A CONTRIBUTION TO THE STUDY OF THE SCOTTISH SKULL. 



391 



that prove the rule," or has there been any special selection of skulls made in our 

 series, especially as regards the female group ? Of such a contingency I am not 

 aware, and the only conclusion I can come to is that there is room for further research 

 as to the relative variability exhibited by the two sexes in different races. 



The variability of the foregoing dimensions and indices may be graphically repre- 

 sented by constructing frequency polygons ; and, from the data given in the tables, 

 it only requires time and patience to calculate the analytical constants of the curves 

 which graphically represent the true nature of the frequency distribution. 



I have constructed frequency polygons for several of the dimensions and indices 

 representing the variability of these in both sexes (figs. 2 to 6). One of the pro- 



FlG. 5. — Frequency curve for cephalic index in 100 male and 100 female skulls. 



minent features of these frequency polygons is the irregularly peaked or multimodal 

 appearance they present, which may appear to point to heterogeneity in the series 

 of crania from which the measurements were taken. Both C. D. Fawcett and Mao 

 donnell urge us strongly to be cautious against using this " mountain range " appear- 

 ance of the curve as an argument in favour of heterogeneity, as the peaks may be due 

 to various other causes, among which the chief are : — 



(a) That we are dealing in most cases with very small random samples. 



(6) That some of the skulls are attributed to, and included in, the results of the 

 wrong sex group. 



(c) That " foreign " skulls are included. 



(d) That the ages of the skulls in the collection are very varied. 



(e) That some of the skulls may have been altered by post-mortem changes. 

 (/) That clustering may occur due to family groups included in the collection. 



