Studies on variation and selection. j;5t) 



In fact the only two Hooded children from this mating are light. 

 76 X 60 is another mating of the same nature, giving the same result. 



This pedigree gives no answer to the question, as to what would 

 be the characters of an Ab rät, an Irish, which when mated to light 

 Hooded would produce only Irish and light Hooded. Such a rat has 

 however been figured by Castle in the only individual pedigree 

 recorded by a picture, and this picture shows it to be an Irish like 

 the AB Irish ones. 



We have now shown the existance of still another gene, present 

 in some rats, absent from others. In rats, which are Hooded, pre- 

 sence or absence of this gene makes the difference between dark 

 Hooded and light Hooded. It will probably be possible to find still 

 more of these gi'nes, by selection-experiments within a family of light 

 Hooded rats. From our pedigrees the first evidence for the existance 

 of our factor B, could only be got from the pedigree of the Hooded 

 family. 



It is further clear, that by calculating the average grades of 

 parents and children, nothing whatever would have been got from 

 these pedigrees. Looked at, side by side with the table of pictures, 

 such a statistical rendering of the facts becomes absolutely meaning- 

 less. We now know several things. If, from one lot of rats, Castle 

 and his assistants could breed extremely light hooded as well as 

 practically self-coloured ones, this may mean that for at least two 

 perfectly independant genes, this family was heterogeneous, either 

 because some rats were heterozygous for one or both, or because 

 these factors were present in some rats, absent from others. And if 

 his material was possibly heterogeneous for the factors we know, 

 which influence extent of pigmentation, it is very improbable that 

 it happened to be pure for other such genes, which we do not yet 

 know separately. We may therefore, with all safety, conclude tliat 

 the selection-experiments of Prof . Castle with Hooded rats do not make 

 it in the least probable that selection could change a gene, as it is evident 

 that he did not take any of the indispensable preliminary precautions 

 which alone could have made his experiments fit to elucidate the 

 question to which he wanted the answer. 



There is one point which needs clearing up. It may seem curious 

 to many, that, if selection in Castle's series acted by eliminating 

 genes, it did not have very much more striking results. 



We, in our two or three generations, had far better results than 

 Castle in six. It must be noted, that Castle did not strictly inbreed 



