1 68 Experimental Zoology 



Cases of blended inheritance especially seem to come under 

 this heading. But so long as we do not know definitely what 

 occurs in these cases, it seems to me arbitrary to speak of unit 

 characters as immutable and quite unnecessary to make this idea 

 a cardinal point of the mutation theory. The behavior of cer- 

 tain characters in heredity shows that they do act as units, and it 

 is a great convenience to deal with them as such, but unnecessaiy 

 to push the matter so far as to hold that they are immutable. 

 If unit characters can be halved, altered, added to, or changed in 

 any w^ay, their immutability does not seem to be an essential 

 point of their characterization, and, as has been said, there is 

 some evidence to indicate that such changes may take place. 

 The idea of immutability is not likely to suggest itself when results 

 of this kind are considered, but in cases where complete domi- 

 nance of one form over the other occurs, the idea of unit char- 

 acters is more likely to arise. If, however, as I think probable, 

 we are dealing here with alternate or contrasted characters that 

 cannot both develop at the same time, the isolation of the char- 

 acters in question is due to their mutual exclusion rather than to 

 their immutabihty. Even if it be true that mutations take place 

 by definite steps without the presence of intermediate forms, it 

 does not necessarily follow that these steps may not subsequently 

 become subdivided by each crossing with the parent form. The 

 central idea of mutation remains, even if the unit- character that 

 marks the steps is capable of being changed. 



The results of hybridizing of forms differing by a single unit- 

 character seem to show that when in the first generation the 

 hybrids (i^j) are strictly like one of the two parents, the hybrids 

 of the new generation (^^2) 2-lso show complete development ^ of 

 one of the two characters in the extracted dominants and of the 

 other in the extracted recessives. This may be called the law of 

 incompatibihty. If, on the other hand, the dominance is incom- 

 plete, i.e. if both dominate in the first generation, there is in- 

 complete dominance in the second generation. This may be 

 called the law of compatibihty. If further facts estabhsh these 



^ In the sense of contrast, not necessarily of actual separation. 



