SKELETAL AGE AT DEATH: THE RELIABILITY OE' CRANIA11 AND PUBIC AGE INDICATORS

SHEILAGH THOMPSON BROOKS

Miiseuni of Anthropology, University of California

Berkeley, California
TWENTY-EIGHT FIGURES

INTRODUCTIOX

The exact chronological age at death of an individual skeleton has recently assumed importance due to the dynamic approach iiow introduced into anthropometrical analyses. It is believed that by utilizing the ages of a skeletal population, estimates can be made of the size of the breeding group, of mortality rates, of childhood survival, of child-birth deaths, etc. Ages of individuals under 25 years can be calculated with rcla tive accuracy, based on the order of epiphysis-diaphysis closure of the long bones; the order of tooth eruption from infancy to about 17 to 20 years; the closure of the sphenooccipital suture before 20, or at the least, 25 years; and the union of the medial clavicular epiphysis between 25 and 28 years. In a skeletal series then, the question is how to make observations on age of death of adult individuals of 28 years and older. Determiiiation of the actual age of an adult skeleton at time of death was a problem which interested T. W. Todd, and one which he and his students attacked by various methods. They wished to obtain a simple and exact means of determining the age of the osseous skeleton by its niorphology, and as anatomists they did not interest themselves i n the age factor as applied to the statistics of ancient populations. The method of determining age by cranial suture closure has always been more generally used, due not only to the greater interest in the skull, but also because the craniumi
3 (i 7

568

SHEILAOH T H O M P S O N BROOKS

is frcquently the best prcservcd portion of the archaeologically rccovered skeleton. When in recent years population dynamics entered into aiithroponie trical analyses of archaeological skeletal material, it was noted that the ages obtaincd from cranial suture closure either produce skewed mortality rates (*enyurck, '47) or do not correlate with the known age of death (Singer, '53). A t the suggestion of Dr. Theodorc D. XcCowii, of the University of California, Departmciit of Anthropology, ail inquiry was begun into the reason why saniples of aborigiiial ("aliforiiia Indians showed a mean age a t death consistently undcr 30 years. The method was twofold: (1) to utilize hoth of Todd's methods of agc dctcrinination, that of cranial suturc closure as well as that of morphological changes occurring at the syrnphyseal surface of the pubis with advancing age, observing whether 01- not these methods arc applicable to races other than the white and ncgro-white with which he had worked ; ( 2 ) to dctcrmine the coi*rclation between these two methods of obtaining age when applied to one iiidividual. The reason for selecting these two nicthods only and not including thc age changes in the scapula (Graves, '22) and the clavicle (Steveiison, '24; Todd and d'Errico, '28) is that the chosen detcrminants can he utilized for the ages froin 18 to ,50 plus years, while clavicular and scapular ages only set a point in time with a margin of perhaps 3 2 years. The first of Todd's methods of dcteimining the age of an individual is based on the progressive. rnorphological changes of the syinphyscal face of the pubis. These changes he divided into 10 phases, the fiwt t h r w phases representing two to three year spans ; the nest phases, through Phase IX, containing -3 years or more ; and Phase S iiicluding all ages over *50pears. 111 a series of papers, published in the early 1920's (see Rihliography) hr descrihed the appearance of thc symphyseal face, at the approximate mid-point of each phase, for males and females of white and negro-white races. The differences I x tween thc groups were not sufficiently distinct to preclude usiiig thc same morphological changes f o r all 1 groups. Shortly after

T H E RELIABILITY O F AGE INDICATORS

569

this, 1925, Todd and D. W. Lyon re-defined the method of age determination by endocranial and ectocranial suture closure, working with the same skeletal material which Todd had used previously. No report has been found of an attempt to correlate the two methods with each other. They were apparently evolved separately and checked only against the known age, which is available for the skeletal series at Western Reserve University (referred to in this paper as the WRU series).
DESCRIPTIOS O F T H E UCMA AND WRU SERIES

In the Spring of 1948, Professor Edward IT.Gifford, Director of the University of California Museum of Anthropology, made provision for the authors research with the archaeological skeletal collection of the Museum (referred to in this papcr as UCMA series). At that time the California archaeological collection consisted of approximately 5,000 catalogue numbers. The minimum requirements f o r consideration of a skeleton were (1) that it be over 18 years of age, as judged by long bone epiphysis-diaphysis union and fusion of the three elements of the hip bone in the acetabulum; (2) that there be present at least the area of the three vault sutures, coronal, sagittal and lambdoid; and (3) that there bc prescrved a syniphyseal surface of one pubic bone. As each catalogue number usually represents an archaeological burial, the hones numbered and classified together are presumed to represent one individual and in cases of doubt, that burial was not included in the series. No consideration was made of area or archaclogical horizon, but all the skeletons are from California. This gave a sample of somewhat under 400 individual skeletons. I n 1953 a second series of about 70 skeletons was analyzed, using the same qualifications, as a check on the method (the histograms of this second series are shown in figure 3, but the series is not included in the tabulation of table 1 ) . After a careful survey and consultation with T. D. McCown, a series of 10 pubes was selected (referred to in this paper as Brooks models), which matched the written de-

570

SHISILAGH THOMPSON B1:OOIiS

scription of Todd of his 10 pubic phases (Todd, 20). They were chosen to represent cross sections taken from the continuum of the morphological change processes of an individual life cycle. Dr. Russell W. Newman modeled this series in dental stone (pl. 3, figs. 18-22 and pl. 4, figs. 23, 24, 26 aiid 27). For cranial age determination the charts published in Todd and Lyons paper (25) were relied on, which indicate expected closure by age f o r each suture. The amount of closure is defined numerically on a 4-point scale : 0 (open) to 4 (complete) closure. Todd and Lyon analyzed separately cndo- and chctocranial suture closure. In working with the archaeologically recovered crania, f requeiitly there is an internal patination of the cranium and even when there is not, it is difficult to see more than the vault sutures in a whole skull. The WRU crania have been sectioned either sagittally or above the orbits, so that endocranial observations are possible. The term cranial when used in reference to the UCMA series or to part of the analysis of the WRU series means all visible endocranial suture closure averaged with ectocranial suture closure. The sex differentiation was based on S. L. Washburns application ( 48) of the ischio-pubic index wherever this was possible, as well as on the breadth of the sciatic notch, the sub-pubic angle, and the general configuration of the skull and mandible. When analyzing the UCMA skeletal series, first the pubic age determinations were made, working through all the catalogue numbers. Then the recording of the cranially determined ages was begun. Not until the work was finished were the two estimated ages of the individual skeletons compared. I n the Spring of 1950 a University of California research grant made it possible to utilize the Hamann Museum collections of skeletons at Western Reserve University. Known age at death, sex, and Todds estimated pubic and endocranial ages are recorded f o r this series. With the help of Dr. William H. Sassaman, Dr. Normand L. Hoerr, and Dr. Idell Pyle all the white female and most of the white male skeletons 011 which Todd had based his results were reviewed. Since the

THE RELIABILITY O F AGE INDICATORS

571

crania are stored separately from the postcranial skeletons, completely independent assessments were made of cranial and pubic ages. Doctor Sassaman was kind enough to loan the original models which Todd had used, illustrated in the dinerican Journal of Physical Anthropology in 1920-21. These were used in conjunction with Brookss models, after first recording pubic ages with the latter models alone. Both of these sets of age determinations varied from each other less than a year or so either way. Photographs of Brooks models (pls. 3 4 , figs. 18-28) and of Todds models (pls. 1-2, figs. 8-17) show that the sets resemble each other closely. The WRU series is not a random sample, but deliberately selected by the author so that each pubic phase has approximately the same number of individual skeletons whose known age falls within that span. This was done as a test check for the correlation of Todds pubic and cranial age, Brooks pubic and cranial age and the known age.
PRELIMINARY ANALYSIS OF THE DATA

Various procedures were used in analyzing these data, but only those relevant to the problem of correlating the two methods of age determination are discussed here. I n coiisidering the UCMA series, it is immediately apparent that there is no way of checking the accuracy of either cranial or pubic ages should they be divergent in one individual, except by an indirect approach. The age indicators apparent in the entire skeleton were eliminated at the outset because the definition of the problem deliberately limited the work to an analysis of only cranial and pubic age determinations as circumscribed by Todd and Lyon. The purpose is to isolate these bones, treating them as single variables and to disregard the rest of the skeleton. The attempt is comparable to the palaeontologists method of relying on a fossil genus or species, which is restricted to one geological stage. These type-fossils are geological age indicators : Can the particular morphological appearance of certain human bones be considered reliable age indicators in the same way as these

572

SHEILAGH TIIOMPSON BEOOKS

type-fossils? The mean ages were calculated for the male and female series, cranial and pubic ages separately, and appraised through ethnological information for their probable errors. Histograms were also plotted showing the disTABLE 1

A g e ckaracteristics of thc Z7CLXAand W R U series

WRU SERIES UCMA SERIES

Male

Female

Male

Female

Observer Total number in series

: : 2

TWT
100 .91

STB
103
.87

TWT
26 .67 .56 .58

STB
82
.72
.41 .40 .38 .33
, .

STB
194 ...

STB
177

Pubic and known age Endocranial and known age Pubic and endocranial age

...
...

3
2
n

.78
.80

.74
.73 .71

...
...

...
...
.38

2 Cranial and

known age Cranial and pubic age

.. .. .. ..

..
..

...
.39 31.8 k 5 0 33.3 & .58

bl

.65
.. ..

Nesn cranial age

Mean pubic age

..

29.7 i .49
39.6 -C .80

..

, .

Mean of individual deviation of cranial from pubic age Mean of individual excess of pubic over cranial age

i 6.5

i 11.6

+ 1.6

+ 9.7

tribution of cranial and pubic ages by sex. These histograms presumably reflect mortality rates among the California Indians archaeologically. The WRU series, having a known age factor, is the base sample that actually confirms the conclusions which were originally demonstrated by the UCMA series.

THE RELIABILITY O F AGE INDICATORS

573

The UCNA series illustrates the sharp deviation between

ages in the female skeleton determined on the basis of cranial

suture closure arid pubic face pattern. The histogram (fig. 2) of the UCMA female series has two points of high frequcncy, the cranial age mode at Phase IV, or 21-26 years, a i i d the pubic age mode at Phases I X and X, or 45 to 50.
cranial Age Pubic Age

-B -

55 5Q

A
g e

1 n

Y
e a
I
S

Todds Pubic Fhases

Fig. 1 Nortality curve, UCMA male.

There is a 10 year difference (table 1) between the female pubic mean age and that of the cranial age. These findings are very similar t o those of fjenyiirek (47), who refers to the slower rate of sutural union in females of an ancient Anatolian population. The female cranial suture closure tends to lag from 5 to 25 years, with a mean of 9.7, behind the age changes observed in the pubic bone. The histogram (fig.3), representing the second series analyzed about three

574

SHEILAGH THOMPSOX BROOKS

years later, has the same configuration as the original VCXA female series. The data obtained from the UCMA male series are not similar to the female series and show a somewhat higher correlation between cranial and pubic ages. There is approximately two years average difference between ages determined
Fubic Age

Todd's Fublc Phases

Fig. 2

Mortality curve, UCMA female.

from cranial suture closure and pubic face pattern (table l), also apparent in the overlapping modes in the histogram of the male series (fig. 1). I n males, the age calculated froni suture closure tends to deviate not more than plus or minus 5 to 8 years (table l ) ,from the age obtained from the pubic face pattern. The histogram (fig 3 ) plotted from the second UCMA series corroborates the original data. If this is a personal error. it is consistent. However, after noting the

THE RELIABILITY O F AGE IFDICATORS

575

similarity between Todds pubic models (figs. 8-17), based on the WRU white male series, and Brooks California Indian models (figs. 18-28), it appears plausible that this is not a personal error, since Brooks pubic types, taken from
Pubic Ago
Cranial Age

- CT]

--

12 A

1-

I : a

1 n

Y
c
a

c
S

Todds Pubic Phases $

10

:L
I

I x x

Fig. 3 Mortality curves, UCMA second series, male and female.

Todds written descriptions, match Todds original models so well. Furthermore, it cannot be easily explained on the basis of racial differences. When Todds URU series were evaluated both male and female cranial age determinations show lack of correlation with the known age. Although Todd did not publish tables of correlations between his estimated

576

SHEILA(+H THOMPSON BROOKS

endocranial age and known age, he did warn his readers that even endocranial age, which he believed to be more rclliable than ectocranial age, had a range of error of at least plus or minus 10 years. The UCXA histogram for the female series shows the highest frequencies of the cranial curve at the ages of 24-26 ;vct\rs. Considering the histogram as a mortality curve, the highest mortality rate occurring during the years of 24-26 is not a t all concomitant with ethnological descriptions of California Indian villages in which a high percentage of wonicn lived to ages w d l over 40 or 45 years (see Rancroft, J)ul)ois, Foster, Gayton, Kroeber, Powers). Moreover it does not fit the high mortality frequency of American Indian fcniales in the age groups over 41 years as cited by Clark Wissler ( '42). The female pubic age curve does not seem more accurate. According to research carried out by Wissler ( '36), and especially by Clements ('31) and Krogman ('35) on American Indian vital statistics, the male pubic age histogram is a closer approximation than the UCRlA female histograms to the type of American Indian mortality curve. The grouping of all ages over 50 years into one category creates a minor distortion that, however, does not affect the major comparisons of the crucial adult ages from 25 to 50. The MWU series, both Todd's and Brooks', were examined from several different points of view. Regression coordinates were plotted for all groups within both series. They a r c not presented here as the table (table 1) of coefficients of correlation presents the same data in a more compact manner. I n effect, Todd's coefficients and Brooks' are very similar for the WRU series. Coefficients of correlation for the UCMA series are included here to complete the tabulation, and they coincide exactly with the more directly derived results. Todd and Brooks show a very high correlation between ages determined from pubic face pattern and known age for males. The figures are nearly 20 points lower in the female group, although the correlation is sufficiently high to be significant. Male endocranial and known age coefficients

THE R E L I A B I L I T Y O F AGE INDICATORS

577

of correlation are not as high as the pubic and known age correlations, but they are not below statistical significance. Endocranial and known age correlations are extremely low in both series of females. Thus, the WRU series corroborates the UCBIA data: the male pubic age seems to be quite accurate, the female pubic age less so, and the cranial age, in both males and females, especially the latter, of dubious reliability.
RESULTS 0% THE SC4TTER DIAGRAMS

Shortly after completing the second UCMA series, it was decided to re-analyze the WRU data before beginning research on the problem of age determination in the whole skeleton. The WRIT series was plotted on scatter diagrams, pubic age being the X-coordinate and cranial age the Y-coordinate. The known age, being an absolute, was plotted as a diagonal line. If the conclusions stated above are correct, the diagrams would show this as clearly as all the other tests had. Regrettably for the conclusions, the WRU diagrams show a much more complex and not very encouraging picture (fig. 4). At the University of California Statistical Laboratory it was decided that the tests previously applied to the WRU material are not valid methods for checking this type of data. The histograms for the UCMA series and the scatter diagrams f o r the WRU series were merely an approach t o the correct evaluation of this qualitative type of analysis. The formulae that produce coefficients of correlation, standard deviations and regression coordinates, tend to distort such data as are presented here, where each individual age is an estimated figure with a plus or minus range of a year or more. Mohamad S. Ahmed, of the University of California Statistical Laboratory, re-analyzed the WRU series (table 2, figs. 5 and 6). He allowed each qualitatively analyzed age determination to represent itself against the known age for that individual on two graphs, one for pubic and one for cranial age. Thus, his conclusions, though startling, are factual and a base for further research into the question of the validity of age indicators.

5'78

SHEILAGH THOMPSON BROOKS

Phase X is not included in the graphs since it has no upper limits aiid is more or less unreliable in an analysis which is detcrmining group limits. The pubic age graph (fig. 5) is plotted with the known age as the X-coordinate and Phases I through I X as the P-coordinate. Rectangles were then drawn "whose sides parallel to the Y-axis, representing the groups, are of fixed length; and the other sides parallel to

I-11-111-m-

- VI

- VII

-VIII

Hale Fubic Ago

48 43 50

F
c
CI

1
e
C

38-

r
n 1 a 1
a

/.

/
. I

IX

-x

.. . ." .. .. .

Female Pubic Age

Fig, 4 Scatter diagrams, WRU, male and female. Roman numerals indicate Todd ' pubic stages. 5

THE RELIABILITY OF AGE INDICATORS

579

the X-axis are chosen in such a way that on the whole the rectangles cover a s many points as possible, without overlapping of the age limits (letter of February 16, 1954, from JL. S. Ahmed). The cranial age graph (fig. 6) is plotted with
Phase Limits Proposed Modifications
Male
Female

-- ---. a

T 50
0

d d s p 45
b

Ix

:.;.;
I
1

i
C

p 40
2
S

e
S

. 35
I !

; 30
5

. . # ( . .

a m
30

25

20

20

25

35

40

45

50

Known Ags in Years

Fig. 5 female.

Scatter diagram of pubic age (STB) and known age of

W U male and R,

the known age as the X-coordinate, and the predicted age based on cranial suture closure as the Y-coordinate. Since in the pubic graph (fig. 5), the side of the largest rectangle used in these solutions is 7 years, hence f o r the purpose of comparison giving the best chance to the method based on

580

SHEILAGH THOMPSON BROOKS

cranial measurements, the predictions based on cranial measurements are taken t o be f 3.5 years., two lines were drawn parallel and 011 opposite sides of the bisector at a vertical distance of 3.5 years from the bisector and the numher of points within this band was noted. This number gives

.
-

WRU Male and Female

Male Female

STB Cranlal-Knorn Age S c a t t e r Diagram

17

20

25

30

35

40

45

50

Known Age in Years

Fig. f i

Scattrr di:igrniii of cr:iiii:il ngcx (STB)

:uitl

kiiowii age of WRU, male

:Ind ftwnlc.

the total number of correct predictions for the method based on cranial (from letter cited above) suture closure. As can be seen in table 2 the percentages of correct prediction for male and female, cranial and pubic age determination, are 1ess.than 50%. Ahmed has suggested that if the

THE REI,I.%HILITP OF AGE INDICATORS

581

limits of the phases were to overlap the percentage of correct prediction of age based on the pubic face in the male would be increased to 72%. This is not the desired solution, since a concisc series of age indicators thc limits of which do not ovcrlap is sought -in fact even Todds overlapping of Phases V, V I and VII by one year each, has been criticized. hhined has also suggested a modification of the age liiiiits (table 2) which shifts Phases V through V I I I makiiig them about three years younger. This is to accommodate the consistent underageing graphically shown in figure 5. When Todds pubic data are graphed the same way, similar disparities appcar (fig. 7 ) . The patterning as a whole is somewhat better than Brooks, hut the percentage of correct prediction is only 5476 for males and 3176 for females. I n part the reasoii for the low pubic percentages is due to the disparateness created by the rectangular uiiits in the graph; the continuum of parallel bands in the cranial graph (fig. 6) creates a slightly higher total of percentage predictions by iiicludiiig all points outside the bisector by 3.5 years. There is a possibility that Ahmcds modified limits would also increase the accuracy of Todds predictions to 60-70%, The potential improvement in these modified limits should not be overlooked, since over 60% correct prediction of male ages by only one skeletal age indicator is an unexpectably high per cent. Recently Singer ($3) published a critical account on the inaccuracies of age determinations through cranial suture closure. His commeiits and conclusions are in accord with the results derived from both thc CCRIA and the WRIJ series of cranial age determinations. The female groups had inuch greater discrepancies than the male groups in thc WRU series between age determined from cranial suture closure and known age. The tendency is for cranial sutures of female iiidividuals to remain open, either throughout life, or to close much later than predictions based on present age determination methods. This trait of unclosed cranial sutures in the female is present also to a certain extent in male

582

SHEILAGH THOMPSON BROOKS

sutures. Frequently, iii both the WRU and UCMA series, skulls were found whose vault sutures were completely united, indicating a minimum age of 46 years, and the circummeatal sutures were open, indicating a maximum age of 37 -years. Jlore often all the sutures were open, giving an age prediction

Knom

As#

i n Yews
nlld

Fig. 7
f~~llllll~.

Scatter diagram of pubic age ( T R T ) and known age of U'RU, mnle

of under 25, when tooth wear and decay, clavicular epiphpsis closure, scapular lipping, and the symphyseal surface of the pubis indicated an age well over 35 pears. I n such cases in the TVRU series, the known age established the relial>ilit;v

THE RELIABILITY O F AGE INDICATORS

583

of these other age indicators and, thus, the lack of validity of cranial suture closure as an age determinant. This evideiicc, combined with Singers data based on Cape Coloured and White German, seems to be more than adequate t o recommend that cranial suture closure be considered a n unreliable tigc indicator, regaidless of sex or race.
T.\BLE 2

Btr.srtl on IfRU series. L i m i t s of pubic phases

~

PUBIC IlIASE

AGE L I M I T S ACCORUIXICI T O TWT

lODIPIED

OVRRLAPPING

LIMITR

LIMITS

I1 TI1
1V

v vI
VII VIII

Ili

17.5-19.5 19.5 -21.5 21 5 2 4 . 5 24.5-26.5 26.5-30.5 30.5-35.5 35.5-39.5 39.5-44.5 44.5-50.5 50.5

17.5-19.5 19.5-21.5 21.5-24.0 24.0-2G.0 26.0-27.0 27.0-33.5 33.5-38.0 38.0-42.0 42.0-50.5 50.5

17.0-20.0 19.0-22.0 20.0-25.0 23.0-27.0 25.0-30.0 27.0-35.0 34.0-41.0 40.0-47.0 45.0-50.0 48.0

l c i c m t a g e of corrcct prrdiction
~

TWT PHASE LIMITS

MODIFIED

LIMITS

OVERLAPPIKG

C:IANIAI. SUTURE (LO 8 URE

Male STB Eeinnlc S T B Male TWT Frmnle TWT

30 35 54 31

(i 1 39

72 45

45 2G

.. ..

..

.. ..

ALTERNATIVE PATTERNS OF METAMORPHOSIS O F THE PCBIC SPMPTIYSEAL FACE

This limitation of the value of ages determined from cranial suture closure makes it necessary to inquire why the ages determined from pubic face pattern for the WRU series, both Todds and Brooks, have such a low percentage of correct prediction. A close look at figures 5 and 7 show that many of the individual ages, especially in the male sample, ai*c too young by only a few years. Ahmed believed that the

584

SHEILAGH THOMPSON BROOKS

modified age groups would correct this error of underageing. There a r e other factors in utilizing the symphyseal surface of the pubis as a n age indicator which Todd ( 23) discussed. I n his early work, 1920-21, Todd delineated the progressive morphological changes that occur on the symphyseal face of the pubis, from age 18 to 50 plus years. The actual models (figs. 8-28), both Todds and Brookss, were chosen to repren sent the midpoint of each of his 10 phases. I both series, some models deviate from the original description of the central tendency, but in utilizing the models allo\mnce may have been made for such deviations. In addition to these described phases, which a r e perhaps the most usual appeartiiices f o r these ages, there a r e alternative morphological patterns with other contours to the pubic face. Some of these alternative patterns have been described by Todd (an), one is described below, a i i d there a r e others which have not yet been analyzed. Todd grouped his 10 phases into three general categories. The first includes Phases 1-111, the post-adolescent stages. The second consists of Phases IV-VI, during which the outline of the symphyscal surfacc is built up. The third covers Phases VII-X, the period of g a d u a l quiescence and later break down (Todd, 20, 21). To the phases which represent build-up, IV through VI, Todd later (23) added other patterns of metanioyphosis. These alternative morphological patterns he called the anthropoid strain, regressive strain and a n intermediate type. They a r e not successivc stages of any of the 10 standard phases, but rather morphological changes alternative to the more usual patterns. OIIC more variation from the norm has heen found and is described here; it is called the rough type as opposed to the more frequent smooth type, which occurs in the original series. This rough type was discovered while working with the WRIT material, so an absolute check on age was obtained. A subsequent re-examination of the California Indian material proved the rough type to he present here also.

THE RELIABILITY O F AGE INDICATORS

585

The rough type had been previously mistaken for a n anomalous Phase X, so the observational difference between the two patterns rests on the recognition of the difference bet w e n the break-down of old bone and the appearance of a iicn surface of smooth bone moving dorso-ventrally over the previous young surface of a Phase I V symphyseal face. The rough type first appears clearly in late Phase V or early Phase VI. During Phases V I arid early VII, there is to be seeii a smooth layer extending from the dorsal edge about half way ovei the symphyseal face (fig. 25). There has been h i l t up the characteristic third period edge a t the dorsal rim a s wcll as at the cranial and caudal margins, if it has extended to thcm. Where this smooth encroaching layer eiids, there is a n abrupt declivity, the length of which is usually crenulatcd. Not yet covered by this escarpment, which extends from the cranial to the caudal margins of the symphyseal face, can be seen the somewhat rippled surface which is characteristic of the post-adolescent period. This young syniphyseal face generally call still be seen caudad, niei*giiig into the inferior ramus of the pubis, and craniad, clsteiiding tmvard the pubic tubercle ; there is no definition of thc ventral edge, although it is beveled as Todd describes f o r Phases I11 and IV, and neither cranial nor caudal margins a w delineated. There a r e other patterns of pubic metamorphosis which have not been recorded. In particular among female pubic holies, there is a pattern which gives the appearance of age due to the excessive slciideritcss of the symphyseal face. The complctc change of the syrnphyseal face from Phases I V through 1711, gives ample opportunity for morphological variations, though certainly none of the phases of the pubic nictarnoi*phosis is static. Todds anthropoid and regressive types as well as the rough and smooth types are quite diffcreiit in their patterning. Further work is needed t o record the frequencies, racial and sexual, of the alternative patterns of puhic metamorphosis, as well as to check the occurrence of other patterns. Thcrc have heen noted frequently certain

586

SEIEILAGH THOMPSON BROOKS

apparent pathologies, such as a puffy and swollen appearance of the symphyseal face, the persistence of the junction lines of the ischial and iliac epiphyses well into adulthood, and a n apparent static quality with certain features of a young phase being carried into an older one. Long before Todd defined his methods of age determination an error mas made by believing that skeletal age could be calculated from one bone, or perhaps from the cranium, alone. The fact that individual bones are riot isolates seems to have been ignored. The laboratory skeleton was once t i p a r t of a living entity on which debilitating changes with time, accidents and death have left their marks. Therc is no doubt that certain parts of the slielcton can be looked upon as age indicators because of their physiological-anat omical functions. The cpiphysis-diaphysis union, which is known to be a valid sub-adult age indicator (Gisculich and Pyle, 50) has a causative relationship with growth hormones arid the physiological factors of increasing age. Todd (21) has said that the pubic symphysis represents an incomplete epiphysis-diaphysis union. The fusing bony nodules which occur in Phases I11 through V, are rpiphyseal in origin. This is why Todd chose the symphyscal surface of the pubis as a n age indicator. It is also why through modification of the age limits of phases and arriplificatioiis of the phases therriselves, the symphyseal surface of the pubis can be made into onc of the mom reliable age indicators for adult skeletal material. Age of a n individual, where there is a whole 01 partial skeleton, should never be dctcrmined on the basis of one age indicator, no matter how reliallle. There are a1ways possibilities of arthritis, boiie tul)erculosis, osteoniyelitis, etc., which can alter a single hone or a whole skeleton. Problems of alternative p a t t e n s of pubic metaniorphosis a r e still unsolved. They should not hc analyzed as isolates, however. Future research should deal with all the changes to the skeleton created by advancing age in the individual over 18 years. The method of age cletcrrriinatiori of an individual over 18 years would thcn inchidc the union of t h e

THE 11ELlbl3ILI'CY O F AGE INDICATORS

587

iliac a i d ischial epiphyses, fusion of the medial clavicular epiphysis, age changes in the scapula, the age patterns of thc pubic face and of other articular surfaces, wear and decay of the tceth, arthritic and other agc changes in the vertebral column, and also cranial suture closure. Certain items in this list can only be utilized when considering skeletal material under 30 years, other items are usually applicable only to individuals over 35 or 40 years. Cranial suture closure, especially in male skeletal material, may be used as a secondary confirmation, with caution. The age changes of the symphyseal surface of the pubis, when correlated with the other skeletal agc indicators, will probably be considered to be the major component in age determination. When the pubic face does not comply with Todd's illustrated phases and the remainder of the skeletal age indicators place the individual between the ages of 25 and 37 years, therc is present one of the alteriiative morphological patterns. The frequency of such occurreiices is relatively high in both sexes and even when these patternings have been established, the pubic phase is only a part of the whole skeleton.
SI'MMARY

1. The validity of the closure of cranial sutures and morphology of the symphyseal surface of the pubis are considered a s age indicators. 2. Both methods are utilized independently on thrcc skeletal series: the University of California Museum of Anthropology archaeologically recovered California Indian skeletons, two series analyzed separately ; the confirmatory data based on a series of skeletons from the Hamann Museum Collection at Western Reserve University. 3. The results of all three series are basically the samc: male age estimates derived from the morphology of the pubic articular surface have high coefficients of correlation with the known age; female age estimates are slightly lower in their known age correlation; and, cranial age estimates, even

ciidocraiiially determined, corwlated poorly aiid sliould lie used only with caution. 4. The WRU series, both Rrookhs aiid Todds original data, are plotted on scatter diagrams, the known age as the ordinate and the pelvic age as the abscissa, and the percentage of correct prediction determined. These percentages are 54% for Todds data and 30% for. Brookss. The diagrams also show a consistent overageing by pubic age estimates and when cranial age estimates are used, A constant undcrageing.

CoiwLmross 1. The age phases of the morphology of the syinphpseal surface of the pubis should be modified according t o the limits proposed by 31. S. Ahi-ncd, as these corrcct the overageing resulting from the original phase limits. 2. Should the articular face of the pubis present A inorphological appearance not in accord with the standard dpscriptioris of Todd, one of the alternative patterns of metamorphosis niay be present. Todd has described three a l t e i ~ i a tive patterns and another is iiicluded in this paper. I n ally series, should a variaiit type of symphyseal face appear with regularity, frequency of occui*reiicefor male and female should be recorded. 3. Age estimates determined by cranial suture closure arc unreliable. As an age indicator in fernale slwletons, cranial suture closure is doubtful, in males it should only he utilized as confirniatory t o other iiidicators of adult age. -1. Even employing the modified limits of the puhic age phases, an age estimate for a skeleton should not he based on pubic morphology alone. The skeleton, in z&.o, w a s R unit a i d when under analysis in a laboratory, should still be considered as a whole. No one age indictitor is adequate.
LITERATURE (LTICI) B.ZNCROPT, HUBERT O I ~ H 1883 The Works of If. I. Baxcroft. Vol. I. T l ~ r Native Races. Wild Tribes. A. 1,. R:iiieroft, Snn Fr:meisiCo, pp. 311-iio. CLEMENTS,FOREST 1931 Racial diffrrenrrs in mort:ility :ind morbidity. H u m a n Biology, 3: 397-419.

THE RELIABILITY O F AGE INDICATORS

IHJUOIS, CORA 1935

589

Wintu ethnography. Uiiiv. of Calif. Publ. in Am. Arcliac010gy and Etli1lolog~,36 : 1-147. FOSTEK, GEORGEM. 1944 A summary of Yuki culture. Aiitliropological Rerords, 2: 155-244. (;ITTON, A. H. 1948n Yokuts and Western Mono etlinograplihy. I. Tolare Lake, Southern Valley, and Central Foothill Yokuts. Aiitliropological Records, 1 0 : 1-141. 1948b Yokuts and Western Mono cthnographp. 11. Northern Foothill Yokuts and Wester11 Mono. Anthropological Records, 10 : 142-301. GRAVES, WILLIAMW. 1923 Observations on age changes in tlic scapula. Am. J. Phys. Anthrop., 5: 21-34. GXEULICII,WILLIAN WALTER,AND S. IDELL PYLE 1950 Ruiliograpliie Atlas of Skeletal Development of the Hand and Wrist. Stanford Uiiiv. Press, Stanford. TJ. 1925 Handbook of the Indimis of Califorilia. Bureau of KROEBER, ALFRED Am. Ethnology, Bull. 78, Washington, D. C. KROGnrAN, WILTON M. 1935 Vital d a t a 011 the populatioii of the Seniinole 111dians of Florida and Oklahoma. Ininan Biology, 6 : 335-349. IOTVERs, STEPHEN 1877 Tribes of California. Contributions to Sorth Binerzcnn Ethnology, Vol. I1T. Covt. Printing Office, Washington, D. C. YEXYFREK, MUZAFFER SUI.EYJlAN 1947 A note on the duration of lifc of the ancient inhabitants of Anatolia. Am. J. Phys. Antlirop., n.s. 3 : 55-66. SINGEH, RONALD 1953 Estiniatioii of age from c r a n i d suture closure. J. of Forensic Medicine, 1: 52-59. STEVENSON, H. 1924 Order of epiphyseal uiiion i n man. Am. J. Phys. PAUL Anthrop., 7 : 53-93. TODD, T I ~ O N A WINGATE 1920 Age changes in tile pubic bone. I. The male S white pubis. Am. J. Phys. Aiithrop., 3: 285-334. 192la Age changes in the pubic bone. 11. Pubis of iiiale ncgro-whit(. 1. hybrid. 1 1 Pubis of white female. IV. Pubis of fenialc iiegro-wliitc hybrid. Am. J. Phys. Aiithrop., 4: 1-70. 1921b Age changes in the pubic bone. V. Mammalian pubic mctnniorphosis. Am. J. Plips. Anthrop., 4 .- 333-406. 19210 -4ge changes in the pubic bone. 171. Interpretation of variations in syniphyseal area. Am. J . Pliys. Aiitlirop., 4 : 407-424. 1923 Age changes in the puliic syniphysis. V I I . The anthropoid strain in huniau pubic symphyses of the third drcade. J . A4~i:it.,j 7 : 274-294. 1930 Age elianges in the pubic bone. VITI. Roentgenograpliic differentiation. Am. J . Phys. Anthrop., 1 4 : 255-272. JR. 1924 Endocranial suture closure. TODD, TIIONASWINGaTE, AND E. w. LYON, Part I. Adult males of white stock. Am. .J. Phys. Anthrop., I : 325 384. 1 1925~1 Cranial suture closure. P a r t 1 . Ectocranial suture closure in adult males of white stock. Part 1 1 Endocranial closure i n adult 1. nialcs of negro stock. Am. J. Phys. Antlirop., 8: 23-71. 1925b Suture closure. P a r t IV. Ertocrnnial closurc. in adult males of negro stock. Am. .T. Pliys. Anthrop., 8: 149-168.

590
WASHBURX, 1,. S.
11.5.

SHEILAGH THOMPSON BROOKS

1948

Sc.x diffcreiices iii tlic pubic boiic.

.\in.

J. Pliys. Aiitlirop.,

6 : 199-207.

1949 Sex diffcreiices iii the pubic boiir of Jhiitu and Bnshninn. Am. J. Pliys. Antlirop., 11.8. 7: 425-433. WISSLER,CLARK 1936 Distribution of deatlifi timong Anirriran Indians. Hiininn Biology, 8: 123-231. 1942 Human cultiirnl lcvcls. In : P r o b l ~ w s Ageing, R. V. Corndry, of cd. Willianis and Wilkim, H;iltiniore, pp. i7-90.

PJIATE 1
ESIT.ANATION OF FIGUHRS

T. W.Todds niodels of the syinpliyseal suifacc of t h e pubis.

8

L o w w riglit. Pliasc I (age 18-19).

Center riglit. Phase I1 (age 20-21 j . Upper riglit. Plinsc I11 ( a g e 20-81 j .

9
10

11 L o ~ c Icft. Phase TV (nge 2S-86). r

12 Uppcr left. P1i:ise V (:igr 27-30),

THE RELIAEILITY O F AGE INDICATORS

SHEILACH THOMPSON BROOKS

591

'1'. W. Todtl 's iiiotlels of tlie synp1iysc~:il surface of tlie pubis.

13 T,o\\.er I c ~ f t . P~I:ISO ( a g ~ VI 30-35'1. 14

13
I(i

Vpper k f t . 1'li:isc T I 1 (:igc 3,5-39).

Upper right. 'lr:ise V I I I (agc. 39-44).

Cc1itc.r right. Plinscl TX (agc 4 - 5 0 ) .

1 7 Lower riglit. Pli:isc Y (:rye 50 aiitl iip\v:ird).

THE RELIAEILITY O F AGE INDICATORS

SHEILACH THOMPSON BROOKS

593

S. T. J 3 r o o k i ) ~ moclclr of tlic syiiil)liyw:il s i i r f t i c c of tlic piibis. Ages nwtlificd accordiiig t o 31. S. Aliiiwd.
18

L o \ \ w lcft. Ylinsc 1 (age 17..5-19.5).

19 Upper left. Pliiise I1 (age 19.5-51.5).

70 T,oiwr riglit. Plinsc 111 (ngr 31.5-24.0).

21 22 Cciitcr riglit. Plinsc IV ( n g r 24.0-26.0).

Uppcr riglit. P1i:ise V (agr 26.0-27.0).

THE RELIAEILITY O F AGE INDICATORS

SHEILACH THOMPSON BROOKS

PTJATE 4

8. T. Hrookxs iiiocli4s of tlie ~ p i p l i p e n surface of tlie piibis. l Agiw luoclifiiul :iceording to M. 8. AlimmL
23 Lower left.
l%tlsl~

vr, smooth type

(r~ge 27.0-83.5).

24 Chiter Icft. Pliriw VII, siiiootli typc (agc 3 . - 8 0 . 353.)

25

Upper lcft. Plirixc! Vl-VlI, rough type (nge 28.0-3i.0).

25 1,ower riglit. Plinso VlII (:igc 38.0-42.0). 4

27 Ccwtc-r right. P l i n s IX (age 42.0-50.5).

28 Vpper right. Pli:iae X ( : q o 50 and upward).

TII1.2 RELIABILITY OF AGE lNUIC.\TORS

s t 1 E i r . w ~ IIOAII~SOX T BROOKS