70 



Scientific Proceedings (118). 



of untreated rats from untreated parents and treated grandparents 

 gave litters that were 13.3 per cent, smaller than their control 

 litters. So there appears to be a reduction in the average litter 

 size in each generation that is about ten per cent, of the respective 

 controls. However, in no case is the reduction statistically signifi- 

 cant. This is probably due to the small size of each sample, 

 since the combination of all the generations gives a difference that 

 is 3.6 times its probable error and so is to be considered significant. 

 Alcohol appears to have caused a modification in litter size that 

 persists for two generations after the original treatment and is not 

 increased by a second generation of treatment. This is the result 

 to be expected from a definite germinal modification. 



Number of Litters. — The numbers of litters are based upon the 

 production during equal periods of the test pairs and their own 

 particular control pairs. In some cases the period was longer than 

 in others, but opportunities for the tests and controls to produce 

 litters were equal in each case and therefore equal in the totals. 

 The number of litters is, accordingly, purely relative. 44 treated 

 pairs produced 32 litters, whereas on the basis of their controls 91 

 litters were expected. This was a reduction of 65 per cent, db 

 3.37 which is 19.2 times its probable error and so, significant 

 beyond all question. Treated rats from treated parents produced 

 14 litters when 22 were expected on the basis of their controls. 

 This is a reduction of 35 per cent. ± 6.91 and is 5 times its prob- 

 able error. Untreated rats from treated parents produced 33 per 

 cent. =h 8.20 more litters than the controls, and the untreated rats 

 from untreated parents and treated grandparents 55 per cent. 

 ± 8.4 more litters than the expectation. Both these differences 

 are fully significant. The number of litters was strongly reduced 

 when the rats themselves were treated, but, just as soon as the 

 alcohol was further away, the number of litters at once increased 

 and the test animals produced significantly more litters than their 

 controls. The obvious interpretation of this result is that alcohol 

 has acted as a selective' agent by preventing those females from 

 having litters that bore weaker determinants for the production 

 of litters. This accounts for the apparently odd fact that two 

 generations of treatment made less difference than a single genera- 

 tion of treatment. The offspring of treated animals are a selected 



