240 HEREDITY IN RELATION TO EUGENICS 



would be well if all were included in all genealogies. Another 

 desideratum is abundant photographs of the persons whose 

 biographies are given; especially, strictly full-face and profile, 

 to facilitate comparisons; and two or three photographs at 

 successive ages would be still better than one. 



Attention should be paid to the form of the pedigree. The 

 commonest form is that which begins with the first known 

 male ancestor bearing the surname. His children are given, 

 but in the later generations only the offspring of males are 

 named. Few genealogies attempt either to trace the lines 

 going through females or to give the ancestry of the consorts. 

 A second form of pedigree begins with the author or some 

 other one person and gives an account of all of his direct 

 ancestors in ever expanding number toward the earUer 

 generations. This method is scarcely more valuable than 

 the other from a scientific point of view, based as it is upon 

 the exploded idea that inheritance is from parents, grand- 

 parents, etc. 



The ideal genealogy, it seems to me, starts with a (pref- 

 erably large) fraternity. It describes fully each member 

 of it. It then describes each member of the fraternity to 

 which the father belongs and gives some account of their 

 consorts (if married) and their children. It does the same 

 for the maternal fraternity. Next, it considers the fraternity 

 to which the father^s father belongs, considers their consorts, 

 their children and their grandchildren and it does the same 

 for the fraternities to which the father's mother belongs. 

 If possible, earlier generations are to be similarly treated. 

 It were more significant thus to study in detail the behavior 

 of all the available product of the germ plasms involved in 

 the makeup of the first fraternity than to weld a chain or 

 two of links through six or seven generations. A genealogy 

 constructed on such a plan would give a clear picture of 

 heredity, would be useful for the prediction of the charac- 



