EXPERIMENTAL EVOLUTION. 113 



mutants is a matter of great importance, to the investigation of which 

 this Station is devoting much attention, both in animals and plants. 



MODIFI ABILITY OF THE GERM-PLASM BY ALCOHOL. 



In earlier Year Books have been outlined plans for studying the in- 

 heritance, if any, of the effects of alcohol. In making this study some 

 rats of a litter of an inbred (and therefore probably homozygous) 

 strain of rats were subjected to alcohol; others of the same litter were 

 not. It was to be expected that the first generation of offspring (de- 

 rived from germ-cells in the body of the alcoholized parents) would 

 show the effects of this alcoholism; any defect would then be ascribed 

 to a modification of those germ-cells. But if the grandchildren also 

 are modified, that would indicate that the germ-plasm of the reproduc- 

 tive cells of the treated grandparents had been modified. It can not 

 be denied that the effect might possibly be due to a modification of the 

 cytoplasm of the germ-cells of the treated individuals; but this is, 

 perhaps, less probable. 



Maze-behavior of the grandchildren of alcoholized rats: With the 

 cooperation of Miss Vicari and of other members of the staff, Dr. 

 MacDowell has during the year completed the first formal report on 

 the experiments of the hereditary effect of alcohol on rats, started in 

 1914. This paper has been submitted for pubhcation in the Journal 

 of Experimental Zoology, under the title, "AlcohoUsm and white 

 rats. I: Influence of alcoholic grandparents upon maze behavior." 



In this paper there are compared, with great care, various parallel 

 series of measurements of reactions in the maze of the grandchildren of 

 alcoholized rats and the grandchildren of non-alcoholized rats that were 

 sibs of the alcoholized ones. A comparison was made of the time taken 

 in learning the maze during the entire 24 trials of the original training; 

 also, separately of the first 12 trials, of the second 12 trials, and of all 

 but the first 3 trials. After having learned the maze, the rats rested 

 for a period and were then again tested in the maze to find out how well 

 they retained what they had learned. The two series were compared 

 by giving each rat 12 trials in this retention test. The 24 original 

 learning trials and the 12 retention-test trials were combined for a 

 comparison. Furthermore, the daily record (of 3 trials per day) was 

 compared for the two series. 



To the above comparisons of learning and retention ability of de- 

 scendants of alcoholic and non-alcoholic rats were added the following: 

 The distance covered in running each trial; the speed, or distance per 

 second; the different types of errors made; the number of trials before 

 the first perfect trial, and the time spent in running perfect trials. 



Dr. MacDowell reports further: 



"Whenever the numbers appeared large enough to warrant the calculation 

 of standard deviations, these have been obtained and the probable errors of 

 the averages and of the differences between the test and control averages 



