134 MISS MAUDE L. CLEGHORN OX THE INHERITANCE OF 



produced by Experiment B were entirely different from each 

 other, for they resembled the maternal parent in each 

 case, all the layings of the Nistri g and Ital.-Jap. c5' being 

 multivoltine, and all those of the Ital.-Jap. § and Nistri 

 c? proving univoltine. 



Result of the Experiment as regards the Size of the Cocoon. 



Experiment A. — The parent cocoons of the Ital.-Jap. S 

 and Nistri 5 cross weighed 2*9 and 1-5 grains respectively. The 

 cocoons of the first generation of this cross resembled those of the 

 Nistri {i. e. the maternal parent) more than those of the Ital.- 

 Jap. parent, being rounded at both ends and very thick but 

 of the loose texture of the Mstri, the firm texture of the Ital.- 

 Jap. being entirely unrepresented. They were uniform in 

 size and shape, but varied in weight from 2*5 to 4'1 grains. 



In the second and the immediately succeeding generations, the 

 cocoons were not so uniform in size and shape. Many were large 

 and rather pointed at the ends, whilst they were all thinner and 

 firmer than those of F^, and I found that in the earlier generations 

 moths which gave complete multivoltine layings had nearly 

 always cut out from cocoons which were about 3 grains or less in 

 weight. I did not make use of these 3-grain cocoons for rearing 

 purposes, even though they were far better than the original 

 multivoltine Nistri cocoon, but I carefully selected the best out 

 of the most multivoltine layings of cocoons weighing from about 

 3'5 grains to 5 grains in weight. The cocoons in all the generations 

 of the cross were far superior to the original Nistri cocoon, and in 

 many of the generations they are also sujperior to that of the 

 Ital.-Jap. Hybrid. 



Up to Fg many of the cocoons were 4 grains in weight, in Fg, 

 F^, Fj, many were over 4*5 grains, while some weighed 5 grains 

 and over. In F^ and Fg there were no 4-grain cocoons, and on 

 the whole the cocoons of these two generations were very poor 

 compared with those of the preceding generations, but they were 

 nevertheless still superior to the original Nistri cocoon. As the 

 layings from which these cocoons were produced were nearly 

 entirely multivoltine, the cocoons appeared to be also becoming 

 more multivoltine in character. The rearings in F^ from which 

 Fg hatched were all entirely multivoltine, while the Fg silkworm 

 were better again and seriposited cocoons, nearly all of which 

 were over 3*5 grains, many over 4 gi-ains, and some nearly 5 grains 

 in weight (with chrysalis and outer fluff removed). The layings 

 of the moths from these cocoons were also entirely multivoltine, 

 which showed that it is possible to obtain a 4- or 5-grain cocoon 

 from entirely multivoltine layings in Fg after a cross without any 

 recrossing. Care was, however, always taken to have the parent 

 moths as distantly related as possible. (Table 1.) 



In this expei"iment I found that, after the direct influence of 

 the fresh cross seemed to have disappeared, every third generation 

 produced the best cocoons, for the cocoons of Fg, F^, F^.^, F^^ 



