Notes on some Buslivian Crania and Bones. 233; 



employment of the lighter material, millet seed. The crania from 

 the South African Museum have been cubed with millet seed; those 

 previously examined and at present in various English museums 

 were cubed with shot. 



Many other measurements of crania have from time to time been 

 described, for which the reader is referred to any standard text-book 

 of anthropology. Further measurements on a previous series of the 

 Bushman and Hottentot skulls will be found in the Journal of 

 the Anthropological Institute of Great Britain and Ireland, 

 vols, xxvii. and xxviii. ; these series have been drawn on for 

 comparison in the present paper. 



METHODS OF INTERPRETATION OF THE MEASUREMENTS. 



It was until recently the custom to compare skulls by a study of 

 the relative magnitude of one dimension in terms per cent, of some 

 other dimension, the resulting value being described as an index. 

 This presents the disadvantage that the index varies according to the 

 magnitude of both components, so that if both should vary together a 

 similar index might be yielded by two skulls or groups of skulls of 

 very different actual magnitude. This method is, however, still in 

 use, and will be first considered. 



In the earlier papers on craniology it was customary to record only 

 the average indices, then the possible indices were divided into a 

 few — usually three — groups and the percentage proportion of a series 

 of skulls falling into each group was noted. Later still it became 

 the rule to arrange the skulls in a series, stating the actual or per- 

 centage number, presenting each index so that variations in 

 distribution around the average or mean point could at once be 

 noted. This would at once show whether the average corresponded 

 to the index of maximum frequency of occurrence, or whether it was, 

 on the other hand, a mean between two or more widely separated 

 groups, in which case it was considered that the series of skulls 

 under observation was not homogeneous but comprised numbers of 

 different racial groups. If such seriation tables be shown graphically 

 in the form of curves in which the ordinates represent the percentage 

 frequency of each index and the abscissge the indices, it will be seen 

 that numerous peaks appear on the curve. The greater the number 

 of skulls in the series the smaller will be these peaks and the 

 smoother the curve, so that large numbers must be available before 

 any very definite interpretation can be given to such peaks as 

 remain. 



