1^0 Arend. L. Hagedoorn and A. C. Hagedoorn. 



his rats, and especially, that lie did not select them very 

 rigourously. 



We read for instance, that from light Hooded rats, he obtained 

 some without any backstripe. He then did not continue with rats 

 of this cathegory only, but he chose as parents for the next gene- 

 ration some without backstripe, and some with a reduced backstripe. 

 This example shows, that he did not, as a matter of principle, continue 

 only from parents, having the chosen character best developed. 



Castle, throughout his paper, emphasizes that selective breeding 

 is necessary very slow work, and that, therefore, the effect of selec- 

 tion may escape us "Mendelians", looking only for gross differences. 

 But this slowness of the effect of selection, in those cases in which 

 it must have results, is only apparent. In Castle's selection-series 

 this slowness is certainly unduly flattered. 



It might be said by some selectionist that Castle's experiments 

 with rats are not the only ones, which seem to prove the modification 

 of genes by selection, and that, even now this one case happens to 

 be disproven, other similar cases still stand. To this we can only 

 answer, that all those cases are very similar, and that especially they 

 all have the same defects. In no one case have the experiments been 

 started with material which was strictly inbred, and which had only 

 the slightest claim to genetic homogeneity. 



It is evident, that it would be unfeasible to work over all those 

 cases which have been published as proof for the supposition that 

 selection can modify a gene. Such work would be hopeless; while 

 we should be engaged in breeding-experiments, patiently disentangling 

 the genetic factors overlooked, the authors of these cases would be 

 busy getting up new instances for us to disprove. 



It is easy to see, how one can, by selecting the animals which 

 vary in one direction, make a family which lacks such factors as by 

 their cooperation to the development tend to influence it in the 

 opposite direction, if the material is impure for such factors. 



And it is also easy to see, why, in a strictly inbred family 

 (brother to sister, parent to child) such selection has a much more 

 speedy effect than in mixed breeding. 



But it is obvious that, when such a change is accomplished, 

 selection in the opposite direction will never succeed in bringing the 

 lost genes back, and therefore it would be impossible in an inbred 

 family, to bring it back to its starting-point. Still, this would be 



