VISIBLE AND TNVISIBLK CHARACTERS IN^ SILKWORMS. 139 



(2) Out of 26 la.yings of pure mnltivoltine Nistri 2 s and F., c? s 

 all the eggs liatclierl, thus all being multivoltine. 



These results give a clue to the gametic composition of the F., 

 moths, for they show : — 



(1) that 19 1\ $s were dominant in the univoltine character 



and 2 in the multivoltine character. 



(2) that all the E., c? s which may have been univoltine were 



recessive in the univoltine character. 



Table 3, illustrating descent of the univoltine and multivoltine 



CHARACTER IN THE CROSS BETWEEN THE UnIVOLTINE ItAL.-Jap. $ 



AND Multivoltine Nistri cJ • 



Univoltine Ital.-Jap. $ X Multivoltine Nistri (? . 



Layings aJl univoltine Fi 



11 ? motlis 21 (? moths. 



7 layings univoltine. 2 multivoltine. 1 half multivoltine F^ 



62?s.34<?s. 



37 layings univoltine. 2 almost univoltine. 2 multivoltine. 2 almost multivoltine. F3 



34 ? s. 47 (? s. 



14 layings univoltine. 7 almost univoltine. 4 almost multivoltine F4 



(Discontinued.) 



Unequal Sex-ratios. 



I had noticed that in the eai'lier generations of the Nistri $ 

 and Ital.-Jap. S cross (which were reared a year previous to that 

 of the Ital.-Jap. 5 and Nistri J cross), the number of males 

 and females seemed very unequal, and that in one generation 

 males predominated, and in another, females. However, these 

 remarkable fluctuations graduall}^ decreased in the latter 

 generations. 



In my second series of experiments, which were with the Ital.- 

 Jap, 5 ^nd Nistri c? ; I noted the exact number of males and 

 females in each generation, and found that in this reciprocal cross 

 the sex-ratios were also very unequal, but just the reverse in the 

 character of the predominating sex, for in F^ of the Nistri 5 

 cross there were many more females than males, while in F, of 

 the Ital.-Jap. $ cross there were more males than females. 



The unequal sex-ratios could be accounted for by supposing 



