.5 1.3.5.7
(i) 1+
3 + 3.6 + 3.6.9+ 3.6.9.12 +...,
1 1.4 1.4.7 1.4.7.10
(ii)
1+ 4 + 4.8 + 4.8.12+ 4.8.12.16 +—;
3 3.9 3.9.15 3.9.15.21
(iii)
1 8 + 8.16 8.16.24+ 8.16.24.32+—;
1 1.5 1.5.9 1.5.9.13
(iv) 1+ + + + +.-.
6 6.12 6.12.18 6.12.18.24
8. The cubic polynomial P(x) assumes the values 2, 3, 2, 11 respectively for
x= 1,2,3,4. By writing
P(x) E A+ B(x-1)+ C(x-1) (x— 2)+ D(x— 1) (x — 2) (x-3) findP(x).
Show how to fit a quadratic polynomial to coincide with the values of
y = sin x at x = 0, in, n.
9. Find a linear approximation to J(3 — x) in the neighbourhood of x = —1,
and interpret your result graphically.
Show that V3.9,-1:1 1.975.
10. Find the sixth term and the rth term of the series whose first five terms are
4, 1, 0. 13, 76
on the assumption that the rth term is a polynomial in r of as low a degree
as possible.
422
19. Complex numbers (2)
1. INTEGRAL POWERS OF
COMPLEX NUMBERS
Theorem 19.1. (De Moivre's theorem for positive integral exponent.) For any
positive integer n
(cos 0+j sin 0)" = cos nO +j sin nO.
Proof.
(cos 0 + j sin 0)n-1 = cos (n — 1) + j sin (n — 1) 0
(cos +j sin OP = [cos (n — 1) + j sin (n — 1) 0] (cos 0+j sin 0)
(cos d+j sin OP = cos nO +j sin no
But
(cos 0+j sin e)1 = cos 0+j sin 0
and the result holds for all positive integral n, by induction.
Ex. 1. Illustrate de Moivre's theorem using the Argand diagram for the cases
n = 2, 3, 4. What can you say about 0 if (cos 0+j sin OP = 1 (n a positive
integer)?
Ex. 2. Express J3 +j in modulus-argument form and deduce the value of
G/3 +jr.
Theorem 19.2. (DiMoivre's theorem for negative integral exponents.) For any
negative integer n,
(cos 0+j sin 0)" = cos nO +j sin no.
423
COMPLEX NUMBERS (2) [19
Ex. 3. Illustrate de Moivre's theorem, using the Argand diagram, for the cases
n = —2, —3, —4. What can you say about (i) zn+Z-n, (ii) z6 — z-6 where
IzI = 1?
Ex. 4. Evaluate (V3 + j)-9.
The results proved in Theorems 19.1 and 19.2 are frequently useful in
deriving further results. We shall illustrate, in the next examples, some of
the techniques most commonly employed.
But
z6+z-6 = (cos 60+j sin 60) + (cos 60—j sin 60), by de Moivre's theorem
= 2 cos 60,
and similar results hold for z4 + Z-4and z2+ Z-2. Thus
32 cos6 0 = cos 60+6 cos 40+15 cos 20+10.
S=
1 2x cos 0 +x2
-
Exercise 19(a)
1. Simplify
(i) (cos -17
. T+ j sin 1704; (ii) (cos +j sin 170-3;
. - j sin 1706 ;
(iii) (cos 17 (iv) (sin 3
6 77+j cos *706;
(v) (cos + j sin *703(cos 47+ j sin 47)4 ;
(vi) (1 +j tan *704; (vii) — j Cot *70-4;
(viii) (1 + cos 20+j sin 20)-4.
2. Express in standard form
(i) (cos +j sin 17012 ; (ii) (1 + j)6 ; (iii) (.s/3 -D4(1 +jV3)6;
(iv) (cos 0+j sin 0)n/(cos ¢-j sin O)m.
3. Express in terms of sin 0: (i) cos 40; (ii) sin 50.
(iii) sin 60
4. Express in terms of cos 0: (i) cos 40; (ii) cos 50;
sin 0 •
5. Express tan 40 in terms of tan 0.
6. Prove that cos 70 = 64 cos7 0-112 cos50 +56 cos3 0-7 cos 0.
Write down the seven roots of the equation
64x7 -112x5 +56x3- 7x = 0,
and also the seven roots of the equation
64x7 - 112x5 + 56x3-7x-1 = 0.
7. Express, in terms of cosines of multiple angles:
(i) cos5 0; (ii) sin4 0; (iii) cos7 0.
8. Express, in terms of sines of multiple angles:
(i) sin5 0; (ii) sin' 0; (iii) sin30 cos 0.
9. Express sin50 cos4 0 in terms of sines of multiple angles.
2. RATIONAL POWERS OF
COMPLEX NUMBERS
In this section we shall denote a general rational number by p/q, where p, q
are integers and q> 0.
Given a real number r > 0, there is just one positive qth root of r9, that
is, there is just one a > 0 such that rP = a3. We write
a = r2vq or a =
*Ex. 12. Prove that, if p, q are integral and r > 0, there is only one positive
real number a such that ry = a3. (Show that, for b > 0, ry = b4 = b = a.)
*Ex. 13. Prove that, if q is odd there is just one real number a such that, for real
r, r5= ag and that, if q is even, there are either two or no such numbers.
Now let us consider the problem of finding a complex number w such
that zr' = w3, where z is the complex number r(cos 0+j sin 0), r > 0.
Suppose w = s(cos c6+ j sin 0); then
rv(cos p0 + j sin p0) = sq(cos qcb +j sin qq5).
(p, q being integers, we may apply de Moivre's theorem to both sides.)
Now two non-zero complex numbers can be equal only if they have the
same moduli, and arguments differing by an integral multiple of 2rr. Thus
ry sq,
p0+2krr = q95.
427
COMPLEX NUMBERS (2) [19
Whether q is even or odd, we may take s = rIvg (see Ex. 13) and
0 p+q2k7r,
Ex. 14. What are the four fourth roots of: (i) 1; (ii) 4; (iii) —1?
rIva (cos eq +j q
It is important to realise that zvia is only one of the qth roots of zr'. We
shall sometimes use the notation VZ2) for zwq; in particular, Vz = z.
*Ex. 15. Under what circumstances do (zP)liq and zvig represent the same
complex number ?
Ex. 16. Verify that the definition given above for z 2)/q yields the correct value for
where r > 0 is a real number.
rvig,
Example 4. Find (-1 + j)1and the other fifth roots of the complex number
(-1+j).
Writing z = —1+ j
= V2(cos +j sin in)
and w5 = z,
21ar in+ 21aT1
we have w= {cos in + +j sin
5 5 I•
Distinct values for w are given by k = 0, 1, 2, 3, 4; k 0 gives (-1 +Di.
Thus
(— 1 +j)i = 2116(cos +j sin -21570.
The remaining four fifth roots of (-1+j) are 2-Ncos j sin E-g),
2•116(cos 117T +j sin IP), 2i+cr(cosF j sin fin), and 2*(cos j sin in).
428
2] RATIONAL POWERS
Ex. 17. Plot the positions, in the Argand diagram, of the affixes of the complex
1 j
number - V and of its five fifth roots.
-V2 + 2
Ex. 18. Find the three cube roots of j, and verify that their sum is zero. Plot their
positions in the Argand diagram.
*Ex. 19. Show that the nth roots of unity are represented by the vertices of a
regular n-sided polygon inscribed in the unit circle izi = 1.
*Ex. 20. If p is a prime number, and if 6 is any pth root of unity other than 1,
show that the complete set of pth roots of unity may be written as 1, 6, 62, ..., 6P-1.
Discuss possible generalizations of this result for the case where p is not prime.
(If you have difficulty in proving the general result, consider the particular cases
p = 3,p = 4.)
429
COMPLEX NUMBERS (2) [19
Exercise 19(b)
1. Find in standard form the three cube roots of:
(i) 8; (ii) - 1; (iii) -j; (iv) (1 +j)3.
2. Express in the form a+ bj, giving a, b to 2 significant figures in (ii)-(iv)
(i) AA3 -4.0; V(1 +2j); (iii) AA -3-0; (iv) A/(1-3.0.
3. If co is a complex cube root of unity, show that the cube roots of z3 are
z, zo), zo)2.
Find the three cube roots of - 2 +2j and deduce surd expressions for
cos (77112) and sin (n/12).
4. Find, correct to 2 significant figures, the real and imaginary parts of (2-j)1/6
5. Simplify:
(i) V(cos 0-j sin Or V(cos 0+j sin 0);
(ii) (sin 0-j cos 0)113;
cos 0- j sin B\
(iii) ik(cos 30+j sin 301'
(iv) (1 +j cot 0)114 ;
(v) {(cos 0+j sin 0) (sin 0-j cos 0)}114.
6. Plot in an Argand diagram the four fourth roots of 16 and, on a separate
diagram the six sixth roots of 64.
7. Solve the equation z4-z2+ 1 = 0.
8. Solve the equation 1 + z+z2 + z3 + z4+z3 = 0.
4. COMPLEX POWERS OF
COMPLEX NUMBERS
(This section may be omitted at a first reading)
If e is the base of the natural logarithms, and if y is a real number, it
seems plausible to assume that, provided ei1has a meaning,
d(eiy) = j
But w = cos y+j sin y also satisfies this differential equation (by direct
verification). Also, when y = 0, z = 1 = w and thus z = w for all real y:
eiv = cos y +j sin y.
Since = ex . Or,
it follows that ex-1111 = ex(cos y +j sin y). (1)
*Ex. 21. Show that e-iv = cos y — j sin y and deduce that cos y = +e-iv),
1 .
sin y = 2j (ew — e-iv). Deduce from these expressions and the infinite series for
e the series expansions for cos y and sin y.
Ex. 22. Show that cos y = cosh jy and that j sin y = sinh jy.
*Ex. 23. Prove that ev. = eZ= <=> z2 = 2knj, k an integer.
*Ex. 24. Discuss de Moivre's theorem in the light of the expression of a complex
number in the form re.
= e-ig
0.208,
432
3] COMPLEX POWERS
= can 2-1-170+j(i7-1n A/ 2)
Exercise 19(c)
1. If a, b, r, s are real numbers and
u = reie, v = seie,
find:
(i) 1 au+ bv1;
(ii) an argument of the complex number au+ by.
2. Express in the form a + bj:
(i) el±bri; (ii) er-FIrD.
3. Express in the form a+ bj;
(i) In(V3 - j); (■
13 -
4. If z moves once anticlockwise around the unit circle in the Argand diagram,
starting at the point -1, describe the motion of the point representing
5. If z is a complex number and sin z, cos z are defined by
1
sin z = (eiz - e-iz), cos z +
2j 2
prove that:
(i) sin2 z+cos2 z = 1; (ii) sin 2z = 2 sin z cos z;
(iii) cos ( w- z) = sin z; (iv) cos (z1+ z2) = cos z1cos z2-sin z1sin z2.
6. Show that:
(i) cos (x+yj) = cos x cosh y- j sin x sinh y;
(ii) sin (x+yj) = sin x cosh y + j cos x sinh y.
7. Show that:
sin u+j sinh v
tan 1(u+ jv) =
cos u+ cosh v •
If x+ jy = c tan i(u+ jv), express x2+ y2 +c2 in terms of u, v and c. If v and c
are positive constants show that the locus of the point (x, y) referred to Cartesian
axes is a circle of radius c cosech v. (tan z = sin z/cos z.) (London)
433
COMPLEX NUMBERS (2) [19
8. Give a sketch of the representation in the Argand diagram of the two sets:
A = {z e C: = 1, — 4ir < Im z < irr};
B= {wEC: z e A, ezw—ez+w+1 = 0}.
Miscellaneous Exercise 19
1. Solve the equation z8— z4 + 1 = 0 and mark the positions of the roots in the
Argand diagram.
2. Prove that, if n is a positive integer,
(cos 0+j sin 0)" = cos n0+ j sin nO.
By putting n equal to 5 in this formula, or otherwise, prove that
A15-1
sin = (0 & C)
10 4
3. Prove that cos 70 = cos' 0(1 — 21 tang 0+35 tan40 —7 tang 0).
Find the real part of (1 + cos 0+j sin 0)n
(London)
(1 + cos 0 —j sin O)n•
4. Express each of the complex numbers z1= (1 + j),‘12, z2 = 4( — 1 +DA/2 in
the form r(cos 0+j sin 0), where r is positive. Prove that z? = z2, and find the
other cube roots of z2in the form r(cos 0+j sin 0). (0 & C)
5. Solve the equation z4 + 2z2 + 3 = 0, giving the real and imaginary parts of
each root correct to 2 significant figures.
6. If a +16' = ,1{(a+jb)(c +jd)} where a, b, c, d, a and ,8 are real, find the value
of a2 in terms of a, b, c and d. (London)
7. Prove that the roots of the equation (z + 1)" — (z — 1)n = 0 (n 3) all lie on
the imaginary axis. Illustrate the result geometrically in the case n = 3.
8. Given that
1 ,,2 1
cot 1 1
D=
1 1
1 1 co 0)2
where w is a complex cube root of unity, prove that D2= — 27. (0 & C)
n-1 n-1
9. Evaluate E cos (a+ r/3) and E sin (a + rfl).
r---0 r=0
10. If z = cos 0+j sin 0, show that z+ 1/z = 2 cos 0 and find the corresponding
result for z — 11z.
Prove that
cos" 0 = 118[cos 88+8 cos 60+28 cos 40+56 cos 20+35].
to
Evaluate (cos" 0+ sing 0) d0. (London)
fin
434
3] MISCELLANEOUS EXERCISE
3 PPMII 437
20. Mappings in the Argand diagram
*Ex. 1. By writing a = h+kj and z = x+yj, prove conversely that any equation
of type (2) above represents a straight line.
438
STRAIGHT LINES AND CIRCLES
Ex. 2. What is the equation of the perpendicular bisector of the line joining
the points which represent the complex numbers 0 and 2-3j?
*Ex. 3. Show that
a*z+ az* = 2klal2, k real,
represents a straight line and describe its relationship to the line
a*z+ az* = 21a12.
Interpret geometrically the quantity Re (a*z).
Ex. 4. Where does the line (1 +2j) z+ (1 —2j) z* = 12 cut
(i) the real axis; (ii) the imaginary axis?
y
X
0
Fig. 20.2
Ex. 5. Show conversely, that for real a = 0 and real c, the equation
azz* +b*z+bz* + c = 0
represents a circle, provided ac < 1612.
Ex. 6. What is the equation of the circle, centre 1— 2j and radius 3 in the form
(3)?
Ex. 7. Show that 2zz* + (3 —j) z+ (3+j) z* +1 = 0 is the equation of a circle,
centre — +j) and radius V2. Find the centre and radius of the circle with
equation:
(i) zz* — z(1 — 3j) — z*(1 + 3j) + 6 = 0;
(ii) 4zz* — z(2+ 4j) — z*(2 — 4j)+ 1 = 0.
3-2 439
MAPPINGS IN THE ARGAND DIAGRAM [20
We now consider the images of straight lines and circles under three
simple functions f, g, h from the set C into the set C given by
(I) f(z) = w, where w = z fl;
(II) g(z) = w, where w = az;
(III) h(z) = w, where w = z--1;
a, ft being complex constants.
(I) w =
This is clearly a translation of the whole plane by an amount equivalent
to the position vector corresponding to the complex number ft (see
Figure 20.3).
Fig. 20.3
(II) w = az.
As shown in Chapter 17 this maps P -›- Q where OQ = lalOP and
LPOQ = arg a; that is, it represents an extension by the factor I al
followed by an anticlockwise rotation of magnitude arg a (Figure 20.4).
Fig. 20.4
The effect of this mapping upon straight lines and circles may be deduced
analytically as follows:
a*z + az* = c
a* (-1+a (-1 =c
a a
(aa)* w + (aa) w* = clal 2
and 1Z -b1 = r
w -b = r
a
lw-bal = dal.
Thus straight lines are mapped into straight lines and circles are mapped
into circles under the transformation w = az.
*Ex. 9 Show that the centre of the z circle maps into the centre of the w circle
under this transformation and that the ratio of the two radii is loc1: 1.
Ex. 10. Illustrate geometrically the effect of the transformation w = (1 + j) z
upon the circle lz- 11 = 1.
(III) w = z-1.
We must restrict the domain of this function to the whole of C with the
number 0 deleted. To describe the transformation geometrically it is
useful to define the inverse of a point P with respect to a circle. Given a
441
MAPPINGS IN THE ARGAND DIAGRAM [20
circle, centre 0 and radius r, the inverse of the point P is the point P'
on OP such that OP .0P' = r 2.
*Ex. 11. Show that the transformation w = z-1maps the point P into the re-
flection in the real axis of the inverse of P with respect to the unit circle Izi = 1
(see Figure 20.5).
Fig. 20.5
*Ex. 12 Show that a straight line through 0 maps into its reflection in the real
axis under the transformation w = z-1. What lines map into themselves under
this transformation?
*Ex. 13. Show that, under the transformation w = z-1, the straight line through
A and perpendicular to OA maps into a circle with its centre on the line OA*,
where A, A* are the affixes of the conjugate complex numbers a, a*.
*Ex. 14. Show that, under the transformation w = z-1, a circle, centre B, which
passes through the origin maps into a straight line perpendicular to OB*,
where B, B* are the affixes of the conjugate complex numbers b, b*.
*Ex. 15. Show that, under the transformation w = z-I, a circle, centre B, which
does not pass through the origin maps into a circle with its centre lying on the
line OB*, where B, B* are the affixes of the conjugate complex numbers b, b*.
*Ex. 16. Given a circle, centre B, not passing through 0, show that the trans-
formation w = z-1does not in general map B into the centre of the image circle.
*Ex. 17. Show that a diameter of a circle is mapped onto a diameter of the
image circle under the transformations w = z +fl and w = az. Show also that,
under the transformation w = z-1-, a diameter through the origin is mapped
onto a diameter of the image circle.
Example 1. If the point z lies on the circle Izi = 1, find the locus of the
point w where w = j/(z+j).
Method (i)
Consider the sequence of transformations:
fi: z ---> u = z+j,
f2: u -> v = -
1,
u
f2: w = jv,
upon the circle Izi = 1. Their effects are shown in the following sequence
of diagrams (Figure 20.6). (For A, recall the result of Ex. 14.)
The locus is seen to be the line w = + Aj or w + w* = 1.
443
MAPPINGS IN THE ARGAND DIAGRAM [20
0 0
4
fa .f;
Fig. 20.6
Method (ii)
w(z+j) = j
_ j(1— w)
z—
w
Thus
Iz I = 1
I wl = Li(1 -w)I
iwi =
ww* = (1 — w) (1 — w*)
w+ w* = 1, with the same conclusion as before.
w=
jz — 2
maps the unit circle Izi = 1 into a circle with centre on the real axis and radius
V2.
444
TRANSFORMATIONS
Method (i)
Write
2z — 3 + j 2(z + 2j)— 3 — 3j 3 — 3j
w = jz-2 j(z+ 2j) =-2J z + 2j '
upon the circle Izi = 1. (Recall the result of Ex. 15 for deducing the effect
fl
0
f3
Fig. 20.7
445
MAPPINGS IN THE ARGAND DIAGRAM [20
of f2. Notice that, by Ex. 17, the line joining the successive images of ±j
remains a diameter of the corresponding circle.)
Thus, the image is a circle, centre -H(1 —j)+ (3 +j)] = 2 and radius
11(3+1) — (1-1)1 = 412 + 211 = V2.
Method (ii)
2z— 3 +j
w=.
jz— 2 '
(jz — 2) w = 2z — 3 + j
2w-3+j
z= .
jw — 2 •
Thus 1z1 = 1
= 12w-3
(jw — 2) ( — jw* — 2) = (2w — 3 + j) (2w* — 3 — j)
ww* —2j(w— w*)+ 4 = 4ww*— 6(w+ w*)— 2j(w — w*)+ 10
ww* —2(w+ w*)+ 2 = 0.
This equation certainly represents a circle. Furthermore, since the
interchange of w and w* does not affect the equation, it is symmetrical
about the real axis and thus its centre lies on the real axis. It cuts the real
axis at points given by w = w*,
i.e. x2 — 4x +2 = 0,
i.e. x = 2 + ,4,
giving a radius of A/2, as before.
Ex. 19. With the notation of Example 2, if z moves anticlockwise round the circle
IzI = 1, starting at z = j, how does w move around the circle
ww*-2(w+w*)+2 = 0?
Example 3. The function f: C C is defined by f(z) = (z — 1)2. Prove that
the image of the unit circle Izi = 1 under this mapping is a closed curve,
consisting of all points w with the property that r = 2(1 + cos 0), where
r = lwl and 0 = arg w.
Since Izi = 1, we may write z = cos 0+j sin q; then
z 1 = cos g 1 +j sin .75
— -
446
TRANSFORMATIONS
Exercise 20
1. The function f has domain {z E C: jzi = 1}. Find the range off in each of the
following cases:
(i) f(z) = z+ 3; (ii) f(z) = z+j; (iii) f(z) = 2z; (iv) f(z) = jz;
(v) f(z) (1+j) z.
2. In Question 1, if the point z is regarded as moving anticlockwise around the
circle Izi = 1, starting at z = 1, describe the motion of the image point w = f(z)
in each of the cases (i)—(v).
3. The function f: C --> C is defined by f(z) = z3+ z2+2z + 1. Find the values of
z which remain invariant under this mapping and illustrate your answer by refer-
ence to the Argand diagram.
4. The function f: C —> C is defined by
f(z) = az2+ bz+ c,
where a, b, c, a C. If f(1) = 0, f(0) = 2j, f(j) = 1 +j, show that just one z re-
mains invariant under this mapping, and find its value.
What two complex numbers map into 3 + j ?
5. As z moves once anticlockwise around the unit circle I zl = 1, starting at
z = 1, describe the motion of the point w where:
(i) w = z2, (ii) w = z3; (iii) w = jz4, (iv) w =
6. Show that the affixes of the complex numbers 1, — 1 + V3j), -1-(— 1 — V3j)
form the vertices of an equilateral triangle. Describe the effect of the transforma-
447
MAPPINGS IN THE ARGAND DIAGRAM [20
tion w = jz +1+ j upon this triangle, illustrating your result by a sketch of the
Argand diagram.
7. Answer the same question as in Question 6 for the transformation
w = (1+j) z+ 1 +j.
8. Show that, under the mapping defined by w = j/(z —1), the interior of the
circle lz— n = 1 is mapped into the exterior of the circlelw I = 1. Show that two
points remain invariant under this transformation and locate their approximate
positions on an Argand diagram.
9. Show that, under the mapping defined by w = (j — jz)/(1 + z), the interior of
the circle 1z1 = 1 is mapped into the half-plane Im (w) > 0.
10. Show that, under the mapping defined by w = (j + 2jz)/(1 —z), the image of
the set {z e C: Izl < 1) is the set {w e C: Im(w) > —
11. Find the equation, in the form a* z + az* = b, of the line joining the points
z= 1 and z = j.
Find the image, under the mappings defined by w = z/(z + 1), of the set of
points represented by the interior of the triangle with vertices z = 0, z = 1, z = j.
In nos. 12-17 find the image of the circle Iz1 = 1 under the given transforma-
tion. If the image is a circle, find its centre and radius.
2 2 +z 1+
12. w = . . 13. w . . 14. w =
j— 2z j —z 1 jz
—
15.(1+j) z— 1 jz +1 —j
w= 16. w= .
(1—j) z+j (1+j) z— 1
2z— j
17. w =
1+z .
18. Prove that, if (z— 8j)/(z+ 6) is purely imaginary, the locus of z in the Argand
diagram is a circle with centre at the point 4j — 3 and radius 5. (0 & C)
19. Show that the set of points {z} satisfying
arg b) = 0,
a
where 0 is a real number in the interval —77. < 0 < IT and a, b are complex num-
bers, is represented by the arc of a circle through the affixes A, B of a, b.
What is the condition for a point z to lie on the other arc of this circle?
20. Show that, if k is a real number not equal to 1, the set of points {z} satisfying
z— a
=k
z— b
is a circle, with its centre the point C on AB such that CA/CB = k2, and that A
and B are inverses with respect to this circle.
What can you say about this circle (i) if k = 0; (ii) if k is very large?
448
EXERCISE 20
Draw on the same diagram the system of circles obtained for various values of
k by taking a = 1, b = —1.
What happens if k = 1?
21. Show that the transformation defined by w = 1/(3 + j— z) maps the circle
lz— 2— j1 = 1 into the straight line Re (w) = 4. What transformation maps the
line Re (z) = 4 into the circle lw-2—j1 = 1?
22. Show that the transformation defined by w = 1+ z2maps the unit circle
I z1 = 1 into the unit circle 1w — 11 = 1. Draw a sketch of the Argand diagram,
construct the point 1+z2and deduce the above result geometrically.
23. Show that the transformation defined by w = 1/(1 — z)2maps the unit circle
Iz1 = 1 into the curve with polar equation 2r(1 + cos 0) = 1. Draw a sketch of
this curve, indicating on it the images of the points 1, j, — 1, —j.
24. Prove that if
1
x+jy = A+114,
then the points on the Argand diagram defined by making A constant lie on a
circle, and the points defined by making p constant lie on a circle.
Prove also that, whatever be the values of the constants, the centres of the two
systems of circles obtained lie on two fixed perpendicular lines.
25. In the transformation defined by w = 1/(z—j), describe the motion of the
image point w if z moves anticlockwise around the rectangle with vertices 0, b,
b+aj, aj, starting at the origin (where a, b are positive real numbers).
26. The points P and Q represent, in the Argand diagram, the complex numbers
z and 1/(z2+ 2). The point P describes a quadrant of a circle, from the origin
along the real axis to the point z = a, round the arc of the circle Iz1 = a to the
point z = ja, and back along the imaginary axis to the origin. Describe the path
traced out by Q (i) when a = 1, (ii) when a = V2.
27. Find the transformation of the form
az +b
w=
cz+d'
which maps 0 into j, j into 0 and 1 into —1.
Show, that, if this particular transformation maps the complex number g.into
the complex number n, then it also maps 7 into g.
Can you generalize these results in any way?
28. Show that, if the image of the complex number z under the transformation
w = z+z-1is real, then Izl = 1.
Deduce a geometrical construction for determining the roots of the quadratic
equation z2 +az+1 = 0 where a is a real number in the interval —2 < a < 2.
Extend your results to deal with the quadratic equation z2 + az + b = 0, where
a, b are real numbers such that a2 < 4b, by considering the transformation
w = z+ bz-1.
449
21. Quadratic equations and
quadratic functions
w = z+ —
2a'
a process usually known as 'completing the square':
az2+bz+c = 0
Z + -Z — (a + 0)
a a
b)2 =
b 2— 4ac
(z+—
2a 4a2
—b ± V(b2— 4ac).
z—
2a
If b2 = 4ac, only one root is obtained, otherwise equation (2) has two
distinct roots. For uniformity it is useful, in the case b2 = 4ac, to say that
(2) has two coincident roots, or a repeated root; with this convention, every
quadratic equation has two roots.
If the two roots of equation (2) are denoted by a and ,3 then az2+bz+c
has linear factors (z a) and (z—,8) and we may write
—
find:
1 1 (ii) 00 + /62;
(i) a+
——;
)3
(iii) a,/?3+a3/3; (iv) a3+,33;
(v) a6 ±fl5 ;
(vi) la-161.
From (3) and (4), 2j
a+ ft = . = 1+j,
1+.1
1—j .
aft = 1_F —J.
1 1 a+11 1+
(i) + = = .i = 1+j.
a fl aft —J
(ii) Since a, i3 both satisfy the given equation:
(1 +j) a2 — 2ja + 1 — j = 0,
(1 +j) /32 — 213+ 1 —j = 0,
whence, by addition,
(1 + j) (a' +132) — 2j(a + fi) + 2 — 2j = 0.
= —j(4j), by (ii),
= 4.
(iv) Since (1 +j) cz2— 2ja + 1 — j = 0
we have (1+j) a3-2joc2 + (1 —j) a --- 0;
similarly, (1 +j) A3 — 2132 + (1 — j) = 0
whence, by addition,
(1 + j) (a3+,83) — 2j(a2 +132) + (1 — j) (cx + fi) = 0.
Thus 3
(1+j) (a +/33) = 2R4i) — (1 — j) (1+j), by 00,
= —10,
giving a3+ R3 = —5+5j, on multiplying both sides by 10 —j).
[Alternatively, a3+ /33= (a +,q) (0C2— afi + /3)
= +13) {(cc +fl)2 — 3afi}, etc.]
(v) As in (iv)
(1 +j) (a4 +/34) = 2j(a3 +133) — (1 — j) (a2 +/32)
= 2j( — 5 + 5j) — (1 —j) (4j)
= —14-14j
giving cc4 = 14.
Thus, (1+j) (0:5+ /65) = 2i(cc4+ /69 — (1 — j) (cx3+ A3)
= 2j( — 14) — (1 — j) ( — 5 + 5j)
= —38j
giving a5+162 = —19-19j.
[Alternatively,
(co +,62) 0,3 + A3) = cc5 ± n5 + ct2162(ot o‘
p) etc.]
(vi) (a —/3)2 = +fly — 4ocfi
= 6j,
la — AP = 6,
-fil = V6.
on multiplying out and rearranging terms; and this last equation has roots
co,fl2.
Alternatively, the required equation is
w2 —Ew+F = 0,
where E= ± fl2 = (ce, + fly _ 2c96% F 0,2,2 = @fly,
Ex. 3. If a, ,8 are the roots of the equation 3z2 — 5z+ 3 = 0 form the equation
(i) with roots 2a, 2fi; (ii) with roots a — 1, fl— 1; (iii) with roots co, IA
*Ex. 5. Explain why the proof of Theorem 21.1 breaks down if a, b, c are not
restricted to the set R. Explain also why the condition Im (a) = 0 was added in
the enunciation of the theorem.
*Ex. 6. Prove that a quadratic equation with real coefficients either has no non-
real root or two distinct non-real roots.
Ex. 7. If a, ft are two non-real numbers with the property that
Im (a + fi) = Im (afi) = 0,
prove that a, ,8 are complex conjugates.
Ex. 8. Explain why the proof of Theorem 21.2 breaks down if a, b, c are not
restricted to the set Q. Explain also why the condition q 0 was added in the
enunciation of the theorem, and where the fact that Jr is irrational is used.
Exercise 21(a)
1. Write down the sum and product of the roots for each of the following
equations:
(i) z2 — 3z — 7 = 0; (ii) 2z2 — 4z + 11 = 0; (iii) (1 + j) z2— z + (1 — j) = 0;
(iv) (z + j)2= (4 — j) z; (v) (z — 1 + j)2+ za = (2jz — 1)2.
2. Write down the equations with roots:
(i) —2, 3; (ii) — (iii) 3 — V5, 3 + V5; (iv)2— V2,3 + 2V2; (v)3 — 4j, 3 + 4j;
(vi) y1 —0), 1(1 +jV3); (vii) 1 +2j, 1 —j; (viii) V2 —j, 1 V2j. —
3. Solve the following equations, giving the real and imaginary parts of the roots
correct to 2 significant figures. Check your solutions by calculating the approxi-
mate product of the roots (using a slide rule):
(i) 2z2 — 3z— 7 = 0; (ii) 2z2 — 3z + 7 = 0; (iii) z2 + (1 + j) z+ (1 2j) = 0.—
5. oc, ,8 are the roots of the equation 2z2 — 9z— 4 = 0. Find the values of
(i) (1/oe) + (1/,e); (cch69+( 67a2); (iii) ; (iv) Ice2— fl21•
6. cc, /3 are the roots of the equation 3z2 — 2z— 7 = 0. Find the values of :
(i) (a +k) (fl+k); (ii) +ft3; (iii) (1 +a)-1+ (1 +fi)-1.
7. cc, /3 are the roots of the equation (z — a)2= 4zb. Find:
(i) 0,2+,62; (ii) (1/a) + OA; (iii) l a —fl.
455
QUADRATIC EQUATIONS AND FUNCTIONS [21
8. The equations
x2 a2y2 = a 2b22
62 y = mx+ c
are solved simultaneously to give two pairs of solutions x = x1, y = Yi and
x = x2, y = y2• Find -1(x1+ x2) and i-(Yi + YD.
9. If a, fi are the roots of the equation 2z2 — z— 7 = 0, form the equations with
roots:
(i) a — 1, 16-1; (ii) 10a, 10fi; (iii) (1/a), (10); (iv) a2, fi2.
10. If a, 16 are the roots of the equation (1 +j) z2 — 2jz + (2 — 3j) = 0, form the
equations with roots:
(i) a — 2j, ft— 2j; (ii) +j) cc, (1 +D fi; c'e2,
11. If a, fl are the roots of the equation 2z2 — 5z+4 = 0, form the equations with
roots:
(i) a —fl, fl— cc+ifi, fi+ja;
ce 2 +
12. The roots of the equation 4z2 + az— 37 = 0 differ by 1; find the possible
values of a.
13. The roots of the equation z2 — az + 9j = 0 are A and jA. Find the possible
values of a.
14. The roots of the equation (2—j) z2 + (3 + j) z— 4 — 5j = 0 are a and A'. Form
the equations with roots (i) (1/a), (1M); (ii) a*, ,8*.
15. If the roots of the equation az2+ bz + c = 0 are a, le prove that
— aw2+bjw+ c = a(jw— a) Ow—M.
16. If a, b are real and non-zero and z2 + az+ b = 0, where z is non-real, prove
that IzI = 1 b = 1 and I al < 2.
17. Discuss the application of the method of completing the square to the vector
quadratic equation
ar.r+b.r+c = 0 (a * 0).
18. The two equations
z2+az+b = 0,
z2+Az+B = 0
have a common root. Prove that
b(a — A)2— a(a — A) (b — B)+ (b — B)2 = 0.
19. If one root of the equation az2+ bz+ c = 0 is the square of the other, find a
relation connecting a, b, c.
20. If one root of the equation az2+ bz + c = 0 is j times the other, find a relation
connecting a, b, c.
21. Prove that, if one of the roots of the equation z2 + za + 1 = 0 has unit modu-
lus then a is real.
456
3] QUADRATIC FUNCTION
Then y = atx2+-
b x+— b2 +- – b2 1
a 4a2 a 4a2f
= \ 2 + 4 a c b21
x+—b
a{( 2af 4a2 f . ( 1)
(i) If a > 0 and b2 < 4ac, then y > 0 from (1), since
(ii) If ax2+bx+ c > 0 for all x, then certainly a > 0, for we may choose
x such that
( b)2 4ac – b2
x — > 0.
+2a 4a2
But, if b2 > 4ac, the equation ax2+bx+ c = 0 has real roots, a, ft say, and
ax2+bx+c = a(x– cc) (x – fl),
from which it follows that y < 0 if a < x < ft. But we are given that y is
always positive; thus b2 < 4ac.
*Ex. 11. If the roots of the quadratic equation ax2+bx+c = 0 are a, A prove
that a2(a—f)2= A. If the coefficients are real, what does this tell us about the
reality of the roots of the quadratic equation?
(ii) I al >
By a similar argument, the range is
{ye R:y<a or y fib
t If y = 0, the equation is no longer a quadratic, but the subsequent inequality is
still satisfied for all a.
458
3] QUADRATIC FUNCTION
(iii) a = 2:
E = 4y +1 and thus the condition E 0 gives us the range
{y E R: y
(iv) a = —2:
By a similar argument, the range is
{y R: y <
Ex. 12. Illustrate the solution of Example 3, by drawing rough sketches of the
curve y = (x + a)/(x2+ ax+ 1) in the separate cases a = —4, a = —2, a = 0, a = 2,
a = 4. [Express y--1in the form z = y-1 = x+ (x + a)-1and deduce that the graph
of z has turning points at x + a = ± 1; hence sketch the y curve.]
Miscellaneous Exercises 21
1. If a, /3are the roots of the equation z2 — 5z-11 = 0 form the equation
(i) with roots cc + (/3/cc), + (al13); (ii) with roots (a +16', ja +,3).
2. Ha, ft are the roots of the equation az2+ bz+ c = 0, find the roots of the equa-
tion
a2z2+(2ca— b2) z+ c2 = 0 in terms of cc, /3.
3. If a, b, c, d are real numbers, prove that the roots of the equation
a—z b
=0
c d— z
can be complex only if b and c are of opposite signs.
4. Show that, if x is real,
< 3x2+x+3
5.
x2+x+1 --
5. Show that, if x is real, the expression
2x2 — 13x— 7
x2 — 3x
cannot assume any value between 25/9 and 9.
6. Show that, if x is real, the expression
x2 — 4x + 3
x2 — 6x+ 8
can attain all real values.
7. If cc, ft are the roots of the equation ax2+ bx+ c = 0 express the roots of
the equation
ac(x2+1)— (1)2—2ac) x = 0
in terms of cc and /3.
Prove that, if x,, x2are roots of the equation
(x2 + 1) (a2 + 1)— nax (ax +1) = 0,
then (4+1) (4+1) = naxi x2(xi+ x2). (0 &C)
459
QUADRATIC EQUATIONS AND FUNCTIONS [21
Hence, or otherwise, prove that, if the quadratic equations have a common root,
it is a repeated root of one of them. (0 & C)
460
22. The parabola and rectangular
hyperbola
Ex. 1. Show that, if both x and y are given as linear expressions in a parameter t,
then the equation connecting x and y is linear and thus represents a straight line.
Ex. 2. Find the Cartesian equation of the line given by the parametric form
x = 3t-2, y = 2t+3;
and find a parametric form for the line with Cartesian equation
3x-5y-7 = 0.
If the expressions for x and y involve non-linear expressions in a para-
meter t, the set of points so defined generally constitute a curve. For
example,
= {(x, y): x = t, y = t 2; t real}
represents a curve, since the expression for y in terms of x is not linear.
461
PARABOLA AND RECTANGULAR HYPERBOLA [22
Ex. 3. Show that the Cartesian equations of the curve P defined above is
y = x2.
Distinguish between the curve P and the curve I' defined by
= y): x = t2, y = t4; t
Sketch r and r for — 3 < x < 3.
Ex. 4. Show that the set of points
{(x, y) e R: x = t20, = t2- 1; t real}
constitute a line segment. (Only in such artificial cases does a non-linear para-
metric expression yield a straight line.)
The expressions for x and y need not be algebraic: the circle with centre
(0, 0) and radius 2 may be expressed in the parametric form
x = 2 cos 0, y = 2 sin 0.
Ex. 5. Find the equation of the curve with parametric representation x = a sec 0,
y = b tan 0.
is the position vector of P with respect to O'. Put another way, if 0' is
taken as the origin, with axes 0' x' , 0' y' in directions defined by i, j then
462
2] COORDINATE AXES
the coordinates (x', y') of P with respect to the new axes and origin 0'
are given by
x' = x—h, y' = y—k.
Fig. 22.1
x = x'+2, y = y'-3,
and the given equation becomes
(x'+2)2 +(x'+2) (y'-3)-2(y' —3)2 —(x'+ 2)-14(y'— 3)— 20 = 0,
which simplifies to
x' 2+x'y' —2y'2 = 0.
But the expression on the left of this equation factorizes:
(x'+2y')(x'—y') = 0;
and thus the equation represents the pair of straight lines
x'+2y' = 0, x'—y' = 0
through 0'.
463
PARABOLA AND RECTANGULAR HYPERBOLA [22
Exercise 22(a)
Find the equations of the curves given parametrically in Questions 1-6.
1. x = cos 0, y = 2 sin O.
2. x= t 2, y = t3.
3. x = 1 -t,y = 1+t+t 2.
4. x = cos 0 -sin 0, y = cos 0+ sin 0.
ate at
5. x = ,y=
1 + t3 1+
6. x = t + 1/t, y = 1 + t2.
7. Show that the parametric form
1 - t2 2t
x= y=
1+t2 ' 1 + t2
represents the curve x2 + y2 = 1,
less the point (-1, 0).
8. If the axes are translated to (1, -4) as new origin, find the new coordinates of
the points:
(i) (0, 0); (ii) (3, -4); (iii) (3, 4); (iv) (5, -5).
9. By translating the axes to pass through (1, 1) as new origin prove that the
equation
x2 - 4y2 - 2x+ 8y- 3 = 0
represents a pair of straight lines, and find their gradients.
10. Prove that the equation
2x2 +3xy- 2y2 - 17x + 6y+ 8 = 0
represents a pair of straight lines intersecting in the point (2, 3).
11. Show, by a translation of axes to pass through (p, q) as new origin, that the
equation
ax2+2hxy+ by2+2gx+2fy+ c = 0
is transformed into the equation
ax' 2+2hx'y' + by' 2+ d = 0
if p, q are chosen so that ap+hq+g = 0,
hp+ bq+f = 0.
464
2] COORDINATE AXES
3. THE PARABOLA
A parabola is defined as the set of points in a plane whose distances from
a fixed point, S, are equal to their distances from a fixed line 1 not con-
taining S: see Figure 22.2. For the two typical points P, and P2,
SP, = P1 MI, SP2 = P2 M2.
Fig. 22.2
Fig. 22.3
that the coordinates of P referred to our axes of coordinates are (x, y);
let M be the foot of the perpendicular from P on to I.
Since XO = OS, S is the point (a, 0) and 1 has equation x+a = 0.
Thus
SP2= (x — a)2+ y2,
PM2= (x+ a)2,
and P = {(x, y): (x— a)2+ y2= (x+ a)2}
= {(x, y): y2= 4ax}.
466
3] PARABOLA
Thus, referred to its axis as the x axis, its vertex as the origin and S as the
point (a, 0), the equation of a parabola is
y2= 4ax. (1)
*Ex. 9. Show that x = at 2, y = 2at is a parametric form for the parabola
y2 = 4ax. Need a be positive in this equation?
From the equation derived above it may be seen that all parabolas have
a similar shape but differ in size (depending upon the distance of the focus
from the directrix). The chord of the parabola through S and perpendicular
to the axis is called the latus rectum; the length of the latus rectum deter-
mines the size of the parabola.
*Ex. 10. Show that the length of the latus rectum is 2SX.
*Ex. 11. What is the length of the latus rectum of the parabola y2 = 4ax?
Answer the same question for the parabolas y2= — ax and x2 = 3ay.
Ex. 12. Using the same axes, plot, on graph paper, the parabola y2 = 4ax, for
a = 2, 1, I, -116, — 1, — 2.
*Ex. 13. By writing the equation
y2 — 4ax-2ay— 3a2 = 0
in the form (y — a)2 = 4a(x+ a)
and translating the axes to the new origin ( — a, a), show that it represents a
parabola of latus rectum 4a. Show also that, with the original coordinate system,
the vertex is the point (— a, a), the focus (0, a) and the directrix x +2a = 0.
Ex. 14. Show that the equation
y2 — 2ax-2ay+$a2 = 0
represents a parabola and find its focus and directrix. Answer the same question
for the equations:
(i) y2 + 3ax + 6a2 = 0; (ii) x2 + 2ax-8ay-7a2 = 0.
The chord joining the points Nat?, 2at1), P2(ad, 2at2) of the parabola
is easily obtained. Its gradient, m, is given by
2at2-2ati
m—
all— at?
2
tl + t2'
and thus its equation is
y-2at1 att.)
= t2
Ex. 20. Show that, if the normal to the parabola y2 = 4ax at the point
P(ap2, Zap)
cuts the parabola again at the point Q(aq2, 2aq), then q is a root of the quadratic
equation in A
pA2 +2A — 2p — p3 = 0.
What is the other root of this equation? Find q in terms of p.
468
-
3] PARABOLA
2 tan a
Thus tan fi =
1— tang a
= tan 2a
and the result follows.
Property (ii) is usually known as the parabolic reflection property since
it shows that a ray of light emanating from the focus will be reflected by a
parabolic mirror along a line parallel to the axis.
Fig. 22.4
Ex. 21. Explain the use of parabolic reflection in car headlights and reflecting
telescopes. Explain, with the aid of diagrams, the ' headlights on' and 'headlights
dipped' positions.
y2 = 4ax.
470
3] PARABOLA
Any line through the focus, S(a, 0), other than x — a = 0, has the form
y = m(x—a).
(Notice that m is a parameter: different values of m give different lines
through the focus and all such lines, with the single exception mentioned
above, may be put in this form.)
The line y --- m(x — a) cuts the parabola y2 = 4ax in points with y co-
ordinates y, and y2given by the roots of the equation
y2 = 4a(7 11- + a),
Notes on Example 2.
(i) The use of symmetric functions of the roots of a quadratic equation
is frequently applicable to problems involving the parabola and has the
advantage of retaining symmetry in the calculation.
4-2 471
PARABOLA AND RECTANGULAR HYPERBOLA [22
R parametrically.
(iii) If LL' is the latus rectum of the given parabola, the mid-point of
the particular chord LL' is S which must therefore lie on the required
locus, giving us a check on our working. Again, the symmetry of the locus
about the x axis was to be expected from the symmetry of the parabola
about its axis.
Ex. 22. Show that if the tangent at the point P(at 2, 2at) passes through the point
R(h, k) then t is a root of the quadratic equation
aA2 —kl+h = 0.
Deduce that the coordinates of the meet of tangents at Pi(att, 2at1) and P2(at2, 2at2)
is R(at1 t2, a(ti + t2)).
Exercise 22(b)
1. Find the latus rectum, vertex, focus, and directrix of each of the following
parabolas:
(i) (y-2)2 = 4(x— 3); (ii) (x— 3)2 = 2(y+ 1);
(iii) (y+ 1)2 = — (x-1); (iv) (x + 2)2 = 3(y + 3);
(v) (y +3)2 = 4(x-5); (vi) (x-2)2 = 8(y-2);
(vii) (y-4)2 + 2x = 0.
2. Find the equations of the parabolas having the given points as foci and the
given lines as directrices:
(i) (2, 1), x = 0; (ii) (-3, 2), y = 3.
3. Find the equations of the parabolas: (i) with focus (4, 3) and vertex (1, 3);
(ii) with focus (-1, — 7/4) and vertex (— 1, —1).
4. Find the equations of the tangents to the following parabolas at the points
given:
(i) y2= 2(x + 1), (7, —4); (ii) (y + 2)2 = x, (9, 1);
(iii) y2 — 3y— 5x = 0, (2, 5); (iv) x2 — 2x— 3y— 21 = 0, (— 3, —2).
5. Show that the line x+ y— 2 = 0 is a tangent to the parabola x2 + 8y = 0 and
find the point of contact.
6. Show that the line 5x+ 2y + 8 = 0 is a tangent to the parabola
x2 -2x-4y = 0,
and find the point of contact.
7. Any element P of the set S of points in a plane has a position vector of the
form
r = (t —1) i+ t 2j,
where t is a scalar parameter. Show that S defines a parabola and determine the
position vectors of its vertex and focus.
472
3] PARABOLA
8. A ball is thrown obliquely into the air in such a way that its position vector
at time t is given by
r = lOti+ (10t — 16t2) k,
where i is a horizontal unit vector and k is a unit vector pointing vertically
upwards. Show that the path of the ball is a parabola, and determine the distance
of its directrix from the origin.
9. Any element P of the set S of points in a plane has a position vector of the
form
r = cos 20.i— (1 + sin 0).j,
where 0 is a scalar parameter. Show that S forms part of a parabola with its axis
parallel to i and sketch the set of points S.
Questions 10-15 refer to Figure 22.6. 0 is the vertex, S the focus of the parabola,
AM is the directrix, ANPM is a rectangle. The tangent at P cuts the axis at T
and the normal at P cuts it at G. PT and SM intersect at Z. You may use any
method which you regard as suitable.
Fig. 22.6
10. Prove that SZ and PT are perpendicular and that Z lies on the tangent at
the vertex. Suggest how, with the aid of a set square, this property may be used
to draw a parabola.
11. Prove that ST = SP and deduce that SPMT is a rhombus. (Notice that this
gives the parabolic reflector property.)
12. Prove that ST = GS and deduce that NG is of fixed length.
13. Prove that NT is bisected at 0.
14. Prove that SZ2 = SO . SP.
15. Prove that TZ = ZN and that GP = 2SZ.
473
PARABOLA AND RECTANGULAR HYPERBOLA [22
27. Prove that the equation of the normal at the point P(at 2at) to the parabola
y2 = 4ax is tx+ y = at 3+2at,
and that it meets the parabola again at Q(aT2, 2aT) where T = — t— 2/t.
The tangents at P and Q meet again at R; prove that, if P is a variable point
on the parabola, the locus of R is
y2(x+2a)+4a2 = 0. (0 &C)
28. The tangents at two points P and Q on the parabola y2 = 4ax meet at R.
Find the equation of the locus of R in each of the following two distinct cases:
(i) when the sum of the ordinates of P and Q is 4a;
(ii) when the sum of the squares of the ordinates of P and Q is 4a2. (0 & C)
(The ordinate of a point is the y coordinate.)
29. The chord PQ of the parabola y2 = 4ax subtends a right angle at the vertex,
P and Q being the points (at 2, 2at) and (aT2, 2aT). Prove that tT+ 4 = 0 and that
the locus of the point of intersection of the normals at P and Q is
y2 = 16a(x — 6a). (0 & C)
30. Show that the differential equation
d2y/dx2 = a,
where a is a constant, defines a family of parabolas (that is, any solution of the
given equation represents a parabola). If a = 6, find the equation of the parabola
of the family which has its vertex at the point (1, —1).
31. Prove that, in general, a unique parabola with its axis parallel to a given
direction may be drawn through three given points. What exceptional cases
may arise?
Find the equation of the parabola with axis parallel to the y axis and passing
through the points (— 1, 5), (0, 2), (1, 3).
32. A man rows a boat across a river flowing at a constant speed v. If his speed
in still water is also v and he always aims at a point on the further bank exactly
opposite his starting point, show that his path is a parabola.
> X
Fig. 22.7
Ex. 24. Show that, if any line through 0 cuts the rectangular hyperbola at P
and Q, then 0 is the mid-point of PQ. (This corresponds to a property of the
centre of a circle.)
Ex. 25. Trace the position of the point P(ct, c/t) as t varies from —co to +co.
The chord joining the points P1(ct1, OD and P2(ct2, OD on the rectangu-
lar hyperbola has equation
c c(t2 — t1)
Y t, = cti t2(t — t2) (x — cti)
1
476
4] RECTANGULAR HYPERBOLA
which reduces to
x + ti to, = c(ti+ 4).
The tangent at P1may be obtained immediately by letting t2
x+ ti y = 2ct1.
*Ex. 26. Obtain the gradient of the tangent at P1by differentiating the equation
xy = c2.
*Ex. 27. Obtain the equation of the chord P1P2by the method of Ex. 18.
Ex. 28. Show that the equation of the normal to the curve at P1is
trx– y = c(tt – 1)/t,.
a+k f = 2h,
which may be rewritten as
(11—P) (k-413) = ice 16•
477
PARABOLA AND RECTANGULAR HYPERBOLA [22
Ex. 29. The locus sought in Example 3 may be definitely only part of the curve
(x— -fa) (y— 3-16) = +aft. Show this by taking A as the origin.
Exercise 22(c)
1. The tangent to a rectangular hyperbola at the point P meets the asymptotes
at Q and R. Prove that P is the mid-point of QR.
2. An axis of symmetry cuts a rectangular hyperbola at A, A'. P is any point of
the hyperbola and N is the foot of the perpendicular from P to AA'. Prove that
PN 2 = AN .NA'
3. With the notation of Question 2, if 0 is the centre of the hyperbola and G is
the point of intersection of the normal at P with the line AA', prove that
ON = NG.
4. With the notation of Questions 2 and 3, if Z is the foot of perpendicular from
O to the tangent at P, prove that OZ .0P = 0A2.
5. With the notation of Questions 2 and 3, if the tangent at P meets AA' at T,
prove that ON .0T = 0A2.
6. Show that the equation
(x — a) (y — b) = c2
represents a rectangular hyperbola and find its centre and asymptotes.
Find the centre and asymptotes of the rectangular hyperbola
xy—x+3y-7 = 0.
7. A variable line passes through the point (1, 1) and meets the x and y axes at
P, Q respectively. The rectangle OPRQ is completed. Prove that R lies on a rect-
angular hyperbola, and find its centre and asymptotes.
8. The tangent to a rectangular hyperbola at the point P meets the asymptotes
at Q and R. If 0 is the centre of the hyperbola, prove that QR = 20P and that
the area of the triangle OQR does not depend upon the position of P.
478
4] RECTANGULAR HYPERBOLA
Miscellaneous Exercise 22
1. Prove that the line px + qy = 1 is a tangent to the parabola
y2 = 4a(a— x) if a(p2+q2) = p.
A line 1 and a point 0 not on / are given; P is a variable point on 1 and Q
divides OP in a fixed ratio. Prove that the line through Q perpendicular to OP
touches a fixed parabola with focus at 0 and directrix parallel to 1. (0 & C)
2. The coordinates of the mid-point of the line joining the points (au2, 2au),
(av2, 2av) are (X, 1'). Express u+ v and uv in terms of X and Y.
A variable chord of the parabola y2 = 4ax passes through the fixed point
(b, 0). Prove that the locus of the mid-point of the chord is a parabola, and
find the coordinates of its vertex and focus. (0 & C)
3. The normal at a point P meets the parabola again at Q. Find the length of
PQ in terms of the acute angle 0 between the normal and the axis of the parabola.
4. Two parabolas, y2 = 4a(x+ b) and x2 = 4a(y+ b), where a > 0, are given.
Prove that each point of intersection lies either on the line y = x or on the line
y+x+4a = 0.
Hence, or otherwise, prove that, if the parabolas have four real points of
intersection, then b > 3a. (0 & C)
5. Lines are drawn through a variable point P of the parabola y2 = 4ax, making
angles a and 7r— a with the axis of the parabola. The lines meet the parabola
again at Q and R. Prove that:
(i) the point of intersection of the tangents at Q and R lies on the parabola
y2= 4a(x + 4a cot2 a);
(ii) the line QR touches the parabola y2= 4a(x — 4a cot2 a). (0 & C)
479
PARABOLA AND RECTANGULAR HYPERBOLA [22
6. Prove that the line y = mx + aim touches the parabola y2= 4ax for all values
of m. Deduce that, if two tangents to a parabola are perpendicular, then their
intersection lies on the directrix.
7. Two directed lines meet at A, and points P and 13' are taken on the two lines
so that AP+ AP' is constant, sense being taken into account. It is required to
prove that PP' touches a fixed parabola. Prove this by first showing that the
equations of the lines may be taken in the form
y = (x + c) tan a and y = — (x+ c) tan a,
where 2a is the angle between the lines and
AP+ AP' = 2c sec a,
and that the coordinates of P and 13' may be taken in the form
(t cos a, t sin a + c tan a) and (— t cos a, t sin a— c tan a).
Deduce that PP' touches the parabola y2 = 4cx tang a. (0 & C)
8. Prove that the mid-points of parallel chords of a parabola lie on a straight
line (called a diameter of the parabola) parallel to the axis.
Given a parabola traced on paper, find a construction for its focus.
9. If a point is marked on a rectangular sheet of paper, one of whose sides is AB
and the paper is then folded in such a way that AB passes through the point,
show that the crease will always touch a fixed parabola.
10. The tangents 1, m from a point P to a parabola meet the directrix at L, M
respectively. The other tangents to the parabola from L, M meet m,1 at Q, R
respectively. Prove that QR passes through the focus, S, and that the angle PSQ
is a right angle.
11. A straight line passes through the point Pi(xi, yi) and has gradient m = tan a.
Prove that a general point of the line has coordinates (xi + r cos a, +r sin a),
where r is a parameter, and that the values of r giving the points of intersection of
the line with the parabola y2 = 4ax are the roots of the quadratic equation
r2 sin2 0+2r(yisin 0-2a cos 0)+A— 4ax, = 0.
Deduce that, if Piis the mid-point of the chord QR, then the gradient of QR
is 2a/y1and its equation is
yiy-2ax = y1-2ax1.
12. Two chords P1Q1, /32 Q2of a parabola intersect at K. Prove that the value of
the fraction
(KA.KQI)/(KP2.KQ2)
depends on the directions of the two chords, but not upon the position of the
point K.
[This result, known as Newton's Theorem, holds for all conics (see Chapter 27).
Use the analysis of Questions 11 to prove the result for the parabola.]
13. Follow through the work of Questions 11 and 12 to prove Newton's Theorem
for the rectangular hyperbola xy = c2.
480
MISCELLANEOUS EXERCISE
14. Three points P(ap2, 2ap), Q(aq2, 2aq), R(ar 2, 2ar) taken on the parabola
y2 = 4ax are such that PQ subtends a right angle at R. Show that
(p+r)(q+r) +4 = 0.
Show that every chord of the parabola which subtends a right angle at R inter-
sects the normal at R at the same point F. Show also that, as R varies, F describes
another parabola. (London)
15. The chord AB joining points A(cti, OD and B(ct2, 02) on the rectangular
hyperbola xy = c2is of constant length 1. Show that, as the position of the chord
varies, the centroid G of the triangle AOB, where 0 is the origin, moves on the
curve (9xy _4c2) (x2 ± y2) pxy.
Find the area of the triangle AOB when the coordinates of G are (c, 2c).
(London)
16. Prove that the equation of the tangent PT at the point P(ct, c/t) on the
rectangular hyperbola xy = c2 is
x+ t2y = 2ct.
The perpendicular to PT from the origin 0 meets PT at Q and the normal at
P meets the hyperbola again at R.
Prove that:
(i) as P varies, the locus of Q is
(x2 + y2)2 = 4c2xy;
and
(ii) c2.PR = OP3. (0 & C)
17. Prove that the equation of the chord joining the points P(ct, c/t) and
Q(cT, c/T) on the rectangular hyperbola xy = c2 is
x+tTy = c(t+T).
Mis the mid-point of PQ and PQ meets the x axis at N. Prove that OM = MN,
where 0 is the origin.
The line through N parallel to OM meets the hyperbola at Rand S whose para-
meters are t, and t2; C is the mid-point of RS. Prove that:
(i) CM is parallel to the y axis;
(ii) t, t, = — tT. (0 & C)
18. Prove that the equation of the tangent at the point (11, k) on the parabola
y2= 4ax and the equation of the tangent at the point (H, K) on the rectangular
hyperbola xy = c2are, respectively,
ky = 2a(x+ h) and Kx+ Hy = 2c2.
Find the coordinates of the point of intersection, P, of the parabola y2 = 4ax
and the rectangular hyperbola xy = 4a2A/2 and prove that the tangent to the
parabola at P is the normal to the rectangular hyperbola at P.
Prove that, if this normal meets the rectangular hyperbola again at Q, then
the abscissa of Q is —4a. (0 & C)
481
PARABOLA AND RECTANGULAR HYPERBOLA [22
19. Prove that the normal to the rectangular hyperbola xy = c2 at the point
P(ct, t) meets the curve again at Q(— eIt3, — ct 3).
The circle on PQ as diameter meets the hyperbola at R; prove that PR passes
through the origin. (0 &
20. Given a point P on a rectangular hyperbola, prove that one and only one
real chord PQ can be drawn which is normal to the hyperbola at Q. Deduce that
there is only one (real) chord AB of the hyperbola which is normal at both A and
B and locate the two points A, B.
21. Points P, Q, R, S are taken on the rectangular hyperbola xy = c2. Prove
that PQRS is a rectangle if and only if the parameters of the four points are of the
form t, — t, —t-2. Deduce that it is impossible to inscribe a square in a rect-
angular hyperbola.
22. The tangent to a rectangular hyperbola F at the point P meets the asymptotes
at Q and R. 0 is the centre of the hyperbola and 0 QSR is a rectangle. If SQ, SR
cut P at Q', R' prove that Q'R' touches a second rectangular hyperbola whose
asymptotes coincide with those of P.
23. Prove that all chords of a parabola which subtend a right angle at a point
P on the parabola pass through a fixed point Q.
Q is called the Fregier point of P. If a given chord subtends right angles at
two points, P and 13' of the parabola, prove that the join of the two Fregier
points, Q and Q', is parallel to the given chord.
24. Find the equation of the tangent to the parabola SI: y2 = 4ax at the point
P(at 2, 2at).
If this tangent cuts the rectangular hyperbola S2: xy = k2at real points Q, R
and M is the mid-point of QR show that M lies on part of the parabola E:
2y2 + ax = 0.
Sketch in the same diagram the curves Si and S2 and the set of all such points
M in the case a > 0. Mark in your diagram the line other than x = 0 which
is a tangent to both S1and S2.
482
23. Polynomial equations
z4 +(3+2j) z2 —(1—j)z+4 = 0
it is not at all obvious whether or not the equation has a solution. Indeed,
two questions arise:
(i) Does every equation possess a solution? (That is, is there a complex
number which satisfies the equation ?)
(ii) If a solution exists, can it be found by processes similar to those
employed for solving a quadratic equation?
The answer to the first question is ' Yes' but the proof of this is rather
difficult (but not beyond the understanding of a really enthusiastic pupil : see,
for example, Hardy's Pure Mathematics, Appendix II or Courant and Rob-
bins What is Mathematics?, Chapter 5). The result was first proved, like so
many other central theorems of mathematics, by C. F. Gauss and is gener-
ally known as ' The Fundamental Theorem of Algebra': explicitly, it states
that any polynomial equation with complex numbers as coefficients (and
in particular, any equation with real coefficients) has a root which is a
complex number (R is regarded as a subset of C in this result). We shall
content ourselves with assuming the truth of this theorem.
The answer to the second question is ' Yes, if the degree of the given
equation is less than five' but 'No, in general, if the degree is five or more'.
The first part of this result is comparatively easy to prove and is due to a
number of Italian mathematicians of the Sixteenth Century; the second
part is very much harder and is due to two mathematicians, E. Galois
(1811-32) and N. H. Abel (1802-29). (The truly dedicated reader may care
to follow the matter up in, say, Birkhoff and Maclane, 'Survey of Modern
Algebra. Brief readable accounts of the life and work of Galois and Abel
may be found in Bell, Men of Mathematics, Volume
483
POLYNOMIAL EQUATIONS [23
Ex. 2. Solve the cubic equation z3+3z2-4 = 0, given that it has an integral
root of multiplicity 2.
*Ex. 3. Assuming that every polynomial equation has a root, prove, by induction,
that a polynomial equation of degree n has n roots, where a root of multiplicity
r is counted as r roots.
*Ex. 4. By writing P(z) (z - a)* Q(z) prove that, if P(z) = 0 has a root of
multiplicity r, then P'(z) = 0 has a root of multiplicity (r- 1).
Prove that the converse is false.
*Ex. 5. If P(z) = 0 has a root of multiplicity r > 2, what can you say about the
equation P"(z) = 0?
*Ex. 8. Prove that, if a(' and anhave opposite signs, the equation P(z) = 0 has
a real root.
We conclude this Section with two worked examples. The first shows
how one equation may be transformed into another in such a way that
485
POLYNOMIAL EQUATIONS [23
the roots of the two equations bear a given relationship to one another.
The second example indicates how calculus methods may be used to
solve problems on polynomial equations.
(2, -32)
Fig. 23.1
k < 0: the equation has two unequal real roots and two complex con-
jugate roots.
k = 0: the equation has three unequal real roots, one of multiplicity 2.
0 < k < 5: the equation has four unequal real roots.
k = 5: the equation has three unequal real roots, one of multiplicity 2.
5 < k < 32: the equation has two unequal real roots and two conjugate
complex roots.
k = 32: the equation has a real root of multiplicity 2 and two conjugate
(complex) roots.
k > 32: the equation has two pairs of conjugate complex roots.
Ex. 10. What are the roots of the equation in Example 2 in the cases
(i) k = 0; (ii) k = 5; (iii) k = 32?
Exercise 23(a)
1. Form the equation with roots 1, —2, 3 —0, 3+0.
2. Form the equation with roots 1+2j, 1— 2j, — 2, 3.
3. Form the equation with a root —1 of multiplicity 3 together with a pair of
roots +j, —j.
4. Solve the equation
z4-2z3 +3z2 — 4z+ 2 = 0,
given that it has a root of multiplicity 2.
5. Use the Remainder Theorem to solve the equation
z3-3z2 — 8z +30 = 0,
given that it has an integral root.
6. The equation
24z4 +100z3 +126z2 + 27z— 27 = 0
has a root of multiplicity three. Solve the equation completely.
7. If a, fi, y are the roots of the equation
z3 — z— 4 = 0,
form the equation with roots:
111 a+1/1+1y +1
(i) (11) , fi , 7
488
1] PRELIMINARY OBSERVATIONS
In words, the sum of the roots one at a time is — a,lao, two at a time is
+adao, three at a time is — adao, and four at a time is adao. For brevity,
we denote a + y + 8 by Ea, afl + ay + a8+ fly ,88±y8 by Zaft, etc.;
no ambiguity arises provided we know the degree of the equation under
consideration (in this case four).
Ex. 11. The roots of the equation
2z3 3z2— 5z -13 = 0
—
are ex, ft, y. Write down the numerical values of /a, Zafi, afiy.
Ex. 12. The roots of the equation
z4 — 4jz+ 1 = 0
are a, A y, 8. Write down the numerical values of /a, Zafl, Zafly, aflya.
*Ex. 13. If the coefficients of a cubic equation are all real, it follows that
Ea, /aft and fly are all real. Explain how this can be so, even if the equation
has complex roots. Illustrate your answer by considering the equation z3— 1 = 0.
Now suppose that the general polynomial equation
ao zn aizn-1 a 2zn-2 + an = 0
has roots al, a2, ..., an. Then it may be shown by induction that the results
proved previously in the particular cases n = 2, 3, 4 hold generally.
Zai = — di /a°
Za ja; = +a,lao
Eat c6i ak = — as/a,
Consider the product (Ea) (Eat): each term such as a2,6 occurs just once,
but the terms a,67 appears three times, as a .,67, /6.7a and 7. aft. Thus
Ea2,6 = EaEaft — 34'7
= —18.
(iii) Since a is a root of the given equation, we have
a3 — 4a — 6 = 0
and similarly for i6 and y. Adding these three results
Ea3 4Ea — 18 = 0.
-
Exercise 23(b)
1. If a, y are the roots of the equation
z3-4z2 +z+2 = 0,
find:
(i) Ea2; (ii) Eat; (iii) Ea4; (iv) Ea2fl;
(v) Ea2/32; (vi) (a + 2) (fl + 2) (y + 2); (vii) El/a;
(viii) Eafi(a2 +,82); (ix) El/ale; (x) E(cc + 08)2.
2. If a, y, 8 are the roots of the equation
z4-4z3 — z + 9 = 0,
find:
(i) Ea2; (ii) Ea2fl; (iii) El/a; (iv) Ea3;
(v) Ea3fi; (vi) I(a+fi+y)2; (vii) E(a+ 1) (13+ 1).
3. If a, fi, y are the roots of the equation
z3+pz+q = 0,
find:
(i) Ea4; (ii) El/a4; E(ce +fl —27).
4. Solve the equation
2z3 -3z2 — 23z+ 12 = 0,
given that its roots are successive terms of an arithmetic sequence.
5. Solve the equation
32z3 + 48z2 + 6z— 5 = 0,
given that its roots are successive terms of an arithmetic sequence.
6. Solve the equation
24z3 — 382.2 + 19z— 3 = 0,
given that its roots are successive terms of a geometric sequence.
7. Solve the equation
2z3 — 11z2-5z+ 50 = 0,
given that one of the roots is twice a second root.
8. Solve the equations:
x+y+z = — 4,
yz+zx+xy = 1,
xyz = 6.
494
2] RELATIONS BETWEEN ROOTS
If the roots of (1) are z1, z2, z3and the roots of (2) are w1, w2, w3, then the
geometrical interpretation of the transformation z --> (w — al)la0 is that it
transforms the triangle z1z2z3into the similar triangle w1w2 w3which has its
centroid at the origin (w1+w2+w3 = 0). (See Figure 23.2.)
W2
Fig. 23.2
if we can find values p, q satisfying (4) then we shall have a complete solu-
tion of the cubic equation.
From (4), p3and q3are the roots of the quadratic equation
Z2 + GZ— H3= O. (5)
Let a, /6 be the roots of the quadratic equation (5). Take p„ q, as any
cube roots of a and 13; then
14+0 = a+161= —G,
Pi qi = aft = — H3.
If H = 0, the original cubic is immediately soluble; if H 0, then
(ART = 1,
Hl
giving Pigs 6
—H '
where e = 1, a) or (o2, andpi and qile are suitable values to take for p and q.
In summary, to solve the equation
w3+3Hw+ G = 0 (H 0),
let a, ft be the roots (repeated if necessary) of the quadratic equation
z2+ Gz— H3 = 0,
and let p = z1, q = 0e where e = ca01(— H). Then
p wp (.02q, (02p + (6)
are the required roots for the cubic equation.
For
(w — + (w — (cup + w2q)) (w — (Op + cog)) w 3 -3wpq — (pa ±q3),
and p, q have been so chosen that
pq = —H, p3+q3= —G.
pq = 2.
Thus, p3and q3are the roots of the quadratic
Z2 +6Z+8 = 0,
498
3] CUBIC EQUATION
Fig. 23.3
Ex. 19. Show that w2, w, may be found by constructing equilateral triangles on
—p, —q as base.
499
POLYNOMIAL EQUATIONS [23
From now on we shall suppose that the coefficients of (2) are real: then
either one root is real and the other two are conjugate (complex) numbers,
or all three roots are real. From (5), since G, H are real, either p3or q3 are
both real or else p3and q3are complex conjugates, in which case we must
have IpI = lql.
Case (i). p3and q3both real (and unequal)—in which case we may take
p and q as both real and the geometrical construction (Figure 23.4) shows
that, of the solutions of (2), one is real and two are complex conjugates.
Case (ii). p3and q3are complex conjugates (and thus p, q may be taken
as complex conjugates). In this case, we must have 11,1 = lqi and the geo-
metrical construction (Figure 23.5) shows that all three roots of (2) are
real.
F'g. 23.4
Since this exhausts all the possibilities, we have arrived at the famous
paradox that, if only one root of (2) is real, then p and q are both real, but,
if all three roots of (2) are real, then they arise from complex values of p
and q (the irreducible case). Put another way, if all three roots of a cubic
equation are real then the working of Cardan's solution necessarily in-
volves complex numbers p and q at the intermediate stages.
Ex. 20. Explain geometrically how (i) a root of multiplicity two, (ii) a root of
multiplicity three, arise.
The expression
A = [(w2 — w3) (ws w2)12
500
3] CUBIC EQUATION
w3+3Hw+G = 0.
From Example 7,
A = — 27(G2 + 4H3).
Consider now the various cases that may arise (remember that we are
considering cubic equations with real coefficients).
(i) One real root a, two conjugate complex roots )6' + jy.
*Ex. 21. Show, conversely, that the sign of A determines the nature of the roots
of the real cubic equation w3+ 3Hw+ G = 0.
The case A > 0 leads to three real solutions via complex values of p
and q: the consequent working in Cardan's solution is frequently rather
5 PPMII 501
POLYNOMIAL EQUATIONS [23
Exercise 23(c)
Solve the cubic equations in Questions 1-5, using a trigonometric substitution
if possible and otherwise employing Cardan's solution.
1. z 3— 2z + 4 = 0.
2. 4z3 5z + 6 = 0.
—
3. z3 — 3z 1 = 0.
—
4. 12z3 — 9z-- 2 = 0.
5. z3 — 9z + 12 = 0.
6. Reduce the equation
z3 — 6z2 + 1 8z — 22 = 0
to the form z3 + 3Hz+ G = and hence solve it completely.
0,
7. Show that the equation
az3+3bz2+3cz+ d = 0
is reduced to the equation
p(w — q)3= q(w — p)3
by the transformation w = az+ b where p, q are the roots (supposed unequal)
of the equation
(ac b2) A24" (a2d 3abc + 2b3) A (ac b2)2 = 0.
— — — —
Miscellaneous Exercise 23
1. Prove, by graphical methods or otherwise, that a cubic equation (with real
coefficients) of the form
ax3+bx2+cx+d = 0
has at least one real root.
Prove that, if A is real, the equation
x3-3x2 +3x+A = 0
is satisfied for one and only one real value of x. (0 & C)
2. By considering two alternative expansions of the expression
In {(1+ax) (1 + fix) (1 + yx)}
(a, y all real), prove that
+fl4 + 74 = p4 A -2q
+4pr + 2q2,
where a, y are the roots of the equation
x3+px2+qx+r = 0.
3. In the equation
x4+ ax3— 20x2+ bx— 576 = 0,
5-2 503
POLYNOMIAL EQUATIONS [23
a and b are to be chosen so that two of the roots are equal and the sum of the
other two roots is zero.
Find all possible values of a and b, supposing them chosen so that the roots
of the given equation are all real. (0 & C)
4. Find the value or values of a for which the roots of the equation
2x3 +6x2 +5x+a = 0
are in arithmetical progression. (0 & C)
5. Given that —2+ jk is a root of the equation
2x4 +8x3 +11x2 +4x+5 = 0,
find k and the other three roots of the equation.
6. Given that a, /3, y are the roots of the equation
x3 — px2+qx— r = 0,
and that sn= an+fln+yn, prove that
si = P, s2 =p2- 2q, s3 = ps2 —qs1+3r.
Calculate s2 and s3for the equation
x3 — 12x2 + 30x-20 = 0.
Show that (s2) 4 and (s3)1are both near to 9 and verify that 9 approximates to a
root of the equation. (0 & C)
7. If co is a root other than 1 of the equation og = 1, prove that the other roots
are (02, idS, (04, (05, (06.
Prove that 1+ CO+ 612+ 613+ (04+ (05+(O6 = 0.
If a = co+ wo, 11 = co2+ ws, y = w3+ 01, prove that the equation with roots
a, ft, y is
z3 +z2 -2z— 1 = 0.
Hence, or otherwise, find the values of
(i) cos tir+ cos fir, + cos lg. (ii) cos PT cos -17r cos tir. (0 & C)
8. When f(x) and g(x) are given polynomials having no common factor, prove
that the values of the constant k for which the equation
f(x)— kg(x) = 0
has a repeated root are given by f(a)/g(a) where a is a root of the equation
f(x) g'(x) — f /(x) g(x) = 0.
Hence, or otherwise, find the values of k for which the equation
x3 — 3x2 + 3kx— 1 = 0
has a repeated root, and find also all the roots of the equation for each case.
(0 & C)
9. Discuss the reality of the solution of the equation
2x3 +3x2 -36x+k = 0
for different real values of k.
504
MISCELLANEOUS EXERCISE 23
10. If f(x) is a polynomial in x which has a factor (x- CO2, prove that (x- a) is a
factor of f'(x). Prove conversely that, if f(x) and f'(x) have a common factor
x- a, then (x- a)2is a factor of f(x).
Solve the equation 4x
4 + 4x3-31x2-66x - 36 = 0
given that it has a repeated root. (0 & C)
11. If a, fl, y are the roots of the equation
x3+px+q = 0,
find the equation with roots ,B/y+71,6% y/oc+ c6/7, ah8+ 16' loc. (0 & C)
12. The sum of the kth powers of the roots of the equation
x4+px2+qx+r = 0
is denoted by sk. Prove that
s2 = -2p, ss = -3q.
sn+4 +PSn+2 qSn÷i rs„ = 0 (II 0),
and deduce that = 2p2 - 4r, s, = 7q(r - p2).
Given that the roots a, b, c, d of the equation are all real and that 5, = 0,
prove that q = 0 and deduce that
(a+ b) (a + c) (a + d) = 0. (0 & C)
13. Show that, if z = 2(- H)i cos 8, where z is a root of the cubic equation
z3 + 3Hz + G = 0, then
cos 30 = ,s/(-G2/4H3).
Deduce that, if H < 0 and G3+ 4H3< 0 the cubic equation has three real
roots.
By using the expression cos 30 = -1-(e
310 + e-30),
solve the equation when G2+ 4H3 > 0,
treating separately the cases
H < 0 and H > 0.
505
24. Vector products and their
applications
*Ex. 4. Show that, if a A b = 0, then either a and b are parallel or one or other
or both of a, b is zero.
*Ex. 5. Show that the vector product is anti commutative; that is, that
-
anb =—b A a.
*Ex. 6. Verify that i A i = j A j = kA k= 0 and that
jAk=—kAj=i, kAi= —iAk=j, iAj=—jAi=k.
Ex. 7. a is a vector of magnitude 2 units pointing due east; b has magnitude 3 units
and points N 60° E. "e' is a vertical unit vector. Find a A b and b A a.
506
1] VECTOR PRODUCT
So far, apart from the result of Ex. 2, the vector product appears
remarkably dissimilar to the product of ordinary numbers. However, there
is one rule which vector products obey and which, in a sense, justifies the
use of the word 'product': it is distributive over vector addition; that is
a A (b + c) = (a A b) + (a A c).
We shall deduce this very important result in a manner reminiscent of
that used to prove the analogous result for scalar products:
a . (b + c) = (a . b) + (a . c).
In Chapter 11, we saw that the scalar product of a = OA with a unit
vector ñ could be interpreted geometrically as the (scalar) projection of a
Fig. 24.2
in the direction fi and it was this fact that enabled us to deduce the dis-
tributive law for scalar products. Consider the plane, II, perpendicular to fi
and through 0; let the foot of the perpendicular from A to II be D; then
the vector OD is called the (vector) projection of a on the plane H (see
Figure 24.2).
It follows from the definition of the vector product that
OD = la A fil,
but OD a A fi,
since OD and a A ft are perpendicular to one another.
As in Chapter 11, Section 1, the vector projection of a on H is unique
and, if the projection of a, and a2on H are OD, and OD„ then the projection
of OB = a, + a, is OC = OD, + OD, (see Figure 24.3 and consider the
mid-point of A,A2).
In words, the (vector) projection, in a plane, of the sum of two vectors is
the sum of the separate projections of the two vectors.
507
VECTOR PRODUCTS [24
Now consider 9 A (a1+ 22): from what has just been demonstrated,
lit A (al + a01 = 101 A aj-F(fi A a2)!.
But 9 A (a1+82) is perpendicular to the plane OBC and thus to OC, 9 A a1
is perpendicular to OD1, 19 A a2 is perpendicular to OD2 and thus, by
rotating the parallelogram 0 D1CD2through a right angle in the plane H,
it follows that
ñ A (al + a2) = fi A al -Ffi A a2.
B
Fig. 24.3
With the distributive rules behind us, it is an easy matter to express the
vector product a A b in terms of components in three mutually perpen-
dicular directions i, j, k. If a = chi+ a2 j+a3k and b = b1i+b2j+b3k, then
a A b = (aii+ a2 j + a3k) A (bl i+b,j+ b,k)
= (a2 b3 — a, b) i + (a3bi — b3) j + (a, b2— a2bi) k,
on repeated application of the distributive rule and using the results proved
in Ex. 6. The result is best remembered written formally as a determinant:
aAb = a1 a2
b/ b2 b,
Since a A = lal Ibl sin 0, we see that the magnitude of the vector
product gives us twice the area of AOAB (see Figure 24.4). The vector
508
1] VECTOR PRODUCT
Ex. 10. OABC is a tetrahedron; prove that the sum of the vector areas of the
triangles OBC, OCA, 0 AB is the vector area of the triangle ABC.
Fig. 24.4
Fig. 24.5
areas of LADB and ABDC (notice that we letter the triangles in the same
sense). Thus
ABCD = -(d A b+b A a+a A d)+-1-(d n c+c Ab+b Ad)
= 2(b A a+a A d+d AC-1-C Ab)
= A (d—b)—c A (d — b)]
= -1-(a—c) A (d b)
—
= A (UC— Ala)
= 1(ic A) (a A c), and the result follows.
—
Ex. 11. The position vectors, relative to some origin 0, of the vertices of a
convex polygon PiP2P3...Pnare denoted by Pi P2 n •••, Pn• Extend the definition
, ,
of the vector area of a triangle to the vector area of a convex polygon and show
that the vector area of P1P2P3...Pn is
i(Pi. A Pz +P2 A Ps + • • • ±Pn--1 A Pn +Pn A PO.
Prove the result of Example 1 by considering the vector areas of ABCD and
OPQ.
we have a1 a2 a3
a.(b A C) = b1 b2 b3
c1 C2 C3
*Ex. 12. Prove that a. (b A c) = b.(c A a) = c.(a n b) (that is, the scalar triple
product is unaffected by cyclic interchange of the letters) and that
a. (b A = (a A b).c
(the . and A may be interchanged).
510
2] PRODUCTS OF THREE VECTORS
Fig. 24.6
*Ex. 18. Prove that the vector triple product is not associative; that is
an(bAc)*(aAb)Ac.
Ex. 19. Prove that a A (b A C) b A (C A a)+c A (a A = 0.
Exercise 24(a)
1. Prove that (a— b) A (a + b) = 2a A b and interpret the result geometrically in
terms of areas.
2. What can you deduce from the equation r A a = r A b?
3. Prove that a +b+c = 0 b A c= c A a= a A b and interpret the result
geometrically.
IsittruethatbAc=cAa=aAba+b+c= 0?
4. Find a A b in the following cases:
(i) a = 3i—j+k, b = 2i+j+k; (ii) a = 5i+4j+3k, b = 3i+4j+5k;
(iii) a = i—k, b = j+k; (iv) a = 21— j — k, b = 2i + j + k;
(v) a = ai+ck, b = bj.
5. If a = 3i +j+ 2k, b = i +2j+ 3k, c = i — 3j-4k, find a.(b A c). What can
you deduce about the points A, B, C?
6. If a = i+j, b = 2i+j, c = i+k, form the vector products a A b and
(a n b) n c and mark the position of the five vectors a, b, c, a A b, (a n b) n c
in a rough sketch.
7. The points A, B, C have position vectors a = 4j + 3k, b = 2j — 3k,
c = 3i-6k. Find the unit vector perpendicular to the plane 0 AB and deduce
the perpendicular distance from C to this plane.
8. With the notation of Question 7, find the area of AOAB and hence the volume
of the tetrahedron OABC.
9. If A, B, C are non-collinear points with position vectors a, b, c, prove that
the vector b A c+c A a +a A b is normal to the plane ABC.
If d is the position vector of the point D and a = 5i+j+k, b = i+j+3k,
c = 31+ 4j — k, d = 8i+ 3j + 2k, find:
(i) the unit normal to the plane ABC;
(ii) the area of the triangle ABC;
(iii) the volume of the tetrahedron ABCD.
512
2] PRODUCTS OF THREE VECTORS
18. ABC is a triangle and points L, M, N are taken on BC, CA, AB respectively
such that
BL _ a CM _ , AN
LC MA = 11' NB = v.
Prove that ALMN = (1 + Apr) AABC and deduce the theorem of Menelaus, that,
if LMN is a straight line, kuv = —1.
513
VECTOR PRODUCTS [24
19. The sides (and sides produced) of the parallelogram ABCD separate the plane
into nine regions (see Figure 24.7). Prove that, if a point 0 is taken within
region I, II, VI or IX the area of OAC is the sum of the areas of OAB and
OAD but that, if 0 is taken within any other region, the area of OAC is the
difference of the areas of OAB and OAD.
(The reader who has studied mechanics may be able to interpret this result in
terms of moments.)
Fig. 24.7
20. ABCD is a tetrahedron. The lengths of AB and CD are a and b and the angle
between AB and CD is 0. A plane is drawn parallel to AB and CD and meets the
edges CA, CB, BD, DA at P, Q, R, S. Prove that PQRS is a parallelogram and
that its area is
ab
— x(h — x) sin 0,
h2
where x is the distance of the plane from AB and h is the distance between AB
and CD.
Hence, or otherwise, prove that the volume of the tetrahedron is
labh sin O.
3. APPLICATIONS TO COORDINATE
GEOMETRY
Throughout this section we shall assume that an origin 0 and a right-
handed set of coordinate axes Ox, Oy, Oz specified by the unit vectors
j, k are given. The position vector of the point A... with respect to 0 will,
as usual, be denoted by a ....
Consider first the line through A in the direction of the unit vector 1
If R is any point on this line, then AR = Ail and thus
AR A = 0,
giving the equation of the line in the form
(r — a) A 01 = 0.
514
3] APPLICATIONS TO COORDINATE GEOMETRY
(r—a) A =0
— ai)i+ (y — a2) j+(z — ajk] A (a rnj + nk) = 0
(y — a2)n— (z — a3) m = (z — a3) 1— (x — ajn = (x— ai)m— (y — a2) 1 = 0
<4-
x— a,= y— a,_z—a3
Fig. 24.8
The distance from the point B to the line, the length BN (see Figure
24.8), is
Kb — a) A 0I.
If B is the point (b1, b2, b3) this gives
x- 1 y-2 z— 3 . 83
1 2 3 141.
d = AB A AC
is perpendicular to this plane. Thus
d = (b — a) A (c — a)
= b A c+c A a+a Ab
and the equation of the plane is now found easily.
515
VECTOR PRODUCTS [24
Example 2. Find the equation of the plane through the points A(1, 1, — 3),
B(2, —1, 10), C(-1, —1, 1).
AB = — 2j+ 13k,
AC = — 2i— 2j+ 4k,
AB A AC = 18i-30j— 6k,
and thus the vector 3i— 5j —k is perpendicular to ABC. Since A lies in the
plane the equation of ABC is therefore
3(x-1)— 5(y-1)— (z+ 3) = 0
which gives, on simplification,
3x— 5y—z = 1.
*Ex. 21. If r = xi+ yj+ zk is the position vector of the point P show that the
volume of the tetrahedron PABC is
(r—a).[(b— a) n (c — a)].
(See Ex. 15.)
Hence obtain the equation of the plane ABC of Example 2 in determinant
form, and verify that the expansion of this determinant gives the same equation
as that obtained above.
Example 3. Find the equation of the plane through the point (2, — 5, — 1)
perpendicular to the line of intersection of the planes
x+7y+z = 0, x-5y-2z = 0.
The line of intersection is perpendicular to both of the vectors
i+ 7j +k and i— 5j-2k
and thus its direction is given by
(i + 7j+k) A(i -5j-2k) = — 9i+ 3j— 12k.
Thus 3i —j+4k is normal to the required plane, whose equation must be
3(x — 2) — (y + 5) + 4(z + 1) = 0
or 3x—y+4z = 7.
Example 4. Find the length of the common perpendicular to the two skew lines
x _ y+1 _ z+2 x-2 _ y-6 _ z-3
1 2 — 4' —1 — 4 — 2
The direction of the common perpendicular is given by
(i+2j+4k) A (—i+4j+2k) =-121-6j+6k.
516
3] APPLICATIONS TO COORDINATE GEOMETRY
= (21+ j — k)/V6.
The lines pass through the points A(0, —1, —2), B(2, 6, 3) respectively and
we have the length of the common perpendicular equal to the projection
of AB in the direction il
= (b a) . 0
—
3x+5y+8z = —1 Pt,
2x+3y+4z = 3 P2,
4x+2y+5z = —2 P3.
These three equations represent three planes, PI, P2, P3; their solution con-
sists in finding the point Q common to each plane (assuming such a point
exists).
From equations P1and P2we obtain (3/31+P2), the equation of the plane
through the line of intersection of P1, P2 containing the origin. Thus
11x+18y+28z = 0 P4
is a plane containing 0 and Q. Similarly, (2P, —133),
2x+8y+11z = 0 P5
is a plane containing 0 and Q.
Thus OQ is perpendicular to both lli+ 18j + 28k and 2i + 8j + 11k; that is
Ex. 22. Use the method outlined above to find a necessary and sufficient con-
dition for the existence of a non-trivial solution to the set of three homogeneous
linear equations:
aix+biy+ciz = 0,
a2 x+b2y+c2z = 0,
a3 x+b3y+c3 z = 0.
Exercise 24(b)
1. Find the equations of the planes through the following sets of points:
(i) (1, -1, 1), (2, 1, 2), (3, 1, -1);
(ii) (1, 1, 1), (0, -3, -2), (-1, 3, - 1);
(iii) (1, 1, 5), (1, -1, -3), (2, 1, 1);
(iv) (4, 0, 1), (7, 3, 5), (-10, -4, -1);
(v) (a, b, c), (3a, b, 0), (4a, 2b, c).
2. Find, in each part, the equation of the plane through the given point and
perpendicular to the line of intersection of the given planes:
(i) (1, -1, 3); 4x-y+4z = 0, 12x- 5y+8z = 0;
(ii) (-5, -4, - 1), 3x+4y+3z = 15, 9x-2y-12z = 3;
(iii) (2, 3, 4), 5x+ y+2z = 7, x+y+z = -4;
(iv) (9, 3, 1), 3x+2y+ 2z = 0, 5x+2y+z = 4;
(v) (1, 1, - 1), 3x+2y- 4z = 0, 4x+ y+3z = 0.
3. Find the length of the common perpendicular to the following pairs of skew
lines:
x _ y-7 _ z-6 x+3= y-6 z+4
._-__ •
(i)
-3 2 2 ' 2 -5 6 '
x-4=y-7 _ z- 3
(ii) x - 6 = y - 3 = z-1,
3 5 2 '
x+1 _ y+1_-z-2
(iii) x- 8 = y-7 = -
2' 2 - -1 2•
4. Find the length of the common perpendicular to the two lines
x-4 =y = z + 3 x-6 y+1 z+5
and =
1 -2 1 3 -3 -1
What deductions can you make from your answer?
5. Find the equation of the plane containing the points A(6, 3, 0), B(-2, 0, -5)
which is parallel to the vector i+j+ k.
6. Find the equation of the plane containing the points A(3, 5, 10), B(4, 3, -3)
which is parallel to the vector 3i + 5j - 6k.
7. Find the equation of the plane parallel to the two lines
1(x-6) =-1-(y-6) = = -(y-2) = z+4
which passes through the point (2, 0, -1).
518
EXERCISE 24
519
2 5. Continuous probability distributions
1. DENSITY FUNCTIONS
In Chapters 7, 10 and 16 we discussed probability distributions in discrete
outcome spaces, that is, in spaces whose elementary events we could count
by putting them into one-to-one correspondence with the set N of natural
numbers. Associated with each elementary event siof the outcome space
S was a number xi, usually an integer—the value of our associated random
variable X.
Random experiments occur, however, in which it is mathematically
more convenient to take as outcome space a set which cannot be counted
and as random variable a subset X of the set R of real numbers. For
example, suppose we take a point P at random on a line AB of length 1 m.
Since we can only measure lengths to within some given limits of accuracy,
say 0.5 mm, we could take as values of a (discrete) random variable the
number of intervals of length 0.5 mm P is from A, giving 201 possible values
in all (0-200). Mathematically the situation may be simpler to analyse if
we take as values of our (continuous) random variable the supposed `exact'
length of AP, measured as a real number of mm; that is
X= fx e R: 0 x 5 100}.
If we are dealing with a subset of R, calculus methods are available and it is,
for example, easier in general to integrate a function than it is to sum a
series.
Given a continuous random variable X, it will generally be impossible
to assign a probability to a specific value of X; rather we assign a prob-
ability to each interval
E = {x e X; x, < x < x2}
as follows. Let f be an integrable function such that
(i) f(x) 0, for all x e X;
(ii) f f(x) dx = 1;
B
50 100
Fig. 25.1
0 if x < 0
kx if 0 < x < 50
f(x) =
k(100 —x) if 50 < x < 100
0 if x?..- 100,
where k is some number. Figure 25.1 shows the graph of y = f(x).
f(x) = 0 if x < 0 or x > 100 simply means that we are restricted to a
choice of P within AB. f(x) = kx if 0 < x < 50 and f(x) = k(100 — x) if
50 < x < 100 ensures that our probability distribution is symmetrical
(reflecting a feeling that we are as likely to take x to be between, say, 35 and
45 as we are to take x to be between 55 and 65) and that the probability of
selecting a point P within an interval of some specified length becomes
progressively greater the nearer the interval is to the centre of AB (see
Figure 25.1 again).
t For consistency we shall take R to be the range of X throughout this chapter.
521
CONTINUOUS PROBABILITY DISTRIBUTIONS [25
30 60 100
Fig. 25.2
Now suppose we wish to find the probability that P lies between 30 and
60 mm from A. We define the event E by
E = {x E R: 30 < x < 60}
and obtain
50 60
Pr (E) = i To- xdx+ f 1„(100 — x) dx
30 50 „
= = 0.54 (see Figure 25.2).
(Again, integration is heavy in the circumstances.)
A different mathematical model, that is, a different choice for f, leads to
a different estimate of the probability of the event E as we shall now show.
We shall suppose a functionfdefined thus:
0 if x < 0
f(x) = kx (100—x) if 0 < x < 100
0 if x > 100
522
1] DENSITY FUNCTIONS
k[50x2 -3x3]000 1
3
k=
500000'
Fig. 25.3
rf(x) dx
523
CONTINUOUS PROBABILITY DISTRIBUTIONS [25
s Lcc,
f(x) dx = 1,
fQ
j _. f(x) dx =
2. DISTRIBUTION FUNCTIONS
The density function, f, of a continuous probability distribution bears the
same relation to probability as mass density does to mass:mass density
is the mass per unit volume, probability density is the probability per unit
interval. It must be remembered that f(x) itself is not a probability, but
8p = f(x) Sx is.
We now introduce a function, closely related to the density function,
whose values do represent probabilities. Suppose the domain of our ran-
dom variable is R and that we are given a density function f: R —> R. Then
we know
(i) f(x) > 0,
+00
(ii) L oo f(x) dx = 1.
Let E be the set of real numbers less than, or equal to x, that is, the set of
real numbers to the left of x on the real axis:
E = fy R: y
Then Pr (E) =
--
f(x)dx rx
depends upon x and we may define a function F: R —> R where
Pr (E) = j'x'f(x) dx
x,
= F(x2) — F(xi).
Since the sets E and {x e R: x < x1} are mutually exclusive, this illustrates
the result that the probability of the union of mutually exclusive events is
the sum of the probabilities of the separate events.
we have
f:kx(1 — x) dx = 1
k = 6.
(ii) If x < 0, F(x) = 0; if x > 1, F(x) = 1.
If 0 < x 1,
F(x) = f: 6x(1 — x) dx
= [3x2 — 2x3]o
= x2(3 —2x).
Thus F is the function F: R --)- {y e R: 0 ( y ..5 1} defined by
0
{if x < 0
F(x) = x2(3 —2x) if 0 ..<, x < 1
1 if x > 1.
525
CONTINUOUS PROBABILITY DISTRIBUTIONS [25
F(x)
(i)
Fig. 25.4
of F, which usually has a shape very similar to that shown in Figure 25.4,
is called a cumulative probability curve (or, occasionally, an ogive curve).
Ex. 6. Explain how the cumulative probability curve may be used to find the
median and upper and lower quartiles.
Example 2. Find the mean of the random variable X whose density function,
f, is defined by
if x < 0
f(x) ={04x(1- x2) if 0 < x < 1
0 if x > 1.
We have
= e(X) = f xf(x) dx
-00
Since
+0° +co
(x -11)2 f(x) dx = (x2-2,ux + #2) f(x) dx
+ CO OP + OP
= co x2f(x) dx-2,u f co xf(x) dx+ It2 f(x) dx
+co
= x2f(x) dx - 2,a . A2.1,
+op
we have a2 +112 = x2f(x) dx,
*Ex. 7. Show that the variance of the random variable in Example 2 is given by
Exercise 25(a)
1. f(x) = kx-2 if x 1,1(x) = 0 if x < 1; find the value of k and hence find
Pr (x < 2).
2. f(x) = ke-2° if x 0, f(x) = 0 if x < 0; find k and hence find
Pr (1 < x < 2).
3. If f(x) = k sin irx if 0 x 1 and is zero otherwise, find k and hence find
Pr (x >
4. Find the distribution function for each of the density functions in Questions
1-3.
Find the mean, median, mode and variance for the distribution with the density
function defined as in Questions 5-8.
5. O if x < 0
f(x) = x) if 0 < x < 2
O if x > 2.
6. if x < 0
(6x +1)/4 if 0 < x < 1
f(x) = {0
0 if x > 1.
7. O if x < — 4
f(x) = {-fir cos rrx if — i ---5 x ---S. 4
0 if x > 4.
8. (0 if x < 0
f(x) =
Ae-Ax if x 0.
9. If the density function, f, of a distribution is given by
0 if x < 0
if 0 < x < 1
.i(x) = 4- if 1 < x < 2
4(3 — x) if 2 < x < 3
0 if x> 3
find the distribution function, F, and sketch the graphs of f and F.
Find the density and distribution functions, g and G, of the new random vari-
able Y, where y = x2, and sketch the graphs of g and G.
10. The density function, f, of a probability distribution is given by
0 if x < 0
x if 0 x <.. 1
f(x) =
2— x if 1 < x < 2
0 if x > 2.
Find the density function, g, of the new random variable, Y, where y =
if x 0, y = 0 if x < 0 and sketch the graphs off and g.
Find the mean and variance of X and the mean of Y.
11. A probability density function of a random variable Xis defined as follows:
x(x— 1) (x-2) for 0 x < 1
f(x) = {A for 1 <x< 3
0 otherwise,
where A is a suitable cons ant. Calculate the expectation is of x. What is the
probability that x is less than or equal to A? (M.E.I.)
529
CONTINUOUS PROBABILITY DISTRIBUTIONS [25
13. The probability density function p(t) of the length of life, t hours, of a certain
component is given by
p(t) = ke-kt (0 < t < oo),
where k is a positive constant. Show that the mean and standard deviation of this
distribution are each equal to 1/k.
Find the probability that the life of a component will be at least to hours.
Given that a particular component is already tlhours old and has not failed,
show that the probability that it will last at least a further tohours is e-kto.
An apparatus contains three components of this type and the failure of one
may be assumed independent of the failure of the others. Find the probability
that:
(i) none will have failed at to hours;
(ii) exactly one will fail in the first tohours, another in the next tohours and a
third after more than 2t0 hours. (M.E.I.)
tribution. f and F are thus given, for the uniform distribution over the
interval a < x ' b, by
1 ) if x < a
1
f(x) = if a.....x:c.b
b a
0 if x > b;
if x < a
x a
F(x) = CO
b_ a if a < x < b
1 if x > b.
Ex. 11. If, in Example 3, the mid-point of the chord is taken to have a uniform
distribution over the interior of the circle, prove that the probability that the
chord will exceed the length of the side of the inscribed equilateral triangle is 1
4.
= 2a tax2— ixs]ga
= 1012.
As x varies from 0 to a, y varies from 0 to its maximum value, a2, and,
532
3] CONTINUOUS DISTRIBUTIONS
0 if x < 0
F(x) = x/2a if 0 < x < 2a
1 if x > 2a
and, by the symmetry of the curve y = x(2a — x) about the line x = a
(see Figure 25.7), we have, for any point (x, y) on the curve,
Pr (z e R: z < y) = Pr (z E R: z < x)+Pr (z E R: z > 2a— x)
= F(x)+[1— F(2a— x)]
x 2a—x
= +1
2a 2a
x
a
Show that Z has a triangular distribution (that is, a distribution with density
function whose graph is triangular).
Like most problems concerned with independent uniformly distributed
random variables, it is best to consider this question graphically. Since
X, Y are distributed independently, the choice of two numbers, x
and y, may be represented by plotting the point R(x, y), all points
of the square OABC (see Figure 25.9) being equally likely. If z = x+y,
Y
A B
N P C
>x
Fig. 25.9
for given z all possible points R lie on the line segment PQ, where P is the
point (z, 0). For 0 S z 1, PQ cc z; for 1 < z 2, PQ cc (2 — z) and
thus the density function, f, of Z, whose value for given z is proportional
to z, is triangular. In fact
0 if z < 0
z if 0 z 1
f(x)
(2 — z) if 1 < z S 2
0 if z > 2.
534
3] CONTINUOUS DISTRIBUTIONS
Ex. 12. Find the distribution function, F, for the random variable Z of Example 5
and sketch its graph.
Ex. 13. If, with the notation of Example 5, the random variable W is defined by
w = x—y, describe the distribution of W.
Fig. 25.10
numerical value of x; that is, the probability is of the form cb(x2) Sx. On the
assumption of the symmetry of the errors made above, the probability of
the 'y error' is 0(y2) Sy and the probability of lying in a small area
SA = Sx Sy
around P is 0(x2) 0(y2) SA, on assuming that the errors occur independently.
But on these assumptions the probability may also be written
95(x' 2) 0(Y/2) 8A, (approx.)
where new axes Ox', Oy' are taken with Ox' passing through P, as shown
in Figure 25.10. Thus, since x'2 = x2 +y2 and y' = 0 at the point P,
q5(x2) 0(.Y2) = 0(x' 2) 0(Y' 2) = 0(x2+ Y 2) 0(0)
and 0 is a function with values satisfying the equation
0(x2) cb(y2) = 1c95(x2+ y2).
6-2 535
CONTINUOUS PROBABILITY DISTRIBUTIONS [25
Fig. 25.11
x 122
—
Y = 5-2
0 5 10 15 20
Fig. 25.16
540
3] CONTINUOUS DISTRIBUTIONS
binomial distribution B(20, 3-) are proportional to the lengths of the cor-
responding vertical lines. The outline shape is strongly reminiscent of the
normal curve. Even in a skew case, where p + 1, the outline is still approxi-
mately normal : see Figure 25.17, in which the length of the vertical lines are
proportional to the probabilities of obtaining the scores 0, 1, 2, ..., 20 in
the binomial distribution B(20, 1).
0
1 5 10 15
Fig. 25.17
kinr)Pr(1—P)n-r3
\
Example 8. A coin is spun 250 times and turns up tails 139 times. Does this
provide any evidence of bias?
Take as the null hypothesis
H: 'the coin is unbiased'.
We have to assess the probability of getting a result as bad as, or worse
than, 139 tails on the assumption that H is true; that is, we have to deter-
mine
Pr(139 or more tails IH) + Pr (139 or more heads I H).
The calculation of this from the binomial distribution B(250, -D would
be prohibitively laborious; we therefore use the normal approximation
to the binomial distribution as our model where
,a = 250 x = 125,
U2 = 125 x i = 62.5,
giving a- = 7.906.
We seek 2 Pr (x > 139),
where x is distributed B(250, 1), and this is approximately
2 [I /138.5-125\1
k 7.906
2[1 — 0(1.71)]
0.087.
There is thus insufficient evidence of bias at the 5 % level.
542
3] CONTINUOUS DISTRIBUTIONS
Fig. 25.18
Consider the calls received at a telephone exchange, the rth call occurring
at time t,., where the time, t, is measured from t = 0. We shall assume that
the number of calls constitutes a purely random process; that is, the number
of calls received in any interval (t, t') is independent of anything that has
occurred previously. Furthermore, we shall assume that the purely random
process is a stationary process: that is, the number of calls received in the
interval (t, t') depends only upon its length, not on its position.
Ex. 15. Discuss the validity of the assumptions made above in the light of what
you imagine would be a typical exchange.
Ex. 16. If p„(t) is the probability that there are n telephone calls in the time inter-
val (0, t) and if St is sufficiently small for there to be a negligible probability
of more than one call in the interval (t, t +St) show that
P.(t+80= (1 — ASO MO+ AP._1(t)St
dp.
and deduce that --d7 = Pa)•
Ex. 17. Prove by induction that the probability, 1,40, defined in Ex. 16, is
given by
e-Al(At)n
P.(t)= •
n!
Ex. 18. Show that the distribution of the number, n, of calls received during a
fixed time interval is Poisson and account for this in terms of the Poisson dis-
tribution being a limiting form of the binomial distribution.
Exercise 25(b)
1. The random variable, X, has uniform distribution in the interval 0 x <10.
Find Pr (X = x, where x2 — 5x + 4 > 0).
2. The line AB has length 10 cm. A point P is taken at random on AB, all points
being equally likely; what is the probability that the area of the circle of radius
AP will exceed 10 cm2 ?
3. The line AB has length 10 cm. An interval of length 2 cm is marked at ran-
dom on the line, the positions of the interval being uniformly distributed. What
is the probability that it will contain the mid-point of AB?
4. A circular disc of radius 10 cm is placed on a table. Another disc, of radius
3 cm, is now placed on the table so that it is at least partially in contact with the
first disc. On the assumption that the permissible positions of the smaller disc
are uniformly distributed, what is the probability that it covers the centre of the
larger disc?
5. A point P is taken at random on the side AB (and between A and B) of the
square ABCD, the positions of P being uniformly distributed. If PC cuts BD at
X, find the probability (i) that BX < -113D; (ii) that BX < 1BD.
6. Figure 25.19 shows a square wooden frame
A B
ABCD with a square hole A'B'C' D' cut sym-
metrically in it. AB = 50 cm, A'B' = 30 cm.
A' B'
A ball of diameter 5 cm is dropped on to the
frame. Assuming that its centre falls within
the square ABCD and that it is equally likely
to meet the plane of the frame at any point
within ABCD, what is the probability that the
ball will pass straight through the hole without D ' C'
If a dose of 17.5 units is administered to each of 100 insects, how many will
be expected to die? What is the probability that just two insects will survive?
(M.E.I.)
18. Packets are advertised as containing 500 g of sugar. Tests carried out show
that 6.7 % of such packets contain a mass greater than 508 g, while 0.6 % have
a mass less than 492 g. Estimate the average mass of the contents of a packet
on the assumption that a normal distribution constitutes an acceptable mathe-
matical model.
19. A firm advertises that their runner bean seeds give a 95 % germination rate.
Of 200 such seeds, 14 fail to germinate. Have you cause for complaint?
20. A man claims that he can forecast rainy or dry weather 48 hours ahead.
Careful records are kept over 80 days and his forecast is found to be correct
47 times. Is his claim justified, or is he merely lucky?
21. A 'chance of failure' distribution is given for time t by the probability den-
sity function
p(t) = (1/a) e-tea (0 < t < co).
Show that a is the mean time of failure and that the variance is a2.
Two components in a machine have failure time distributions corresponding to
means a and 2a respectively. The machine will stop if either component fails and
the failures of the two components are independent. Show that the chance of the
machine continuing to operate for a time a from the start is e-312. (0 & C)
Miscellaneous Exercise 25
1. AB is a rod of length 2a. A point P is taken on AB, the points being uniformly
distributed, and the stick is broken at P. Find the expected value of AP2+PB2.
The two parts of the rod are placed on a table in such a way that A and B
are aV3 apart. What is the probability that the triangle APB will be obtuse?
2. The mean deviation about p for a distribution with density function f is
defined by the equation +co
= lx-F1 f(x) dx.
Prove that the mean deviation of a random variable distributed normally about
with standard deviation a is approximately fisr.
Rods are manufactured with a mean length of 18.4 cm and standard deviation
025 cm. If only those rods with a length greater than 18.4 cm are retained, what
is their average length on the assumption that the rods were originally normally
distributed about 18.4 cm with standard deviation 0.25 cm?
3. If X, Y are independently and uniformly distributed between 0 and 2, find
Pr (xy < 1).
4. The side of a square is uniformly distributed between 0 and 1. Find the den-
sity and distribution functions for the area of the square and sketch their graphs.
Determine the expected value of the area and its variance.
547
CONTINUOUS PROBABILITY DISTRIBUTIONS [25
5. The probability density function for the life x of a motor car tyre is given by
f(x) = Ae-As
for x 0, where A = 0.04 and x is measured in units of 1500 km. What is the
probability that a single tyre will last more than 30 units?
A car has four tyres in use and they are all of the same age. What is the prob-
ability that all of them will need replacing before 30 000 km? (M.E.I.)
6. A random variable X has the probability density function
exp -gx1(cr,13)}
f(x) -
6,0{1+ exp [ - gx/(c•N/3)]}2
for all real x. Show that the distribution is symmetrical about x = 0. Determine
the cumulative distribution F(x) = Pr {X < x}.
You are given that the variance of X is 0-2. Compare the value of F(o), F(2cr),
F(3c•) with the corresponding values for the normal distribution with zero mean
and variance Gr2. (M.E.I.)
7. A random variable X is distributed normally with expectation ,u and standard
deviation cr. Find the mode and median of the distribution.
What is the proportion of the population lying between the points of inflexion
of the curve of the probability density function? (M.E.I.)
8. The diameters of some machined components are distributed normally with
mean 5.00 cm and standard deviation 0.05 cm. Find the expected proportion
of the components which will be outside the range 4.925 to 5.075 cm and the ratio
of the expected proportion in the range 5.025 < d < 5.050 to the expected pro-
portion in the range 5.050 < d < 5-075.
It is desired to adjust the mean of the process so that there are, on average,
twice as many components in the range 5.025-5-050 cm as in the range 5.050-
5.075 cm. Show that this can be done if the mean is adjusted to a value between
5.00 cm and 4.95 cm, and find the value by trial to the nearest 0.01 cm.
(M.E.I.)
9. The probability density function of a distribution is given by
e—xxA-1
f(x) = (x 0, A an integer > 0).
(A - 1)!
Find the expectation and variance of X. If AX) = 4a find 6'{(X-,u)3}.
,
F(Y) = () f e-x'/2clx.
o
548
MISCELLANEOUS EXERCISE 25
Deduce that Y has a X2distribution with one degree of freedom, defined by the
density function f, where
0 (y < 0)
f(y) 1 v 112 p-v/2 0).
,1(27r)' .-
12. Two chords are drawn independently at random in a circle. What is the
probability that they intersect?
13. Two men, A and B, are allowed a lunch-break of an hour. A is at liberty
to leave the office at any time between 12.00 and 12.45, while B may leave the
office between 12.45 and 2.00. What is the probability that they will both be out
together, on the assumption that the permissible starting times for both A and
B are uniformly distributed?
14. A needle is pivoted at the point (0, a) and is rotated. When it comes to rest,
the point P at which its axis (produced if necessary) cuts the x axis is marked.
The random variable X is defined to have as its value the x coordinate of P.
On the assumption that the angle that the needle makes with the y axis is
distributed uniformly between ± -17r, prove that X has a Cauchy distribution,
defined by the density function
a
f(x) — ( co < x < co).
71(22 +X2)
By considering the forms that these equations take under the steady-state
condition pk(t) = pk(t at), where k is zero or any positive integer, or by any other
-
means, obtain and the mean queue size (in terms if A and p) in the steady
state. (C.S.)
19. X is a continuous random variable with mean p and variance o2. A is a posi-
tive constant. Prove Chebyshev's inequality
Pr (I X- pl > Acr) < 1/A2.
Show how this inequality may be sharpened, in the case A = 2, if the dis-
tribution of X is assumed to be (i) normal, (ii) uniform.
20. Spacecraft land on a spherical planet of centre 0. Each is able to transmit
messages to, and receive messages from, any spacecraft on the half of the surface
of the planet nearest to it.
(i) It is known that spacecraft have landed at points A and B on the surface
of the planet. Show that the probability that a space craft, landing at random on
the planet, will be able to communicate directly with the spacecraft at A and B is
(it - 0)12rt
where 0 is the angle AOB.
(ii) What is the probability that three spacecraft, all landing at random on the
planet, will be in direct contact with each other? (C.S.)
21. Engine crankshafts are manufactured so that the diameters, in centimetres,
form a normal distribution with mean 5 and standard deviation 0.03. Crank-
shafts with diameters less than 4.94 or greater than 5.06 are rejected. The ac-
cepted product is classified into three grades of size 4.940-4.988, 4.988-5.012,
5.012-5.060.
Show that:
(i) f xe-°" dx = -
Find the mean and standard deviation of the population and the probability
that a randomly chosen member of the population is greater than 2. (M.E.I.)
24. Initially a machine is in good running order but is subsequently liable to
breakdown. As soon as a breakdown occurs repairs begin. If the machine is in
good order at time t then the probability that a breakdown occurs in a small
interval (t, t+ d t) is a dt, and if it is under repair at time t the probability that the
repair is completed in time (t, t+ d t) is fldt. Let p(t) be the probability that the
machine is under repair at time t.
Write down an equation relating p(t + dt) to p(t) and hence show that p(t) is
551
26. Numerical solution of equations
1. ACCURACY
When equations arise in practical problems, their solutions are generally
required only to some given order of accuracy. General algebraic methods
of solution (for example, for the cubic equation), even if they are available,
are often less suitable than approximate methods for deriving numerical
solutions. Thus, to find the real root of the equation
x3 — 2.7x— 5.3 = 0
by Cardan's method would require burdensome calculations and no such
method exists to solve, for example, the equation
x5 — 2.7x— 5.3 = 0.
Because of its great practical importance, the estimation of roots of
numerical equations has been extensively studied; in this chapter we shall
confine ourselves to some of the most elementary methods available: the
reader who wishes to study the subject further should consult one of the
standard texts on numerical methods (for example, Henrici, Elements of
Numerical Analysis). It may be mentioned in passing that this subject allows
more scope than most others in elementary mathematics for students to
devise their own methods; although these may often prove less efficient
than the standard techniques, much profit will be gained by developing
them as far as possible—apart from the obvious enjoyment of producing
something original.
Before considering the solution of equations it is worth pausing to con-
sider some problems connected with the accuracy of calculations. In
numerical calculations, errors occur through rounding-off to a given number
of significant figures. If a number is to be rounded-off to N digits and the
discarded digits form exactly half a unit, round-off to the nearest even digit.
For example, to 3 significant figures,
3.864 3.86; 21990 220 x 104;
0.03765 3.76 x 10-2; 21.55 216 x 10.
*Ex. 1. Suggest any advantage you see in adopting the convention above. Can
you suggest any reason why even rather than odd digits have been selected?
552
1] ACCURACY
Once numerical data have been rounded-off, their subsequent use in
calculations introduces further errors which will be cumulative. It would
be pleasant to be able to calculate exactly the error in any numerical answer.
This is usually impossible, but we may be able to produce a positive
`acceptance' for the answer; that is, a quantity which we are sure is more
than the absolute error in the answer. For example, suppose we round-off
a given angle 0, measured in degrees, to 3 significant figures to give 37.2°.
The actual angle lies between 37.2°-0.05° and 37.2° + 0.05°. From four-
figure tables, tan 37.2° = 0.7590 and the difference for ±0.05° is
4(0.7618 - 0.7563) 0.0014.
Thus tan 0 = 0.7590 + 0.0014;
in other words, tan 0 is 0.7590 with an acceptance 0-0014.
*Ex. 2. If possible inaccuracies in the tangent tables are taken into account, show
that tan 0 = 0.7590 ± 0.0015.
Now suppose y is a rational approximation to the number x; thus
x = y + e,
where e is the error involved in writing y for x. The absolute error is
I el while the relative error is I elly (more correctly 161/x, but the two forms
are nearly equal for small relative errors and the form Ici ly has the advan-
tage of possessing a known denominator).
Usually we do not know the exact error, only an upper limit to the
absolute error—the maximum absolute error. For example, if a number is
rounded to 4 decimal places, the maximum absolute error is 5 x 10-5.
Similarly, we have a maximum relative error. In a calculation in which we
know the maximum absolute errors in the given numbers it is important
to be able to estimate the maximum absolute error in the answer.
Suppose that x1is written as yiwith maximum absolute error I el l
(we use modulus signs to emphasize that maximum absolute errors are
positive) and that x2 is written as y2with maximum absolute error 1 621 .
Then the maximum absolute error in writing y, +y2for x„ + x2 is 1611 + 16 21,
while the maximum relative error is
(I ei +1621)/0).+ YD.
Ex. 3. Show that the maximum absolute and relative errors in writing y1-y2 for
x1- x2are respectively I 611 + 1 621 and (I elI + 1 621)01-Y2).
*Ex. 4. Neglecting terms such as 161 621, show that the maximum absolute and
relative errors in writing y1y2 for x1x2are respectively
IY1 621 + IY2 611 and I ei/Y11 + 162/Y2
Show also that, writing yily, for x1/x2they are respectively
I ei/Y2 I + 62/Yil and I el/Yi I + 162/Y21.
553
NUMERICAL SOLUTION OF EQUATIONS [26
example, if the correct answer is 0.649 999 9 and our numerical answer is
0-650 000 1 then, to 1 significant figure our answer is 0.7 but the correct
answer is 0.6 although our calculation was only 0.000 000 2 in error.
Fig. 26.1
*Ex. 10. Show from graphical considerations how the computation of 412 /0
enables one to tell whether the approximation by linear interpolation is likely to
be too big or too small.
By direct computation, f(1) = —1, f(2) = 8 and the real root lies be-
tween x = 1 and x = 2, probably nearer to x = 1.
By linear interpolation, a second approximation is given by
x =1
Since f'(x) > 0, .f"(x) > 0 between x = 1 and x = 2 (see Figure 26.2)
the root x = 19 will be an underestimate.
Fig. 26.2
Ex. 11. Prove that, if all of the coefficients are positive, then the equation
ao x" F ai xn-1+ ... + an = 0
can have no positive roots.
Prove further that, if the coefficients of all odd powers of x are zero and the
remaining coefficients are positive, then the equation can have no real roots.
557
NUMERICAL SOLUTION OF EQUATIONS [26
has three sign changes and thus either one or three positive roots. Again,
the equation
xs - 6x6 - 4x5 + 3x4 - x2 + 2 = 0
+ - - + - +
T
has four sign changes and thus has zero, two or four positive roots.
By writing -x for x one may similarly find an upper limit to the number
of negative roots.
Ex. 13. Prove that P'(x) has k sign changes if a,„1> 0 and (k- 1) sign changes
if < 0.
Ex. 14. Prove graphically that, if P'(0) > 0, no roots of the equation P(x) = 0
lie between x = 0 and the least positive root of P'(x) = 0 and that if P'(0) < 0,
at most one such root of P(x) = 0 exists.
Ex. 15. What happens in Exs. 13, 14 if = 0?
Ex. 16. Complete the proof of Descartes's Rule of Signs using mathematical
induction.
558
2] LOCATING ROOTS
Exercise 26(a)
1. If x is rounded-off to 3 significant figures to give the numerical value 48.7,
give the value and range of acceptance for (i) sin x°; (ii) tan x°, using 4-figure tables.
2. If x, y are rounded-off to 3 significant figures, their values are x = 17.2,
y = 5.16; give the range of acceptance for (i) x+y; (ii) xy; (iii)
3. Compare the relative accuracy obtained from your square root tables for
x = V3 - V2 with that obtained for x = (V3 + V2)-1.
If V2 = 1.41421... and J3 = F73205... find as accurately as you can a value for
V3 - V2.
4. Given that V130 = 11.4018 and V132 = 11.4891 find as accurately as you
can a value for V132 - V130.
5. The triangle ABC is right angled at B. AC, AB are measured to the nearest
millimetre, their lengths being found to be 7.4 and 4.4 cm respectively. Use the
theorem of Pythagoras to calculate the length of BC, stating what confidence
you have in the reliability of your answer.
6. If x = y+ e, use the Taylor expansion to find the absolute and relative error
in taking tan x to be equal to tan y.
7. With the notation of Question 6, find the absolute and relative error in
taking (1 +x2)-1/2 to be (1 +y2)-112.
8. If x = 4-6 ± 0.05 and f(x) = x3- 2x +1, and if x is taken to be 4.6, find the
relative errors in (i) x; (ii) f(x). Comment upon your result.
9. Find graphically, using the method of enlarging scales, the least positive root
of the equation
cot 7-1 = 1 + x
(2
correct to 2 decimal places.
10. Solve graphically the equation 2x = 1 + In x.
11. Solve graphically x3 = 10.
12. Find 21/6, correct to 2 decimal places, by a graphical method.
13. Locate the three roots of the equation
2x3 - 6x- 3 = 0
and find their approximate values, using linear interpolation.
14. Locate the two real roots of the equation
x4 +x2 + 10x- 24 = 0
and find their approximate values, using linear interpolation.
15. Use Descartes's Rule of Signs to prove that the equations
(i) x6 - 2x3 - 3 = 0; (ii) x4 -x3 + 5x2+2 = 0; (iii) x6 + x4 - 4x3 + 5x + 2 = 0
each have at least two complex roots.
559
NUMERICAL SOLUTION OF EQUATIONS [26
Fig. 26.3
*Ex. 18. Illustrate graphically the iterative process of Ex. 17 by sketching the
graphs
y= x and y= 5+51x.
Ex. 19. Explain geometrically why convergence is more rapid for finding the
positive root from xr.+.1 = 5+ 51x, than for finding the negative root from
x„+1 = ix,2—1.
Ex. 20. Explain why, in both the iterative processes considered so far, the suc-
cessive approximations oscillate from side to side of the exact root.
= Acc).
But xr+1 = f(x,),
giving
cc — er+1 = f(c6— er)
f(x)— e„ f '(a), on using Taylor's expansion.
Thus er-F1 erAcc).
Successive applications of this result give
r+1;
67+1 err(Ix) er-i{ '(42 "'
" GUM
thus, provided If'(a)! < 1, < I eol
562
3] ITERATIVE PROCESSES
Ex. 21. Explain graphically what happens to the iterative process if f'(a) = 1.
*Ex. 22. Illustrate the significance of the condition If(a)1 < 1 for the con-
vergence of the iterative process
= f(x,)
by drawing four graphs with:
(0 no) < 1, f"(a) < 0; (ii) f'(a) < 1, f"(a) > 0;
(iii) f'(a) > 1, f"(a) < 0; (iv) f ' (a) > 1, f"(a) > 0.
As we have already seen, the iterative procedure outlined above con-
verges to the required root rather slowly. A more powerful iterative pro-
cedure is provided by the Newton — Raphson process. Consider the equation
f(x) = 0
and suppose, as before, that a is an exact root, with x0the first approxima-
tion and error eo; that is
cc = xo + 60.
We then have f(a) = 0,
© f(x0+ co) = 0,
and thus, on using a Taylor expansion and regarding e0as being sufficiently
small for us to be able to ignore e2, and higher powers,
f(x0)+ eo f'(x0) 0.
This yields an expression for the error term
f(X0)
, (p
ovided
r f'(x0) 0),
° f (x0 ``"
and we may take as our next approximation
xi = x0 fixo)
fuo) .
and the curve y = .f(x) is seen to have two stationary points, a maximum
at x A/1.7, and a minimum at x +V1.7. But — A/1.7 —1.3 and
f( 1.3) < 0 therefore the curve cuts the x axis just once and the equation
—
thus has only one real root (which is clearly positive). (See Figure 26.4.)
Fg. 26.4
The next step in the solution is to locate the root: to do this we substitute
integral values for x until we discover a sign change. In order to write
down the differences Of, A2f, ... we tabulate our working, writing the values
off in a vertical line:
x f of 6,2f oaf
0 —8
—4
1 —12 6
2 6
2 —10 12
14
3 4
The root we seek lies between 2 and 3; using linear interpolation we
take as our first approximation
xo = 2+14 x 1 ^ 2-7.
Since 4/(2) > 0, the curve y = f(x) is increasing between x = 2 and
x = 3 and since i2f(2) > 0, it is increasing at an increasing rate: we
564
3] ITERATIVE PROCESSES
deduce that 2.7 is an underestimate of the exact root (see Figure 26.5,
which has not been drawn to scale).
Since f'(x) = 3x2– 5,
f(2.7)
xi = 2 7
f (27)
+1.817
1
= 27
16.87
2.808.
Fig. 26.5
The process is now repeated, taking, x1= 2.808 as our second approxi-
mation. (The process is much facilitated by the use of a hand-calculating
machine: if one is available, recall the method of nested-multiplication for
the evaluation of polynomials—see Chapter 18.)
f(2.808)
x, = 2'808–
0.1
= 2.808
18.65
= 2.802.
Since we may regard this as a more accurate approximation to the re-
quired root than 2.808, we take
x = 2.80
as the root to 2 decimal places.
Ex. 23. Use a calculating machine to obtain the root of the equation
x3 — 5x — 8 = 0
correct to 4 decimal places.
Fig. 26.6
Ex. 24. Show by graphical considerations, that, once the process starts to con-
verge to a root, it always does so from one side. Try to produce an analytical
argument to support this assertion.
Ex. 25. Explain by a graphical argument how inaccuracies may arise in the
neighbourhood of two nearly equal roots.
xr+i ,
.,.r
f(xr)
—
f'(x0)'
where the variable denominator I (xr) is replaced by the constant f (x0).
The convergence is less rapid but nevertheless fairly good.
er+1 = (167,.
Similarly, if xr+i = 3a4+a24,
1
xr +er =-
a
, 3 1 3xr 34 3
a —xr
—
Exercise 26(b)
Find the real roots of the equations 1 8, using any suitable iterative process. Use
-
linear interpolation to find the first approximation and give your final answer to 3
significant figures.
1. x3-3 = 0. 2. x3 — 100 ----- O.
3. x6 -5 = 0. 4. x2— 3x — 11 = 0.
5. x3 6x2 + 10x 9 = 0 (1 root).
— — 6. x3 3x2 — 3x-7 = 0 (1 root).
—
Miscellaneous Exercise 26
1. Find the greatest root of the equation
x3 — 3x + 1 = 0,
correct to 3 decimal places.
2. The roots of the quadratic equation ax2+ bx — 1 = 0 are calculated from the
recurrence relation
1
=
xr+1 ax,.+ b•
Interpret this process geometrically, and prove that
er+i = — a X2e,. ,
where X is the exact value of the root being calculated and X = x,.+ 6,..
With the help of reciprocal tables, use this method to solve the equation
5x2 — 3x— 1 = 0.
3. Give a sketch showing the general shape of the graph of y = sec x for values of
x from x = 0 to x = VT.
Deduce from the graph that large roots of the equation
x cos x = 2
are approximately equal to (n+ i) 7r, where n is a large integer; and prove that
closer approximations are given by
x = (n+}) 7T±
— (2n + 1) IT'
where the positive sign is taken when n is odd, and the negative sign when n is
even. (0 & C)
4. A root of the equation
sin3 ix°+cosx°- 400 = 0
is close to 60. Find the value of the root, correct to 0.1. (0 & C)
5. Draw an accurate graph of y = sin x between x = 0 and x = 7r. Draw in
the same diagram the lines y = mx for m = s. f, s.f, 1. Determine the values
of x where these lines cut the graph of y = ix sin x, giving your answers in the
form kg, where k is correct to 2 decimal places.
570
MISCELLANEOUS EXERCISE 26
Use the values obtained to draw a separate graph of y = -Pr sin x/x between
,
A 1
and that
2K2
Find an approximation to the root correct to the term of order K-4. (0 & C)
14. The equation x2 -ax + b = 0 has roots a, /3. Form the equation with roots
fi2 and, if the equation with roots ae, /62' is x2 -a„x+b„ = 0, write down
the equation with roots a2"-", 162'1.
If a = A.16 where IA I < 1, show how the formation of a sequence of such
equations may be used to give successive approximations to ,6 and hence a.
Under what circumstances does this method provide a very satisfactory
method for solving quadratic equations? Discuss the magnitude of the errors
involved and consider the cases of (i) equal, (ii) complex, roots.
Discuss the application of this method of solution to equations of degree
higher than the second.
15. Prove that the equation ex = Ax has a real solution if and only if A e,
but that x ex = # has a real solution for all real values of it.
Solve the equation xex = 1 to 5 decimal places.
16. xr, xr-F2 are successive approximations to a root of the equation f(x) = 0
obtained by the Newton-Raphson process. Writing Cr for the correction
-f(x0if '(x,), so that Cr+1-f(xr+ Cr)If'(xr+ Cr) show that
17. It is required to find the nth root of the number a by the Newton-Raphson
process. Show that this can be done by solving the equation xn+k- axk = 0.
Using the analysis of Question 16, show that most rapid convergence is
obtained by choosing k = -1(1- n). Use this method to obtain a value for .;/2,
carrying out two iterations.
572
27. The ellipse and hyperbola
the value of e. The reader will no doubt already have noticed that, with
e = 1, the conic defined is a parabola. More generally, we define the follow-
ing types of conic:
(i) if 0 < e < 1, the conic is called an ellipse;
(ii) if e = 1, the conic is called a parabola;
(iii) if e > 1 the conic is called a hyperbola.
The shape of the three types of conic are shown in Figures 27.2 (ellipse),
27.3 (parabola) and 27.4 (hyperbola). (The dotted lines in Figure 27.4 are
573
ELLIPSE AND HYPERBOLA [27
not part of the hyperbola: they are, in fact, asymptotes and are included
as an aid towards drawing the curve.)
Ex. 1. Explain why the set of points
{(x, y): x2= 4[(x — 1)2 + (y — 1)2])
represents an ellipse.
2. THE ELLIPSE
The ellipse has been defined in Section 1 as a conic with eccentricity e < 1.
Before attempting to obtain the Cartesian equation of an ellipse, it is
worthwhile to consider a little of the geometry of the curve, in order that
we may be able to choose the most suitable coordinate axes. Suppose that
S is the given focus and 1 the corresponding directrix and let K be the foot
of the perpendicular from S on to 1.
If e is the given eccentricity, then there are two points A, A' on the line
SK which belong to the ellipse—namely the points dividing SK internally
and externally in the ratio e: 1 (see Figure 27.5 and notice that A' lies on
KS produced, since e < 1). Let 0 be the mid-point of AA' and set OS = s,
OK = k. Then, if AA' = 2a,
SA = a— s, A'S = a+ s,
AK = k—a, A'K = k+a,
and thus, by the definition of A, A' as points of the ellipse,
a — s = e(k — a)
a+s = e(k+a).
Solving these two equations we have
s = ae, k = ale.
574
2] ELLIPSE
Let us now take 0 as the origin, OK as the x axis and the perpendicular
to this line through 0 as the y axis. Then, by what we have just shown, S
is the point (ae, 0) and 1 is the line x — ale = 0. Let P(x, y) be any point of
the ellipse; by the definition of the ellipse
sp2 = e2pm25
K
A ' 0 S A
Fig. 27.5
Fig. 27.6
AA', BB' are respectively the major and minor axes of the ellipse; a is
the length of the major semi-axis, b is the length of the minor semi-axis
(b < a, hence the word ' minor '). From the original definition of b, we have
the important equation connecting a and b:
b2 =a2(1 — e2) .
Ex. 5. If a circle is regarded as a special case of an ellipse with equal major and
minor axes, show that it corresponds to a conic with zero eccentricity.
Ex. 6. A chord drawn through a focus perpendicular to the major axis of an
ellipse is called a latus rectum of the ellipse. Show that the length of the latus
rectum is 2b2/a.
Ex. 7. Show that the eccentricity of the ellipse
x2 y 2
fs 6 = 1
is sand deduce the coordinates of its foci.
Ex. 8. Show, by translating the coordinate axes, that the equation
(x 1)2 (y + 2)2
=1
3 2
represents an ellipse with centre (1, — 2). Find the coordinate of its foci and show
the eccentricity to be I.A/3.
The ellipse arises in nature as the typical orbit of objects moving under
the action of an attractive force varying inversely as the square of their
distance from a fixed point. For example, planets describe ellipses (with
slight variations and discrepancies due to the attractions exerted on each
other) with the sun at one focus—hence the use of the letter S for the focus.
Similarly, for the motion of satellites around the Earth, the centre of the
Earth being situated at a focus (again with a slight discrepancy, this time
576
2] ELLIPSE
because the Earth is neither exactly spherical, nor uniform). The aesthetic-
ally pleasing shape of the ellipse has long been admired: it can be seen, for
example, in the elliptic arches of some bridges which with their reflection
in still water yield complete ellipses.
Ex. 9. The eccentricity of the Earth's orbit is approximately -617, while that of the
planet Pluto is 1. Compare the shapes of their orbits.
Ex. 10. An astronomical unit (a.u.) is the mean distance of the Earth from the
Sun. (It is approximately 500 light-seconds.) At its nearest approach to the Sun,
the planet Mercury is distant about 0.308 a.u. from the Sun, while its maximum
distance is about 0.466 a.u. Calculate the approximate eccentricity of the orbit.
*Ex. 11. Show that x = a cos 0, y = b sin 0 gives a parametric representation for
points of the ellipse b2x2 a2y2 = a 2b2.
A rod has three points P, Q, R marked on it, where PQ = a, QR = b. If
P, Q are constrained to move along two fixed perpendicular lines, show that R
moves along the arc of an ellipse. (This is the engineer's paper trammel method
for drawing ellipses.)
*Ex. 12. Draw in the same diagram the ellipse
b2x2 + a2y2 = a2b2
577
ELLIPSE AND HYPERBOLA [27
It follows from the preceding remarks and the result of Ex. 14 that
properties of parallel chords of circles, mid-points of chords of circles, etc.
are preserved by T and thus correspond to identical properties of the
ellipse. For example, since the mid-points of parallel chords of a circle lie
on a straight line through the centre, the same is true for an ellipse. The
Fig. 27.8
line obtained is called a diameter of the ellipse. Furthermore, the two lines
parallel to the given chords passing through the intersections P, Q of the
corresponding diameter with the circle are tangents: the same is true of the
ellipse (see Figure 27.7).
Again, in the circle, the centres of chords parallel to PQ define another
diameter, RS, such that all chords parallel to RS are bisected by PQ. The
same property is thus true of the ellipse: P'Q' and R'S' are called conju-
gate diameters; they have the property that each bisects all chords parallel
to the other (see Figure 27.8).
The geometrical interpretation of T is as follows. Given a point P, draw
578
2] ELLIPSE
Fig. 27.9
*Ex. 15. Show that T is a non-singular linear transformation, that is, that T-1
exists.
dy_
_ _ b2x (y * 0).
then
dx a2y
Deduce that the equation of the tangent at the point (x1, yi) on the ellipse is
x— x1 = y — y, = a
a2y,
— b2x1
and show that this may be rewritten in the form
xxi yy .
— - +-- =1.
a2 b2i
*Ex. 17. Show that the equation of the tangent to the ellipse at the point
(a cos 0, b sin 0) is
x cos 0 y sin 0.1,
+
a b
and that the equation of the normal is
ax sin 0 by cos 0 = (a2 b2) cos 5 sin 0.
— —
580
2] ELLIPSE
The ellipse, like the parabola, has a rich geometry. We shall now prove
two of its most famous properties. (Since both are focal properties, neither
is derivable from the circle.)
(i) Focal distance property
For any point P on the ellipse with foci S, S' and major axis 2a,
SP+S'P = 2a.
Take the ellipse in the standard form
b2x2 +a2y2 = a2b2
and let M, M' be the feet of the perpendiculars from P on to the two
directrices (see Figure 27.12).
Fig. 27.12
*Ex. 19. Prove that SQ+S'Q < 2a for all points Q within the ellipse and that
SR+ S'R > 2a for all points R outside the ellipse.
*Ex. 20. Prove a converse of the above result, namely that if S, S' are fixed
points and P is any point such that SP+ S'P is constant and greater than SS',
then the locus of P is an ellipse with S, S' as foci. (If the constant distance is
taken as 2a, show that a unique ellipse exists with S, S' as foci and major axis
of length 2a; then use Ex. 19.)
*Ex. 21. If a, b are complex numbers such that la— bl < c, where c is a real
positive constant, describe the set of points
{z C: lz — al + lz —1)1 c}
in the Argand diagram.
Ex. 22. Explain the theory underlying the following well-known mechanical
construction for an ellipse. Two drawing pins are stuck in a sheet of paper and a
loop of cotton is placed loosely around the two pins. The loop is made taut by
the point of a pencil which is then made to trace out a curve on the paper,
keeping the loop taut at all stages of the construction.
(ii) Reflection property
The tangent and normal at any point P of an ellipse bisect the angle SPS'.
Take the ellipse in standard form and let P be the point
(a cos 0, b sin 0).
With the notation of Figure 27.13, the equation of the normal, PG, is
(see Ex. 17)
ax sin 0 — by cos 0 = (a2 b2) cos 0 sin 0,
Fig. 27.13
Ex. 23. ABC is any triangle and P is any point on the external bisector of the
angle BAC. Prove that
BP+ CP ...-- BA+ CA,
with equality only if P coincides with A.
Example 2. Prove that the chord of contact of tangents from the point
C(h, k) to the ellipse 2
Y2
+— 1
a b2
ky ,
is the line a2 _i__ — I .
b2
Deduce that, if the tangents at the extremities of a variable chord through
the fixed point D(a, ,8) meet at R, then R lies on the straight line
ax fiy
712 4- --bi = 1.
From Ex. 16, we know that the equation of the tangent to the ellipse
at the point P1(x1, y1) is
xxi yyl i
a2 ± b2 ="
Now suppose that P1(x1, y,), P2(x2, y2) are the points of contact of tangents
from C(h, k) to the ellipse. Then the coordinates (h, k) satisfy the equations
xxi yy, ., xx2 yy2 „
, 2 = ' and
a2 ± b2 a2+ b2 i —
for all positions of the point (h, k). Thus (h, k) lies on the line
fiy ,
+ =
a2 b2
Fig. 27.14
Then the normal at P has equation ax sin 0 — by cos 0 = (a2 b2) cos 0 sin 0
(see Ex. 17). Thus we may write down the coordinates of the points
H (y = 0) and K (x = 0) and hence of Q (see Figure 27.14):
(a2 b2
H: cos 0, 0) ,
a
a2 b2
K: (0 sin 0),
' b
/a2 b2 a22 b2
cos 0 sin 0).
a '
584
2] ELLIPSE
Exercise 27(a)
1. Find the eccentricities of the following ellipses:
x2 v2 X2
— = 1
1; (ii) — +y2 = 1 (iii) 3x2 +5y2 = 1.
8 6 4
Find its eccentricity, the coordinates of its foci and the equations of the cor-
responding directrices.
14. The tangents at the points P, Q of an ellipse, centre 0, meet at the point
T. Prove that OT bisects PQ.
15. The parallel chords P1 Q1, P2 Q2of an ellipse are bisected by the diameter UV.
Prove that P1Q2 and P2 Q1 meet on UV, as also do P1P2and Q1 Q2.
16. UV is a diameter of an ellipse and P is any point on the ellipse. Prove that
the diameters parallel to PU, PV are conjugate.
17. With the notation of Question 16, the tangent at U meets PV at T and the
tangent at P meets TU at M. Prove that M is the mid-point of TU.
18. The diameter UV of an ellipse bisects the chord PQ. UP and VQ meet at
X, UQ and VP meet at Y; prove that XY and PQ are parallel.
19. What is the locus of the mid-points of chords of an ellipse which pass through
a common point?
20. UV is a diameter of an ellipse and the tangents at U, V are u, v respectively;
the tangent at any point P meets u, v at X and Y. Prove that X and Y lie on con-
jugate diameters of the ellipse.
21. Prove that tangents at the extremities of a focal chord of an ellipse meet on
the corresponding directrix.
22. The tangent to the ellipse x2 y2
— + -= 1
a2 b 2
at the point P meets the axes at Q and R. Find the locus of the mid-point of QR.
23. The perpendicular from the centre of an ellipse with focus S to the tangent
at a point P meets SP produced at Q. Prove that the locus of Q is a circle, and
find its centre and radius.
24. Define geometrically the eccentric angle 0 of a point P on the ellipse
x2 y2
a2 ± -
b2 = 1 (a > b)
and P(a cos 0, b sin 95) is a point on the curve. Calculate the lengths PS, PS' in
terms of a, e and 95, where e is the eccentricity of the ellipse and verify that
PS +PS' is constant.
The tangents to the ellipse at P cuts the x axis at T, and the normal at P cuts
the x axis at N. Prove that
(i) OT. ON = a2e2; (ii) PT/PN = tan 0/(1 e2)
—
3. THE HYPERBOLA
A hyperbola is a conic with eccentricity e > 1. The analysis of the hyper-
bola follows closely that for the ellipse and there are many striking
similarities between the geometry of the two curves. To obtain the standard
form of the equation of the hyperbola we first derive certain geometrical
588
3] HYPERBOLA
results, just as we did for the ellipse in Section 2. Suppose that S is the
given focus and 1 the corresponding directrix and let K be the foot of the
perpendicular from S on to 1. There are two points, A, A', on the line
SK which belong to the hyperbola—namely, the points dividing SK
internally and externally in the ratio e: 1. Since, for the hyperbola, e > 1,
1 will lie between A and A' (see Figure 27.15, and compare with Figure
27.5). Again we take 0 as the mid-point of AA' and set OS = s, OK = k.
Then, if AA' = 2a,
SA = s — a, A'S = s+ a,
KA = a—k, A' K = a+k,
A' 0 K
Fig. 27.15
589
ELLIPSE AND HYPERBOLA [27
x2 y2
= 1.
a2 a2(e2-1)
Since e2 > 1, a2(e2—1) > 0 and we may write
b2 = a2(e2_ 0.
Fig. 27.16
(Notice that this is the first slight point of departure from the correspond-
ing analysis for the ellipse. Notice also that, for a hyperbola, b can be greater
than a.) With this notation the equation may be written as
x2 y2
a2 b2 '
which is called the standard form of equation for the hyperbola.
*Ex. 24. Show that the hyperbola is symmetrical about both the x and y axes
and deduce by an appeal to symmetry, the existence of a second focus S'(— ae, 0)
and second corresponding directrix.
*Ex. 25. Prove that all chords through the centre, 0, are bisected at 0 and that
the length of the latus rectum (see Ex. 6) is 2b2/a.
The form of the complete curve is shown in Figure 27.17. Notice that,
if P(x, y) lies on the hyperbola,
v2
x2 = a2 (1 +, ) .... a2,
2
590
3] HYPERBOLA
and thus no part of the curve lies within the interval -a < x < a. It
follows that the points B (0 , b) and B'(0, - b) do not lie on the curve in
contrast to the corresponding points of the ellipse. A (a, 0) and A' ( - a, 0)
are called the vertices of the hyperbola; AA' is the transverse axis of the
hyperbola, while BB' is the conjugate axis (see Ex. 26).
Fig. 27.17
that is x -y = 1
a b x y.
+-
ab
x y
As x, y both tend to co, ( - --) tends to zero and the equation of the curve
ab
approximates to
x y
- - = 0.
a b
This line is thus an asymptote to the hyperbola; by a similar argument so
also is the line
x
-+- = O.
a b
*Ex. 26. Show that the equations of the asymptotes may be written in the form
x2 y 2
- - -
a2 b2 = 0
and prove that the acute angle between them is 2 arctan (61a).
591
ELLIPSE AND HYPERBOLA [27
this form are said to be conjugate; show that conjugate hyperbolas have the same
asymptotes.
A=
1 1
V2 V2
s((cosin (-1
vr) —sin (-410\
(— in) cos (—lir)/
which represents a rotation through an angle — (see p. 243, vol. 1).
It follows that, under T, the image of a curve C has precisely the same
appearance and geometrical properties as the curve C itself, but is rotated
through an angle of —17r. But, since
1
x =- (x' — y'),
2
1„
Y= +Y
we see that any point of the set
{(x, y): xy = c2}
maps into a point of the set
{(x, y): x' 2 —y'2 = 2c2}.
592
3] HYPERBOLA
Ex. 28. Prove that the eccentricity of a rectangular hyperbola is V2 and that the
latus rectum is equal to the distance between the vertices of the curve. (A rect-
angular hyperbola is sometimes called an equilateral hyperbola. All rectangular
hyperbolas are similar to one another.)
is(a(lIt+t)12, all/t— t)/2). Examine how this point moves along the curve as t
varies between —co and +oo; in particular, explain what happens as t approaches
the value zero from below and from above.
Ex. 31. Show that x = a sec 0, y = b tan 0 gives a parametric representation
for all points of the hyperbola
b2x2 — a2y2 = a2b2
and explain how P(a sec 0, b tan 0) moves on the hyperbola as 0 varies from 0
to 27r. In particular, what happens as 0 approaches the values 17r, lir from below
and from above?
Ex. 32. The circle. x2+ y2 = a2
is called the auxiliary circle of the hyperbola
b2x2_ a2y2 = a2b2
(see Ex. 12). Sketch in the same diagram a hyperbola and its auxiliary circle. P is
a point of the hyperbola with positive coordinates, N is the foot of the perpen-
dicular from P to the transverse axis and NQ is a tangent to the auxiliary circle
at the point Q with positive coordinates. The angle NOQ = 0. Show that Q
is the point (a cos 0, a sin 0) and that P is the point (a sec (- b tan 0).
593
ELLIPSE AND HYPERBOLA [27
Ex. 33. Show that the equation of the tangent to the hyperbola
x2 y2 ,
a2 — b2 = I
at the point P(a sec 0, b tan 0) has equation
x sec 0 y tan 0 _ 1
a b — •
Fig. 27.18
It is not difficult to see that SP— S'P = — 2a if P lies on the branch of the
hyperbola enclosing S, and that SP—S'P = +2a if P lies on the opposite
branch.
*Ex. 34. Prove that, if S, S' are fixed points and P moves in such a way that
SP S'P is constant, then the locus of P is a branch of a hyperbola.
—
594
3] HYPERBOLA
*Ex. 35. Interpret the conditions ISP— S'Pl < 2a and ISP— S'Pl > 2a geo-
metrically.
Ex. 36. Devise a mechanical construction for a branch of a hyperbola with given
foci and length of transverse axis, based upon the relation SP — S'P = 2a.
*Ex. 37. By adapting the method used for the ellipse, prove that the tangent
and normal at the point P of a hyperbola with foci S, S' bisect the angle SPS'.
The given line and conic meet in points with x coordinates given by
ax2 +fl(mx + c)2 = 1,
that is, by (a +/3m2) X2 + 2ftmcx +(/3c2— 1) = 0.
The line y = mx+ c touches the conic ax2 +fiy2 = 1
<r>, n2m2c2 = (cc ±/6,712) (fl C-2 1) (the condition for double roots of the
quadratic equation above)
a fle2 = ± flm2
Ex. 38. An elliptic lamina moves in a plane in such a way that it touches each
of two fixed perpendicular lines. What is the locus of its centre?
596
3] HYPERBOLA
Exercise 27 (b)
1. Find the eccentricities of the following hyperbolas:
(i) 4x2 -1y2 = 1; (ii) 2x2—y2 = 1; (iii) (x-1)2 — 2(y-2)2 = 1.
2. Sketch roughly the hyperbola
+y2 1
and find its eccentricity, the coordinates of its foci and the equations of the
corresponding directrices.
3. Sketch roughly the hyperbola
,16(x — 1)2— i(y + 1)2 = 1,
find its eccentricity, the coordinates of its foci and the equations of the corre-
sponding directrices.
4. Show that the equation
x2 -4y2 +6x+8y+1 = 0
represents a hyperbola, and find its eccentricity, the coordinates of its foci and
the equations of the corresponding directrices.
5. Find the equations of the asymptotes of the hyperbolas
(i) 9x2 _ 4y2 = 12; (ii) 3x2—y2 = 1.
6. Find the equations of the asymptotes of the hyperbolas
(i) x2 — 4y2+ 2x+ 8y— 5 = 0; (ii) 2x2 -4y2 — 8y— 5 = 0.
7. Find the equation of the hyperbola with asymptotes
2x —y = 0, 2x+y = 0,
which passes through the point (2, 3).
8. Find the equation of the hyperbola with asymptotes
3x-2y-9 = 0, 3x + 2y + 3 = 0,
which passes through the point (2, — 4).
9. Find the equation of the tangent to the hyperbola 2x2 — 3y2 = 6 at the point
(3, —2) and prove that the line x+y + 1 = 0 also touches the curve.
10. Find the equation of the normal to the hyperbola 3x2—y2 = 2 at the point
(1, 1) and find the x coordinate of the point where it meets the curve again.
11. Find the equation of the tangent to the rectangular hyperbola
x2— y2 = 3
which passes through the point (3, — 3).
12. Prove that a hyperbola intersects a line parallel to one of its asymptotes at
just one point.
13. Find the equation of the rectangular hyperbola which has the points (4, 0),
(— 4, 0) as foci.
14. If the eccentricity of a hyperbola is e, prove that the eccentricity of the con-
jugate hyperbola is a/,,/(e2 -1).
8 PPMII 597
ELLIPSE AND HYPERBOLA [27
show that the result above is true also for the hyperbola
b2x2_ a2y2 = a2b2.
598
31 HYPERBOLA
Repeat the above process to prove the following more general result: if a chord
PQ of a hyperbola meets the asymptotes at P', Q', then PP' = QQ' .
Suggest a further generalization of this result by considering a second rect-
angular hyperbola xy = d2in place of the asymptotes.
x
0
Ex. 39. Describe in words and draw sketches of the following sets of points:
(i) {(r, 0): r sin B = a}; (ii) {(r, 0): 0 = e};
(iii) {(r, 0): r = a sin a cosec (0-0.
(Hint: rewrite the defining equation and think of the Sine Rule.)
(iv) {(r, 0): a < r < b, 0 < 0 < (v) {(r, 0): r sec 0 <
Ex. 40. Find the polar equation of:
(i) the line passing through the points with polar coordinates (1, 0) and
(1, in);
(ii) the circle through the points with polar coordinates (2, fir) and touching
the initial line at the origin;
(iii) the circle of radius 1 with centre at the point with polar coordinates
(A/2,
This equation is the standard form for the equation of a conic in polar
coordinates with the origin at a focus and the initial line along the major
axis.
Ex. 41. Sketch the parabola r(1 +cos 0) = 2.
Ex. 42. Show that the equation r(2+ cos 0) = 2 represents an ellipse and sketch
the curve.
Fig. 27.21
The polar form for a conic is of great value in dealing with properties
of focal chords. We conclude this section with an example illustrating its
use in this context.
Fig. 27.22
(ii) PQ = PS+ SQ
1 1
1+ V2 cos 0+ 1— V2 cos 0
21
1-2 cost 0'
1PQ1 = 12/ sec 201,
and similarly 'UV 1 = 12/ sec 201.
Ex. 43. PQ is a focal chord through the focus S of a conic of semi-latus rectum
1. Prove that
1 1 2
PS+SQ =i •
602
4] POLAR EQUATIONS OF CONICS
Ex. 44. PP', QQ' are perpendicular focal chords of a conic; prove that
1 1
,+ ,
PP QQ
is constant.
Ex. 45. Show that the polar equation of the directrix corresponding to the focus
0 is
er cos 0 = 1.
Show also that a polar equation of the form
1/r = a cos 0 + b sin 0
represents a straight line.
Deduce (with the help of Exercise 27(a), question 11 (second part)), that
the equation
1
- = cos (B— cc)+ e cos 0
r
is the equation of the tangent to the standard conic at the point with vectorial
angle a.
Prove that the tangents at the ends of a focal chord of a conic meet on the
corresponding directrix.
5. SECTIONS OF A CONE
This section is included for the attention of readers interested in the
historical development of the geometry of the conics.
The conics, or conic sections, were first studied, as their name suggests,
as sections of a right circular cone. Their history extends back to the time
of Ancient Greece: they were first extensively studied by Apollonius
of Perga (247-205 B.c.) who wrote a treatise on their properties. The focus-
directrix property, which we have made the basis of our definition, was
not discovered until later.
The various types of conic arise as sections of a cone made by planes
making various angles with the axis of the cone. For the purposes of the
following definitions, the cone is taken to extend infinitely in both direc-
tions from the vertex. (See Figure 27.23.)
First observe that any plane perpendicular to the axis of the cone (and
not through the vertex) cuts the cone in a circle, which may therefore be
regarded as a particular type of conic.
If the plane is oblique, not parallel to one of the generating lines of the
cone and cuts only one half of the cone, the resulting section is an ellipse
(Figure 27.24).
If the plane is parallel to one of the generating lines of the cone, the
resulting section is a parabola (Figure 27.25).
If the plane is oblique and cuts both halves of the cone, but does not
603
ELLIPSE AND HYPERBOLA [27
pass through the vertex of the cone, the section is a hyperbola (Figure
27.26).
Ex. 46. Show how a pair of straight lines arises as a conic section.
V is the vertex of a right circular cone with its axis vertical and we con-
sider the section of the cone by the plane H. A sphere may be drawn to
touch the plane H at S and also to touch the cone in a circle lying in a
horizontal plane H'. The vertical plane containing V and S cuts H and
the cone at A and A' and H' and the cone at C and C'. 1 is the line of inter-
section of H and H' and A' A meets 1 at K. P is any point on the cone lying
in the plane H and VP touches the sphere at P'. M is the point of 1 such that
PM and AK are parallel, NPQ is a horizontal section, N lying on AA' and
Q lying on VA.
Fig. 27.27
Miscellaneous Exercise 27
1. Show that the equation
x2 +2y2 -2x+12y+8 = 0
represents an ellipse, and find its eccentricity and the coordinates of its centre.
Prove that
3x-2y+2 = 0
is a tangent to the ellipse and find the point of contact.
605
ELLIPSE AND HYPERBOLA [27
2. The tangents drawn from a point P(xi, y1) to the circle x2 + y2 = r2 touch the
circle at L, M. Show that the equation of LM is xx,+ yy, = r2.
If P moves on a hyperbola of eccentricity e with its centre at the origin, show
that LM touches a hyperbola of eccentricity e', where
11 1
= 1.
e2 e'2
Draw a figure for the case when e < Al2 and the circle touches the given hyper-
bola. (London)
3. Show that the equation of the tangent at P(a cos a, b sin a) to the ellipse
b2x2 + a2y2 a2b2
is bx cos a+ ay sin oc = ab.
The tangent at P cuts the x and y axes at A and B respectively and the normal
at P cuts the x and y axes at C and D respectively. Find the ratio PC: PD.
If AD and BC meet at E, prove that BE is perpendicular to AD and hence, or
otherwise, find the equation of the circle through A, B, E. (London)
4. Prove that the conics
x2 y2 X2 y2
a2+b2
— = 1 and +
a2 + A b2 + A
=1
have the same foci, whatever value A takes (provided A — a2or — b2).
Find the equation of the rectangular hyperbola whose foci coincide with those
of the ellipse b2x2 + a2y2 = a2b2.
Show further that two ellipses which have the same foci cannot intersect in
real points but that if an ellipse and a hyperbola have the same foci, they inter-
sect orthogonally at four real points.
5. P is any point on a rectangular hyperbola, centre 0 and foci S, S'. Prove that
OP2 = SP. SI'.
6. P is the point (a cos 0, b sin 0) of the ellipse
b2x2 + a2y2 = a2b2
and Q(a cos 0, a sin 0) is the corresponding point of the auxiliary circle. Prove
that the perpendicular distance from S to the tangent at Q to the circle is equal to
SP.
7. Write down the equation of
(i) an ellipse which has its minor axis along the y axis and touches the x axis;
(ii) an ellipse which has its minor axis along the x axis and touches the y axis.
Two such ellipses are given. Assuming that they meet in four (real) points,
write down the equation of any conic through these four points, and prove that
its centre lies on a certain rectangular hyperbola. (0 & C)
8. A tangent to the ellipse b2x2 + a2y2 = a 2b2
=
x2 y2
T A,
12 ba
where A. < 1. Prove that the tangent to S at any point P meets S' in two points
Q and R which are equidistant from P.
If A = —1, prove that the tangents to S' at Q and R meet on S at P', the
reflection of P in the origin. (0 & C)
10. The asymptotes of the hyperbola
x2
—— ya
—
a2 b2 = 1
are 1 and l', and P is a point on the hyperbola. The perpendicular from P to 1
meets / and l' at X and Y, and the perpendicular from P to l' meets 1' and 1 at
X' and Y'. By expressing the coordinates of P in parametric form, or otherwise,
prove that, for all positions of P,
a2b2
(i) PX.PX' = •
a2+b2 '
a2b2
(ii) PX.PY =
a2— b2;
a2b2(a2+b2)
(iii) PY.PY' = (0 &
(a2 — b2)2
11. Show that, if the point (x1-1-r cos 0, yi+ r sin 0) lies on the central conic
ax2+ by2 = 1, then r satisfies the quadratic equation
r2(a cosy 0 +b sine 0)+ 2r(axicos 0 + byi sin 0)+a4+ by;. — 1 = 0.
Deduce that, if (xi, y1) is the mid-point of the chord PQ, then PQ has gradient
— ax1/by1.
Prove that the locus of mid-points of parallel chords of a central conic is a
diameter of the conic.
12. Use the analysis of Question 11 to prove Newton's Theorem for a central
conic: if PQ, RS are two chords intersecting at X, then the ratio
PX. XQ
RX. XS
depends only upon the directions of the chords PQ and RS, and not on their
positions.
13. Given the outline of an ellipse, show how you would construct the centre,
the axes, the foci and the directrices.
14. Prove that, if tangents are drawn to a hyperbola from any point of the
conjugate hyperbola, their chord of contact touches the opposite branch of
the conjugate hyperbola and is bisected by it.
607
ELLIPSE AND HYPERBOLA [27
15. Find the equation of the perpendicular bisector of the line joining the points
yi), (x2, Y2).
A fixed circle has centre C and radius 2a. A is a fixed point inside the circle and
P is a variable point on the circumference. Prove that the perpendicular bisector
of AP touches the ellipse whose foci are at C and A, and whose major axis is of
length 2a. (C.S.)
16. A point P is taken at random inside an ellipse of eccentricity e. Calculate
the probability (in terms of e) that the sum of the focal distances of P should be
not greater than the distance from a focus to the opposite end of the major axis.
(C.S.)
(Note: the area of an ellipse with major and minor semi-axes of lengths a and b
respectively is rrab—a result easily deduced by the usual calculus methods, or
alternatively by considering the effect on areas of the linear transformation T of
Section 2 of this Chapter.)
17. E, is a circle, centre 0, radius b, E2 is a circle, centre A radius a (a < b) which
touches /1internally. Describe the locus of the centre, P, of a variable circle, Z,
which touches Elinternally and E2 externally.
If E, is a straight line, E2 is a circle, centre A and radius a which touches Ei
and P the centre of a variable circle E which has Elas a tangent and touches E2
externally, describe the locus of P.
18. Prove that at most four normals may be drawn from a point A to a central
conic with centre 0.
Prove further that, if the normals at the points -P1, -P2, -P3, -P4 on a central conic
intersect at A, then -P1-P2-P3-P4 lie on a rectangular hyperbola which passes through
0 and A and has its asymptotes parallel to the axes of the given conic. (The
rectangular hyperbola of this question is known as the hyperbola of Apollonius.)
19. Prove that the vector equation (referred to the vertex as origin) of the tangent
to the parabola y2 = 4ax at the point p = at2i+2atj is r = p+Au where u is
the unit vector (ti+ j)/(1 + t2)i.
Prove that the locus of the meets of tangents to the parabola y2 = 4ax which
cut at a fixed angle a is a hyperbola of eccentricity sec a.
20. Prove that, if the chord PQ of any conic with focus S, when produced, cuts
the corresponding directrix at R, then SR is a bisector of the angle PSQ. Deduce
that, if the tangent at P cuts the directrix at T, then the angle PST is a right angle.
21. If X is any point on the tangent at the point P of any conic with focus S,
and if H, K are respectively the feet of the perpendiculars from X to SP and the
directrix corresponding to S, prove that SH = eXK, where e is the eccentricity.
608
28. Further matrices
Ex. 1. Prove the assertion that if T maps a point P (other than the origin) into
itself, then it maps every point of the line OP into itself.
Ex. 2. Prove that, if T maps two points, P and Q, into themselves, where 0,P, Q
are distinct and not collinear, then T is the identity transformation, that is, the
transformation with matrix
We may now investigate the answer to the question: Even if T does not
map any point other than the origin into itself, is there a line 1 through the
origin such that every point of l is mapped into a point of 1? If such a line
exists, we say that 1 is an invariant line and that 1 maps into itself under T.
Ex. 3. If S is a point such that T(S) e OS, prove that every point of the line OS
is mapped into a point of OS.
Suppose that S is a point other than the origin, with the property that
T(S) E OS; write OS = s. Then As = As and thus
(A — AI) s = 0.
Since s 0, we must have det (A — AI) = 0 (see p. 294). This is a quadratic
equation in A, the characteristic equation of the matrix A, with, in general,
609
FURTHER MATRICES [28
two distinct roots, Al and A2, called the eigenvalues of the matrix A. Any
non-zero solution, s1, of the homogeneous equation
(A — s = 0
is called an eigenvector corresponding to the eigenvalue A1; similarly, we
can find eigenvectors corresponding to the eigenvalue A2.
A = (4
6 2
5)
that is
(6 4) (y1 — (0
0) •
On solving these equations we obtain x = 2 a, y = — 3,1t, giving as an
,
eigenvector
2,tt\
=
— 3i1) •
In particular, a unit eigenvector is given by
2[03 \
= k 3/03) •
—6
4 x)
_ 32) (y (0
0) ,
that is,
v\
giving s2
= (2v) •
— (1/V5
2/v5) .
map into points of the same line (but, apart from 0, not this time into
themselves): 2x —y = 0 is a second invariant line under the transforma-
tion.
(these eigenvectors being chosen for their simplicity, although any non-
zero multiples of s, and s2would do just as well) has an important property.
Since As, = s, and As2 = 8s2
it follows that the product AP is a 2 x 2 matrix with s, as its first column
and 8s2as its second column. Thus
1 0
AP = P(
0 8)
But, since det P = 7, P is non-singular and we may multiply both sides
of the above equation on the left by P 1to obtain
1 0
P-1AP = (
0 8) •
The matrix A has been reduced to diagonal-form. Notice that the ele-
ments of the diagonal are precisely the eigenvalues of A.
611
FURTHER MATRICES [28
for
1 1 0\ 1oc\ 1 a\
V) 8) k,e) — ‘8,e)
as required.
Thus, when considering the transformation T, it is simpler to express
vectors in terms of the eigenvectors s1, s2rather than in terms of the more
usual base vectors i, j. For example, the point Q, where
q = 2s1 + is2,
maps into the point Q', where
q' = 251+452.
Figure 28.1 shows the effect of the transformation T upon Q:
OQ = 2s1 +-1-s2 = OP+PQ,
OQ' = 251+452 = OP'+P'Q',
where OP' Q'P is a parallelogram.
P'
Fig. 28.1
612
1] 2x2 MATRICES
83)
= (P-1AP)3 = (P-1AP) (P-1AP) (P-1AP)
0
( 1
= P-1A(PP-1) A(PP-1) AP
= P-1A3P
As p 1 \ RA
k0 83)
=1/2 /1 0\ /2 - 1
7 k -3 2)2)k0 512) k3 2)
=1 2 512\ /2
7k/ - 3 1024) k3 2)
= 1/1540 1022\
7 k3066 2051)
/220 146
438 293) •
More generally we have
A 1 2 1\ /1 0 \ /2 - 1 \
= 7 k -3 2) k0 8n) 3
1 4+3.84 -2+2.84\
= 7 (-6+6.8n 3+4.8n)•
Ex. 5. Verify by direct calculation the form given for A" above and also prove
the result by mathematical induction.
*Ex. 6. If si and s2are eigenvectors corresponding to the distinct real eigen-
values A1, A2 of a matrix A and if P is the matrix with first column s1and second
column s2, and assuming that P is non-singular, show that
p iAp (A 0 )
-
0 A2
A=
- 1)
613
FURTHER MATRICES [28
bb2 (00) 0;
(0 1)
(bb: 4) (0) b4
Theorem 28.2. If the eigenvalues of the 2 x 2 matrix A are distinct then the
matrix P is non-singular.
Proof det P = 0
s1= ks„ where k is a non-zero number,
As,— kAs2 = 0, on multiplying by A,
A,s, —Aks, = 0
(Ai — A2) sl = 0, since ks2 =
Al = A2, since s1+ 0.
Thus, since det P = 0 Al = A2, we have Al + A2 det P + 0.
Theorem 28.3. If siand s2are unit eigenvectors of A then s1and s, are per-
pendicular if and only if P is orthogonal.
Proof
_ (x) s2 _ (x2)
Write
Y2
xi x2
then P
= ( I. Y2
pip = (xi Yi\ i yi2
x+ x1X2 ±Y1Y2)
and
kx2 Y2/ \Yi x2' = X2 xl Y1Y2
2
x2+ y22 •
and P is orthogonal.
616
1] 2 x 2 MATRICES
Theorem 28.5. The reduction to diagonal form P-'AP for a symmetric matrix
A + kI may always be effected, and P may be taken to be orthogonal.
Proof. Notice first that, since s1and s2are non-zero,
s1perpendicular to s2 <=> si s2 = 0.
Now suppose that si and s2are unit eigenvectors of A.
As, = A,s,
sAs„. = Aiss„ on premultiplying by s;,
(sAs„)' = A1(sis2), on taking the transpose of each side,
ra slAs2 = A,sis2, by Ex. 8 and noting that A' = A,
A2 Si S2 = A1s1s2, since As2 = A2 s2,
-4-> (Ai— s2 = 0
s;.s2 = 0, since Al + A2 by Theorem 28.4.
Thus, s1and s2are perpendicular and, by Theorem 28.3, P is orthogonal.
Furthermore, by Theorem 28.2, P-1exists and the theorem is complete.
(Notice that orthogonal matrices are always non-singular.)
1 3
Ex. 12. If A = (3 1), find an orthogonal matrix P such that
P'AP =
°
2)
and interpret P, P' as matrices of a rotation transformation.
617
FURTHER MATRICES [28
Exercise 28(a)
1. Find the eigenvalues of the matrix
(5 2
A=
98
and the corresponding unit eigenvectors. What lines in the plane map into
themselves under the linear transformation T with matrix A?
2. Find the eigenvalues of the matrix
2 3\
A
= (8 0/
and the corresponding unit eigenvectors. What lines in the plane map into them-
selves under the linear transformation T with matrix A?
3. Find the eigenvalues of the symmetric matrix
(23 36\
A=
36 2/
and hence find an orthogonal matrix P such that P-1AP is diagonal.
4. Find an orthogonal matrix P such that the matrix P-1AP is diagonal, where A
is the symmetric matrix
5 —1\
- 5/ •
(ac
are 1 and A, where 0 < IAI < 1.
Find A for the matrix
A= 1\
and describe geometrically the linear transformation with matrix A" where n
is a large positive integer.
12. Find the eigenvalues of the matrix
23 —33\
A
= (14 —20/
and deduce that
An = 11.2"}1-21 —33.2"+33
7.2"}1-14 —21.2"+22
If B = An+An-1-1-An-2 + +A+I
find an explicit form for B.
What are the eigenvalues of B?
Ex. 14. Show that, if T(S) = S where S is not the origin, then T maps each point
of the line OS into itself.
Ex. 15. U, V, W are distinct non-collinear points in space with T(U) = U,
T(V) = V, T(W) = W. Prove that, if the plane UVW does not contain the origin,
then T is the identity transformation.
Prove that:
(i) A is similar to A;
(ii) if B is similar to A, then A is similar to B;
(iii) if B is similar to A and C is similar to B, then C is similar to A.
7x -5y + 10z = 0,
621
FURTHER MATRICES [28
Ex. 17. Verify that, in the notation of Example 2, P-1AP is a diagonal matrix.
Ex. 24. If 1 -2 - 2
P= -2 1 -2 ,
-2 -2 1
prove that P is orthogonal and interpret P as the matrix of a linear transformation.
*Ex. 25. If u is a unit vector, show that it is possible to construct an orthogonal
matrix with u as its first column.
*Ex. 26. If P, Q are two 3 x 3 orthogonal matrices, prove that:
(i) P--1is orthogonal; (ii) PQ is orthogonal.
Interpret both these results geometrically.
Ex. 27. Find a result corresponding to Theorem 28.6 for 4 x 4 orthogonal
matrices. (Be careful when you take determinants !)
where P is an orthogonal matrix with sias its first column and a, ... 72 are
constants (see Ex. 25). Premultiplying by P-1we then have
Ai oci. oc2
P-1AP = (0 fli 1 32 .
0 7, 7
But (P-1AP)' = P'A'(P-1)' = P'A(P')' = P-1AP (see Ex. 8), since
A' = A and P' = P-1. It follows that P-1AP is a symmetric matrix:
A1 0 0
P-1AP = (0 A 13) •
0 fi2 72
1612 fi2)
(8 72
to diagonal form. Thus the matrix
(1 0 0
Q = 0 q1 q2
0 r, r2
reduces P-1AP to diagonal form:
Ai 0 0
Q-1P-1APQ = (0 it2 0).
0 0 it3
Finally, we may write PQ = R where, by Ex. 26 (ii), R is orthogonal.
Ex. 28. Is the converse result true that, if P is orthogonal and P-1AP is diagonal,
then A is a symmetric matrix?
625
FURTHER MATRICES [28
Example 3. If
11 —1
A= k2 3)'
find A8.
The characteristic equation of A is
A2-4A+5 = 0
and thus by the Cayley—Hamilton theorem
A2 — 4A+ 51 = 0.
But A8 .=_ (A2 — 4A+ 5)f(A)+ccA +ft,
where cc, ft are integers. Since A = 2+j =- A2 -4A + 5 = 0 we have
(2 +j)8= cc(2 + j) +fl.
Using the Binomial Theorem and equating real and imaginary parts,
28 -28.26 +70.24 -28.22 +1 = 2cc+fl
1 8.27-56.25+56.23-8.2
f16[16-112+70-7]+1
=a
= 2cc+fi
116[64 — 112 + 28 — 1] =a
f cc = —336
tiG = 145.
Thus A8----- (A2 — 4A + 5)f(A)— 336A +145
and A8= —336A +1451, since A2-4A+ 51 = 0,
= 1-336 336\ + (145 0\
k — 672 —1008) k o 145/
= I 191
— 336\
k — 672 —863) •
Alternatively, one can 'build up' A8:
A2 = 4A-51, as before;
A4 = 16A2 — 40A + 251, using the commutative rule,
= 16(4A— 51)— 40A + 251
= 24A-551;
A8 = 242A2 — 48 . 55A + 5521
= 242(4A — 51)— 48 . 55A + 5521
= —48 .7A— 5(576 — 605) I
= — 336A+ 1451,
and proceed as before.
627
FURTHER MATRICES [28
Exercise 28 (b)
1. Find the eigenvalues of the matrix
3 4 1
A= (2 5 1
2 3 3
and determine corresponding unit eigenvectors.
2. Find the eigenvalues of the matrix
(1 -2 1
A = 3 -4 1 .
3 -7 4
What does the fact that zero is one of the eigenvalues tell us about A?
Find a matrix P such that P-1AP = D, where D is a diagonal matrix.
3. Find the eigenvalues of the matrix
31 -2
A= 4 0 -2
( 4 -1 -1
and hence write down the value of det A.
Find A-1and determine the eigenvalues of A-1.
Suggest and prove a general result about the eigenvalues of an inverse matrix.
4. Find the eigenvalues of the matrix
1 2 -2
A= 6 4 -6
(6 5 -7
and corresponding unit eigenvectors. Find a matrix P such that P-1AP = D
where D is a diagonal matrix.
Show that, if A is the matrix of the linear transformation T, then T maps all
points of the plane 2x +y - 2z = 0 into points of the same plane. Find the equa-
tions of the other two planes through the origin which have this property.
5. Find the eigenvalues of the symmetric matrix
-1 -6 -4
A = 1 (- 6 -2 -2
-4 -2 3
and hence find an orthogonal matrix P such that P-JAP = D, where D is
diagonal.
628
EXERCISE 28
6. Find the equations of the planes through the origin, which map into them-
selves under the linear transformation T with matrix
(1 3 —2
A=31—2.
3 4 —5
7. P is a 2 x 2 orthogonal matrix with det P = 1. Show that the eigenvalues of
P are ej0 and e-i°, and interpret O.
Show that, although the geometrical property of a 2 x 2 orthogonal matrix P
is characterized by its eigenvalues if det P = + 1, this is no longer true if
det P = 1.—
tr A = k+A2 +4.
Prove also that tr (A + B) = tr A + tr B and that tr (,itA) = u tr (A).
9. A and B are both 2 x 2 matrices. Prove that A and B have the same eigenvalues
if and only if both det A = det B and tr A = tr B. (See Question 8 for definition
of trace.)
10. Prove that, if A is a non-singular 3 x 3 matrix, then
det (A — AI) = det A — A det A tr (A-1)+M tr A— A3.
11. If A, B have a common eigenvector s, with corresponding eigenvalues A, it,
prove that s is an eigenvector of (i) A + B; (ii) AB. What are the corresponding
eigenvalues?
12. A is a 3 x 3 skew-symmetric matrix (A = — A'). Prove that
det (A— AI) = — det (A + AI)
and deduce that if A possesses a non-zero eigenvalue a, then — a is also an eigen-
value.
If f(A) = det (A2 —AI), prove that — f(A)/A is the square of a linear poly-
nomial in A.
13. If B = P-1AP and s is an eigenvector of A prove that P-ls is an eigenvector
of B.
14. If A is an eigenvalue of the 3 x 3 matrix A prove that Ar" is an eigenvalue of
An. Deduce that, if
A* = a2A2 + alA + a,I,
then Ar = a2 A2+ai A+a0.
15. A matrix A is said to be nilpotent if there exists a positive integer n such that
An = 0. Prove that, if A is nilpotent, then all the eigenvalues are zero.
16. Find A4 + A2 + I if
A=( 3 5\ .
—1 —2)
9 PPMII 629
FURTHER MATRICES [28
17. If
A— (12
5 "52)
Al 0 0
P-1AP = 0 Al 0 .
0 0 A2
show that Tleaves all the points of a certain plane containing the origin fixed and
find the equation of this plane. Show also that every plane containing the line
x y =z
5 3 2
is mapped into itself by T.
Find a diagonalizing matrix P for this particular matrix A.
El = 13-1(0
NT
0
0
Nr+1
0
0
0 P,
(dr+2
J
for some matrix P and integer r.
Prove further that, if 52 t I, then the matrix
1+52+522
is singular.
28. Show that the linear transformation T with matrix
1 3 —2
A= -1 5 —2
(—1 4 —1
leaves just two lines through the origin invariant and find their equations.
Explain geometrically why it is impossible to find a matrix P such that the
matrix
P-1AP
is diagonal.
632
29. Further coordinate geometry
Prove that a straight line not parallel to the y axis cuts the curve in either
one or three real points and that, if a straight line not through the origin cuts
the curve at P, Q, R and the tangents to the curve at these three points cut the
curve again at P', Q', R', then P' Q' R' is a straight line.
Any line not parallel to the y axis has an equation of the form
y = mx + c.
This cuts the curve in points with parameters given by the roots of the equa-
tion
at3= mat2+ c,
that is, by at3—mat2 — c = 0.
Since this is a cubic equation in t, it has either one or three real roots,
due regard being paid to their multiplicity. This proves the first result.
Now suppose that the parameters of the points P, Q, R are respectively
p, q, r. Then, if the equation of the line PQR is y = mx+ c we see that the
roots of the cubic equation
at 3— mat 2—c = 0
are p, q, r; thus qr+rp+pq = 0.
But the line does not contain the origin; thus p, q, r 0 and we have
1/p+ 1/q+ 1/r = 0.
Again, since the tangent at P cuts the curve at P', parameterp', we have,
as above,
(1/P)+(l/P)+(1 /P') = 0,
giving, = — 2/p, similarly 1/q' = — 2/q, and 1/r' = — 2/r and we have
(1/p')+(l/q')+(l/r') = —2[(1/p)+(1/q)+(l/r)] = 0.
It follows that p', q', r' are the roots of a cubic equation of the form
ats — m' at2— c = 0
(where m' = p' +q' +r' and c' la = p'q'r') and thus the points P', Q', R'
lie on the line
y = m'x+ c'.
The shape of the semi-cubical parabola is shown in Figure 29.1. 0 is
called the pole and Ox is the axis of the curve. The Cartesian equation is
x3= aye. A connection between the parabola and the semi-cubical para-
bola is obtained in Example 2.
Example 2. Given a parabola and a point C, prove that, in general, either one
or three normals may be drawn to the parabola to pass through C but that,
if just two such normals may be drawn, then C lies on a certain semi-cubical
parabola.
634
PARAMETRIC FORMS
Fig. 29.1
This cuts the parabola at points with parameters given by the quartic
equation
(at 2 -a)2 + (2at -
fi)2 == 72.
But the coefficient of t3in this equation is zero and thus the sum of the
roots is zero. Since three of the roots are, by definition p, q and r, and since
p+q 1 r = 0, the fourth root must be zero and we have proved the final
--
result.
Ex. 2. Prove that the normals to the parabola y2= 4ax touch the semi-cubical
parabola of Example 2. Draw in the same sketch the parabola y2 = 4ax and the
semi-cubical parabola 4(x- 2a)3= 27ay2. (It is worth drawing the parabola
accurately and then constructing a large number of normals—which may be
done very rapidly using the geometrical property that, in the notation of Exercise
22(b), NG = 2a.)
Ex. 3. Draw the parabola y2 = 4ax and shade in the set of all points from
which three real normals may be drawn to the parabola.
Ex. 4. The Folium of Descartes has Cartesian equation
x3 +y3 = 3axy.
By considering the intersection of the line y = tx with the folium, show that the
curve may be represented parametrically by
3at 3at 2
x= y
1+ t 3 1+ t 3
Taking t = 0 it is seen that the origin lies on the curve and that the x axis is a
tangent there; by writing u = t-1, show that the y axis is also a tangent at the
origin (which is thus a double point of the curve, that is, a point through which
pass two separate branches of the curve).
Show that the line x- y = 0 is an axis of symmetry and that the line x+y = a
is an asymptote. Sketch the curve.
Ex. 5. Using the method of EX. 4, find a parametric form for the curve
y2 = x2(1 +
and deduce that no point of the curve has an x coordinate less than -1. Prove
also that the x axis is an axis of symmetry for the curve.
Show that the origin lies on the curve and that this point arises from two dis-
tinct values of the parameter. Deduce that the origin is a double point of the
curve and determine the equations of the two tangents there.
636
1] PARAMETRIC FORMS
Exercise 29(a)
1. Show that, if OP = r = at i + at 3j, then P lies on the curve I': x3 = a2y.
Prove that any line in the plane cuts r in either one or three real points.
The tangents at three points P, Q, R of the curve meet F again at P', Q', R'
respectively. Prove that:
(i) if the centroid of the triangle PQR lies on the y axis, so also does the cen-
troid of the triangle P' Q'R' ;
(ii) if P, Q, R are collinear, so also are P', Q', R'.
Prove that, if the points with parameters t1, t2, t3are collinear, then
1+t1+t2 +t3 +t1 t2 t3 = 0,
and conversely.
Show that, given a point P lying on 1', two lines may be drawn through P to
touch the curve at Q and R. If QR cuts r again at S find the parameter of the
point S in terms of the parameter of P.
Show that the roots of the equation t 2+ t — 1 = 0 give the double point of the
curve.
(iii) that if t2 and t3satisfy the above equation then the points
t = tl, t = t2 t = t3
,
are collinear.
A chord through the point (a, 0) meets the curve again at the points P and Q.
Prove that the locus of the middle point of PQ is a curve with parametric equa-
tions
x = — laT2, y = — 1-T(T 2+2). (0 & C)
6. Prove that, in general, three tangents may be drawn from a point C to the
cubic curve x3 = a2y and that, if these tangents cut the curve again at P, Q, R
then the tangents at P, Q, R are concurrent.
7. A rectangular hyperbola is given parametrically by the equations
x = ct, y = c/t.
If the four points of the curve with parameters tb t2, t3, t4lie on a circle, show
that ht2t3t4, = 1. Show conversely that, ifif _1 t2 t 3 t4 = 1, then the four points lie
t tt
on a circle.
A variable circle passes through the fixed points A, B of a rectangular hyper-
bola, and meets the hyperbola again at P, Q. Show that the direction of PQ is
fixed.
8. P, Q are variable points on the parabola y2 = 4ax such that PQ is parallel to
the fixed line x+ ky = 0. The normals to the parabola at P and Q meet at R.
Prove that the locus of R is the normal to the parabola at a fixed point on the
parabola, and find the coordinates of this point. (0 & C)
9. Find the equation of the normal to the rectangular hyperbola xy = c2 at the
point (ct, c/t).
The normals to the rectangular hyperbola at the points P, Q, R, S are concur-
rent; prove that each of these points is the orthocentre of the triangle formed by
the other three.
10. The normals at three points P, Q, R of the parabola y2 = 4ax meet at a
point N. Prove that the centroid of the triangle PQR lies on the axis of the para-
bola.
If N coincides with P, prove that QR passes through a fixed point (that is, a
point whose position is independent of P, Q, R). (0 & C)
11. Find the coordinates of the point P, other than the origin in which the line
y = tx meets the curve x3 + y3 = axy. (t is called the parameter of P.)
If a line meets the curve in three points whose parameters are t1, t2, t3, prove
that ti t2 t3= —1.
If Q is any point on the curve with negative parameter, not equal to —1, prove
that there are two points P1, P2 of the curve (other than Q) such that the tangents
at P, and P2 pass through Q. Prove that OP'and OP2make equal angles (apart
from the sense) with either of the coordinate axes. (0 & C)
12. Prove that the equation of the normal to the curve y = x3at the point
(t, t3) is
x+3t2y = t+30.
638
1] PARAMETRIC FORMS
may be rewritten as
(x+u)2+(y+v)2+(z+w)2 = u2 ±v2 4.14,2—d,
and thus represents a sphere, centre (— u, — v, — w), provided
u2 v2+ w2 > d.
Ex. 10. Find the centre and radius of the circle determined by the sphere
(x— 1)2 + (y — 3)2 + (z — 2)2 = 20
and the plane x—y-3z = 3.
Now consider two spheres, with equations
(r a) . (r a) = c2, (r b) . (r b) = d2.
— — — —
that is, to the line of centres AB and, since a plane and a sphere determine
a circle, we see that two spheres, provided they intersect at all, intersect in
a circle.
641
FURTHER COORDINATE GEOMETRY [29
through the centre, B, of the sphere and is perpendicular to n; thus the locus
of the mid-points of parallel chords of a sphere is a plane through the
centre of the sphere (a diametral plane).
radius to the point of contact has coordinates (4+A, — 5A, 3 +A). This lies
on the given plane if
(4+A)— 5( — 5A)+ (3 +A)+ 20 = 0,
that is, if A = —1. This gives the coordinates of the point of contact as
(3, 5, 2).
Fig. 29.2
*Ex. 11. Show that a parametric form for a cylinder, radius b, with its axis along
the z axis, is
r = b cos Oi+ b sin 61+ Ak.
Notice once again that we have two parameters, 0 and A for this surface.
Ex. 12. What does the surface
x2 y2
+ =1
a2 b2
represent in three dimensions?
A cylindrical spring is a three-dimensional curve lying on the surface
644
2] SURFACES IN THREE DIMENSIONS
where p (the pitch of the helix) and b (the radius of the cylinder on which the
helix lies) are constants. Note that the parametric representation of the
helix involves only one parameter.
*Ex. 13. Interpret geometrically the constant p, the pitch of the helix.
Ex. 14. Show that the helix defined above is one of the two curves of intersection
2lTz
of the cylinder x2 +y2 = b2and the corrugated surface x = a cos . Show also
that the other curve formed is an oppositely twined helix.
The right-circular cone is the surface traced out by a variable line through
a fixed point 0 (the vertex of the cone) and making a constant angle with
a fixed line though 0 (the axis of the cone). If we take 0 as the origin and
the unit vector u = /i+ +nk as defining the direction of the axis we have,
for any point P of the cone,
r.0 = Irl cos a,
where r = OP and a is the angle of the cone (Figure 29.3). In Cartesian
form, this gives
(lx + my + nz)2= (x2+ y2+ z2) cos' a
Fig. 29.3
and we see that the equation of a cone with its vertex at the origin is
homogeneous in x, y, z. We saw in Chapter 26 that the section of a plane
with a cone is a conic.
cone (a generator of the cone) (see Figure 29.4). Then, if LOVT = a, since
OV = 3 and the radius of the sphere is 1,
cos a = 232 .
Now consider any point P(x, y, z) on the cone. A unit vector along
OV is u = and thus, since
VP.0 = IVPI cos a
we obtain
[1(x — 2) — A-(y + 1) — 1(z + 2)]2 = Ex— 2)2 + (y + + (z. + 2)9,
Fig. 29.4
Fig. 29.5
or A1x'2 +A2y'2 = 1.
But A is a symmetric matrix and, by Theorems 28.3 and 28.4, we can find
an orthogonal matrix P such that PAP' = D where
Al \
D
= (0 A2P
Al and A2 being the eigenvalues of A. (Recall that, for an orthogonal matrix,
P-1= P'. Notice too that, to obtain PAP' in diagonal form (rather than
P'AP), the matrix P is obtained by transposing the matrix with columns
which are the eigenvectors of the matrix A; that is, the eigenvectors appear
as the rows.) The fact that P is orthogonal tells us that the new axes, Ox',
Oy' will be perpendicular. The details of the transformation are given
below.
Consider the linear transformation defined by
v = Pu,
x
where v = , the new position vector of the point u = . Then
.Y, vv
(;)
P'v = P-lv = u
and the equation u'Au = 1
transforms into the equation
(P'v)' A(P'v) = 1
or, since (P')' = P, and using the results of Ex. 8, Chapter 28,
v'PAP'v = 1.
But PAP' = D, a diagonal matrix and the equation has been transformed
into
A02) (y
x,')
(x' Y) (O1
or A1x,2 + 42, = 1.
Thus we have tr A = Ai + A„
det A = A1A2;
but tr D = Ai + A„
det D = A1k2,
and thus both the trace and the determinant of the matrix defining the
conic remain invariant under the given linear transformation.
*Ex. 15. Show how the invariance of the trace and determinant enables us to
deduce immediately the nature (ellipse, hyperbola) of a central conic
ax2+2hxy+ by2= 1.
Ex. 16. Prove that, if Q is any non-singular 2 x 2 matrix, then
tr (Q-1-AQ) = tr A and det (Q-1AQ) = det A.
Can these results be generalized for higher-order matrices (the trace still being
defined as the sum of the elements in the leading diagonal)?
Thus 13 ' = 1)
2
= /2 —1\
and P
k
The transformation v = Pu
may be interpreted as a rotation of axes clockwise through an angle
0 = arctan (Figure 29.6).
Fig. 29.6
In other words, in terms of the original coordinate system, the major and
minor axes of the conic lie respectively along
x + 2y = 0 and 2x — y = O.
*Ex. 17. Show that the equation
ax2+2hxy+by2 = 1
(i) represents an ellipse if h2— ab < 0, and a > 0;
(ii) represents a hyperbola if h2 — ab > 0;
(iii) represents a rectangular hyperbola of a+ b = 0;
(iv) represents a pair of parallel straight lines if h2 = ab (h * 0).
Ex. 18. Let us write the conic of Ex. 17 in the form
x'Ax = 1,
(a
where A= x = (x )
h' Y);
u is a unit direction vector and P is a point with position vector p. The point
with position vector p+ Au lies on the conic if
(p+Au)' A(p+Au) = 1;
651
FURTHER COORDINATE GEOMETRY [29
Exercise 29 (b)
1. Obtain the equation of the sphere:
(i) with centre (0, -1, 0) and radius 1;
(ii) with centre (2, - 3, -1) and radius 5;
(iii) with centre (a, b, c) and radius ,N1(a2+ b2+ c2).
2. Find the centre and radius of the sphere:
(i) with equation x2 + y2+ z2- 6x-2y +2z- 5 = 0;
(ii) with equation x2 + y2+ z2 + 4x-8y -2z-60 = 0;
(iii) with equation x2 + y2 + z2 -2ax-2by+2cz+2bc-2ca+2ab = 0.
3. Find the equation of the sphere:
(i) through the points (4, 4, 4), (5, 6, 1), (0, -4, 2), (7, 3, 2);
(ii) through the points (2, 4, 4), (5, 1, 4), (3, 4, 3), (- 3, -1, 0);
(iii) through the points (0, 0, 0), (a, 0, 0), (0, b, 0), (0, 0, c).
4. Find the centre and radius of the circle:
(i) with equations x2 + y2 + z2 - 12x+ 2y - 10z+ 44 = 0, x- y+z = 6;
(ii) with equations x2 +372 + z2 - 18x - lOy + 12z-55 = 0, 2x + 3y - 4z = -1.
5. The position vectors of points A and B relative to some origin 0 are respec-
tively a and b. Show that the locus of a point P which moves in space in such
a way that PA2+PB2 = d2, where d is a constant, is a sphere with its centre at
the mid-point of AB. Find the radius of the sphere in terms of a, b and d. What
can you say about the sphere if d = la-bl?
6. With the notation of Question 5, show that the locus of a point P which moves
in space in such a way that PA2+PB2+PC2 = d2, where C is a point with posi-
tion vector c, is a sphere and find its centre.
Can you generalize this result in any way?
7. Show that the line r = Au, where u is a unit vector, cuts the sphere
(r -3i- 2j - 2k). (r -3i - 2j - 2k) = 6
at two points with parameters given by the quadratic equation
A2-2Au . (3i + 2j + 2k) + 11 = 0.
652
EXERCISE 29
Deduce that the line through the origin in the direction of the vector i+j+ 3k
touches the given sphere.
Prove more generally that the line through the origin in the direction of the
vector /i+ mj+nk is tangential to the sphere if and only if
(31+2m+2n)2 = 1412+m2+n2),
and hence show that the equation of the cone with vertex the origin which cir-
cumscribes the sphere is
(3x+ 2y + 2z)2 = 11(x2"2+z2).
8. Find the equation of the cone obtained by rotating the line
xlp = ylq = zlr
about the line x/a = ylb = z/c.
9. A right-circular cone of semi-angle a has its vertex at the origin and contains
lines in the directions i, i+j and i +j+k. Show that
sect a = 9— 242-246.
10. Show that the equation
yx+ zx+xy = 0
represents a cone, vertex the origin and semi-angle arctan 42. What is the
curve of intersection of this cone:
(i) with the plane x+y+ z = 1;
(ii) with the plane x— 2y+ z = 0;
(iii) with the plane x —2y + z = 1?
11. A circle, S, is defined by the vector equation
(r — a) . (r — a) — c2 = 0, r . n = p.
Show that any sphere passing through the circle S has vector equation
(r a) . (r a) c2 +A [r n —p] = 0
— — —
Find the coordinates of the point in which the helix cuts the plane
1/2x — y + 2,/2z = 0.
Show also that the helix cuts the plane x+ z = 0 at an infinite number of
points, and explain this geometrically.
18. Prove that the mid-points P, Q, R, S, T, U of the edges BC, CA, AB, AD,
BD, CD of a tetrahedron ABCD, in which the opposite edges (BC, AD),
(CA, BD), (AB, CD) are perpendicular pairs, lie on a sphere. Is the converse
result, that if the points P, Q, R, S, T, U lie on a sphere then the opposite pairs
of edges are perpendicular, true or false?
If the sphere cuts BC again at P', prove that BC is perpendicular to the plane
P AD.
19. Two spheres, centres Aland A2 and radii c1and c2intersect in a circle of
radius d. If AlA2 = 1, where 12 = c,2+ cL prove that d = c1c211 and find the ratio
in which the centre of the circle divides the line A1./12.
20. Prove that there is one and only one sphere which contains a given circle
and passes through a given point not in the plane of the circle. Two circles, not
in the same plane, meet at points A and B. A plane meets one of these circles
at C and D and the other circle at E and F. Prove that:
(i) the points C, D, E, F lie on a circle;
(ii) the lines AB, CD, EF are either concurrent or parallel. (0 & C)
654
EXERCISE 29
21. Show that, if the ellipse (x2/a2) + (y2/b2) = 1 lying on the plane z = 0 is
rotated through four right angles about the x axis, the equation of the surface
obtained is
x2 y 2 z 2 1
ai+ b2± b2 = '.
The surface thus obtained is called an oblate spheroid if a < b and a prolate
spheroid if a > b. Both surfaces are particular cases of the ellipsoid.
x2 y 2 z 2
-+ = 1.
a2 -b2 +c2
656
Revision exercise C
1. The polynomials g(x), h(x) both leave the same remainder on division by
(x— a). Prove that the polynomial
xg(x) — ah(x)
is divisible by (x — a).
The polynomial f(x) (degree 4) leaves a remainder of rx+ s on division by
(x— b)3. Prove that f"(x) is divisible by (x— b).
2. If z * 1 and
1 +z
w= 1 —z
show that IzI = 1 if and only if Re (w) = 0.
3. If a < b = an < Irn for all a, b e R where n e Z, what further conditions does n
satisfy? Are these further conditions still necessary if a, b are restricted to positive
real numbers? (S. M.P.)
4. The function f: R R is defined by
f(x) = x2— 21x1.
Sketch the graph off and find what values of x are invariant under!.
5. Solve the equation 3 sin 0+ 2 sin (60°— 0) = 2
for values of 0 lying between 0° and 360°.
6. A random number table consists of a succession of digits each chosen at
random independently from the set 0, 1, 2, ..., 9. Two successive digits are taken
from such a table; show that the probability that their sum is 9 is 1/10.
Four successive digits are taken from the table. Calculate the probability, p,
that the sum of the first two equals the sum of the third and fourth. (M.E.I.)
7. Prove that
a(1 — r")
a+ ar+ ar2+ ...+arn-' =
1 —r '
where r * 1.
Prove that in general
(1+x+x2+...+x2n)(1—x+x2—... 4.x21 = 1±x2+x4+...-1-x4".
State any values of x for which your proof does not apply and obtain the
appropriate results in each of these cases. (0 & C)
8. Prove that, if the points 0, A, C are non-collinear, then the position vector
with respect to 0 of any point in the plane 0 AC may be expressed uniquely in
terms of a = OA and c = OC. (over)
657
REVISION EXERCISE C
25. Three points P, Q, R vary in position on the surface of a fixed sphere in such
a way that pA2 +QA2+RA2
is constant, where A is a fixed point. Find the locus of the centroid of the triangle
PQR.
26. If 0 is real and z = cos 0+j sin 0, find
1
zn+ - and Z. — .
zn
Determine a, b, c, d such that
27 cos3 0sin3 0 = a sin 80 +b sin 60+ c sin 40+ d sin 20. (M.E.I.)
27. Show that the cubic equation
2x3 + 3x2 — 12x+ k = 0 (k real)
has three distinct real roots if and only if
— 20 < k < 7.
Solve the equation completely if k = —13.
28. Two real numbers are taken at random from the set of real numbers
{x e R: 0 x 1).
Find the probability:
(i) that the square of their sum is greater than 1;
(ii) that the sum of their squares is greater than 1.
29. Show that the point
(a(1— t 2)I(1+ t 2), 2btI(1+ t 2))
x2 y2
lies on the ellipse —+—=1
a2 b2
for all values of t.
Deduce that, if the line lx+ my = 1
is a tangent to the ellipse, then
12a2+ m2b2 = 1.
30. Prove that 12 +22 +...+n2 = i-n(n + 1) (2n + 1).
If s1 = 12 +42 +72 +... +(3n-2)2,
and s2 = 22 +52 + 82 + +(3n-1)2
show that s2— sl = 3n2.
and s1+s2 = 6n3 —n.
Hence, or otherwise, find s1and s2. (0 & C)
31. Prove that the medians of a triangle ABC are concurrent at a point G.
0 is an arbitrary point in the plane of triangle ABC and L, M, N are the mid-
points of BC, CA, AB. Prove that the lines through L, M, N drawn parallel to
OA, OB, OC respectively are concurrent at a point H.
Prove also that the line GH passes through 0 and determine the ratio OG: GH.
660
REVISION EXERCISE C
33. For what values of 0 in the interval 0° < b < 360° is the expression
6 cosy 0 — cos
less than one? Give your answers to the nearest tenth of a degree.
Answer the same question for the expression
6 sine 0— sin 0.
34. Express
x-4
(x — 3) (x — 2)
in partial fractions.
If x is so large that (1/x)3and higher powers of 1/x may be neglected, show that
the given expression is approximately equal to (x2 + x-1)/x 3.
35. Show that there is just one real value of the constant A for which the equations
Ax + y + z = 2x+ Ay + 3z = 4x— 2y+ Az = 0 have a common solution other than
x = y = z = 0. Find, for this value of A, the solution of these equations which
satisfies also the equation x +y+ z = 3. (M.E.1.)
36. A and B play a match of five games, each of which must be won or lost. In
each of the first three games the probability that A will win is 2/3 and in the
remaining two games the probability is 3/4. Find:
(i) the generating function for the probability that A wins r games;
(ii) the probability that A will win the match;
(iii) the average number of games that A will win. (0 and C modified)
38. Show that the points 0(0, 0, 0), A(0, 2, 2), B(2, 0, 2), C(2, 2, 0) are the
vertices of a regular tetrahedron. (over)
10 PPMII 661
REVISION EXERCISE C
47. Show that there are two complex numbers z such that
2 — ji = 1
and arg z = 1.7r
and find their moduli.
u 1\ mu ( xn )
52. If A = (1 3) prove that, for large n, xn YnV3
1 1 ' kl/' vn
and interpret your result geometrically.
53. Prove that, for any positive integer r,
tan (2r-10) =- cot (27-10) —2 cot (2r0).
n
Hence evaluate E 2?-1tan (2r-10).
r=0
54. Two players stake £1 each in a game in which eight counters are tossed on
the ground. The counters are painted black on one side and white on the other.
If the number of blacks showing is odd the thrower wins the other stake, if all
counters are of the same colour he wins a double stake, and otherwise he loses
his stake. Calculate the expected gain of the thrower and state what assumptions
you make. (M.E.I.)
55. Prove that n
2=-,}n(n+1) (2n +1).
r=1 r
The x axis from x = 1 to x = 2 is divided into n equal intervals by the points
n 1
(1, 0), (1 + , 0) , (1+-
2 , 0) , (1 + , 0 , (2, 0).
n
Rectangles with sides parallel to the axes are drawn on these/intervals
with one
vertex on the curve y = x2and the rest of each rectangle between the curve and
the x axis. Show that the sum of the areas of these rectangles is
1+ 1\ 4 1
k !if (3 6n)
Show further that as n increases this sum increases but remains less than
56. Show that, if the roots a and fi of the quadratic
ax2+(b+c)x+d= 0
are unequal, the equation
aunun±i+ bun+ cun+i+ d = 0
un+1— a —a
can be put in the form
Uss-a — fi us — ft
where A is given by the equation
A +1 a
b—c
A-1(a — Q) = •
57. If alb —1 is a small positive quantity not greater than 10-N, show that
a\-k 3a+ b
( 1)1 a + 3b
is not greater than 10-3N-4.
664
REVISION EXERCISE C
3x1+2x2+ x3 = c
or otherwise, find the inverse of the matrix
1 0 0
A= — 1 1 0 .
(3 2 1
1 4 —2
Given B = (0 1 3) ,
0 0 1
find B-1and also, by any other method, the inverse of AB. State a set of equa-
tions which this inverse (AB)--1would help you to solve. (M.E.I.)
67. An unbiased six-sided die is thrown. If a six is thrown the die is thrown a
second time and the scores for the two throws are added; find
(i) the chance of a score of more than eight;
(ii) the average score.
If the die is thrown again whenever a six has been thrown, find the average
number of throws and the average total score. (0 & C)
68. The position vectors relative to the origin 0 of the points A, B, P are respec-
tively a, b, p. X is the point of trisection of AP nearer A, and Y is the mid-point
of BP. Z is a vertex of the parallelogram PXZY. Write down, in terms of a,
b, p, the position vectors of X, Y, Z.
666
REVISION EXERCISE C
By comparing the sum of all the terms after the first three with a suitable in-
finite geometric series, prove that the magnitude of the error involved in neglect-
ing these terms is less than (5x3/81) (1 — x)-1 .
Show that, if x 0.1, this error is, in magnitude, less than 0.00007 and hence,
using the binomial expansion, evaluate (0.9)i correct to three decimal places.
76. Prove that the points with coordinates (0, —2, — +), (1, 1, 1) (5, 3, 4) and
(8, 2, 54) are the vertices of a trapezium in which the non-parallel sides are equal
in length. (S. M.P.)
77. The fixed point C and a variable point P have position vectors c, r respec-
tively in three-dimensional Euclidean space. Describe in geometrical terms the
locus I' defined by r.c = Irk I cl cos a,
where a is a given acute angle. If C is the point (a, b, c) and P the point (x, y, z)
express the equation of r in cartesian form and show that, for all real numbers A,
(x, y, z) e I' (Ax, Ay, Az) a F.
Taking the special case with C the point (4, 0, 3), cos a = 4/5, show that F
meets the plane z = 6 in a parabola and that the sphere Ir — cl = 3 touches the
plane at the focus of this parabola. (M.E.I.)
78. Obtain a quadratic approximation for the expression
V(1+ x)/(1 — 2x)
in the neighbourhood of x = 0.
Write down the equation of the tangent to the curve
y = 1/(1 + x)/(1 — 2x)
at the point (0, 1) and draw a sketch showing the relative position of the curve
and the tangent at this point.
79. ABCDA'B'C'D' is a parallelepiped, with faces ABCD, A'B'C' D' and paral-
lel edges AA', BB', CC', DD', etc. The mid points of CC', AB', B'C' , C'D' are
-
B is the matrix
k2
(5 6)
prove that Bn = 1) A +I
for any positive integer n.
668
REVISION EXERCISE C
(n — n n r)
82. Prove that (0 n).
86. Find and illustrate graphically the set of points (x, y) that satisfy simul-
taneously the inequalities
(x-1)3 ix-1,
y2 < (M.E.I.)
(1 a b
87. M=
1 02 b2)
By considering the matrix product M'M, or otherwise, prove that
2 a+a2 b+b2
a+ a2 a2+ a4 ab + a2b2 = 0.
b + b2 ab+ a2b2 b2+ b4
Evaluate the determinant
2 x+y x+y
x+y x2 + y2 2xy
x+ y 2xy x2 +y2
88. Describe geometrically the sets of points in the Argand diagram defined by
(i) iz-1I (ii) lz-11 = 21z-Fil;
(iii) lz-11+ lz+ji = 2; (iv) Re (z+ 1) = jz+jI.
89. The points P(ca, c/a) and Q (cfl, c//f) lie on the rectangular hyperbola
xy = c2, and (X, Y) is the mid-point of PQ. Find X and Y in terms of c, a and
ft. Write down a quadratic equation in t whose coefficients involve only c, X
and Y and whose roots are a and Q.
Determine what points (X, Y) can arise as mid-points of real chords of the
hyperbola. On a rough sketch shade in the regions they occupy. (0 & C)
90. Show how the matrix
(cos 20 sin 20\
A=
sin 20 — cos 28
can be associated with reflection in the line y = x tan O.
If B is the matrix associated similarly with reflection in the line y = xV3,
find the values of 0 for which AB = BA.
Describe how these values could be predicted by geometrical argument
(M.E.I.)
91. A rectangular box has faces OABC,O'A'B'C' and 00', etc. vertical. OA = 2
units, OC = 1 unit, 00' = 1 unit. Taking 0 as origin and axes Ox, Oy, Oz along
OA, OC, 00', find the equations of the planes OA'C', ACM, where M is the
mid-point of B'C'. Hence show that
(i) the line of intersection of these planes is parallel to the face OABC;
(ii) the cosine of the acute angle between planes is 7V6/18.
92. A, B are two fixed points with position vectors a, b relative to some origin 0.
X, Y are two variable points with position vectors x, y. Show that, if
x+y = a and x y -= b,
then 0, X, Y, A are coplanar and OB is perpendicular to this plane. Hence find
670
REVISION EXERCISE C
the most general solution, in terms of a, b and a scalar parameter, of the vector
equations
x+y a, x A y = b.
93. Prove that the matrix
(cos e - sin 0 0
P = sin 0 cos 0 0
0 0 1
is orthogonal and interpret P as the matrix of a linear transformation. Show that,
if x is very small, the matrix
1 -x 0
Q= x 1 0
0 0 1
represents a small rotation, if we ignore powers of x higher than the first. If SS
is a matrix all of whose elements are very small, and if
I+ SS
is orthogonal, prove that the matrix SS is skew-symmetric.
94. Two men, A and B, play a match consisting of separate games, the probability
of A's winning a game being p and the probability of B's winning being q.
(p+q is not necessarily equal to one.) They start the match with n counters each
and the winner of each game receives a counter from the loser; the first player to
win all 2n counters wins the match. If the probability that, when A has k counters
he will eventually win the match, is denoted by 11k, prove that
(p+q) uk = Puk+i+
What is the initial probability that A will win the match?
95. Prove that, if n is a positive integer, then
0 < (2 - V2)n < 1.
Use the Binomial Theorem to deduce the existence of an integer m such that
(2m -1) (2 -V2)„ < 2„ < 2m(2 - V2)„.
96. In the complex equation
zn.± +a,z+ 1 = 0
it is given that the complex numbers a„,., an-2, •.., a, satisfy
1, la„-21 1, ..., 1.
Show that any root of the equation in the complex plane must lie in the annular
region
< izi < 2. (C.S.)
97. The expressions E1, E2, E3are defined as follows:
E1 x+ y -2z -3,
E2 E 2x -3y -4z + 4,
E3 a 3x+ 5y + 3z- 5. (over)
671
REVISION EXERCISE C
If ,k, it are two numbers, explain the geometrical significance of the equation
E1+ AE2+ pEs = 0.
Find the equation of the plane which contains the line
(x+1) = —7(y-1) = —7(z-1)
and passes through the point of intersection of the planes El = 0, E2 = 0,
E3 = 0.
102. If a is one of the complex fifth roots of unity, and x = a — a4, show that
x4 + 5x2+ 5 = 0.
Express the other roots of this equation in terms of a. (0 &
103. The tangent to the cubic curve y = ax3 at Pi(xi, yi) meets the curve again
at P2(x2, y2). (P1is not at the origin.) The tangent at P2 meets the curve again at
P3(x3, y3) and so on. Show that x1, x2, x3, ... are in geometric progression.
The tangent at P1meets the perpendicular from P3 to the x axis at P. Show that
P1is the mid-point of P2P and that the locus of P is another cubic curve. (M.E.I.)
104. S stands for three-dimensional space in which points are specified in the
usual way by coordinates (x, y, z), and P stands for the plane z = 0 in which
points are specified by just two coordinates (x, y). A transformation
(xj) .÷ (0 1 "
1 0 23) (yx
z)
.
maps points of S onto points of P; and a transformation
T2 (;)f
— 22 —1 (;)
—1 —1
maps points of P onto points of S. Show that for T1any given image point in
P is associated with a set of points of S, and describe this set; and that for Tx
the range is a proper subset of S1, to be specified.
Show that for both transformations the line joining a point and its image is in
a fixed direction. Hence describe in geometrical terms the transformations
71T2 and T2 T1. • (S.M.P.)
where A, B, C are positive real numbers and a > b > c lie one between a and b
and one between b and c.
If A is small compared with B and C show that one of the roots is approxi-
mately
A (b — a) (c — a)
a+
B(c — a) + C(b — a)•
673
REVISION EXERCISE C
108. A square matrix is called 'magic' if there exists a number k such that the
sum of the elements of each row is k and the sum of the elements of each column
is k. Prove that, if A is a non-singular magic matrix, so also is adj A.
109. X, Y and (n-2) other players play a game in which each player has an
equal chance of winning each round. The game is won by winning 8 consecutive
rounds of the game. Let
qk = Pr (X wins)Y has won the last k times),
rk = Pr (X winsIX has won the last k times)
for k = 1, 2, ..., 7.
Show that
(n-l)qk +rk = 1,
1
rk =- k-Fi+ (1--) qi, for k = 1, 2, ..., 6,
n
and find the corresponding result for k = 7.
Hence show that the chance that X will win the game is (n7 -1)/(n8 - 1)
if Y has just won one round but did not win the round before that. (M.E.I.)
where j = ,s/( — 1) and a, b, c, d are real numbers not all 0. Show that H con-
tains a subset representing the complex numbers and prove that each element of
H has an inverse in H.
The matrix X is expressed in the form
X = al+ cJ + bK+ dL,
where
1=11 0 j_ (0 — 1) K = tj 0\ , L = (0. j\
ko 1) ' ki ko —,) kio;
Prove that J2 = K2 = L2 = —I and that JK = —KJ = L.
Given that KL = — LK = J and 14 = — JL = K, evaluate XY, where
la' +61 — c' + dj\
Y—
+ d'j —b'jj
by multiplying the expressions for X and Y in terms of I, J, K, L. (M.E.I.)
112. X and Y are independent Poisson variates from distributions with para-
meters it and A respectively. Find
Pr (X = riX+ Y = n)
for r = 0,1, ...,n.
The telephone exchange in a large firm handles both internal and external
calls. It is desired to investigate the distribution of internal calls and equipment
is available for recording the number of internal calls in each set of 10 calls. Cal-
culate the probabilities of 0, 1, 2 and 3 internal calls in sets of 10 if the average
number of calls per set is 7. What is the most likely number of internal calls?
(M.E.I.)
113. Show that if tabular values are given for x = 0, 1, 2, 3 of the function
f(x) E a+ bx+ cx(x — 1) + dx(x — 1) (x — 2),
then the values of b, c, d can be obtained from the differences derived from the
table, e.g. b from f(1)—f(0).
Given the following table for a cubic polynomial g(x), find and check the
values of a, b, c, d required to express it in the form given above; and hence or
otherwise evaluate g(5), g(6), g(7) and g(8),
x 0 1 2 3 4
g(x) 87.25 89.37 91.49 96.85 108.69 (M.E.I.)
114. Three bags each contain N tickets numbered 1 to N. Three tickets are
drawn, one from each bag; find the probability that the sum of the numbers
obtained is 2N.
115. Two circles, centres A and B and radii one unit, are drawn respectively in the
planes r .n = p„ nu, = p2. Prove that they lie on the surface of a sphere if and
only if
(a— b) .(nl A n2) = 0 and (pi — ni.b)2= (p2—n2.a)2.
675
REVISION EXERCISE C
prove that the mid-points of BC, DE, FA are the vertices of an equilateral tri-
angle.
117. The region E0is defined by
ax2+ by2 < 1,
where a > 0 and b > 0. Without assuming any properties of conics, prove that,
if pi, P2 are the position vectors of points of E0, then
(i) — E E0, Ypi+ P2) e
(iii) 2p1 e E, where E is the region ax2+bya 4.
A system of elliptical discs congruent to E0is obtained by translating the centre
0 of E, to all points g whose coordinates are integers. Prove that, if the discs with
centres at g1, g2overlap, then
(iv) for a common t, t E E0 and t g, a E0,
—
Deduce that if E contains no point, other than the origin, with integral co-
efficients, then no members of the system S overlap. (M.E.I.)
118. The log-normal distribution is a probability distribution given by
1
f(x) = exp [ — {ln (x/p)}2/2x]
x.\/(27ra)
for 0 < x < oo. By using the transformation x = pea and comparing with the
normal distribution show that:
(i) f f(x) dx = 1, (ii) f xf(x) dx = pea, (iii) f x2f(x) dx = p2e2ce.
Deduce the values of the mean and variance of the distribution. (0 & C)
127. Let M denote the set of all 3 x 3 real matrices and define an operation *
between members of 931 as follows:
X * Y = X+Y—XY.
We define a matrix X to be quasi-regular if we can find a matrix M e 991 such that
X*M=M*X= 0.
Prove the following results:
(i) 0 is quasi-regular.
(ii) If X is non-singular and quasi-regular, so also is X-1.
(iii) If X, Y are both quasi-regular, then X * Y is quasi-regular.
(iv) If X, Y, Z, are quasi-regular, then (X * Y) * Z = X * (Y * Z).
(v) If Xn = 0 for some positive integer n, then X is quasi-regular.
128. The polynomial P(x) is defined by
P(x + 1) .m4 P(x)+ (1+ x)",
where n is a positive integer. If P(1) = 1, find P(0) and P(— 1).
Prove that:
(i) P(— x) P(— x —0+ (— Xn;
P(x) (— On-1P(— x-1).
Deduce that, if n is even, then the sum of the nth powers of the first k natural
numbers is a polynomial in k, with integral coefficients, which is divisible by
k(k +1) (2k + 1).
What can you say about the sum of the nth powers of the first k natural
numbers in the case when n is odd?
129. In the tetrahedron ABCD, the perpendiculars from A, B to the opposite
faces intersect; prove that the edges AB, CD are perpendicular.
Prove also that the perpendiculars from C, D to the opposite faces intersect.
130. The parametric vector equation of a line 1 through the origin in three-dimen-
sional Euclidean space is
r = tk,
where k is a constant unit vector and t denotes distance measured along I from
the origin. A point P has position vector s. Find the position vector of the re-
flection of P in 1, i.e. of the point Q such that PQ is bisected at right angles by I.
If r = tki (i = 1, 2,) are two distinct lines through the origin, and
Si (i = 1, 2) are the operations of reflection with respect to these lines, prove
that 52S1 = S1S2if and only if the two lines are perpendicular. (C.S.)
678
REVISION EXERCISE C
The following books may usefully be added to the more comprehensive biblio-
graphy of Volume 1.
Armitage, J. V. and Griffiths, H. B. A Companion to Advanced Mathematics
(Cambridge). Surveys the underlying concepts of modern algebra and analysis:
an exacting book for sixth-formers but could be read with profit by a student
embarking upon a university course. Contains an excellent selection of
questions.
Brand, T. and Sherlock, A. Matrices: Pure and Applied (Arnold). Illustrates
many of the interesting applications of matrix theory.
Budden, F. J. Complex Numbers and their Applications (Longmans). An enter-
taining book illustrating applications of complex numbers in both pure and
applied mathematics.
Cundy, H. M. and Rollett, A. P. Mathematical Models (Oxford). Gives instruc-
tions for making many models. Full of stimulating ideas: very strongly recom-
mended.
Durran, J. H. Statistics and Probability (Cambridge). One of the best intro-
ductory texts available; contains an outstanding selection of examples: very
strongly recommended.
Hardy, G. H. A Mathematician's Apology (Cambridge). A leading mathematician
writes about what the subject means to him.
Hardy, G. H. and Wright, E. M. An Introduction to the Theory of Numbers
(Oxford). A famous work on one of the most fascinating of all mathematical
topics—the properties of the integers and their generalizations. Although the
book goes well beyond school level it is charmingly written and contains much
that will appeal to an intelligent sixth-former.
Henrici, P. Elements of Numerical Analysis (Wiley). A very good book on numeri-
cal methods, although rather advanced for sixth-formers.
Hunter, J. Number Theory (Oliver and Boyd). A more modern approach to the
subject than in Hardy and Wright: the text is somewhat exacting as it is
written for university students. Contains some good examples.
Kemeny, J. G. and Snell, J. L. Finite Markov Chains (van Nostrand). Applica-
tions of matrix theory to probability. An interesting but not always easy book.
Kline, M. and others. Mathematics in The Modern World (Freeman). Readings
from 'Scientific American'. Contains a series of fascinating essays on a wide
variety of subjects.
Maynard Smith, J. Mathematical Ideas in Biology (Cambridge). Among much
other interesting material there is an account of the applications of probability
to genetics. The same subject is dealt with in greater detail in D. S. Falconer's
Quantitative Genetics (Oliver and Boyd).
Mendelson, B. Introduction to Topology (Blackie). Perhaps the most readable
introduction available to topology, an important branch of modern mathe-
matics.
680
BIBLIOGRAPHY
681
Answers
CHAPTER 17
PAGE
381 Ex. 3. (i) -1+2j; (ii) 1 -4j; (iii) 7+ 4j; (iv) -8+6j;
(v) -26-18j.
382 Ex. 5. (i) -5+18j; (ii) -99+20j; (iii) 27+36j.
Ex. 6. (i) (2 - j)/5 ; (ii) (5 + j)/2; (iii) (4 - 3j)/5.
383 Ex. 8. (i) 3, - 2; (ii) 5, -12; (iii)
Ex. 9. b, 0.
Ex. 10. (iv) 8-j; (v) (2-j)/5; (vi) 11-5j; (vii) 8-8j.
Ex. 11. R.
Ex. 12. (i) 5; (ii) 2; (iii) 1/J2; (iv) 1; (v) 2 sin ie.
Exercise 17(a)
1. (i) 7+6j; (ii) 9+17j; (iii) 4+3j; (iv) 11+6j; (v) 11+22j;
(vi) 1 +32j; (vii) 0; (viii) (3 - j)/5; (ix) -72+368j;
(x) - 6 - 12j ; (xi) (7 + j)/50; (xii) - (11 + 29j)/37.
2. cos (01-02)+j sin (01-02); cos 201+j sin 201;
cos 2(0i + 02)+ j sin 2(01 + 02).
3. (i) 2 j; (ii) (1 j)/2; (iii) (5 ± jA/3)/2; (iv) ± 1 - j.
384 4. 3-2j; (i) (1 +.0/V2; (ii) 2 5. 3+2j, 1-j.
6. (i) --f; (ii) 2-j, 2+j.
7.1-cos 0 +j sin 0, (1 +j cot -1-0)/2. 10. z, pure imaginary.
12. (i) x2- 4x + 5 = 0; (ii) x2 - 4x + 13 = 0; (iii) x2 - 4x+ 7 = 0;
(iv) x2 - 3(1 + j)x+ 5j = 0; (v) x2 - (5+ j)x+4(1 + j) = 0.
13. (i) (x-1-D (x - 1 +D; (ii) + 2jy) (x-2jy);
(iii) (x+ j A/3 y) (x - j y); (iv) (jx -1) (x+1 - j) ;
(v) (jx - 2y) (x+ 2jy); (vi) (x + 1) (2x- 1 - ji/3) (2x - 1 + jA/3)/4;
(vii) II(x ± 1/A/2 ±j/V2).
14. 2, 3; (i) 2; (ii) x + 1.
15. 1, j, -1, -j; 1, 1+j, j, 0 according as n = 4m, 4m + 1, 4m + 2 or
4m+3.
385 17. -22.
388 Ex. 21. (i) 2, in; (ii) in; (iii) 1, n; (iv) 1, in; (v) 1, -}n;
(vi) 2, -in; (vii) 2, -*71; (viii) isec al, a;
683
ANSWERS
PAGE
388 (ix) 1sec al, (Or -2a)/2; (x) 2 sin la, (a-77)/21
(in the last three parts the principal argument depends on a).
390 Ex. 25. (/3 - 1) + j(A/3 + 1), cos 75° = (A/3-1)/24, etc.
Exercise 17(b)
396 1. (i) 5, -53°; (ii) A/13, 56'; (iii) A/5, -117°; (iv) A/5/2, 63°; (v) A/29, 112°;
(vi) -153'; (vii) 13, -23°; (viii) 41, -103°.
2. (i) cos 30 + j sin 30; (ii) cos (0+0)+j sin (0 + cb); (iii) j; (iv) (1 +j) V2;
(v) j; (vi) (1 - j A/3)/2; (vii) j;
(viii) 4 cos 40 sin Isb(sin (0 + cb)/2 - j cos (0 + .75)/2).
3. (i) (7r +295- 28)/2; (ii) (-37+20-20)/2.
7. Circle, centre 2+j, radius 1. 12. -j.
397 13. (i) (1 +j)/2; (ii) (1 ± j A /3)/2; (iii) j, 1- j; (iv) -1 +
14. 9+6j. 16. (i) (ii) semicircle, radius 1.
17. (1 -j tan 40)/2, (1 +j cot 40)/2. 18. Square.
Miscellaneous Exercise 17
1. (,13 -j) (1 +j),2,12, 2. 8, 2.
3. 1 ± A/3 +j(5 ± A/3).
398 5. -1+2j, (x - 1)2 + (y - 4)2 = f.
399 14. (i) izi = A/7; (ii) 2+j, 3+2j.
400 15. r = 4, 0 = 147.5°.
CHAPTER 18
402 Ex. 1. xs- 9x2 - 3x + 4, 2.
Ex. 2. 25086278.
403 Ex. 3. 2600.
Ex. 4. y = 1957x-7373.
Ex. 5. 2412, 1957, 1190.
405 Ex. 7. - 26, -8.
Ex. 9. -0.099, 0.357.
407 Ex. 13. 2x2 - 5x+ 3.
Exercise 18(a)
1. (i) x2-3x - 5; 6; (ii) x3 - 9x +4; 2;
(iii) 3x3 - 5x2-2x-7; -10; (iv) x3 + 4x2 + 16x+ 59; 231;
(v) 7x4 -35x3 +175x2 - 875x +4372; - 21865.
684
ANSWERS
PAGE
408 2. (i) x2 - 3x - 7; 7; (ii) x3 - x +3; 0; (iii) x4- 4x3 - x2 + 3x+ 5; 2;
(iv) (16x3 - 44x2 - 14x + 11)/8; t;
(v) (9x4 - 63x2 + 6x -40)/9; V.
3. (i) 159.1; (ii) 3F4; (iii) 283.3.
4. (i) (x- 1)3 + 4(x - 1)2 + 7(x-1)+5;
(ii) 2(x - 2)3 -5(x- 2)2 -14(x- 2) + 8;
(iii) (x +3)3 - 2(x +3)2 - 21(x + 3) + 50;
(iv) (x+ 2)4 -8(x+ 2)3 + 24(x+ 2)2 - 32(x + 2) + 15.
5. (i) x3 - jx2 +4jx+ 2; 1 -j;
(ii) x3 + (1 + j)x2 - (2- j)x + (1 + 2j); -j.
.
6. (i) 2.6; (ii) 3.0. 7. (x + 4)4 + 4(x +4)2; minimum; -4, -4, - 4 ± 2j.
8. x5 3.xs + 2x +1; 1.
- 9. A = 0, B = 1. 10. A = A, B = 4A -17.
11. - 6, 139. 12. 86, 68.
409 13. x2 - 8x-8. 14. 2x3-4x2 - x+ 6.
15. n+15n(n- 1)/2+ 25n(n- 1) (n- 2)/3 + 5n(n -1) (n- 2) (n-3)/2
+ n(n - 1) (n - 2) (n-3) (n-4)/5.
Ex. 14. e.g. a = 5, b = - 2. Ex. 15. a = 3, b = 3, c = -1.
411 Ex. 17. (i) 4/(x-1)-3/(x + 2); (ii) 3/2(x- 3)-1/2(x + 3);
(iii) 1/(x- a)- 11(x + a); (iv) 1/(x- a)+ 11(x + a).
412 Ex. 18. (i) In V1.5; (ii) In (f).
Ex. 19. (i) 4(x - 2)-2- 6(x - 3)-2;
(ii) {(x+ 1)-2 - 36(x - 6)-2}/7.
415 Ex. 21. (i) 2I(x - 2)-3/(x - 2)2 - 2/(x +1);
(ii) 1/(x+ 1)-4/(x+ 1)2 - 1/(2x -F 1).
Ex. 22. (i) 1/(x - 1)- x/(x2+ 2);
(ii) 2/3(x - 1)+ (x-1)13(x2+ x+ 1).
Ex. 24. (i) 1/(x - 1)- 1/(x2 - 2);
(ii) 1/(x -1)- {1/(x-8/2)- 1/(x+ V2)}/24.
Exercise 18(b)
1. (i) 2/(3- x)+ 3/(1 - 2x); (ii) 2/x- 3/(x+ 2);
(iii) 7 1(x+ 3)- 5/(2x- 1); (iv) 1/2(1 - x)- 3/2(1 +x);
(v) 2+3/x- 5I(x +1); (vi) 1/(x-1)- 3/(x + 2) + 2/(x-3);
(vii) 1/x-3/(1-x)- 5/(1 +x); (viii) - x + 3/(x- 2)- 7 1(x + 2);
(ix) x2 + 5x+ 19 + 81/(x-3)-16/(x- 2);
(x) 1/x2+1/x- 2/(1 +2x);
(xi) 1/(1 - 2x)2 - 1/(1 - 2x)+ 1/(2 + x); (xii) x/(2 + x2)- 1/(1 + x);
(xiii) (2x + 1)/3(1 - 2x + 4x2)- 1/3(1 + 2x); (xiv) 4/(x - 3)2 +1/(x- 2);
(xv) 1 - 1/3(x + 1) + (x - 4)/3(x2 + 2);
(xvi) 1/(x - 2) - 1/x4- 243 - 2/x2 - 1/x.
2. A = - 13, B =13, C = -11, D = 27.
3. 2/(x- 1)4 + 6/(x-1)3 +4(x-1)2 +1/(x-1)-1/(x+ 4).
685
ANSWERS
PAGE
416 4.1 -1/(n + 2); (3n2 + 18n + 26)/3(n+ 2) (n+ 3) (n+ 4).
5. (i) 1/3(1+x)- (x-2)/3(1 - x + x2);
(ii) 1/3(1 + x)- (1 +V3j)/3(2x- 1 - V3j)- (1 - V3j)/3(2x - 1 + V3j).
6. 1/(x- 2)-1/(x-1); -1-- 1/2(2y -1)+ 1/(y - 1);
- 1/8(4x - 1)+ 1/(2x - 1).
Ex. 25. 0.9802.
Exercise 18(c)
420 1. (i) 1 +3x + 9x2 + 27x3;(ii) + 2x2 + 1- .x3;
(iii) 1 + 4x +12x2 + 32x2; (iv) 1 + 2x2;
(v) 8+6x+ 6X2 '+ *X3•
2. (i) 1 + x+ lx2 +1x2; (ii) 1 - 2x - 2x2- 4x3; (iii) 2 -4x - 512X3;
(iv) 4+16x+ thX2 + 2 04 8 X3 ; (V) 0-+ +*.x2+
3. (i) 1 + tx2 + tfX3 ; (ii) 1 - 8x + 40x2 - 160.0;
(iii) 1+1x+ .x2 +Hx3; (iv) *+*4.-x + + 35 x3.
(v) 2 -11-2,x - ThX2 20 3 6X3.
4. (i) 1 + 3x + 7x2+ 15x3; (ii) + -276x2 + 6.7C3
- ± 2x2-16x3, (iv) 3x- 5x2 + 7x3;
(v) 1 +8x+ 26x2 +64x3.
5. (i) 1 -ix+ 1eX2;
(ii) 1 + x + -}X2 ; 1 + +81x2 ; 1 - + 1 .X2 ;
(v) 1 - 3x +12x2.
Miscellaneous Exercise 18
1. A = A, B = arbitrary, C = -A-1.
3. 1, - (x + 1), 1/(x- 2)- (x + 1)/(x2 + x + 1).
422 4. 1/7(x - 2) - (x + 3)/7(x2 + x + 1); 1/9(x - 2) - 1/9(x + 1) - 1/3(x -F 1)2.
5. j i; A-- (4n + 5)/8(2n+ 1) (2n+ 3).
,
CHAPTER 19
PAGE
423 Ex. 2. -29.
424 Ex. 4. j/29.
Ex. 6. 2t(3-10t2 + 3t4) (1- 15t2 + 15t4- 0).
425 Ex. 8. (10- 15 cos 20+6 cos 40-cos 60)/32.
Ex. 9. 57r/32.
426 Ex. 10. (1 - x cos 0 - xn cos nO + xn+1cos (n-1)0) (1-2x cos 0+x2)-1
Ex. 11. x sin 0(1 - 2x cos 0+ x2)-1
Exercise 19(a)
1. (i) - 1; (ii) -j; (iii) j; (iv) (J3 +j)/2; (v) 1;
(vi) 16( - A/3+ j)/9,s/3; (vii) -(1 + j V3)/32;
(viii) sec40(cos 40-j sin 40)/16.
2. (i) -1; (ii) 16; (iii) - 29(1 + iN,/3);
(iv) cos (n+ m)0 + j sin (n+ m)0.
3. (i) 1-8 sine 0+8 sin4 0;
(ii) 5 sin 0 -20 sine 0+ 16 sin' 0.
4. (i) 1- 8 cos2 0 + 8 cos4 0;
(ii) 16 cos' 0 -20 cos3 0+5 cos 0;
(iii) 6 cos 0- 32 cos3 0+32 cos' 0.
5. 441 - t2) (1 - 6i2 + 0)-1.
6. cos {(2k+1)114}, k = 0 to 6; cos (2kir/7), k = 0 to 6.
7. (i) (cos 50 +5 cos 30+10 cos 0)/16;
(ii) (cos 40-4 cos 20+3)/8;
(iii) (cos 70 +7 cos 50 + 21 cos 30+35 cos 0)/64.
8. (i) (sin 50 -5 sin 30+10 sin 0)/16;
(ii) (35 sin 0 -21 sin 30+7 sin 50 - sin 70)/64;
(iii) (2 sin 20-sin 40)/8.
9. (sin 90-sin 70-4 sin 50+4 sin 30+6 sin 0)/256.
10. (i) 37r/16; (ii) 37r/512.
11. (i) (cos 0+ cos nO-cos (n +1)0 - 1) (2-2 cos 0)-1;
(ii) (sin 0 + sin nO-sin (n+ 1)0) (2-2 cos 0)-1.
12. (cos 0+2m}1cos nO - 2- 2n cos (n +1)0) (5- 4 cos 0)-1.
427 13. 2n cos"' (0/2) cos (n012).
14.4 cosec2(0/2) {sin 0+ (n+ 1) sin (n-1)0 - n sin n0}.
15. (sin2 0- 0 sin (rz +1)0+ sinn+2 0 sin nO) (1 + sin2 0-sin 20)-1.
16. (Et,- Et2 t3 t4) (1 + t, t2 t3 Et, t2)-1•
17. (sin 30 sin3 0- sin (2n + 3)0 sin2n+3 0-sin6 0+ sin (2n+ 1)0 sin2"-F6 0)
x (1- 2 sin2 0 cos 20+ sin' 0)-1.
687
ANSWERS
PAGE
427 18. a = b= c= kff ±
428 Ex. 14. (i) ± 1, ± j; (ii) ± 2, ± 2j; (iii) ± (1 ± j)/A/2.
429 Ex. 18. -j, ( ± V3 + j)/2.
Exercise 19(b)
430 1. (i) 2, (-1 ± V3 j); (ii) -1, (1 ± V3 j)/2; (iii) j, (± V3 - j)/2;
(iv) 1 +j, - (A/3 + 1) + (V3 - I) I}/2, {(A/3 - 1)- (A/3 + 1) j}/2.
2. (i) 2-j, (ii) 1.3 + 0.79j; (iii) -0.28- I.8j; (iv) 1.4-1.0j.
3. 1 + j, - (V3 + 1) + (V3 - 1) j1/2, {(A/3 - 1)- (A/3 + 1) j}/2;
cos 7r/12 = (1 + V3)/2A/2, sin 7r/12 = (V3 - 1)/2A/2.
4. 0.088j.
5. (i) cos 3 -j sin 3;
(ii) cos (20-7016+j sin (20-70/6;
(iii) cos 20-j sin 26; (iv) coseci cjs (ir -20)/8;
(v) cjs (40 - 7018.
7. (± A/3 j)/2. 8. -1, (± 1 ± V3j)/2.
9. ij, -coj, - w2j where (4) is a complex cube root of one.
10. tan krr15, k = 0 to 4.
11. cos 0(1-sin 0)-2, 0 = 2k7r In.
431 15. (a+ b+ c) Poll+ w2c) (a+ (o2b+ wc),
(a+ b + c) (a2+ b2+ c2- bc- ca- ab);
(a+b+c+d)(a-b+c-d)(a+jb-c-jd)(a-jb-c+jd)
433 Ex. 25. (i) jir; (ii) -4j7r; (iii) e-n; (iv) el".
Exercise 19(c)
1. !ar- bsl, 0. 2. (i) +j)/A/2; (ii) e-nr.
3. (i) In 2-jar/6; (ii) ein(cos In 2+j sin In 2).
4. Moves along real axis from en to e-". 7. 2c2cosh v (cos u+ cosh v)-1.
Miscellaneous Exercise 19
434 1. cos (2k + 1)7r/12 +j sin (2k +1)7r/12, k = 0, 2, 3, 5, 6, 8, 9, 11.
3. 2-71cos 2n0(1 + cos 0)-n.
4. (i) 2 cjs (g14), 8 cjs (37r/4); (ii) 2 cjs (117r/12), 2 cjs (- 577/12).
5. ± (-0.61+1.2j), ± (-0.61- I.2j).
6. Eac- bd+(a2 b2)1 (c2+d91}.
9. (i) cos {a +1(n- i)/7} sin 10 cosec I/5';
(ii) sin {a + 1)fl} sin infl cosec
10. 357r/256.
688
ANSWERS
PAGE
435 11. (i) 2 cos 1-0 cjs -1-e; (ii) 2 sin -1-19 cjs (0-02;
(iii) V(cot 1-6+) cjs
13. cb = (2r+ 1)7/15 14. tan {4k- 3)74/4n.
15. (V3+1)/24, 1/(V6 + V2).
16. - (3*+ - (3*(o+ 3*co2), - (3*(A)2 + Ad).
437 23. 1-11(n + 1); A = -C = all-a)-2, B = (1 -a)-1.
24. (x-1) 11 (x2 -2x cos 21or 1(2n + 1) +1).
k=1
CHAPTER 20
439 Ex. 2. (2 + 3j)z + (2 - 3j)z* = 13.
Ex. 4. (i) 6; (ii) - 3j.
Ex. 6. zz* - (1 +2j)z- (1 -2j)z*- 4 = 0.
Ex. 7. (i) 1+3j, 2; (ii) 1-j, 1.
Exercise 20
447 1. (i) lw - 31 = 1; (ii) 1w- = 1; (iii) 1w1 = 2; (iv) lw1 = 1;
(v)iwi= V2.
3. -1, ±j. 4. 1+j; 2+j, -1+j.
448 11. z(1 -j)+ z*(1 + j) = 2. Interior of `shark's fin' bounded by real axis,
lz-li = 1. 1z-,1 ajj = V6/4.
12. Circle, 2j/3, 1. 13. Line. 14. Line. 15. Point j.
16. Circle, 2, ,15. 17. Line. 19. arg = - O.
(z - b
449 21. w = 3+j-1/z. 25. Arcs of three circles through origin, centres
j/2, 1/2b, j/2(a - 1) and imaginary axis.
26. (i) 3ww*- 2(w + w*)+ 1 = 0; (ii) w+ w* =
27. w = (z - j)(jz
CHAPTER 21
450 Ex. 1. 1+ j, (-1+7j)/5.
454 Ex. 2. (i) -1 T?; (iii) V47/2 ; (iv) 4-
Ex. 3. (i) 3z2- 10z + 12 = 0; (ii) 3z2 + z+ 1 = 0;
(iii) 9z2- 7z + 9 -= 0.
Exercise 21(a)
455 1. 3, -7; (ii) 2, 51; (iii) (1 - j)/2, -j; (iv) 4-3j, -1;
(v) (2- 6j)/5, - (1 + 2j)/5.
689
ANSWERS
PAGE
455 2. (i) z2 -z - 6 = 0; (ii) 6z2 -z- 2 = 0; (iii) z2 - 6z+ 4 = 0;
(iv) z2 - (5 + A/2)z+ (2+ A/2) = 0; (v) z2 - 6z+ 25 = 0;
(vi) z2 -z+ 1 = 0; (vii) z2 - (2+ j)z+ (3 +j) = 0;
(viii) z2 - (,/2+ 1) (1 - j)z- 3j = 0.
3. (i) 2.8, - F3; (ii) 0.75 ± 1.7j; (iii) 0.42+0.86j, - F42- 1.86j.
4. (i) 3, -9, 27, -1, -972; (ii) 1, -s, -W;
(iii) j, - 1 + j, 1- 2j, (1 - 5 +1.
5. (i) (ii) (iii) V113/2; (iv) 9013/4.
6. (i) (3k2 + 2k-7)/3; (ii) W; (iii) -4.
7. (i) 2(a2 + 8ab + 862) ; (ii) 2(a+ 2b)/a2 ; (iii) 41{b(a + b)}i I.
456 8. - a2mcl(b2+ a2m2), b2c1(b2+ a2m2).
9. (i) 2z2+ 3z - 6 = 0; (ii) z2 - 5z- 350 = 0; (iii) 7z2 + z- 2 = 0;
(iv) 4z2 - 29z+ 49 = 0.
10. (i) 2z2 -2(1 - 3j)z- 5 - 9j = 0; (ii) z2 - 2jz+ 5- j = 0;
(iii) z2 - (2+ 14j)z -12+ 5j = 0.
11. (i) 4z2 +7 = 0; (ii) 4z2 - (10 + 10j)z+ 9j = 0;
(iii) 4z2 - 19z+ 44 = 0.
12. ± 24j.
13. ± 3(1 +1).
14. (i) (4 + 5j)z2-(3 + j)z- (2 - j) = 0;
(ii) (2 + j)z2 + (3 - j)z - 4 + 5j = 0.
19. (ca2)* +(c2a)* + b = 0.
20. b2 = 2ca.
Miscellaneous Exercise 21
459 1. (i) 11z2 - 8z- 400 = 0; (ii) z2 - 5(1 + j)z+ 47j = 0.
2. a2, ,92. 7. al/J, ft/a.
CHAPTER 22
461 Ex. 2. 2x - 3y + 13 = 0, (5t-1, 3t-2).
462 Ex. 5. b2x2 -a2y2 = a2b2.
463 Ex. 6. (-1, 3), (-4, 4), (- 2, 5), (-3, -4).
Exercise 22(a)
464 1. 4x2 +y2 = 4. 2. x3 = y2. 3. y = x2 -3x+ 3.
4. x2+ y2 = 2. 5. x3+ 313= axy. 6. x2+ y2 = x2y
.
8. (i) (-1, 4); (ii) (2, 0); (iii) (2, 8); (iv) (4, -1).
9. ±4-.
465 12. (- 2, 5).
690
ANSWERS
PAGE
466 Ex. 8. PSZ = 90°, SPM bisected by tangent.
467 Ex. 11. 4a, a, 3a.
Ex. 14. (7a/2, a), x = 5a/2; (i) (-11a/4, 0), x + 5a/4 = 0;
(ii) ( — a, a), y + 3a = 0.
468 Ex. 20. q = — p —21p.
Exercise 22(b)
472 1. (i) 4, (3, 2), (4, 2), x = 2; (ii) 2, (3, —1), (3, i-), 2y = —3;
—
(iv) 8x + 3y+ 30 = 0.
5. (4, — 2). 6. (-4, 6). 7. — i — i +,
473 8.
474 18. Same axis, focus; directrix is tangent at vertex of original parabola.
19. Parabola. 21. y2= 2ax. 22. Parabola, latus rectum a.
475 28. y = 2a, y2-2ax = a2. 30. y = 3x2 4x. —
31. y = 2x2 — x+ 2.
Exercise 22(c)
478 6. x = —3, y = 1. 7. (1, 1).
479 10. 2x— 8y 15c = 0.
Miscellaneous Exercise 22
2. Y/a, ( Y2 2aX)12a2; (b, 0), (b + a, 0).
—
CHAPTER 23
484 Ex. 1. 2 cjs (2k+ 1)/r/10, k = 0 to 4, i.e. IV(10+ 2V5)+ 1(A/5 1)j, etc.
- —
Ex. 2. 1, 2, — 2.
—
Exercise 23(a)
PAGE
488 1. x4- 5x3-4x2 + 16x - 8 = 0. 2. x4 - 3x3 + x2 + 7x 30 = 0.
-
Exercise 23(b)
494 1. (i) 14; (ii) 46; (iii) 162; (iv) 10; (v) 17; (vi) 24; (vii)
(viii) 22; (ix) -2; (x) 30.
2. (i) 16; (ii) -3; (iii) 9; (iv) 67; (v) 32; (vi) 48; (vii) 18.
3. (i) 2p2; (ii) (p4 +4pq2)/q4; (iii) 0.
4. -3, i,4.
5. -1,1. 6. q, , 1. 7. -2, 2-1, 5.
8. 1, -2, -3.
495 9. 2, 3, -4. 10. -1, 2, 3. 11. 3abc = a2d + 263.
12. -3,
13. (i) z3+ 2pz2+ p2z - q2 = 0; (ii) z3 - pz2- q2 = 0;
(iii) z3+ pz - q = 0; (iv) z3 + 4pz -8q = 0.
14. (i) z3 + 3qz2+ (3q2+ p3)z + q3 = 0;
(ii) z3 2pz2+p2z - q2 = 0.
-
692
ANSWERS
Exercise 23(c)
PAGE
502 1. -2, 1+j. 2. (3 ±ja/4-
3. 2 cos 7r/9, - 2 cos 470, -2 cos 27r/9.
4. cos a, cos (;n+ a), cos (;rr+a) where cos 3a - 3 •
5. 31+ 31, 6)31+ 031, 6)231+ (o31.
6. 2- 21+ 21, 2-6)21+6)21, 2-021+U)21.
7. 5+61+61, 5+ (o61+ 061, 5+ w61+ a)261.
8.4, - - -0.
503 9. 2j, j(1 ± V2).
-
Miscellaneous Exercise 23
3. + 8, T- 288.
4. 1. 5. -2 ± j, ± j/V2. 6. 84, 708.
7. 8. 1, (1, 1, 1), (4, -1, -D.
9. 3 roots if -81 < k ‘. 44.
505 10. - 1± J5.
11. q2(y + 1)3 +p3(y + 2) = 0.
CHAPTER 24
506 Ex. 7. + 3e.
508 Ex. 9. i+ 4j 3k.-
Exercise 24(a)
512 4. (i) - 2i -j + 5k; (ii) 8i - 16j+ 8k; (iii) i -j + k;
(iv) - 4j + 4k; (v) - bci+abk.
5. 0; coplanar with 0. 6. -k, j. -
PAGE
518 4. 0; lines intersect. 5. 2x + 3y— 5z — 21 = 0.
6. 7x-3y+z-16 = 0. 7. x+y-2z = 4.
519 8. (i) (4, —2, 3); (ii) i+ 2k; (iii) 6i-5j-3k;
(iv) 6(x-4) = —1(y +2) = —1(z-3).
10. x = =
11. r= c+ v{[bi A (81-0] A [b2 A (a2 — c)]}.
12. 2831(3ccfry). 13. a,I613.
CHAPTER 25
523 Ex. 2. = 0 < x < 100.
Ex. 3. 2 x 10-4,124-.
528 Ex. 8. I.
Exercise 25(a)
1.1, 2. 2, 0417. 3-
4. (i) (x —1)1 x for x 1; (ii) 1 -e-2X for x 0;
(iii) -1(1 — cos nx) for 0 ‘. x < 1.
529 5. 1, 1, 1, 6. t, I, 1, -1
3 92•
7. 0, 0, 0, (77.2— 8)/477.2. 8. 1IA, (In 2)/A, 0, 1/A2.
9. F(x) = 0, x2/4, (2x-1)/4, (6x— 2x2 — 5)/4, 1;
G(y) = 0, ay, 1,-(2,1y
, —1), etc.; g = dG/dy.
10. g(y) = 2y3, 0 < y < 1, 2y(2 —y2), 1 < y < A/2;
1, 1; 8(2/2-1)/15.
11. 97 ileu.
530 12. 1, —4, 6.
13. e-kto, e-skto, 6e-3ktou — e-ktor.
14. 1.5, 41, 0.15.
15. 4/7r, (2/77) In 2, 0.0785, 4-1, 0.854.
535 Ex. 12. -1-z2, —1+2z— lz2.
Ex. 13. Triangular.
Exercise 25(b)
545 1. -j-). 2. 1 — 1/,/(107r). 3. 4. 4. 169
5. 1,1. 6. 4. 7. 4r/77..
546 8. 2r2-16r2/7r2. 9. /2/4, 14/48.
10. a4{4n.— n2— (In 4)2}/477.2. 11. 1, 53, 160.
12. 0149. 13. 2.421, 2.641. 14. 0.004.
15. 0-154. 16. 219, 78%, a = 0156.
17. 10.825, 98, 0.271.
547 18. 502 g.
694
ANSWERS
Miscellaneous Exercise 25
PAGE
547 1. 8a2/3, 1A/2. 2. 18.6. 3. (1+1n 4)/4.
4. 1/(2V.0, 1,
548 5. 0.301, 0.092. 6. 1 + exp [- gx1(cr,13)].
7. 0.683. 8. 0.1336, 1.630, 4.98.
9. A, A, 2A. 10. 183p.
549 12. 13. 15. 1 2A-2.
17. #(1+y)-1, 2/7r,/(1-z2); 2A✓45; 2/rr, V{(77.2 -8)/2n2}.
18. A/Ca-A).
550 19. (i) p < 1/22.4; (ii) p = 0. 20. 1.
21. 4.471, 5.000, 5.029; 0.2280.
23. V65/3, 0-387.
CHAPTER 26
554 Ex. 6. Absolute: 21yieli, lei/2-N/Yi I , (I + I e21)/2A01+ Y2)-
Exercise 26(a)
559 1. 0.7513 +0.0006, 1.1383 + 0.0020.
2. (i) 22.36 ± 0.055; (ii) 88-75+0.35; (iii) 3.333 ± 0.014.
3. 0.31784. 4. 0.043685(5). 5. 5.95+ 0.10.
6. lel sect y. +Y2)1. 8. (i) 0.011; (ii) 0-034.
9. 0.39. 10. 0-23(1). 11. 2.15(4). 12. 1-15.
13. 1.94, -0.56, -F38. 14. F56, -2.56.
17. n even: no real roots; n odd: one real (negative) root.
Exercise 26(b)
569 1. F44. 2. 4.64. 3. 1-38. 4. 5.14, - 2.14.
5. 4.10. 6. 4.13 (5). 7. 2.30, -1-30.
8. 2.65, -1.42, -4.23. 9. 4.193.
10. 1.092, 1.931. 11. 0-8447. 12. 0.7920.
13. 0-60. 14.1- (5k+ 3)-1. 15. 0-877.
570 16. 0.838. 17. 3.591.
Miscellaneous Exercise 26
1. 1-532. 4. 60-3.
5. 0.88, 0.78, 0.69, 0.60, 0-50.
695
ANSWERS
PAGE
571 8. 3.316625. 9. 1.81713. 10. 2.8.
572 12. 1162(28). 13. K+112K3 1/4K4.
- 15. 0.567144.
CHAPTER 27
576 Ex. 7. (± 3, 0). Ex. 8. (2, — 2), (0, —2).
577 Ex. 10. 0.205.
Exercise 27(a)
585 1. (i) (ii) A/3/2; (iii) A/2/A/5. 2. 3, (0, ± 2), y = ± i.
3. (2± /3, —1). 4. 1/A/2, (-2, 1), (-4, 1).
586 5. (2, —1), (2, —3). 6. it. 7. x —2y = 7.
8. 4x+y = 6. 9. (-1, 2), (-2, 1).
13. Mid-point of Os.
587 22. a2y2 +b2x2 = 4x2y2. 23. S, b.
27. x 3y = ± 5A/2.
596 Ex. 37. Arc of a circle.
Exercise 27(b)
597 1. (i) (ii) A/3; (iii) A/(3/2). 2. 20/3, (± 2A/2, 0), A/2x T- 3 = 0.
3. 4i(6, — 1), (-4, —1), 5x — 21 = 0, 5x + 11 = 0.
4. VI, (-3 ± A/5, 1), x = —3 ±4/A/5.
5. (i) 3x ± 2y = 0; (ii) A/3x ±y = 0.
6. (i) x— 2y = —3, x+ 2y = 1; (ii) x± A/2y = A/2.
7. 4x2 —y2 = 7. 8. 9x2 — 4y2 — 18x — 24y— 32 = 0.
9. x+y-1 = 0. 10. x+3y = 4, -H.
11. 2x+y+3 = 0. 13. x2 —y2 = 8.
598 18. (1, 0), y-2A/2x+2(V2+ 1) = 0.
19. 5x-2y-9 = 0, 2x+ 5y — 50 = 0, (20, —125-).
600 Ex. 38. (i) Line parallel to 0 = 0. (ii) Circle. (iii) Line at a to 0 = 0.
(iv) Quadrant of annulus. (v) Interior of circle, centre (a, 0).
Ex. 39. (i) r(sin 0+ cos 0) = 1; (ii) r = 2 sin 0;
(iii) r2 -2r(sin 0+ cos 0)+1 = 0.
Miscellaneous Exercise 27
605 1. 1/A/2, (1, —3), (-2, —2).
606 3. b2/a2, x2+ y2— ax sec cc— by cosec a = 0.
608 15. 2x(xi — x2)+ 2AYI-3,2) = 4+34 -A.
16. iV(1+ 4e + 3e2). 17. Ellipse, parabola.
696
ANSWERS
CHAPTER 28
PAGE
613 Ex. 4. 3, -2.
Exercise 28(a)
2IN113 1/V10
618 1. 2 11 3x+ 2y = 0, 3x -y = 0.
-3/V13/ ' (3/,/10)'
Exercise 28(b)
-5/V33 ( 1/3V3 1/V3
628 1. 1, 2, 8; ( 2/V33 , 1/3V3 , 1/V3 .
2/V33 - 5/3V3 1/V3
(1 1 1
2. -2, 3, 0; A singular, 3 1 1 .
3 4 1
697
ANSWERS
PAGE
(2 -3 2
628 3. 1, -1, 2; det A = -2; 2A-1 = 4 -5 2 ;
4 -7 4
-1, 1,1; eigenvalues AT I.
(1/A/2) (0
4. 1, -1, -2; 0 , f), 1/V2 ,
1/V2 I 1/V2
(1 1 0)
0 2 1 . x+y-z = 0, y-z = 0.
1 2 1
1 -2 2
5. 1, 2, -3;( 12 -1 -2 .
2 2 1
629 6. y-z = 0, x-y = 0, 15x + 13y- 8y = 0.
16. i 16 20\
11; A +it, Att.
\4 -4/
i 370 -455
630 17. A if n odd, I if n even. 18.
k- 260 240/ '
10 9 23 4 -2 -3
20. A3 =( 5 9 14 , A-1 = 1 0 -1 .
9 5 19 -2 1 2
21. (171+ A - A2)/52.
2n}4-15 - 2n-F3- 4 . 3n +12 2n+4-4.3n- 12
22. ( 10.2n-10 -5.2"-2.3"+8 10.2"-2.3n-8 .
- 10.2n + 10 5.2n + 3n+1-8 -5.2"44 + 3n+l• + 8
631 24. 0, ±j tan V. 25. x+2y- 6z = 0;
cei. ce2 5
P =(fli A 3 , where at +2fli - 67z = 0, det P * 0.
Yi. Y2 2
632 28.1x = -fy = z, x = y = z.
CHAPTER 29
633 Ex. 1. (-a, 0).
636 Ex. 5. x±y = O.
Exercise 29(a)
637 3. (2- 01(2t + 1). 4. p+q+r = 0.
638 8. (4ak2, 4ak). 11. (at/(1 + t2), ate/(1 + t3)).
698
ANSWERS
PA GE
639 16. 8:1.
640 Ex. 7. Ends of diameter.
Ex. 8. (1, 2, —1), 4.
641 Ex. 10. (2, 2, —1), 3.
644 Ex. 12. Elliptic cylinder.
Exercise 29(b)
652 1. (i) x2+ y2+z2+2y = 0; (ii) x2+y2+z2-4x+6y+2z —11 = 0;
(iii) x2 + y2+ z2—2ax —2by-2cz = 0.
2. (i) (3, 1, —1), 4; (ii) (-2, 4, 1), 9; (iii) (a, b, —c), la— b + cl.
3. (i) x2+ y2+ z2— 4x —2y +2z — 32 = 0;
(ii) x2 + y2 + z2— 2x — 4z — 16 = 0;
(iii) x2 + y2+ z2— ax— by— cz = 0.
4. (i) (4, 1, 3), A/6; (ii) (5, —1, 2), 9.
5. 4-A/{2d2 — (a— b)2}. AB is diameter.
6. Centroid of /ABC. EPA: = d2; locus a sphere, centre at centroid of
polygon
653 8. (p2+q2+ r2)(ax + by+ cz)2= (ap + bq+ cr)2 (x2+ y2+ z2).
10. (i) Circle; (ii) pair of straight lines; (iii) hyperbola.
11. (1, 1, f), 34/2.
13. x—y+ 5z = 7, (2, —2, 3).
654 14. 12, 70. 16. Circle; rectangular hyperbola.
17. (1/V2, — 1, — 1/A/2). 18. True. 19. 4:4
656 26. Ellipse; 2,/2/3; x— 2y = 0, 2x +y = 0.
27. Rectangular hyperbola. (± 5, ± 1), 5x+y = 0, x— 5y = 0.
28. A/(1); x + (3 ± 2.,12)y = 0.
Revision exercise C
657 3. n > 0 and odd; n > 0 only. 4. —1, 0, 3. 5. 8.2°, 90°.
6.0.067. 7. —1, 1. 8.1:1.
658 10. Cancellation.
11. (1/2, —170, (12, in); --4717..
12. (i) 1/(x-1)-4/(x+2); (ii) (2 +3x)/(1 + x2)— 3/x;
(iii) 1/(1 — x)+1/(1 — x)2 + 1/(1 — 2x).
13. 12.67. 14. (2at 2, 3at).
15. (i) { —1 x 1); (ii) {x > 1}; (iii) {x = 1};
(iv) {— 2 < x < —1}; (v) { —1 < x < 1}; E = B' n C', F = B u C'.
659 16. Line x+y = a+ b. 17. (i) H;
699
ANSWERS
PAGE
0 0 0)
659 18. (2 0 0 .
0 1 0
20. 7g/12, 5g/4, 23g/12.
21. -1+1k, -1/k. 22. x < -5 or -3 < x < 7.
23. 1/(1- 3x)- 2/(1- 2x); -1- x + x2.
24. Pal, 0.
660 25. Plane perpendicular to OA. 26.1, -2, - 2, 6.
27. -1, (-1 ± V105)/2. 28. I, 0-215.
30. 1(6n3- 3n2 -n), -1(6n3 +3n2 -n). 31. 2:1.
661 33. (i) 60° < 0 < 109.5°, 250.5° < 0 < 300°;
(ii) 0° < 0 < 30°, 150° < 0 < 199.5°, 240.5° < 0 < 360°.
35. A = -2, (1, f, D.
36. (i) (2t+ 03 (3t+ 1)2/432; (ii) W ; (iii) 3.5.
38. (i) 1/V33; (ii) 3/V33.
662 42. 3. 43. -2 < x < 0, 2 < x < 7. 44. (2, -3, 7), 2.
663 47. A/2, 2A/2. 50. 5.02, 103.26, no (1 % exceed 106.52).
51. 2, 6.
664 53. cot 0 -2"" cot (2'W). 54. 3/128.
665 60. (i) R; (ii) {x E R: V2 < x < V2}.
- -
669 83. 1. 84. (i) (1, -2, 1); (ii) no solution; (iii) (A, -3A, A + 1); 8; 16.
85. (- Oln.
670 87. O. 88. (i) Line; (ii) circle; (iii) ellipse; (iv) parabola.
89. c(ct + 13)12, c(ct+,8)1(2c93); cYt2- 2XYt+cX = 0.
90. 63 (3n -1)77 .
700
ANSWERS
PAGE
670 92. x = Aa + (a A b)/Ia12, y = (1 - A)a - (a A 13)i1a1 2.
671 94. pn 1(p" +qn). 97. 2x + 9y + 5z -12 = 0.
672 98. 2.06. 99. 5x + 3y + 4z = 0, y-z = 0, x - y = 0.
100. Half-turn about (0, J3).
673 102. a4 - a, a. - a3 - a2.
104. Line in direction 3i+ 2j+ k; plane x+y- 4z = 0.
674 110. -0.96, 2.88, 5.08.
701
Index