230 



Gates. 



the blades to any extent, but the midribs were usually deep red on 

 their dorsal surface (a rubricalyx character). The rosettes also per- 

 sisted longer and contained more leaves than ever appear in grandi- 

 flora rosettes under the same conditions. Fig. ii resembles rubri- 

 calyx somewhat more than does fig. 12. 



On the other hand, the leaves were much more broad-pointed 

 than in rubricalyx, and in most of the plants (fig. 12) the leaf shape 

 was certainly nearer grandiflora, especially in the distal part of the 

 blade. A striking feature was that the rosette leaves were much 

 larger, and hence the diameter of the rosette much greater, than in 

 either parent. Their average size was also distinctly greater than 

 in the reciprocal cross. Of the 147 rosettes only 58 formed stems and 

 came into bloom. The latter numbered 34 with red buds and 24 with 

 green buds, showing the heterozygous fRr) condition of the rubri- 

 calyx parent for this character. Considering the relatively small 

 numbers, this is not a wide departure from the expected equality. 

 The provisional detennination of the rosettes gave 42 R, 71 r, and 37, 

 most of which were probably i?'s, doubtful (i). The adult characters 

 of these F^ hybrids will be referred to in connection with the reci- 

 procal cross. The absence of dwarfs from the F^ of grandiflora x 

 rubricalyx must be accounted for by the dominance of the tall stem 

 (TJ of grandiflora over the dwarf character (i); for the rubricalyx 

 parent was heterozygous (Tt) for this character, and if self-pollinated 

 would have given dwarfs as did its sister plant (No. III. 10, pedigree i) 

 The absence of dwarfs from the F^ of rubricalyx x grandiflora must, 

 however, be accounted for by the dominance of the tall stem of the 

 rubricalyx parent over the dwarf condition latent in the race of grandi- 

 flora used as male parent. 



If we now turn to the Fj of rubricalyx x grandiflora (Culture 



88 

 No. — ) we find a family of 67 plants. As rosettes they were very 



uniform, differing only in the presence or absence of R. Fig. 13, photo- 

 graphed July 13, shows one of these rosettes. It is, like the reciprocal, 

 intermediate but nearer the rubricalyx parent, which it resembles in 

 shape (except that the blade and the points are distinctly broader) 

 and in crinkling. The dorsal midribs are also red, as in rubrinervis 



(1) Of the rosettes recorded as i-, six which bloomed afterwards proved to be 

 R, and 11 of the rosettes recorded as ß? proved to be R. On the other hand, 

 6 rosettes recorded as R turned out on blooming to be ;-. Hence 17 were gained to 

 the R's on blooming, while 6 were lost to them. 



