478 COBB AND BARTLETT: INHERITANCE IN OENOTHERA 



The actual behavior of the various cultures is very satisfac- 

 torily accounted for by means of this formulation. That the 

 revolute-leaved mutations breed true for revoluteness is shown in 

 table I. In the same table will be found the evidence that the 

 revolute-leaved mutations give matroclinic progenies when 

 crossed with the f. typica of the corresponding strain. It is the 

 prevalence of matroclinic inheritance in these mutation crosses 

 that makes a purely Mendelian explanation of their behavior 

 impossible, and that requires the conception of a and jS gametes. 

 The data in regard to matroclinic inheritance are in some respects 

 inadequate, and will be amplified as rapidly as possible. 



The failure of Lexington C, and the other strains that resemble 

 it, to throw revolute-leaved mutations is attested by the accumu- 

 lated data for hundreds of cultures, aggregating many thousands 

 of plants. Since the evidence is negative, it would be idle to 

 assemble here the data supporting the conclusion that Lexing- 

 ton B is the only strain in which such mutations occur. The 

 data regarding the mutability of Lexington E are in part pub- 

 lished (see footnote 2). 



According to the scheme above, one would expect the 

 cross typica C X mut. formosa to act in a manner quite differ- 

 ent from its reciprocal, which we have been considering. In- 

 stead of a Mendelian segregation of flat-leaved and revolute- 

 leaved plants in the F2 generation, one would expect only segre- 

 gation with regard to the capability of occasionally undergoing 

 the mutative change from flat to revolute. Segregation of this 

 ability would mean that revolute-leaved plants might appear in the 

 the F2 generation, but, if present, each one would be the result of 

 a separate, individual change ; they would not be due to the in- 

 heritance of a changed condition, as was the case in the revolute- 

 leaved recessives in the 3 : i segregation described. 



Such a character, the capability of mutation, is not easy to 

 deal with, since it is impossible to be sure that a particular parent 

 might not show mutabiUty if more of its seeds were germinated. 

 The degree of mutability does not appear to show great uni- 

 formity from generation to generation, as may be seen by refer- 



