LAMARCKIANA SEM1GIGAS. 555 



of flowering, the mutants belonging to the same type were side by 

 side. Thus the different groups could easily be distinguished, and 

 also the slight deviations in the individual characters within each 

 group could be subjected to a careful comparative study. Most of 

 the 81 mutants flowered during July and August, and ripened their 

 fruits in the fall. They were stout, healthy plants, whose characters 

 could be compared with those of the older mutant races, all of which 

 flowered at the same time in the experimental garden. 



This comparison was begun in the stage of rosettes and con- 

 tinued through the summer, until the ripening of the fruits. As a 

 rule the rosettes show their type distinctly at the time of planting, 

 or, on weaker plants, some weeks later. Then the young stems 

 develop with a less differentiated foliage, but the spikes of flower 

 buds are seen to develop once more the marks of the groups. Moreover, 

 the deviations from the main types are usually small during youth, 

 but become more pronounced with advancing age. 



This article reports only the results of the cross semigigasxvelutina. 

 The reciprocal cross would not produce a similar deformity. We may 

 conclude this from a study of the effect of the pollen of our plant 

 in other different crosses, made in 1922, partly on other species, 

 partly on the parent type and some of its mutants. For each of 

 these crosses 10—15 or more flowers were pollinated, and yielded 

 40—60 or more viable seeds. In May the rosettes had reached the 

 stage in which the marks of laeta and velutina are easily distinguished, 

 but no velutina were found. Thereupon 12 specimens from each 

 cross were planted out and cultivated during the whole summer, 

 until they began to ripen their seeds. In the crosses with 0. biennis, 

 0. grandiflora, and 0. blandina they were exactly like the laeta 

 from the corresponding crosses with Lamarckiana. In Lamarckianax 

 semigigas they repeated the marks of the mother, and in latax 

 semigigas they were partly lata and partly Lamarckiana. No other 

 types were observed in these cultures, which embraced in all over 

 400 rosettes and 60 flowering plants. The high degree of mutability, 

 shown by the cultures derived from the egg cells of semigigas, failed 

 in those from the pollen. From this experiment, therefore, we con- 

 clude that the pollen of our semigigas contains, among its viable 

 gametes, only or almost only such of a sufficiently pure laeta type. 



Doubling of chromosomes 



Stomps (11, 12) was the first to introduce the study of 0. mut. 



