Wilson — Unsound MendcUan Developments. 417 



founH. Cinnamon agouti and silver-fawn gave cinnamon agouti, dilute 

 cinnamon agouti (the new colour), chocolate, and silver-fawn in the 

 ratio 9:3:3:1/ Thus dilute cinnamon agouti differs from each of the 

 others in one factor, and it can only be Pqr. If we now write down the 

 eight groups., we see that they are really such a set as might be produced in 

 the second crosses from two parents differing from each other in thi-ee pairs 

 of characters. We may therefore set down in addition the numbers for each 

 group. 



In this case the two methods of analysis arrive at the same result ; but 

 the presence and absence method arrived at this result because no effective 

 recessive was disclosed to raise confusion among the absences, and there was 

 therefore no need to introduce a factor more than the case could contain. 

 Had such been disclosed, the presence and absence theory would have 

 had to give it a name ; its dominant would then have had two recessives 

 — its absence and the disclosed character— and confusion would have resulted. 

 So long as the presence and absence theory introduces no superfluous factor 

 it works like the Mendelian theory itself, although those who use it may 

 imagine they are working with the other. 



Since the presence and absence theory is unsound, it follows that any 

 theory depending upon it is also unsound, and that work done upon the 

 presence and absence or upon any dependent tlieory will have to be revised. 

 In this connection it may be suggested that the first essential is that more 

 attention be given to the logical consequences of the Mendelian formulae. 



It is not necessary to give less attention to micrographic and internal 

 aspects, but it is necessary to give more to macrographic and external. 



By way of illustration, several tentative solutions of cases suggested by the 

 data given in Professors Bateson's and Punnett's volumes might be put 

 forward. It must be understood clearly, however, that these solutions are 

 only tentative because the complete data have not been available. 



' Journal of Genetics," vol. i, No. 2, p. 177. 



3q2 



