TWO PAIRS OF FEMALE TWINS 237 



little more than one-half the size of one normal yolk for birds 

 of this kind. The immediately preceding eggs and the succeed- 

 ing one were similarly much undersized. And, further, this 

 female produced eggs at an abnormally slow rate during the entire 

 year. 



The third double-yolked egg was produced by a generic hybrid 

 (T. orientalis X St. alba), whose reproductive record seems to 

 be fairly normal for her kind. She was mated to another female 

 when the double-yolked egg was produced. 



The fourth and fifth of these double-yolked eggs were also 

 produced by generic hybrids. The fourth was from a Stig. 

 senegalensis X St. alba hybrid. It was produced only five days 

 after a preceding clutch^ — a thing of most unusual occurrence; 

 in another case she produced two clutches, of one egg each, only 

 four days apart. ^ The fifth egg was from a complex generic 

 hybrid (of three species, orientalis, risoria, alba), which produced 

 eggs during only one season, and only three clutches after the 

 abnormal one containing the double-yolk. 



From the above records it seems clear, therefore, that double- 

 yolked eggs of doves are practically restricted in their production 

 to hybrids from wider crosses or to birds showing striking repro- 

 ductive abnormalities or to both of these. The history of the 

 few cases of double-yolks that are known would indicate, then, 

 that such eggs would not be expected to appear in the series in 

 which the two cases of twins were found. 



ON THE CAUSE OF THE FORMATION OF IDENTICAL TWINS 



At the beginning of this paper it was stated that on the basis 

 of the present data a suggestion could be offered as to the reason 

 for the occasional separation of the blastomeres which leads to 

 the production of identical twins. Possibly the data do not really 

 provide such a suggestion; but, knowing that the two eggs that 

 produced two pairs of identical twin ring-doves were from yolks 



* These eggs may of course be considered as of the same clutch; in this case the 

 abnormality would consist in the time interval being a four-day period instead of 

 the normal forty hours. 



