EXPERIMENTAL EVOLUTION. 117 



"A more striking result of the comparison of the two generations is that the 

 differences between the tests and controls are greater in the generation farther 

 away from the alcohol. This is especially the case in the first three days of 

 the training, where the differences are great in the second generation, but in 

 the first generation are practically non-existent. In the later days the actual 

 differences between the tests and controls in the two generations are not very 

 unlike. However, when these differences in the later days are considered, it 

 appears that they are not significant in the second generation, while in the 

 first, due in part to the larger numbers, they are fully significant. Thus there 

 are significant differences in the first part of the training for one generation 

 and in the second part of the training for the other generation, although in 

 both generations all the significant differences are in favor of the controls." 



The above indicates that the effect of alcohol was certainly not 

 purely somatic, due to alcohol in the blood of the mother entering the 

 vessels of the fetus, for if this were the case, the grandchildren should 

 show less rather than viore difference between the tests and controls. 

 There is some effect handed down; whether it originated in the germ- 

 cells of the embryos of the offspring of the treated mothers, or in the 

 germ-cells of the parents themselves, can not be decided from the evi- 

 dence at hand. The only data on this point appear to show that the 

 germ-cells of the fathers alone w^ere not sufficiently modified to affect 

 the offspring. A small group of 5 rats, whose father w^as treated and 

 whose mother was normal, compared with 7 normals from the same 

 pair of grandparents, do not show any effect of the alcohol treatment; 

 on some days the averages of the test rats are higher, on some days the 

 averages of the control rats are higher, wdth no preponderance either 

 way. If this result has any significance it may be due to two things: 

 either that the amount of alcohol taken by one parent w^as not enough 

 to modify the offspring, or else the susceptible age is while the individual 

 is in utero. Unfortunately, no further generation was raised from these 

 rats. It appears to be entirely possible that the difference in first days 

 of training in the two generations studied may be due to two different 

 effects of the alcohol treatment; there may be working both (1) a direct 

 somatic effect that does not modify the first part of training, but which 

 tends to make the rats move more slowly in the last part, after the 

 habit is well learned, as well as (2) a germinal effect that does not make 

 itself manifest till the next generation, and then modifies the nervous 

 system rather than the muscular mechanism and hinders the learning 

 process that goes on in the early trials. 



^'Maze-behavior of rats from parents and grandparents treated with alcohol. — 

 Miss Vicari has made simimaries of the training records of a set of rats that 

 came after two alcoholized generations. Of these there were 9 tests and 9 

 controls. When the time data were averaged for each day of training sepa- 

 rately, the differences in every case in the original training were plus, that is, 

 the tests took more time. In the trials of the retention test given a month 

 later, the controls took more time on the first and third days; the tests took 

 more time on the second and fourth days. Since the number of rats is so 



