198 THE PHYSIOLOGY OF REPRODUCTION 



Moreover, the difficulty experienced by Wood l in superimposing 

 the complete hornlessness of Suffolk sheep upon the white face 

 of the Dorset horns, is probably another example of the physio- 

 logical correlation subsisting between different, and apparently 

 unconnected, structures. Originally, this case was regarded as 

 one of simple superposition, and Bateson 2 describes the hornless 

 character as having been transmitted " pure," but subsequently 

 many of the so-called hornless sheep were found to have grown 

 scurs. The explanation which I tentatively suggest is, that the 

 character of pure hornlessness was somehow or other incompatible 

 with the pure white-faced character, these two characters being 

 ordinarily indications of two sorts of metabolism, in just the same 

 kind of way as the beef -producing quality and the milk-producing 

 quality seem to be to some extent incompatible in cattle. 

 I am inclined to go further, and to suspect that many of the 

 other Mendelian cases, when examined more critically, will show 

 that no one character can be superimposed upon another, in 

 experimental breeding, without altering, though perhaps only 

 very slightly, the character upon which it has been superimposed. 

 It is only when the amount of alteration is minimal that 

 the transmission of pure characters is apparent, according to 

 Mendelian expectation ; but experimental evidence has shown 

 that there are considerable numbers of such cases. It is a 

 legitimate field of work for the biometrical school of biology 

 to determine by statistical methods the extent to which varia- 

 tion occurs as a result of attempted superposition of characters 

 which in their " pure " state are physiologically incompatible. 

 Furthermore, a latent character may be regarded as one, the 

 outward manifestation of which is incompatible with the existing 

 kind of metabolism, but which is capable of reappearance as 

 soon as the conditions become favourable. But because it is 

 helpful to assume that latent characters are present in some 



1 Wood, "The Inheritance of Horns and Face-Colour in Sheep," Jour. 

 Agric. Science, vol. iii., 1909. 



2 Bateson, Mendel's Principles of Heredity, Cambridge, 1909. No doubt, 

 however, it is arguable that the scurs themselves represent unit char- 

 acters, and that if the scurs are of different kinds, these also represent 

 unit characters (which have hitherto somehow remained "latent"), and 

 that if they occur with different degrees of development, these again are 

 unit characters. And so on. 



