Anda di halaman 1dari 470

tlNITED STATES DEPARTMENT OFC()MMERCE • Luther H. Hodges, Secretary NATIONAL. BUREiIlU OF STANDARDS • A. Y.

AHla, D",ecfor

Handbook of Mathematical Functions

--

With

Formulas, Graphs, and Matheinatical Tables

Edited by

Milton Abramowitz and Irene A .. SteguD -:;;::'

National Bureau of Standards Applied Mathematics Series • 55

Issued June 1-964

Tenth Prin~~g,. Deeeznbes- 1972, With correction.

Fa., we· by Ui ... · Supel:jjll;~_t. Of Doc!u.m.enb •. U.S. Gc!. .. ~_~t Pr~nt~rig OffiCe WjQhi .. gt .... , ·.D.C • .&04.020 -l':ric .. sn.ss d-om.tic p .... t;paid. ~. 110.50 GPO 8oo""to_

The text relating to physical constants and conversion factors (page 6) has been modified to take into account the newly adopted Systeme International d'Unites (81).

ERRATA NOTICE

The original printing of this Handbook (June 1964) contained errors that have been corrected in the reprinted editions. These corrections are marked with an asterisk (~) for identification. The errors occurred on the following pages: 2-3,6-8.10.15,19-20,25,76,85,91,102, 187, 189-197,218,223,225,233,250,255,260-263,268, 271-273,292,302, 328,332,333-337,362,366,415,423,438-440,443, 44'fi, 447, 449, 451, 484, 498, 505-506, 509-510, 543, 556, 558~ 562,571,595,599, 600, '722-723" 739, 742, 744, 746,752, '756,760-765,774,777-785,790,797., 801, 822-823, 832, 835, 844, 886--889,897,914,,915,920,930-931:,93'6, 940-941,944-950,953, 960, 963,989-990,1010, 1026.

Originally issued June 1964. Second printing, :N"ovembel' 1964. Thil"d prihting, March 1965.

Fourth printing, December 1965. Fiftb printing, August 1966. Sixth printing, November 1967. Seventh printing, May 19u8 . .Eightll prttitmg, 1969. Ninth printing, November 1970.

Libra ry of Congress Catalog Card Number: 64--60036

II

QA J.f1 ./tJ-!

G-.b

Preface

The present volume is an outgrowth of a Conference on Mathematical Tables held at Cambridge, Mass., on September 15-i6, 1954, under the auspices of the National Science Foundation and the Massachusetts Institute of Technology. The purpose of the meeting was to evaluate the need for mathematical tables in the light of the availability of large scale computing machines. It Wa-s the consensus of opinion that in spite of the increasing use of the new machines the basic need for tables would continue to exist.

Numerical tables of mathematical functions ate in continual demand by scientists and engineers. A greater variety of functions and higher accuraer of tabulation are now required as a result of scientific advances and, especially, of the increasing use of automatic computers. In the latter connection} the tables serve mainly for preliminary surveys of problems before programming for machine operation. For those without easy access to machines, such tables are, of course, indispensable.

Consequently, the COnference recognized that there was a pressing need for - a modernized version of the classical tables of functions pi Jahnke-Emde. To implement the project, the National Science Foundation requested the National Bureau of Standards to prepare such a volume and established an Ad Hoc Advisory Committee, with Professor Philip M. Morse of the Massachusetts Institute of Technology as chairman, to advise the staff of the National Bureau of Standards during the course of its preparation. In addition to the Chairman, the Committee consisted of A. Erdelyi, M. C. Gray, N. Metropolis, J. B. Rosser, H. C. Thacher, Jr., John Todd, C._B. Tompkins, and J. W. Tukey.

The primary aim has been to include a maximum of useful information within the limite; df a moderately large volume, with particular attention to the needs of scientists in all fields. An attempt has been made to cover the entire field of special functions. To carry out the goal set forth by the Ad Hoc Committee, it has been necessary to supplement the tables by including the mathematical properties that are important in computa.tion work, as wen as by providing' numerical methods which demonstrate the use and extension of the tables.

The Handbook was prepared under the direction of the late Milton Abramowitz, and Irene A. Stegun. Its success. has depended greatly upon the cooperation of many mathematicians. Their efforts together with the cooperation of the Ad Hoc Committee are greatly appreciated. The particular contributions of thes-e and other individuals are acknowledged at appropriate places in the text. The sponsorship of the National Science Foundation for the preparation of the material is gratefully recognized.

It is hoped that this volume will not only meet the needs of all table users but will in many eases acquaint its users with new functions.

ALLEN V. AsTIN, Director

June 1964 Washington, D.C.

In

Preface to the Ninth Printing

The enthusiastic reeeptionaccorded the "Handbook of Mathematical Functions" is little short of unprecedented in the long history 6f mathematical tables that began when John Napier published his tables of logarithms in 1614.- Only four and one-half years after the first copy came from the press in 1964, Myron Tribqs, the Assistant Secretary of Commerce for Science and Technology, presented the 100,OOOth copy of the Handbook to Lee A. DuB ridge, then Science Advisor to the President. Today, total distribution is approaching the 150,000 mark at· a scarcely diminished rate.

The success of the Handbook has not ended our interest in the-subject, On the contrary, we continue our close watch over the growing and changing world of computation and to discuss with outside experts and among ourselves the various proposals for possible. extension or supplementation of the formulas, methods and table's that make up the Handbook.

In keeping with previous policy, 3, number of errors discovered since the last printing have been corrected. Aside from this, the mathematical tables and accompanying text are unaltered, However, some noteworthy changes have been made in Ohaptef2:Physical Constants and Conversion Factors, pp.6-S. The table on page 7 has been reviaed to give the values of physical constants obtained in a recent reevaluation; and pages 6 and.S have been modified to reflect changes in definition and nomenclature of physical units and in the values adopted for the acceleration due to gra.vity in the revised Potsdam system.

The record of continuing acceptance of the Handbook, the praise that has. come from all qtl,arters,and the fact that it is one of the most-quoted scientific publications in recent years are evidence. that the hope expressed by Dr. Astin in his: Preface is being amply fulfilled.

LEwIS M.. BR,ANSCOMB, Director National Bureau of Standards

November 1970

rrra

Foreword

This volume.is the result of the cooperative effort of many persensand a number of organizations. The N ationsl Bureau of Standards has long been turning out mathematieel. tables and has had under consideration, for at least 10 years, the production of a compendium like the present one. During .It Conference on Tables, called by the NBS Applied Mathematics Division on May 15, 1952, Dr. Abramowitz of that Division mentioned preliminary plans for such an undertaking, but indicated the, need for technical advice andfinancial support.

The Mathematics Division of ~e National Research COUD@ hasalso had an active interest in tables; since 1943 it has published the quarterly journal, "Mathernatical Tables and Aids to Computation" (MTAC), editorial supervision being exercised by a Committee of the Division.

Subsequent to the NBS Conference on Tables in 1952 the attention of the National Science Foundation was drawn to the desirability of financing activity in table production. Wi th its support a 2-day Conference on Tables was called at. the Massachusetts Institute 9f Technology on September 15-16, 1954, to discuss the needs for tables of various kinds. Twenty-eight persons attended, representing scien tists and enginee-rs using tables as well as- table producers. This conference reached consensus on several conclusions and rscommendations, which were set forth in the published Report of the Oonference. There was general agreement, for example,"that the advent ofb;igh-speed computing equipment changed the task of table making but definitely did not remove the need for tables", It was also agreed that Han outetanding need is for a Handbook of Tables for the Occasional Computer, with tables of usually encountered functions and a set of formulas and ta'bles for interpolation and other techniques useful to the occasional computer". TheReport suggeeted that the NBS undertake the production of such a:aandbook and that the NSF contribute financial assiatance, The Conference elected, from its participants, the following Committee: P. M. Mo1'8~ (Chairman), M. Abramowitz, J. H. Curtise, R. W. Hamming, D. H. Lehmer, O. B. Tompkins, J. W. Tukey, to help implement these and other-recommendations.

The Bureau of Standards undertook to produce the recommended tables and the National Science Foundation made funds available. To provide technicalguidance to the Mathematics Division of the Bureau, which.carried out the work, and to provide the NSF wj~h independent judgments on grants for the work, the Conference Committee was reconstituted as th~ Committee on Revision of Mathematical Tables·of.the Mathematics' Division of the National Research Council. This, after some changes of membership, became the, C ommit tee which is signing this Foreword. The present volume is evidence that Conferencea can sometimes reach conclusions and that,their recommendations sometimes get acted on.

v

VI

FOREWORD

Active work was started at the Bureau, in 1956. The overall plan, the selection of authors for the various chapters, and the enthusiasm required to begin the task were contributions of Dr. Abramowitz. Since his untimely death, the effort has continued under the general direction of Irene A. Stegun. The workers at the Bureau and the members of the Committee have had many diseussioaa about content, style and layout. Though many details have had 'to be argued out as they came up, the basic specifications of the volume 'have remained the same as 'Were outlined by the Massachusetts Institute of Technology Conference of 1954.

The Committee wish-es here to register its commendation 01 the magnitude and quality of the task carried out by the staff of the NBS Computing Section and their expert collaborators in planning, collecting and editing these Tables, 'and its appreciation of the willingness with which its various suggestions were incorporated into the plans. We hope this resulting 'Volume will be judged by its users to be a worthy memorial to the vision and industry of its chief architect, Milton Abra,.p1owitz. We regret he did not live to see itspublication.

P. M. MORSE, Chairman. A. ERDELYI

M. O. GRAY

N. G.METROPOLIS J. B. ROSSER

H.-C. THACHER •. Jr. JOHN TODD

O. B. TOMPKINS J. W. TUK:EY.

Contents

Preface .. Foreword . Introduction

1. Methematical Constanta DAYlD S. mPMAN

2. Physical Constants and Conversion Factors A. G. McNISH

3. Elementary Analytical Methods . .

MILTON :!sa:UfOWITZ '

4. Elementary Transcendental F.unctions

Logarithmic, Exponential, Circular and Hyperbolic Fun~tion~ RUTH ZUCK;I!lR.

5. Exponential Integral and Related Functions . . .

WALTE"R GAUTSCHI and WILLIAM F. CAHILL 6. Gamma. Function and Related Functions.

PHILIP J. QA VIS

7. Error Function and Fresnellntegra.ls W AL'rER G A UTSCHI

8. Legendre Functions . . . . . , .

IRENE A. STEGUN

fl. Bessel Functions of Integer Order .

F. Wd J. OLVER

10. Bessel Functions ofFractionaI Order.

H, A. .ANT-o.sr:EwIC~

11. Integrals of Bessel Functions. . . .

YUDELL L. LUKE

12. Struve Functions and Related Functions .

MILTON ABBAMOWlTZ

13. Confluent Hypergeometric Functions Lucy JOAN SLATER

14. Coulomb Wave Functions

MILTON ABRAMOWITZ

15. Hypergeometric Functions

FRITZ OB·ERHETTINGEB.

16. Jacobian Ellip~ic Functions and Theta. Funct.ions L. M. MILNE-THOMSON

17. Elliptic Integra.ls '. . . . . . . . . . . .

L. M. MILNE-THOMSON

18. WeierstrasS Elliptic and Rela.ted Functions .

THOMAS H. SOt;rTHARD

19. Parabolic Cylinder Functiona . . . . . . .

J. C.P. MILLER

Pa.ge
ttl
V
IX
1
5
9
65
227
253
295
331
355
435
479
495 -
503
1}37
555
567
587
627
685
va VIlI CONTENTS

Page

20. Mathiec Functions . . . . 721

GERTRUDE BLANCH

21. Spheroidal'Wave Functions.

ARNOLD N. LOWAN

22. QrthogonalPolynomials . .

TIas W. HOCHSTRASSl!l:B,

23. Bernoulliand Euler Polynomials, Riemann Zeta Function EMILlE V. HAYNSWDRTH and KARL GOLDBERG

24.. Combinatorial Analysis . . . . . . . . . '. . . . ..

K.. GOLDBER.G,. M. NEWMAN and :E. HAYNSWORTR 25. Numerical Interpolation, Differentiation and Integration.

PHiLIP· J. DAVIS· ana IvAN POLONSKY 26.!?,obahility Funetiona. . . . . . . .. . . .

MARVIN ZELEN and NORMAN c_ SEVERO 27. Miscellaneous Functions .

lIUilNE A.STEGUN

28.. Scales of Notstien . . .. .

S. PE·AVY and A. SCHOPF 29. Laplace Transforms .

Subjec.t Index .'.

Index Of Notations . .. .

751

771

821

875

925

997

1011

1019 1031 1044

Handbook of Mathematical Functions

with

Formulas, Graphs, and Mathematical Tables

Edited by Milton Abra.mowitz and Irene A. Stegun I. In-troduction

The present Handbook has been designed to provide scientific investigatora with a comprehensive and self-contained summery of the mathematical functions tha.t arise in physical and engineering problems. The well-known Tables of Functions by E. Jahnke and F. Emde has been invaluable to workers in these fields in its many editions- .during the past half-century. The present volume extends the work of these authors by giving, more extensive and more accurate numerical tables, and by ~ving larger collections of mathemetical properties of the tabulated

'(unctions. The number of Junctions covered has also been increased.

The classification of functions and organization of the chapters in tIDS Handbook is similar to that of An Index of Mathematical Ta.bl~ by A. Fletcher, J. Q. P. Miller, and L. Rosenhsad.! In general, the chapters contain numerical tables, graphs, polynomial or rational approximations for automatic computers, and statements of the principal mathematical properties of the tabuI ated functions, particularly those of compute-

tiona! importance. Many numerical examplea are given to illustrate the use of the tablesand also the computation of function valuee which lie outside their range. At the end of the text in each chapter there isa short bibliography giving books and papers in which proofs of the ma.thematical properties stated in the chapter may be found. Also listed in the bibliographies are, the more important numerical tables. Comprehensive lists Of tables are given in the Index mentioned above, and current information on new tables is to be found in the N ational Research Council quarterly Mathematics of Cornputetion (formerly Mathematical Tables 'and Other Aids

to Computation). -

. - The mathematical notations used in this Handbook are- those commonly adopted in standard t~_xts, particularly Hig1er Transc~Ddental. Functions, Volumes 1~3, by A. ErdelYI, W_ Magnus, F. Oberhettinger' and. F. G. Tricomi (McGrawHill, 1953~55). Some alternative notations have alao been listed, The introduction of new symbols has been kept to a rninimumvand an effort has been made.to avoid the' use of conflicting notation.

2. Accu.-acy of the Tables

The number of eignifleant figures given in each table has depended to some extent on the number available in existing tabulations. There hse been no. attempt to make it uniform throughout the Handbook, which would have been a costly and laborious undertaking. In most, tables 8:t least five significant~ures have been provided, and the tabular intervals have generally been chosen to ensure thabIinearinterpolation will yield-fouror five-figure accuracy, which B,uffices in moet . physical applicafions. Users requiring higher

"

I The moot ~t. 'thll stxth, with F. l.oo9oh added 118 co-author, wBil' J;iublJilhed In 1961) by MCOraw-Bill. U'.S.A .. IWd Teubnlll', Germany ..

i The sMond edltloll, with L. :T. Comrie ilddeilll8 co-~uthor Was publlsbe.d III two "plumes In 196:2 hoY AddlBon-Wesle,y, U.S,A .• BIld 8clenttllc Com-

piitlnil Berv:lee Ltd .• GrIlSt,llrltllln. '.' .'

precision in their interpolates may obtain them by use of higher-order .interpolation procedures. described below.

In certain tables many-figured function values are given at irregular intervals in the argument. An example is provide_d by Table 9.4. The purpose of these tables is to furnish "key values" for the checking ?f ,Progr9JILS .for automatic computers;

no question of interpolation anses. .

The maximum end-figure error, or "tolerance" in the tables in this Handbook is ,{o of 1 unit everywhere. in the case of the elementary functions, and 1 unit in the case of the higher functions except in a few cases where- it has been permitted

to rise to 2 units. .

X INTRODUCTION

3. AtniliaryFunctions and Arguments

One of the objects of this Handbook is to provide tables or computing methods which enable the user to evaluate the tabulated functions over complete ranges of reel values of their parameters. In order to achieve thisobj ect, frequent use has been made, of aexiliary functions to remove the infinite part of the original functions at their ~~~&ritie8, and auxiliary arg~ents to cope with infinite ranges, .An example will make the procedilla clear,

The exponential integral of positive argument is gifen by

J% e" Ei(:z;) = ._du

-",u

. :t Xl. za

='Y+In:Z;+1.ir+z:21+3.31+ ...

""~[1+~1+~+~+ ... J<:C-4<X»

The logarithmic singularity 'precl~des d~ect interpelation near x=O. The functions El(X) -In x and x-1[Ei(x)-ln x-'YL however, are wellbehaved a:nq readily interpolsble in this region. Either Will do as 'an auxiliary function; the latter was' in fact selected as it yields slightly higher accuracy when Ei(z) is-recovered. The function x-1[Ei(x)-ln x-'Y] has been tabulated to nine decim als for the range O'::;:Ii .:::;;I:. For 1'::; x .::; 2, Ei(;r:) is sufficiently well-behaved to adrnit direct tabulation, but for larger values of z, its exponential character predominates. A smoother and more readily interpolable function for large x isu-~EHx); this has been tabulated for 2'::;x'::;1O. Finally, . the range 10 .::; x. '::;!Xl is covered by use of the inverse argument :r:~l. Twenty-one: entries of u-'tEi(z), corresponding to x-1=.1(-.005)0, suffice toproduce an interpolable table.

4. interpolation

The tables in this Handbook are not p;oviiled with differences or. other aids to interpolation, because -it was felt that the space t.hey require could be better employ~d by the tabulation of additional functiona, AdirnttedJy aids could have been given with9ut eonsuming extra. space by increasing the intervals of tabulation, but this would have conflicted with the requirement that linear interpolation is accurate to four or five figures.

. F~or app).ications. in whicl: linear interpolation 18 msuffiClently accurate It 18 "intended that Lagrange's formula or Aitken's method of iterative linear interpolation 3 b-e used. To .help the user, there is a-statement at the foot of most tables of the maximum error in a linear interpolate, and the number of function values needed in La,grange's formula. or Aitkenls method to, interpolate to full ta:bular accuracy.

:As an example, consider tliefollowing extract from Table 5.1.

:t xezE1(x) x xe"E1(iz:)
7. 5 .89268785!1: 8.0 .89823 7113
7. 6 . 89384: 6312 8.1 .89927 7888
7,7 .89497 9666 8.2 . 9002·9 7306
7,8' .89608 8737 8. 3 .90129 60')3
7.9 .89717 4302 8.4 .90227 4695 The. numbers in the square brackets mean that the maximum .error ina linear interpolate is 3 X 10-6, and that to interpolate to the full tabular accuracy five points musb be used in Lagrange's and Aitken's methods.

'A. C.Ai"tken_i. On :lntel'pQJatlon by lteratiqn of proport)on<lil'arts, with, oat tl!e nSIi oM1IleJ::el\oos. proo. Eil:lnbI1Tgh~llth, 600.3, 56-71l{19B2).

Let us sUep9se that we wish to compute the value of xe"&(x) for x=7.9527 from this table. We describe in turn the application of the methods of linear interpolation, Lagr8iDge and .Aitken, and of alternative methods, based on differences and Taylor"s series.

(1) Linear interpolation. The formula for this process is given by

i1>=(l-p)]o+pjj

where JOI flare consecutive tabular values of the function, corresponding to argumentsXo! :til 1'8-spectively; p is the given fraction of the argument

interval .

p= (x-xo) /(Xl-Xo)

and jp the required interpolate. In the present instance, we have

10= .89717 4302

11=.89823 7113

p=.527

'The-most convenient way to evaluate the formula. on &' desk calculating machine is to set fo and i1 in turn on the keyboard, and carry: out the multiplications by I-p andl.cumulatlvely; a. part~8.r check is then :Q!_ovide by the multiplier disl reading unity. We obtain

1.627=(1-.521)(.89717 4302)+.527( .. 89823 '7113)

. = .89773 4403.

Since it is known that there. is a. pOl3Sibleer;r~r of 3XlO~6in the linear formule, we round off tb?s result to .,89773. The maximum possible erf?r m this a.nswer is composed of the error committed

INTRODUcrION

by the laat rounding, that is, .44:03 X 1O-~, plus 3XlO-6, and so certainly cannot exceed .8XIO-5• (2) Lagrange's formula. In this example, the relevant formula is the 5-pomt one, given by

f=A_2(P)}-JI +A_1(P)f-1 + Ao(P)jo+AI (P)jl

.. . +A~(P)fll

Ta,bles of the coefficients AJ:(p) are given in chapter 25 for the range p=O(.Ol)l. We evaluate the formula, for p=.52, .53 and .54 in turn. Again,_ in each evaluation we accumulate the A.t(p) in the muUiplierregister since their sum is unity. We now have the following subtable.

:t: z&%(:t:)
7.952 .89772 97'e;7
10622
7.953 . 89714 0879 -2
10620
7.954 .89775 0999 The numbers in t'he third and tourth columns are the first and second differences of the values of xe"'Et(i) (see below); the smallnesa of the second difference provides a check on the three interpolations. The required value is, now obtained oy linear interpolation:

j,,=.3(.89772 9757)+.7(.89774 0379)

= .89773 7192.

In cases where the correct order of the Lagrange polynomial is not known, one of the preliminary mterpolations may have to be performed with polynomials of two or more different orders as a check On their adequacy .

(3) .A.i~ken 's method of iterative linear interpolation.The scheme for carrying out this process in the present example is as follows:

n 2:,. lI .. =:U"&(z) 1/0, " Y~.l.1I !lO.I,i. " 'Yo, 1.2.3, n 2;,,-2;
"0 8.0 .89823 7113 .0473
I 7.9 .897114302 . 89773 44034 -.0527
2 8. 1 .89927 7888 . 8977'4 48264 . 89773 71499 .1473
3 7.8 .896088737 ? 9OZ20 2394 . 89773 71938 -.1527
4 8. 2 .90029 7306 4 98773 1216 16 89773 71930 .2473
5 _7.7 .89497 9866 2 35221 2706 43 80 -.2527 Here

1 lyO

110.,.= %,,-%0 1/"

- 1 IYo.l

YO.I ... =;--=x y

. l1li 1 0.11

1 jYo.1. .... "'-I ....

!lo. I •.•.• ,"'-I, ...... =:::-=:;;- y.

..... fl N",. 0.1 .•••• ",,-I ••

x -xl

,1ft

:&,,-z

If the quanbities x,. -x 'and Xm -x are used ars multipliers when forming the cross-product on a desk maehine, their accumulation (x",-x) - (3im -x) in the multiplier register is the divisor to be used at that stage. .An extra decimal place is usually carried in the intermediate interpolates to safeguard against acoumulation of rounding errol'S.

. The order in which the tabular values are used is immaterial to some extent, but to achieve the maximum rate of conver~enee and at the same time minimize accumulation of rounding errol'S, we begin, as in this example, with the tabular argument nearest to the given argument, then take the nearest of therema.i:ning tabular arguments, and 80 on.

The number of ta.bular values required to achieve a, given prec~ion emerges. naturally in the course: of the Iterations. Thus m the present exampla six values were used, even though it was known in advance that five would suffice. Theextra row confirms the convergence and provides a valuable check.

(4) Difference formulas. We use the central

difference notation (chapter 25), .

:to In
II/III flljl
:1:1 II
lila,;.' li'f1/3
zs' Ii ""1, lI'!t
a/m B'f6li
:t3 h o~h
_6/111
:1:. it Here

'4lfS'=/I-fo, lJfs/'J=f'1.-iI, ... I,

"'fl=6fa/1-6fm=b-2il+fo BI/&n=6'%-lJIh=!a-3h+3fl-jQ liYz="'hl2-otjal~,=f,-~a+6fa-4!l+fo and so on.

In the present example the relev_ant part of ~he difference table is as follows, the differences being written in units of the last decimal place of the function, as is customary. The smallness of ~he high differences provides a cheek on the function values

:z; xe"'E1(x}

7. 9 . 89717 ,4302 8, 0 . 89823 7118

"lj -2 2754

-2 2036

Applying, for example, Everett~s interpolation formula.

f,,=(1-p)fo+EJ(p)8'fa+'~MPWfo+ ...

.. +pf1+ Fa.(P) ffJfl + F,(PWfl +

and taking the numerical values' of th~ interpolation coefficients E2(P) , E,('P}, FJI(p) and ~j(p) from Table 25.1, we find tna.t

ll'lT 1tUlJ U ... -1·10..1.'1

10tf,5i7= .473($9717 4302) +.061-196(2 2754) - .012(34)

+ .527 (89823 7113) + ~063439(2 2,036) - .012 (39) =89773 7193.

We IDfl,Y notice in passing that Eveeett/s fornmlashows that the error in it linear interpolate is approximately

.E~(p)l)2jD+ F2(P)~fl "'" U1Mp} + E2(P) ][02/0+@fI1 Since the masimum value of IE2(p)+F2(P)1 in the range .o<p<l is ~, the maximum. error in a linear interpolate is approximately

1, '. . 1

16162fo+62fd, that'ls'16If2-fl-fo+f-ll.

(5) Taylor's series. In cases where the successive deriv:atives of the tabulated function can be computed Iairlyeesily, Taylor's expansion .

f{x) =f(x(J) + (:I; -xo)1':7°)+ (X-ZO)2f'~\-;;D)'

+ (x-xo),p" (:1)0) + 31

can 'be used. We first compute as many of the derivatives P") (X6) as are significant, and then evaluate the series for, the given value of z. An advisable oheckon the computed values of the derivatives is to 'reproduce the adjacent tabular values by evaluating the series for :1(=:2:-1 and a::l.

In the present example, we have

!(l:)=xe:OEI(X)

~~) = (1 +:rl)f(x) - 1

f" x) = (l+:rl)f'(x.) -x-if(x)

1''' x) == (1+x-1Jf" (x) -2x-~f'{x) +2J;-3t(X),

With·zQ=7.9 and x.-Xo=.0527 our computations are as follows; a.D extra' decimal has been retained in the valuea of the terms in the series to safeguard

against accumulation of rounding errors. .

k f< tl (x~)./ k! (x ~xD),tf{k) (zo)/k I

o .89717· 4302 .89717' 4-302

1 .01014 0669 .00056 6033 '3

2 -.ooua 7621 -.00000· 3159 5

3 .00012 .1987 .00000' 0011 9

.89773 7194

5. Inverse Interpolation

With linear interpolation there is no difference in principle between direct and inverse interpolation. In cases where the linear formula provides an imiufficiently accurate answer, two -methods are available. We may interpolate directly, for example, 'by La~ange's formula to .prepare a new table at 8. fine interval in the neighborhood of the approximate value, and then apply. accurate inverse linear interpolation to the subtabulated values. Alternatively,. we may 'USe .Aitken's method or even possibly the TayIOr'.s· series method, with the roles 'of function and argument interchanged.

I~ is important to realize that the accuracy of an Inverse interpolate may be very different from that of iI. direct interpolate. This is particularly true. in regions where the function is slowly varymg, for example, near a msaimum or minimum. 'The maximum precision attainable in an inverse interpolate can be estimated with the aid of the formula

dj Ax~t1f/ ax.

in w¥ch III is the maximum possible error in the function values.

Example. Given xe"E1:(i) =.9, find z from the table on paige X.

(i) Inverse Iinear in terpolstion. The formula for pis

s= (jll-jO)/(fl-jr.}.

In the present example, we ha.ve

P· .9-.$9927 7888 72 2i12

.90029 730!i-.89.927 7888 101 9418 .70.83'57,

The desired xia therefore

Z =:1:0+ P(Zl-XO) =8. i + . 7063~7(.1) =.8.17088 57

To estimate the possible error in this answer, we recall that the maximum error of direct linear interpolation in this table is Aj=3X 10-6• An approximate value for dj/dx is the ratio of the first difierence to the argument interval (chapter 25), in this case .010. Hence the, maximum error :in x is approximately 3XI0-s/(.QIO), tha.t is, .0003.

(ii) Subtabula.tion m.ethod.To improve the approximate value of x just obtained, we interpolate directly for p= .70, .71 and .72 with the aid of Lagrange's 5~point formula;

x xB'"E1(:i:) 8 tf1
8. 170 . 8.9999 3683
1 0151
8.171 .90000 3834 -2
1 0149
8.172 . 90001 '3983 Inverse linear interpolation in the new table givea

.9- .89999 3683 6223

p .00001 0151 ' .

Hence x=8.17062 23-.

An estimate of the maximum error in· this result

IS

(iii) Aitken's method. This is carried outin the same manner as in direct interpolation.

INTRODUCTION Xlll
n y,,=xe"E1(z) ;i:" Zo ... ~.!.. .. :'O.I.~.11 ~.I;2.a." lI .. -,y
0 . !.JO029 7306 8. 2 .00029 7306
1 .899Z77888 8. 1 8. 17083 5712 -. 00072 2112
2 .90129 6033 8.3 8. 170~3 1505 8.17061 9521 . Q0129 6033
3 .89823 71i3 8.0 8. 17113 8043 2 5948 8.17062.2244 -. 00176 2887
4 .902274695 8. 4 8. 16992 9437 1 7335 415 8. 17062. 2318 .00227 4695
5 . 89717. 4802 7. 9 8. 17144 03B2 2: 8142 231 265 -. 00282 5{j~8 The estimate of the maximum error in this result is the same as in the subtabulation method. An. indication of the en-or is also provided by the

6. Bivariate Interpolation

Bivariate interpolation is generally most sim-ply performed as a sequence of univariate interpolations. We carry out the interpolation in one direction, by one of the methods already described, for several tabular values of the second argument in the neighborhood of its given value. The interpolates are differenced as a check, and

discrepancy in the highest interpolates, in this case :to .U,8.4. and :to ,1,2; ,&.

interpolation is then carried out in the second direction.

An alternative procedure in the case of functions of a complex variableis to use the Taylor's series ezpansion, provided that suceessive derivatives of the function can be computed without much difficulty.

7. Generation of Functions from Recurrence Relations

Many of the special mathematical functions which depend on a parameter,called their index, order or "degree, satisfy 8. linear difference equation (or recurrence relation) with respect to this parameter. Exil.mples are furnished .by the Legendr-e function P,,(z), the Bessel function J .. (z) and the exponentia1 integral E,,(x), for which we have the respective recurrence relations

(n+ l)P"TI- (2n+ l):z:P .. +n:P .. -I=o 2n

J .. +1--J .. +J .. _1=0 z

ParticulrJ.i'ly for atrtomatic work, recurrence relations provide an important and powerful computing tool. If the values of P",(z) or J,,(z) are known for two consecutive values of n, or En(z) is known for one value of n, then the function may be computed for other valuea of n by successive applicationa of the relation. Since generation is carried out perforce with rounded values, it 'is vital to knowhow errors may be propagated in the recurrence process. If the errors do .not grow . relative to the size of the wanted function, the process is said to be stable. If, however, the relative errors grow and will eventually overwhelm the wanted function, the process is unstable.

It is important to realize that stability may depend on (i) the particular solution 01 the difference equation being computed; (ii) the values of x Of other parameters in the difference equation;

(iii) the direction in which the recurrence is being applied. Examples are as follows.

l3_ta-bUity-mcTll:asing 11. P,,(:z:), .p.;::(x).

Q,,(:.), Q:'(:z:} (z<l) Y .. (:z:), K .. (:z:}

J ...... -j,i (%c). I ..... -}4(:z:) E .. .(x) (n<~).

Stability-decreaSing 11. P,.(x.), P:{x) (x:<l) Q,,(x), Q:(!l:)

J .. (x), J,.(x)

J,,+}<l{:Z:), l,.+~(x)

E .. (x) (n>x)

F,.(." p) (Coulomb wave function)

Illustrations of the. genera.tion of functions from their recurrence relations are given in the pertinent chapters. It is also shown that even in cases where the. recurrence process is unstable, it may still be used when the starting values are known to sufficient accuracy.

Mention must also be made here of a refinement, due to J.C. P. Miller, which enables a recurrence process whi'ch is stable for decreasingn to be applied withQuta.ny knowledge of sta.rting values for large n, Miller's algorithm, which is wellsuited t~ automatic work, is described in 19.28. Example 1.

8. Acknowledgments

The production of this volume has been the result of the unrelenting efforts of many persons, all of whesseontrfbutions have. been instrumental in accomplishing the task. The Editor expresses

his thanks to each and every- one. ,

The Ad Hoc Advisory Cemmittee individually and to,gether were instrumental. in. establishing the baSIC tenets thatserv:edas a guide m the1orma~ tion of tbeentire work. In particular, special tha.nkaa.re due to Professor Philip M. Mortie for his ,oontin:uoU2-enC0Ul'age;ment and support. Professors J. Todd and A.ErdeIyi, panel members of the Conferences on Tables and members of the Advisory Committee have maintained an undiminished interestjoffered .many suggestions and carefullyreedall the chapters.

Irene. A, Stegunhasse:rvedf,rliectivelya:s associate editor" sharing __ p each stage ~f. the planning of the volume. WIthout her un tiring efiorts~ completion would never have been possible.

Appreciation 113 expressed for the, generous COOp era tion of publishers' and l1uthors in granting permission for the use of their sourcemateriel, Ack:n:owledgplen ts for tabular material taken wholly or iri. part from published works are given on thefirst-page.ofeachta.ble. Myrtle R. Kellington corresponded with authors and publishers to obtain formal permission for inoluding their m.aterial, maintsJ:ued uniformity throughout the

bibliographie r¢erenees andassisted in prep aring the mtroduetorymateria]. .

Valuable assistance intheprepara tion, checking and editing of the- tabular material was received iroID: Ruth ~. Capuano, ~lizabeth F. Qodehoy, DaVId S.Llepman, Kernut Nelson,Bertha H. Walter and Ruth Zucker.

EquaJIy important has been ·the untiring cooperation, B.SsistanC8j. and patience or th~ members of the NBS st'8.ff in. ha:ndijng themyri_ad of detail. '):lecessarilyatten4ing the publica,tio.n of a volume. ofthism.aguitll.de. Especially appreciated have been the_ helpful discussions and services from the members of the Office of Technic~ Information in t.he areas of editorial format, graphic .lU't layout, printing detail, pre-printing reproduction needs, as well as attention to promotional detail and financial support. In addition, the clerical and typing stili Of the Applied Mathematics Divieionment commendation for their efficient and pa.tientproduotion of manuscript COR:?" involvi.·n g com. p~cat.ed.· .teohnical notetien ..

Fm.ally, the continued support of. Dr .. E. W.

Cannon, chief of. the Applied Mathematics Division, and the advice of Dr. F'. L. AJti assistllJit chief,8;s well as of the many mathematiclans in the Division, is gratefully acknowledged.

M. ABR;4MOW~TZ.

2. Physical Constants and Conversion Factors

A. O. McNISH 1

Contents

Page Table 2.1. Common Units and Conversion Factors. . . . . . . . .6 Table 2.2. Names and Conversion Factors for Electric and Magnetic

Units. . . . . . . . . . . . 6

Table 2.3. Adjusted VaJues of Oonstanta . . . . . . . . . . . .. 7

Table 2.4. Miscellaneous Conversion Factors. . . . . . . . . . .. 8

Table 2.5. Conversion Factors for Oustomary U.S. Units to Metric

Units. . . . . . 8

Table 2.6. Geodetic Oonstants . . . . . . . . . . . . . . . . .. 8

.1 Natio_nal Bureau of Standards.

2. Physical Constants alld Conversion Factors

The tables in this chapter supply some of the more commonly needed physical constants and conversionfactors. *

The International Sys,tem of Units (8I) established in 1960 by the General Conference of Weights and Measures under the Treaty- of the Meter is based upon: the meter (m) for length, defined as 1650763.73 wave-lengths in vacuum. corresponding to the transition 2Pl[)-5ds-of krypton 86 ; the kilogram (kg) £0.1' mass, defined as the mass of the prototype kilogram at Sevres, France; the secon-d (8) for time .• defined as the duration of 9192681770 periods of the radiation corres-ponding to the transition between the two hyperflne levels of cesium 133; the kelvin (K) for temperature, defined as 1/273.16 of the thermodynamic temperature of the triple point of water; the ampere (A) for electric current, defined aa the current which. if flowing in two infinitely long parallel wires in var::¥o separated by one meter, would produce a force of 2 X 10-1 newtons per meter of length between the- wires; and the~ candela (ed) for luminous intensity, defined as the luminous intensity of 1/60000() square meter of a perfect radiator at the' temperature

of freezing platinum. .

All other units of 81 are derived from these base units hy~ assigning the value unity to the proportionality constants in the defining equations (officialsymbols for other $1 units appear fn Tables 2.1 and 2.2). Taking 1/100 of the

~£ee also "Preface to NInth FrlnUllg," "Ilg~ IlIa 8.111:1 pag1l II.

meter as the unit for length and 1/1000 of the kilogram as the unit for mass gives risesimilarly to the cgs system, often used in physics and chemistry.

SI, as it is ordinarily used in aleetrornagnetism, is a rationalized system, i.e., the electromagnetic units of 8I relate to the quantities appearing in the so-called rationalized electromagnetic equations. Thus, the force per unit length between two current-earrying parallel wires of infinite length separated by unit distance in '1!ctCUO is 2f = p.ni,i~/4'll', where jLu has the value 4?1" X 10~:;Hjm. The force between two electric charg~si'n vactto is correspondingly given by f= qlq~/4'IT(ur\. t!(, having the value l!~,c2, where e is the speed of light in meters per second. «(ro~8.854 X 10-i~F /m)'

Settingp." equal to unity- and deleting 4". from the denominator in the first equation above defines the egs-emueystem. Setting (10 equal to unityand deleting 471' from the denominator in the second equation correspondingly defines the cgs-esu system, The ega-emu and the cgs-esu systems are most frequently used in the unrationalized forms.

Table 2.1. Common Units and Con.version Facto .. s,. CGS System. and SI

eGS Name

SJ Name

Factor

Quantity

lOS 107 10'

.newton (N} joule (J) watt. (W)

dYM erg

Force Energy Power

Table 2.2. Names a.ild Conversion Factors for Electric and Magnet~c Units

Q:aant~ty

S1 emil esu ernu-SI esu-SI
name name name factors faetors
ampere (A) abampere S.1;atampere 1,0-1 ,..,3 X. 10'
coulomb (Gl abcoulomb statcoulomb 10·,1 ,....·3 >< loa
volt (V) abvolt statvelt lOa -(.113) X 10-·~
ohm (n) abohrn statohrn lOS ...:. (1/9) X 16~11
henry (H) centin::!ete.r ~ __ ~ ___ ~ or. __ •• __ ~ _ ••• __ 109 ,..., (1/~JX .10~1 !
farad (F). .. __ .. ,--,._ .. - ... -.- ..... centimeter 1,070 -9 X Ion
A'm~l oersted __ ••• ___ • ___ •••••• _r.·r~_ 41"X 10~l •. -s X IDS
A gilbert --_ ... '_ .... -~ .. --'-'-~ .• ~.- 41< X 10~1 _3fl!)6:
weber (Wb) maxwell ~.~~~ .. -;.--+- .. --.-.-.- t08 -(1/3)X 10-1
te,sla (T) gauss (n) --,. -~---.----. ----_. --_ .. 101 -(1I3)X 10-"
__ ••••• _., ••• __ •••• 7 ••••• • ..... ." ..... ~.--~ .. ~~ .... "----- - - - --_. -~---. -_. - .. - .. -- .. ~ 10~5 -3 X 105 Current

Charge..

·PQren:tjaJ

Resistance

Inductance Capac·itance Magnetizing foree Mag:netomotive force Magn!!ti!)fiu.x .,

Magnetic fluxdensi:ty ~lectric displaeement

________________________ J_ ~, ~ _L ; ___

E~.a:mple: If the vahle assigned to acurrent is 100 -arnperes its value in abamperes is 100. X 10-1 = 10.

P!JYSICAL CONSl'ANTS AND CONVERSION FACTORS

7

The values of constants given in Table 2.3 are based on an adjustroentby Taylor, Parker, and Langenberg, Rev. Mod.Phys. 41, p.375 (19S9). They are being considered for adoption by the Task Group on Fundamental C6nstants of the: Committee on Data for Science and Technology, International Council of Scientific Unions. TheuneertaU1ti~s given are standard errors estimated :from the experimental data included in the adjustment. Where applicable, values are ba!;ed on th~ unified scale ofatomlc m941l\les in which tl;leato.mie mass unit (u) is defined as 1/12 of themass o.f the atom of the~ 120 nuclide.

i

Table 2.3. Adjusted VaIues of CODstan't8

Constant

Symbol

(uncorrected for diamagnetism,

B.O) , , 1Jo·'P/P.N 2.792709 17

Gas. constant , , .R. 8.31434 35

N.ormalvolwnie. perfect gas ,.. V. 2.241 36 39

Boltzmann constan"t , _ k 1.380.622 5,9

First radJationcQnstant (!3.,..hc) 01 4.992 579 38

S.6cond radiation constant c, 1.438833 6.1

Stefan-BoltJm1~,nn constant _... a:' 6.669 61 96

G;rav:i~tional constant · .... ·······._' .... 1 G 6.673 231

Speed of light in vacuum - - C

ElementfJ.J:'Y charge e

Avo~dxo consta:tlt ··· · .. ·, - 1 ~ ~

AtOmlC mass unlt " 1 u

Electron :test lt1a.s~ ~ .. - m.

. .... I

Proton rest lila.SS ~ -~

. .. I .. ~

I

Neutron rest mass _ ! m;. ~

. ~

Fara.day constant , i F

P)an~1!:: constant 1 h .x

Fine structure eonstant , 1 Q.

, l/r;I;

Charge to ms,.Bs ~tio for elecbron., ~/m.

Quantum-charge ratio __ .- .. _..... hie

Compton wavelength ofeleetron .... Xc >.c/2 ...

Compton wavelength of proton ... _ 'Xc,p

·J,.c.~12 ...

Rydberg constant _.............. Roo

Bohr radius _.......................... "'"

Electron radius _ .. _....................... r.

Gyromagnetic ratio of Ji:toton ........ 'Y

'Y/27r

(u::~;re.~~.~ ... ~~~ .. ~~:~~~~:~ { i ~:/2~

Bohr magneton ". Poe

Nucie,armagneton "...... P.N

Proton moment _...... p.p

P.;J"'N.

Value

2;9979260 1.6021917· 4.803250 6.022169 1.660531 U09 &58 5.4.86930 1,672614 1.007 276 61 1.674920 1.00866520 9_6~8 670 2.892599 6.626196 1.0~459.19 7.297351 1.370 360 2 1.7'588028 5.272759 4.135708 1.3795234 2.4263096 3.861. 59.2 1.3214409 2.103139 1.09737312 5.2917715 2.817989 2.6751965 4.257707 2.6751270 4.257597 9~274'096 5.050951 1.4106203 2.792782

Unit

Uneer-I.- -----------------r--------------------taiDty . Sy,ste.me Inte:t'national Gentime,ter·gram.-seeond

! .. ('81) (CGS)

±10 XIOs mls XI01Q r:mls
70 10-19 C 10-:0 cm1/tgl/2 ..
21 I~' ~ •• ~- •• ~ ••••• r .... ~ ...... ~.4_.,_~ .•.•. ~;". ~ ••. ~.- •• - 10-10 <:m3/tgl/2S71 t
40 IOU mol-l 1023 mol-l.'
11 1Q-S1 kg 10·iH g:
54 10"31 l<g 10-n g
.34. 1O~4 u 10-~ u
11 10-21 ~g io-» g
8 I 100 u 100 u
11 10-27 kg 10-H g-
10 :1,00 u 100. u
54 10" C!mol lOS em:L/~gl/2mol-U
.16 r,~ ..... , •• '.~ .~ •• ~ !.~- ............ -.~- .... 'f;" •.••• 10H cmS'/'g-l/~s-lmol-l t
50 lcrs~ J 0 S 10-iT erg- s
80 , 1O-8~ J oS lO-iT erg· s
11 10-~ ... -~ ... ~ .................. ~ ...... ~~ .. 10-3 -,~,,,,-,,,,-,,,,,-,,-,---,---,,-,-
21 10i ............... - .......... - .. 10; .. , •• __ ••••• ~H.-. _. + ~~ ~~_
54 1011 C/Itg lQ1 em 1/2/gY2 .,
16 ~ +"~ •• ~ ... ~.~ ..... ~ •• - ......................................... -,. 1011 cmJI2g-11~s~1 t
14 100-15 J • siC 10-1 cm8!20g1/2s-1 >I'
46 •••••• _ ••••• ' •• ·.·r'.' ...... .,. .... _ ........ r .. __ ••· __ · __ ,_.,- 10-11 cmllZgl/2 t
74 10-12 m 10-10 em
12 10~13; m 10-11 c:m
9Q 10-a m 10-lB ern
14 10-16 m 10-1~ em
11 10.1 mol 10.5 em-!
81 10-a m 10-9 cm
13 io-» m 10,13 em
82 109 rail·· s-lT-i 104 rad·5-1G-l ..
13 101. HzIT 103 s-lG-l ..
82 108 red· s-lT-l 10' rad« s-lG-1 ..
13 107 Hz!T 103 s-IG-l ...
65 10-H J/T ro-» erg/G ..
50 10~27 J/T lo--n erg/G ..
99 lO-H, JjT 10-~3 erg/O"
17 10' ....... ·._· ••• ·_·+r+ •• +~~··· .. 100 ".'I~·· ~~. ~ ..• ,~ ............. 100 ........ ~~_,o+.~ ........ ,.,_ ........ ,_ 100 ... ~ ..... , ....... -~ ............
100 J . K-I mol-t lor erg.' K-J mol-l
10~1 m3/mol 104 cm.1fmol
10-28 JIK 10-18 erg/K
10-24' J·m lo-n erg· em
t(}-' m- K 100 em·K
IO-s W . m-!lK-~ lO-~ e;r;g' • mn:-2s-1 K -~
10-11 N· mR/kgl 1()-3 dyno em';gS ;l:Based on 1 std. dey; applies to lastdi:gii:.:; i»;- preceding cohmm, *Electromagneticaystem. tEleetl'osta1;i:esystem.

,.'

,:;ru,

8

PHYSICAL CONSTANTS AND CONVERSION FAC'I'UK::i

Tab1e 2.4:. Miscellaneo'us Conversion Factors

= 9 .. 806 65 meters per second per second* =1.01325 X 105 newtons per 'square meter" = 1.01325 X 106 dynes per square centimeter" - 4.1840 joules"

= 4.).868 joules'"

= 10-B cubic meter" = 10-io meter*

= 105 newtons per square meter"

= 10f! dynes per square centimeter"

= 10';';l meter per second per second" ~ 1 centimeter per second per second" = 1.496 X 1011 meters

= 9.46 Xl-01S meters

= a.08 X 1016 meters

-=- 3.26 light years

1 curie, the quantity of radioactive material undergoing 3.7 X 1010 disintegrations per second",

1 roentgen, the exposure ofx- or gamma radiation which produces together with its secondaries 2 .. 082 X 109 electron-ion pairs in 0.0.01 293 gram ofair' ..

Standard gravity, go

Standard atmospheric pressure, Po

1 thermodynamic calorie,' cal, 1 IT calorie", cal,

1 liter, I

1 angstrom unit, A. 1 bar

1 gal

1l:j.stronomical unit, AU 1 light year

1 parsec

The index of refraction of the atmosphere for radio waves of frequency less than 3 X 1010 Hz is given by (n~ 1)106= (77~6/t) (p + 4810e/t), where n is the refractive index; t,tempe.rature in kelvins;p, total pressure inmilllbars; e, water vapor partial pressure in millibars,

Factors for converting the customary United Geod~ticconstants for the international

States units to units of the metric system are (Hayford) spheroid are given in Table 2.6.

given in Table 2.5. The gravity values are on the basis of the revised Potsdam. value, They are about 14 parts per million smaller than previous. values. They are calculated for the 'surface of the geoid by the international formula.

Table .2.5.F~u~tors for Converting Cmtom.ary U.S. Units to 51 Units

1 yard 1 foot

0.914 4 meter*

0.304 8 meter* 0.025 4 meter" 1609.344 meters '"

Table 2.6. Geod.eticConstants

a =6 378 388 m;t = 1/297.; b = 6 356 912 m

1 inch

1 statute anile

1 nautical mile (inter- t 852: meters" ' national)

Latitude

Length (If l' of longitude

Length. of l' of latitude.

m,1$" 9.780350 9 .. 783800 9.793238 9.806154 9.819099 9.828593 9.83.2· 072

1 pound (avdp.) 1 oz. (avdp.)

1 p.ound force

1 slug

1 poundaI

1 foot pound

OA53 592 37 kilogram" 0.02834952 kilo.gram 4.44822 newtons 14,593 9 kilograms 0.138 255 newtons 1.355 82 joules

0'" .15 30 45 60 75 90

.MeteTs 1842.925 1844.170 1847.580 1852.256 l·85·6.951 1860.401 18,61.666

Meter8

1855.398 1792.580 1608.174 ~314.175

930.047 481.725 o

Temperature (Fahrenheit)

1 British thermal unW

32 + (9/5) Celsius temperature" 1055 joules

1 Used principally py chemists,

11 USed principaJly by engineers.

3 Variou~ definitions are give:n for the British thermal unit. This represents a rounded mean value differing

from none of the mote in:ipo~.nt·defi:nitioil.gby more than 8 in lOt. . . .

* Exact value.

L

r

I'

3. Elementary Analytical Methods

MILTON ABRAMOWITZ 1

Contents

Elementary Analytical Methods , ...•.....

3.1. Binomial Theoremand Binomial Coefficients; Arithmetic and Geometric Progressions; Arithmetic, Geometric, Ha.rmoiiic and GeneralizsdMeens . . . . . . .

3.2. Inequalities. . . . .. , , . . . . ,_

3.3. Rules for Differentiation and Integration.

304. Limits, MaximaandMiniraa ..

3.5. Absolute end Rela·tive EITorS • . 3.6. Infinite Series. ., . . . . . . . 3.7. Complex Numbers and Functions 3.a.Algebra.ic Equations . . . . . . 3.9. Successive Approxima.tion Methods

a.lO.Theorems on Continued Fractions.

Numerical Methods • • .• .. . . ." ~ • • . . • .. • • . •

3.11. Use and ExtensiQn of the Tables 3.12. Computing Techniques ,

References ,. • ,. .:. • • • • •

Table -3.I.Powersa.nd Roots

nt, k=1(1)10, 24., 1/2, liS} 1/4, 1/5 n=2(1)999,Exact or lOS

Pa,ge 10

10 10 11 13 14 14 16 17 18 19

19 19 19

23 24

The author acknowledges the asslstan:ce of Peter J. O'Hara and Kemiit C. Nelson in the preparatio'n and che.eldngof the. table of powers alid roots.

1 NatianalBureau ofStanda:rdS. (Deceased.)'

9

3. Elementary Analytical Methods

3.1. Binomial Theorem and Binomial Coefficients; Arithmetic and Geomet'ric Prngressionsj Arithmetic, GeoDletri.c, Harmonic and Generalized Means

Bmomiat Theorem

3~1.1

(a+b)n=a"+(~) a"-Ib+(~) a"-2b2

+C)a"-3b3+ ... +bn

(n a positive integer)

Bim~mial Coefficient!!. (se~ chapter 24-)

3.1.2

(n)_C' n(n-I} ... (n-k+l)

* k ~n " k!

n!

(n~k)!kl

(~)=(n~k)=(-l)~ (k-~-l)

(n+ 1)=(71,)+('. n )

k k, k-l

(~)=(~)=1

. 1 +(~)+G)+ ... +(:)=2'1 3.1.7 1-(7)+(~)- -. - +(_l}n (:)=0

3.1.3

3.1.5,

3.1.6

Table of Binomial Coefficient8 (~)

3.1.8

~k 0 i 2 3 4 5 6 7 8 _g iO 11 12
~
1 ____ 1 I
2 ____ 1 '2 1
3 ____ 1 3 3 1
4 ____ 1 4 6 4 1
5 ____ 1 5 10 10 5 1.
6 ____ 1 6 ,15 20 15 6. 1
7 ____ I 7 21 35 35 21 7 1
8 ____ 1 8 28 86 70 56 28 8 1
9 ____ 1 9 36 8'4 126 126 84 36 9 1
10 ____ 1 10 45 120 210 252 210 J20 45 10 i
11 ____ I 11 55 165 330 ,462 462 330 165 55 11 1
12 ____ 1 12 66. 220 495 i92 924 792 495 220 66 12 1 For a more extensive table see chapter 24.

• Seep a,ge: II.

10

3.1.9

Sum of ArithlDlltic Prog,ressioD to n Terms

a+(a+d)+(a+2d)+ ... +(a+(n-l)d)

1 n ( ,

=na+2n(n-l)d=2 a:+l).

last term in series=l=a+(n-l)d

Sum or Geometric Progression to n Terms

3.1.10

8r1,-a+ar+ar2+ . _ . +ar"_l=a(!-rn) -r

lim sn=a,!(l-i:)

n-7'"

Arithmetic Mea.D of n Quantities .A

a.i.u

A al+a2+ .. - +a,. n

Geometric Mea.n of 11 Quantities G

3.1.12 G= (alQ.z ... q,,) II"

Ilarm.onic Mean of n Quantities H

(a~>O,k=-1,2, ... ,n)

Generalized -MellO

3.1.14 3.1.15

(1 n )111

M(t)= -~ a:t

n t-l

M(t)=O(t.<O, some at zero)

3.1.16 lim M(t) = max.

t-» ...

3.1.17 lim M(t)-min.

t---J:-=

3.1.18 3'.1.19 3.1.20

limM(t)=G

1-j.G,

M(l)=A

M(-l)=H

3.2. Inequalities

Re1ation Between Arithmetic, Geom.etric, HannoniC! and Generalized Means

3.2.1

A;i G.;;:::H" equality if and only if al' :a2=- ... =a;,

3.2.2

min. a<M(t)<mux. a

11

3.2.3

min. a<G<max. a

ELEMENTARY ANALYTICAL ME'l1HODS

Minkowski's Inequality for SUDlS

equality holds if all al< are equal, or t<O and an a~ is zero

3.2.4 M(t)<M(s) ift<s unless all al< are equal, or 8<0 and an at is zero.

Triangle Inequaliti(l'8

3.2.5

[chl-Iazl::; lal+azl ::;1411 +la21

If;akl:5 ±! ak 1

M .1:-1

3.2.6

Chebyshev's lu.~alit1 If ll1~a2~a8~ ... ~a ..

bl~b2;;:: b3~ . . .~ bll

3.2.7

Bolder's Inequality Cor sums

equality holds if and only if Ib.tl--:c!atl,,-I (c=con-

stant>O). If p=q=2 we get . .

3.2.9

[± a;i:hJ2S;:E.1l a~ ±. H, (equality for a~=cbk.'

. k=.l k-l k-l

C constant) .

Holder's Inequali'ty for Integrale

If .!.+-!=1 . ·>1 >1

P g_ ,p , q

3.2.10

f~ . [ill JVi>[ r~ ] 1/.~

Ja.1j(:i')y(x)!dX$ ,,If(x)I''dx JIJ Ig(xWdx

equality holds if and only if Ig(x)l=vlf{x)lp-l (c=constant>O).

If p=q_=2 we get

Schwan'lI inequalitY

3.2.11

[ib f(x)g(x)dxJ II s f.b [j(x) Nx ill [g(x) Fda;

If p>l and i:h" .b~>O for all k, 3.2.12

equality holds if and only if bt,-·cat (c=ieonstant>O).

Miokowski'8 lnequ.a1itj' for Integral!9

If p>l, 3.2.13

( fll )1/'11 (I.~ )11-:;

Ja If(x)+g(x)I"dx· ~a jJ(x)i"dx

equality holds if and only if g(x)=cj(x) (c=const!}nt>O).

3.3. Roles for DUietentiation and Integration Derivati ves

3.3.1

d () du . t t

-d cu =c T' e cons an

z. u,x

d du dv

- (u+v}=~+dx dx ax.

3.3.3

d dv du

d-'- (uv)=u d--:+v d----"x xx

3.3.4

.!:_ ( /.) vduldx-udv/da;

d· u v 2.

X . V

3.3.5

d dudv

dx '11, (v) = dvdx

3.3.6

d (.") • (V dy, +1 dP)' .

- u· =u -- nud» . u d» dx

Leibniz's Theorem for DUl'erentiation of an Integral

3.3.7

d J:~~CJ

.J~ f(x; c)dx

f.W a(,)

f.b(") () . db da

=. :;.. j(X, c)dx+f(b, c) de' -j(a,c) d-

ace) u~ . . C

12,

Leihniz's Theorem for Dilre~ntiation of a Product

ELEMENTARY ANALYTIC.u. METHODS,

3.3.8

dn ... d"'U (n) d"'-lU d;v, 'en) d"~'J,u, d'lv

dx" ('11//))= dx" v+ 1 dX,.-l dx + 2 a,x,tI-~ dx2

(n) dit-·,u d/» d'.'v

+ .. ,+ . d fI-' d~+ ... +u -d II

T x' x' x

3.3.9

dX=lf!l dy dx

3.3.10

3.3.11

d:x =_[<Py tJ,y -3 (d2y)2] (dy)-6

d'll dx: dx d:rJ dx

Integration by Parts

3.3.12 Judv= uv·- J'VdU

3.3.13 f uvdx= (fUdx ) v- J (fUdX ) ~~ dx,

Integrals of Rational Algebraic Ftuu:~tioJ1fl

(Integration oonstants are omitted)

(n~-l)

3.3.15

f dx 1

+b=-In lax+bl

ax , a

The following formulas are useful for evaluating

J P(x)dx h PC)' 1 '_1 d

(ax2+b:t+c)!I were x 15 a po ynomlwl1D

n> I. is an in teger,

3.3.16

f dx 2 ,2ax+b

(a:!f+b;l:+e) (4ac_b2),jarctan (4ac-b2)f

(b2-4ac<O)

1 In 12ax+b- CbZ-4ac)tl

(b2-4ac)t 2ax+b+ (b2-4ac)t

(b2-4ac>O)

3.3.17

-2 2ax+b

3.3.18

3.3.19

f xdx 1'2 bJ dx

a:!f+bx+c 2a 10 lax +bx+cl-2q. a:z:2+bx+e

3.3.20

J dx. _=_l_ln Ie +dxj (q.+bx) (c+dx) ad-'be a+bt

JdX 1, bx

3.3.21 a2+ b2x2 ab arctan--;;:

3.3.22

3.3.23 f a2~~2x2 2~b 10 1~+t~1

3.3.24 I (X2!tall)2 2~3 arctan ~+2a2(;+a2)

3.3.25

Integrab of Ir.rational Alsebraic Functions

3.3.26 •

J dx 2 [-d(a+ bX)]1/2

I(a+.bx) (e+dx)JI/~ (-bd)V2 arctan b(c+ax)

-J .' (2bda:+oo+ bC) (b'>O. d<O)'

(-bt.l)1l2 arcsm be-ad '

=(b:)1/2ln l[bd(a+bx)p12+'b(c+dx)I/~1 (bd>O)

I dil 2 - [d(a+bX)]1/2

(a+bx)m(c+dx) [d(bc-;-ad)j1/2 arctan (bc-aiC)

3.3.27

3.3.28

3.3.29

(d(ad-be) <0)

3.3.30

1 . ld(~+bX)1/.2-[d(ad-b~)IIl!l1

[d(ad.-bc)pI21n d(a+bx)1/2+[d(ad-bc)pI2

(d(ad-be) >0)

3.3.31

f [(a+bl:) (C+d:x)]l12ax

(ad-:bC)~~b(c+dx) [(a+bx)(c+dx)]lIZ

(ad- 01.1)2 f ,dx

- Sbd ,Ha+6x)(c+dx)]11ll

3.3.32·

J[C+dXJ,V2, 1",

a+b:c dX=b [(a+bx) (C+dX)]112

(ad-Oc) f dx

26 (a+bx)(c+dx)]I'.\I

3.3.33 f(az2+~:+C)lli

=a-1/2ln !2al/~(ax'+bi+c)1/2+2ax+bl.(a>O)

3.3.34

.

-a.-liZ arosinh (2aX+b)

, .,' ' (4q;c-b2) 1/2

(a>O,4ac>b2)

=a-1/2ln Izai+bl(a>O, b'-4ac)

( ) l/~ . (2ax+b)

=- -a - "arCSill " ,

, , (li2---,4aC) In

(a<O, b2>4<u:, 12a~+b'I«b2-4ac)Uf)

3.3.35

a.3~36

3.3.3,8.

f dx f .dt h '

, x(ax2+bx+c)lIi- (a+bt+ct2)l12w ere t=l/x

3.3.39

f xdx

(a~+bx+c)ll2' . ~" f

~-- -- "-rl (' x2+b+ )1(.2 b. S', d,x,

~ -2a (a;t:Z+ bx+a) 1/2

3.3."40 f (x2:a2) & In 1:&+ (x2±a~il

3.3.41

3.3A~2

3.3.44

f" dx 1,a

, .. ' arccos-

x(x2-a2)1 a x

J dx . x

C' IJ' '2" 1 aresm+ a -:I; i' ,a,

3.3;46

3.3.47

3.3.48

. x-a arcsm -, -~

,," a

3.3.49

f' dx

, (ax2+b)(cr+d)t

1 ,". , x(a:d-bv)i

[b(ad-bc))' a!_ctan [b (cx2+d) Ii

3.3~50

1 , In I [o'(cx2+d">'li+ x (bC-ad)il _ 2[b(bc-ad)]1 ' [b(CX2+d)li-x(bc-ad)!

(bc>ad)

3~4. Limits, Maxima and Minima,

l:rI,d~tenninateFOl"in,!i (L'HospitaI's lui,e)

3.4.1 Let f(x) and g(x) be difie1'entiable on an interval as x<b forwhichg' (x) ¢O.

If

or If

and,if

limj(x)=O and lim g(x)-O

,Hb- ~-"

1imj(x)~ <P and lim g(rt) = ,9')

~~- :t~-'

.Iim f',(i)=l then lim j(~) =~. Z'4b- q (x).r-~b:- g(z)

Both b and l may be 'finite or infinite.

Muima and. -Minima

ELEMENTARY ,ANALYT~tAL lI1E'l':E:ODS

.Approximate . Value"

3 5 2 (b) th "l t-' - . '. f .. _. ax.6.x

•• _. . €lore .a. ·_~we error 0 .: Xo IS u.X=-~-

:c Xb

3.5.3 (c) the 'percentage ertrjr is 100 time;sth.e 3.6.5

relative error.

14

3.4.2 (1) Fu,nctions oj():ne Va?"iable

The function j, j(x) has it 'maximum at -:P=-XQ if j'(xo)=O and f"(~)<O. and a minimum at :1,=000 if f'(Xo) =0 and _f"(xo»O. Points :to for which l' (xu) ~o are called stationary points.

3.4.3 (2) Functwns oj Two Variables

The functionJ(x; y) has It maximum or minimum for those values of (xo. Yo) for which

oj' ?J."

O~=O'I ~=O,

. _ '. I 'rij/ox(yy

and for which , ?;rjj'iYy2

(a) fez, y) has a maximum

.0'1· ,?l'j .'" "

If b~<O and by2<O at (xo,·Yo),

(b) }(x1 y) has a minimum

. ?Jif ,.' o2j.

if or>Oand Oy2>O at (x!). Yo).

3.5. Absolute an,dRelative Errors

(1) If Xo isanapproximation tothe true value of z, then

3.5.1 (a) the absolute error of xti is .:l;l;=Xo-X, x - Xo is the correction to z,

3.5 .. ~ (2) Thf>absolute error of the sum or differenceof several numbers is n.t_ most equal-to the 'sum of the absolute errors of the individual numbers.

3.5.5 (3) If j(Xl, X2, ••• ,x .. ) i~a. function of (tI,X2, •• 'J x" and the absolute error in :i:~ (i=l, 2, .. 11.) is .6.XI, then the absolute error injis

3.5.6 (4) The relative error of the product or quotient of several factors is at most equal to the sum of the relative errors 6f theindividual. factors.

3.5.7

(5) If y=j(re) I the relative error _By=~:,~"J;/~2 ax

3.5,.8 3.5.9

(a + b)t F"$(j,~+ kal·-1b (l+E}(l+l1) ~l+e+1] i+E 1+'I]~1+~-'1

3.5.10

3.6. Infinite Series

TayloJ:"'s Formula fora Single Vliriable

3.6.1

. .' hZ .

j(x+k)= f(x)+hf' (;I;) +2! j" (x)

h,,-1

+ +: .: r:» (. ) +R' . ... .. (11.-1)'1 .:v"

3.6.2-

R.'n,=:.'·"',.'j_ (n)( ,+e'lli) Ii)' (1 I) )ii1j<l( +e: h)

'" . (n-l)! . - i - ..... x ." 2 •

(0<01• z(x) <1)

3~6.3

"/" (I (l-t)n-lj\ri)(x+th)4t

(n-1)IJo .

3.6.4

f(x)=}(a)+ (x 11 a) j'(a)+ (x;aPjll{a) + (x-a)?"~lj(n_I)(' )+R

... + (1/;-1)! .. a·"

(a<~<x)

La.gr.ange'-s EXpansion

Ii- y j(':JJ), y_C)j(xo), I' ($0) #-0, then 3.6~6

. ~ (:Y-Yo)" f elf-I { a-xo } "] x=xo+ .f={ . kt . Ldif-1- j(x.) -Yo'p~"o

3.6.7

.. ("II . )1'[ A~l ( {'" "". y')]

- ',1;1-110" W", , '.. ""~wo

+f-i k!dw-19 (z) j(z)~yo .' ;1;;".1"0

where g(3:) is any function indefinitely dlfferentiable ..

Binondal Series

3.6.8

(l+x)a=t (~) ill'

ELEMENTARY ANALYTIC:AL METHODS

15

3.6.9

~_ 0:,{0:-1) 2 0:{0:-1)(a-2).

(l+x) -l+ax+ 2' . x + 31 x3+ ... ,

3.6.10 {1+x)-1=1-x+;t2_x3+x!l- ...

3.6.11

z XZ XS 5x4 7x~ 21x~

(1+x)t=1+Z-g+i6-12S+256-I024+" .

(-l<x<l)

3.6.12

(1+ )-.=1_£L 3X2 _5x3+35x4_ 63x6 . X . 2' 8 16 128 256

231x8

+ 1024 - . . . (-l<x<l)

3.6.14

(1 +:i)-t'=~-~ x+~ x2~ ~~ xa+2!~ x{

91. 5 728 fi

-729 z +6561 x - ...

Asymptotic Expansions

""

3.6.15 A series ~ a;!:x-k i:;l said tQ be an asymp-

~~o

totie expansion 01 a, function lex) if

'ii-I

j(x)-L:J ak,r};;=O(x-") as X--H~

k=O

for everyn-::l, 2, . . .. We write

CD

}(x) '" ~ a,tx-l<.

The series itself may be either convergent or divergent,

Opera1:iO,ll8 Whh Series

Let sl=l+a,x+ai;c2+aaxS+a"x'+

82= 1 + b,x+bir+baX'+b,i'+ . . . 83=1 +~lx+C2r+C3w+c~x4+ ...

Operatien ,Cj Ci c~ C:t
3.6.16 83=.sl' -a. a~-a2 2alfl.J-a3-,a~ 2ala3-3~a2-ll;.+~+i:!_t
3.6.17 . -4 -2a. 3af-2aa 6a.a2 - 2a3 - 4a~ 6ala3+3a~ - 2a. -12afa2.+5af
8g·='8,
3.6.18 sa=slf 1 1 . 1 2 1 1 + 1 ~ 1 1 1 '+ 3 l 5 6
-2a1 2a2-gal 2a1'-4al~ 16al za'-ia,aa-gat 16l',a1-128al
3.6.19 Sa=sl"H 1 321 3 1 5_ 33 2 1 1-5 t. 35,
-2Q• -a.,--ai -aja2 - -as - -'ct, ialaa+jG2-2a.I-16alat+ 128a1
8 2 4· 2 16
3.6.20 Sa:=8~ nal 1 . ,cLG3(n-l) na.-+!-cla,(n-l) + ~n(-n-l)a~
2(n-1)c.a1+na2*' + !c"aHn-l) (n - 2) ,.
.6 +na~ +~ ('110 - 1) (n - 2)'CIa.lllJ
:2 .' .
1
+ 24.{n -1)(11.,-2) (n--3)cla!
3.6.21 Sa=811la G.+b) b2+a.bJ +a2 ba+a..b2+a,b.,+aa ·bl+al~3+0gb2+a3b. +a,
3.6.22 8a=sds2 al-bl 02 - ~blC) +bt) 'a8- (blC2+bJcI+bs) OJ - ~blca + b3c: +haCI + hi)
3.6.23 lIa=exp (II. -1) a. +12 03+ a.CZl + ~aT + +1t+1 ,+1 ..
a;, 2G1 a. alaa' 2a, 2;02a, 24af
3.6.24 8a=1+1n 81 1 1 a.1 - i(a301 + 2a~+ 31l1ea) ill
a) ai-2!lIC1 aa -~(a2el + 2ajcil) 16

Reveraion of Series

ELEMENTARY ANALYTICAL METHODS,

Polar Form

3.6.25 Given y=ax+bz2+oxa+dxl+ex~+fx6+gx7 +

then

x=Ay+B?f+Olf+DlI+E1t+Fy&+GyT+ ...

where

aA=l

a,sB=-b

a,sO=2b2-'M

a7D=5abc-a2d-'5b3 aOE=6a~bd+3a2c.l+14.~-ase-21ab2e all F= 7 as be+ 7 aBed +84q.b3c - a'J

-28d2bcz-'42bli-28a2b2d

a 1$0= 8a}bj +8a'ce +4a'd2+ 120aWd

+ 180a;Jb2c2.+ 132bO-a~g-36aah2e -72a3bcd-12aaca-3.30ab~c

Kwnm.er's Tl'llDJIformation of 'Series

CD

3.6.26 Let>:, ak-sbeagivencQDvergentseriesapd t=O

..

'5: Ck-C be a given convergent series with known

t=O. '

sum C such that lim a.t=~~o.

IH ... Ct

Then

Eult'l"I"1! _Transformation of Series

'"

3.6.27 If:E (-l)l:,o,l:=ao-al +~:- ... is a con- 1:"0

vergent series with sum s then

Euler.Maclaurin Summation Form. i.lIa

3.6.28

'11-1 fro 1 1

{.;tfk= s. J(k)dk-2 U(O) +f(11.) 1+12 [1'(11,)-]' (0)]

__ 1 [J"'(11.)-f'''(O)]+_l_ [pVl(11.)-j<l'l(O)]

720 .' 30240 '

_ 1 [I(vm,,) j(vnJ(O)]+

1209600 ' ,n - . . .

3.7. Complex Numbers and Functions

Carteeian Form.

3.7.1

3.7.2 3;7.3

z='re'·~=r(C06 9+i sin 8) Mod'!flus: lei = (XZ+y2)t=T

3.7.4 Argument: i1rg z=arctap (yjx)=8 (other notations for arg z_are am a and ph a).

3.7.5 3.7.6

3.7.7 3.7.8 3.7.9

Real P8J.'t: x=f!lz=r cos e

Imaginary Part: y=..Fz=r sin (J

CumplcJ: Co .. j,ugate of ~

-

arg z: = - arg .2

Multiplication and Division

3.7.10 3.7.11 3.7.12

3.7.14

3.7.15

3.7.16 3.7.17

3~7.18 3.7.19

3.7.20

POW~I!

=r" cos n8+irn sin n(}

(n=O,± l,±2, ... )

z2=r-y+i(2zy)

il=i'-3x1f+i(3:z::ly-yS)

z'=i;~- 6xY+yf+i(4ry-4:r;if)

3.7.21 z!i=xii-lOx-3y2 + 5xy'+i(5x'y-l0x~'if+y5) 3~7.22

Z"=[ZIJ_(~)Z"-V+(~) x"-4y'_ ... J

+i (7) i"-ly_{~) x"-3y3+ ... ],

(n=I,.2, ... )

ELEMENTARY ANALYTICAL METHODS

3,. 7.23un+t . X1.tn-YV .. ; V ,,+,1 =:tVn + 1/U"

f1l gtl and J Z'.' ere called harmonic polynomials.

1 . z x-iy

t: IzI2=XZ+y2

3.7.24

3.7 .. 25

Boola

3.7.26

Ii -'I!'<8~'I!' this·is the principal root. The other root has' the opposite sign. The principal root ~ given by

3.7.27 z'-=[i(r+x)'j'±i['Hl'-x)]l=u±iv: where 2u'1I=Y and where the ambiguous sign is ta.ke_n to be the same ELS thesjgn of y.

3.7.28 zli"=rflne~8fn, (principal root if' -11"<8$:11'). Otherroots Me rtbietCH2.-.tlf1l (k=l, 2,3, ... , n-l).

Inequalities

3.7.29

Complex FlUl-ctions. Cliucby-Riem'lIllD ,EquatioDs

1(2) J(x+iy)=U(i,y)+iv(z, y)whereu(x,y),v(x,y) are real, is analytic at those points z=x+iy- -at which

3.7.30

If z=rel'8

t.

3.7.31

Laplace'!! Equation

The functions u{x,_ y) and v(x, y) are called harmonic functions .andsa.tisfy Laplace's equation:

Cartesian Coordinates

3.7.32

Polar Coordinates

3.7.33

17

3~8. Algebraic Eq'Q.ationa

" Solution 0' Qua:dratic Equations

3.8.1 Given az2+bz+c=O,

(b) 1. -

Zl,a=- 2G ±2a gl, q=b2-4ac.,

.zl+za=-b/a, Zl.z2=c/a If q>O, two real roots,

fJ. 0; two equal roots,

q<OJ pair of complex conjugate roots,

Solution of Cubic Equations

3.8.2 GiveIit+<Z-.i~+alz+ao=OJ let

q_=! a _! n~.·T=! (an-- 3n-) -!. n!. . 31 9.-' 6 1""" '"\J 27 -~

If t+r2>O, one.real root and 8; pair of complex conjugate roots,

0+.,&=0, all. rootsreal and at least. twoare equal,

1+,-2'<0, all roots real (irreducible case).

Let

then

If ?i, 22, Za are the roots of the cubic equation al +Zg+ Z3=-a2

S91lltion of Quartic Equations

3.8.3 Given ~+a8z2+lZ'lIzz+<ltz+ao=O, find the, real root Ul of the Qupic equation

U3_~1b2+ (a1t'ta:-4a:o)u- (a~+aoa:-4a0Ct2)=O and determine the four roots of the quartic as solutions of the two quadratic equations

v~+[~3=F(~+ul-a2YJv+ ~1~[(~1)2 -aoT~o

l~

ELEMENTA.l\Y AN,ti.YTlCAL METHODS

Method of lterati.on. (Succetl8ive Substitut'ion)

Han roots of the cubic equation arc real, usc the value .o'f HI which gives real coeffi.ciGltts in tIlt' *ql,ladrn:tic equation and selectsigns so that if

z'+a3i'+<l2z2+aIZ+ao= (Z2+PIZ+ !b) (z2'+ p;z+qs), then

3.9. SUCCe8sbe Approximation Methode

Geu.eral Comment8

3.9.1 Let X=XI be an approximation to:t~E where fW=O and both il:1 and E are in the interval

a:S;x.;E;;b. We define e

(n=l,2, ... ).

Then, if l' (x)';;:: 0 and the constanta c" are negative and bounded, the sequence x'" converges monotonically to th¢ root~.

I!c".= c= consta.nt<O aad f' Cx) >0, then the process converges but not necessarily monotonically,

D,egree o.f CODvetgenee of an A.PPl"oximn.tion P~88

3.9.2 Let XII x:,xa, ... be. an Infinite sequence of approximabions to a Dumper f. Then, if

(n=1,2, ... )

where A and ·kare independent of-a, the sequence is 'said to have convergence of at' most the kth degree (or order or index) to E. If k = I and A<ltbe convergence is linear; ifk-2 the convergence, is quadratic.

R.~guLt Falsi (FaiR Position)

3.9.3 Given y=/(x) to find .~ such that f(~)-O, choose Xa and XI such that !(:Co) und !(x]) huve. opposite signs and compute

Then continue with $2 and either of Xo orXj for which f('rQ) or !(Xi) is of opposite sign to f($2).

Regula falai is equivalent to inverse linear inter-

polation. .

3.9;4 The iteration scheme tk+1 = F(;r,,,) will converge to a zero of :1;,= F(x) if

(1) IF'(x)l:S;q<l for a:S;x:S;bl

(2) a<x(l±IF(xo)-~ol<b.

-. 1-q -.

Newton's Method of Slici:.eil8ive ApproQl'D.ation.s

3.9.5

Newton's Rule

If X=Xk is an approximation to the solution x=E of !(x}=O then thesequenoe

willconvergs qua.dratica.lly to x=E: .(if instead of the condition (2) above),

(1) Manotonio convergence, f(XO)J" (xo) >0

. .

and l' (x), I" (x) do not change sign. in. the

interval (~, £), or

, .

(2) Oscillatory convirgence, f(Xo)l"(xo) <0 and f'(x), fll(X} do' not change sign in the interval (XO,:I;I), :tQ:5~:5Xl'

New ton 'II Meth.od Applied to R_I nth Roote

3.9.6 Givenxll=N, if x" is an approximation z- Nil n then the seq uenee

will converge quadra:ticaUy to z.

. 1 (N )

Ifn=2, :l;Hr=Zx_.t+;Z;t ,

- Aitken's 1i~·Proee8e for Accelera.tion of Sequence8

3.9.7 If XtJ X,t+h XJ;+.2 are threesueeessiveIteeates in a- sequence converging with an errer which is approximately in geometric progression) then

(Zk-Xk+l)2 l:J.2Xk

X~XHZ~Z:+l. 62xl:

is em improved estimate of z. In fact, if Xk=:i:+* O().~) then i-x+O(,-I;), 1'-1<1.

3.10. Theoremson Cont,~Dued Fra.ctioDS

ELli!MENTARY ANALYTICAL METHODS

19

Definitio~

3.10.1

(1) Le~

If the- number of terms is finite, f is culled n, terminating continued fraction.. If t.he. number of terms isinnniteJ is called un infinite continued fraction and the terminating fraction

j A"b+(l.t «ii, ~ ".= Bn, =" b1+ bz+ ... b,.

is called the nth convergent 0.( J.

(2) If lim BAn exists, the infinite continued frao-

'n-+m -n 'ill.

~ionj is said to be convergent. If aj=l and the "tare integers there is alwuys convergence.

Theorems

(1) rIaj and Q1 lLfe positive. then 12 ,,<}ZlI+2'1 J'zn.-l>hn+l'

(2) tfJ,.=4n/

B ...

A",=b"A"_-I+a.nA"'-2 Bn= b7lB",-I+a:"B,,-a where A_I=l, Ao=bD,.B_1=O, Bo=1.

(3)

(4) (D)

/I

AnBn:_t-An_1Bn: (_l}":-l II Q,k k=1'

For every n~ 0,

J,,=oo+ cia, Ci.C2a., .c~C3aa ... ,C" . ...,lc;,a"

. c1b, + c:lb2+ c3b3+ c,.b."

(6) 1+ba+b~b3+'" +blloj .. , b ...

1 . b2 ba b"

1- &2+1- ba+1- ... -bll+1

.!.+..!_+ ... +.l=_L ~. . , . U~-l

ui "U2 u" "U1- 'l,tt+U'2~ -Un.-l+'!f·,.

l_~+~ ... +(-1)" . x" .

a(l aoalaOalCZ:! ·fLofLla.2, •. an;

1 aoxalX an-IX'

- ao+ al-x+ tl:!-z+ .. , +a,,-:t

c

2',.0

I.B.

1 .• :6

1,0

:1iI

.S

.a

FIGti'REl 3.1. y=x". 1 1 ±n=O,·s' 2' 1,2,5.

Numerical Method~

3.II. Use and Extension of the Tables Example I. Compute zlYand x·1 for x=29

uaing Table 3.1. .

= (1;45071 4598.Wl') (4,.20707' 2333 ·lOIt) ~6.10326l-248, 1.0:7

X47=(iiP)2fx

__ (1.25184 9008·10S5)z/29 =5.4.03882547 ·10~

Exam:ple 2. Compute X-Bl~ forx=9.19826. ·(9.198:;l6}114=(919 .. 826/100)11'= (919 B26)l/i/lOt

Linear interpolation m Table 3.1 gives. (91 9.826)1/4 ~_5.507144 .'

By Newton's method for fourth roots With

N=919.826,.

![. ' .• 919.826 +3 (55.o7144).J-5 5.07143845

4 (5.,507144)3 ... __ .' .

Repetition yields the same result. Thus,

xil:-=5.5.o714 384.5/1Ql= 1.74151 17'96,

.:ir3/~ = xi /3: = .189:)3 .0568.3.

3.12. Computin.g Techniques E.x'aDlple_ 3. Solve the quadratic equation zZ-lB.2x+.056 g-iven the coefficients as 18.2±.l,

~See PAge n.

20

ELEMEN'TARYANALYrriCAL METHODS

.056 ± .00 1. From 3.B.1 the solution: is, x=l(18.~± [(18.~)2-4(.05fi) ]t)

=H18.A± [,331.016]1)=1(18.g± 18.19,39) = 18.1969, .OO~

The smaller root may be obtained more accurately from

;4< .05,2/18 . 19.6\t= .0031 ±, .0001.

Example 4. Compute (-3+.0076i}i.

From 3.7 .26, (-3 +.0076i)l=u+iv where

Thus

r-[( -3)~+ (.0076):ll]!= (9,00005776)1-=3.000009627

v_[l.OOOOO Q~27-(-3)T=1.7~205 2196

y .0076 " ,,' .

'It 2v2(1.732052i we) .00219 392926

We note that the principal square root- has been computed.

Exa)nple 5. Solve the cubic equationa;3-18.1x -34.8=0.

To USB Newton's method we first form the tal>ll;l of f(x}=x3~18.1x-34.8

fez) 4 -43.2 5

6 7'

x,

72.6 181.5

We obtain by linear inverse interpolation: 0-(-.3)· xo=5+72.6_ (_ .3) =5.004.

Using Newton's method, PCx)=3x2:"'_18.1 we get Xl'''''''Zo- f(xo)/!' ($1))

5 0'0'4 (-.07~_15 9936) ·~.5 0'05'26

"'" " 57 .02O_O~8··

Repetition yields ,x~=5.00526 5097. Dividing f(x) by ;1.:-5.00526 509,7 gives x.a+5.00526 5097x +6.95267869 the zeros of which are -2.50263 ,2-549 ± .83036 ·800i.

Example 6. Solve the quartic equation x4-2.37752 4922:);8+6.07350 5741x2

-11.J7938 023'x+9.0526.5 5259=0.

Resolution In'lo Quadratic Factors

(xl,+ PIX-f--qlj (v+ P~+ql) ,

by,lnverse lnterpoia:tion'

Starting with the trial vIl1ueqi = I.we .cempute successively

1 9.053 -1. 0,93 -1. 284,
'z 4. 5Z6 -2. 543 .165
2, 2 4. 115 -3.106 .729 5.383 :032 -2.02$

We seek thtl,t value of ql for which yf!b) = o.

Inverse interpolation in ,Y(Ql) gives y(ql) ",,0 for. 21:::::2.003. Then,

ql

2.003 4. 520 -2. 550 .172 . ,011

.Inverse in terpolabien oetweent'li =2.2 and qt = :2 .. 003 gives ql:_____,2.0041, find thus,

P'l

.00078 552 .00001 655 -.00075263

a 00>11 4.51706, 76'40 2.0042 4,516:84. 2268 2. 0043 4. 5Hi61 6903

:...2.5'5259257 . 17506 765 -2.55282851 .175,30 358

-2. M306~47 .17553 955

Inverse interpolationgives ql =2.00420 2152, and we get finally,

I

PI

<,22

2.00420 21.52 4. 5168;3 7410 - 2. 55283 358 . 17530 8659: -.00000 onu

ELEM,EN'I'A~Y ANAt'YTiCAL MJ!iTaUJJ~

Texts

References

[3.1] R. A. Buckingham Numerical methods (Pitm:!f,D Pu.b1,is'biil.g O\lrp., New York, N.Y., HIB7).

[3.2] T. Fort, ,FinJt~ di:ff'erences (Clarendon Preas, Oxford,

England, 1945). '

[3.3J L. Fox, ThEl use ~n:d cpnstruction of matberaatical tables, Matbema.t,i,clll Tables, vol. 1, National PbysicalLa.boIatory (Her Ma:jesty's Sta.tionery

Office,Londo,D, England, 1956). '

[3.4] G. :II. Hardy, A course of pure mathematics, 9th ed. (Gsmbridge Unrv.Press, Cambl'icige,England, and The Macmillan Co., NewYork, N.Y., Hl47). [3.51 D. R. Hartree, Numerical analysis (Cl8.rendon Press, Oxford, England, 1:952).

[3.61 F. B. Hildebrand, Introduction to numericill a.na,lysis, (McGra.w-Hill Boak Co" 11lC., New York, N.Y., 1956).

[3.7] A. S. Householder;, _Prl!;wjpl_e,!l;of.nUIIierical analysis (McGraw-Hill Book Co., Ine., New York, N.Y., 1953)_

[3.8] L. V. Kantcir0witschaiid V. LKrylow, N iihetung:smetnodender H.oh~r~n A~lysjS ('fEB Deutscher Verlll,g der Wissl;lns(1)a,lten,' Berlin, Germany, 1956;tra.nsla:tedfrom Ruaai&n, Moseew, U.S.S.R" illm. .

[3.91 K. Knopp, Theory and application of infiniteeeriea .. (Bla.ckie and Son, .. Ltd., London, Englal).d., 1951). [3.10] Z. Kopal, Numerical-analysis (John 'Wiley &·8ons,

Inc., New York, N.Y., 1955).

. t3.111 C. Kowalewski, Interpolationund genaherte Quadratur (E. 'G. TeuPI}er, Lelpzlg, Germ.any,. 1932). [3.12J K. s, Kunz, Numerical analysw (McGraw-Hill Book Co., Inc., New YQik,N.Y., 1957).

[3:13] C. Lauczos, Applied analysis (Pre;n:t;ice_;.Ha.Il,. Ine., E.nglewQpd Cliff's, N.J.,. 1956}.

[3.Ul 1. M,Langman, Note oua method for computing . infinite ioteg;ra.ls of oscifutory functioas; -Proc. Oambridge Philo.s.SOc. ,52, 764 (1956).

[3;151. S, E, Mikeladze, .Nnmerical methods of mathem-aticallitJ.e.lyeis (Russian) (00$. lzdat. Tehn.Teor.Lit., Moscow, U.S.S.R., 1953).

l3.1aJ W. E. Milne, Nu,merica.l ca.lclllu8 (Pri.nceton Ilniv, Press, Princeton, N.J., 1,949).

[3.17] J:,. M. rv,[ilne-'rhoIIlisoIl, Thecalc.ullls fJ£ finite differ~enee8 (Mli.omillB.Jl.,and Co., Ltd., LendonEnglsnd, 1951).

(3.18] H. 'Mineur, Techniques. de .ca.leul numerique (Librairie PolytechniqueCh. B~ranger, Paris,

France, 1952). .

_l3.19] National Physical Labora.tdry, Moderncomputiog methods, Notes on Applied SCience No. 16 (Hei:

Majesty's Stationery . Offi.c6,LbD'don,Engl:a.nd,

1957). _

[3.20] J. B. Rosser, Transformations' to speed. the OO[]vergenc.e fif series, J. Resea.rch. NB8"46_, 56-64 (195'1)'.

[3.21) J. B.Sca.r.borough, Numerical ma~hematical analysis, ~d ed. (The Johns Hopkins Press,Ba1tirriore, Md.; Oxford Uoiv. Press, London, England, 1.95.5).

[3.221 J. F. Steffensen, Interpolation (Chelsea Publishing Co., New York., N.Y., -1950).

[3';23] H. S. Wall, _Analytic t_heoryof continued fraatiolls, CD. Va.n Nostrand ce., Inc., New York, N.Y., 19~~ .

[3.24) E. T. Whittaker a.nd G. Bebinaon, Tbe ealculus of . obeervatlons, 4th ed, (Bla_ok;ie and.so~, Ltd., London, Engl8Jld, 1944).

[3.25} R. ZU):mUhl,· Praktische Ma.1ihema.tik(SprlngerVerlag, Berlin, Qetma.ny, 1953). .

Mathematical Tablea.and CaUectiOD,8 of Formulaa

[3,261 E_P. Adams, Sniithaornan ma.~hetIl8.tical formulae and ta.Qllls of elliptjc f~ctjons,3d rep~t (The Smithsania;n Instit-utionr Wl:lflhingtan, D.e., 1957).

[3.271 L. J. Cotncie, Barlow'l? tables or aqua-tee, cubes, !!qu:a.reroots, cube roots and reeiproealll Df all integel'S upto 12,500 (Ohemical Publishing Co., 100.,. New York, N.Y., HiM).

[3.281 H. B.. Dwight, Tables of integrala and other mathematica.l data, ad ed. (The· MacmiJla.n CQ. .. ,New

York, N.Y., 1957). .

[3.29] Gt. Brita-ill H.M. Nautical Almll<nac Office, Interpo!&tion and alliedtabl~s (Her Majesty's Stationery Office, London, England, 1966) ..

[3.;30J B. O. Peirce, A shortta.ble of int~gr&l.ai4th ed. (Ginn end 00., Boston, -MBBB., 1956).· ..

13.31.] G. Sch:uh, Formelsarnmlungzur praltt4sohen Mathematik (de Gruyter and Co., Berlin, Germany,

194p),. .

4. Elementary Transcendental Functions Logarithmic, Exponential, Circular and Hyperbolic Functions

RUTH ZUCXEIl,l

Contents

Mathematical Properties. . . 4.1. Logarithmic Function. 4 .. 2. Exponential Function. 4.3. Circular Functions . .

4.4. Inverse Circular Functions. 4.5. Hyp arb olio Funetione . . . 4.6. Inverse Hyperbolic Functions

~s.ge "67 67 69 71 79 83 86

Numerical Methods .

4.7. Use and Extension of the Tables .

89 89

References . . . . . . , . . . . . . .

93

Table 4.1. Common L_ogarithms (100 ::;x,~ 1_350) loglO Xi x=100(1)1350, 10D

Table 4.2. Natural Logarithms (O~:z:~2.l) In XI x=0(.001)2.1, 160

95

10C)

Table 4.3. Radix Table of Natural Logarithms. 114

In (l+x)J -In (I-x), x=10-"(lQ-n)10-n+~, _'n-=10(-1)I, 25D

Table 4.4. Exponential Function (0 ~ 1:z:1 ~ 100). .

e"', ±x=O(.OOl)l, IBD) x=0(.1)5, 15D

x=5(.l}lQ) 12D, -:1:=0(.1)10, 20D ±:r;=O(l)lOO, 198

116

Table 4.5. Radix Tabl-e of the Exponential Function . . . . , . , . . 140 e", «=, x=iO-n(1O-:n)10-,.tJ, 71.=10(-1)1, 250

Tahle4.6. Circular Sines and Cosines for Radian Arguments (O:::;x ~1.6). 142

sin z, cos x, 7:=0(.001)1.6, 23D .

Table 4.7. Radix Tabla of Circular Sines and Cosines ..... , , , . 174 sin a:, cos x, x=10-"(lO-"')10-lI+1, n=10(-1)4,25D

Table 4.8. Circular Sines and Cosines for Large Radian Arguments

(0 SX ~ 1000). . . . . . . . . . . . . . . . . . . . . . _ 175

sin x, cos x, :t=0(1)100, 23D, :t=100(1) 1000, 8D

1 N ation.aJ Bureau of Standards.

61S

-4~lementary 1 'tans cend ental Functions Logarithmic, Ex:pon~ntial, Circular and H yperholio Functions

4.1. Logarithm.i.e. Fune.tion

Mathematical Properties

Logarithmic Identi ties

lntep-al Rep~ntatiqn

4.1.1

In e= f! dt J1 t

-'I"

!y

o

FIGURE 4.1. Branch c'Utjor ln a and :f. (II not an integer or eero.)

where the path of integration does not pass through the origin or cross the negative real axis, In z is il. single-valued function, regular in the a-plane cut along the negative. real axis, rea] when e is positive,

4.1.2 In z=ln r+iB (-.-<BS;").

4.1.3 r=(:z;2+y:l)t, x=r Cos 8, 'y=.r sins, B=arctan '!!.. , x

The generallogarithinic function: is the manyvalued function Lns defined by

4.1.4

1'dt

Ln Z= -

1 t

where the path does not p~8 through the origin.

4.1.5

Ln (1'tll)=ln (refl)+2kwi=In r+i(O+2k?r),

k being an arbitrary integer. In z is said to be the principal branch of Ln z.

4.1.6

Ln (Ztzz}=Ln ZI + Ln Zz.

(i.e., every value of Ln (ZlZ2) is one of the values of LB ZI+LB Z2.}

4.1.7

In (zlz:z}-ln zd~ln Z'l

(-'II"<arg zl+arg Z2::;'1I")

4.1.8

4.1.9

In 21=ln ZI-In Z2 Z2

(-'II"<arg Zl'-arg zz:$1r)

Lnz"'=n Ln z

(n integer)

4.1.10 4.1.11

In zl'=nln z

en, integer. -1f'<n arg z$7r)

Special V s.lues (see chapter 1)

In 1=0

4.1.12 4.1.13 4~1.14

In, 0=- co

In (.-l)=ri

4.1.15 In(±i)=±tn-i

4.1.1~ In e=l, t is the real number such that

i~~t=l

. ( 1)~ .

4.1.17 e=lim 1+- =2.'71828 18284 ...

71-+'" n

(88e4.2.21)

Logs.rithms to General Base

4.1.18

log, a=iln zjln a 10 z=logb Z

g" log~a:

4.1.19

4.1.20

. 1

logab=-l - og~a

log~z=ln z

4.1.21

4.1.22 loglO z=1n ,zjln lO=loglO e In z'

=(.43429 44819 ... ) In z

67

68

ELE¥ENTARY TRANSCENDENTAL FUNCTIONS

4.1.23 In z,-ln lOIog1oz__:.(2.3Q25850929 .. .) Iog1oz (log. x=ln z, called natural, N apierian, or h_yper. bolie logerithms; log., x, called common or Briggs

logarithms.) ,

Series: Expansions

3x ,II.n (l-x)l< '2

4.1.35

(O,<x:$.5828)

4.1.36

In x:$x-l

4.1.37 In x:Sn(xll"-l)for any positive.r,

4.1.24

In (1+z)=z-!z2+lzs- ...

(l,zl:$1 and 2~-1)4,.1.38

I]n (l+z)I:$-ln (I-rzl}

(x>O) (1?1<1)

4.1.25

lnz=(': z 1)+~(Z /Y+~(z/Y+ ...

(~z~l)

4.1.26

In z=(z-I)-Hz-I)ll+t-(z-l)B- ...

(lz-1Istl . 2~O)

4.1.27 .

In Z=2[(~+i)+~(;+~J+k(:+~Y+· .. J

(&lz~ 0, ~ Z ¢iO)

4.1.,28 In (:+~)=2 (~+3~3+5~6+ .. .)

(lzl~l,z~±l)

4.1.29

In:(z+a}=lna+2 [(2a~z)+~ (Za-~zY

t~ (2a~zY + ... J

(a>O, Biz?. -a¢iz)

Limiting V dues

4.1.30

lim x-a In x=o

4.1.31

(a constant, ~a>O) lim xc. In X=_O

4.1.32

( ,m 1 )

1, im L; -k-1n "' =v (Euler's constant)

m-t0> 11;-1

- = .57721 56649 ...

(see chapters II 6 and 23)

Inequalities

4.1.33

x

l+x <In (l+x)<x

(x>-1, x¢iO)

x< -In (l-x)< I X X

(:t<l, xr!O)

4.1.34

Continued Fractlons

4.1.39

z z z 4z 4z 9z

In (I+z).=l+ 2+ 3+4+ 5+ 6+ ...

(z in the plane' cut from -1 to - a:I )

4.1.40

In (1 +Z)= 2z ~ 4Z2 9z2 ••• 1-z }- 3- 5- 7,-

(z in the out plane of Figure 4,7.)

Polyno~ial Appro,rluiation!l ~

Jw s x ::;.JlD

loglo x. a1t+aat3+E(X), t=(:c-I)/(x+lJ

4.1.41

I1iI=.86304

aa=·36415

4.1.42

~<x< 'iO

-v LU - - -Y1U

loglo x=cht+ast3+ast5+Cht7 +aet9+e(x) t=(x-l)!(x+ I}

1£(x)I$1O-7

a,=:094376476 ag=.191337714

4.1.43

al=.868591718 aa= .28933 5524 a6. .177522071

O:$x$l

In (1 -l-z) =alx+~x2+a.az3+a4:t4_+a.x~+E (z) k( x) lsI X 10-6

al=.99949506 a2= - .49190896 aa= .28947 478

a4= - .13606275 a6= .03215 845

~'T.l1e approximations 4.VH- to. 4.1.44 are if-om C_. Hastings, Jr., 1I,.ppro;-::imations for digital c~mputers; . ~tmcet0n Doh:. Press, Prineetou, N.J., 1955 (wIth permission).

4.1.44

ELE¥ENTARY TRANSCENDENTAL FUNCTIONS

69

In (1 +a:)=C+Jx+a2x2+aaiB+a,i4+a6x,6tasx6

+a7x1 tagxs+e ex)

al = . !l99,99 642:39 It:.!- -. 49987 41238 ~ = . 33179 90258 a,=-.24073 38084

a5= .16765 40711 a~---= -. O!J532 93897 a,= . 0360884937 cz,s= -. 006453544,2

Approxitnation in Tenn8 of Chebyshev Polynom:illlill s·

4.1.45

T,. "'(:I:)=C08 'i1fJ, cos: 8-2x-1 (see chapter 22)

In (1 +x)= ~ A"T" "'Cx)

11"'0

n All

o . 37645 2818

1 . 34314 5750

2 -,.02943 "7252 3 . 00336 7089 4 - ,. 000:43 ,3276 5 . 00005 9471

n A~

6 -. 00000 8503 7 . 00000 1250 8 -.00000 0188 9 . 000000029

10 -.00000 000.4 11 . 00000 000 1

Difi'erentiationFonnulae

4.1.46

,d 1 -d -in z=',zz

4.1.47

d':',ln z=-(-1)-",-I(91,-1)!'z-"

de" --

4.1.48

f dZ=ln z

_ z

4.1.49

fIn z dz=z ln.a-sa

4 .. 1.50

4.1.51

Jz" (In z)'li dz

,2,,+1 (lai)m '12.+1

tn,' f zn (In z)~ -ldz

110+1 _ - '-

(n,e-l)

, 'TheapprflxiIDation, 4.1.45 is from C. W. Glenshaw, PolynomiaJappro:xima.tioIlB to elementery fuu!}tione. Math. 'I'ables Aip.s Comp, 8) 14a,....147 (1954) .(wltP"'peI1- mlssion) .

JdZ'I'l' -'-l-=n n s .iZ n z

4.1.52 4.1.53

Jln [z+(:z2±1)llclz=z In [ztCz:.:i·±1)']-(z2±1)'

4,.1.54

fz"ln [z+(z2±l)ildz-:~ In [zt(z2±I)ll 1 f·zi.+1

- 11,+1 (z,2±1)i de '(n;¢-l)

'4.1.55

Ddinite lntegrah r I In !. dt= -1r',/6 Jo I-t .

4.1.56

r1 lns .. Jo I+t dt=-.,(l-j12

J.e Idt ""'"li(x) '(8ee5.1.3) ant

4.1.57

4.2. Exponential Function

Series Expansion

4.2.1

.::;l z~ zli, ez=e:xpz:-l+1l+21+SI+ . .. (z=;c+iy)

where tis. the real number defined in 4.1.16

4.2.2 Ln (ezp z)=z+2kri (k any integer)

4.2.3 .In (exp Z)=2 (-1r<Jz:s'lr)

4.2.4 ezp (In z)=exp (Ln z) =,Z:

, d .

dz exp z =exp z

4.2.5

4.2.6 4.2.7

Defutjti~n of~n:eral Powers If N =0.", then z=LogJ) N ai"---'exp (? In a)

4.2.8 Ifa=)alexp(ia:rga) (-1!'<arga$1f)

la'i =lal;;e""1l "18" <I

&rg (a-)=V Inla.l+x·arga

4.2.9 4.-2.10 4.2.11

Ln a,'=z hi a fo:r one of the values of Lna"

4.2.12 In a"=x In a (a rea.! and positive)

4.2.13 Je:l-ez

70 4.2.14

.arg(e)=y

ELE.MENTARY TRANSCENDENTAL FUNCTIONS

Special Value!, (see chapter 1)

4.2.16 a~b~= (ab)~ C-ir<arg a+arg b s.1I')

FIGURE) 4.2. Logarithmic and tXponentf,aIIy,nctions.

Periodic Properly

4.2_.17

(k lilly in teger)

·&ponen.tial IdentitieS.

4.2.18 4.2.19

The restriction (:-'1r<JZ1S1r) can be removed if Zz is an integer.

Limiting Values

4.2.20

limz(;<r"=Q (.Iargzl S.j1l"-E<!1i", dcom:!tant) l:if-loo

4.2.21 lim (1 +~)' "'=e'

~.,. 1n.

4.2.22 4.2.23 4.2.24 4 •. 2.25 4.2.26

8-2.71828 1.8284 ...

",'

e. =.ct)

±~ e2=±i

e2rki=1 (k any integer)

.&pon~~l Ineqqali;t;iea If xis real. a;nddiHerent from. zere

Continued Fracti.on"!iI

4.2.27 4.2.28

4.2~29 4.2.30

4.2.31

4.2.33

4.2.34

4.2.35

4.2.40

., 1 z z Z Z s •

e=-. -----.-.-.-----.

'1- 1+2- 3+ 2- 5+2~ ...

t:

e -l-x<l-:z::<e-a: (z<l)

gZ>l+z

e"'<11.Z (x<l)

:2: .

l+x < (1-.e-il')<z (z>-l)

:z;«e~'-l)< __!_ (fu<l)

. . I-x'

·:,Z

1+x>e1!:j-% (:z::>-1)

e"'>1+ x~ (n->O; x>O)

n.

4.2.36 e~>( 1+~)">e~ (x>O,y>O)

4.2._37 e-"<l-~ (0 <til 5 1.5936)

4.2.38~lzl<le=-11<~1.z1 (0<1 zl<1) 4.2.39 le"''--lls.el~l-iS.lzlel=1 Call z)

=1.+. z .. 'Z2/4 .. 3zW·15z2/4.35~~/4(4nt'-1) (j.'.I<' <XI ..• )

. (I-Bf2)+ 1+ 1+ 1+' . . 1+ .... z

• . .. ~n. nla ~ (n+ 1)z2z (1I.+2)z 3s . . -

e -ell.-I(z) nl- (n+l)+ (n+2)- (n+3)+ (n+4)- ('11+5)+ (11.+6)- ... (lz'l<w)

(For en(z) see 6.5.II)

4.2.41

4.2.42

ELEMENTARY TRAN:SCENDENTAL FUNCTIONi3,

e;2f> aretsn ~-1+~ d2+1 a,z+4 a2+9

. z-a+ 3z+ 5z+ 7z+ ....

(z in the cut plane of Figure 4.4.Y

·4.2.,43

.PolyilomiaI ApPI:oritnations;' O:$x«;ln 2-.693 ....

e-'~= 1 +alx+Clillx2+,,(:i::) 1,,(x)1 :::;aXIO-3 a1=-.9664 a2=.3536

4 .. 2 ... 44 O:S;x':::;ln 2

e-Z= 1 +<hX+a2.:z:2+a3w3+.ajX"+ f{X)

_..-

IE(x)l.:::;axlO-G

a,=-.99986JW a2= .49829Z6

as=- .1596332 a~= .0293641

4.2.45

OS~Slu2

e-;:;=1 +alx+a2x2+aax"+a~x4-+a6:t

+a6x6+a7Z;7 +f (x}

aL - - .9999999995 a2- .4999999206 a3= - .1666653019 ai= .04165 73475

a~:-- .0083013598 ae= ,0013298820 a7=-.0001413lt3-l

4.2.46 ;;

1(J= (1 +alz+a2x2 +a3xB+aj:t') 2+E (e) 1·,,(z) ts 7 X 10-1

ch=1.1499196 c:ta.= .6774323

aa= .20800 30 a_4-.126S.o 89

4.2.47

O;$:t.S1

10""= (1 +(hx+tl::i:c2+aaXs+a4xi+a6z6

+ GsXe + Ci?x7) 2+ E( X) I ,,(x)1 <5 X 10-8

al =1. 15129277603 az~ . 66273 088429 a-a= .25439 357484 Gt= . 07295 173666

a6=. 01742 1119.88 as=· 1)0255491796 a.., = . 00093 264267

i The approxim.a.tioDB:4.;2.43 to 4 .. 2.45 are from B. CarlBon, M. G'61dstem, Rational approxb::ilatioD of functions, -1,08 . Alamos Bci.elltine Labo.ratory LA~1943; Loa Alamos; N. Mex., 1955 (with permission).

5 The approXj.rrta.tions 4,2..46 to 4 .. 2.4-' are frc;iin O. Hastings, Jr., Approxlmationsfot digital cempnters. :f:rincetoD. Univ. Pre~s, Princeton, N.J.,. 1956 (with. permlBslon).

4.2.48

Approximations in Terms of Chebyshev Polynomials G.

n o 1 2 3 4 5 6 7 8

4.2.49

4.2.50

4.2.51

4.2.52

4.2.53

4.2.~

T!{x) -cos nO, cos (J.2x-l (see chapter 22)

..

e"= 2: An:T~(x)

·.n=.O

..

e-~= 2: A_,.T!(t}

.11.-11

AlI

1. 75338 7654 . ·85039 1654. , 105208694 ,00872 2105 ,000543437 .000027115 .00000 1128 .000000040 ,000000Q01

n A ..

o . 64503 5270

1 -.31284 1606 2 . 0.38704116 3 -. 00320 S68~ 4 . 00019 9919

5 -. OOOQO 9915

6 .00000 Q415 7 -. 00000 0015

DiJflll'entiat;ion Formulaa

. J eP'dz=e'!"/a

1

S· e"

2"~dz=aa+l [(az)"-n (a.g)"--l +n(n-l) (az) "-2

+ ... +(-l)"-'n!(az)+(-l)'%!l (n,~O)

4.2.56

J. e.~% til!.a I·· e··~ .

__:__ dz- . +-- -- dz

. zn·· - (n-I)z .. ""l· n:-1 zri-t

(See chapters 5, 7 and 29 for other

involving exponential functions.)

4.3. Circular Function!!

4.3.1

4.3.2

(1i>1)

integrals

sin. e

ej~-e;~n-

2i

.~ The apprcximatioas, 4.2.48 axe from C. W'. C1ens~aw, Po.·lY,iiXnrihil. aPJl.r.()xirIl.!lttoDStO. . eLemen._ ta.ry f .. unctions, Math; TaMes Aids Comp, 8.1 143-147 (1954}.(Wlth. permlasion).

72

4 .. 3.3

4.3.4

4.3;5

4.3.6

4.3.7 4.3.8 4~3.9

·u---

4.3.10 4.3.n 4.3.12

EL;EMENTARY TRANSCENDEN'rAL FUNCTIONS

tan sm s

anz=-· -

eOS:Z

-c 1

esc 2=-.'.Sln z

1 sec z=--

- - cos ,z

. 1

coil 2~-. -. --

- tana

PeriodicPropentiee

sin (z+2br) sin z (k any integer)

cbs, (e +2.br )'=c(iS Z

tan (z+br)=tan z

y

III .fl I

:I 1

.1:1 I

z,o \1 j, I

:\ :1 ,/

t_~ IV' J P I

r- ;.~ I

<---~~ ./ - .... ~,

' ..... ,

-\

\

- sinx -- C,~C~

--,- cos ~ ., ..•. ~ $ec x

--.- tcnx ~- COI·X

Relations Between Circular Functions

sin~ z + COS' .Z= 1

sec~ z-tari2z= 1

Neg-a tive.AngleF6Fmulas

4.3.13 sin ( -z)= =-sin z

4.3.14 COs (-Z)=005 Z

4.3.15 t,B.ll (- s) = - tl;tn z

Ad.ditio,D For:mulas

4.3.16 sin (z,+z,)-SlnZl cos ;;,il+cos 21 sin ,22

4;.3.17

4.3.18

ilalf~AD;gle Formulas

z .. (l+C_ .. os Z). l

4.3.21 cos :____: ±

- 2 --, 2

4_.3.22 tan:~. ±(1-COS 2)i

2 1+1308. Z

I-cos z SIn z

sinz I+cos .2

'I'he ambiguit-y in sign may -be resolved with the aid of a diagram. .

TranSformation of, Trigonometri.c Integeals

2 dU=l+z2'dz

, 2:e

4.3.23 am u=l+~i

Multiple.;.Angle FOrlDliIlas

. 22' - 2tan.z

sm - z-= . SIn Z COBZ· '

1+ts.n2z

4.3.25 cos 2z=2. QOsl! z-I=1-2 sin2z

4.3.24

2 '2. I-tan2,g:

=oos e-SID· Z .

. ... ... 1+tan2 z

2

2 tan z

4.3.26

ta.ri.,2z

I-ta.n2 z coV z--.:.l cot s-~a.n s sin 32=3 sin~-4 ain3e

4 •. 3.27

4.3.28 cos 32=- 3 COE! 3+4 coss :2

4.3.,29 sin 42=8 6083,2 sin z-4 cos asina 4.3.30 cos 4z=8 cos! z--8 cos:s z+ 1

Producte of SiDes and C.osines

4;3.;n 2 sin Z1: sin 2'2- cos (~1- Zil) - cos (~i + Z2) 4.3.32 2 cos Zi coaz~=cos (Zl- Z2)+OOS (Zl+22) 4-.3.33' 2 sin Zl cos z2=siD: (Z,-2.2) +am (Zl+22) Additi.on.and Subtraction of Two Cil'cular FunctioDil

ELEMENTARY TRANSCENDENTAL FUNCTIONS

4.3.38

$in (Zl±Z2) cos 31 COS Z2

41.3.39

sin. (zd: z ,)

cot Zl±COt 22=' _ .

sins, sin 2'2

Relations Between Squares of Sines and Cosines

4.3.40

sin2 zl-sin2 22=Sin (2'1+2:2» sin (Z1-Z2)

73

4.3041

coaz 21-eos2 z2=-sln (:rh+Z2) sin (31-22)

4.3.42

4-.3.43

Signs of ~he Circular Functions in the Four Quadrants

Quadrant sm cos tan
cae sec cot
I + + +
II + - -
ill - - +
IV - + - 4.3.44

Functions of Angles in Any Quadrant in Terms of Anglesin the First Qua.dnmt. (0 ~8 ~~, k any integer)

-_ -8 !±8 11'±8 3-11' ±8 2k1l'±8
2 2
S1ll ________ -sin 6 cos 6 =Fsin 6 -cos 6 ±sin 8
~os ____ --_- _ :cos6 =Fain ,8 -COB 9 ±sin e +cos_9
tan _______ -'tan 8 Toot 8 ±t-an 8 =Feat 8 ±tan 6
gc-------_ -csc 8 +see 8 =Fese 8 -sec () ±csc 8
sec _______ sec 8 =Fcso 8 -sec ~ ±csc 8 +.sec 8
~t _______ -c9t8 =Ftan 'J :1:: cot ,8 =Ftan 8 ±cot 9 4.3.45

Relations Between Circular (or Inverse Circular) Functions

sin x=a cos x=a tan,x=a esc x=a, sec x=·a cotz=re
sm x ______ a (1-a2), a(l +a2)-t a-I a-1((i-l)i (1+a2)-l
cos x ______ (1-a2)t a (1+a2)-t a-1(a2-1)t a-1 a(1+a2)-~
tan z ___ ~_ ~ a(1-a2}-j a-1(1-a2)l a (a2-1)-j Ca2-1)t a~l
csc :1:_ ~ ____ a-I (1-a2)-t a-1{1+a2)l a a(a2-1)-t (1 +a2)i
sec $ ______ (1-a2)-1 a-I (1+a2)i a(a2-1)-t a a-let +a2)t
cot $ ______ a-1(1-a2)-t a(1-a2)-t a-I (a2'-l)t (a:l-l)-t a (o~x~~) illustration: If sin x=a, cot x=a-1(1-a2)t arcsee a=arccot ,(a2~ l)-t

74

ELEMENTARY TRANSCENDENTAL FUNCTIONS.

,[3/2 ..;2/2 1/2 4.3.49 sm z=-i sinh iz
-{d/3 1 .fa 4.3.50 cos z=cosh'iz
..f2 2..{JtS 4.3.51 tan z=-i tafih iz
2
4.3.52 esc z=i each iz
2-./3/3 .J'2 2
4.3.53 sec z=sechiz
,fa 1 -iB/3 4.3.54 cot Z='l, coth i.e 4.3.46 Oircular Functions .for Cel't.ain Angles

1r/6 1r/4 'lr/a
30° 45° 60°
<,
1/2 .J2/2 ,f3j2 0 71:/12
0,0 15°
-_.- ---
am 0 -J2(4Z-1)
4
COB 1 ~ (.J3+1)
tan 0 2-{3
eSC 00 ../2(-/3+1)
aec 1 "f2(..f3 - 1)
cot ·00 2+·.Jj 5w/12 1r/2 71r/12
750 900 1050
8m ~ (~3+1) 1 .;; (-./3+1)
;J2 0 t(~-l)
cos 4 ,(,[3-1)
ta.n 2+,J§" 00 -(2+-./3)
cae -/2(..ja-l) 1 {2(,fa-l)
sec ..,fi{.,f3 + 1) ro -../2(.,[3+ 1)
cot 2-..;3 0 -(2--J3)
311'"/4 5'1f"/6 11:1r/12
135° 1500 1650 -1/2

4.3.47

e'=e*',II=e% (cos 'y+i sin y)

De Moivre's TheiJ,tem

4.3.48 (cos z+isin z),=cos, I'z+i sin I'Z (-1r<dfz-::;'11" unless.» is an int'eger) Relation to Hyperbolic' FunctioWl (!lee 4.5.7 to 4.5.1%)

Circular-Functions in Tel"ml! of Real and IUlaginary Pal"ts

4.3.55sinz=sin xcoshy+i COS X sinhy 4.3.56 cos .z=cos ::& cosh y-i sin x sinh y

4.3.57

sin 2x+i sinh.2y tan z

cos 2x+ cosh 2y

sin 2$._i, sinh 21/ cot z

cosh2y-cos 2:1:

4.3.58

2:/3;3

-.,fd ModulUll and Phase (Argu:ment) .0£ ~.cular Func;:p:o:D!I

-2

-1

11" 1800

o

sin ";2/2 1/2 - -JZ (..[3-1)
4
cos --JZ/2 -{3/2 -.J2{~3+])
4
t&il -1 --{j/3 -(2-Ta)
csc -./2. 2 -/2(+3-+1)
seQ -.J2 -2.,f3/3 -..j2{,f3 -1)
cot -1 -{3 -(2+v'3) o

-]

co

4.3.60 4.3.61

[sin .el = (sin2 x+sinhi'll)l

=[t (cosh 2y-cos 2x))l argsin z=.al'cta.D. (cot :t tanh y) [cos zl= (cos; x+sinh2y)t =Il (cosh 2y+cos 2x)Jl

arg CGS z= -arctan (tan x· taph y)

I . I (00.8. h 2y- 00& 2x)t

tan z= .

cosh 2y+cos 2x

_ ,. (sinh 2Y)

a.r:g tan z =a.l'ctan .' 2 '.

... . sm x

4.3.62

4.3:63

Series E:qIan8ioDi!

4.3.65

.ZS Z" ;;;:7

sm Z=Z"";-31+51-71+ ...

4.3.66

(Iz[ < co)

([zl<m)

4.3.67

ELEMENTARY TRANSCENDENTAL FUNCTIONS

75

ZS 2z~ 17z7

tan z=z+S+15+ 315 + ' ...

+ ( _1),,-122"(22"'-1) Bh 211.-1 +

(2n) I Z .•.

4.3.68

_1 a 7', 31 ~

esc z-;+6+360 z3+15120 a + ...

+(--1)"-12(2l"-1_1)B2n 2 .. -1+

(2n) 1 2 ...

4.3.69

4.3.70

In sin_ z=± (-1)"22n-IB211z211

Z 71=1 n(2n) 1

4.3.72

"'- (_I}"2211-1(22n-I)B, , "

In cos 2' = :8 ,2" 2~"

!i-I n(2n) I

Inequalities
4.3.79 sin x>~ (-~<x<~)
x 7r 2 2
(Ial<~) 'l.a.so sin x::; x ::; tan x (O::;X$~)
4.3.81 sm z (0 $x s 'lr)
cos x$----$l
x
(lzl<7I") 4.3.82 < sin 1I'X <4 (O<x<l)
7r 1l:(1-:t)--
4.3.83 [sinh U) $)sin z) $ cosh Y
4.3.84 Isinh yl $lc08 zl $ cosh 11
(iz)<~) 4.3.85 lese zl$csch/yl
4.3.86- [oos 2'j ::; cosh) a-/
4.3.87 J sin zl::; sinh] a I
4.3.88 [cos 21< 2, [sin zl::;~ /al (lzl<1)
Clzl<1t) (lzI<tn-)

Infinite Peoduete

4.3.89

."" (_ Z2)'

sin z=z IT 1--'

k~J k2r

4.3.90

_ <D( 4i2)

cos Z=.i:~l 1-(2k-l)2w-2

Expansio,n in Partial Fractions-

4.3.73 4.3.91

1 .. , 1

cotz=-+2z ~ 2 Jclr

z .1:-1 z - -

(Z~O,±'II",±2r, ... )

ta.n z .. ' (-l},,-12S"(22!l-1_1)B

In --= 2: " 2",: aU

z ,,-1 n(2n) 1·

(lzl<tt) 4.3.92

where Bn and E" are the Bernoulli and Euler

numbers (see chapter 23).

LiDlitiDg Values

4.3.74

I, sin X 101---=1 $-+0 X

4.3.75

lim tan x=l %-+0 x

4.3.76

lim ,x

nsm -=x

11-+'" • n

4.3.77

Ii x

m 'n tan -=x'

11-+,'" n,

4.3.78

lim x

, c09-=1

~.. n

.. 1

csc2 z==~, ( k ')2

11-- .. ,z~ ?r

4.3.93

(z~O,±iI',±2"" ... )

1 a> (-1)'"

csc,z=-+2a); 2 Jc21r-'-l

z t=i z, - ,

(z~O,±7I"!±2r, .. ,)

Continued Fractions

z ZZ Z2 Z2 (11")

4.3.94 tan z=l_ 3- 5- 7- .. , a #2±n1l'

4.3.95

76

Polynomial Approximation.s 7

ELEMENTARY TRANSCENDENTAL FUNCTIONS.

4.3.96

a~=-.16605

4.3.97

u2=-.1666666664 a~= .00833 33315 ·0.:8= - .00019 84090

ag= .00000 27526 UIO=-,OOOOO 00239

4.3.98

cos x=1+~x2+u~'+~(x) IE(z)1 s9X 10-'

a,z= - .49t?7Q

a~=.03705

4.3.99

o<x<~ , - -2

Hz) I s2X 10-11

a2= - .49999 99963 a4= .0416666418 Ue= - .0013888397

a8---' .00002 47609 alo= - .00000 026Q5

4.3.100

~=.31755

7 ThiOl approximations 4.3.96 to 4.3.103 are from R Car.I· SO(l; M. Goldstein, Ratioulll approximation Qf functions, Los Alamos Scientific Laboratory LA-19.43. Los Alamos, N. 'Mex., 19fi5 (with permission).

4.3.101

tan z .

-X-=l +a~i2+a4x4+a6Xe+a'8x8+alO:i:lo

+a12x'l!l+e(x)

IE(x)J':52XlO-~

'.

a2= .33333 14036 a4=.13339 23995 uo=.0533740603

as=.02456 50893 alO- .00290 05250 a!z=.00951 68091

4.3.102

*

a4=-.024369

4.3.103

a2=-.33333 3~410 a4= - .02222 20287 aa=-.OQ21177168

aa= -.0002078504 alO= - .00002 62619

Approximations in TeI'rD~ of Chebysbev Polynomfals B

4.3.104

T! (x) = cosmO, cos 8= 2x--l (see oh apter 22)

'" ..
sin ~x=x LJ A..T:C x2) cos i;rx=::E A"T:(r)
11=0 1'1"".0
n A" n A~
0 1.27527 ~962 0 .47200 121.6
1 -.28526 1569 1 - .49940 3258
2 .OO~11 8016 2 .027992080
3 - .00013 6587 3 -.00059 6695
4 .00000 1185 4 .00000 6704
5 - .00000 0007 5 -.000000047 ~The appreximattons 4.3.104 are from C. W. Clenshaw, Pol~'nomiE11 app,rQxi~!!-tioll<i to elementary functions, Ma.th. Tabh>sAid~ Comp. 8, 143,.-14.7 (1954) (with permission).

·See I>agen.

4 .. 3.105

4.3.106

4.3.107

4.3.108

4.3.109

.4.3.no

4.3.111

4.3.112

4.3.113

4.3.114

D.ifferenUa,tion Formulas

ELEMENTARY' TRAN,sCEN'DENTAL FUNCTIONS

77

d ,

-d sm a=cos z , ,~

Ii '

-d cos z=-am z z'

lztan. Z-Bee~ z

d

-d' esc z= - esc z cot ·z e.

d ,

dz see.a=esec z tan. z

dd (lot z=-csc2 z z

dn, '('.,1)

dz" sin z=sm z+r,

..

dn '( 1)

dz" COS Z= cos , z+zn1l"

Integration Formulas

fSin edz=-cosz J cos z de: sin z

ftan zdz=-ln cosz=ln seea

4.3.115 4.3.116

fcs.c zdz=in ta.n~=ln .(cscz'-col.z)=! In1-cos z ,', ,',', 2 " '. 2 1+008 Z

4.3.1l7

f s8czdz=ln (sec s + tan z) =In tan (~+~}--:gd -1(Z) -lnvarse Gtid6rmannian Function

11" gd z=2apctan 6"-2"

4.3.U8 J coted.z=ln sin 2=-1n esc z 4 .• 3.119

f z,. sine dz=-z" cos e+n f 2',,-1 cos zdz

4.3.120

-sin Z 1 feose

(n-l)zIl1 +c 1. zn-l d?

, ' ,

(on> 1)

I z

4.3.121 sm2 zdz=-z cot a-l-In sine

-2 cos.t 1

(n-l) sin" .1 Z (n-1) (n-Z) sin" 22

(n-2)f z dz . +(n":'l) .sin"-2 z (11)2)

4.3.123

4.3.124

fCOSZdz= 2"

ens z . 1 fsin Z dz ·(n> 1)

(n-l)z,,·1 1'1.-1 .Z,,_1

4.3.126

J zde

coa"z

zsin z 1

(n-I) cosn lZ '(n-1) .(11,-2) cos" 2z

(n-2)f'ZdZ +(~-l) . c9s"-~Z (rn>2)

4.3.127

f. ' ,sinm+l z cos .. -l·z

sin" z cos"'z dz=' ,'" '

. m+n

(n-I)J

+ ' sin'" 3COS"-2 z dz

(m+n)

sinm-"l z 008,,+1 Z

=-

m+n

em-I)f' ..

+ ' .". sm"'-2 Z cos" ede

(m+n) .

(m¢-n)

4;3·.128

J ~nm ::as~z

1

(n-l)sinm 1 Z·COSII. 1 Z

m+n-2f dz

+ n--l ,sinmz cos .. 2 e

(1/,>1)

-1

ffl+n-2f az

+ m-l sinl'ri"2 z cos" Z

(m>l).

I tan,,-l Z f' , ..

4.3 • .129 .. tan"zdz 'n-1 - te..n·1t-2zdz (n¢l)

, r " oot"-1 Z J" ."

4.3.130 J cot"zdz=- 11,-1 " cotn-2zdz (n¢l)

78

ELEMENTARY TRANSCENDENTAL FUNCTIONS:

4.3.131 (2:) b

It de 2·t ato.ilZ + ' (a2>bZ)

+b '" (ai-b·'2:.)l a,rc aD (.a',2-'b,2)l '

a' )HDZ

4.3 . ..132

f dz

l±sin z

"'.3.134

I' dz z 1 +0013 z ta:n 2"

4.3.135

f dz z

" -cot '_'2'

1-oos'z

•• 3.136

J&~ sin. b.z de _ a~~b2 (a sin bz-b cos bz)

fl.S.137

f tf~cos bz dZ=rtZ~b2 (6- cos bz+b ai'll bz)

4~3.138

Jeln' sin" be de

e(U sin·,,-I bz. .

, '~+2bl (a. Sill bz-no COtS bz)

an' ,,' "

4.3.139

f" ~ 008"-1 be ' , .

cu' COS1; be &2=' aZ+n2bi (a 008 bz+nb 10m be)

'n(n-IW~f'

+ . 2+ ~b't ~ C08"-2 bed« a . n

Defud.te1:ntegra1&

4.3.140

f"

Jo sin mfsiil nt dt=Q.

(m~n, m anda integers)

e sin2'nt =, r ',11' cos2'nt dt-! Jo Jo Z

(n an. integer, n:r!O)

4.3.141

So'" sintmt dt=~

=0

(m>O) (m=,O) (m<Q)

4.3.142

4.3.143

11'

=-"2

["cos at-cos ht dt~ln (b/a),

Jo t

4.3.1441

f"" . i'" . 1~'

sm t2dt = , cost2 dt='-. -,',

o '. ' , ,0 - 2 2

4.3.145

1"'/2. i1l'/2 '11"

In sin t dt', , In cos t #=-'2 In 2

o 0

i" ~~~t dt=~ e~m

(See chapters Sand 7 for other integrals involv-

ing eireular fun~tiOI):8,) ,

, (See [5 .. 3] for ,Fourier transforms.)

4.3.146

Formulas .for Solution of Planeftight Triang~tl:S

B

b

c

o

If A Band 0 are the vertices (G the right angle), ~nd a, band c the sides opposite respectively,

. A a 1

Sln=-·=--.

C I:;sc.d

.Ab 1 cos .a=:d seC.A

A a 1 Lan =-.b' =-A

". cot,

versine A=vers A~ I-pQs A

ceversine A=covei:s.A= l-s;ll A

haversine A=ha.vA=l Vel'S A exeecant A=exsec A~seo A-I

79

4.3.148

ELEMENTARY TRANSCENDENTAL FUNCTION,S

4.4. Inverse Circulal' Functions

FOJ'DlUW tor Solution or Pla~e Triangle!!

B

A

c

In a triangle with angles, A, Band 0 and sides opposite a, band c respectively,

abc

sin' A=sin B= sin ()

a=b ens C+c cos B

4+ b tan teA + B) a-b tll.O t(A-B)

be sin A. . b' ,~

&reI!. .2 " [8(8-0,)(8- )(8-c)1~

s=!(a+b+c)

4.3.149

Fonnu1ae for Solutiou of Spherical Triaus:les

a

If A, Band 0' are the three angles and a., band c the opposite sides,

sin A sin B sin (J

-'.-' -'=""'";---b' =-, --

sm s sm sme

cos a-cos b ooa e-l-sin b sin C COilA cos b GOS (c±8)

eoss

where tan 9=t&n b cos A

COS A=-COB B cos O+sin Bam o cos a

Defi.nitions

4.4.1

. f" dt

arcem Z= Jo (1-t3)i

(z=x+iy)

4.4,2

... ' fl dt 11" "r' esin e

a. ceos e= J 2 (l-tZ)l 2 ..,

*

f; dt arctan z= Jo 1+t2

The path of integration must not cross thE) real axis in the case, of ~.4.1 and 4.4.2 and the imaginary axis in the case of 4.4.3 except possibly inside the unit circle. Each function is singlevalued 'and regular"in the a-plane cut .along jhe real axiafrome- c::o to -1 an-d + 1 to + u> m thecase of 4.4~1 uI_ld 4.4~2 a~d a1o~g ~:he, imagin.arl·axis from $ totc::oand -t to -, co In the case 0 ·4.4.3.

Inversecirculat functions ~are also written arcsin z=sin-1 a, areeos Z=009-1 :t, arctan e =ta.n-1 z, ....

When -l:::;z:::;l;arcsin, x and arccos z are realsnd

4.4.4 -t-r ~ arcsin x:::; J1I"t 0 :::; arceosz :::; 11"

7r !Jlz>O

-4.4.5 arctan z+arccot s=±"2 flz<Jf

4.4.;6 4.4.7 ,4:4.8 4.4.9

arccsc z=arcsm liz arcsee z= 8l'CCOS lIz arccot z=a.I'ctan liz

aresee z+arccsc ,z=i1r

(see 4.3.45)

IV

o

o +1

-i

arclan:z

arcSin 1 and al'CCQ.IZ

iy

jy

+1

o -I

-10 +1

arccotl!

On::CIC z and «cnez

FIGURE 4.4. Branch cuts lor i'ntler8e cire'Uw,r junctions,

80

ELEMENTARY T~8~NDENTALFUNCTIONB

Fun(l~za~nud h9per.;y

The general solutions of the equations

/

sin t-z

COS t=z

tall t=z

are respectively

4.4.10 t=Arcsinz= (-1)' arcsin z+k'll'

4.4. II

t=ArccCiS z= ±a;rccos z+2br

4.4.12

t=Arcta.n z=arc~an z+hr

where fc is an arbitrary integer.

4.4.13 Interval containing principal value.

11 z posjtiV'e X negative

or zero

(z2¢-1)

arcsin a; and arctan x 0 5:y 5:.1t /2 -11"/2 $y<O

*arceo_s x andarcsee x 0 5:y 5:11"/2

*a;reeot x and arecsc zO:$ y.$ 'Jr/2 -w:/2 $V<O

-Q'c:,lfl':I;

-- _._ Q~"ctOi,:Ii:

~~~ orC1Ci'n I. -->g.rCol;:.1;.1i

~ le· ... ~ ........ orc:"c::~: -~~n;:c:ot·1I:

FIGUR~ 4.5, Inversecirc'/il;ar juru;tians.

".4.14 4~4~15 4 •. 4.16 4.4.11 4.4.18 *4 .. 4.19

arcsin (-z)=-al'csin 2 8I!;COS (-z)=:Ir-arecosz

arctan (- z) = - arctana areeso (-z}=-arccSC.2 arcsec (-.Z}=1I"-arcsec z arccot (-z)=-a:rccotz

Re.Ia:tion to Inver&e Hypel'bollcFunciiOIl8 (Bee 4.6.14 to 4.,6.19)

4.4.20 4.4 • .21 4.4.22

Arc$n Z= -i Arcsinh iz Arccos z- ± i Arccosh z

Areta!!. z=-i Arctanh iz (22¢-1)

Arc esc z=·i Arccsch ie Ar-esee e= ± i Arcsech e Arccot z=i Arccoth iz

_ .

4 .. 4.23 4.4.24 4.4.25

Logarithmic Repre.eentatioDS

4.4.26 Arcsin x=-iLn [(1-x2)i+i'X] (:r~l)

4.4.27 Arccos :r=-i Ln [x+i(1-x2).) (x2:::;1)

i l-i:z:; ii+x 4.4.28 Arctan x---'';;j .. Ln 1 +. ---2. Ln i: -

~. ~ x .. t-X

(x real)

. '_ [(x~-] )f+iJ

Arccsc X=-~ La_x .• (xt;::: 1)

4..4.30 Arcsecx=~i Ln[1+i{~2-1)tJ (x2~1)

4.4.29

4.4.31

Addition 'lind Subtraction of Two lnyetee Cireul", ..

Functions

4.4.32

.Arcsin 21;± Arosinz2

=A.rcsin [zl(1-zfyt±Z2(I-z?-)f]

4.4.33

ArccOs 2t±Arecos 2:

= Arccos { Zlz::=F[(l-zi) (l-zi)l~} 4.4.34

A.rcta.n zl±Arct-an z2=Mctan (t~±Z~)

. . . 21Z2 .

4A.35

ArWZ:l ±Arccosz~

=.Aresin f ~tZ.~ ± [(1- zn (l-:zi)]t} =Arccos [z2(1-zf)t=j= .ih(l-zi) f]

4.4.36

Arctan Zt±Arccot.z2

Ar t· (Zl.z2.±.1) ,t_ ... t ( .Z~-=FZl )

= can ·=.Q.I.CCO .. . ..

Zg =r Z12:1 Z2-± 1

Invene Circular FUDm.oD8 in T6nneof Real and _lIiI.a~ Paria

4.4.37

Arcsin z=b+< -1)1 arcsin {3

+(-l),tnn [~+(a2-l}I]

4.4.S8

Arccos z=2lcr±. {arcCo8~-i In [a+(at-l)t]}

81

ELE;MEN'rARY TRANSCENDENTAL FUNCTIONS

4.4.39

( 2x )

Arctan z=,k!lr+tul'ctan 1 ~ 'i

-x-y

i [Jj2+(:y+ 1)2J ,2 _ +4 In x2+(y---' 1)~ (z rf 1)

where k is an integer' OF zero and

a-it (x+ 1)2+ y2Jl+H (x-l)>l+y2l; ,{3--.".j [(?,+1)2+y21;-i(x,-1)2+Y~lt

Series Expa~ona

4.4.40

• Z3 1 ·3z.5 1· 3 . 5z1

arcsin z:._z+Z.3+2.4 .5+2.4.6.7+' .. (lzl<l)

4.4.41

• • ( ) 1r ( ) I [ ~ 1 ·3· 5 . ' . (Zk-1) kJ

arcsin 1-2'-2 2z l+(,;t 2z.t(2k+l)k! z

. (lzl<2)

4.4.42

Z3 Z5 27

arctan z=z- 3+5-7+' .. (Iel::;l a;ndzZ~-l)

11' 1 1 1 '

=Z-z+3z3- 5z5+' .. ()21>1 andz2;¢=-1)

Z [ 2 22 2·4 ( (:,2 )2 ]

=1+z2 1+31+z~+3.5 1+22 + ...

(z2#--1)

Continued Fractions

Z Z2 432 9z2 16z2

.4.4.4:3 arctan 2'=1+ 3+'5+ 7+ 9+'"

(2 in the cut plane of Figure 4,4.)

arcsm z

4.4.44 ~

,"" 1- Z2

z 1· 2Z2 1·2i2 3· 4Z2'3 .432 1- 3- 5- 7- 9-'"

(z in the cut plane of Figure 4.4.)

Polynomial Approxilnations ,

4.4.45

I~(;t) I :S;5XlO-5,

au= 1.57072 88 uj=-.21211 44

a2= .07426 10 U3= - .01872 93

o The approxhnatfons 4.4.45 to4A.'i7 9.F,e from Q. Hastings, Jr., 4;ppl'oximatiQ"nB for digital computers. Princeton Univ, Press, Princeton, N.J., 1955 (with pesmissiun),

4.4.46

ai'csin:z:=~- (1- x) t(ao+t1.J.a::+a2~2+q3t1 +a;4x!lc+a5x~+a&:te+a7t7)+E(X)

If(x)152XiO-S

ao= 1.57079 63050

0;,= .03089 18810

al=-.214~9 8$016 a6=~·01708 81256

a2= .0.8897 89874 a6= .00.6670.0901

a3=-,o.5Q17,43046 1'L'1=-,o.012624911

~~~ -1~x51

ai. = .99986 60

~= - .08513 30

as=-.330,2995 ae= .0.20.83 51

a5= .i8014 10

are tan 'x

x 1+.28x2+E(x)

lE(X) I ~5XlO-3

4.4.49 n

arotan X Xl

~=-.33333 14528 ai= .19993 55085 a6 = - .14208. 89944 0;8= .10656 26393

11,10= - .07528 90400 ai~= ,042,90 9.6138 aI4=-.0161657367 o;1S= .00286 62257

10 The approximation <i.4.48 is from C, H!lStiD~, ~r., Note 143, Matb,. Tables Aids Compo 6, 68 .(1953) (W1th. permission)' . .

II The apprcximation 4,4.49 is Irom B. 9arIson, M. Goldstein, E,ation.al approximatioIl-:of functions, Los Ala.mos Scientific Laboratory LA-1943, Los Alam95, N. Mex., 1955 (with permisaion) .

82

ELEM.ENTARYTRANSCENDENl'ALFUNCTIONB

App .. o~bnationa in Terms of Chebyshev PolynomiaJs 12

4.4.50

cos 9=2x-l

(see chapter 22)

CD

arctan z=il::E A"T;{;:I)

:n-'O

n, Ail

o .88137 35'8'1

1 ~.l0589 2925 2 . 01li3 5843 3 -.00138 1195 4 . 00018 5743 5 -. 00002 '62-15

n AI>

6 .' 000003821

7 _,.. 00.000 0570 8 .0000'0 0086 9 -.00000 0013

10 .00000 0002



For :c >1, use arctan x=l'!l'-arct~n (1#)

4.4.51

...

arcsin x=x:E A"T:,C2x2)

n=O

'"

a.rCCQS x=;#-x L .d"T:(2x2)

'71-1'1

n o 1. 2 3 4

Ail 'n

1.05123 1959 5

., 0549.4 6487 6

· 004080631 7

· 00040 7890 8

· 000046985 9

AI> .000005881 .00000 0777

. 00.000 0107' .00000 0.015 .00000 0002

For i.J2~x::;l, use arcsin x=arccos(1-x2)J, arc-

COil, x=e.rcsin (l-i2)t, .

DiJJerentiatiol1 Formulae

4.4.52

d , ( ")'"

- arcsin 2= l-z, -,

dz .

4.4.53

s. . aroeos Z~ - (1- z~)-j

dz '

d 1·

d~ arctan 2= 1 +Z2

4.4.55

d -1

dz arccotz=l+t2

4.4.56

d

dz ElJ'csec 2

1

11 The ~pill'oxim.tioD!i4-.4.50to 404'.51 ate from C. W.

Clerishsw,Polynoinial approximattona to elemeptsry Iuncticns, Matb. Ta.bles Aids Com.p.S, 143-147 (191)4)

(witb permiesion). .

.c:: ............. _ 1''-

d 1

d-. &fccse z=-. ( ... ~ 1)' 22 Z-, ,

. . ".

4.4.57

Integratioft. Fonirulas

4.4.58

J arcsin z dz=z arcsin z+(1-z2)i

4.4.59 I arccos a dz=z arecos z- (1- z2)1

4A.60 Jaretan Z dz=z·arctan 2-;10 (1+22)

4.4.61

J arceso z dz=z, arocac z±1n [Z+(Z2_~PJ

[ O.·< ate .,..Z<i]

-~<Ei.rccse z<O

4 .• 4.62,

farC$ec Z 4.z=z'Q.rcsec, z=fln [z+(z2-1)j]

," [0<. a ..... cz<;J •..

~<a.rcsec z<1I'

. 4.4.63

f arccots dz=za.rccotz+i in (1 +:;:2)

·4. "U,4

I~ arcsin z dZ=(~i ~) arcsin. z+i (l-i~)~

4.4.65

C . ,,+1 1 I ,,+1

JZ" arcsin z de ~+la;rcsin z- n+l (/--:21)1 dz

(n?!,-I)

4.4.66

fz 8.rCcOSZ dZ=(~2_i) arccos z-~ (1-22)1

(n~-l)

4.4.68 .

Jz arctan z dz-~ (;+Z2) arctan z-~

83

ELEMENTARY TRANSCENDENTAL FUNCTIONS

(n~-1)

4.4.70

(n~-l)

~.5~ Bype.l.'bolic Functions

DefinitioDs

4.5.1

sinh e

2

4,.5.'2

cosh z

4.5.3

tanhz=sinhzlcosh z

4.5.4 4.5,,5 4.5:6

each z,= ljsinh z

sech Z= 1/00sh g'

(loth Z= l/tanh a

\~ ..

\~'h'~

\

\ Z.O

, ,.

"-

-I~";~ :I

• __ ..... ci1h ."

---.'ontil:

-.-- I:K!l1

.•. ;.: •. ,fttl.lI

-..........., toth *

FiGUR1!l 4.6. Il,yperb,olicjunetions,

. Relatiun to Circular Functions (see 4.3.49 to4.S ... 54) Hyperbolic formulas can. be derived from. trig'On,Qmet:tic identities by rep1acins- z by iz

4.5.7

eoshe " cos iz

4 .• 5.8 4.5.9 4.'5.l0 4.5.11 4.5.12

e6th .f;!=i, cot iZ:

tanh Z= - i tan iz

cseh z=i csc.ss

sech ,3=8ee. iz

4.5.13

Periodic Proper:ties sinh (z+2k7ri) = sinh z

(k any integer)

cosh (z+2k:ri)=cosh 'z

tanh (z+k1l'i) = tanh 2

4.5.14 4.5.15

Re1ationll Between.·Hyperbolic .Functions

4.5.16

cosh" z-sinh: Z= 1

tanhli z+sech2 2=1

4.5.17 4.5.18 4.5.19 4.5.20

coth~ 2-csch2 ;:=1

cosh s-l-ainh z=e~

cosh z-sinhz:.......r'

Negative Angle ForioUla"

4.5.21 4.5.22 4.5.23

sinh (-oi) =-sinh e cosh (-z) =cosh etanh (-z)=-'ta.nh e

Additioit Form.u1as

4.S.24sinh (?I+Za)=si:mZ1 cosha,

+ cosh Zl sinh 212

4 .•. 5.25 cosh (zl+z2)=cosh 21 cosh eD

+ sinh Zl sinh Zt

4.5.26 tanh (ZI +Z2) = (tanh Zl + tanhz,!;) I

(1 -l-tanh 21. tanh ~)

4.5.27 coth (Zl+z2)=(eoth;31 oath 2'2+1)(

(coth 'z;.+enth 21)

Half-An~le For.Dlulas

4.5.28

sinh ~=( cosb, 2Z -1 Y

84 4.5.29

ELE~ARYTRANSCENDENTAL FUNCTIONS

h Z (COshZ+l)t cos -=

2 2

4.5.30

tanh ~=(COSh 2-1)t cosh 2'-1 sinh 2

2. . cosh 2 + 1 sinh 2 cosh $+ 1

4.5.31

4.5.32

4.5.33

4.5.34

4.5.35

4.5·.36

4.5.37

4.5.38

Multiple;'Angle. Fo.l'tnWa.s

. h. inh h 2 tanh Z

sm' 22=2 s .. 2 C()S ZIt h2 - an ·2

cosh 2z~2 cosh! z-I=2 sinh! 2+1

2 tanh 13 tanh 2z 1 + tanh~ 2

sinh 3z=3 sinh 13+4 sinh" 2

cosh 3e--3 cosha+s cosh" 2

sinh 42= 4 sinh! z cosh z+ 4 c-osh3 z sinh z

cosh 4z=cosh4, z+6 sinh! 2 cosb- s-l-sinh' z

Products ofHyperholic Sines and Coliinl'lll'

2 sinh Zj sinh Za=cosh (ZI + Z2)

-cosh (ZI-Z:)

4.5.40 2 sinh ;<It cosh Zz=sinh (21 +Z2)

-l-sinh (ZI-Z2)

Addition and Subtractio.n of Two Hyperbolic Fqpct;j.ons

4.5.41

. h -l-sinh 2 ._1... (21+22) h (21-23)

sm ZI SID ZIl= SlI!.Ul -2- COB . -2-

4.5.43

coshz1+coshz2=2 cosh (ZltZ2-) cosh (Z~2 22)

Relations Between Squares o.f H~perbolic Smeli'! and C<Wn1lB

4 •. 5.47

sinh" zj-sinh2 zz=Sinh'(21+Z2) sinh (ZI-Z2) =oosh2 ZL-C0Sh2 22

4.5.48

sinh! z~+COSh2 z2=cosh (21+Z,2) cosh (Zl-~) =c08h2 ZI -l-sinh" Z2

Hyperbolic.'Fu.netions in Te:nns of Real and ·lmaginary

. Part.s

4.5.49 sinh z=sinh x_?Os V+i cosh X sin ,y 4.5.50 cosh z=cosh x cos y+i sinh z sin y sinh 2x+isin 2y

4.5.51 tanh-a c-osh 2x+cos,2y

sinh 2x- i sin 2y 4.5.52 coth e c-osh 2X:-C08 2y

.De Moivre'8 l'heo~m.

4.5.53 (cosh a-l-sinh. z)ft=cosh nz+sinh n2

Modulps and Phase (Argument) of HyperhoHc FunotioDs

4.5.54

[sinh zl = (sinh" z-l-sin" y)t =L1<cosh 2X-C0S 2y))!

arg sinh z=arctan '(eoth :t tan y)

4.5.55 4.5.56

[cosh zl=(sinh2 X+C082 y)t

= [HC08h 2x+cos 2y))~

".5.57

arg cosh z=arcta:o (tanh z tan y)

4.5 .. 58

I tanh zl=(coSh 2x-cos 2Y)' cosh 2:z::+ cos 2y

4.5.60

ELEMENTARY TRANSCENDENTAL FUNCTIONS

Relations Between Hype_~bo1ic (Of Inverse Hyperbolic) Functions

85

sinhx=a

cosh x=a

tanh x=a

csch x=a

sech_z=a

roth Z=(l,

sinh X ____ ! a (o,~-l)l o,(l-a2)-! a-I a-l(l-a~)~ (a2-1) -"t
cosh x _____ (1+a')' a (1-0,2)-, o,-1(1+a2)* a-I 0,(0,2_1)-.
tanh x _____ a(l +a~)'-i 0,-1(0,2_1)' a (l+a~)-i (1-a~)1 a "",1
csch :1: _____ a-I (a2-1}-i a-1(1-a2)+ a a(1-o,2)-. (a2-1)t
sech x _____ (1+d2)-6 a-I (1-0,2)1 a(1+a1)-' a a-1(a2-1)1
coth x _____ a-I (a2+ 1)' o,(a2-1r' ,0,-1 (1+0,3). (1-a2)-t a Illustration: If sinh _z=a, coth x=(1.-1(a2+ 1)1 arcsech a=e.rccot.h (l-a2)-t

4.5.61 Special Values of the Hyperbolic Functions 4.5.66

h 1 Z2 5 -, 61 5+ + E2f1. 2"+

sec Z= -"2+242 -7202 .•. (2'11,.)12 . _.

(Jzl<~)

.4.5.67

1 _ z z~ 2 Ii 22"B2f1. In-I

coth z='i+3- 45 + 945 z- ... + (2'ib)! z- + ...

_ (lzl<1I")

where B; and E"are the nth Bernoulli and Euler numbers, see chapter 23.

InfiniJ.(l Products

0 'II'" 3'11'"
Z 2'1> 11"1. 2 '/, co
_- --
sinh 1iL ______ 0 i 0 -'I, co
cosh z _______ 1 0 -1 0 -00
tanh 0 0 -' 1
z ______ oo'!. - co-'!.
csch z _______ co -'!. c:o i 0
Bach z _______ 1 CD ---1 co 0
eoth z ______ ., co 0 ee 0 1 4.5.62

Series Expansions

, i-3 zJ Z7

sinh z=z+31+5'+7!+ .. - (jzl<oo)

4.5.63

(Jzl<"" )

4.5.65

+ +22";(22ri--l)B2", . 21i.,...1+

. . . (2n) I z . , .

(lzl<~)

4.5.68

4.5.69

4.5.70

4.5.7'1

4.5.72

4.5 •. 73

4.5.14

Co.nti:nued.F.raction

Differentiation Formulas

j_ sinh z=cosh '2 de

d ,

de cosh z=smh s

d

dz tanh z=s6ch:z

.!! csch z=-csch .z Goth z dz

86

ELEMENTARY T-RANSCENDENTAL FUNcTIoNS,

4.5.75

dd sech z=-sech z tanh .2: . z

4.5 .. 76

d h h"

- cot z=-csc2z,

ds." .

Integration Fonnnlae

4.5.77

f sinh zlz=coshz J cosh Z dz~sinh z

4.5.18

4.5.79

Jtanh z dz=m cosh ;2

4.5.80

f . z

csch z de: :......10 tanh 2

4.5.81

Isech a dz'-'arctan (shill z) oj oeth z dz-lo sinh g

4.5.82

4 •. 5.83

Jzn sinh e dz=z: cosh z-n f z,,-lcosh zdz

4.5.84

fz,.: coshz dz=z" sinh z-n Izn-I sinh e dz

4 .. 5.85

Is-inh'" zdish"'z dZ-____!_ sinh",+lZ 008hn-1 Z m+n

. 11,-1 f . b' .. h d

+ __ ._. 8m 0.", zoos' "-2Z2

·m+n

1 . .'. .

=-- srn:h",-l aeosh .. ,,+1,.

m+n .. ". '"

J d.z

4.5 .. 86 sinhmz coshn z

-1

1

m-l sinh';; 1"14 cOShri-'l Z

m+n-zf dz

m-l ' sinh"'-z,z cosh" z

(mrH)

1

1

=~-1 sinh"'-l s.cosh" J 2

+m+n-2S'. .dz

n-l 8mh~ Z c013hn-2 z

(nr!l)

4.5.87

J . tanh,,-l Z f" . '

tanh", z dZ=n"'-l +t~nhn,...~ 2 dz

(fl.,.:;1)

4.5.88

J cothll-l Z f

coth" z dz=- 'n-l +. coth 11-22 as

(nrf 1)

(See cha.pters, 5 and 7 for other integrals involving hyperbolic functions.)

4.6. Inverse Hyperbolic Funetions DefinitioD8

inh r: dt. arcs ,z=J 0 {1 + t~)!

(z=x+ill)

4.6.1

4.6,.2

I" dt arccosh z- Jl (t2-1)f

rt dt arctanhz-' Jo 1--£2

The' pa.ths of :integration must not cross the fol-

lowing outs. .

4.6.3

4.6.1 imaginary axis from -i CD to +i andi toim

4.6.2 real axis from - 00 to + 1

4.6.3 real axis from ~ 03- to -1 and + 1 to

+00' .,. '. .

Inverse hyperbolic functions. are &180 written sinh-1 2, arsinh z, si'rsinhz, etc.

'4 .• 6.4- 4.6.5 4.6.6

areesch z=arcsinh l/z aresech .Z= srceosh 1/2 arccoth z=arctanh liz

iy iy iy
+i
':)1, -·OCI +1 X :t
0 0 -1 0 +1
-i
Clr~sinh % orCCO$h z orctcoh z
iy ,iy iy
+i )( ... ~ +1. += -I +1 x
0 ... j 0 0
oi:c()sch z ercsech z orccolhz FWURE4.7. Branch cuts jar inverse hyperbolu functions 0

ELE;M:ENTARY TRANSCENDENTAL ,F{JNC'I'lON8

87

4.'6.7 srctsnh z-=arccothz±i"lI"i

(see 4.5.60) (according as Jz~O)

Fwndamental Property

The general solutions of the equations z=-sinh t

z,=cosht

z=tanh t

are respectively

4.6 .. 8 t=Arcsirth $= (-l)~ arcsinh z+lc'll"i

4._6.9 t~.Arccosh z=-±arcco:sh z+2k1li

4.6.10 t~ Arct.anh z=-lU'ctanh.z+k1l'i

(k) integer)

Functi'o~ of Negatlve Arguments

_4.6.11

arcsinh (-2) =-~nresinh 2

r

*4.6.12

arccosh ( -z) =1I'i,~&rccoshz

4 .• 6.13

arctanh (-z}=-o,fct_anh z

Relatlonto Inverse Circqla.-rFu-nctione (see 4.4.20 to 4.4.25)

Hyperbolic identities can he derived from trig-

onometric identities byreplacing ;;: by iz. .

4.6.14

Arcsinh i= -i Arcsin 'iz

4.6.15

Arccosl; z= ± i Arccos z

4.6.16

Arctanh ;; = - i Arctan ie

4.6.17 4~6.18 4.6.19

Arccscb z=i Arccsc ie

Arcsech Z= ± i Aresec ?

Arccoth i-:- iArcc6t iz

Logaritbmie. Represents. tions

4.6.20

arcs-inti :i:=ln [2:+ ("X2+ 1);1 !linrcosh x=ln [z-} (:c~ -1) I]

4.6.21

4.6.22

. I -l+x

aretanh x=! n l.:z;

4.'0.23

. P (1 )i]

arecseb x,-----ln C;+, i2+1. .

(x >"'0.)

4.6.24

[1 (1 rr-

arcsech $=ln .;:+ x2:-t, (O<x'::;; 1)

th. 1} x+ 1

arcca .• X=- 'n -_ -

2 x~l

4.6.25

._6_0', -a.o

*

~.O

4.,0:

~~~,~Ctil;( Qtt;~O'I~1"~7'- I

I.~·

-- ---'CfCSClflhJ;: • _." •• ,arIl:C~:!i1'1 ~ ~ - - GI.I;:fOf1h :.., - - - Q/(~~!I:;f1:ti1 ._ ... _ .... ,O'F'!.~~~'",. ---- .c~C911'1oi;

\ '\ ." V \ ~.o Ii·, ..

II \ ,a~

I ~l:::'

~·l'Cim'l:l'l j: l ~".lt

! ,.;~

FIGUR'E4,8. Inverse hyperbolicju/nctions ..

Addition and Subtraction of Two Inverse Hyperbolic Funetione

4.6.26

Arccosh Zl± Arccosh Ih

4.6.28

4.6.29

=A.rrsinh{zI2g± [(1 +zD.(Z5-1)]!} =ArcC'(lsb rz~(l+3n!±21(Z~-n'J

4.6.30

.s." page 11.

88

St;ries ExpansjoDs

ELEMENTARY TRANSCElN".DENTAL FUNCTIONS

4.6.31

inh ' 1 ..:.~+1.3 5

arcsin: 2=2-2. 3':- 2.4. 5 Z

4.6.42

.!i arccoth 2=(1- z~), -I

dz. .

1·3·5 7 f

-=.2.4.6 . 72 + . . . 4.6 .. 43a:rcsinhz gz=,z arcsirih~-(1+z2)~

(lzl<l)

1 1·3

~lrt 22'+2. 222 2.4.42,4

1·3·5 +2.4.6.626

(lzl>1)

1 1·3

arceosh 2-1n 22- 2 .. 2;;:2 .2.4, 4e,'

1 ·3·5 2.4.6.6#°-

(lel>1)

za zs, a" 4.6.33 arctanh 2=z+3+5+7+ ...

1 1 1 1

4,,6.34arc.coth 2=-+-3", 3+5',5+-7. 7+ , ...

z Z 2' ?

(lzI>l)

(lzl<l)

Continued FractionS,

;;: . zt 4;;:2 922' 4.'6.35 arctanh 2=1_ 3- 5- 7- ...

(z in the cut plane of Figure 4.7,)

4.6.36

B.:Tcsilih Z 2: 1· 222 1 ' 222, 3 ' 4Z2 3 .422

";l+z~=l+ 3+ ' 5+ "7+ ~

DjfferenJil.-tion .F orm ulas

4.6.37

4.6.38

d , b (2 1)-'

--d arCCQs' Z~2--'

, Z '

4.6.39

"!!a.rctanh z= '(1- Z2).-1

d2 ' ' , "

4.6.40

d 1

d~an,::e!;ich ? = 1= z (1 + Z2) t

(according as 9Pz~O)

d. ' 1

dzarcsech 2=1= z{l-zZ}t

4.6.4J

4.6.44 Jarccoshzdz=zarccosh Z_(Z2_I)t 4.6.45 Jarctanhzdz=za:rctanh z+!ln (l-e2)

f arceseh. z dz~~ arcesch !t'± arcsiahs (according as 9i!zE<,O)

4.6.46

*

4.6.48

f arcseeh z d'$= z arcsech a ± aresin s

J arocoth z clz=zarccoth z+lln (z~-l)

*

4.6.47

4.6.49

J", 2z2+1.z ' ,

e a.rcsrnhzdz=--4-. - arcsinh z-'4 (z2+1)1

4.6.50

J' . . .2'''+), 1 J' .2,,+1 ,

z"aresmh zaz 'n=Fl arcsinhs-e- n+l (l+zz)td~_

(n~-l)

4.6.53

J ' z~-l, ", 2 sarctanh.e dz=---y- sretanh z+2'

4.6.54

f· ' 2,,+1 1 f" Z,,+L

z"'arctanh z dz =--, . srctanh il- ,'-- - -, -, 2dz

- ' , , , n+ l' , , n+ 1 1- z

(n~-l) 4.6.55

(according as g.fz~O)

(n~-l)

89

4.6.57

ELE¥ENTARY TRANSCENDENT.At. FUNCTlONS

4.6.59

f 21 1

2 arcsech 2 dZ="2 .arc_sechzT'2 a-,z2)1

(according B.s8fz=<O)

4.6.58

r 2,,+1 1 f Z"

JZ" arcseeh Z dZ=n+l areseeh z±n+ f (1_z2)ldz

(n.~-I)

f 22-1.. z z arccoth z dz=~ arccothz+2

4.6,60

I· Z·+I 1 f Z·II+ 1

. 2" arceoth zdz= n+l arecoth z+n+l .. z2_1dz

(n~-l)

4.7. Use and Extension of the Tables

NUinerical Methods

NOTE: In the examples given it ia assumed that the argumentsare esact.

Example I, Computation of Common Logarithms.'

To compute common logarithms, the number must be expressed in the form e- 1011, (1 5 z< 10, -= :$qS'=) .. The common logarithm of x-lOll consists of nn integral part. which is caned the eharactenistic anda decimal part which is called the maneissa, Table 4.1 gives the common logarithm ofz,

.009836 9.836 .10-3 3.9928185=(-2~OO71815) _09836 9$36.10-22.99281 S5=(-1.0071if15j . {lS36 9.836.10-1 1.9928.1 85=(-0.0071815)

9,836 98~36 983.6

9.836 ·10~ 0.9928-185

9.836 ·10! 1.9928185

9.S36·1()2 2.9928185

Interpolation ia Ta.ble 4.1 between 983 nnd 984 gives .99281 'S5 us the mant-issa oi ,9836.

Note that 3.99281 85=-3+.99281 85. When q isnegativ'e the common logarithm- can be expressed in thenlternative forms

Iog1o (.009836)=3.99281 85=7.99281 85-10

, .

= -2.00718 15.

The la.st form is convenient for conversion from comme~ logarithms to natural logOJ'ithms.

~ Th~ mv~sl!; of.loglo~ is ~alled the antilogarithm .ft. x! and IS wrrtten a.nt.llog x or log-I e, The logarithm (.If the reciprocal of a number is called the colog8.l'lthm,written colog.

Example 2.

Compute :z:.:-'H forx=9.19826 tQ 10D using the Table of Common Logarithms,

. FrO!!l Table 4.1~ four-point Lagrangian interpolanon grves 10glo{9.19826) =.9~370 56812. Then!

-£ loglo(x)---, -.7.227792609=9.2772207391-10. Linear inverse interpolation in Table 4.1 yields antilog 6,2772·2).=.18933. For 10 place accurncv su btabulation with 4-poin b Lagrangian inter-

polants produces the table .

N logloN A !l~'
.18933 .. 27721 94350 .229379
.18934 .. 27'724 23729 -13
2 ;29366
.18935 .. 2772653095 By linear inverse interpolation X~3H = .18933 056$5.

ExampleS •

Convertlog., x to In x for x=,00983ft

Using 4.1.23 and Table4.l, In (.009Sa6) In 10 (.agio (.009836) =2,.30258 5093 (-2.0071815) = -4.62170 62.

Exatnple 4.

Co!npute In z for ·z=.00278 t96D. .._

Using 4.1.7, 4.1.11l1n« Table 4.2, in (.00.278) == In (.278·1O-~)=ln (.278)-2 In lQ= -5.885304.

Linear interpolution bctweenx= .002 and ::e=.003 would give 10(.00278)=-5.s-98. To obtain 5 d~cilTI!i.lplnec accuracy with linear interpolation it is necessary _thll t. x> .175.

Example 5.

Compute In x. f9f x=1l3L718 to 8D, Using 4.1.7, 4.1.11 sud Table 4.2

In 1131.718=ln(1l::3~18113J )

=In 1131.718+1 1 131 +1 105

. 1131 n i , n

=11l(l.OOO~3 4$36)+lu L131+31n 10.

ELEMENTARY TRANSCENDENTAL FUNCTIQNS

Example 25.

Compute arcsec2.8 to 5D. Using 4.3.45 and-Table 4.14

,_ " (z2-1).

aecsec Z= arcsin

z

, " [(2.8)1l_1]1

arcsec 2.8=arcsm 2.8

=8l'C_Bin.93404 97735 =1.20559

or using ,*.3.45 and Table, 4.14 ercsec e=a.-rctan (e2-1)i

arcsse 2.8=a:rctari 2.61533,9366

"""~-arcta;n .;3823595564)

from 4.4.3 and 4.4.8, =1.570796-.365207'

=1.20559.

Example 26.

Oompute aretanh X; for x=.96035 to 6D. From 4.().2~ and Table 4.2

93

0.-.1-. 60 1ln' 1+.96035 _J,. In 1.96035

aretallll. .9, 35=2", 1.- .96035'- .03965

=i In 4,9.44136 191

= i(3.90078 7350) = 1.950394.

Example 27.

Gom pu te arcccsh :doi" -Z= 1. 5368 to6D. U Bing Table 4.17

arccQsh xarcc_osh 1.5368 852346 (z2-1)1 [(1;5368}~-11f .'

arccosh 1. 53ij8 = (.852346) (1.361754)1

= (.852346) (1..166,942) =.-994638.

Exru:nple 28.

Compute arecosh-z for x=31.2 to 5D.

Using Tables 4.2 and 4.17 with 1/:x,=1/3L2 = .03205 128205

~cCOBh 31.2-ln 31.2-.6'92886 arccesh 31.Z=.692886+3-.440418=4.1:l330.

References

Terl8

[4.1] B. CarIeon, M. Golds:teiD,Rfl.~ional approximation of functipIl6, Los Alamos Scientific Laboratory LA-1943 (LOs AlaIiiQH,R Mex., 1955) ..

[4.2] C. W. Clenshaw, Polynomial .. approximatiollll to elementary functions, Math. Tables Aids Camp.

8, ).43-147 (1954). .

[4.3) C. W, Clelll!haw, .A.. note on the. summation of . Chebyshev series, Math. Tables .Aids Comp. 9, 118-12-0 {1955J.· .

[4.4] o. rr. Har?y, A cO?I'8e Ofpuremath~matic~ 9th ed, (Oambndge U~IV. Pr~,. Cambncjge.,. J!in.gla.nd, and The Mal}~ Co., New York} :N,Y., 19~7). [4.51 C,. Haa. tin@l.,.". ,JJr., Appro~ma,tions for: digital computers CJ:'rinceton Univ. Press, Pnnceton, N.J.,

1955}. ...

[4.6] C. Hgstin~, Jr., Note 143, Math. Tables Aids Camp_.. _6,68 ('19.53).

[4.71 E. W.· Hobsen,' A ~rea.tise <?n pllWetrig-onj)In~try, 4th ed, (05II:\-bndge Unlv. Press, Cambridge, .England, 1918)_. . . . [4.81 H. S. Wall, Analytic theory ofcontmued fractions

.. (D_ Vap Nostrand 00., Inc., New York, N.Y., 1948)'.

Table!!

[4.9] E. P. Adams, ,Smithsonian mathematical formulae a,nd tables of -eUiptlc functi9!ls,.-3dreprint {The SmithsciIii&1;l Ip,s'titution, WaShington,. D.C.,

1957). ..

rUm H. Andoyer,N"ouv6!les' tables- trigon?ll:l~triqtlel:l fOIidam_entaJes (Henn.an:n et' fila, _Pa.rol) France, 1916}.

[4.11] British Association for the Advancement of Science, Mathematica.l Tablell, vol, I. Oircular a.nd hy:perbolie functioliS,. e-x,ponential,. sine ,lI,Ud cos4Je integrale, factorial function a:nd allied funetions, Hermitian p(oba.bility fun.ctlons,3d ed. (Cambridge Univ. Pr~ss, Cambridge, :mngland, 19fil).

f4.12} Chemics.) Rubber:ComP.anY1 8.t&nda.rdmath.ematlcal table!!. 12th ed. .(CherwceJ Rubber Publ. Co., Clev'ela.nd, ObiG, 1959).

[4.131 L.J_. Comrie, Chamhtll'S' six-figure mathematica.l tables, vol. 2 (W. R. Chambers, Ltd., London,

England; 19<19). .. ,

14.14J H. B. Dwight, Tables Of. integI'als and oth~r m5the~ . maticaldat&,,3d re v, ed, (The Macnullan Co.,

New York; N.Y., 1957)_. ..

[4. is} W. Grobner and N. '. H.· ofrei.t.~r, .Integr~tafel, unbestdmmte und bestunmte Integra.le (SpringerVerlag, Wien und ITIti6Pl'u!lkl~UStri$, 1949-1950),

r 4.16J Har-va.r~ Comp:zta.tion La.boratory! Tables of the function Sl'CSID .2 (Ha:rv5~ Uruv. Press, Cambridge, Masa., 1956).z=x+iy, O~:i;:S;4751 o _:::;; 11 :::;;47-5, 6D. varying interv&l.6,

[4.17] Harvard Oomputation Laboratory, Tabl~ of inyeI'$e hyperbolic functiolll,3 (Harvard Uniw, Press, Cambridge. Mass., _ 19:49). a.rotw-:z;, 0 ::::;;1;< 1; arcBinh ,3:, O,5;i<3.5; a:rcc.os,h, X, .. 1:S::'l:<~.5j arcsinh!ll, arcooah .:z:, 3.5;5$ $22980, 9D, varyIng fntervals.

[4.18] N ationalBureau of Standards, Tables 01 W·,. Appl!ed Ma.th. Series 27 (U.S. Governme.nt. Prl9-tmg Office, WlisliiDgton, D_C.,. 1953Y. . .t=O(.OOOO~)l, 10 D. Radix table of 10,,"10-")10=1(1)999) p =.3 (3) 15, 1:5D;

94

ELEMENTARY TRANSCENDENTAL FUNCTIONS

{4.19] Nationa,l Bureau or Standards, Table of natural 10g8!ithms fo~. arguments. between 'zer<? and five to sixteen deeimalplaees, 2d ed., A;pP!ied Math. Series 31 (U.S. <#Jvernment PrintIng office, W~hinJtf,on, D.C., 1953). x=0(.0001)5, 1.6 D.

[4.20] National ."Bureau of Standards Tables 01 the ex,. ponential function e":, 3d ed., AppUed Math. Series. 14 (U.S. Government Printing Office, WIJ.8~gton D.C., 19f:jl). z=o -2,4999(.0001) .9999, 180, :1:= 1(.0001) 2.499~ 15D, :1:=2.5(.001)4.999, 15D, x-5(.01)9.99, 121), z= - .000099(.000001) . 000099, 18D, z= -100(1).100, . 198 z=-9X 10-"(10-")9 X 10-"; 1'=10, 9,8,7,180; values of e and lIe, 25560.

[-l.2l] National Bureau of Standards, Table of the descending exponentii!1J,. z=·2.5 to· z= 10, Applied Matb;. Senes 46 (u .$. Government Printing Office, W81Ihington, D.C., 1955). z=2.5GOOl)10, 20D.

[4.22] National Bureau of Standards, Tables of sines and cosines for radian arguments, 2d ed., AppJi.ed ¥ath. Series 43 .(U.S. Government PrintIng Office, W81Ihington, D.C., 1955). sin z, COB :1:, z=0(..001)25.2, 0(1)100, SD, z=10-"UO-")9X 10-"; ~=5,. 4, a, 2, I, l_5D, :;;=0,(.00001) .01, 12D.

[4.23) National Bur-eau of Standards, Tables of circular and h~bolic sines and IlO!'lines for radian argumentS, 2d ed., Applied Math. S.eries 36 (U .8. Government PrintiDg Office, W81Ihington., H.C., 1953). sin z, C08 :z:,ainh z, cosh :Il, :i;=Or.OOOl)

1. 9999, 0(.1) iO, 9D. _ .

[4..24) National ..Bure'!'u of Sta.ndardB, Table of clrc~ar and hyperbolio tangents and cotangents forradllm argumsnta, 2d _Iiri!ltin~ (Celumbia Uni;Y. Press, New YOl'k, N.Y., 1947). ta.n:l:, eot a, tanh:!:, co,h x; z=O(.OOOl):2, 8D or 88, :1:=0(.1) 10,

10D. .

[4.25) National Bureau of Standards, Table. of sines 'and cosines to fifteetldeoirnal plaoes at hundredths of ad.eSI:~' AP.p lied Mat~. Series 5 (U.S. Govern~ent Prmtin.g Office, :Washington, D.C., 1949). BID:Z:, C08 :li, :z;= 0'1 (.01 0) 90°, l~D; supplementary table of sin Z C08 x, x=I°(1")89 , 30 D.

[4.26] National Bureau of Standards, Table. of secants and ceeeeanta to nine significanb figurea at hundredths of a dfilgl'oo, Applied Math. S~ries40 (U.S" Government Printing ~ce, Washington; D.C., 1954).

[4.27] National Bureau of St~dar~. Tables of funetioD8 and of 'zeros of lun.ctioDII, Oollected short tabl~ of the Comp:utation LaboratOry, ApP.lied Math. Series 37 (U.S. Government Printing Office, W81Ihington, D.C., 1954).

[4 .. 28] National Bureau of Stande.rds,Table 9f arcsin :c (Columbia Univ, Press, l'o!ew York, N.Y., 1945). arcsin z, z=0(.0001).989(.00001)1, 120· !!<p.xili~ table of f(v)=[!..--arcsin (l-v)]/(2t1)H, 11-0(.00001).0005, ian.

{4 .. 29] Natjonal Bureau of Standards, Tabl~ of arctan z, 2d ed.L-.. AI>~l.ied Math. Series 26 (U.S. Government. rrintmg Office, Wasbingtpn D.O., 1953). :t =;0 (.001)7 (.01) 50(,1)300(1) 2006(10) 10000, 12D.

[4.30) National Bureau of Standards, Table of hyperbolio aines and coaines. :t=2 ~ z~10, Applied Math . S~ries 45 (U.S. Government PriDting Office, Washington, D.C., 1956). z=~(.OOl)lO, 96.

[4.311 B. 0: Peirce, A short. table or integrals, 4th ed. (Gmn and Co., Boston, Mass., 1956).

[4 .. 32) J. Petera, Tea-place logarithm table, vols, 1, 2 (together 'With an appendix of ~athem~tical tables);. (Berlin, 1922; rev. ed., Frederick UDg~ Publ, oc., New York, N.Y., 1957). .

[4.33] J. Peters, Beven-place vilJ.u68 of trigonometric. functions for every ;housa..D,dtb of '80 degreeCBerlin-Friedona.u, 1918; D. Va.n Nostrand Co., Inc. No,!" York N.Y., 19~).

[4.34] L. W. Pollak, Rechenta.feln2l.ur harmoniachen Analyse (Jol:umn Ambr08i~ Barth, Leipzig,

Germany, 1926). . .

[4.35] A. J. Thompeon, Standard ta.ble of logarithms to twenty decimal p_lace8, Tracts for Computers, No. 22 (Cambridge Univ. Press, Cambridge, England, and New York, N.Y" 1952) ..

[4:36] J. Todd, Table of arctangents of ra~ional numbers, NBS Applied ~atb. Series 11. (U.S. Gov.ernment Pr,inting Office, Wa8hin~n, D.C:,> 1951). ~t&n min a.nd arocot mIn, O<m<n:s,I00, 12D; reduotdons of arctan mint O<m<n:$l'OOireductiO,D8 of-arctan l' for reaucible.n~2089.

(4.37] U.S. Department of Commetce, Coast and Geodetic E!urvey, Natur-aleines and cosines.to eig"ht decimal places, Special Publication No. 231 err.s. GoverDra~ht Printing Office, VI ashIDgtop, D .. C., 1942,) .•

[4.381 C. -E. Vap- Orstl'aQd, 'rable$ of th.e exponentIal function and, of the circular sine and cosine to -radian argumenta, Memoirs of -the National AC,lJ.(!emy of Sciepces ,J4, 5th.l\l1e!.iloi:r (U.S. Government Printing Office, Washington, D. C., 1921).

[4.391 B. V. Vega, Logarithmic tables ~f numbers, and trigonometrical functions (0. E .. Steehert &; Co., New York, N.Y., 1905); loglo z, :1:= 1(1) 100000; _ logarithms. of the trigonometrical functions for every ten seconds.

"See page II.

5. Exponential Integral and Related Functions

W,I\:L':r'ER GAUTBClill AND WILLIAMF. CAmL,L ~

Contents

Pa.ge

MatheJ:Datical Properties.. 228

5.1. Exponential Integral 228

5.2. Sine and Cosme Integrals 231

Numerical Methods . . . .. . .233

5.3 .. Use and Extension ofthe Tables 233

References . . . . . . . . . . . . . . 235

Table 5.1. Sine, Cosine and Exponential Integrals' (O:5.z::; 10) . 238

li-1Si(x), x-2[Ci(x)-ln x-,),]

x-l[Ei(Z)_ln x-"Y],x-1[E1,(x)+In x+1'l, x=0(.,01).5, lOS Si(x), Oi(x), 10D; Ei(x), E1(i),9D; x=.5(.Ol)2

Si(x), Ci{:z), lOD; :te-:Z Ei(:z:), xerr. Ei(X), 9Dj. :z:=2(.1)10

~

Table 5.2. Sine, Cosine and Exponential Integrals fo,r Large Argumetl.ts

(10::;:1:::;ro). . . . . . . . . . . . . .. . . . ... . ... 243

:J;f(t),. 9D; z2g(x), 7D; xe-ZEi(x), 8D,; :tC'E1(x), ien

J(x) , -.si(x}cos x+Oi(:t) sin ,:t, g(:t)=-si(x) sin x-Ci(x) cos x .X,..I'= .1 ( - .005) 0

Table 5.3. Sine and Cosine Integrals for i\.rguments n (0 ::;x:::; 10) 244

Si(rx), Cin(p), ;I;,=0(.I)1O,7D

Table 5.4. ExponentiaJ. Int~gra.1B E,,(z) (0S"xS'2) .

E~(z}-xIn X, E .. (x),. :n=3, 4, 10,20, x=0(.01).5 .En(;IJ), '11:=2,. 3,4, 10, 20, x=.5(.Ol}2, 7D

Table 5.5. Exponential Integrsls E .. (x) for Large Arguments (2::;x::; =J . 248

(:i:+n)e~EJJ(xL n=2,3, 4,10,20, z-1=.5(- .. 05).1(-:-.01)0, sn

245

Table 5.6. Ekpo.Q.ential Integral for Comple,x:.Argum.eMs (lzl<29) . .249, ze~Ei(2), z=x+iy, :1;=-':19(1)20, 11=0(1)20, 6D

Table 5. 7. Exponen tialIntegral for Small Oomplex Argumen ts (121 < 5) . 251

~2El(Z), zx+iy, ~=-4(.5)-2J y=0(.2)1, en

El (2)+ In 2, z=x+iy, :J::= -2 (.;5)2.5, y=O(.2)l, 6D

Thl!' authol'S a.cknowledge th"!la!!s~tance of David S. Liepmen in the preparation and checking of the tablee, Rpbert. L. D~rahfor thecomputatian of Tableo'S.2, and Alfred E. Bep,m for tl:i.e eomputetion of Tahle5.6.

I Guest worker, Na.tkmaJ Bureau of Standard.s, from the America.n University. (PresentlY

purdue University.) . ,

I National Bureau oC StlWdards.' (Preeent~y N ABA.)

227

5. Exponential Integral and Related Functions Mathematical Properties

5.1. Ewonentiallntegral

Definitions

5.1.1

(Iarg 21<11")

5.1.2

r'" e:' ('!z e'

Ei(x)=--J . -dt=..LJ -dt

. -$ t _., t

5.1.3

liC ) - t: d. t. -Ei(ln ) x -To hi t- . x

(z>l)

S.I.4

(71,=0, 1, 2, ... ; !?lz>O)'

5.1.5

5.1.6 fJn.(2)=f1 t"e-Udt -1

(n=O, I, 2, ... )

In 5.1.1 it is assumed that the path of integration excludes the origin and does not crOSB the negative. real axis.

Analyti.c continuation of the functions in 5.1.1,

5.1.2, and 5.1.4 for n>O yields multi-valued-functions with branch points at 2=0 and Z= 00.3 They are single ... valued functions in the a-plene cut along the negative real ~.' The :function ll(z), the logarithmic integral, has an additional branch point a.t Z= 1.

Interrela tiona

5.1.7

E1(-z±iO)=-Ei(z) =t=i1l", -Ei(z)=tlE1(-x+iO)+E1(-z-iO)]

• Some authors [5.14], [5.16] usa theentiref1.lllction 50l (l-e-')dt/t 88 the basic function and denote it py Ein{z). We have Ein(z)..=E1{z)+ln z+.,..

·t Va,dotls &)lthors definetheintegrsJ. J~w (e. ,/t)cU in the s-plane cut along the positive real axi8 and denote it also by Ei(i). For %=z> 0 addi:tioIial notations such as' Ei (z>. (e.g., in (5.10], [5.25]), E"'(.:~) (in [5,Z]), Ei"''(z), (in [5.6]) are then used to designate th_e principal value of the integral. Correspondingly, ,E'I(X) is often denoted by ":"'Ei(-x).

2251

518 ( )- , -,,-1 -If (1+ _j__~+ +3")

•• OW z -n.z e ~'2!_. .. -rtf

5.1.9

..,2 "

(J (,)- , -II-I{ ~[1 +". +( 1)"·z]

" 2 -n.z e' . -z 2!-··· - nl

_ 22 .2"

-e ~ (1+2+21+ ... + nl)}

'J

3

~i(.J

-3

FIGURE 5.1. y=Ei(x) and y=Et(x).

FIGURE 5.2. 11= En (x)

n=O, 112,3,5,10

229

EXPONENTIAL INTEGRAL AND RELATED FUNCTIONS

y

5 0-0 ".1 n'2 n,'3 n·4 no5 n'6

4

2

0:

.5.

FIGURE 5.3. y=a,,(z) 11.=0(1)6

FIGURE 5.4. Y= {J" (x)

11..=0, 1,2, 5, 10, 15

SerieA Expanaiona

5.1.10

_ ~z"

Ei(x) ='Y+ln x+ 2J -I

_Inn

(jarg 21<1;-)

5.1.12 E,,(z)

n-1 1

.p(l)=-"),, .p(n)=-'Y+ 2:-

m-l77l

Clarg 21<.".) (11.>1),

'Y = .57721 5664·9 . . . is Euler's eonstant.

Symmetry Relation

5.1.13

R.ecurrence Reilatioll!l

5.1.14

(n-:-l,2,3, ... )

5.1.16

z,B .. (.z)=(-1)f1i·-e~r+n,B"_l(z) (11.=1,2,3, ... )

Inequalitiel[5.81J [5.4]

5.1.1,7

n-l <.

-E .. (x) <E,.+l(X) E,,(x) (x>0;.n=1,2,3, ... ) 11.

5.1.18

E~~x)<En_l (:z:)E"+l (x)

5.1.19

x~n < e7iE,.(:z:) ~:z;+!-1

(x>O; 11,=1, 2,3, ... )

5.1.20

i In (l~)<e;El(x)<ln (l+i)

(X>O; n=l, 2,3, ... )

'Continued Fractlon

Special Valuee

5.1.23

5.1.24

5.1.25

e-' 2

aoCZ)=-J Po(z)=-sinh e 2e

230

Derivatives

EXPONENTIAL INTEGRAL AND R.ELATED FONdTIO?f~

5.1.26. d~;2)=_E._l(Z) 5.1.27

d~ d~-l

dz" [e'E1(z)]= dz .. -1 [e2' EtCz)]

t(-l)"(n-l)!

z" (1i=I,2,3, ... )

D~f;ip.ite and Inde6ni~- Inte8Tals

(For more extensive tables of integrals Bee [5.3], [5.6], [5.111, (5.12], [5.13], For integrals involving E .. .(x) see [5,9].)

5.1.28

:5.1.29

5.1.30

So'" t-ib

--_ eifJ1dt=e4bE (al.)

o t~+b2 . 1 U

5.1.31. _

r'" t+ib(tI,-1 -01&( E'( J.) ,

J 0 t2+b-2 e .,.,t= e - 1 all +I.'ll")

5.1.32

J:'" e-CU-e-b'. b

. dt=1n-

o t a·

5.1.33

So'" EW)dt=2 In 2

5.1.34

J:'" e-d' ElI(t).dt=

(-It-1pn (1+a) + ~ (-l)1'al>]

a ~ k .

(a>-I),

5.1.35

r1 ea' sin bt . b

Jo t dt=r-afctan a +JE1(-a+ib)

(a>O,b>O)

5.1.36

(I e-a1sin bt b

Jo t dt=arctan a +..FE1(a+ih)

(a>O I b real)

5.1.37

fl ea~(l-cos at) dt=! In (1+ b:\+Ei(a)

Jo t Z ~J

+9t'~(-a+ib) (a>O,b real)

5.1.38

ile-ar(l~COS bt). dt-~ In (1+ !~-El(a)

+9l'E1(a+ib) (a.>o, a real)

5.1.39

5.1.40

fO; e'-l

Jo. -~-dt=Ei (x)-lnx-,,(

5.1.41

f e'Z. i . .

aZ+r d.X=2a [e-aEl(-a-",x)-e~El(a-tz)l

-l-const.

5.1.42

f a:~Izz2 dx=-~ [e-aEl (-a-ix)+etlE1 (a-ix)] -l-const,

5.1.43

f .,+8" 2clx=-! J(e!aE1(-z+ia))+C:OD$t, (a>O)

a x a .

5.1.44

f a~%t7 dx=-Pi'(eiaE1(-x+ia»)+const, (a>O)

Relation to IncoDlpJete Gamma Functio.n (see 6.5)

5.1.45 5.1.46

ElI(z)=zfI-1r(1-n, z) a,,(z)=z-n-1r(n+l) s)

5.1.47 ~ .. (z)=z--l[r(n+l, -z)-r(n+l, z)]

Relation to Spherical DeMel Functioll8 (see 10.2)

5.1.48 ao(z) =-/!z Kj(z) I ~o(z) =~ Ii(Z)

5.1.49 al{z}= /2 KS/2(Z), .81(Z)=- !211'18/2(2)

1V.z 1{ z

EXPONENTIAL INTEGRAL AND RELATED FUNCTIONS

231

Nwnber-1'heo:retic Signi6eaneeqf li (:Ii)

(Assuming Rie;mannJshyp()th~is that all nonreal zeros oft(z) have a real part of i-)

5.1.50 li ex) -1r.(x) - 0 (.jX lne) (x~ III ) T(Z) is the number or primes lesstha:n or equal to x.

y

150

200

50

ji(xj .-

... ,

.> lI"(d ."

",

",

...

'"

",

'"

'" '" ",

,.-

",

'" '" /'

'" /'

/.

100

FrGURE:5.5. y=li(x) and: Y=lI"(z)

5~1.51

EtI(z)-- e~· {1_~+n(n~1)_ n(n+1~(n+2)+ ... }

(Iarg z!<j?r)

5.1.52

E(x)=e-t {1+ n 7l(n-2z)

.. ·z+n (x+n)2+ (:ll+n)4;

+n(6$~~!:ftj-~·') + R(n, x)'}

-.36n-·,sR(n,x):$(1+ X+~_l)n-' (x>(»

Po1ynoDliaiand Rational ApproximatioWl I 5.1.53·0$;1:$1

El (x) +In x.=ao+Oix+a.r+a¢'+a¢,+a.a6+E(z) ]E(z)j<2XlO-7

J The. approximation 5.1 •. $3. is from E. E. Allen, Note 169, MTAC 8,240 (1954); approximations S.l.S. and 5.1.56 are: from C .. Ha.etiu8IJ, Jr., .Appr.oximatiOrni for digital computers, Princeton Univ. Press, Princeton, N.J., 1955; approximation 5.1.55 is moDi C. H.a,stipga, Jr., Note 143, IylT.,4..C 1, 6S (195'3) (with permw,ion).

Go=-.57721 566 al= .99999 19.3 G:l = - .24991 055

aa= .05519 96.8 a.,= -.00976 004 a6~ .0010,7 857

5.1.54

at=2 .. 3347.33 t_b- .250621

61=3.330657 6,= 1.681534

5.1.55

~-4.03640 a,=1.15198

h1 ±::.5.0a637 bs=4.19160

5.1.56

al= 8 .. 57332 87401 <Z:I =18.059016913 0 as= 8,.63476 08925 a4= .26777 37343

b1= 9.57332 23454 b~=25.63295 61486 63=21 . .09965 30827 b,= 3.95849 69228

5.2.1

5.2. Sine apd Cosine Integrals DefinitiOD8

. i'sin t ,'. 8Hz) = -: - at

o t

5.2.20

Ci(z)='Y+lo z+ f,l cost-ldt

. Jo t

5.2.3 7

Shl(;;:)= r- sm;t dt Jo

S.2.41

. r' eoah t-l

Chl(Z)='Y+ln z+Jo' t. . dt

(Iarg z]<r)

~ Some authors [5.14], [5.161 use th.e entir.e function L"(l-COS t)d!./t sa the basic funllti.on and denote it by Cin(i). We have

Ciu(z) = -Oi(z)+ln t+,..

7 The notations Sih (:e) = r" sinh .~. d!.le., Jo

Cinh(z) = r' (cosh t-l)d!.# have also been proposed [5;"14.] J.

232

EXPONENTIAL INTEGR,AL AND RELAT,ED FUNCTIONf3_

"

5.2.11 Ci(z) +El (z)= 502 e -ot COl I sin (2 sin .t)dt

5 212 f() I" sin.t d't f" 6-" dt (n;, >0) 5.2.22 Si(ix)=-2i,IEi(x)+EI(x)] (:1;>0)

' ••• " :e=JQ t+z ,= Jo t2+1' ,·';nZ ' '

5.2.5

si(~)=Si(z)-~

AUxilia,ry Functio~

5.2.6 5.2.7

!(z)=Ci(z) sin z-si(z) GPS Z g(z)=-Ci(z} cosz-si(z) sin z

Sine and Cosine Integral~ in Terme of Auxilia,ry FUD.etioD8

5.2.8

Si(z}=~-J(z) cos z-g(z) sin e Ci{z)=f(z) sin Z-[l(2) cos-a

5.2.9

Integral Repl'~ntat;joJi8

T

5.2.10 si(z)==- f1ir:coet cos (2 sin t)d't 1;' o.

r~ COS t, fa> te-ZI

5.2.13 g($)= Jot+z dt= Jo t2+1 dt

Si(xj

1.0

FIGU1'lE 5.6. y:=:Si(x) and y:-Ci(x)

Series Expansion,,!

5.2.14

, ... (-1}"':#t1+'

Sl(Z)=t,; (2n+l)(2'n+l)!

Si(z)=1r ~ J~+I (~)

,5.2.15

5.2.16

. ' '" (_1)lI:z2n

Ci(z)='Y+1nz+:E 2n(2)1

R=l " , n

5.2.17

• 0> ZZ~+I

Shl(Z)=~{21}+1)(2n+l)j

Chi(z) ='Y+ln z+ ~ 2n'(~)!

5.2.18

SYIPmetry Belatione

5.2.19 5.2.20

Si(-z)=-Si(,z}, Si(zJ=Si{z)

Ci(-z)=Ci{.z) -·i1r Ci(Z)=Oi(z)

ReI_ation t,o &po~e.nt:iallntegral

5.2.21

-

Si(z) =;i [E, (iz) - EI r- iz) ]+~

<lug zl<~)

5.2.23

Ci(z)=-~ [Ej(iz)+E,(-iz)]

5.2,.24 Ci(iz)=~ [Ei(X)-E,(x)l+i~

Value Ilt I~nit)':

5.2.25

lim Si (x) ~

.;r"cg,

Integrals

(For more. ext ensiv e tables of integrals see [5.3J, [5.6], [5.111, [5.1:n [5_.13].)

5 6 f.'" sin t d .. ( )

.• 2 •. 2' ,_ . t=-Sl Z

• t

(Iarg zl<1I")

i'" cos t .

5.2.21 . =:: dt_~--Ol (2')

• t.

1'" .... 1

rat 81 (t)dt= -~ arctan (L

o a

5.2.29

5.2.30 1"' cos t ei (t)dt= i'" sin tsi (t)dt=-i

EXPON.ENTIAL INTEG-RAL AND RELATED FUNaI'WNf)

5.2.31 ill) CP (t)dt= l'" si2 (t)dti

5.2.32* Sa'" Ci (t) sci (t)dt~ In2

5.2.33

11(1 ~t -7 cos bt dt~ In (1 +~) +Gi eb)

+8iEj{a+ib) (~real, 0>0)

Asymptotie Expan.sions

5.2.34

(Iugzl <11')

5.2.35

( .). 1 ( 31 51 7! )

9 z ....... Z2 1- Z2+Z'- Z6+' . "

(jarg zl<1I')

Rational Approximatiotl!! e

5.2.36

l'::;x< CP

. _ . 1 (:z~+alXa+a2) J(x)=-x i~+blX~+bt +t(x)

IE(X)I<2X 10-'

a1=7.241163 ~=2.,463936

b,=9.0~8580 b'J_-_7.1S7433

233

5.2.37

'. 1 (X'+alg;2+~)

g(x) = :&2 ,X.+Olx2+b'2' +f(X)

jE(:t)/ <10-1

0.),=7.547478 ~=L564072

b1=12.723684 b2= 15.723606

5.2.38

j (x) =~. .(iC8talze +~xt +0.3;1;2 +(4)+ E (e)

-' Z 'x8+blXB+b~z·+OilZ~+b4

jc:(x)I<5XlO-7

~ = 38.027264 llj=265.187033 I!s= 335.,677320 a,= 38.102495

Oi - 40.021433 ba= 3.22 .624~n 1 oa=570 .2~6280 h, -= 157 .105423

5.2.39

. ' 1 (x'+CLlx8+(kx.+a:ax2+a,) g(X)=x2;tB+hlZG+b2X'+baX2+b, +e (x)

IE(z)l <3X 10-7

_~l= 42 .. 242855 ~-302. 757865 aa =352.018498' a,= 21.821899

b1 ,-,-- 48.196927 b2= 482.-485984 ba=1114.97S8.85 b4=449.690326

Numerical Methods

5.3. Use and Extension of tbe Tahl~8

Example 1. Compute Oi (.25) to5D. ]j'rom Tables 5.1 and 4.2 we have

OJ (.25) - h1(.25) -'Y-.249350 . (.25)2 . . J

Ci (~25) = (.25)~(- .249350) + (-1.38629)

+ .577216.;_ -.82466.

EX8rppi.e 2. Compu teo Ei (8)00 58.

From TallIe S.l we have xe~itEi (x) =1.18185 for x=$. From Table_ 4.4, e8=2.98096XIOa:. Thus Ei (8) =440.38 ..

-:See p_.,e n,

8 From. C. Hnstings, Jr., Approximations for digital computers, Princeton Univ. Press, Princeton, N.J., 1955 (with permission).

&!ample 3. Compute Si (20) to 5.D.

Since 1/20=.05 from Table 5.2 we find j(20)=.0497'57,g{20)=.Q,02464. From Table 4 •. 8, sin 20=.912945, cos .20=.4080S2. Using5.2.8

Si(20)=i-J(20)COB 20-g(20)sin 20 =1.570796-.022555=1.54824.

Example 4. Compute FJ",(z), n=l(l)N, to 58 for x=L275, N=lQ. -

If .:1: is less thaa about five, the recurrence relation 5.1,14 can be used in increasing order ofn withoutserioua loss of accuracy.

By: quadratic interpolation in TaMe 5.1 we get EI (1.275) = .14'08099) and from Table 4.4, e-I.'Z76 =.2794310. The recurrence formula. 5.1.14 then yields

234

EXPONENTIAL INTEGRAL AND RELATED FUNCTIONS

. .

7i E .. (1.27.5) E .. (1.275)
1 .1408099 6 .Q430168
2 .0998984 7 .03.74307
3 .0760303 8 .0331009
4 .0608307 9 .. 0296'534
5 .0504679 10 .0268469 Interpolating directly in TableS." for "1ji= 10 we get E10(1.275)=.02(J8470 88 a check.

Exampl~5. Compute En(x),n=l(l)N," toSS for ~=IO, N=IQ.

If·,as in. thiseX'8IP,ple, ;l: is appreciably larger tha-n five and N 521, then the recurrence relation 5.1.14 m..ay be safely used in decreasing order of 11.([5.5]). Fi'-6ni Table 5.5 for ~-1=.1 we get (rt 1 O)~El0(X) = 1.02436 BO that EI(I(1 0) =2.32529 X1'0-o, Using this 8.8 the initial value we obtain column (2).

lQ3E .. (10) 1 ()liE" (10)
n (1) (2)
1 .41570 .41570
2 .3830.0 :38302
3 .35'500 .354~8.
4 .33000 .33041
5 .31000 .30$98
6 .28800 .29005
7 .27667 .27325
8 .25333 .25822
9 .25084 .24.4.72
10 .22573 .23253 From Table 5.2 we get $e:tE1(z)=.915633 so that E1(1(l)=4 .. 15697X10-e as 8. check. Forward recurrenee starting with ~ (1 0}=4·.1570 X 10-8 yields the values Incolumn (1). The underlined

:figures a.re in. error. .

Example 6. Compute E,,(z), n=l(l)N, to5S [or X= 12.3, N-20.

If N is appreciably larger than ;2:, and z appre .. eiably larger than five, then the recurrence relation 5.1.14 should be used in the baekward diree~ion to genera te E,,(x) for n<nQ, and in the forward cfuection togene:rate ErI(z) for n >no, where

fto=(t).· I

From 5.1.52, with nc~12,:r=1.2.3, we have

e-u.a .. .

EItfJ (z) - 24:3 (1 +.02032- .00043 - .0000 1)

.' =1.91038XlO-7•

Using the reeurrence relation 5.1.14,8.8 indicated, we g(!lt

n 1 OIiEn (12.3) 1 08ElI (12.3) ''11.
12 .. 191038 191038 12
11 .19·9213 . 1'834;98 13
10 .208098 .176516 14
9 .21.7793 .170042 15
8 .228406' .164015 16
7 .240073 1583.97 17
6 .252951 .. 153144 18
5 .267234 .148226 19
4 .. 283155 .143608 20
3 .300998:
2 .321117
1 .343953 From Tables 5.2 a-nd 5.5 we find Ei (12~3)= .343953 XlO-&, E2(M2.3) =.143609 X 10-!l as a. check.

Example 7.. Compute all (2) to 68 for 'Ii = 1 (1) 5. The recurrence formula 5.1.15 can be used for an z >0 in inereaeing ordetof'n without loss of

accuracy. From 5.1.25 we have ao(2)=~e"'2 =.0676676,80 we get

fa, .... ,,(2)

o .0676676

.1 .101501

2 .169169

3 .321421

4 .7105,1'0

5 1.84394

Independent ealculation with 5.1.8 yields the same result for (16(2).

The functions ao·(i) and C¥1(:l:) can be obtained from Table lO.8 using 5.1.48,5.1.49.

Example 8. Compute ~",(rt), n.=O(l)N to 68 lor x=l, N=5.'

Use the recurrence relatien 5.1.16 in increasing order of 'It if

x>.368N+.184 In N+.821

and in decreasing order oIn otherw.U!9 I5.5].

From 5.1.9 withn=5 we get ,6s(1)=-.324297 correctly rounded to oD. "Uaing the zeeurrence formula 5.1.16 in decreasing order O£,I nand ClID"ying- 9D we get the- values in column (2).

.6 .. (1) P .. (l)

n. (l) (2)

o 2.35040.2 2.35040 2389

1 -.73575 9269 -.73575 8.880

2 .87888 iJ849 .87888: 4629

3 -.44,9509n2. - .449507;383

4 .55236 3499 .55237 2854

5 - ;32434 3774 - .32429 7

Using forward recurrence instead, starting with

EXPONENTIAL INTEGRAL AND RELATED li'UNCTIONS

.60(1)=2 sinh i=2.350402 and &;ge.mca.rryip.g 9D, we obtain column (1). The llnderJiJ?e.dfigmes are in errol' . The above shows that, tbreesignificant figuresere lost in forward recurrence, whereas about three significant figures are gained in backward reeurrencel

.A.nalterna.tiveprocedu:re is to start with. an arbitrary va.luafor 'It sufficiently large (see also [5.1)),. To illustra.te, starti.)lg -with the value zero atn= 11 we get

n ,6, .. (1) n /J .. {l)
11 O. 5 -.a2~2g7
10 .280560 4 .552373
9 -.2.06984 3 -.449507
8 .319908 2 .878885
7 -.253812 1 -.735759
6 .404621 0 2.350402 The functions 130(:£) and 13r(~) ca.nbe obtained, from. Table 10.8 using 5.1.48, 5.1.49. •

Example 9. Compute, El(z) for, z=3.2578 +6-.8943i.

From Table 5.6 we have for zQ=zo+iYIJ=3+7i

ZaetOEl (20) = .,934958 + .,095598i,

e~DEI(ZO) = .059898- .107895i.

f(z) f(Zo+.:lZ)-·}(Zo)+j'i~) As

fit (:.0) '2

+---m- (liz) + ...

Tat3

[5 . .11 F. J. Ct>r1:ia.tP, On the computa.tion <if auirl1ia.ry fUnctions fOr t·wQ..-aentfi)" Integrals by means of a.: hjgb---apeed (:omputer,. J. Chem. Phys, 24-.452-453 (1956).

[5.2J A. Erd61yietal., Higher tranecetLdental fJmctioDe., vol, 2. (McGl'll-w-H;1t. Hook Co., Ine., New York, N.Y., 1958).

[5.~1 A. ErdByi et al., Tabl~ of integral transforms, vols.

I, 2 (McGraw-am Book Co., Ino., New York, N.Y., 1954).

[5.4) W . Gautaehi, SOme elements,ry inequalities l'$tirig to the. gamma and Incomplete g.amm.a function, J .. Math. Phys, 38;77-81 (1959)..

[5.;51 w._. Gautsc:hi, RecU+B_ive computation. of c-ertain ·inte8;r'als,. J.. Assoc. Comput. M.ach. 8,' 21--40

(1981). . .

Refer.encea

235

withAz- z- ~=.2578 - .1057i. Thus with 5.1.27 we get

Fe

o 1 2 3

. 05,9898 -. 107895i .008174 +.0l2795'i -.00185Il +.000155i . 000088 -. 000212,

. 059898 -. 107895i .003460 + .. 002435i -. 00009~ +. OQOllOi

-.000003 -.000004.\

1(2)=.063261 -.105354i e-'=.031510 -.022075i El (2)= - .000332 - .004716i

Repeating, the calculation with Z{,= 3 +6i and .1.2=.2578 + .8943i we get the same result,

An alternativeprocedure is to perform biva;ria.te inte;rpo1a.tioIJ.. in the real and imagillary parts of u1E1(z).

Example 10. Compute Et (:::) forz= -4.2 +12.7i.

Dsing the formula at the bottom of Table 5.6

sE ( ) .7n093

e . 1 2 ""'-3.784225+ 12.7i

.278518 .010389 -_

+ -1.90572+ 12.7i+2.tI900+12.7i

= -.0184106-.0736698i

Et(z) ~-1.87133'-4,70540i.

[5.6) W. Grobne.rBiDd N. Hofrejter, Ip,tegra.ltafel (Springer-Verlag, Wlen and Innsbruck, Austria, 1949-50).

[5.7] C .. Raatinga, Jr., Approximations for digital'computers (Princeton Univ. Press, Princeton" N:J .• }'955) ..

{S.8]E. Hopi, :Mathematical problems of radi.a.tive equilibrium, Cambridge Tracts in Mathematics and Ma.thema.tical Physics, No. 31 (Oa;mbddge Univv Prese, Ca.mbridge, Engla.nd, 1934) ..

15 .. 9] V .. Kourganolf, Basic methods in tra.nsfer preblema (Cbfmd Univ. Press, London, Engla.nd. 19-52).

L5.1QJ F. Ldseh and F. Behobltk, Die Fakultii.t und ver-

. wa.ndte Funktionen (B. G. TeUbner, Leipzig, Germany, 1951).

[5.11] N. Nielsen, Theerte dell IntegraUogarithmus. (B. G.

. Teubner, Leipzig, Germany, 1906).

236

EXPONENTIAL INTEGR4L AND RELATED FPNCTrONS

[5.12) F.Oberhettinger, Ta.fueUen sur FoUrier Trans:fOrmation (Sprmger. Verlag, Berlin, G6ttIngen, Heidelber'g, Germany, 1957).

[5.18] L M. Ryahik and I. S. Grl!odstein, 'I'ablea of series, products and'integreJs (VEB Deutscher Vedag

der WissenechafteD, Berlin, Germe.ny, 1957). '

[5;14] S. A. Schelkunofi', Proposedeymbole, for the modifled co,sine and exponential iIitegT:aJ., Quart. ,Appl. Math. 2.90 {1944}.

[5.15] J. Todd, Evaluation of the,eXponentiaJ integral for large complex ,arguments" ,J. Research NBS 52, 313--317 (19t!4) RP 25Q8.

[5.16] Jr. 0, Trico:rni. Funsioni ipargeQmetric.be confhienti (EdlzIoIii Ctelilon~l?, ROlfie~ It,aly. 1954).

Tables

15.171 British Asso[:iaiion for the Advancement of 8ruence, Mathematicsl Tables, veil. T. Circular and hyperbolic funetions, exponentiel, sine and cosine integrals, etc., 3d ed. (Cambridge UDiv. Press, Cambridge, England, 1951). EI(x) -In :(:, - El (x) -In. x, Ci(x.)-iil x, Si(.:c) , z=O(.l}o, UD; Ei'(x),

x-5(..1)15, to-US, Ei{X), x=5(.1)15, 13-HDj Si(2;), Ci(a:), iI;,=5{.1)20(.2)40j. 100.

[5.181 L, Fo:!:, Tables of Weber parabolic cylinder functions and other functions fot l$J'ge I!J'gJlJDJ;mts, Mathe._matJcal Tables, vol, 4, National Physical Labnre-tory (Her Majesty's ,Stationery Office; London, England, 1960). e~'"Ei(x)" e"E'I.(:li), X~I= o (.OOl).1,

lODj !(x) , ,g(x),x-I=O(.OOl).l, IOn. -

[5.191 J. W. L.Glaisher, Tables of the numerical values of the' sine..integr;a.l, eosfne-integral and expo_oentiaJ integral, Philos. Trans. Roy. Soc . . London 160,367-388 (1.810). SH::z:) , Ci('i), EHx), -E1{x), x:""O(.OI)1,18D, x=I(.1)5(1)15, llD.

[5.20] B.S. Gourary and ¥. E. Lynarn, Tal;>les of the auxiliary molecular integralsA .. (x) and the auxiliary functjOIis C .. (x) , The Johns Hopkins Uoiv. Applied Physics Laboratory, eM Report 905, Baltim.ore, Md. (19_57). a~(x),nlt,,(x), x=.05(.05) 15, n=O(l) 18,98.

[5.21] F. E'. Harris, Tables of the exponential integral Ei (x) I Mati:l.Tables Aids Compo 11,9-16 (1957). EJ (x) I e''EJ'(z), Ei(l:), e-"Ei(x), X.= 1(1)4(.4)8(1)50, 1&-19S. .

(5.22) Harvard- UniverSity, The A:nnaJa of the- Computation Laboratory, vola. IS, 19; Tables of tlie generalized sine- and eosine-integral functions, parts I, II (Harvard Univ. Preas, Cambridge, MMS.,

1949). B(a,:z:}= f.'-'siDUih;,C(a,x}= f.'-'l-00sUdx

.Jo 'It Jo· 'U . "

'6D S (' ") p: ain u. . 0_

-; . sa, x= Jo' .. u.-em xdx, m:;{a; .:z;) =

fc~ sin u, . "'):" COiJ·'U. .

-- cos :z;dx, 6D;. CeCa, :z;}= --'.-··sm :r;dx,

(I 11 .. 0 u

'. f%COS-U;

Cc{a, x) = Jo --u-(l-cost)$, 6,Dj'U.=..,!:i:'+a2,

0~a<26, O~x~25.

[5.23J Harvard UniVersity, The AlinalS of the06rnputa.tion La'boratory,-vQL '21; TabJes of the generalized elqlonential-integrru fu.net'ioD,l! (HarvlUd Univ. Press, Ca.mbridge; Mass;, 1949). B(G, x)=

1% 1-e- 1x-e~ sin·u. .

-- dx, Ea(a, x) =. .. dx Ec(a x)-

o u .... Du I ,-

rx 1-6-,,0 cos U

Jo'U _ th, 6D; u=..,!iJI+riJ, OSa<10,

O~:z:<lO.

{5.24] _A. V. Her.shey, Computing progre.,tns for. the complex exponential :integral, U.S. .Navel PrOving Ground, 'Dahlgren, Va., NPG R~port No .. 1646 (1959) ... -.Ed..,...z),. z=z+iy, ':1:= -20(1)20, y=0{t)20, 13S.

{5.25J E. Jahnke ~nd F. Emde, Tables of functions, 4th ed. (Dover Publications, Ine., New York, N.Y., 1945). -Et(:t:), Ei(x), x=IY( .. {)l) 1(,.1)5(1)15, 4- 58j 8i(z), eiCx),. ;c=O(.Ol)1(.1)5(1)1'5(5)100(10) 200(.100) io=», generally 4-58; ma.xiraa· and

mii:J.ima of Ci(z) .and s1(:z:), 0<x<16,58. - .

[5.261 :a:. A. KarpOv and '8, N.Ra;>;umovskil, Tablitsy integral' nogo logarifina. (Izdat. Akad. Nauk SSSR., Moscow, U.S.S.R., 1956). H(x) , :1:= 0'(.,0001) 2.5 (.001) 2.0 (.01)200 (.1) 500(1) 10000(10) 25000, 78,; li{it)-ln II-x), :1:=.95('.0001)1.05., 6D.

(5 .. 27) M. Kota_ni, A, Amemiye., E. Iebiguro, T .. Kimura., Ta:ble of moleculwintegrals (Ma.rUZeIi, Co., Ltd., Tokyo, Japan, 195!». a,,(z),.z= .25(.25)9(,5).l9{1) 25, n=0(1)15, 118; fJ,..(:z:), x=:25(.25)S(.5)19(1) 25,11.=0(1)8, US.

[5:28J M.¥ashiko, Tables of generalbed expouential-, sinew andcosine-.i.ntegr.ais, Numerical Computation Bureau Report No.1, T(:lkyo, Japan (1953). E1(z)+ln 10:1= C.a;) +In ~-iS.m., z=~"', e='O(.05) 5, a=OO(2C1}600(1")90°,6D; ze':EI!'(z) = A;;. ('1/) exp

(i<l> .. ('I)]" z= ~tico, n= .01 (.01) .2; a= OO{2~)60" (1"). ~IO°., s-:6P.

[5.29} G.F. Miller, _ Tablfll! of generalized exponential integrals, Mathematical TaibleB", vel, 3, National PhySical· Laboratory (Her Majesty'a Sta.tionery Office, London, Engla.lld, 1958). (x +n)!FE .. (:z:), x=O(.Ol)l,n= 1(1)'6,x=0(.1)2Q, n= l{l} 24,. X:-I=, 0(.001).05; 1&=1(1)2;4;' 8D.

[5.30] J. Mill¢r,J. M. Gerha.uaex, an,d.F. A.Mat~e.Q, Quan:tum. ¢hemisti:y iD,tegt!i.ls I!,Dd ta.blel:i (Unjy. ()f 'J;'ex{ia Pres!,O, Au&tib, Te.x., HI59). a,,(z), x=.125 {.l25)"25, n= 0(1) 16" 148 j{J.(:z:},:i:= 0(.125)24.875, 11.=0.(1)16, 12+-1~.

{S.SI} J, MUlex and R. P. Hlll'St,SimplUied cllolcu1ation of the exponential integtw, Math. Tables Aids COmp. 12, 1B7-193 (1958:). e-'!'Ei.(:Z:), Ej(i) , e'"l!lt(z), E1(:z:},:z:>= .2(.0.5) 5 (.1) 10(.2) 20. (.5)50(1),80; 168 .

[5.32] NationaJ Bureau of Standards,. Tables of sine,

. coaineand exponential integrals, vol, .I (U,S.

Gov-el'nment Printing Office, W.aShington, D...C'J 1940.). 8i(z) , Ci(:z;) , Ei'(x) , El (:r;) , x=O:(.OOOl)2,

:r;~O(.1)lO;9D.. .

[5.33) Nationsl BUreau ofStandarda., Te.bles ofsiIie, castile a;nd exponential integrals, vol, II (U;S. Government PnntiD;g Office, W9.I!bington, D,C" 1940,). 81(x) , Ci (x) , Ei{x) • .lJMz), z=0(.001)10" 9-10 D or B; Si Cx} , Ci{z}, Z= 10(.1)40, IOD:; Ei (:z:) , E1(t) ,x=' 10(.1) 15, 7--n8.

EXPONENTL\LJl'frEGRAL .um RELli.'I'ED FUNCTIONS

.237

[5.34JNational Bureau of Sta.ndards, Table of sip,e and cosine integrals for argumente from IOto 100, A-pplied Math. SerIes 32 (U.S. Government Printin,;g Office, Washington, D.C., lQS4). Si{x) , Ci(x) ,x= 10(,(\1) 100, Lon.

[5~35JNa,.tions.1 Bureau ofStandaJ"rls; T.abli!s of furicii-Oris and of zero~ of functions, Collected shortt·ahles ()fth~ Computation Laboratory, Applied Math. ~eries37 (U.s. GovernmElnt Printing Office, W sshingt on, D.O., 19'54). E,,(x), n=O(l}20, x=O(.Oi)2{.i)lO, 4.-98; .E~(iI:)-xLn .x, z~O(.Ol) .5, 78; E~(x)+ ~ In x, x;=;O(.OI).I, 78.

[~.361 National Bureau of Stimda£ds, Tables of the ex'po.llential integraJ, for complexergumenta, Applied Math. SeriesSl (U ,8. Government Printing Offiae, Wa.stiington, D~C., 1958), E1{z)+lnz, 6D, i,=O(.02)1, y=O(,02)l, 2:=-1'(.1)0, '11= 0(.1)1, E1(z),6D, :£.=0'.02)4, y=0(,02)3(,05)lO) :z:=O(1}20, 11=0.(1)20., ;;:=-3.1(.1)0., y=0{.1}3.,1, :z;= -4.5'(.5)0., y=a(.1)4(.5)10; x",,-10(.5)-4,5, y=0.(.5)10, X= -20.(1)0.,,,=0.(1)20.; e'E1(t}, en, ':c=4(.1) 10, s= 0.(,.5) 10..

[5.371 8, OperlandElr, Ta.t}!lllen von Exponentia.Ji'uuktionen und-Irrtegraleu zut Anw~ndung auf Gebieten dl»' Thetmody·na.mik, Halbleitertheorie und OasJtinetik (Akad.emie- Verlag, Berlin, Germgny,

, AE vr " (-IlE)'kT' (-t:..E)

1959). 11:1" /lE' ezp kT '6E ezp kT '

CAE)' k (T ,(-aE). AE (ilB)

EL kT'tlEJo eXP1cT .' aT, k,Te:xP kpX

E· (t:..!\ ,tlE (tlE) E (6E) E· " '(. )

'kTJ' 1- kTexp kTlloT ; t:..=.2,22,

T= 25(25) 1000, T=15a(1O)390, HS;. :lr1• e,x:p(-:r:-1),:z:>exp (-X-I),. E\ (:v-I) , L% exp (-t-1)dt, z-I e~(Z'"I)El (:r:-1) , l-z-1 e:q:i (x-I)E1 Crl) ; x-.01 (.()OOl).l, 5-68.

[.5,38] V. .I. Pa.gurov:.a., Ta.bles of the eilpollenMa.l i.nte~.al

E.(:z;) =f- e-"""U-'du, TraneIatedfromthe Rus-

. 1

sian by D, G"Fry (Pergamon Press, New York, N. Y.; Oxford, London, England; Pads, France, 1961). E,,(x.), n=O(I)2.0, x=O(Ji1)2C1)10, 4- 9SjE2(x)-x In x,:c=O(.Ol)5, 'lB, E$(;z;)+iz1ln.x, :2;=0(.01).1, 7Bj e·E .• (x), n=2(l)lO, :z;=la(.1)20, 7D; eoE.(:c), 1'=0(.1)1, ~=,01(.Ol}7(.05)12(.1)20, 7.801' D.

[5;39J T.a.blitey integral'nQgosm1.i.sa. i kosiIi,Uila (IZda.t, Akaa N a.ukSBBR., Moscow, U .S.S,R.; 1954), Si(:i;)" Ci(:t} , :1:=0(.0001) 2{.Oo.l) 10 r.oosnco, 7D; Q,J(t,t) -in:1;, z=O(.OQ01) .01, 7D.

{SAO] Ta;hlit-sy l.nte;gral'no[ .pokaz&tePnol funktsfi (Izda.t, Aka._d, Nauk SSBa., Moscow, U,S.S,R., 1954). E,i(x), El'(X) , x=O(,()OOl) 1.3 (,OOl),3'{.oaQ,5) 10.(.1)15, 7D,

[~.411 ,D. K Trubey, /L table of three exponential intfr grels, Oa.kRio.ge Na.tiqnal L.abora.tory J=teport 2760,0&k Ridge, Te.nn. (June 1959). lilt{x) , E~(x), E3(x)., x==O(.OO05) .1(.001)2(.01) 10(,1)20, 6S.

6. Gamma Function and Related Functions

PHILIP J. DAYI"l

Con,tents

Page

Mathematical Properties. 255

6.1. Gamma Fun.ction. !:l55

6.2; Beta Function . ...258

6.3. Psi (Digamma) Funcsion. 258

6..4., Polygamma Functions. . 260

6.S. .Incomplete GatnIlla. Function. 260

6.(t. Ineomplete Beta Function . 263

Num.erical Methods. . . . . . . . 263

6.7~ Use and Extension, of the Tables. 263

6;8. Summation Of Rational Seriee by Means of Polygamma Funet_ions~~64

References. . . . . . , . . . . . . . . . . . . . . . .. . . 265

Table 6.1. Gamma, Digamma andTi-igamma Functions (1 ~-xS2) 267

f(x}, In rex) .u». i/!'(tr:), x:-1(.005)2, 10n

Table 6.2. Tetra,ga;mmaand Pentagamms, Functions (1::::; x:f: Z) 271

,p"CX), ifi!3l(x), X= l(.Ol)2, 10])

Table 6.3'. Gamma alld Digamma Functions for Integer and Hslf-

Integer Values (1':::;. n.':::; 101) . - . -. . . . . . . . . . . . . . .. 2'12

r(n), 118 1/1(11,), lOD

l(r(n), 98 ~11i(2n')tnn+t]e-", 8D

r(n+H, 88 In n-y..{n), 8D

n:-l(l)lQl

Table 6 .. 4·, Logarithms of the GammaFunction (1 ~ n':::; 101).

~274

logio rCn), 8S loglo r(n+i), 88

logIo r(n+i), 88 In r(n)-Cn-j) In.n+n, 8D

loglo r'n+~), 88

n-l(l)lOl

I Nat.lonal Bureau of Standards.

253

6. Gamma Function and Related Functions

Mathematical Propertfes

6.1. Gamm.a (Factorial) Funntion

Euler's Integra]

6.1.1 r(z)= I'" t,-lc1dt (EWz>O)

=],f So'" t~·,..le-k~dt (~'z>O,E,ilk>O)

Euler's Formula

6.1.2

. nln' .

r(z)=lim (+1) ( +) (e ~O,-1,-2, ... )

Il<-t"' Z .2 • . • . Z . n

Euler's Infin.i te Product

"(= lim .. [l+!+!+!+" ... +l_ln mJ

m~'" 2 3 '4 m.

=.57721 56649 ...

'Y isknownas Ewet's constant and is given to 25 decimal places in chapter 1. r(z) is single valued and analytic over the entireoemplez plane, save fgr the points z= -n(n= 0, 1, 2, ... ) where it possesses simple poles with residue (-1) ~!nl. Its reciprocal 1 If' (2) is an entire function possessing simple zeros at the points .2= -n{n -0, 1, 2, .. .).

Hankel '8 C9nt.ou': Inte~a1

6.1.4

1 i l' ( t)-~ -tdt

-'-'=-. - e "

T(z) 2'/r (:

y

I

I

, " .,

.

2 \ •••••• /

\

·2

n

'If

FIGURE 6.1. Gamma til/notion.'" ---,y=r(x), ,.. - - - ,y=ljr(x)

6.1.9 r(3/2)='ii=.88622·692M .. ~ =(i) I

6 I'( +-. 1). I·MI,13 ... (4n-3) r(1.)

.LI0 n" 4n .4.

r(1)=3.62560 99082 ...

The pa~h of 'inte;gration C statts at + cO on the

real axis, circles the origin inthecounterclockwiss 6.1.11 r(n+iJ 1+7·10 .~ ... (3n,-2) r(l)

direction and returns to t.hestal'ti1;lgpo'in't.

Factorial and n Notations r.(t)=2 .. 678.9385347 ...

6.1.5

II (z)= zl= r (e+ 1)

Inte:gl';t Values

6.1.6 r(n+l}=I·2·3 .... (n-l)n=nl 6.1.7

1 1

~ r (-.e) DC -n-l) I

(n=O, 1, 2, ... )

l"ractional ValQes

6.1.8

r (1) = 2 fO> e-Il! dt=1C'i= 1.7724538509 . .. = (-f)!

~o .

6.1.12 r(n+t) =1.3.5.7 . 2~ (2n-l)r{t,) 6.1.13 T(n+i)2.5.g.1l.g,; (311,-1) T(i) r(J)~1.35411 79394· ...

6.1.14 r(n+V=3.7.U.15 ~~. (4n-l) r{i)

r (-1) = 1.22541 67024 .

----

255

256

. Reeul'l'en.ce Form ulas

QAM:I\!A FUNCTION AND ~ELATED FUNCTIONS

'1r

6.1.15 r(2+1)=.2r(2)=z!=2(2-1)1 6.1.16

r(n+2) =(n-l+z) (n-Z+i) ... (1.+2)1'(1+2) = (n-l+z)!

=(n-l+z)(n~.2+~) ... (1+z)z!

Reftection Formula

6.1.17 r(z)r(1-;o:) = __'zr(~z) r(i)=1t esc 1r,'Z

Duplic.ation F(jrm~

TripJica~on Fonnultt _

6.1.19 .I' (3z)= (211") -I 33--t 1'(z)r (z+l)r(z;:t-t)

Gauss' Multiplication Fonnula

:Binomial CoefficieDt

6 ( Z)- ,z!, 1'(z+l)

.1.21w wI (z-w)! P(w+l)I'(z-w+l)

6.1,.22 (2)0=11

. _- .. - ... . ,.' 1''(~+n)

(z),,=z(z+1)(z+2) ... ,z+n.-1)=r(z)

Gamma Function in the C:(unpl~ Plane

6.1.23 1'(Z)=I'{z)j ill r(i).=hi 1'(2)

6.1.24 arg T.(,z+l)=arg I'(z)+arctan ~

Ir(x+iY)[·~· .... [. . _ y~ J-1

6.1.25 ,1'(3<) = ,l!.o 1 I (x+n)i

6.1.,26 jr(x+iy)I ~1I'(x)1

6.1.27

arg r(x+i,y)=W(x)+ ::E .... _ .: (,. _+Y-arctan-. +Y ).

n-o X nx· n

(x+iy¢0,-1,-2; ... )

where f(z)=I"(z)!T(.z)

6.1.28 r(l+iy)=iy r(iy)

6.1.29 r(~y) r( -iy) =l 1'{iy) I~ y sinh 'try

.1rY 6.1.31 r(1+iy)r(l-iy)=lr(1+~yW sinh 1l'Y

r(l_+' )1'(.3_') . 1r-J?

,4 'IIJJ :& ~y cosh 'IrY+1, Sinh 1l'Y

6.1.32

Po-we .. Series

6.1.33

In r{l+z)=-ln(l+.z)+z(l-").)

+:t (-I)'''[r(n)-l]z'''/n (lzl<2)

,,-2

t(n) is the Riemann Zeta Function (see ehapter 23).

Serie.5 E:tpanllion·~· for l/r(i)

1 '" (jzl < «J")
6.1.34 -c- -' -.' == ~. c,.zl:
r(z) !:~1
k tA:
1 1.00000 00000 0'00000
2 0.57721 56649 01$329
3 -0. 65587 80715 2fJ2538
4 -0.04200 2.6350 34095.2
5 O. 16653 813113 822915
6 -0- 04219 771H5' 555M3
7 -0.00962. 197'15 278770
8 0.00721 89432 466630
9 -0.00116 516'75 918591
10 ~.OOO21 5,2416 7411~9
11 0.00012 8(.)50'2 823882
12 ~O.OOO02 01348 547i!07
13 ---:-0.00000 12504 934821
14 0.00000 11330 272320
15 -0,00000 02056 338M 7
16 0.00000 00061 160950
17 O. 00000 00050 020075
18 -0.00000 00011 812746
19 O. 00,000 OO_OO_! 043427
20 O. 00000 00000 077.823
21 ~O .. 00000 00000 036968
22 O. 00000 00000 005100
23 -0.00000 000'00. 00"0206
24 -0. 00000 00000 000054
25 0.00000 00000 000014
26 0.00.000 00000 000001 . ~Tbe ooefficic.nts Ct are from H. T. baVi$, T:l!-bles of higher mathematical funetions,2 vols., Principia Press" Bloomington, Ind., 1933, 1935 (with permis.sion); with corrections due t:o B'_ E. Salz_61'L

257

Polynomial Appro~ati()DS s

GA¥MA IroNCTION;WDRELA.TED FUNCTIONS

Error Term. for~pb>tic ~o.n

6.1.35

1'(x+ I) =x!= 1 +alx+azZ2+aax3+a~~+asX6+ E(X) If(~) 1 ~5XHJ-~

aj= -. 57486 46 a~~ .9'51'2663 aa'= - .09985 88

a.= _ 42455 49 a5=-.10106 78

6.1.36 O::;;:t':Sl

r(x+ 1) =x,I=1+b1x+b2x2+ +baXIi+E;(X)

If(Z) 1 5;SXIO-7

b, = -. 57719 1652 b~= .98820 5891 ba- -. 89705 6937 b~= .918206857

b5'=-.75670 4078 08= . 48219 939'4 b1=-.19352 7818 Os= . 03586 8343

Sttdi:ng"s F orm:ula

6.1.37

t ,[ 1 1 139

1'(z) '""'e-'z'-- (21f) 1 +12:::;:+288Z~ 51840z3

571 ]

'248832Qz'+' ..

6.1.38

x!=~2\1r x"+!exp (- x+___!_) 12_x, ,

6.1042

If

RlI(z)= In r (:::;:)-(z-,) In z,+z-j. In (27)

then

IR ( )1 < . )B,z,..+2IK(z)

,.2 - (2n+l).(2n+2}lzl an.+l

where

For It real and positive, Iln is less in.absolutevalue than the first term neglected and has the same z!llgtl.

6.1.43

!?lIn I'(iy)=91ln PC-ttl)

=lln( :~)

y S "1I'Y

'Vi In (21r)-~-iln v,

6.1.44

(;ri>o, 0<8<1) Jln r(~Y)=lLtg r(iy)=-ar,g I'(-iy) =-.Fln p(-iv)

Aay:mptotic-FormuJas

6.1.39-

r(az+ b) ...... ..,I2; e-~(az)<1t:H-:i

6.1.40

In FEz) rv. (z-!) IDZ- z+* In (211-)

+ :E . .B'nn .. _

m;..l 2m(2m-l}zllml

Y 1 y.l .. ~. (-1)"-1 B~fi

....., ny--,II'- ~ (2n-l) (2n)y2fi:"'1

(y-+ =)

6.1.45 lim(21r) -ilI'($+~Y) 1~~Ii'llyl!-z-l

11'1-t1l>

1· b_aI'{n.+a) 1 <n~ on r(n+b)

For Bnsee chapter 23

6.1.41

In .I'(a) -(z~tJ IDZ-,'Z+t In- (21T)+_!_-~

. . 12$ 360z3

8 From C. Hastings,Jr., Approxima.tions for di_gi.ta] qom'pute~) Princeton Untv. Press, Princeton, N,J., '1965 (with permieaiOJl).

6.1.47

Z~-d P'(z+a) ...... 1+(a-b)Ca+b-1)

I'(Z+~) 2.2

1 (a-b).(', . ,.,)1

+ 12 . 2 . 3(a+b-1}3-a+b-lz2+'"

as z_oo. along any curve joining z . 0 and z= 0;0 ,

providing zre -q" -a-I, .... ; z¢ -5.1 b-l,

258

CQntinued Fracti.on-

GAM¥A FUNCTION ..uro RELATED FUNCTIONS

y

6.1.48

In r(z)+z-(z-i) In 2--!ln (21i')

au al a:iaa a~ as =21-2+ z+ z+ z+ z+ ...

1 1 53. 195

110= 12' a1=30' a2=Z10' as=S71'

22999 29944523 109535241009

a{- 22737J a6 19733-142' aiJ 48264275462

(f?lz > 0)

Wallis' Foemula!

2 i1r/2(sin)2n ._._ 1· 3· 5 ... (271.-1) - . z dx= - .-. ). --

1r 0 .. COB 2··-4· 6 ... (271.

(2ri) I 1 (2n) = 2'''{n!):t 22n n

r(n+t) rlr(1t+l)

'" _1_ ·[1 1 I 1

7rtnt. 811,·· 12S>n?

6.1.50

t: [. e: .. I_.e-!IJ dot

In r(z)= Jo . (z-l) e-I l-e I t (&i s > 0)

=(z-!) In z-z+tln 2'11'

+2 i<O ar:~~(~!z) elt:

(gl::J > 0)

6 •. 2. Beta. FUnction

6.2.1

11 So"· tz~l

B(e ··ui)· = ... t·-1 ·(I-t) u.-t dt= _ . - - . dt

, 0 .. - - 0 (l+t}s+1D

6.2.2

l~!'J

=.2 (~in t)iIJ'-l (cos t)-:lUl-1 dt



(elz> 0, £fw> 0)

B(. ) r (z)r(w) BC· a)

z,w r(z+w) . - 'W,s

6.S. Pst (Dig-auu:n.a) Function 6

6.3.1

Vi(z)=dlln r(z)]!dz= I"(z)/r(z.)

4 Some authors employ the spe(l~1 dOl,lble factorl~l notation as follows;

C2n) ! :! =2·4·6 .. _ (2n)=2"n· i

(2n~i) I [=1.3·5 ... (2n-l)= . .".-i 2n r(n+\)

6 Some .&)1tbo.l'S write.p(.z).= :'" ln r(z+ 1) and simllerly for thep:olygaIilrt:lafi.l:actjons.

S
4
II
Z
...--
1 /
/
-'5 -2 -I 0 'V2 3 4
-I /
-z I
-5
-4 -5

.FIGURE 6.2. Psi {l1/nj~#on. y=,p(:c) = dIn rf':c)/ib:

Integer 'Values

(n~2)

6.3.2

6.3.3

'f(t) = --y-2ln 2= - 1.96351 O{)260 21423 ...

6.3.4

¥i(it+1.}=-1-2ln2+2(1+!+." +_ .. _1_)

"!I" 3 2n-l

(n~lJ

~eCU1TenCe Fo~u1a~

6.S.5

6 •. 3.6

.. 1 .. 1 + ...

,p(n+e) (n-l)+z + (n-2)+z

1 1 . )

+2+z +11-2 +l,V(l+Z.

ReBeetionFonnula

GAlVIMA. FU.NCTIONAND RElL~TED FUNCTIQNS

259'

6.3.7 !Jt(l-z)-j-(z)+1r cot 'lrZ

Dqpijcati.on Formula

0.3.8 4-(2z) =~lfo(z)+i'" (z+!) +ln2

Psi: Function in the COlnplex Plane

6.3.9 .pCi) =¥(z)

6.'3 .. 10

f1l4-(iy)= fft,p( -iy)= &1..,..(1 +iy) = !Jl4-{I-iy)

6.3.11

yr1/;(iY)=flr1+! 11' eoth TrY

6.3.12

..P ~{f+iy)=t'lr tanh iII .;f,p{1+i'Y)=-2~ +} 'll'coth~' '_Y:E(n~+yZ)'-l

. n;=.l

6.S.13

Series Expmsi.oD8

...

6.3.14 "'(1 +z)= -1'+ 2::i (-1) "t(1i) ZIa-1 (:Izi <1)

. .. n=2

6.3.15

,p(1+z)=iz-l-vcot 712-(1-2.2) -1+1-'Y

..

- ~[r(2n+I)-11zZ"(l.z·I<2)

.. =~

6.3.16

·"'Z

,pO+Z)=-'Y+L: (+) (z~-1,-2,-3, .. ,)

91=1'11. '11. Z

6.3.17

fJloJ;(l +iy) =1-7-1 ~ll

'"

+~ (-1)~+11t(2n+ 1)-I]y2,i;,

n~l

(lyl<2)

'"

=-1'+y2 L: n-1(n2+yg)-1

n-l

(- 00 <s«. co)

6.3.18

';(z) ",In z-~- i> Bin

2z 11--1 2nz211

1 1 1. 1

=In z-2":-12z2+120z'-252z&+ ... ,

(Z~1xI in] a.rg .zl < 11")

111 =InY+12yZ+12Oy~+252116+ ....

(y4co )

Extrl:!mae of r(z) - Zeros of fex)

x,. r (.lf1,)
0 + 1. 462 +0. 886
1 -0. 504 -'a: .. 545
2 -1. 573 +2 .. 3.02
3 -2:. 611 -0. 888
4 -3. 635 +.0. 245
5 -4.653 -0. 053
6 -5. 667 +0. OO~
7 -6. (j7$ -0. 001 6.3.20

xo=I.4616321449 68362 r(Xa)= .88560 31944 10889

:1:,..= -1(, + (In '11.)-1.+ o[ (In n)~:l]

'Definite Integra~

6.3.21

6.3.22

\t'(z)+'Y= r= e~':e~~j dt= ~l 1-;:.!:t-1 dt

~o e ~~ ,

f'" (1 1)

1'= _' - -'-. . dt

o e'-1 'tel

l~·( 1 _t) d.t

= ---e-

o l+t t

~FromW. Sibaga.ki,Theory and a.ppliootionsQf the gamma. function,. Iwananrl Syote:n, Tokyo,. J a.pan, 1952 (with penniseloa).

260

6.4.1

_ _ d" .d"-fl

1/1(11)(2)--- 1/1(2)=-- _- .. In r(2)

. dzll d2"+~

*

=(_1)"+1 C"'tlle-Udt Jo l-e-'

"'("'{(:Un=O,l~ , , ,), is a single valued Dioolytic function over the rlltirecomplex plahesave at, the points Z= -m(m=O,l,2, ' . , ) where it pos~ Be5SeS polesof order (n+ 1),

InteR"" -Values

6.4.2

.f<nl(i)= (-l)Jt+lnlt(n+ 1)

(n= 1,2,3, . ,)

604.3

+(JJJI(n+l)=(-l)*m! [ -tCm+l)+l

1 - 1']

+2*+1+'" + n!"H

F ... actional Vailies

6.4.4

t/i(":) (:1) =(-1),*lnJ(2"+1-1)t(n+ 1)

(n=1,2, ... )

6.4.5

.. .-

1/1' (n+l}=!-r-4 '5: (2k-1)-2 t=1

Reeul"tt'nce Formula

6.4 •• 6

ReSection Formula

6.4.7

Mu] tiplication Form.ula

6.4.8 IjI<1i)(mz}=8lnm+-. 11 5l.1 J/lt"'·C2+-._k_ .. ·)

m"+ t=il VI.

.-

8=1, n==O 0=0, n>O

7 -i' is known aa the trignmma(UBction. {I", ",,(~J. ~IH nre the tetrn-, penta-, nnd hexagamma functiona respectivclv. Somell.utli9TS write ,,(z) =d no r (z + OVdl, nnd similarly lor the polygamma fanetlons.

·see page u.

604·.9

~1"l(l+z);::::;(-l)"+l [n!rcn+n

(11.+ 1).! .. (11.+2)1 .... . t ]

- 11 f(n+2)z+ 2!· t(>n+3)z - ... _

t~zl<l)

,

6.4.10

'"

1/I1"l(2)=( -l)!I+ln!~ (z+k)-!l-I

_. ~·o

(2 ~O,-1,-2, ... )

Asymptotic Formulas

6."'.ll

.p(r';(z)_(_1)1I~1 [.(n.-l)'+ .'n! .

e" 2z .. +1

oj (2h+n-l) 'J. . . - .. - ..

+~ Bit. (2k)!z2.l:+11 (2_= 10 1 arg 2 1 <"')

6.'.12

, 11 (111

.p (,2)-"2+221+6z3-30z5+ 42z7-30z~+'"

(z- w in 1 :8tg a] <"')

6.4.13

II 1111135

'!_ (z) --zl-z-a-2-z4+ij?- 6Z9+1021Q-6212+' .•

(z- ee in 1 args 1<11")

6.4.14

(8)-). 2- 3 -2 1+ 4 3 +10

.p . (z "" za+ 2~+ zs- Z7 3zQ-zrr .zu-

(z~ = in la.rg 21 <;,r)

6.5. Incomplete Gamll);a Fu.nCtion lsee also 26.4.)

6.5.1

1 1:r·

P(a z) = -. _ ... -. e-'trJ-1 dt

, rea) 0

6.5.2

,,((a, z) =P(a, e) rCa) = !j"e-1t--1 dt

6.5.3

f(a., x) =r(a1-'Yca,x)"",i"e-llll-ldt

6.5.4

_ x-a...

"(*{a, "1")= x-ap(a, J) =r(~) "(a, ::r)

-y* is 11- sing-Ie. valued analytic function of a and x possessing .ne finite singularities.

GAMMA FUNCTION AND RELATED FtlNCTIONS

a

261

5r---~--~--------~--------~-------4---------r--------~------~--------'



o

,* O(Y<O.2$

• )', .

FIGURE 6.3. Incomplete gamma junction. -y+(a,x) = ':1:(-'0,), rz /!;-t~..--ld~

, ra Jo "

From F.G.,fiicomi, Bulla funzione gamma incompleta, Anna.li di M;a.tematka. IV, 3R, 1950 (with permlssion).

262 6.S.5

6.5.6

G:AM:.M;A FUNe'l'10N AWn RELATED FTINCTIOlS"S

'Y Ci.;Z2)=2 i:J;e.~j~ dt="j; arfa.:

Probability Integral of the x!~Distri.butioD

(pear8(in'.$ FonD of the In.c.omplete Gamma Function)

1 i .. -./P+l

I( , .. ) = . . . ~lt'P dt

. 'U,P. PCp+l) '. 0 e

=P(p+l, u4p+l)

6.5.7

6.5.8

6.5.9

6.5.10

s.s.n

E;.. (x)= .. r "'~-~~t~"'dt-:t>l~lr(l-n,x) ,. Jl

Incomplete Gamma Fun.ction ass COnfluent l,IyPe:rs~in,e1;ri9 Funetio~ (8e~~hapter IS)

6,.5,12 "Y(ajx)=a-txlle-zM(I,l+a,x)

=a--:1zlZ M(a; 1+a,-.x)

6.5.13

p (nJX}=~-(l+X+;;+ ... +(:~-:)!) e-~

=l-e;,,_l (x) e"'';

For relation to the. Poisson distribution, see 26.4.

6.5.14

6.5.15

y"'{ -n:, :n)=x'"

6.5.16

6.5.17

6.5.18

. (-1)" [... . _",11-1 (-l)ij!]

6.5.19 r (-n,x) =~ E1(x)-e .~ xI+1 .

6.5.20 r(a,i;z}=e,r~<I [O(x;a}-iS(x,a)]

ReeureenceFoemulae

6~5.21

6.5.22

6.5.23

Derivativei a.ndDUf~nt:i,aI Equations

6.5.25

01 (a,~) ClZ

6.5.26

:~" [x-"r Ca,X)}= ( -1) "x-a-lIr(a+.n,x)

(n=o, 1, 2, .... )

6.5.27

6.5.28

$edee Developments

6.5.29

.. _ z ..

'Y""(a, z}=r~E. rca+n+l)'

1!~O ...

1 '" (_z)1I

r (1,) ~ (a+n)nl

(lzl<Ql)

GAMMA F1JNCTWN A..""IfD RELATED FW'fC'l'IONS

6.5.30

'Y(a, x+Y)-;i(a, x)

=e-~z~-l i:. (a-l)(a.'-~n' . , (il-ni[l_e-lIen(y)]

11=0

(lyl<lxD

Continued Fraction

6.5.31

. . .(. 1 l~a I 2-a 2 .)

r(a,x)=c'X~x+ n_: -x-f-1+i!:+" ,

(x>O,lal< CQ.)

Asymptotic F:xpaneions

6.5.32

r'(a,z)",",zrt-1e-'[I+a z 1 +(a-l~~a-2)+, : oJ

(z- 00 iIJ.larg zl<;1F)

SUPPO$(' R" (a,z) ='!lU+1 (a,2) + ' .. . is the remniudernfter -n terms in this series, Then if a.,2 [l,t'{l reu:l, we have for n>a.-2

awl sign R."(a,z)",-sign Untl(a,z) .. .'" (-1) "-",4t" 'Y(a,z)-".~ (a+n)n!

6.5.33

6.5.34

6 .. 5.3:5

263

Definite lritegrab

6.5.36

f'" e-a'r(b,ctJdt=qb)[.l-._______£_.b bJ_ k .' a. ~+~

(91(a+c»O,f]tb>-l)

6.5.37

So'" ta-1P(b.,t) dt r(a:Q)

6.6~ Incomplete Beta Function

B",-(ajb) = J:"ta-I(l-t)~-ld-t.

6.6.2 Ix(a,b)=B:;(a,b)/B(a,b)

6.6.1

For statlstical applications, see 26.5.

Relation to Binomial Expansjon

6.6.4 J,,(a,n-a+l) = tt (j) Zl(l-p)~-i

For binomial distribution, see 26.1.

Recurrenc'c Formulas

6;6.5 Ii(a,b)=x]$(a.-l,b)+(l-x)J,.(a,b-l)

6.6.6 (a+b-aa:}l';(q;,b) •

=a(1-x)I2:(a+ 1,b-l)+bI~(a,b + 1)

6.6.7 (a+b)I",(a,6).=.alx(a+ 1,b) +bl,,(a,b+l}

Relation to Hyptlrgepm.etricFonction

(z_" <Xl. in I Rtg ;:::1<!1r) 6.6.8 B%(a,b) =a-1x<lF(a,J -b; a+ 1 t x)

Num.erical 'Methoc:l_s -

6.J. Use and Extension of the Tables

Example 1. Compute f(6.38) to 8.s. Using tlie rocurrcnee rolariou 6.1.16 owl Table 6.1 We ha'li'c',

r(6.3.8)= [(5 .. 38) (4:.:38) (:3.38}(2.3B) (l,38)]r(1.38)

=2;J2:43671.

Example 2. COiTIIH)tC' In r(56.38), using Table 6.4 nnd linear interpolation in /2< "-e have

1n r (56,38)= (56.38-i) In (56,38)- (56,38)

. . +j2 (56.38)

The error of linear interpolation in the table: of the hu\ction]2 is smaller than 10-7 in this region. HC'llCQ .•• j~(56.38) = .. 92041 67 and In r(5B.3S) = 169.85497 42.

Direct intnrpolation in Table 6.4 of loglO fen) eliminates the nec('~sity of employing logarithms. Howovor, tilt' CITe}!' of linear interpolation is .002 SQ that logw f(n) is obtained with 0. relative 01'1'01' of lO~!,

264

GAMMA ,FUNCTION :AND RELA'l'ED FUNCTIONS

'Example 3. Computetf(6.38) to 8S. Using the recurrence relation'6;8.6 and Table 6.1.

, 1 1, 1 ,1 1 "

¢(6.3S)5.~8+ 4.38+3.38+ 2.3S+1.3S+f(1.38)

= 1.7727"5 59.

Example 4. Compu.te 1jI(a6.aS) . Using Table 6.3 we have,,.!t(56.,38) =In 56.38-ja(56.38).

The error of linear interpolatien in the tl;l.ble of the function 13 iasmaller than ax 10""7 in this region. Henc6,.Ja (56 ,38) == ,Q0889 53 and,p(56 .38) = 4;02-3219.

Example 5" Compute In J'(l-i}. From the reB,ection prm,ciple 6.1.23 and Table 6.,7,

In r(l-i)-ln r(l+i) =-.6509+.3Q16i.

Example 6. Compute ill r(t+f1.,). Taking the logarithmof the recurrence relation 6.1.15 we have,

In I'C!+ti)=ln r(!+ti)-ln Ci+ii) =-.2~419+.03467i

- (t In i+i arctan 1) =.11239-.75013i

The logarithms of complex numbers are found from 4.1..2.

Ex,ample 7. Compute In r(3+7ijusingthe duplication formula. 6.1.18. Ta.king the logarithm of 6.1.18, we have

~t ID21r=~ .91894

(t+'7i) In 2= 1,73287+ 4.85203i ill r (!+ii)=-3.31598+2.32653i In r(2+ii)=-.2.660,47+2.93869i

In r(3+7i)=-5.16252+1Q.11625i

E;xanl:ple 8. Compute In r(3+7i) to 5D using the asymptotic formula 6~1.41.. We, have

Then,

(2.5+7?;) lti(3+7i) ~-3. 0857779+ 17.1263119i - (3+7~) =-3.0000000- 7.0000000i

-'fIn (2~)= .. 9189385

[12(3+7i)]-l= .0043103-.- .OlQ0575i'

-[360(3+7i)3]-1= .0000059- .OOOOCi'22i

.ln r{3+ 7i) = -5. 16252 + 10. 11625i

6.8. Summation of Ratlenal Series by Mean.s ofPolygamma Functiene

An infinite series whose general term is it rational function .of the index may always be reduced to 'a finite: series Qf psi and polygamma functions. The method will he illustrated by writing the explicit formula when the denominator centeins a triple root'.

Let the ,gene.ra! term. of an infinite. series have theform

pen).

where

where pen) 1:S8. polynomial of degree m+2r+3s-·2 at most and where the, constants a'l 13.~, 'and 'Yfare distinct, Expand Un in partialfraetions as follows

'"

Then; we ,may express :E <Un in terms of 'the

'1'10--1

eenstantsappearing in this partial fraction expan-

sion as follows .

Higher order-repetitions' in the denominator are handled similarly. U the denominator contains

265

GAMMA FmlCTION ,AND RELAT;HlD Fm;rCTIQNS

only simple ordouble roots) omit, the, corresponding lines.

Rxatnple9. Find

... 1

8=~, ' , "

n-I (n+ 1)(2n +1){4n +1)

Since

,Ii , 1 , " I (n+l) (2n+l)"(4n+l)n+l-n+1 + n+t-'

we have

ThUs,

8= ~ !.f (2)+ !fCH) - H{l i) = .047198.

Example io,

Find 8' '~-nC2(8:+ l)~'t'

S. 1 16, 16 1 1

moe 1i2(8n+l}2-n + 71+1+ n2+ (n+i}2'

we have,

Therefore

8= 16lji(1)-16,,(li) +.r'(l) +1/'(11) =-.01349_9"

E:J.:~ple 11.

II> 1

Evaluate 8=~ (11.11+ 1) ('ilf+4) (see also 6.S.la).

! (n~i-nl D

'i (1 1)

-12 n+2i-n-2i .

i-i -i, i

Hence, a;=-, a,,=--,I'7_=-,'a;,=:---,

-. 6 -~ 6 ~_. 12 • 12

and therefore

$= 6£ (1P(I+i) -4-(I-i)1 + l~ [yi(1+2i)-1f(1-~i)l.

By 6.3.9J this reduces to

8=~ ,..1' !f(l +i) -i ..Iy,.{1 +2i).

From. Tabl_e 6~8j 8=.13876.

References

Texts

[6_1] E. Artin; Eihfqb,tun:g 'in die Theorie der Gamm&funkti011 (Leipzi~1 Germ..!UlY, 1931).

[6.2J P. Eo Bohmer, Differenzenglelchungen und be~ , etim.mte Integrale, chao 3,4_, 5 (K. F.Koehler, Leipsig, Germany, 1939).

[6.3] 0;. Doetsch, Handbuch del' La.place-Traneformlltion, vol. II, pp, 52-61 (Btrkhaneer; Basel, SWitzerland, 1955).

[6Al A. ErMlyi et -al., Higher transcendental {unctions; vol. I, eh, 1, eh, 2, sec. 5; vol .. 2., eh. 9 (.MoGra.w~ Hill Book oe, Ine., New York, N.Y., 1953).

[6.5] C. Hastings, Jr., Approx:irila;tions for digita:-l computers (Princeton Univ.Press,Priilce-ton, N.J.,

19~. .

[6JjlfF. L().Seh aI).d P_ $_chobJjk;j Dier FaJrult!i.t und verwaJidteF~tionen (B. G~ Teubner, Leipzig, Ge.NlJ.:any, 1951).

[6.7] W. S.i,bagalti, Tb,eory'iihdapplh}ations (if the g_a.mma function (Iwanami Syoten~, Tokyo, Japan, 1952).

[6.81 E. T. Whittaker and G. N. Watso.n, A course of modern analYBiI!, ch. i2, 4th ed. (Ol'im!:lridg{j Univ. Press, Cembridge, England, 19_52).

T.B'bles

[6.9] A. Abramoy, Tables of 10 ref) f()r complex argu-ment. Trailslated trom the Russian by D. G. Fry- (Pergamon Pte.B5, New York, N.Y., 1960).

In r(z+iy), z=0(,(1l,)10, I/=O(.Ol)4, en, .

[6.10-1 Ballistic Research Laboratory I A table of the factorial numbers and their reciprocals frotn 11 througb 10001 to 20 signi.ficant,digite. Technical Note No. 381, Aberdeen ProviD,g Gr,ound, Md., 1951.

[.6.11) Britdsh Association for the Advanc_ement of BCi:enctl:l Ma.thema.ticaitable.s, voL 1, 3d ed.,pp. 4o-M (Cambridge Univ, Presa, (ja.mbridg~, England, 1951) . The ga.m_ril,t;.. and polygamma functions.

Also 1+ ~Clogla (t)ldt; 2:_=O(.Ol)l,1OD.

[6.12] H. T: Davis, Tables of the higher mathema.tical functions; 2 vola. (Principia Press, Bleomington, 'Ind., 19'33, 1935). Extensive, mM)' place table!i of tbe gamma andpolyga.mma. fUDctii}n$ up to .pc.) ex) and 9:f :t!)J.eit logilJithmll_

[6.13] F. J.Du~e, Nonvellestables de logili n! &33 dec'i,.. males depuisn=l jusqu'a n=3000 (Kundig, Gelleva,BWit:z~land; Index Generalis.; Paris, France, 1927). .

266

GAMMA FUNCTION AND RE.LATED litr.NOTIOl'l"S

[6.14} N ational Bureau of Standards, Tables of 71,1 and r(n+l) for the first thousand values of 11, Applied Ma:tb. Series 16 (U.S. Government Printing Office, Wa.shingtQD., D.O., 1951).nl, 16Sjr(n+!), 88.

[6.15J National Bureau of Standa.rds, Table of Coulomb wave functiops, vol. I, pp. 114-135" Applied Math. .Series 17 (U.S. Government PriD.ting Office, Was'):llngton, D.C., 1952).

9t'lr'(l +i!l)/r(l + i"],!I=0(.005),2 (.01)6 (.02)10(.1) 20 (.2) 60 (.5}110, lOD; arg reI +"1]) .'11=0(.411)1 (.02} 3 (.05) 10(.2)20(.4)30(.5)85, 8D.

National Bureau of Sta.ndards, Table of the ga.mmer fUIl,CtiOIl for complex a.rgwnents:, Applied ¥ath.

Series· 34 (U.S. Government Printing Offi~e, Washington, D.C., i954).

[6.16]

In r(x+i1f}, x=O(.l}lO, y=O(.l)lO, i2D.

Contains an extensive bibliography.

[6.17] National Physical Laboratory, Tables of Weber parabolic cylinder functions, pp. 226-;233 (:Her Majesty's Stationery Office, London, E;ngland, 1955).

Real and im.aginary partsof la TCilii+ fia) 1 Te,-=; 0 (1) 3, a= 0(. 1)5(,2)20, 8D; (Ir(l+!ia)/rti+tiQ)I)-IlJ

a ... 0(.02) 1(.1)5(.2)20, 8D,

[6.18) E. S. Pearson, Table of the logarithms of the complete r-functjon, arguments 2 to 1200, Tracts for Computers No. VIII (Cambridge Univ. Press, Cambridge, England, 1922). Log1o,r(p), p=2(.1) 5(,2)70(1) 1200, IOD.

,[6.19] J. Peters, Ten-place logarithm tables, vol. I, Ap-

. pendix,' pp. 58-68 (Frederick Ungar Publ, Co., New York, N.Y., 1957). nl, n=l'(l)60i exact; (nl)"-1, n=1(1)43,54D; Log1q(nl). n=l(l) 1200, l8D.

[6.20] J. P. S"E&rueY8J1d M. V. Wilkes, Table of the reciprocal oftba gamma function for complex argnment (Univ. of Toronto Pless, Tort;ltltoJ, Ca.na.da., 1950). z=-.5(.Ol).5, y=O(..Ol)1, 6D.

[6.21] M. ZYczkoV'l"sk.i, Tablice f':lnkcyj eulera i.pokrewnych (Panstwowe Wyda.wnictwo Naukowe, Warsa.w, Poland, 1954) . Extensive tables of integrals involving gamma and bet!l- funetions.

For references to tabl.\la.r material on the incomplete ga.JEma. and ineornplete beta functiona, see the l'efetences in chapter 26.

7. Error Function and Fresnel Integrals

WA.LTER GAUTSCm 1:

Cpntents

Table 7.1. Error Function and its Derivative (0 _:sx':::;2) (2/:J;)e-;.z2,-erf x=.c2/~) il! e~I'Jdt,. x=0(.01)2, IOD Table 7.2. Derivative of the Etr<Jr Function. (2~x~lO) (2/,j;)e-ff, .z=2(.Ol}10, 88

Table 7 .• 3. Complementary Error Function (2.~x~cQ).

xe<l:2erfc. X=·(2/...j;)xei;Zf. ... e-~·zdt, z-2=.25(- .. 005)O, 7D-

.. ~.

Page 297

297 299 300 302

304 304

308

3_l0

Math~mat;ical Propentdes . . . . . . . . .

7.1. Error Function .

7.2. Repeated In tegrals of the Error Function 7.S. Fresnel Intsgrsls . ,.. . , ,

7.4. Definite and Indefinite Integrals .

NumerleaIMethods .

7.5. lise and Extension of the Tables .

References . . . . ., ", .. , ..

312

316

er£c~, 11.=1(1)10, 15D

Table '7 .4. Repeated Integrals of the Ettor Function (0.:s;x:S5)

21'r (~+l)ill erfc x=2"+1r (;+1)/,;10> (tn1z)JI e-1idt :1:=0(.1)51 n=l(1)_6, 10, 11, 68

Table 7;5. Dawson's Integral (O.:s;x.:s; co).

317

319

rz, .

xe-z-2 J, e,·'ldt, .;v-i--:-.25(-.005)O,9D

I Guest worken, Na.tiona.1 Bureau of S.ta,ndll.r~,fr(lmThe American University. (Presently PurdueUniver.sit:y.)

295

7. Error Function and Fresnel Integrals

Mathematical Properties

7.1. Erl'or FUnction

De6rdtions

1.1.2

f 2 f." j2d rf'

er c Z= r: ' e - 't= l-e, Z

-y'1r . "~

7.1.3 w{Z')=e-i\1+~i'e'2dt)=e-'2'erfe (-iz)

In7.1~2 the path of integration is subject to the restriction a.rg i-a with Idl<'~ as t-4oo along the path. (a=~ is permissible ifElt'remainB bounded to the Ieft.)

7'.1.4.

( ),_i f' '" e-12d.t -, 2iz ,i'" e,-12dt

w z -- -. --.- _- -2--':;:

, 'lr _'" z-t '/f' '0 z-t

(.Jfz> 0)

.7

.s

,6,

.3

.2'

.1

o

-

.• :6, .EI r.o 1'.2· I:~ 1.& La' 2.0 ~,:? 2.~

FIGlJ"lI,E 7'.l.y=e1;·L"'e-'·cU. p ... 2 (1) 6

Senes EXPllD,SlOD8

2 '" (_1)'112211+1 1.1.5 erfz- ..;;.~, . ,n! (2n+ 1)

7.1.6

2 ~ '" 2"

=.0.-.:. e'-'~ Xi, " ' , ,: 'Z~lI-1;1

..;; fI-() 1·3 .. ' . (2n+l)

7.1.7

7.1.8

." (i'z)"

w(z) ,:E '. (', , )

"'-0 r~+l ,

For ]",-t(Z}, see chapter 10.

Symmetry RelatioDs

7.1.9

erf (-z)=-erf a

1.1.10

7.1.U

w(-.e)=2e-'~ -We?) w(z)=w( 'z)

7.1.12

.1

.6

.5

297

298

ERROR FUNCTION AND FRESNEL INTEGRALS

y

FIGUBE 7.3. AUitude Ohart of w{z}.

7.1.13

fuequallti~ (7.1lJ, [7.17]

(For other inequalities see [7.2].)

CoiltinQed F'1:actioDII

7.1.14

2e.2 .J.'" e-12dt-__!_ 1/2 _1_ 3/2 ~ ... (-Blz>O)

. ~ z+ z+ z+ z+ z+

(..I"z#O)

a:~~) and H~1i) are. the Zeros and weight factors of the Hermite polynomials. For numerical values see cbapter 25.

Value at Infinity 7.1.16 erf s-s-l (z~.oo in /a.rg z/<i)

Marlmmn and ln8ection Points for Dawson's Integral [7.31 J

F(z)'=e-~ J:~ el2dt

7.1.17 F(.92413 88730 )=.54104 42246 .

7.1.18 F(1.50197 52682 )=.42768 66160 .

Deriva uves

7.1.19

d"+l " 2 ._ ..

dZ"+l erf z=(-~) Ii H .. (z)e

7.1.20

W(II+%} (a) +281(1(11+1) (z).+2(n;+ l)w(ft)(z) =0

(n=O, 1,2, ... )

(11-=0,1,2, ... )

2i w'(z)=-2zw(z)+ .fi

(For the Hermite polynomials H,.(z) see chapter 22.)

WCOl (a )=w{z) ,

Relation to Confluent Hypergeometric Function (see chapter 13)

7.1.21

erfz=2Z-M(~'~' _z.a)=2~ e-" M(l'~2 32)

. ;,fi 2 2 ",.otr

The Noniial Distribution Function With Mean m· and S.ta:il.d~d Deviation IT (see chapter 26)

. (t-TJl)1

1 f;ll --. 1 ( (a:-m»

--= e 26·~ dt=2" 1 -l-erf 1I.,;2

1I.,j27r·· - ...

7.1.22

Asym,ptoticExpaD8ion

7.1.23 .

s -0) 1·3 ... (2m-l)

~ze" erlc z ...... 1+ L-l (-1 m (2Z2)'"

m~l .

(2~CD, /arg z/< :)

ERROR FUNCTION AND FREB~EL LNTEGRALS

299

If R,,(z) isthe remainder after n terms then

7.1.24

R (.)=C-l.)iI i.' 3 ... (2n-1) 8 . " z . (2;::2)" ,

(OJ ( t)-n-t

8=)0 e-I 1+ Z2 dt

(Iarg ~I<i) (Iarg zl<~)

For x real, R,,(:c) is less in absolute value than the first neglected term Mid of the same sign.

Rational Approrlm.ati0l18 ~ (0 ~z< ... )

7.1.25

erf x=1- (al.t+~t~+ast8) e~T.J +E(X) /

IE (:c) 1 ~2.5XI0-6

1 t=-· - l+px

p=.47047 a1=.34802 42

aJ=r- .09587 98 aa=.74785 56

7.1.26

erf x=l- (art +a2tJ+ast3+a,t'+a~t!) e-irl +E(:I:) , 1

t=l+p:t(

IE{X)j Sl.5XlO-7

p=.32759 U aJl=.25482 959·2

a2= - .28449 6736 aa= 1.42141 3'741

a4=-1.45315 2027 £h=1.06140 5429

7.1.27

~=.278393 aJs=.000972

a2=.230389 a,=.078108

7.1.28

erf x=l

1

[1+alx+~:z;2+ ... +aJexapft+E(x) If(i) 1 S3XlO-7

a,.=.0705230784 a3=.009.z·705272 ct~= .00027 656'72

~= .O~28 20123 a,= .OODU) 20143 aJ8= .00004 30638

. 2 Approxima.tiOns 7 .1.2S-7.1.~ are from C. H 8.!itings, Jr., Approximations for digital computers'. Prineetot; Univ, Press, Prtneeton, N. J., 1955 (with pel'Iilission).

Infi;nite S.eries A'pp~xbp.ation for Complex El'l'Or FtmCti()D [7.19]

7.1.29

-~

erf (x+iy)=erf x+;1J"X [(I-cos 2xy)+i sin 2XyJ

where

f,,(x,y)=2x-2x cosh ny GOS 2xy+nsinhny sin 2xy g .. (x,y)=.2x c08hny sin 2xy+n si.n.hny cos 21,y

. IE(x,y)1 :=:010-1°1 erf (x+iy)'1

7.2. Repeated Integrals of the Error Function

Definition

7.2.1

(11,=0,1,2, ... )

. I f 2.t·o . IT erf

1-er 'C ,Z :.;; 6.-, le . e Z= c z

-Differential EquauoD

d2Y+2z dy -2n =,0

dz2 dz Y

y=AiI'l erfc z+Bill eric (-2) (A and B are constanta.)

"1.2.2

EEpreasil;lD ,!S a Single Integra]

7.2.3

'" err . 2 J'" (t-z)" -12dt

1 C Z c .1' e

",,11" z. 11,

Power Serieli a

7.2.4

Recurren.ee Relations

7.2.5

i" erfc z=-~ in-1 erfc z+21 jlld erfc z

n 11,

(11,=1,2,3, ... )

7.2.6

2(11,+1) (n+2)i,,+2 erfc z

= (2n+ 1 +2Z2)i"erfc z-~ in-2 eric z ('11.=1,2,3, ... )

• The terms in this series corresponding to k=n+ 2; n+4, 11+6, ... a,re understood to be zero.

300

ERROR FUNCTION A]jD FRESNEL INTEGRALS

7.2.7

in eric 0 1

s-r (~+l)

(11,= -I, 0, 1,2, .. .}

1.0

FIOUlll'J, '7>4. -Repea.ted ln~gra~ of the Error Functwn. p-2'r(pl) i'efta:

71";0,1, 2; ~8. U,22

Derivatives

d"'rf' "'In

7.2.Q """'-'-lne, C Z~-l!l- er o z

dz. '

(n=O, 1, 2, ' ... )

7.2.9

7.2.10

Relation to Hh,,(iIS) (Me 19,;14) i"eric 2=(2" \r)! Hh",(.flz)

Relation, to Hennite polyD,om.ials (see, chapter 22)

"'-:PI

7.2.11 (-1)"i" eric z+irt eric (- Z)=2:~'-nIH1t(iz)

Relation 'to the Confluent :Bypergeom.eiric 'Function (see chapter 13)

7.2.12

'ii 'erf '_ -za '[ , 1" M"(' n+l 1, 2) iverre z-6 2"r(~+I}' ~2-' 2"' e ,

z '(12. 3 2)']

-, '(n+. 1)- M 2+1'2' z , 2,,-lr _"_,_'

2

Relation to Parabolic Cylindc'I' Functions (ace ch-llpter 19)

7.2.13

7.2.14

Asymptotic Expanillion

',,- ,,2 e-z2 '" (-l}'II(2m +,n.)I

1 erfcs ...... _'- (2),.+1 ~, " 1(2 )2'm

-y:ll" ,.2' ", .. 0 n.m, Z '

(Z~OO; lafg zl <3411") 7.,3. Fresnel Integrals

De6nitiQn .

7.3.1

The following functions are also in use: 7.3.3

OI'(Z>=.../; i~ costZ4t, 02.(::c)=.v~"lri~C:7 teU

7.3.4

r . 1 fi'Sin,t

81 (x) = sin t2-d,t, S2(X) . I<L. r; dt

'., . 0 -v211" 0 -vt

Auxiliary Function!!

7.3.5

7.3.6

g(z)=[~-O(z) ] 60S (~. Z2)+[~-S(z)]sin (~. Z2)

Interrelations

7.3.7

ERROR FUl'i!'OTION AND FR.ESNEL INTEGRALS

7.3.8

7.3.9 O(z}=~+f(z) sin (~Z2)-g(e) COB (~ S2)

Series E;s:panriolis

7.,3.11

7.3.12

O( ) -" (:! 2)";'" (_l)n~fI,{"+l

2, -cos ,22 ~ 1 . 3 ... (4n+ 1) a '

('If ) co (-1 ~,n1r2n+ I

• ,2, ..j; , , 411+3

+sm "22, ,~1. 3 .'. ,.(4n+3) 2 ,

1.,3.13

,,"' (-1)"('lI"/2)~"H

8('z)-" ' S4"+3

, ~~ (2~+1)!(4n+3)

FWURE 7.q. Fresnel Integrals., .. ~ct..),v"'8Ci)

301

7.3.14

, C'1r' ) '" (_1)711r2rttl "

B{e)=-cos2s2 ~ }.3. : .' (471+3) .zgn+~

+ . (!. ,2) ~ (-1)"1%'2'1 , 4nH

S.1U 2 z· ,£;61.:3 ... (411,+1) z,

7.3.15 0:.(Z)=JI/2.(Z)+JiI2(Z)+Jg/2(z)+ •. ,

7 •. 3.16 S2Xe)=JSI2(Z)+J7/2{z)+Jllli(Z)+",

For Bessel functions Jf>+1I2(Z) see chapter 10.

Sym:roeb'y Re}.ations

7.3.17 O(--z)--O{z), S(-z}=-S(z) 7.3.18/ 9Ciz)=iC(i), S(iz)=-iS(z)

7.3.19 a(i)=o(z)~ 8(z)=8(z)

Vahle atlnJinity

.' 1 O(:t)~,

7.3.20

Dedvatives

dj(i) , C":)

'dx =-1rxg X ,

dgJx) ='/'(x}(x)-1

itx '

7.3.21

Relation to ErxorFunction (see 7.,1.1, 7.1.3)

7.3.22

O(g) +i8(a) =lti ed [~ (l-'i)g]

=lti {l_e~iziw [f {l+i)Z]}

7.3.23 g(x)=£?{lfi w [f (l+i)X]}

.;7.3.24 j (x)-:,Y {lti w [~ (l+i}Z]}

Relation to COnfllltlut Bype~geometric FIlDction (see «;h.liP~l' 13)

7.3.25

""") '8'()' M' ,,(13 .71'2) v{;!l +'1., z =~' .2' 2"' '1. "2 z

Relation, to Spherical Be8seIF~ction.s (see chapter 19) 7.3,.26 O2(3) =~lS'J _t(t)dt, 8a(z) =~J:·JI(t) dt

302

Asymptoti.c Eqlall!!ione

ERROR FUNCTION AND FRESNEL INTEGRAL,5

7.3.27 ·f() .l·.+~. '" (_1')1>11.3 ... (4m-l)

lI'ZZ '" L..J..... (. Z).2m

m-l .'1!'2.

(z--t 00, I·a.rgzl< ~)

7,.3.28

'. <D .",1 ·3 ... (4.m+1)

'ltzg(z) ....... 2:. (-1) . ( 2)Zm+l

m-[): '/1'2

(z~oo., larg 21< i)

If R;'!) (z)) R~&l (z)are the remainders afte:r'!l- terms in 7.3.27, 7.3.28,. respectively, then

7.3.29

R(J)( )=(_l)~l. 3 ... (4n.-l)fj(f)

" a r==. . (1I'Z 2).2 11' •

1 f"" e-lt2.,,~! . (. . . .11")

(J(II r(2'll!+~) Jo' l+(:::ydt larg zl<4

7.3.30

R('){~\=·'(~I)",l . 3 ... (4_n+l) a(f)

U' \~ . (·Z)·llii V",

. 11'2

,fI' r (~+il rl:(~),'tO·rg zl<i)

'1I'e ..

7.3.S1 181111<1,18(1'1<1 (Iarg z!:Si) For x real, R;'!l (x) and R~f)(X) are less in absolute value than the first neglected term and of the same sign,

Ra.tioD.a.I ApprQIimatioDEl4 (0:::::;.];$00)

7.3.32

f· ( ) 1+.926x + ( .) x, 2+ 1.792 x+3.1 043fl . e X

7.3.33

. . 1

9(:11) 2+4. 14.2x+ 3A92x2 +6.67 Or.+ Ef{x)

IE(x) 1:5 2X 10-3

(For more accurate approximations see [7.1].) 7.4. Definite and Indefinite Integrals

For amore extensive list of integrajssee [7.51, [7 81, [7.15]..

7.4.1

tApproximations 7 .. 3.32, 7.S.33 are based on those given in (L Hastings'. Jr., Approximations for calculating Fresnel integrals,Approximation Newsletter, Apdl'1958; Note. 10. ({)iee also MTAQ.10, 173, U56.)

7.4.2

iUl .... . llaw: ,oL(!C . b

e- (at2+2bt+e)dt =~ .. - eo.· eric' r:

D 2 a . ~a

(Bla>.9)

7.4.3

1· 3 ... (2n-l) Fi.

2"+ La" "Va

r(n+l) 2a"+i

7.4.5

7;4.6

1.4 •. 7

(fJla> 0)

7.4.8

7.4.9

f" e-'Hdt Jo ..[i(t+z)

7.4.10

:z ea~ eric .Jii;Z.

(&la>O,2~O,larg zl<'II')

i~ e-a,'ldt ... '[ i·.,f1U2 1 .. ..'] _. -·-.=e,-<i? . .J; erdt- .... Er(ar)

D t+x .. o. 2 .

*

(a>O"x>O)

1.4 .• 11

l!

i'" e-otl tIt 7r.,.? ,

Il+ 2··=.2~ B.eNc ..;ax

o tx x

7.4.12 rl ral2d .. t .......;'! ea[1-(erf-1a)·2]

Jo t~+1'4· .. "y ' ..

(a>O)

7.4.13

(x. real, y>O)

.·Seepage n,

303

ERROR FUNCTION AND FRESNEL INTE<1RALS

7.4.14

fa> (x. -. -t)._·e-'%dt .. -

1f..fw(x+iy)

_'" (X-t)2+1f

7.4.15

r"W-(x2~1I)]e-j2dt' .• !: Blw{x+;ill) J 0 t4-2(;tl___.:.y2) t2:+(x2+t)2:2 .y-'J,;C

7.4;16'

r "'2X'J1e~12dt .w ...I w(x+iy)

Jo t"'-2(x2-t)t2+(:C~+ll)l! 2y-ix

(:treal,y>O)

7.4.17

i'" . la' a [at erf bt dt=-eMl2 eric ""-b

, 0 a 2

704.24

1r [1 ,(a f'Z)Ji.

+22-S2\1~

7.4.25

r,,; C~I{i - 12- { [1 (12ab)]

J 0 t2+ bll4t=1I" "V b 2-0 "V --;;:- cos (ab)

+[~-s( ~) ] sin tab)'} (§4Ja>O, e1'b>O)

7.4.26

-[4-o(.J~)lsin (ab)} (9la>O, Plb>O)

7.4.27

7.4.,20

7.4.22

7.4.23

! .. r~1 sin (t~dt=-v~{[~- (J (~.f;) ] Cos (~)

+[~-s (~~~)J sin (~)} (.Bla>O)

[1 (eL)]. (a~)}

-_ 2-0; .. sin 2;7\'" .

7.4.28

La> e-a'SCt)dt .. ~ {[~-a(;)Jcos (~:)

+[~- s 00] sin (~)} (&la>O)

7.4~29

i'" e-(lIO( {¥)dt- za(~~a)t..jall+l (91a>O)

7.4.30

I <II e-f>IS( f2!l- .. dt 1

v : J2a(~a,2+1+a)L.ja?+1

(&t'a>O)

704.31

J:"'·{[l . J' [1 J2} . I

'. ~-C(t)_ + .. --S(t) ,dt--

a 2 .·2 11:

7.4.32

re-(a.:t~+~+~')dx_!· !~e·b3_:a~ erf (..;ax+'!")+eonst.

J' 2~~ . fo .

(a~O)

304

ERROR FUNCTION AND FRESNEL INTEGRALS

7.4.33

f e-atxl :dx= ~ [ e2M ~rf ( ax+~)

+e-24~ eri (ax-~) ]+c,onst.

(a¢O)

7.4.34

7.4.38

f cos (a#+2bx+c)dz

_ /1r { (b2-ac) [ _ /2 ]

=" 2a 'cos -a;- a "a7r (ax+o)

J 1 2

7.4.35 erf xdx=x erf x+,fir e-:Z: -l-eonst,

7.4.36

ii2

f e~ erf ox dx=~ [erlZ erf Dx-e4b' ed (bx- 2~ )]

-l-const. (a FO)

7.4.37

f e4Z erf Ix. dx=~ {ea: en -vI.

+CODst.

(aFO) ,

f sin (az2+2b,x+c)dx

=~ {coseZ a ac)s[ ~! (aX+b)]

-sin (02 a ac) o[ {!; (ax+o)]}+const.

7.4.40 f G(x)ih==:tO(x)-~ sin,(~ x2)+const.

7.4.41 fS(x)dx=xS(X)+; cos (~x2)+oonst.

Num.erical Methods

7.5. Use, and Extension of the Tables

E~a:mple 1. Compute eri .7'15 and e--(.ml! using Taylor'!:! series.

With the aid of Taylor's theorem and 7.1.19 it can be shown thlllt

erf (xo+ph)=erf Xo

+ ~ e,-~ph [1-Ph;7;o+jP:iM(2~-1) J+E

e-C"ll+P1l12=e-~ [1-ZphJ:o+p2h,2(2:zfo-1)

-~ p3h3~(2~-3)]+'1

where lel<L2X10-]O, 1111<3.2X10-1o il h=lQ~2, Ipl~t. With :1:0=.74, p=.5 and using Table 7.1

erf .745=.704,6780779+(.5)(.00652 58247)X

[J - (.005) (.74) + (.00000 $3333) (.0952)] =.70792 8920

e-(.145)2=f (.6525824665) [1-.00-74

+ (.OQ0025) (.0952) + (.00000 00833) (.74) (1.9Q48)] =.574057910.

As a. check the computation was repeated with Xo=.75"p=-.5.

Example, 2. Compute erfc x to 58 for 2:=4.8. We,have 1/r=.0434028. With Table 7.2 and linear interpolation in Table 7.3, we obtain

eric 4.8= 4~8 (1.11253) (10-10)(.552669) .;; =(1.13'52)10-11•

308

ERROR FUNartON AND FRESNEL INT;EGlL\LS

References

Tert&

[7.11 J. Boersma, Comp1).tation of Fresnel integra.lsj Math. CO.lilp. 1,J,; 380(1960).

[7.2] A. V. Boyd, lnequ,alities for MillB' ra.tio, Rep.

StatiSt . .Appl. Res. Un. Jap.Sci. Engra. 6, 44-4.6 (1959) .

[7.3J O. Emersleban, Numerische Werte des Fehlel'~ integrals ftll' ,;:n;, Z. A:o.gew; Math; MeeCh. 31, 393"'394 nssn. . [7.4] A. Erd6lyi' at iI., Higher transcendental funationa, vol. 2 (McGl'aw~Hill Book Co., Inc., New York, N.Y., Toronto, Canada, London, England,. 1953). [7.5] A. ErdeJ.yi et al., Ta.bles of integr.altraDsform.s;

. vol, 1 {McGraw-Hill Book Co., Inc., New York, N.Y., Toronto, Canada, London, England, 1954,). [7.6] W'. Ga.utsclii, N'ote on bive,riate.lineil.l' 4lterpolatiOll fo.r a.na.lytic functions/Math .. Tables Aids Camp. 13, 91-96 (1959)"

[7.71 W. Ga.utschI, Recursive coroputa.ti9l'l of the repeated integr.a1s of thee.tro;r function, .Math. OOIl;lP. 15, 227.:..23.2 (1961).

[7.8] W. Gr.tibMr snd N. Hoire-jter, htegra.1't-&fel (Sprj,nger-Verle.g, Wienand Innsbruek, Austria, 1949'-50).

[7.91 D. R. Ha;rt:ree, Borne properties and a.pplica.tions of

. the repeated integrale o.f the error function, Mem.

PlOC. Maneheeber :tit. Philoa. Soc. 80, 85-102 (1936).,

[7,101. ,C. Hastings, Jr., Approxima.tions for djgita.lcom~ puters (Prince-ton Univ. Pr-ess; Princeton, N.J.; 1955).

[7.1l} y, Komatu, Ele~et\tary m.eq~ilJ.i:lijes for Mills' ratio, ·Rep. StatiBt. AppL Res, Vp·Jap. Sci. Engl'l!. 4, 69--70 (1955-57).

[7.12] E. Kieya.zig, On thezer.Clsof the FreBllel integr-als, Ce:nad.. J. Math. 9, U8--131 (1957).

[7 • .13] Th, Laible, Hobenkartedes FeblerintegraIs, Z.

Angew. Math. Phys .. 2, 48h486 (1951).

'[7.141. F .. Ll5sch lind F.Sch0blik, D.ie Fa:kultii.t (B. G.

Timbner, Leipzig, Gern:umy, 1951).

(7.15] F. Oberlietti:nger, 1'abellell ZUl Fourier T:rariSfO.I'matio)). (Springer-Verlag, Berlin, Gtittinge.o,. Heidelberg, Gel'IDa.ny, 1957).

[7;.16] J. R philip;. The flliic:tJon inv eric e, Austral.

J. Phys, 1.S, 1&-20 (19~0).

[7'.17] H. O. Pollak, A remark on "Elementary inequalities faT Mills' .ra.tio" by yo.saktl Ko.matu,· .Rep. Statist. Appl. :Res. Un. J.ap. Sci. Engrs, 4, 110 (1955~57).

[7.18] J. B. Rosser, Theo:ryan(iiLppuca.tiotlof L'e-z2dx~nd

So ~ e-p~uidy l" ,..,i~dz· (.Mapleton House, Brooklyn; N.Y.,. 1M8).

[7.19] H. E. Salzer, Formulas· for oaleu1a.tmg the error function of a complex variable, Math. Tables Aids Comp, 5,67-70 (.1951).

[7.20] H. K Salzer, Complex. zeros of the etTor func.tion, J. Fra.nklin Inst, 260, 20!)-211 (1955).

[7.21] F. G. Tl'icomi, FUnzioni .ipergeometrichecaniiuenti (Edimoni CUInonese, Rome, Italy, 1954).

[7.22] G. N. Wa~Qn, A trea.tiBe on the theory of Bessel functions, 2d ed. (Ca.mbridgf) Univ. Press, Lond~n,

Engle.n:d" 1958) ,... .

Tables

[7 .. 23J ·M. Abramowitz, Table· of the integra.1l" e.-iiI du, J.

Math. PhYB, 30, 162:-163 fi9'51}. x=0(.01)2.5, sn.

[7.24] P.O. CleOimow·and Oars. M. MUnford,. A table. of

I .i,ui.! ., >. ~~.

'V (;'II')e' . EP·"-· iJ).. for complex values of p,

p •

Phil es, Ti'll,ns.E,Ciy. Soc. London {A.l, 245 .• 189- 2:11 (1952). 1,,1 =O(.Ol}.8,arg p=0"W)45", 40. [7.2-5] R. B. Dingle, Doreen Arndt and S. K. Roy, The' integrals

. C~(x} =- (PI}-ISo'''' ~.(!!,+~)~Ie~·de

ap_d

- Dp(Z)=(PI)-l!o'" EP(e'+:ll~)~2CI4~

and th!\IT tabulation, App_l. Sci. Re.s. B 6 .• 155:-164 (:L956). C(:t), Sex), :t={).(:!-)20, iao.

[7"::16J V. N. Faddesva and N. M. Tetent'eV', Ta.bles of

values of thefu~otion w{z) =(l-'~ (I + 2~. r ~ 8 Pdt)'

.. . 'V 'II' Jo

foroomplex argument. Translated from the Ruasie,n by D. G.Fry (Pergamon Presa, New York. N.Y.,. 1961}. 'w(z),Il=:t+iy; :1),11·=0(.02)3; :1:=3(.1)5; y=O(.1)3;:z;=O(.1)5,y=3(.1)5; 6D.

(7.21] B. D. Fried and S, D. Conte, The ple.sm;a,disP!lrsio.n funnticn (Academic Press, New York, N .. Y. and London, Eng~nd; 1961). i-,fi1D(z) ,i·{illl (z), i!'=.a;+ill;· x== 0(.1)9 .. 9, y= -.9.1{.1)10; z=s var, (.1)9.9.11= -10(.1)--9.2; 68.

[7.28J K. A. Karpov, Tablitsy funktsii w.(z) =e-'; P &'~tk v

. Jo

komplekanoi oblast.i (Izda.t. Akad. Na.uk SaSR., Moscow, 1,LS.S~R.,1954).

,."';z; :Z;= 0(.001) 2 (.01) 10; 5D;

z=pelli 6."";'· 2.5°{2.-5°) 30"(1.25°)3S0 (.625·)40"; P=(J8 (.OOl)p; (.01),!!;' (,.0002)5, 0 '5_Pe 5.I,Sp;' :$;5,

5Dj

z=iy; y=0(.OOI)3(.OO02)5, 56.

[7.29] K. A. Karpov, Tablitsy funldi!ii' F(Z)=1i1&"tb: v .komple·kanoi oblasti (bdat. And. Nauk. SSSR, Moscow, U,S.S.R., ;1.958). ~r=pe'~; , e=45,o{.3125") 4$. 75°{.62-5°)55" (1.25~)65° (2.5"}90"" p=p,{.OOl)p~(.Ol)p;'. 0:$;",:<1'; ::;p~' ~5, 5Dj

21=:i;; .,;=0(.001) 10, 5E\.

[7.30] J. KaY6; A table of thefirsli ele-v:en repeatedinte. graJIl of the·ar-ror function, J. Ma;th. Phys. 34" i19--125 (1955) . inerfc·:I), x = o (.Olh2{.05).1-( . .1) :,\, n=-1(l)l1, 6D.

[7.31] B. Lohmander and S. Ribtsteu, Table of the.function

I1=e-~L;i; e~dt, Kungl.Fya:iogr.. SSllsk. I Lund Fo.rb.28.t 45-52 (1958).x=0(.01).8(;02)5, z-bO(.005) .2, 10D; z= .5(.5) 10, 20D.

Contains also 2'OD vsJuee for. maxim.UlXl 8ndip_~ fieotio.n pointS.

ERROR FUNCTION AND FRESNEL INTEGR~S

309'

[7.32] W. Lash Miller and A.. R. Gordon, Numerical ev-aJuatiGln !!If infinite series and integrals which arise in certain problems of linear hea.t flow, electrochemical cilliilsion, etc., J. Phys. Chern. 35, 27SS-:2884 (1931).

F(x)=e-z2 So· et2dt; x=Q(.01) 1.99, 6D;

x =2'{.Q 1) 4(.05) 7.5(.1)10.(,.2) 12, 88.

[1.33] National Bureau of $tandard!;, Tables of the error function. and its derivative, Applled Ma.th ... Series 41, 2d ed. (U.S. Government Printing Office, Washington, D.C., 1954).

(21-../"r)e-"~, erf x, x=O(.OOOl) 1(.001) 5.6, 15Dj (21..J7r)e-:r1, eric x, x=4(.01)10, 88.

[7.341 T. Pea.rcey, 'l'ab1e of the Fresnel in~gral (CambridgeUniv. Press, London, England, 1956).

C (/ff:} s (-Ij). x=0(.01)50, &-7D.

[7.35] Tablltsy integraiov Frenelya. (Izr,lat. Akad. Nauk ElSSR., Moscow, U.S.S.R., 1963). C(z), S(x), x= 0(.001)25, 70; S(x}, :1:=0(.001) ;58, 7Sj C(x) , x=O(,()01) .101, 78.

[7.36) A. van Wiingaa.rden. and W. L. Schean, Ta.ble of Fresnel integrals, Verh. Nederi. Aka.d. Wetensch., Ald. Natuurk. Sec. 1, 19, No.4, 1-26 (1949). O(:z:) , Sex}, :z: =0 (.01) 20, on. (Also ~oIltains aumerdeal values of the coefficients in Taylor and asymptotic expansions.)

ERROR FUNCTIO:r:.fAND ll'R.ESNEL INTEGRALS

COMPLEX ZEROS OF TIn: ERROR FUNCTION

1. 1 2 3 <\ 5

:&n

1. 4:.50b~ pl;6 2.24465 928 2.83974 105 3.33546074 3.76900 557

YIn 1~ ;l!1I;

1.8809:4 300 b 4.15899 .840

2', 61~57 514 7 4.51631 940

3 •. 17562 BlOB 4.84797 03.1

3,646174.38 (} 5.15876 791

4.06069 7n 10 5.45219 220

ert ·Zn,.=erC (-Zn) =erf 2.,,·=erf (-~) =0

329

Table 7.10

Y,l 4.43557 144 4.1B044764 5.10158804 5.40333 2&4 5.6B883 744

x,Ll~·· ( l)Jn(·.An-~.)

:;;:;- .. 4.j~-- 'F (n>Ci)

.Y1I 2 2 .. 2V~4n-~)

From H. E. Salzer, Complex zeros of the error function. J. Franklin Inst, 260.209-211, 1955 (with permission),

CQMPLEX ZEROS OF FRESNEL INTEGRALS

O(z;..) = 0 I!n =~'n +~l?b~

SCt:l.) ""0 .~,=;~+iV~

n

o 1 2 3 II

~I§~, 0.0000 1.7437 2. ssrs 3 .• 1208 3.8759

Ihi 0.0.000 0.3057 0.2529 0.2239 0.2047

!1~1o 0.0000 2. 0093 2.8335. 3.4675 4.0026

Table r.u

1/; 0.0000 0.2886 O~2443 O.21a5 n.200.8

5 4.36110,.191)9 4.4742 O.lB77

o O· 0.,-·· o. _ (). C . .0 ~. 0._ .

s(z..) = s( -z..) = s(~)= $( -.zn) =,5'{.z,,) = SJ ~".Zn) =$(-nI0=,S(1~~) =0

u. _In(''.J~)

.~;w;- ~~ ... 1 ..

""'.'1;'70·-"

r-

MAXIMA AND MINIMA OF FIt.:ESNEL INTEGR4L5 Tllble7.12

M .. =O(.J411+1) mn=O(.J4n+3) M;=SC"'4n+2) m;=S(.J4n+4)

7~ o 1 2 3 4

5

j1(.on D.779893 0.640807 0.605'721 0,588128 0 •. 577121

0.569413

m",

0.321.056 O.3B.03S9 0.404,260 0.417922 0.42703.6-

0.433666

. 1 ,,2(41l+J,)2-3 ,Mn~2+ :l'3(4n+l)~IZ

M' .1. 1!'2(43+2)2-3 n-2+,,3(4ft+2)5{2

1Jf; 0.713972 O. 62,89.4.0 0.600}61 0.584942 0.574957

0.567822

1 .. ~(41l.+3)L3 1I~n-2 .. S(4n+3)$j2

• 1 16.-2{n+i)z-3 m"~2 . 32 .. 3(n+l)6f.!

m;.. 0.343415 0.387969 O.408}Ol 0.420516 0.42B877

0.435059

From G. N. Warnon, A treatise, on the theoi-y of Bessel functions. 2d ed. Cambridge Univ, Press, Cambridge, En,gla(\d, 1958· (with permission).

8. Legendre Functions

IRENE A.STEGUN 1

Contents

Page

Mathematical Properties. 332

Notation . . . . , , . 332

B.l. Differential ~quation 332

8.2. Relations Between Legendre Functions. 333

8.3. Values on the Cut. . 333

B.4. Explicit Expressions . 333

8.5. Recurrence Relations 333

8.6. Special Values . . . 334

8.7. Trigonometric Expansions 335

8.8. Integral Representations 335

8.9. Summation Formulas , 335

8.10. Asymptotic Expa.n.sions 335

8.U. Toroidal Functions f' .336

8.12. Conical Functions. . . 337

8.13. Relation to Elliptic Integrals 337

8.14. Integrals. , . . . . . . . 337

Numerical Methods , . , . . . . . 339

8.15. Use and Extension of the Tables 339

References -. , . . , .'. . . , . . , , 346

Table 8.1. Legendre Function-First Eind P,.(z) (xS1) 342

x=O(.Ol)l, n=O(1)3, 9, 107 5....,8D

Table 8.2. Derivative of the Legendre Function-First Kind P~(x)

_ (x:Sl), .. , . , , .. , , , . . 344

x=O(.Ol)l, n=i(1)4., 9, 10, p-7D

Table 8.3. Legendre Function-Second Kind Q,.(x) (x ~ 1). 346

x=O(.Ol)l, n=O.(1)3, 9, 10, SD

Table 8.4. Derivative of the Legendre Function-Second Kind Q~(x)

(x::;; 1), . . . . . . . . . . . . . . . . . . . . 348

x=O(,Ol)l, 'it 0(1)3, 9, 10, 6-:8D

Table 8.5. Legendre Function-First Kind P .. (x) (x;::: 1) 350

x=I{.2)10, n. 0(1)5, 9, 10, exact or 68

Table 8.6. Derivative of the Legendre Function-Firat Kind P~(x)

(x ;:::i), . . . , . . . . , . , , . . . . , . . . . 351

x=1(.2)10, 11,=1(1)5,9, 10, 68

Table 8.'1. Legendre Function-Second Kind Q.(x) (x ~ 1) 352

z=1(.2)10, n=0,(1)3, 9, 10, 68

Table 8.8. Derivative of the Legendre Function-Second Kind Q~(x)

(x~l). . . . . . . . . . . . . . . . . . . . . . . . . 353

:1:=1(.2)10, 11,=Q(1)3, 9, 10, 68

The author acknowledges the aasistanee ~f Ruth E. CapUAnO, Elizabeth F. Go-defroy, David S. Liepman, and Bertha E. Walter in the prepsratlon and ebeeking of the tables aDd examples.

J. N&ti.onal Bureau of Standards.

331

8. Legendre Functions

Mathematical Propeeties

,Notation

The conventions used are z=x+iy, X, y real, and in particu!ai',x always means a real number in the interval -l::;x~+l withces 8=x where 8 is likewise a real number i n and m are positive integers or zero; " and ~ are unrestricted except whereotherwise indicated.

Othernoto.t.ions .are :

, _., n!P,,(x) P"(x) for (2n-1)t!

P1jr/l.(x) for (-l)"'~{x) 1':::($) for (-1)~P:(:zj)

P:(:t) for (-1)'" ,(2n+i){n_:m)! P:(x)

2(n+m)! :1

~~(z) for Pt(z) I O-~{z) for Q:(z.)

O:(z) for el'''IQ~(Z)

QII( ) f ,sin (11+1.4)71"" QI'( )

.z or ..2

" , sm V7I" ' , ,

(&lz>1)

Various other definitions of the functions oeeur as well 9.B mixing Of definitions.

S.l. Dift'erential Equaiion

8.1.1

('1'2) _d'lw 2- dw+,[ ( +'1) ~2] 0'

-z_ -, -- z--~ , JI" 'l"". -- W='

, ,dz2 dz 1- Z2

Solut:ioD8

(Degree II 'and order p. with smgularitiesat Z= ± l,eD as ordinary branch points-e-u, v arbitrary complex constants.)

P~(ZC), Q:('1) ~A88OCiated Legendre FunctiDns (SpbeJ:". ied Hil:rlJuinics) of tb~e Fir:stand Second J(jnds '

[srg (z± 1))<71"",

)argzl<'iT

(::2 -1)1,,- {;z-lP"(z+ l)tl'

(For PNz), .u=O" Legendre polynomials, see chapter 22 . .)

8.1.2

P~(z) r(l~.u) [:+i],jl FC-v, v+l;l-~; 12 z) (11-zl'<2) (For F(a, b; c; s) see chapter 15.-5

(lzl>1)

Alternate FQriDf!,

(.Additional forms may be obtained by means of the transformation formulas of the hypergeometric function, see [8.,lJ.)

{F( II P. 1+'V.u.1. 2)" F· (. 1. iI P. 1+'V p..3 .. :l)}

8.1.4 P:(2)~~'("~ W~ r G~~~~) r (~~~~('r (~:;~) r2(~~;)

8.1.5. P~(z)

8.1.6

(11-21<2)

2 The functions Y:"(8,rp):: ::} P:_(cos,e) called surface hal'IiH:iriics of the lirst,kin.d, _tess~_ra.l for m-<nand scctqral for 111=.11. With 0:$0 :::;:11", 0:::; 'f':$ 211", they are everywhere one valued. andcontdnuous functions on. the .surfaoe of the unit sphere J:~+!i~+ ,el=] where :t"=sin f) COB ~,y~sin ti8,in,~ and z=co:;f).

·See iillg,~ n.

332

LEGENDRE FUNCTIONS

,333

Wronskian

8.1.8

eC;'''22!'f (p+~+2) r (P+~+l) (1-Z2)re-~+2) r(p~~+l)

8~1.9 W.{P,,(z), QJ!(z)}=~(z2-1)-1

W{P~(z), Q~(z) J

8.2. RelatiensBetween Legendre Functions

Negative Degree

P~"-l (z)=P:(z)

8.2.1 8.2.2

Q~"-l (2) = [ -7I'eC,.~ COS P7rP:(Z)

+Q~{z) 'sin ['7I'(1I+.u) ]}js,in- [;,-(p-I-1)]

Negative Ar:gulDent (..f" '~Q),

8.2.3

P:(-z)=e:Fh"P~(2)-t e-1;'r'sin ['7I'(p+Ii)]Q:(z)

, . '71' '

Negative OFd.er

8.2.5

Degree ,.u+-j and Order ,II+! fJrz>O

_,11_' ( z )_ (z2-1)1!"-e-'''''Q~(z)

6.2.7. P -~-t (z2-1) 1J~ -:(1'11' }1/2 r (v + JL +1)

1).2.8

Q'=~1 ((Z2~ 1) 112)'.

=-i{t1l')l/2T{ --1'-",) (z2-1)1Ite-"'P~(Z)

8~3" ValJi.es on the Cut

8.3.1

(-l<x<l)

·SeepsOD.

(Upper and lower signs according as Jz~O.)

8.3.2

P~(x)=e±jjl'''P~(;t±iO) *

8.3,3

*"

=i;.--le,-!jA."'[e -,!tj>""Q:{::c+ iO)

-eif""Q~(x-iO)l *

8.3.4

Q~(x) = ie. -f,.'1O"re~I(":"Q: ex +iO) +e_!'~rQ~(x-ib) 1 (Formulas £or P~(i) and Q~(~) are obtained with the replacement of ::::-1 by (l-x)e±'r, (Z2_1) by (l-x2)e.±h, z+ 1 byx+ 1 forz~x±iO,)

8.4~ Explicit EXpressions (x=cos 8)

Po(X} = 1

8~4.1 Po(i) =1

8.4.2

" i (2+1)

Qo(Z) =21n :;:-1

Q, (,), 11 (l+X)

, OX: =2 n ,1 x

=XF(l i.·~·,X2)

2"" j"2" I '

8.4.3 PJ(z)~z

8.4.4

QI (2) =~ In G+ ~ )-1

8.4.5

., Xl (1+%) Ql(X)=~ n --- -1

, 2 I-x

P~(z) =i(3z2-1)

Pa(X) =!(3r-l)

=1(3 00829+ 1)

3z , 2

'(,3Ji2,-1)' ,,(1 +:1:,') 3.:t

--. In -- --.

,4, I-x 2

6.5. Recurtenee Relations

(Both p~ and Q~ satisfy the same recurrence relations.)

Varying Orde~

8.5.1 P~+I.(z) = (z2-1) -+ { (v-p.)zPt(z) ~ (11+ -,,)P~_! (z)j

334

LEGENDRE FUNCTIONS

8.6.9

8.5.2

(Z2-1)dPd~(_Z):=(P+,/.L) (1{-fJ+l) (zZ-l)JP~-l(Z)

" ,2

, -,/.L?P~(2)

3.5.3

(v- Jl+l)P~+1 (z)= (2p+l) 'ZP~(z) - (lI+ll)p~-dz) 8.5.4 (z1-1) dPJ;z)-vZP~{Z)-(jI+.u}P~_t(Z)

Varyi'Qg Qrder and J)e.gl'ee

8.5~5 P~+l (2) =P!-l (z) + (2v+ 1) (z2-1)1p~-I(z)

8.6. ~pec.ial Values x=o

8.6.1 P~(O)

=2".,...-1 cos [!:n{v+,u)]r(tJl+il-l+l)/r(tv-tfJ+l) 8.6.2

Q~(O}=

-2,,-I1r,t sin [!1i(p+.u}JrClv+t.u+iJ/r(lJl-~1-l+1) 8.6.3

[dP~(X)l"=

4::r: .. o

2,,+11r-t sin ,Il1r(l' + .u)] r( fl'+!.u+ 1 )lr(lr- !.u+t) 8.6.4

[dQ~(X)l =

dx ",-0

21';rJ cos [17l"(v+ .u)] J'( tv+ !.u+ O/l'( !v- ,,,,+t) 8~6.5

W{PI'(:)" Q"( .)}""" 22JlT(!v+tfJ+J)rC!iI+l.u+l}

• a: J • X ,"'-0 r(tv-!1-'+1)f(tv-i.u+i)

p=m=l" 2, 3, ' ..

8.6.6

P;(z) = (z1-1)'1'" dIflP.{z),

" " dz'!'

P: (x) = (-l}~(1_X2}!md"'~~x)

8.6.7

~(x) =( -l)"'(l-.xZ}t'" d!"r;;~X) .u=±!

8.6.8

P!(z) = (z2-1)-1/4(21r) -112 {[z+ (z2-1F{2J"+1 +[z+(z~-1)1t2r·-t 1

·See page II_

8.6,10

Qt{z)=i(ln.pi2{z2-1) -l/4{.a:+ (z.2-1)1/2]-.-~

8.6.11

Q;l(z)= -i(2'1Tf'2 ~:+~)~IJ4[Z+ (z~_l}ln]-.-i '"

8.6.12

P!(cos e}=(t:r)"""t (sin 6)-t eels ((p+J)8] 8.6.1.3

-

Ql(CI)S 8J=-(!7T")1 (sin 8)-~ ~in [(v+U8]

,8~6.14

P;I(cos 8)=C~i)-;(p+j)-1(sirJe)-! sin [(p+!,)8}

8.6.15

\Ie Q;t(cos 8)-(271) t (21'+1) -1 (sin 9)-; .eos ~ (p+~)81

8.6.16

~" 2-·(~~-1}i·

P; {z) r(v+l)

p;"(cos 8)2.;i:~ Ii·

8.6.17

,u=o, p=1t

(RodrigueJi' Formula)

8.6.18

8.6.19

where

W,,-l(X)= 2t_nl P"-I(J;)+3~==~) Plld(X)

2n-9" ""

+5(n-2) Pn,_;~(x)+ ' , -