( 8 ) 



change of conditions. A. betularia, when apparently a pure 

 race, had probably had in its ancestry very many not only 

 crosses with doubledayaria, but actual changes from betularia 



[xi 

 to doubledayaria and back again. Such an alternation of 

 inheritance probably went back not to the beginning of 

 In tularin as a species, but much further, to a time when the 

 present genus, subfamily, or even family, was represented by 

 one ancestral species, or even further. 



All dimorphic or polymorphic forms might be represented 

 as two (or more) forms combined together, as dominant 

 and recessive, the dominance not being conditioned as in 

 tin Mendelian relation purely by inheritance, but by the 

 environment. 



Levana and prorsa had been shown by Mr. Merrifield to 

 be quite interchangeable at an early larval stage. As the 

 change of conditions necessary had an annual cycle, there 

 was the appearance, without the fact, of an alternation of 

 generations. The regular change however kept each form 

 ready (<> appear at once ; but in the case of our melauic 

 forms, or of ehrysippus, the changes of environment were not 

 annual but rather secular, so that in the intervals the form 

 that is for the moment recessive receded more and more, and 

 conceivably might he eliminated, but in the result a change of 

 conditions operated gradually in bringing it to the front again. 

 That this semi Mendelian character of dimorphism goes 

 back far into the ancestry of dimorphic species was clear from 

 the circumstance t hit seems fairly obvious, thai each form of 

 a dimorphic (or polymorphic) group is naturally selected apart 

 from its associate, and (by selection) resents the natural 

 tendency for the two forms to coalesce by inheritance from 

 each other. Sexual dimorphism might possibly be the most 

 ancestral form of dimorphism, and from it other forms might 

 have arisen. At any rate, it fell in with the same vh 

 dimorphism that he had attempted to sketch, but which 

 doubtless would need a long, essay to illustrate clearly, and 

 without which it would not perhaps be easily appreciated 

 that the same mechanism exists whether the dimorphism be 

 apparently alternative or secular, whether the distinct forms 



