Davenport and Wkeks: Inheritance of Epii^epsy 25 



turbed, the expected proportions ; but if, on the other hand, alco- 

 hohsm is a cause of defective offspring then the proportion of 

 defective offspring should exceed expectation. We do not pre- 

 tend to have sufficient data, and data of the right sort, to settle 

 this question ; but what we have is sufficient to give some support 

 to one hypothesis or the other. In Table II we find a great pre- 

 ponderance of defective offspring — 87 per cent., instead of the 

 expected 50 per cent. In Table V the matings of alcoholics with 

 other "tainted" consorts yield 23 defectives in 71, or 32 per cent, 

 instead of 25 per cent. The two fraternities from two alcoholic 

 parents give, together, 25 per cent, epileptic and the same of neu- 

 rotic offspring. In the alcoholic X normal matings of Table VI, 

 where only 25 per cent, defectives are to be expected, actually 29 

 out of 80, or 36 per cent, are defective. We see, accordingly, a 

 constant excess beyond expectation of epileptic and feeble-minded 

 offspring from alcoholic parents. In so far our results support 

 the view that alcoholism, to a certain extent, is a cause of defect ; 

 that 10 or 20 per cent, more children in any fraternity are defective 

 than would be were it not for alcohol. However, a word of cau- 

 tion must be added. It is not improbable that some of the alco- 

 holics are actually feeble-minded and any such would tend to 

 increase the average of defective offspring because of their inher- 

 ent defective germ cells and quite apart from any poisoning effect 

 on the germ cells of alcohol. The hypothesis that alcohol is a 

 "race poison" deserves more critical, especially experimental, 

 study on the lower animals. 



Second. Have we any reason for believing in an increase in 

 the proportion of epilepsy to the population at large? Certainly 

 the care of epileptics is an increasing burden to the state ; but it is 

 obvious that that is largely because the class is better cared for 

 by the state. 



To answer the question completely for any state like New 

 Jersey we should find the proportion of epileptics in the popula- 

 tion bom between 1870 and 1900 and compare it with the propor- 

 tion of epileptics in the population born between 1840 and 1870. 

 No census adequate to give these proportions has been made. An 

 approximation to the required comparison would be furnished if 

 we had a random sample of the population of each of the two 

 epochs and knew the proportion of epileptics in each sample ; but 

 we cannot, as yet, meet that condition. The best thing we can 



