250 BOTANICAL GAZETTE [October 



the preliminary name of 0. biennis Chicago.^ I shall deal with this 

 one under the name O. saligna, and designate the new inconstant 

 mutants of 0. Lamarckiana as 0. cana, 0. pallescens, 0. Lactuca, 

 and 0. liquida. As far as investigated, they all follow the rule 

 that in every generation they split up into two ordinarily almost 

 equal groups of typical specimens and of atavistic individuals 

 which, in all cases, exactly dupHcate the characters of 0. Lamarck- 

 iana* Moreover, they show a relatively high degree of mutability. 



With one of them, 0. cana, I have made a number of crosses with 

 allied forms, in order to ascertain that it behaves in the same manner 

 as 0. scintillans, and that the same conception of heterogamy must 

 be applied here also. In this mutant the pollen carries only the 

 hereditary qualities of 0. Lamarckiana, and the specific marks of 

 the mutant are handed down to their progeny through the ovules 

 only .5 This conception of heterogamy may be considered to hold 

 good for the other inconstant types also. 



The same behavior is found in 0. lata, but since this form never 

 produces any fertile pollen in my cultures and has to be fertilized 

 by 0. Lamarckiana in order to produce seeds, the evidence which 

 it affords is less stringent than that given by the self -fertile dimor- 

 phic races. 



Oenothera Lamarckiana mut. cana. — -Among a number of 

 dubious mutants from O. lata which were cultivated as biennials 

 in 1906-1907, a plant was noticed in the third generation of that 

 family with narrower leaves of a gray color, evidently constituting 

 a new type. It was very vigorous, reached a height of about 2 m., 

 and was self-fertilized. It will be designated as 0. cana from 

 lata no. i, since the first family of 0. cana was derived from it. 



Next year the same mutant type was recognized among the 

 young rosettes, issuing from different samples of seeds of O. lata 

 (fig. i) . All in all there were 5 specimens of O. cana. In order to 

 determine the frequency of this mutant I have made two cultures 



3 Gruppenweise Artbildung, p. 52. fig. 18 and pi. 6. 1913. 



< In the wild condition such a splitting would evidently cause a race to die out 

 after a few generations, especially since the atavists are very fertile and much stronger 

 than the mutant form. As a, matter of fact, inconstant wild species of this type 

 are not known. See The mutation theory, Vol. I, p. 380. 



s Gruppenweise Artbildung, p. 273. 1913. 



