Davenport and Weeks : Inheritance of EpiIvEpsy 3 



of cases an epileptic was known to have epileptic ancestors 

 ("direct heredity") or other relatives ("collateral heredity"). 

 This proportion was then taken as the index of heredity. The 

 index naturally increased as the study of the family was extended 

 and so it is not surprising to find that the indices of heredity of 

 epilepsy, determined by different workers, varies from 21 per 

 cent, or lower to 75 per cent. 



In this paper our data have been analyzed by the method 

 commonly employed by biologists and known as the Mendelian 

 method. This method assumes that the inheritance of any char- 

 acter is not from the parents, grandparents, etc., but from the 

 germ-plasm out of which every fraternity and its parents and 

 other relatives have arisen. The bodies of persons as we know 

 them serve as (imperfect) indices of the nature of the germ- 

 plasm from which they spring. The relation of soma and germ- 

 plasm is as follows: 



1. If the soma lacks a unit character upon which normal de- 

 velopment depends that is prima facie evidence that the repre- 

 sentative of that character is absent from its germ-plasm; con- 

 sequently such a person cannot transmit the character in question. 



2. If the soma has the unit character for normal development 

 that is evidence that the germ plasm has the corresponding de- 

 terminer. But either one of two cases is possible, (a) the deter- 

 miner was derived from both sides of the house, so that it is 

 double in the germ plasm (duplex, designated below by 2) and all 

 the germ cells have the character; or else, (b) it came from one 

 side of the house only, in which case it is single in the germ plasm 

 (simplex, designated below by i) and half of the germ cells have 

 and half lack the character. The condition in the case when the 

 determiner is absent may be called nulliplex (designated below 

 byo). 



A moment's consideration will show that six kinds of matings, 

 disregarding sex, are possible. These matings, together with 

 the sort of offspring that they may be expected to yield, are indi- 

 cated in the following table: 



I, o X o. All without the character of full mental development. 

 II, o X I. 50 per cent, devoid of the character; 50 per cent, simplex. 



III, 0X2. All with the character, simplex. 



IV, I X I. 25 per cent, with the character absent, So per cent, with it 

 simplex; 25 per cent, with it duplex. 



V, 1X2. 50 per cent, with character simplex ; 50 per cent, with it 

 duplex. 



VI, 2X2. All with character duplex — mentally strong. 



