516 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1961 



In the almost complete absence of analytical methods related, in the 

 Mendelian sense, to the genetical content of the data, much effort has 

 been devoted, with great effect, toward increasing the efficiency of 

 more empirical methods of statistical analysis, that is to say, analysis 

 in terms of phenotypes. The whole accepted edifice of physical classi- 

 fication of human populations depends, in fact, upon the results of 

 such analysis. 



The genetical analysis of the blood groups and biochemical charac- 

 ters has enabled populations to be compared much more effectively than 

 could have been done on the basis of the observed characters (or pheno- 

 types) alone. A knowledge of the genetics involved has also, in some 

 cases, made possible a fairly full analysis of the mechanism by which 

 these characters, through natural selection, become adapted to the 

 environment. A full analysis of the modes of inheritance of the 

 external body characters might be expected to have similar conse- 

 quences for these characters. However, it is now clear that continu- 

 ously varying characters such as skin color and stature are each mider 

 the control of a large number of genes, known as polygenes. Gener- 

 alized methods of analysis of observations on such characters have 

 been devised by Darlington and Mather (1949), but a full analysis in 

 terms of individual genes is not at present in sight. 



In view of the greatly increased discriminatory power which 

 genetically based methods would almost certainly confer, much further 

 effort is needed, but few geneticists appear to be aware of the need, 

 and very few indeed have contributed at all substantially to the 

 subject. In the case of stature, Fisher and Gray, as we have seen, 

 many years ago extracted virtually all possible genetical information 

 from the data then available; since then Tanner and Healy (Tanner, 

 1954; Tanner and Healy, 1956) have extended such analysis to some 

 more recent data, but no other work of importance has been done. In 

 the case of skin color, however, a considerable amouiit of work has 

 been done recently, especially by G. A. Harrison (Harrison, 1957; 

 Harrison and Owen, 1956-57) with promising results. 



I believe that a more fundamental genetical analysis of mammalian 

 and human morphological characters is possible than has so far been 

 achieved, and I would commend this difficult problem to any younger 

 geneticists who may hear or read this lecture. Such an analysis would, 

 I hope, lead in time to much more efficient methods of dealing with the 

 raw data of morphological anthropology. Hitherto the data them- 

 selves have proved intractable genetically : not only this, but also their 

 sheer volume, even that of the reliable and well-standardized data 

 alone, has discouraged any attempt at a systematic comprehensive 

 analysis. If only methods for their analysis could be devised which 

 were genetically sound, statistically efficient, and practically con- 

 venient, then the modern availability of electronic calculators, capable 



