224 Gates. 



characteristic red blotches appearing on the earher rosette leaves. 

 The later rosette leaves, at the time stem-formation began, were more 

 broadly pointed than is typical for grandiflora, and there was a slight 

 amount of crinkling present, (see fig. 6). Six of the seedlings omitted 

 the rosette stage entirely (fig. 7). 



My culture of race No. 40 in 1909 numbered 43 plants, derived 

 from the self-pollination of one typical grandiflora individual. Of these, 

 38 were tall grandfiloras and 5 were dwarfs, having short internodes 

 but grandiflora foliage-characters. One of the dwarfs self-pollinated 

 yielded in 1910 a uniform dwarf race of 23 plants. The race of grandi- 

 flora used in my cross rubricalyx x grandiflora therefore contained tbe 

 capacity for producing dwarfs in the ratio of about 5 in 43. This matter 

 will be referred to again later. Dwarfing was therefore introduced into 

 the cross grandiflora x rubricalyx from the rubricalyx parent, and into 

 rubricalyx x grandiflora by the grandiflora parent. It later reappeared 

 in the F2 of both these crosses. A point which further demonstrates 

 that the new character exhibited by rubricalyx is dominant, is the follo- 

 wing. Both red- and green-budded plants appeared in the F^ of both 

 crosses above-mentioned, showing that the rubricalyx parent was hetero- 

 zygous for red in both cases ; while the dwarf character being recessive, 

 did not appear until the Fo in both cases. The rubricalyx parent of 

 the cross grandiflora x rubricalyx was therefore heterozygous for 

 both the red and the dwarf characters. If we adopt the current Men- 

 delian terminology, denoting the rubricalyx character by R and its 

 ab.ience frubrinervis ) by r; and similarly denoting tallness by T and 

 its absence (dwarfing) by /; then the rubricalyx individual in question 

 would be denoted by RrTt. 



The pedigree of the ancestors of the cross rubricalyx, x grandi- 

 flora is given in pedigree 2. The rubricalyx in this cross was also hetero- 

 zygous, therefore Rr, while the grandiflora was capable of producing 

 dwarfs and taUs, in the ratio 5 : 38. Dwarfs appeared also in certain 

 F2 families of this cross. 



C. Collateral experiments. 



In addition to the pedigrees of the parents in the two reciprocal 

 crosses, I have given in pedigreö i the results yielded by a number of 

 collateral families, so as to show their relationship to the general course 

 of the experiments. These families may be briefly referred to here. 

 Unfortunately, many of them were planted out too late in the season, 



