[ 233 ]

HEADFORM AND HUMAN EVOLUTION

By A. A. ABBIE, Department of Anatomy, University of Adelaide
INTRODUCTION

The progressive rise in mean cephalic index, which is such a striking feature of human evolution, has lately attracted much attention, particularly from Kappers (1935a, b, 1936a, b, 1937, 1939, 1945) and Weidenreich (1940, 1945). Kappers considers the various aspects of the problem only as they affect man. Weidenreich, on the other hand, adopts a broad biological approach which adds very materially to our knowledge from all points of view. However, his conclusion (1945, p. 50) that 'Brachycephalization, therefore, means a better adjustment of the skull to the erect posture by adopting a more complete globular form' carries little conviction for there is no evidence that the most brachycephalic peoples are better adjusted to the upright posture than are the most dolichocephalic. Weidenreich's belief that increasing brachycephalization is associated with human evolution generally is certainly correct, but some other explanation should be sought for the coincidence. Ever since its inception the cephalic index has remained the major criterion of human affinities and development. On the other hand, anyone who has measured a series of skulls must have realized that the loci from which the index is calculated vary from skull to skull, and that the index expresses something different for every skull. Nevertheless, Morant (1928, p. 355) claims that 'only one character has been found which is capable of making an absolute distinction between more than one of the groups [i.e. of humans] and all the others: that is the cephalic index'. Thus the value of the index seems to be well established, but it can be accepted only as a total response to a number of variable factors, sofime of which are probably still quite obscure. This is the attitude adopted in the present inquiry which seeks some unifying principle under the apparent diversity of a multitude of observations. The problem of headform is considered in its relation to evolution, race, size of jaws, heredity and environment, internal form and capacity, stature, sex and individual development. The fact that such a unifying principle does seem to emerge justifies this addition to the already extensive literature on the subject.
MATERIALS

Data relating to the Australian aborigine are employed as a control on a number of the various factors considered here. The cranial data comprise measurements on both adult and juvenile skulls in this Department and, especially, in the Adelaide Museum. For adults, fifty male and fifty female skulls were measured: they represent about 10 % of the total material available, are representative of all parts of Australia and form a strictly random sample since they were taken just as they came. The fact that the mean cranial
Anatomy 81
16

Table 1. Length, breadth, cranial index, facial angle and alveolar index of fifty male and fifty female Australian aboriginal skulls
Male
A

Female
A

Means Standard deviation Corr. coeff. L./B. 0 530 (P <0.001) Corr. coeff. F.A./Alv. I. Corr. coeff. L./F.A. 0-096 (n.s.) Corr. coeff. 0-153 (n.s.) L./Alv. I.

Cranial Facial index angle 81 66*5 79 66*7 67*2 86 67-2 80 67*2 75 132 67-0 68-0 77 126 67*0 84 68 3 129 67*2 77 68*4 128 67*4 68-5 75 124 67-4 69-0 78 1'95 132 67-7 69-2 80 202 138 68-3 74 69*2 187 128 68*5 77 69*2 173 68*8 119 75 69-3 176 121 68*8 69-3 77 184 127 69-0 179 124 69-3 83 187 129 69-0 173 120 78 69-4 190 131 69-0 178 124 78 69*7 192 133 69*3 172 120 76 69*8 192 133 69-3 96*9 173 122 70*5 79 183 127 69*4 98*0 181 128 80 70*7 190 132 69-5 101-0 178 126 70*8 80 188 131 69-7 110.0 170 121 71.2 76 182 127 69*8 98-2 184 131 71-2 80 182 128 70*3 97*9 182 130 71-4 70 185 130 70 3 101 9 186 133 71*5 87 190 134 70 5 108-6 172 123 77 71'5 194 137 70-6 109*5 176 126 75 71-6 187 132 70-6 108*1 176 127 76 72-2 187 132 70-6 100-0 180 130 72-2 76 193 137 71*0 96*1 177 128 72-3 73 196 140 71-4 100*0 184 72 133 72-3 182 130 71*4 97*9 167 121 72-5 82 183 131 71-6 100-0 171 124 72-5 77 195' 140 71-8 110*2 172 125 72*7 81 185 133 71*9 103-8 181 132 72*9 78 182 131 72-0 100*0 176 129 73.3 79 183 132 72-1 106-9 174 128 73-6 81 174 126 72-4 103*9 182 134 73-6 73 175 127 72-6 100*0 167 123 73-7 79 186 135 72-6 100*0 183 73-8 135 76 188 137 72*9 102-0 176 130 73*9 81 187 137 73-3 104-0 188 139 73-9 79 187 138 73-8 166 -100*0 123 74*1 80 180 133 73-9 105*2 169 126 74-6 80 176 130 73-9 100 0 168 126 75-0 76 .189 140 74*1 100-0 165 124 75-2 78 176 131 74-4 105-4 169 127 75-2 76 193 144 74-6 105-3 168 128 76-2 75 190 143 75-3 162 126 77-8 79 100*0 187*080 131-300 70*224 78*460 102-306 176-100 125-480 71-306 78-020 6*521 5-489 2-631 3-489 3*700 6-582 5-015 2.678 3-310 0 543(P<0.001)

Length (mm.) 187 199 192 186 193 197 188 192 190 184

Breadth (mm.) 120 130 127 124 129

Cranial index 64-2 65-3 66*2 66*7 66-8

Facial angle 74 76 77 80 78 76 77 75 78 80 78 84 76 76 76 81 75 80 75 87 84 76 78 80 81 73 73 76 75 84 85 80 83 82 75 77 81 73 75 78 80 78 79 83 78 82 78 73 83 81

Alveolar index 102-9 102-8 97-1 101*9 106-8 105*8 97 0 104-7 102 0 95-2 103*0 101-8 103*1 103*1 103-9 100.0 105*3 103*1 105*0

Length (mm.) 170 189 186 183 174 172 180 174 181 174 172 185 185 176 179

Breadth (mm.) 113 126 125 123 117 117 123 119 124120 119 128 128 122 124

Alveolar index 100 0

100 0 109 9 107-7 109-6 100 0 105-3 108*2 106*5 103*3 107-2 106-4

100*0

102-1 100 0 104-1 105-2 109-4 102*2 105-5 104-5 10141 10241 107*5 95-0 103*3 108-3 106-3 106-3 113-7 106-5 105-6 105*3 101.1 104*2 100 0 95 7 103*2 103-2

100.0,

J10-6
98-9 100*0 100*0 100 0

101*1

103*2 104-4 103*2 104-2 103-822 3-827

-0-620(P<0 001)
0-016 (n.s.) 0-046 (n.s.)

-0-619(P<0 001)

Headform and human evolution

235

indices (male: 70*2; female: 71-3) conform closely to figures given by Turner (1884), Hrdlicka (1928) and the pooled data of Martin (1928), given here in Table 6, supports the belief that the sample is representative. Herskovits (1926) believes that the correlation between head length and head breadth is an index of group homogeneity; the highest coefficient he gives is 0586 for Teneriffian males and for Batatelu. If his view is correct this group of aborigines is exceptionally homogeneous, having a length-breadth correlation coefficient of 0 530 for the male and 0 543 for the female, with the probability of less than 1 in 1000 that the result is due to chance. These data are recorded in Table 1. Throughout this paper the customary level of significance has been adopted, namely, that for a statistical index to be real-i.e. unlikely to have occurred merely as the result of chance-the probability of its occurring by chance must be jess than five times in a hundred trials (P < 0.05). The juvenile material comprised forty-two skulls, ranging from newly born to late adolescent. The ages were estimated according to Campbell's (1925) observations. The data are presented in Table 2. Information relating to the stature and cephalic index of fifty living male and fifty living female aborigines is compiled from the work of Burston (1913), Campbell & Lewis (1926) and Campbell & Hackett (1927). They represent respectively groups from the Northern Territory, South Australia and Central Australia. The figures were taken strictly as they came: all the female figures from Campbell and Lewis and from Campbell and Hackett were recorded with an equal number of the males; the totals were completed by adding to each an equal number of males and females from Burston. The results are given in Table 3, The ectocranial and endocranial indices comprise the measurements of twenty-one skulls given by Weidenreich (1945, Table 1) plus the corresponding figures for twenty others from the collection in this Department. The data, and the provenance of the skulls, appear in Table 4. For comparison with the aboriginal data measurements relating to stature and cephalic index (Table 5) and cranial index (Table 6) were collected from a number of sources, but chiefly from Martin (1928). Finally, to obtain a figure for the mean cephalic index of the newly born European the heads of forty infants were measured. One female proved to be of part-aboriginal origin and was omitted from the series. The remainder comprised sixteen males and twenty-three females; their ages ranged from 1 to 22 days and none gave any evidence of persistent birth moulding.
FACTORS WHICH MAY BE CONCERNED IN THE DETERMINATION OF HEADFORM

(1) Evolution The progressive rise in the mean human cranial index from early man up to the present is too well known to require more than brief attention; moreover, the evidence is fully surveyed by Weidenreich (1945). Outstanding features of
16-2

236

A. A. Abbie

the rise are that it seems to have been practically universal, and everywhere approximately contemporaneous. Another important point is that while the mean index started low there have always been individuals, or even small
Table 2. Cranial index, facial angle and alveolar index offorty-two Australian aboriginal juvenile skulls
Dental epoch Prior to eruption of first permanent molar
Estimated age Term 18 months 18 months 2-5 years 2-5 years 2-5 years 2-5 years 5 years 5 years 6 years 6 years 6 years 6 years 7 years 7 years 8-10 years 8-10 years 8-10 years 8-10 years 8-10 years 8-10 years 8-10 years 8-10 years 11 years 11 years 11 years
11 years 11 years 11 years 11 years 11 years 12-15 years 12-15 years 12-15 years 12-15 years 12-15 years 12-15 years 16-20 years 16-20 years 16-20 years 16-20 years 16-20 years

Prior to eruption of second permanent molar

Cranial index 80-0+ 72-8 76-7 75-3 78-0 72-0 67-4 74.1 71-8 80-4 75-1 71-1 67-4 71-4 74-8 73.5

Facial angle 83 80 91 90 82 89 83 79 83 83 84 86 80 83 78
81 77 80 85 80 82 82 84 84 79 85 80 82 84 82 81 78 80 84 80 77 84 79 76 83 82 83 (1) 84-4 (8) 81-6 (13) 81-4 (14) 80-8 (5)

Alveolar index 100-0 100-0

Remarks

Mean cranial index

Birth moulding Nor basion

94.7

Basion damaged
No basion No basion

100-0
101-2

80-0+ 74-8 (2) 73-2 (4)

73-0 (2)

102-6 102-3
102-4 102-4

No basion Prosth. damaged No basion No basion

73.5 (4)
73-1 (2)

No base 96-6 102-9 96-8 96-7 102-4 100-0 97-8 95.5 96-7 102-2 99.0 100-0 105-7 101-1 100-0 96-6 106-4 101-1 98-9 103-4 100-0 103-2 101-1 106-8 104-4 100-0 100-0 (1) 99-2 (4) 100-3 (11) 100-7 (14) 103-1 (5)

72-1 (8)

78-1 69-9
70-5 63-6 69-6 77.9 73.3 72-2 73-6 75.7 69-9 73-6 71-1 70-6 72-2 75-3 71-8 77-1 76-3 73-5 72-4 71-1 72-0 72-0 70-1 74-2

72-4 (8)
Known female

Prior to eruption of third permanent molar

Arbitrary subdivision based on skull length under 171 mm.

74-4 (6)

Arbitrary subdivision based on skull length

171 mm. or over

71-9 (5)

Means for 5-year periods

Birth 1-5 years 6-10 years 11-15 years 16-20 years

80-0+ (1) 74 0 (8) 72-6 (14) 73-2 (14) 71-9 (5)

Figures in brackets represent the number of observations upon which the means are based

groups, with an index well above the mean (Weidenreich); subsequent elevation of the mean seems to have been due rather to expansion of such groups in situ than to the introduction of a new factor.

Headform and human evolution

237

(2) Race The term 'race' is objectionable, both on politico-sociological grounds and because it is used equally for the whole human species and for sections of that species. It is hard to find an adequate substitute, but wherever practicable the term 'group' is employed here instead. The fact that isolated groups may by inbreeding attain stable and characteristic forms of head (compare, e.g. the Armenian with the Australian aborigine) is undoubtedly the basis for the belief that headform is an indication of affinity. Thence arose the view that the progressive brachycephalization in Europe betrays progressive admixture with brachycephalic Asiatics. This socalled 'substitution theory' is denied by Hug (1940) and by Weidenreich (1945). Weidenreich points out that there is no evidence that the supposed home of this brachycephaly actually harboured brachycephals at the alleged time of invasion, nor is there any evidence of contemporaneous brachycephalization in the regions over which the invaders must have passed. Moreover, the presumptive descendants of the Asiatics can have inherited only their rounder heads for there is no trace of Asiatic influence in the facial skeleton. Gerhardt (1938) denies that the cephalic index has any racial significance: he believes that brachycephaly is due to a number of factors and has often appeared independently, probably by mutation. Kappers, too, seems to favour the idea of mutation. At all events, it must be agreed that there is little justification for attributing the increasing brachycephalization in Europe, at least, to Asiatic infiltration. (3) Relation to jaw development Since the dolichocephaly of early man was usually associated with marked prognathism, and since the evolutionary record shows a progressive diminution in both features (e.g. Morant, 1935) and since, further, the Australian aborigine exhibits definite increase in both during his ontogeny (Table 2, Fig. 1), the possibility that skull length and jaw length are related cannot be ignored. It is conceivable, indeed, that the longer skull serves as a mechanical balance for the heavier jaws. In pursuing this matter it may be borne in mind that even in the dolichocephalic and prognathous negro the part of the head behind the occipital condyles weighs 18 % less than that in front of the condyles (Schultz, 1940). The possibility of some such relationship was checked in the Australian skulls, both adult and juvenile. The official measure of prognathism is the 'facial angle'-the angle the nasion-prosthion line makes with the Frankfurt horizontal. However, this angle is a measure of alveolar prognathism (Parsons, 1930) or prodenty (Weidenreich, 1945), rather than of true facial prognathism so, as a check, the 'alveolar index' of Flower (basion-prosthion/basion-nasion) was also determined wherever possible. There is (Table 1) fair correlation between the two expressions (male - 0-620; female - 0 619), with a probability of less than 1 in 1000 that this is due to chance.

238

A. A. Abbie

Any relationship between head length and prognathism should give a good correlation with either facial angle or alveolar index. In fact (Table 1), there is no significant correlation in either case or in either sex. In the adult female, indeed, although the mean cranial index is higher than in the male, prognathism as measured by either criterion is greater (Table 1, Fig. 1). The difference in facial angle is not significant, but the difference in alveolar index is, since the probability that it is due to chance is less than 1 in 20. Greater female prognathism occurs in other groups (Martin, 1928) and, as it has some significance, must be a secondary sexual character. In the present context,

--4
-

Ivi.

Age in years Adult means Fig. 1. Developmental changes in cranial index (smoothed-out curve), facial angle (broken line) and alveolar index (dots and dashes). Data: Tables 1 and 2.

this character strengthens the evidence of the correlation coefficients that there is no relationship between head length and jaw length. These two features must, then, vary independently. Such independent variation is in accord with the more general observations of Davenport (1926), Huxley (1932), Burkitt (1946) and others, but is contrary to Weidenreich's (1940) belief (p. 435) that 'Phylogenetic enlargement of the brain alone is sufficient to transform " automatically" the entire skull with all its peculiarities.' Further evidence of independent variation in the skullsome of it collected by Weidenreich himself-will be given later. In the mean-

Headform and human evolution

239

time it may be noted that the series of dogs upon which Weidenreich's views are based is far from representative. Certainly, the bulldog, King Charles's spaniel and pekinese have both round heads and short jaws, but many others in the 'dwarf' series-dachshund, miniature schnauzer, Scottish terrier, wirehaired fox terrier, cairn, sealyharm, etc.-which have short heads have decidedly long jaws. Short-headedness is correlated with dwarfism but it is not necessarily associated with shortening of the jaws.

(4) Heredity The fact that isolated inbreeding can produce a fairly stable and constant mean cephalic index, and the further fact that interbreeding between diverse stable groups introduces modifications in the descendants, indicate that hereditary transmission is a factor to be reckoned with. However, the inheritance of cranial characters seems to be a complex affair (Fawcett & Pearson, 1898; Frets, quoted by Kappers, 1945 and Weidenreich, 1945; and others), probably because so many independent variables are involved in the process that it cannot readily be expressed in simple Mendelian factors. In the case of the cephalic index, Weidenreich (1945) states that head length and head breadth vary independently, a view supported by the observations of Dunn (1928) on the cranial characters of pure and mixed Hawaiians and by those of Spier (1929) on native- and American-born Japanese. Further evidence of the complexity of this problem emerges from the measurements made by Sullivan (1919) on the 'Samar' united twins of whom Lucio had a cephalic index of 80'8, Simplicio an index of 85-5. However, while the genetics of headform are not simply resolvable, it must be conceded that within a homogeneous group under stable conditions the cephalic index remains constant for a long time, and this must be due mainly to hereditary transmission. (5) Environment Boas (quoted from Martin, 1928; Kappers, 1935b; and Boas, 1940) was probably the first to adduce substantial evidence in support of the view that change in environment may influence the headform of the new generation. The suggestion has been attacked, notably by Pearson & Tippett (1924), but has received so much independent support (see Guthe, 1918; Hirsch, 1927; Klein, 1935; Kappers, 1935b, 1945) that it must be taken seriously into account. To summarize briefly: the observations indicate that those with a high mean native cephalic index tend to show a fall in the first American-born generation; those with a low index tend to show a rise; intermediate groups betray little change. The best substantiated example of a significant fall is in Jews-from a native cephalic index of 83 to an American-born one of 81-4; an example of a rise is in Sicilians-from a native mean of 77-7 to an American-born mean of 81-5. Swedes, Scots and Italians, with native indices near 80 vary little, if at all. Similar observations are recorded by Klein (1935) on Askenasim

240

A. A. Abbie

Table 3. Stature, head length and breadth, and cephalic index of fifty male and fifty female Australian aborigines. Data: Burston (1913), Campbell & Lewis (1926), Campbell & Hackett (1927)
Male

Female

Stature (mm.) 1538

1555 1567 1587 1590 1591 1597 1602 1611 1616 1616 1626 1627 1635 1639 1644 1647 1648 1651 1651 1657 1663 1664 1671 1673 1674 1686 1689 1693 1693 1698 1705 1707 1708 1724 1730 1730 1731 1733 1740 1747 1761 1762 1768 1772 1779 1830 1833 1875 1876 1684-200 78-120

Head length (mm.)

Means Standard deviation Corr. coeff. stat./H.L. Corr. coeff. stat./H.13. Corr. coeff stat./C.I.

187 181 185 188 190 185 142 142 199 186 145 184 142 196 139 189 146 193 137 200 143 192 132 195 145 187 142 185 139 172 128 202 147 198 146 148 195 186 140 196 144 195 146 209 148 194 146 182 148 197 150 192 136 190 143 189 140 180 145 142 190 194 147 195 136 205 143 188 140 203 143 199 144 197 153 197 137 199 136 199 145 197 146 193 137 199 146 196 145 201 130 195 136 201 146 192-940 141-940 7-150 5-121 0-456 (P<0-01) 0-015 (n.s.) - 0.373 (P < 0.01)

Head breadth Cephalic Stature index (mm.) (mm.) 142 1426 75 9 142 78*5 1459 140 75-7 1463 138 1475 73-4 134 1487 70 5

Head length (mm.)

181 142 179 193 133 176 140 171 125 76-7 1487 187 143 71-4 173 125 1508 174 78-0 1510 130 77-2 1510 177 133 70 9 1510 176 136 77-3 1512 181 134 1516 182 71-0 129 71-5 1520 182 126 68-8 1522 184 136 74-4 1525 175 138 75*9 1531 175 135 75-1 1533 175 134 74-4 1537 192 130 72-8 1538 186 135 73-7 1539 174 124 75 9 1540 178 136 1542 75-3 188 135 1548 73-5 186 131 74 9 1550 185 137 70-8 1553 175 138 75-3 1555 188 137 81*3 1555 173 131 76-1 1559 188 137 70-8 1560 184 138 75-3 1561 178 128 1561 74-1 188 138 1562 80-6 183 136 1563 181 74*7 132 75-8 1564 188 137 1564 69*7 183 138 69-8 1567 188 130 74-5 186 1580 137 70 4 1580 178 136 72-4 1582 182 142 77*7 1589 173 142 69-7 1592 179 132 68-3 1593 180 132 72-9 1594 182 132 74-1 1606 180 132 71-0 1609 192 145 73-4 1620 174 134 74*0 1622 192 136 64-7 1628 185 132 69-7 1644 174 132 72-6 1722 187 134 73-648 1551-460 181-420 134-380 5-952 4-646 3*233 51*504 0-161 (n.s.) 0-053 (n.s.) - 0-086 (n.s.)

Head breadth Cephalic (mm.) index 134 74'0

79.3

68-9 79-6 73-1 76-5 72-5 74-7 75-1 77-3 74*0 70 9 69*3

73.9

78-9 77-1 76-6 67-7 72-6 71-3 76-4 71-8 70 4 74*1 78-9 72-9 75-7 72-9 75 0 71-9 73-4 74-3 72-9 72-9

75*4
69-2 78-0 82X1 73-7 73-3 72-5 73-3 75-5 77 0 70-8 71-4

73.7 76*4

75.9

71-7 74-134 3*002

- Headform and human evolution

241

Jews born in Amsterdam and by Spier (1929) on Japanese born in America (calculated adult mean). The outcome of the mass of confusing, and sometimes conflicting, evidence which has been published on this subject seems to be that a change in environment-a change usually for the better-may be sufficient to jolt the hereditary mechanism out of its accustomed groove. The alteration effected is not uniform: high indices are downgraded, low indices are upgraded, those in between are unchanged. Thus all tend to converge upon a common mean within a restricted range of about 80-82. Much remains to be done on this subject but the results to date suggest that a favourable environment induces a trend towards a common headform with an index at the lower limit of brachycephaly.
(6) Relationship of ectocranial index to endocranial index Weidenreich (1945) could find no correlation between these indices in a series of twenty-one skulls he had measured. Harris (1926) records a similar discrepancy in gorillas. This line of investigation seemed worthy of further pursuit, however, and Weidenreich's series was extended by the addition of such data on twenty skulls in this Department. The information is given in Table 4. The figures for the two indices were then plotted (Fig. 2)-the ectocranial indices in ascending order, the corresponding endocranial indices on the same scale but in a separate curve. As is to be expected, the endocranial curve-despite occasional fluctuations-is generally above the ectocranial curve; and the discrepancy between the two diminishes fairly constantly from the longest skulls to the shortest. Most interesting, however, is that through a limited ectocranial range of about 79'8-82 the two indices conform much more closely than anywhere else in the series. While the number of observations is relatively small, and the concordance is far from exact, there is a clear indication that within this range the skull form approaches most closely to that of its contents.

(7) Relationship of cranial index to cranial capacity from Apart cranial index, capacity depends upon such factors as height of skull, thickness and rounding of the bones, etc. and lies rather beyond the scope of this inquiry. Weidenreich (1945) denies any correlation between headform and capacity and quotes a number of examples to support his view; at the same time, there is some hint of such a correlation in the work of Bolk (1903) and Schwerz (1912). These workers both found that in a series of skulls the mean cranial capacity tends to a maximum in the upper mesocephalic range. (8) Stature A generalization known as 'Ammon's law' (Kappers, 1945) postulates a negative correlation between stature and cephalic index-the greater the stature the lower the index, i.e. the longer the skull. This has been confirmed by Pittard (1905), by Boas (1940) and, with some difficulty, by Kappers (1945).

242

A. A. Abbie

The correlation was checked on the aborigine data collected by Burston (1913), Campbell & Lewis (1926) and Campbell & Hackett (1927). The results (Table 3) show that for the male aborigine the correlation coefficient for stature and cephalic index is -0-373; this is not a high correlation but, since the probability that it is due to chance is less than 1 in 100, it is significant. Similar
Table 4. Ectocranial and endocranial indices of forty-one skulls; twenty-one measured by Weidenreich (1945), the remainder measured for this investigation
Skull H. 8oloe8108, V African African Australian aborigine Australian aborigine Australian aborigine African Chinese Sinanthropus, X Sinanthropus, XII H. rhodesienasi Sinanthropus, III Sinanthropus, XI African H. neanderthalensi8
Ectocranial index 66-3 68-5 69-2 69-5 70-0 70-3 70-4 70-5
71-8 72-3 72-3 72-8

Endocranial index 76-2 70-0 72-0 74X4 73X1 79 5 71-8 74-3 74-2 76-8
78-8 78-3 76-7 71-7 78-6 78-8

Authority Weidenreich This paper

Weidenreich This paper ,, Weidenreich

72X9
73-5 73-8 74-0 74-5 74-7 75-0 75-1 75-2 76-3 76-7 77-1 77-1 77-2 77-4 78-9 79-6 79-8 79-8

EkhBM , V African La ChapeUfe Tamil Chinese European Pithecanthropus, II European Tabun, I Gibraltar European European European African F.T. European foetus African Chinese European European Chinese European European Chinese Chinese-Tamil cross European

Ehringsdorf

80-0 80-3 80-7 81-8 82-6

82-8 85-9 86.9 93-4 94-6

80-3 79-5 79-0 73-5 75-6 79-8 81-8 80-4 80-8 83-9 76-3 81-4 86-7 79-6 81-8 79-6 81-6 80-2 81-8 84-4 86-7 88-1 92-0 95-1 95-0

This paper Weidenreich ,, ,, This paper Weidenreich This paper ,, Weidenreich , This paper Weidenreich ,,
,

This paper Weidenreich This paper

Weidenreich This paper ,, Weidenreich This paper ,, Weidenreich

correlations were determined for stature/head 'breadth and stature/head length: there is no significant correlation with head breadth but the correlation with head length is of the same order as that with cephalic index. The generalization holds weakly for the aboriginal male, then, and depends mainly upon head length. In the aboriginal female, however, there is no

Headform and human evolution

243

significant correlation between stature on the one hand and head breadth, head length or cephalic index on the other (Table 3). Kappers (1945) was equally unable to establish any correlation between stature and cephalic index for the females of other groups. Turning to mankind in general, similar calculations were made for the males of fifty other groups and the females of thirty-eight other groups (Table 5). In 969594-

93-

92t
91

90#
89-

88.
8786 8 85
Ca

i*

~8483
82-,
80 9 79.0
.*4 *99

21

~77

678-

AX 'r

76-

751
72- J/ 66 71 I 70 * 696867-

73-',~~~~~~~~~

661

Fig. 2. Comparison of ectocranial.indices (continuous line) with corresponding endocranial indices (broken line). Ectocranial indices in ascending order. Data: Table 4.

each case the figures quoted are group means derived from what seem to be the most reliable sources. In the males the correlation coefficient is 0-211, in the females 0-1321. Neither figure is statistically significant although the male correlation remains higher than the female, and the sign of the coefficient has changed-this, if it means anything, indicates some tendency for increasing stature to go with rounder-headedness, a tendency in harmony with what

244

A. A. Abbie

appears to have been the course of human evolution generally (see Le Gros Clark, 1945). In any case, it is clear that 'Ammon's law' has significance only within a homogeneous group, a fact already insisted upon by Pittard (1905) and Boas (1940). The 'law' does not apply to females, probably because the shorter growing period does not permit such full skull differentiation as in the male: as Boas (1940, p. 48) points out, '...the most characteristic forms of each type are found in the adult male'. In the case of the skull it is clear that the additional lengthening upon which 'Ammon's law' is based is achieved late, during the few years of further differentiation after female growth has ceased. Since increasing stature is, in general, associated with increasing head length it is to be expected that dwarf types, in whom differentiation stops prematurely, will be rounder-headed than those more fully differentiated. This is clearly so in the observations made on dogs by Weidenreich (1940) and in the human pygmy material collected by Kappers (1939, 1945), and the mean cephalic index for the pygmies (stature less than 1500 mm.) in Table'5 is 80-9 (range 79.0-8380). In this connexion the females of the taller groups must be regarded as modified dwarfs. However, the association of rounder heads with shorter bodies is only part of the story, as Table 5 shows. For example, the Northern Andamanese, with a mean stature of only 1486 mm., have the same mean cephalic index (82.5) as the Sara who have a' mean stature of 1817 mm.the highest recorded. Moreover, the whole story of human evolution associates rounder-headedness with increasing stature. Thus, there are two kinds of brachycephaly: one, characteristic of infants, dwarfs and females, represents incomplete general differentiation; the other, associated with tallness, expresses the human evolutionary trend. This was recognized by Macalister (1898) who called the two kinds primary and secondary brachycephaly respectively. Failure to appreciate this distinction has led to much unnecessary confusion.

(9) Sex It is well known that females are usually rounder-headed than the males in the same group. This observation rests on ample evidence and depends, as already shown, upon the shorter period available for female differentiation. However, the mean male index does sometimes exceed that of the female. Martin (1928) doubts the accuracy of observations of this kind, and the only exception admitted by Kappers (1945) occurs in Armenians. At the same time, there is little justification for doubting figures simply because they conflict with an accepted hypothesis and the data employed here (Tables 5 and 6) are taken just as published. At the outset, the relationship was tested on the skulls of Australian aborigines, fifty of each sex (Table 1). The mean index for the female (71.3) exceeds that for the male (70.2) by 1 1 a significant difference, for the probability that it is due to chance is only 1 in 20. The individual values were then plotted,

Headform and human evolution

Table 5. Stature and cephalic index of males of fifty peoples and offemales of thirty-eight peoples
Male
A

245

Female
A

People Veddah Buduma Kadir Australian Aborigines Bedawin South African Bastards American Eskimos
Ainu

Polar Eskimos Senoi Batwa Mataco M'Baka Chinese Kamschatkans Tunguse Semang Ituri
Ostiaks

Cephalic index 70 5 72-5 72-9 73-7 75 0 75-8 77*0 77-3 78-0 78-0 78-1 78*1 78-1 78-3 78-5

Mawamba Pygmies Tapiro Fuegians

Shoshoni

Fan Chiriguan Aueto Jukagires Asiatic Eskimos Danes Kalmuck

Ossetes

Lithuanians White Russians Polish Jews Aeta Tschuktsches Sara Northern Andamanese Polynesian Jakutes Southern Andamanese Balkan Tatars Hawaiians Samoyed Pamiri Tehuelche Armenian Turki Kirgish Californian Indians Means Standard deviation Corr. coeff. stat./C.I.

Stature (mm.) 1533 1755 1577 1684 1660 1684 1577 1567 1570 1520 1522 1638 1671 1639 1601 1565 78-7 1528 79 0 1390 79*0 1563 79*2 1408 79 5 1449 79.5 1577 79.5 1661 79*5 1698 79.5 1634 80-2 1581 80-2 1560 80-4 1673 80-7 1691 80-7 81-1 1634 1690 81-5 81-5 1656 1652 81-5 1610 81J8 1465 82-0 1622 82-0 1817 82-5 1486 82*5 82-6 1720 1624 82*7 1482 83.0 1657 83-3 1713 83-4 84*4 1550 1683 85-0 1780 85-0 1685 85-8 1675 86-0 1651 89-4 1709 89-7 1614-74 80*282 91*37 3-836 0-211 (n.s.)

Cephalic index
-

Stature (mm.)

72-3 74-2 744

76-7 74*5 78-4 77-4
-

1620 1430 1552

-

1570 1497 1471 1450

Source of data Haddon (1929) Martin (1928) ,, Table 3 Haddon Martin

77-7 77-8 77 9
-

1437 1529 1568

1499 1465
-

Haddon Martin
,

77-4 79 3

,, Haddon Martin ,, Haddon

-9

79 0 77-6

1441 1356

78-9 79.5 80-5 81-4 81*8 80*0 79*7 81-5 82-6 81-9 82-0
82-2 82-9 81-8 80-0 81-9
-

1473 1528 1589 1517 1521 1470 1518 1592 1598 1573 1546 1544 1506
-

Martin ,, Haddon Martin ,,

, ,

9, ,, 9,
,

,, ,,

1520 1676 1385

-

80*9 82*7 83-5 84-2 83-9

-

1512 1402 1546 1626 1430

,, ,, ,, Haddon Martin ',, ,, Haddon Martin

,,
Dunn (1928) Martin Haddon , Haddon Martin

9,

84-1

1529
-

86-3 89-9 80-22 3*536 0-1321

1503 1580 1514-97 69-74 (n.s.)

,,

246

A. A. Abbie

Table 6. Cranial indices of males and females of sixty-six peoples. Data from Martin (1928) except where stated otherwise
Cranial index People Pericues Torres Straits Fidschi Islanders Loyalty Islanders
Australians Babase

Punjabi Tamils Duke of York Islanders Papuans Ambitle

Veddahs

Male 66-1 68-3 69-6 69-8 70-3 70.4 70.7 70-8 70 9 71-0
71*1 71*2 71-3 71-3 71-4 71*5 72-5 72-5 72*9 73-2 73'2 73-4 73-7 73-7 73-9

Eastern Eskimos New Caledonians

Hottentots Greenland Eskimos

Sinhalese
Kaffirs

New Irelanders Naqada Cuenca Tutan Germans Friterpen Tasmanians

Female 68-5 70 3 68*9 71*8 71'6 71-6 72-3 70 3 71-8 73*0 71*1 72-7 71-4 74-4 74.9 71*9 72*0 72-9 75-1 74*6 75*0 76-3 74-1 74-5

Remarks

Martin,

mean

of three figures

Hill (1941)

Martin,

mean

of two figures

75.9
74.9 75-6 75 0 76-5 75-0 72-9 73.9 750
78-2 76-5 77.4 7741 78-4 77.5 77-2 77-6 77-8 79 0 77-5 78*9

Eskimos Portuguese Micronesians Eastern Tschuktsches English Ama Zulus Maoris Anglo-Saxons Marquesans Thebans
Tahitians

Groterpen Hawaiians Teneriffians

Ainu

Quipuzcoans Fuegians Friesians Californians

Guanches

Cretans Japanese Gallier Etruscans Asiatic Woguls Prussians Kanakas Old Pompeiians Andamanese Italians Alsatians
Bashkires

74-2 74-3 74.4 74-4 74-7 74.8 74.9 75-0 75 0 75-1 75-1 75-4 75-5 75.9 76-0 76-2 76*9 77 0 77.3 77-3 78-0 78-3 78-4 78-5 78-8 79-2 79-3 80-3 80-6 80-8 80-8 81-0

Mean, Wunderly (1939), Hrdlilka (1928)

Mean, Martin, Turner (1884)

80-9 79.7 75-4 78-2 78-4 78-9 80-3 81-0 82-7 80-0
81-2 81-0

Headform and human evolution

Table 6 (continued)
Cranial index
A

247

People Javanese Kalmucks Romanians

Wurttemburgers Bohemians Bavarians Swiss Tirolese Telengetes Means

Male 81-5 81-6 82-3 82-6 83-1 83-8 84*4 85-0 86-0 75-73

Female 80-9 81-0 83-2 82-7 83-5 84-2 84-7 84-9 87-3 76-64

Remarks

Martin, mean of two figures

each sex in a separate curve but to the same scale, in strictly numerical order (Fig. 3). Within the range of these observations (64.2-77.8) the curve for the female lies consistently above that for the male. Taken in conjunction with similar findings on aborigines by others, this indicates that the mean female cranial index really does exceed the-mean male index. The matter was then put to a general test by taking the mean male and female cranial indices of sixty-six other groups (Table 6): the difference between the mean female figure (76.64) and the mean male figure (75.73) is 09-quite comparable with that for the aborigines. Then similar curves for the two sexes were plotted. In the first place the male figures were plotted in strictly ascending order with the corresponding females in a separate curve on the same scale (Fig. 4). Despite some fluctuations, the female curve runs mainly higher than the male, the disparity diminishing progressively from the bottom to the top of the series. However, over the limited range of 78-82 the female fluctuations virtually cancel out and the two curves may be considered practically coincident. In this respect the similarity between this curve and that comparing ectocranial and endocranial indices (Fig. 2) is quite striking. To emphasize the sexual distinction the male and female figures were next plotted, independently of any racial affinity, each in strictly numerical order (Fig. 5). Now most of the fluctuations disappear and the female curve is more consistently above the male curve; there is still a progressive convergence from bottom to top, and the tendency to coincide in the lower 80 range is preserved. Comparison of Fig. 3, which represents individual aborigines, with Fig. 5, which represents means for different groups, reveals that the distinction really is one of sex: the aboriginal curve might very well have been taken from the lower part of Fig. 5, the indices simply do not go high enough to disclose the convergence found in the more extended series. Elimination of the sexual distinction over the 78-82 range can be brought out more clearly, perhaps, by isolating this group and comparing the frequency with which the female exceeds the male and vice versa with the corresponding frequencies above and below this range (Table 7). It will be seen that while above and below the 78-82 range the female exceeds the male in over 85 % of

248

A. A. Abbie

cases, within this range there is no difference between the sexes. Even allowing for the small number of observations, the correspondence at 78-82 is very striking. A point of interest is that skull measurements illustrate the sexual difference much better than do head measurements. In the case of the cranial index of the aborigine (Table 1) there is a significant difference of 1 between the sexes, but with the cephalic indices (Table 3) the difference between female mean (74.1) and male mean (73.7) is only 0*4, which is not significant. Also, taking mankind in general (Table 6), the mean cranial index for females exceeds that for males by 0 9-comparable to the difference found in aborigines; but with the cephalic index (Table 5) the female mean (80.2) is practically the same as
79 78 77

76/
75-

696867-

K__

66

64-5
Fig. 3. Cranial indices of fifty male (continuous line) and fifty female (broken line) Australian aborigines. Each point represents an individual; each sex plotted in ascending order. Data: Table 1.

the male mean (80.28). It seems probable that reduction of the sexual difference in whole head measurements depends upon greater development of the temporal muscle in males, for Schultz (1940) found that such development in older chimpanzees adds considerably to the breadth of the head. In any case, the observation justifies Morant's (1928) preference for purely skeletal measurements. As already noted, the generally rounder head in the female simply reflects a shorter period of differentiation. This, however, does not apply to such secondary sexual characters as prognathism for, according to Boas (1940), secondary sexual characters are present in childhood. Moreover, Huxley (1932) and Reeve & Huxley (1945) have shown that in other animals secondary sexual characters with a high growth ratio may impose a localized distortion upon the total growth gradient.

Headform and human evolution

249

To summarize: the female is generally more brachycephalic than the male, a distinction which appears better from skull than from head measurements. The rounder female head is an example of primary brachycephaly-relative dwarfism or incomplete differentiation and, unlike prognathism, is not a secondary sexual character. The difference tends to vanish over a limited cranial index range of 78-82.

(10) Individual development The fact that the child's head is generally rounder than the adult's is recognized in the term 'infantile' commonly applied to those with a high

A
'I

!Ii\

I.
0.

Fig. 4. Cranial indices of males (continuous line) and corresponding females (broken line) of sixty-six peoples. Each point represents a group mean. The males are plotted in ascending order. Data: Table 6.

cephalic index. However, while the rounder head is more like the infant's in many groups it is not so in all, nor is the process of head development from infancy to adulthood always comparable from one group to another. The aboriginal skull collection in the Adelaide Museum contains one of a newly born foetus which, unfortunately, retains a good deal of birth moulding. As it is, the skull has a cranial index of 86, but even the most generous allowance for the distortion leaves the index still over 80. Starting from this point (Table 2), the index drops quickly to the lower 70's during the first few years 17 Anatomy 81

250

A. A. Abbie

(Fig. 1), and then more slowly until about the age of 10; after a minor rise in early adolescence there is a progressive decline to the adult mean. Earlier cessation of growth puts the adult female mean naturally on the curve some way in front of the adult male mean. This is the type of curve usually visualized for dolichocephalic people. However, in groups where the mean adult index is high, e.g. in the upper 80's, the curve of skull maturation is different. In these cases, according to Ranke (1905), Martin (1928), Klein (1935), Kappers (1945) and others, the cephalic index of the newly born is in the lower 80's, being less than the adult mean. Now, in development, there is a rise during the first 2 or 3 years to a level above the adult mean; after that there is a gradual fall, often with a slight elevation in early adolescence, to the adult standard. Thus, the mean cephalic index at birth is much the same for both the dolichocephalic and the brachycephalic: compare, for example, the cranial index of just over 80 for the aboriginal foetus with the mean cephalic index of 81V1 for the newborn of a predominantly brachycephalic group (Ranke, 1905). Aeby (quoted by Vogt, 1864, p. 30) confessed himself unable to distinguish the skulls of foetal negroes and Europeans, Macalister (1898) drew attention to the similarity of the foetal skulls of British, Egyptians and Hindus, and other examples could be quoted. It seems clear that no matter what may be the mean cephalic index of the parents, the mean index of the newborn falls within a very restricted range which lies close to the lower limit of brachycephaly. This view was checked on the heads of thirty-nine European infantssixteen males and twenty-three females, aged from 1 to 22 days and without head moulding. The mean cephalic index for the males was 78-39 (range 74-3-82.6), the mean for the females was 78-4 (range 73.684.2). Sexual difference is negligible and the combined mean is 78f4. To this may be added a mean of 82f0 for five full-time, undeformed formalin-fixed European foetuses. In addition, Ranke (1905) has published figures for twenty European infants. (age 1-21 days): eleven males gave a mean index of 81-5, nine females a mean of 80-7; again the sexual difference is slight and the combined mean is 81-1. The mean of the sixty-four observations on Europeans recorded here is 80-5. This may be compared with the cephalic index of 76-6 (well within the European range) of the part-aboriginal infant omitted from the above series, with a cranial index of 80 + for the full-time aboriginal skull and with a mean index of 81-1 for three European foetal skulls. These observations all combine strongly to support the view that in the matter of headform there is a foetal norm, restricted to a small range of indices about the lower 80's, which seems to be common to the most diverse groups and to both sexes. This may be considered the primary norm for the whole human species. Whether or not this holds for other primates is yet to be

determined.
It appears further that where the mean adult index lies clearly below or above the primary norm a major characteristic deviation in the adult direction

Headform and human evolution

251

occurs during the first 2 or 3 years of post-natal life. This establishes a secondary, or group, norm from which there is a gradual decline to the adult mean. Skull differentiation thus shows four distinct stages: a common human stage at birth, a characteristic group stage at 2-3 years, an adult female stage in late adolescence and an adult male stage at the end of growth. In other words, common human heredity determines headform" until birth, specific group heredity comes into play during the first few years, and the sexual factor determines the time at which skull differentiation ceases. Davenport (1926) has drawn attention to a tremendous acceleration of growth in the period immediately after birth-this appears to coincide with the attainment of
88-

878685848382

8180-

797877767574-

73_
72. 71-

/0,

7069- ,
68-

6766
Fig. 5. The same data as in Fig. 4 but the two sexes are now plotted independently, each in ascending order of cranial index. Males (continuous line), females (broken line).

specific headform; he also mentions a lesser acceleration during early adolescence which agrees with the observed minor increase in cephalic index at the same time. At all events, specific group headform is already determined by the third year and the change thereafter may not be particularly striking. This is probably why Pearson & Tippett (1924), whose series started with 5-yearolds, could find no significant change in headform during ontogeny. It' must be remembered, too, that even the changes of the first 3 years will be clearly marked only in groups which show a decided adult deviation from the primary norm, and it is evident that while the term 'infantile' can be applied to the
17-2

252

A. A. Abbie

rounder-headed members of dolichocephalic groups it should not be used for the rounder-headed of brachycephalic groups where the infant's head is longer than the adult's. Finally, since the adult female clings to a juvenile form of skull it is easy to see why disparity with the male disappears when the adult male index falls within the foetal range of 78-82. All this makes it clear that both the long headed and the short headed are divergent from the primary human norm. In other words, they are both equally specialized, or differentiated-but in opposite directions. Wood Jones (1931) has pointed out that many of the characters which early man shares with apes and monkeys, usually designated 'primitive', really represent differentiation away from the primitive form as found in the foetus, and he calls them 'pithecoid specializations'. These include dolichocephaly. Actually, a similar observation was made long ago by Aeby (quoted by Vogt, 1864, p. 31): 'I have

Table 7. Analysis of data in Table 6 to show the distribution of female-male excess of cranial index in the 78-82 group as compared with the groups above and below this range
Range 660-77-9 78.0-81-9 820-87-9 Totals
Female exceeds male
38 (84.44%) 6 (42.86%) 6 (85.71%) 50

Female equals male

2 (4.44%) 1 (7-14%) 0 (-) 3

Male exceeds female 5 (11-11%) 7 (50.00%) 1 (14.3%) 13

Totals 45 (99.99%) 14 (10000%) 7 (100-00%) 66

Range

78-0-81*9
82-0-87-9 Totals

66-0-77-9

Further analysis of same data Male Female equals or exceeds exceeds male female 40 (88 90%) 5 (11-10%) 7 (50.00%) 7 (50-00%) 6 (85.70%) 1 (14-30%) 53 13

Totals 45 (100-00%) 14 (100-00%) 7 (100-00%) 66

already elsewhere drawn attention to the similarity of all foetal cranial forms. I am now prepared to lay it down as a general law, that a cranial form occupies a higher rank accordingly as it advances by uniform peripheral development from the foetal form; and that it stands lower, accordingly as the growth is confined to certain directions and points. From this point of view, the narrow skull must be considered as a lower type.. ..' If 'higher' is replaced by 'primitive' or 'undifferentiated', and 'lower' by 'specialized' or 'differentiated', Aeby's conclusion is virtually the same as that reached here. And he continues (loc. cit): 'We will not leave unmentioned that possibly the same position may be assigned to the most decided broad skulls.' This could be considered an anticipation of the views presented here regarding the higher grades of brachycephaly. The truly primitive skull-the primary, or foetal, type-has an index close to the lower 80's, at the lower limit of brachycephaly as usually defined. The

Headform and human evolution

253

adult skull may not differ much from this, or it may diverge, becoming either longer or shorter. Thus adult skulls fall into three groups: dolichocephalic, foetal and brachycephalic. The index range called 'mesocephalic' is only a degree of dolichocephaly and has no biological significance. 'Mesocephaly' could be applied to the only truly intermediate range- the foetal standard of 78-82, but long usage otherwise would render such change impracticable. The term is better abandoned: then, brachycephalic skulls are those with an index over 82, dolichocephalic skulls have an index under 78, while skulls with an index between 78-82 should be designated 'paedocephalic'.
(11) Independent variation in the skull Evidence has already been adduced to show that the jaws and the cranium vary independently. To this may be added Begg's (1939) finding that jaw size and tooth size are inherited independently. Other examples of independent variation are head length and head breadth (Weidenreich, 1945) and, probably, head height as well (see Macalister, 1898). Attention may now be drawn to the probability that the front and the back of the cranium also vary independently. A dolichocephalic skull such as that of the Australian aborigine exhibits frontal characters which are obviously far differentiated from the foetal type; brachycephalic skulls, on the other hand, tend to retain the foetal form in this region. But at the back of the skull things are otherwise. Here the dolichocephalic skull protrudes to an opisthocranion well behind the inion, and the expanse from foramen magnum to inion is long and comparatively flat; this is as in the foetus. In the highly brachycephalic skull, however, the maximum occipital protuberance usually coincides with the inion-whence the back ascends abruptly to the vertex, and the expanse between foramen magnum and inion is comparatively short and strongly curved; this is quite unlike the foetus. Thus, the dolichocephalic skull differentiates by frontal extension, the brachycephalic by occipital shortening; there may be all intermediate grades but there is no apparent correlation between the two.
DISCUSSION

It has been shown that headform in various groups is influenced by evolutionary trends, by heredity and probably environment, to some extent by stature and sex, and by individual development. There may be a relationship between headform and cranial capacity, but none could be found between headform and size of jaws. There is substantial evidence that some parts of the skull vary independently. This survey has uncovered an interesting series of coincidences. The discrepancy between ectocranial and endocranial index is at a minimum over a limited range of 79-82. The sexual disparity in cranial index tends to disappear over a similar range of 78-82. In a favourable environment dolichocephals and brachycephals seem to converge upon a common mean of 80-82

254

A. A. Abbie

in the first new generation. It is possible that the maximum cranial capacity lies close to the same range. And, finally, the mean cranial or cephalic index in the newborn of any group-irrespective of parental index-lies close to 80. These points have all been discussed in the appropriate place; put together thus they emphasize a common norm of 78-82, the point from which skull differentiation sets out and towards which, it seems, evolution is tending more and more to return. All this, together with the increasing persistence of such characters as metopism (Bolk, 1921), thinning -of cranial bones and delay in suture closure (Weidenreich, 1940), might well be considered an example of Bolk's 'foetalization '-a progressive tendency to cling longer to the undifferentiated condition of the foetus. However, even if the foetal shape is retained, the skull does show quite a considerable change in size. Males within a homogeneous group show some negative correlation between stature and cephalic index. This correlation does not extend to females or to mankind in general, nor is it evident in human evolution which associates increasing stature with progressive shortening of the head. Increasing stature differentiates away from the foetal norm and is a foetal tendency only to the extent that it may betray persistence of the foetal habit to increase in size. On the other hand, expanding bodily bulk has characterized the history of most of the successful animals for which there are adequate records. Such progressive increase has been attributed by some to 'orthogenesis', but since the existence of this agency has been called into question it seems preferable to seek a more satisfactory explanation. Selection is unlikely to preserve either foetalization or orthogenesis for its own sake: some advantage must accrue and it is necessary to find where this advantage lies. A clue appears in the relation between ectocranial and endocranial index; it will be recalled that the disparity between the two is largely eliminated over the foetal index range. In the foetus the skull is little more than an additional cerebral membrane, conforming very closely to its contents. Adult skulls which preserve the foetal index similarly conform more closely to the brain than do skulls elsewhere on the index scale. The fetal pattern, then, provides the most economical kind of brain protection-a thin rigid cover with the least expenditure of material. The further possibility that this pattern affords the greatest capacity should also be taken into account. The advantages of retaining the foetal form of skull are manifest; the fact that this entails retention also of such other foetal characters as metopism, thinner bones and delayed suture closure may enhance the advantage or may be purely incidental. Certainly, persistence of these features does not contribute to further cerebral expansion, for von Bonin (1934) has shown that phylogenetically the human brain has betrayed no increase in size for a very long time. According to von Bonin the size of the brain has become stable. lHe believes that increasing brachycephalization connotes internal reorganization, and that this has positive selective value. Here it is suggested, instead, that the factor which has selective value is the progressive tendency to maintain a form

Headform and human evolution

255

of skull which may afford the maximum. capacity in the most economical fashion. It seems that while the mean human cephalic index shows an evolutionary rise, this does not imply that evolution has an extreme grade of brachycephaly as its ultimate aim. Excessive brachycephaly is as aberrant and specialized as excessive dolichocephaly and, under favourable circumstances, shows a tendency to drop to a lower grade. The thesis advanced here is that the whole evolutionary trend is towards securing a mean common index of 78-82. This in itself would ensure a rise in mean index for the mean of all the data in Table 6 is still only 76-64 for the female and 75-73 for the male, and the'number of dolichocephalic groups greatly exceeds the number of brachycephalic. The advantages of preserving this form of skull have been fully discussed.
SUMMARY

1. The problem of headform has been considered in relation to evolution, race, jaw development, heredity and environment, internal form and capacity, stature, sex, and individual development. Data on the Australian aborigine were employed as a check on as many aspects of the investigation as possible. 2. The evolutionary record shows a progressive increase in mean stature together with a progressive rise in mean cephalic index. 3. Headform is a feature of race only because isolated groups, inbreeding for long periods, attain stable and characteristic means. 4. There is no significant correlation between headform and size of jaws in either sex. However, the rounder-headed female is significantly more prognathous than the male, and this is a secondary sexual character. 5. Headform is probably determined by so many independent variables that solution on simple Mendelian lines is very difficult. Since stable forms are preserved by isolated groups, however, hereditary transmission must play a part in maintaining them. 6. The available evidence suggests strongly that change to a more favourable environment tends to bring all indices to a common mean of 80-82. 7. The endocranial index is usually higher than the ectocranial, particularly in long skulls, but the difference practically disappears over a range of 79-82. 8. There is some suggestion that maximum mean cranial capacity is associated with a mean index just under 80. 9. There is a moderate negative correlation between stature and cephalic index for adult males in homogeneous groups. Females do not differentiate far enough to attain this correlation, nor is it found in mankind generally or in human evolutionary history. 10. Brachycephaly is found to be of two kinds: primary brachycephalycharacteristic of infants, dwarfs and females-is an expression of incomplete differentiation; secondary brachycephaly-associated with increasing staturereflects the trend of human evolution as a whole. 11. Incomplete differentiation usually leaves the female skull rounder than

256

A. A. Abbie

that of the male, particularly in long-headed peoples. The disparity practically disappears in the 78-82 index range. 12. Irrespective of the shape of the parental head, the newly born have a common mean index range of 78-82. Dolichocephalic groups show a fairly progressive fall to the adult mean; brachycephalic groups show a rise during the first 2 or 3 years to a level above the adult mean, and fall progressively thereafter. 13. Four stages in skull differentiation are observed:, (a) at birth, the primary-or universal human-mean of 78-82; (b) during the first 2 or 3 years, a secondary-or specific group-mean; (c) in late adolescence, the adult female mean of incomplete group differentiation; (d) finally, the adult male mean of complete group differentiation. 14. Sexual skull disparity disappears in the range 78-82 because here the male approaches the 'infantile' condition retained normally by the female. 15. The foetal form of skull conforms most closely to the brain; the difference between ectocranial and endocranial index disappears over the foetal range for the same reason. 16. Adult human skulls may retain the foetal form or they may become either longer or shorter. The long skulls and the short skulls are both equally specialized away from the primitive foetal form. 17. The proper index grouping for skulls is: brachycephalic-over 82; dolichocephalic-under 78; the intermediate, or foetal, range of 78-82 should be called paedocephalic. This grouping has a definite biological meaning. 18. Cranial evolution is not tending towards a high degree of brachycephaly as such. The trend is more and more to preserve the foetal mean index of 78-82, and this in itself is sufficient to raise the whole human mean. This trend is not an example of 'foetalization' per se but reflects an evolutionary urge towards the form of head which seems to offer the greatest capacity at the minimum cost.
I am indebted to the Director of the Adelaide Museum for affording access to its unparalleled collection of aboriginal skulls; to the Director of Obstetrics, Adelaide University, and the Matron of the Queen Victoria Maternity Home for their co-operation in securing the head measurements of the infants; and particularly to Mr Cornish, Lecturer in Statistics, Adelaide University, for the time he has given to analysis of the data.. To all of these I wish to express my

deepest gratitude.
REFERENCES BEOG, P. R. (1939). The evolutionary reduction and degeneration of man's jaws and teeth. Proc. 10th nat. Congr. Au8t. dent. As8. pp. 642-654. BOAS, F. (1940). Race, Language and Culture. New York: Macmillan. BOLK, L. (1903). Kraniologische Untetsuchungen hollandischer Schadel. Z. Morph. Anthr. 5, 609-697. BOLK, L. (1921). tiber Metopismus. Z. Morph. Anthr. 21, 207-226.

Headform and human evolution

257

BONIN, G. VON (1934). On the size of man's brain as indicated by skull capacity. J. comp. Neurol. 59, 1-28. BURKITT, A. N. (1946). The zoological position of man. Med. J. Aust. 2, 613-617. BURSTON, R. S. (1913). Records of the anthropometric measurements of 102 Australian aboriginals. Bull. N. Terr. Aust. 7A. CAMPBELL, T. D. (1925). Dentition and palate of the Australian aboriginal. Univ. Adelaide Publ. No. 1 under Keith Sheridan Foundation. CAMPBELL, T. D. & HACKETT, C. J. (1927). Adelaide University Field Anthropology: Central Australia. No. 1. Introduction: Descriptive and Anthropometric observations. Trans. roy. Soc. S. Aust. 51, 65-75. CAMPBELL, T. D. & LEWIS, A. J. (1926). The aborigines of South Australia: anthropometric, descriptive, and other observations recorded at Ooldea. Tran8. roy. Soc. S. Aust. 50, 183-191. CLARK, W. E. LE GROS (1945). Pithecanthropus in Pekin. Antiquity, 19, 1-5. DAVENPORT, C. H. (1926). Human metamorphosis. Amer. J. phys. Anthrop. 9, 205-232. DUNN, L. C. (1928). An anthropometric study of Hawaiians of pure and mixed blood. Pap. Peabody Mus. 11, 89-211. FAWCETT, C. D. & PEARSON, K. (1898). Mathematical contribution to the theory of evolution. On the inheritance of the cephalic index. Proc. roy. Soc. 62, 413-417. GERHARDT, K. (1938). Zur Frage Brachykephalie und Schmdelform. Z. Morph. Anthr. 37,277-489. GUTHE, C. E. (1918). Notes on the cephalic index of Russian Jews in Boston. Amer. J. phys. Anthrop. 1, 213-223. HADDON, A. C. (1929). The Races of Man and their Distribution. Camb. Univ. Press. HARRIS, H. A. (1926). Endocranial form of gorilla skulls. Amer. J. phys. Anthrop. 9, 157-172. HERSKOVITS, M. J. (1926). Correlation of length and breadth of head in American Negroes. Amer. J. phys. Anthrop. 9, 87-97. HnTL, W. C. 0. (1941). The physical anthropology of the existing Veddahs of Ceylon. Ceylon J. Sci. (G), 3, pt. 2, pp. 25-144. HIRSCH, N. D. M. (1927). Cephalic index of American-born children of three foreign groups. Amer. J. phys. Anthrop. 10, 79-90. HRDLICKA, A. (1928). Catalogue of human crania in the United States National Museum Collections. Proc. U.S. nat. Mus. 71, 1-40. HUG, E. (1940). Die Schmdel der fruhmittelalterlichen Graber aus dem solothurnischen Aaregebiet in ihrer Stellung zur Reihengraberbevolkerung Mitteleuropas. Z. Morph. Anthr. 38,359-528. HUXLEY, J. S. (1932). Problems of Relative Growth. New York: Dial Press. JONES, F. WOOD (1931). The non-metrical morphological characters of the skull as criteria for racial diagnosis. J. Anat., Lond., 65, 179-195. KAPPERS, C. U. ARIkNS (1935a). Stability and variability of central- and palaeo-Asiatic index peaks and some remarks on the Greenland Eskimo. Proc. Acad. Sci. Amst. 38, 686-700. KAPPERS, C. U. ARIENS (1935b). The degree of changes in the cephalic index correlated with age and environment. Proc. Acad. Sci. Amst. 38, 989-1001. KAPPERS, C. U. ARIENS (1936a). The character and spread of associated cephalic and cranial index peaks in Africa. Proc. Acad. Sci. Amst. 39, 33-49. KAPPERS, C. U. ARIENS (1936b). The spread of primitive humanity, and its links with the more differentiated races, as revealed by cephalic and cranial index curves. Proc. Acad. Sci. Amst. 39, 1156-1165. KAPPERS, C. U. ARIENS (1937). The spread of primitive humanity and its links with the more differentiated races, as revealed by cephalic and cranial index curves. Proc. Acad. Sci. Amst. 40, 29-38. KAPPERS, C. U. ARIENS (1939). Index curves of Pygmy and Veddoid tribes. Dravidian influences in the Indian Archipelago. Proc. Acad. Sci. Anst. 42, 687-700. KAPPERS, C. U. ARIENS (1945). On the frequent occurrence of double peaked index curves and the presence of allosexual indices in men and women. Proc. Acad. Sci. Amst. 48, 487-498. KLEIN, W. (1935). The degree of developmental changes in the length-breadth index of the head of Dutch Askenasim Jews. Proc. Acad. Sci. Amst. 38, 1021-1026. MACALISTER, J. (1898). The causation of brachi- and dolicho-cephaly. J. Anat. Physiol. 32, 334--340.

258

A. A. Abbie

MARTIN, R. (1928). Lehrbuch der Anthropologie. 2te Auf. Jena: Fischer.

MORANT, G. M. (1928). A preliminary classification of European races based on cranial measurements. Biometrika, 20 B, 301-375. MORANT, G. M. (1935). A study of pre-dynastic Egyptian skulls from Badari based on measurements taken by Miss B. N. Stoessiger and Professor D. E. Derry. Biometrika, 27, 293-309. PARsONS, F. G. (1930). Alveolar and facial prognathism. J. Anat., Lond., 65, 147-151. PEARSON, K. & TIPPETT, L. H. C. (1924). On stability of cephalic indices within the race. Biometrika, 16, 118-138. PITTARD, E. (1905). Influence de la taille sur l'indice cephalique dans un groupe ethnique relativement pur. Bull. Soc. Anthrop., Paris, pp. 279-286. RANKE, 0. (1905). Beitrige zur Frage des kindlichen Wachstums. Arch. Anthrop., Braunschw., 31-32, 161-180. REEVE, E. C. R. & HUXLEY, J. S. (1945). Some problems in the study of allometric growth. Essays on Growth and Form Presented to D'Arcy Wentworth Thompson, ed. by W. E. Le Gros Clark and P. B. Medawar. Oxford Univ. Press. Pp. 121-156. 'SCHULTZ, A. H. (1940). Growth and development of the Chimpanzee. Contr. Embryol. Carneg. Instn, 28, 1-63. SCHWERZ, F. (1912). Die Allamannen in der Schweiz. Z. Morph. Anthr. 14, 609-697. SPIER, L. (1929). Growth of Japanese children born in America and in Japan. Univ. Wash. Publ. (d), 3, 1-30. SULLIVAN, L. R. (1919). The 'Samar' united twins. Amer. J. phys. Anthrop. 2, 21-24. TURNER, W. (1884). Report on the human crania and other bones of the skeletons collected during the voyage of H.M.S. Challenger in the years 1873-1876. Rep. Chall. Exped. Part I. The Crania (Zool.), p. 10. VOGT, C. (1864). Lectures on Man. Anthrop. Soc. London. Longman, Green, Longman and Roberts. WEIDENREIcH, F. (1940). The brain and its role in the phylogenetic transformation of the skull. Trans. Amer. phil. Soc. 31, 321-442. WEIDENREICH, F. (1945). The brachycephalization of recent mankind. Sthwest J. Anthrop. 1, 1-54. WUNDERLY, J. (1939). The cranial and other skeletal remains of Tasmanians in collections in the Commonwealth of Australia. Biometrika, 30, 305-337.