12 MUTATIVE REVERSIONS IX COTTON. 



that have been brouglit into a condition of uniform expression of 

 characters, but otlier kinds of reactions are not less interesting and 

 important. The uniformity or " breeding true" of a few generations 

 of individuals does not show that a stock is ' ' pure " ' in the sense em- 

 ployed by many writers on Mendelism. The idea that the Mendelian 

 relations of expression determine the "presence'' or "absence" of the 

 characters is a convenient assumption when the typical Mendelian 

 behavior appears, but reversions to "latent" characters show that 

 expression is no complete index of transmission. 



The idea that the ancestry of our cultivated plants is to be traced 

 back to uniform "pure" stocks that transmitted only single sets of 

 characters finds no warrant in the study of the more primitive t^^pes 

 and wild relatives of our domesticated species. Diversity of expres- 

 sion, instead of uniformity, is the rule in nature, and the transmission 

 of the diverse characteristics does not cease when uniformity of 

 expression is enforced through selection. Reversions show that the 

 underlying inheritance of diversity is not completely lost, nor the 

 power of the ancestral characters to reappear, even after long periods 

 of suppression. 



REVERSIONS INTERFERE WITH MENDELIAN EXPRESSION OF 



CHARACTERS. 



The tendency to reversion has to be reckoned as a serious obstacle 

 to the utilization of hybrid varieties unless the external conditions 

 and the processes of reproduction are under much more complete 

 control than with an open-fertilized field crop. 



Mendel pointed out a veiy useful distinction in showing that two 

 kinds of combinations of characters are represented among hybrids, 

 some stable in expression and others unstable. Intermediate char- 

 acters or reversions that arise from divergent tendencies of expression 

 may occur with much regularity in the first generation of a cross, but 

 may afterwards diminish or disappear. Even when the first gener- 

 ation shows uniformly intermediate characters, the later generations 

 tend to revert to more complete expressions of the parental characters. 

 The typical Mendelian relations appear in crosses between strains 

 that differ by definite tendencies to bring certain characters to full 

 expression or to leave them without expression, but it does not appear 

 that tlie analogies of such characters are applicable to all kinds of 

 plants or to all classes of hybrids. 



Some writers on Men(Jelism have supposed that inheritance is 

 governed by protoplasmic determinants, or "units," that are entirely 

 separate and independent , so that some of them can be changed with- 

 out disturbing the others, like changing the letters of a word or the 

 words of a sentence. Individual words from related languages can 



[Cir. 53] 



