126 



Gates, Tetraploid Mutants and Chromosome Mechanisms. 



narrow -leaved form 20 ). Fig. 6 shows the rosette photographed 

 July 10 th , when the central shoot was beginning to develope, and 

 fig. 7 shows the mature plant. This' is the first time I have suc- 

 ceeded in getting one of the very narrow-leaved variants to flower, 

 but it showed a very high degree of sterility, producing no good 

 pollen. Its chromosome number will be determined later. The 

 buds were much smaller than in normal gigas : but were not measured. 

 They were greenish in colour, and the earlier ones showed asym- 

 metry in shape similar to that of lata. The petals were also some- 

 what crumpled, and the sepals rather short. The stamens were 

 sinuous and contained but little pollen, one stamen being found 

 attached throughout its length to a petal. 



Table III. 

 Measurements of typical capsules. 



The above Table gives the measurements of several capsules 

 from each of the races mentioned. The lata capsules are shortest, 

 because they produce fewest seeds. In 0. yiyas Italy the capsules 

 are strongly rhomboidal in cross-section, their greatest diameter 

 being, as in most of the other forms, just above the base. Here 

 again the short length of the mature capsules is undoubtedly to 

 be attributed to the small number of ovules which it matures. 

 As shown in Table II, the ovary before fertilization is longer than 

 in 0. Lamarckiana, while in 0. yiyas Sweden it is very much longer, 

 as is also the ripe capsule. The fact that the capsule of the latter 

 usually has a broad expanded base has already been referred to. 



The Pollen Grains. 



During the past season I devoted some time to the examina- 

 tion of Oenothera pollen, the various types of grains being determined 

 numerically, with interesting results. The pollen of yiyas is charac- 



26) I have repeatedly found that when old seeds are sown and only one or 

 two germinate, the plants produced are almost invariably aberrant. This would 

 seem to indicate that the seeds which produce aberrant individuals are more viable. 

 I have had this experience too frequently to attribute it to mere chance. 



