291 R- Ruggles Gates. 



recessive ; and the new characters in the mutant ruhricalyx behave 

 as a simple Mendelian dominant in relation to its parent nibriiiervis. 

 Again, gigas X Lamarckiana according to De Vries yields an 

 intermediate race which remains constant, though in some cases 

 the behaviour is more variable, depending apparently on chromosome 

 mechanisms. These three types of hereditary behaviour are doubt- 

 less indicative of different types of germinal change involved in the 

 origin of certain mutants, a view which is fully corroborated by the 

 cytological evidence. 



As regards interspecific crosses in CEnothera, the most important 

 paper which has yet appeared is that of De Vries (1 1) on double 

 reciprocal hybrids between certain races of the Linnasan species 0. 

 biennis and O. muricata. In general, it was found that both 

 reciprocal crosses yielded uniform but strongly patroclinous hybrid 

 races which remained true in later generations. These reciprocal 

 crosses, which were therefore unlike, were in 1908 reciprocally 

 crossed with each other, i.e. (biennis x vmricata) x {muricata x 

 biennis and {muricata x biennis) X {biennis x muricata). In both 

 cases the offspring were a uniform race corresponding with the 

 "outside grand-parents." Thus (B X M) X (M X B) gave B or 

 biennis, while (M x B) x (B x M) gave muricata which remained 

 constant. The double reciprocal crosses therefore gave complete 

 reversion to one of the grandparents, while the characters of the 

 grandfather could not be transmitted through the mother and those 

 of the grandmother could not be transmitted through the father. 

 Similar results were obtained in crosses between races of O. biennis 

 from Holland and Illinois, also with biennis and cruciata, biennis and 

 strigosa, bietmis and Hookeri, and biennis and Lamarckiana. In the 

 biennis-muricata series of hybrids, what are called sesquireciprocal 

 crosses were made, thus {muricata x biejinis) x muricata [= 

 (M) B X M] and {biennis X muricata) x biennis [=(B) M x B]. 

 The former of these crosses is again identical with miiricata, and 

 the latter with biennis. Iterative hybrids were also produced in the 

 following manner : (M) B x B and B x (B) M, etc. The iterative 

 hybrids were like their hybrid parent. Thus (IVI) B x B^(M) B 

 and B x (B) M=(B) M. 



An ingenious explanation of these curious results has since 

 been offered by Goldschmidt (23) whose interpretation is based on 

 cytological study of these hybrids. Goldschmidt believes that a 

 condition of merogony exists, the male nucleus developing in the 

 cytoplasm of the egg, the nucleus of the latter degenerating. The 

 cross biennis x muricata would therefore contain in its cells a 

 muricata nucleus and biennis cytoplasm, while the cells of muricata 

 x biennis would be derived from a biennis nucleus embedded in 

 tnuricata cytoplasm. It is of course well known that the so-called 

 "male cell" in Angiosperms is really a male nucleus which has lost 

 its cytoplasm. 



If the explanation of Goldschmidt proves to be correct, then 

 these hybrids will furnish conclusive proof of the predominating 

 influence of the nucleus in inheritance, for in each case the hybrid 

 strongly resembles the parent from which its nucleus was derived. 

 Although the evidence for Goldschmidt's view is by no means 

 complete and final, yet it establishes a certain presumption in its 

 favour. Among the points figured are (i.) the degenerationof oneof the 



