February 4, 1916] 



SCIENCE 



177 



copious diet of tlie same green flagellates produced 

 as high as 95 per cent, of male offspring, thus 

 showing that it is the quantity of the food that 

 regulates the production of the sexes and not 

 the stimulus of a change of food. 



Male-production in Hydatina Favored by Oxygen: 



A. Franklin Shull and Sonia Ladoff, Uni- 



Tersity of Michigan. 



Whitney's experiments of a year ago, in which 

 feeding these rotifers on the green flagellate 

 Chlamydomonas resulted in greatly increased male- 

 production, left room for doubt whether other 

 agents than nutrition might not be producing part 

 of the effects noted. The food cultures were dif- 

 ferently constituted at the outset, and the organ- 

 isms reared in them may have produced secondary 

 differences. We have attempted to test some of 

 the possible factors other than nutrition. So far 

 our results may be interpreted largely in support 

 of Whitney's conclusion; for, while one of the 

 suspected agents has been found to increase male- 

 production, its effect is not so marked as that in 

 Whitney's experiments. The one effective factor 

 discovered is oxygen. Under several different con- 

 ditions, oxygen produced uniform effects of mod- 

 erate degree. 

 On the Inheritance of Size in Paramecium: James 



E. AcKERT, Kansas State Agricultural College. 



A series of experiments with Paramecium cau- 

 datum and P. aurelia was carried on with a view 

 to determining the effect of selection within the 

 progeny of a single individual. In 1911, when 

 these experiments were begun, the excellent work 

 of Jennings had already been reported; but, to 

 test this principle, independently, using large num- 

 bers of individuals, seemed justifiable. In a 

 typical experiment a single Paramecium was iso- 

 lated on a depression slide in a few drops of hay 

 infusion. After several generations there were 

 isolated from its descendants two Paramecia — one, 

 the shortest of the progeny, the other, the longest. 

 The descendants of each of these individuals were 

 kept in separate receptacles under environmental 

 conditions as nearly identical as possible. At a 

 later time all but a few of the animals of each 

 group were killed and measured. In all of the 

 experiments, except one, the images of the Para- 

 mecia were thrown upon a screen with a combina- 

 tion microscope and lantern, giving a magnifica- 

 tion of 3,200 diameters. The usual methods of 

 dealing with statistical data were used in the 

 preparation of the results. In all cases the effect 

 of the selection within the progeny of a single in- 



dividual was negative. In some instances the dif- 

 ference in mean lengths of the groups under com- 

 parison fell within the probable errors of the 

 means; in others the mean lengths of the progeny 

 of the smaller Paramecia were larger than those 

 of the descendants of the larger Paramecia. The 

 conclusion is based upon measurements of nearly 

 6,000 Paramecia. 



The Influence of Selection on the Number of Ex- 

 tra Bristles in DrosopMla: E. Caeleton Mac- 

 DowELL, Carnegie Institution of Washington. 

 Previously it has been shown that the extra 

 bristles that characterize a certain race of 

 DrosopMla are conditioned by a Mendelian de- 

 terminer; that the exact number of extra bristles 

 is not inherited, but varies in relation to external 

 conditions; that, in spite of this, the selection of 

 high variates as parents, continuously raised the 

 averages of the race for several generations, after 

 which no further progress could be determined. 

 The present report carries the selection for in- 

 creased numbers of bristles to the forty-sixth gen- 

 eration. In certain of the later generations the 

 averages have been raised by more favorable con- 

 ditions and by counting only the large flies that 

 hatch at the flrst of a bottle, the flies at the end of 

 a bottle being smaller and with fewer bristles. 

 The upper limits of the distributions would not be 

 influenced in the same way, and so offer a better 

 test for the effect of selection. These upper limits 

 show no tendency to advance after the first few 

 generations. Two series of return selections have 

 been made, from the sixteenth and twenty-seventh 

 generations. These failed to show any lowering 

 of the averages, although carried on for six and 

 eight generations, whereas the initial rise of the 

 averages was immediate. The distribution of 

 extra bristles extracted from a cross with normals 

 is lower than that of the corresponding inbred 

 generation. A race of low grade has been estab- 

 lished from extras extracted from a cross. This 

 race averages about two bristles lower than the 

 high-selected race. If, as formerly proposed, the 

 initial rise in the averages was due to a sifting 

 out of secondary determiners, all the above results 

 would be expected. 



Twinning in Cattle, with Special Reference to the 

 Free Martin (illustrated with lantern) : Leon J. 

 Cole, College of Agriculture of Wisconsin. 

 A study of 303 multiple births in cattle, ob- 

 tained directly from breeders. The records in- 

 clude: 43 cases homosexual male, 165 eases re- 

 corded heterosexual (male and female), 88 cases 



