238 
A First Study of the Burmese Skull 
Dravidians. Finally, I took as the material for a Dravidian type the series of 
40 skulls brought over by Dr Shortt from Madras in 1878, and belonging chiefly 
to the Maravar tribe, a Dravidian stock from the Madura district of the Madras 
Presidency. It was unfortunate that Dr Shortt saw fit to divide his material into 
two halves, one of which went to the museum of the Anthropological Society of 
Paris, while the other came to our own Royal College of Surgeons. This second 
series of 20 was sexed and measured by Flower. Three are given as ?; three are 
indeterminate, and if we may class these with the males we have seventeen of the 
latter. The 20 in France were measured by E. Callamand* by Broca's method, but 
were unfortunately not sexed. Comparing, however, the means obtained by him 
with those obtained by Flower I have come to the conclusion that the 20 in France 
must be either all, or practically all, males, and have put them with Flower's 17, 
making 37 in all. This has been done with the more confidence as I have seen, on 
examining the series in the Royal College of Surgeons by the kindness of Professor 
Keith, that the skulls are very closely alike and form a distinct type with no marked 
exceptions, so that the means may be obtained with as great accuracy from a score 
of them as from threescore of a more mixed race. 
With regard to the whole of our comparative material, I may say that I am 
probably more conscious than anyone else of how much work remains to be done 
on all the races dealt with. The tables of means on pp. 239-240 are such as the 
data available could supply. 
11. Comparison of Burmese with adjacent Racial Series. 
We will begin by comparing our largest group, the Burmans proper, with the 
Malayans, basing of course our judgment of any difference between their respective 
means for any character, on its relation to the probable error of that difference in 
samples of the size usedf. 
We notice how extremely close the two means are in the majority of characters; 
in fact, a significant difference is only revealed by both sexes in the following points : 
the Malayan, though brachycephaUc, has a smaller cephahc index (B/L) than the 
* "Le crane des Noirs de I'lnde," Revue d'Anthropologie, Vol. i. 4th series, pp. 607-625, 1878. 
f The probable errors can be roughly estimated from those given in Table II, where the Burman 
(Type A) is compared with Type C. The means of this latter type are based on about eight individual 
measurements for males, and about seventeen for females. As these numbers vary in the different 
series which we shall compare with our Burmans, so the probable errors will be modified, and in the 
following proportions : 
Males. 
For means based on 14 skulls, the probable error of the difference will be -75 times that given in Table II. 
„ ,, 21 ,, ,, „ ,, „ -65 times. 
54 „ „ „ „ „ -5 
266 „ „ „ „ „ -4 
Females. 
For means based on 6 skulls the jjrobable error will be 1-5 times that given in Table II. 
„ ,, 50 „ ,, „ -75 times. 
278 „ „ „ -6 
