[Vol. XXIII, 1922.] P. C. Mahalanobis : Analysis of Stature. 61 



4'55 + '20. , Blakeman believes 1 the "increased variability in 

 stature to be due to the measurements being taken on the corpse 

 and not on the living subject.' He mentions further' 2 that the 

 average V for males in Pearl's data is 4*11. 



I have thought it best to omit the above series of corpse 

 data for purposes of comparision. It will be observed that the 

 variability is in each case considerably higher than the average 

 variability (which is about 3'6) obtained by omitting them, Thus 

 the only effect of including the "corpse" data would be to still 

 further increase our average variability. We may further note that 

 in most of the above cases, the variability is even higher than the 

 variability of our Anglo-Indian data, which is about 4*06. Thus 

 omission of the corpse date cannot affect our general conclusion 

 that the variability of the Anglo-Indian series is not significantly 

 greater than the average variability of stature for homogeneous 

 material. 



Tocher 3 gave in 1906, a very large series of measurements 

 on the Scottish Insane, numbering 4381 males. 



Schuster * in 1910 gave V for different age-groups of Oxford 

 undergraduates. For reasons already explained I have taken 

 the average variability for the whole group of 959 individuals. 

 In an editorial note to the above, 5 some results for 493 Scottish 

 (Aberdeen) undergraduates are quoted. I have calculated the 

 coeff. of variability in this latter case also. I ma}^ note in passing 

 that the different age-groups of the Oxford data do not give lower 

 values of variability, in fact give slightly greater values than the 

 total in many cases. 6 



Craig 7 gave the results of a very large series of measurements 

 of modern Egyptians. These were classified in accordance with 

 the town or district of birth. 8 The total number in each group 

 is fairly large and this series gives us a very good list of variabi- 

 lities for purposes of comparison. I have retained the separate 

 variability for Aswan, omitting the total variability as the material 

 is not homogeneous. 



Garett 9 has given a series of measurements of the natives of 

 Borneo and Java. The majority were coolies in the employ of the 

 author. Unfortunately the number in the case of each people is 

 not very extensive, and I have been only able to retain the values 



1 Ibid., p. 131. 2 Ibid., p. 132. 



3 J. F. Tocher: "The Anthropometric Characteristics of the Inmates of 

 Asylums in Scotland.,'' Biometrika Vol. 5 (1906), pp. 298 — 350. 



* E. Schuster: " First Results from the Oxford Anthropometric Laboratory," 

 Biometrika Vol. 8 (191 1), pp. 40-51. 



6 Ibid., p. 49. 



6 Thus the lumping together of all age-groups cannot again affect the general 

 validity of our conclusions. 



7 J. I. Craig " Anthropometry of Modern Egyptians," Biometrika Vol. 8 

 (1911), pp. 69—77. 



5 Ibid., p. 75. 



9 T. R. H. Garett: " Natives of the Eastern Portion of Borneo and Java," 

 Jour. Roy. Antlirop. Inst., Vol. XL1I, 1912, pp. 60 — 66. 



