10 THE COEFFICIENT OF MUTATION IN OENOTHERA BIENNIS L. 



production of new forms takes place in pure lines of a species which 

 has come down to us unchanged during at least a century, since 

 the time Michaux discovered it in the United States and sent it to 

 Europe. 1 ) At least there is no direct recombination of characters 

 by actual crosses between different elementary types, such as we 

 ordinarily suppose to occur in polymorphic groups in nature. 



The other alternative, that it is logically inconceivable that 

 hybridism could have presided at the origin of new species, coin- 

 cides exactly with the current conception of the mutability in the 

 evening primroses. New forms originate through the evolution 

 of new characters, as in 0. gigas and 0. rubricalyx; 2 ) or through 

 the loss of existing ones, as in 0. nanella and 0. rubrinervis; or by 

 means of the appearance of qualities, which were probably latent 

 in the parent race, as in 0. lata and 0. scintillans.*) These cases 

 are evidently not recombinations of existing characters. If it is 

 conceded that the hypothesis of a hybrid origin does not apply to 

 them, it is obviously unimportant for the theory whether or not, 

 besides them, there are other instances which may be considered 

 as hybrid recombinations. 0. semigigas, which is a hybrid between 

 a normal and a mutated sexual cell, has never been considered as 

 an argument against the mutation theory. 



In cultures of chrysomelid beetles W. L. Tower has observed 

 hereditary changes which run almost parallel to the mutations of 

 0. Lamarckiana. He started from crosses between Leptinotarsa 

 decemlineata , L. multitaeniata and L. oblongata and obtained 

 constant races. When given proper treatment by changing their 

 environic factors, these races could be made to break up, and they 

 did so in a manner at least partially analogous to that of the evening 

 primroses. 4 ) 



i) The probable origin of Oenothera Lamarckiana. Opera VI, p. 579; 

 see pi. III. 



2) O. gigas is considered to be a progressive mutant on account of 

 its double number of chromosomes and its special behavior in crosses. 

 O. rubricalyx, which arose in the cultures of Gates from rubrinervis , 

 and which I cultivated this summer from seeds kindly supplied by him, 

 is perhaps the most beautiful among all the mutants of O. Lamarckiana. 

 Its red color is something quite new in the group. It behaves as a Men- 

 delian dominant in crosses with its parent species and is therefore ob- 

 viously of a progressive nature; see Gates, R. R., Amer. Nat. 45:600, 191 1. 



3) See Gruppenweise Artbildung, Berlin, pp. 244 — 260, 19 13. 



4) Tower, W. L., Evolution of the chrysomelid beetles. Carnegie 

 Institution of Washington Yearbook no 12:68 — 71, pi. 3, 1913. 



