44 Records of the Indian Museum. [Voi,. XXIII, 1922." 1 



Conclusion. 



(1) The Gaussian curve is quite adequate for graduating a 

 short series of 200 Anglo-Indian measurements. This confirms 

 CD. Fawcett s rule of normal distribution for short series 1 of 

 anthropometic measurements. 



(2) There is some tendency towards Type IV with leiitn 

 kurtos.s. All long series, with the exception^ of Am can P a nd 

 I talian recruits seem to be definitely lepto-kurtic. It is therefore 

 likely that stature distribution is in general slightly lepto-kurtic 

 m character but this small lepto-kurtosis does mft become 2 

 tically significant m small samples. ' 



MeanWo^m ".Ti! T d P ° sitive (Mode bein 8 greater than the 

 recruits Scotll "ales criminals, Italian conscripts, Italian 



recruits Scottish insane and a short series of several offenders 

 among English criminals. It is negative in the case of sever al 

 short series of English criminals, and for one long series vz 

 American recruits. For a short series of Anglo-Indians it is 

 positive but is so small that it cannot be called sgnmcant" 

 Hence we conclude that the small skewness of our present sample 

 is not incompatible with homogeneity sample 



case IflllTf' there l W f ihat {he d "tribution of stature in the 

 caseo Anglo-Indians is of the same nature as in the case of other 

 samples where the material is known to be " homogeneous" In 

 other words, the nature of distribution of stZ TeZs.ot reveal 

 any presence of heterogeneity in the Anglo-Indian population* 



1 Biometrika Vol. i (1902), p 443 



detail eoXcus^n TZ ;^H n p°i Cat r sk a f in v St f°™>S-ejty. For a 

 Biometrika Vol. 4 (1905), pi", : Variation, a Rejoinder," 



