50 



FRANK R. LILLIE. 



which was apparently monozygotic) ; the addition of these gives 

 the following distribution of the sexes for all my cases : $ $ 29, 

 $ 9 6, £ $ 33, 9 2 24 ; or again combining normal and intersexual 

 females of two-sexed pairs together : $ $ 29, $ 9 39, 9 2 24. As- 

 suming an equality of the sexes in the population dealt with, the 

 expected distribution of sex in the 92 pairs of twins would be 

 $ $ 23, $ 9 46, 9 2 23. It is obvious that there is an excess of 

 male pairs and a deficiency of two-sexed pairs on this assumption. 

 The numbers are not very large, but probably sufficiently so to be 

 significant. 



The aberrant ratio for the twins would be satisfactorily ex- 

 plained if the transformation of the free-martin sometimes went 

 so far in the male direction as to render it indistinguishable from 

 a male. I have, therefore, recently studied male pairs with con- 

 siderable care, but have found no evidences of abnormalities in the 

 reproductive organs ; in my experience the structural differences 

 that distinguish a free-martin from a male are always very great. 

 It was natural to expect the greatest modification of the female 

 when both the male and female twin came from the same ovary 

 and developed in the same horn of the uterus, for under these 

 circumstances the vascular anastomosis would presumably be un- 

 usually early and extensive. The two cases of this sort which I 

 have, however, show no greater modification of the female than 

 many others (see Sec. 4). 



It appeared probable, therefore, that the assumption of approxi- 

 mate numerical equality of the sexes in the population was at fault. 

 I therefore suggested to Mr. J. M. Jewell that he ascertain the 

 sex-ratio in foetal calves taken from the same slaughter-house. 

 Accordingly the sex-ratio of 1,000 cases taken at random was 

 determined (Jewell, 1921). For our purpose it is sufficient to 

 notice that a sex-ratio of 123.21 was established for the 1,000 

 cases, including all foetal ages ; the sex-ratio for the sizes included 

 in the twin collection (3.5-30 cm. in length) was found to be 120 

 {i.e., 120 males for every 100 females; 264 entries). 



Taking this figure as the sex-ratio of the population from which 

 the twins came, the expectation of sex distribution in the 92 pairs 

 of twins would be $ $ 27.37, $ 945.62, 9 919.01. This shows a 

 type of sex distribution similar to that actually found in the twins, 



