DEPARTMENT OF EXPERIMENTAL EVOLUTION. 119 



Dr. A. M. Banta has, as stated in the last report, selected inside 

 of one "pure-line" of daphnids the most and the least sensitive to 

 light. The species reproduces parthenogenetically. The data, in- 

 cluding the reaction times of over 10,000 individuals, covering in 

 some cases 50 generations, are now being reduced. 



The Director, working with Miss Annie P. Henchman, studied 

 the variation in the statoblasts of the fresh-water bryozoan Pectina- 

 iella magnijica. These seed-like buds carry marginal hooks whose 

 number varies from 11 to 26. By budding, compound individuals 

 or colonies are produced, each of which may contain from 100 to 900 

 statoblasts. All of the statoblasts of one colony are produced from 

 the same germ-substance without a sexual process to complicate 

 matters. There is a great range of variation inside of any one colony ; 

 i. e., one colony has a mode at 15 hooks, but there are a few stato- 

 blasts with as few as 12 hooks and some with as many as 20. But 

 when we compare colonies that are probably or certainlj^ unrelated 

 the mode and the range in the number of hooks are both quite 

 different. Thus, in another colony, the mode is at 18 hooks and the 

 range is from 15 to 22. Thus variation inside the non-sexually pro- 

 duced colony is less than the variation between colonies. As yet 

 attempts at breeding from given statoblasts have not succeeded. 



METHODS. 

 Statistical Analysis of Data. 



Dr. J. A. Harris has given much time to improving statistical 

 methods and has prepared a memoir on the calculation of inter-class 

 and intra-class coefficients of correlation from class moments when 

 the number of possible combinations is large. 



The Director has published a third edition, slightly revised from 

 the second, of ''Statistical methods with special reference to bio- 

 logical variation." 



Critique of Variation due to Heterogeneity of the Substratum, by J. A. Harris. 

 Dr. J. A. Harris had his attention called to the failure of one series 

 of plantings of beans to give results in harmony with his conclusion 

 that the larger the seed planted the greater the mean number of pods 

 per plant. The series in question gave precisely opposite results. 

 A study of conditions of the planting shows that the seeds of each 

 weight-group had been planted in order across a garden plot that was 

 selected for its apparent uniformity of soil yet there was a demon- 

 strable loss of productiveness from one side of the row to the other, 

 and the larger seeds were planted at the least fertile end. So great was 

 the heterogeneity that the larger seeds actually produced fewer pods 

 than the smaller. In statistical studies on relative productiveness, 

 homogeneity of substratum is clearly of primary importance. 



