DEPARTMENT OF GENETICS. 135 



"Dice-casting and pedigree selection," in which it was shown that 

 practically all of the mathematical and statistical phenomena of 

 schemes of selection can be demonstrated by dice-casting. Specific 

 examples were worked out in the case of regression, pure-line selection, 

 and selection on the somatic basis. 



Selectiox of a Physiological Character in Cladocera. 



Dr. Banta's complete paper on this subject appeared as publication 

 No. 305 of the Institution. Three species of Cladocera were used: 

 two of the genus Daphnia and one Simocephalus. All were tested for 

 their reactiveness to a constant light placed at one end of a long tank 

 of water in which the Cladocera swam. All variations in reactiveness 

 due to environmental conditions were considered and their effects, 

 as far as possible, eliminated. For many generations the Cladocera 

 that were the most responsive to light were selected and bred from to 

 form the plus strain; the least responsive to light were similarly se- 

 lected and bred from to form the minus strain. In most strains (13) 

 the continued selection of more and less reactive individuals for gen- 

 erations did not cause the two lines to diverge in reactiveness. One 

 strain (No. 757) showed a clear effect of selection. The lines diverged 

 and the divergence was permanent (hereditary) . The results indicate 

 that selection alone was impotent to create diverging lines unless favora- 

 ble mutations were occurring, and these apparently did occur in strain 

 No. 757. 



Modification of a Mendelian Ratio by Selection. 



During the past year Dr. Little and Miss E. E. Jones have carried on 

 an experiment with Drosophila melanogaster to determine whether or 

 not the customary 1 : 1 Mendelian ratio in back-crosses of a single 

 pair of characters can be modified by the isolation of different genetic 

 strains incident with selection. 



A pair of fourth-chromosome characters, normal-eyed and eyeless, 

 was chosen. A control series showed that there was normally a slight 

 deviation from the 1 : 1 ratio in back-crosses between normals carrying 

 eyeless and eyeless flies. The observed ratio, which was 1.44±0.02 

 normals to 1 eyeless, is the result of a differential viability of the sort 

 frequently found in cases where a recessive mutation appears to be at a 

 slight disadvantage when compared with the normal type. 



Eight selection lines were started, two of which died too early to give 

 conclusive results. Four lines were selected for an excess of normal- 

 eyed and four for an excess of eyeless. The former are known as the 

 high selection lines, the latter as the low selection lines. A detailed 

 description of the experiment is almost ready for press. 



