i86 BOTANICAL GAZETTE [march 



them grow separately. The experiment succeeds easily if the 

 rosettes have produced one or two roots of their own, however 

 young and slender these may be. 0. biennis nanella shows the 

 same character, and in August I succeeded in isolating from my 

 8 pure line dwarfs 8 rosettes, all of which have since developed 

 into healthy young plants with some long and narrow leaves, 

 followed by almost sessile ones, quite different from the rosettes 

 of normal 0. biennis. 



Moreover, two nanella mutants occurred in the cultures of 

 0. biennis sulfurea which I shall have to describe later. These 

 cultures were grown from self-pollinated seeds of the four sulfurea 

 mutants of Stomps (1913), and embraced over 1000 flowering 

 individuals, the flowers of which were pale yellow without exception. 

 Two of these plants proved to be dwarfs and were transplanted into 

 my experimental garden. Both of them have flowered with pale 

 flowers, have been self-pollinated, and yielded a sufficient harvest 

 of seeds. The coefficient of mutation in this race was therefore 

 o . 2 per cent, which does not differ essentially from the first instance 

 (o.i per cent). These dwarfs are the founders of a new race, 

 0. biennis sulfurea nanella, which I propose to cultivate next year. 

 Its pedigree name would be O. biennis mut. (19 13) sulfurea mut. 

 (1914) nanella. It is a double mutant, such as are quite common 

 in horticulture, and shows the way in which wild species would 

 have to be analyzed. 



I used the pollen of the O. biennis nanella of Stomps, in 1913, 

 for two crosses, which may be briefly mentioned here. In the first 

 place, I fertilized castrated flowers of the pure line of O. biennis. 

 The pollen was not abundant, and I got only 15 good seeds, all of 

 which have germinated and become stout flowering plants. They 

 differed from normal O. biennis in no respect and at no moment 

 during their development. Their self-pollinated seeds will have 

 to be sown next year. In the second place, I pollinated O. Lamarck- 

 iana with the pollen of O. biennis nanella. From this cross I had 

 a culture of 55 individuals, all of which have flowered. One of 

 them proved to be a lata mutant, having besides the lata marks the 

 same characters as its sisters. These were all alike and in no way 

 different from the ordinary and well known type of 0. Lamarck- 



