I. B. J. SoLLAS 20n 



is due to the fact that our present total is made up largely of the 

 offspring of families of the other two kinds. 



In families in which dwarfs of both sexes have already occurred the 

 numbers of the sexes are equal. In families in which up to the present 

 no male dwarfs have been born there is an excess of females. 



In making out the tables of families containing dwarfs of both sexes 

 we are clearly safe in including all families ; but in both the other cases 

 we cannot be certain that any of the families, if they had incrciised, 

 would have remained in the class in which we have placed them. 

 Nevertheless the facts may have some interest. 



Thus in the families containing dwarfs of the male sex only, the 

 proportion of the male to the female offspring approximates to 3:1. 

 It is possible that the 18 normal offspring of unrecorded sex in family 2 

 (see Table 1) w^ould have altered this proportion; but even if we make 

 the unlikely supposition that they were all female we should still have 

 a large excess of males. In family 22 each parent has, when paired 

 with other mates, produced female dwarfs; of the other parents 416 

 and 571 have produced female dwarfs when paired with other mates. 

 Possibly further breeding would show that some of the families in this 

 table are out of place. 



(4) Fantilies coidaiaimj female dwarfs only. 



Offspring 



Totals ... 11? 13 cT 13? 4 



In this table, the female parents in the matings 15, 16, and 28 have 

 never yet given birth to any male dwarfs, the males with which they 

 were mated have had male dwarfs among their offspring. The parents 

 in the matings 17, 21, and 23 have all, when paired with other mates, 

 had male dwarfs among their offspring. 



Sturtevant (/. Exp. Zool. 12, 1912, p. 513) has published an 

 interesting attempt to explain these results, as far as they went in 



