128 CARNEGIE INSTITUTION OF WASHINGTON. 



selection was begun. The effect was pronounced. In two generations 

 the new high strain (X) attained even a higher level than the original 

 high strain and maintained a level as high as that for Strain I. 



In the twenty-ninth generation a low selection was begun in Strain I. 

 The result was the immediate production of a low strain (XII, see 

 fig. 2). At the same time a low selection was begun in Strain X, 

 the high strain, which was itself a return selection from Strain III. A 

 low strain (XIV) was obtained in the second generation of selection. 

 These selection experiments are represented diagrammatically in the 

 figure, the abscissas representing generations of descent and the or- 

 dinates degrees of intergradeness. 



Seven other intergrade strains of Daphnia longispina used in similar 

 selection experiments gave similar results, so that one seems warranted 

 in concluding that in this sex-intergrade stock selection and return 

 selection are equally effective and that a strain may be raised or low- 

 ered in the scale of intergradeness at will by means of "selection." 



Fig. 2. — Curves showing diagrammatically some of the selection experiments with sex inter- 

 grade strains of Daphnia longispina — Strains I, III, X, XII, and XIV. The abscissas represent 

 generations of selection, the ordinates degrees of intergradedness. Roman numerals indicate 

 the strain numbers while the courses of the curves indicate which were selected as high and 

 which as low strains. 



The interpretation of the results of "selection" in Entomostraca is 

 not entirely clear, since there are numerous elements of the problem 

 that are unknown. To these principles we hold fast: Selection in a 

 proper sense selects only what is given; it does not create. Also, any 

 inherited consequences of the selection are carried in the constitution 

 of the gametes, either chromosomes or cytoplasm. If, then, "through 

 selection," a parthenogenetic race has been genetically modified, it is 

 because genetic idiosyncrasies in the desired direction have been 

 afforded. If a rapid change has been possible through selection it is 

 because individuals showing somatic variations carry corresponding 

 determiners for such in their germ-cells. Thus, in making selection 

 of visible, somatic hereditary peculiarities we are, at the same time, 

 selecting corresponding, but invisible, gametic determiners for such 

 hereditary peculiarities. 



Just what the nature of the hereditary change in the gametes is, that 

 the selector takes advantage of in his successful selections, is uncertain. 



