No. 607] 



BATS AND EVOLUTION 



417 



the colored sheen on certain black ones by that name. In 

 every instance there appeared new recessive characters. 

 For every one of them we could see that crossing, re- 

 combination of genes, was the cause, not loss-mutation. 

 But it becomes clear that it is very difficult to be sure that 

 apparent cases of loss-mutation are not due to recom- 

 bination, unless the number of young in the generation 

 in which the novelty appears is ratlier large. If we 

 mate a species A to a species B, and some yellow or 

 long-haired or albino animals are produced in Fo we are 

 rather sure that recombination and not loss of genes 

 causes the novel form, even if the number of young is too 

 small to know whether the new character was found in 

 one animal among every four or among sixteen. But if 

 we mate two animals belonging to one single species, and 

 it happens that each possesses a gene which the other 

 lacks, the two genes having equal influence on the devel- 

 opment but of sucli a nature tliat animals lacking both 

 are albinos, or yellows, the production of a few animals 

 witli new recessive character may easily be looked upon 

 as mutation. In such an iiislance, it will ho found that 

 the two animals which in-odiiccil tlic hchn-ozygote who 

 gave the aberrant >(»uiiu- wonhl ln' fiuuid to be liomo- 

 zygous in respect to tho prc^cnco of •'(ho'" factor. For 

 if wo mate an animal liavin- [o tho now form vvzz. 

 all tho youiiu- will ho (lai-k. and none alhino. ( 'onvorsoly. 

 if Nv<' niato tho ZZ ].aront to Iho alhino, it will also bo 

 foun.l to bo li..n.o/ygou^, all tho young will bo colored. 

 In other word-, tost mating will in certain instances be in- 

 snf'ficicMit i)roor for the occurrence of a loss-mutation. 



Fji tlu" days when wo talked about ''unit-characters" 

 and the factors which "dotorminod" nnit charaoloi'v, it 

 was commonly hold that ei-ossino- in tlu^ wid(^st sens(>, 

 mating of forms with divors(^ i>-onot\ i>e could not c(Uint 

 for verv much in evolution, as it could onl\- riMMiuibino 

 oxi-tin- cliaractcr- and n.)t .Mvate ur^^ one-. We have 

 since learned t., h.ok U!M>n the -en.- as upon thin-s 

 whi(di hoi]) with other factor- in th.^ devdoimuuit to make 

 an organism develop, and wo now know that tho action 



