84 PROBLEMS OF GENETICS 



undoubtedly hereafter be identified in reference to all the various 

 kinds of chemical processes upon which the structure and func- 

 tions of living things depend. The identification of these proc- 

 esses and of the bodies concerned in them will lead to a real 

 classification of Substantive Variations. 



To forecast the lines on which such classification will proceed 

 is to look too far ahead. We may nevertheless anticipate with 

 some confidence that future analysis will recognise among the 

 contributing elements, some which are intrinsic and inalienable, 

 and others which are extrinsic and superadded. 



We already know that there may be such interdependence 

 among the substantive characters that to disentangle them will 

 be a work of extreme difficulty. The mere fact that in our 

 estimation characters belong to distinct physiological systems is 

 no proof of their actual independence. In illustration may be 

 mentioned the sap-colour in Stocks and the development of 

 hoariness on the leaves and stems, which Miss Saunders's experi- 

 ments have shown to be intimately connected, so that in certain 

 varieties no hoariness is produced unless the elements for sap- 

 colour are already present in the individual plant. 



The first step in the classification of substantive variations 

 is therefore to determine which are due to the addition of new 

 elements or factors, and which are produced by the omission of 

 old ones. A priori there is no valid criterion by which this can 

 be known, and actual experiments in analytical breeding can 

 alone provide the knowledge required. Some very curious results 

 have by this method been obtained, which throw an altogether 

 unexpected light on these problems. For example, in order that 

 the remarkable development of mesoblastic black pigment char- 

 acteristic of the Silky Fowl should be developed, it is practically 

 certain that two distinct variations from such a type as Gallus 

 bankiva must have occurred. I assume, as is reasonable, that 

 G. bankiva has genetic properties similar to those of the Brown 

 Leghorn breed which has been used in the experiments which Mr. 

 Punnett and I have conducted. Gallus bankiva was not available 

 but the Brown Leghorn agrees with it very closely in colouration, 

 and probably in the general physiology of its pigmentation. 



