Breeding experiments which show that hybridisation and mutation etc. 247 



Table V. 



Number 

 oi 19 1 2 culture 



Number of Fi 

 parent plant 



Character of buds 

 in parent F, plant 



Tall. 



Dwarf. 



Ratio 



grandiflora x riibricalyx, F2 



XI. 9 



I. I 



green buds 



60 



23 



22 

 50 



2.73:1(1) 

 I : 2.17 



rubricalyx x grandiflora, F2 



X. 3 

 VII. 5 



red buds 

 green buds 



(grandiflora x rubricalyx) x [rubricalyx x grandiflora). 

 Cult. 77 . . . .1 XI. 9 X VII. 5 [green buds x green budsj 45 I 8 I 5.6 : i 



reciprocal cross the conditions are quite different, only two Fg families 

 out of nine containing dwarfs. This result would also be expected 

 from the fact that the parent, grandiflora, which introduced dwarfism 

 into the cross, only produced dwarfs in the ratio of 7.6 : i. The actual 

 ratio of T and t in three of these four families (Cults. 53, 66, 60, see 

 table \') and also in the double reciprocal cross (Cult, yj) which con- 

 tained dwarfs, are in full accord with the expectation based on these 

 facts. Thus the two families from rubricalyx x grandiflora give T : t 

 ratios which are not significantly different from each other and whose 

 average, 6.37 : i, is not very different from the ratio 7.6 : i in the 

 grandiflora grandparent. The ratio obtained in families 60 and 66 

 are therefore explicable on the simple hypothesis that the tendency 

 to produce dwarfs in about the ratio of 1 in 8 is inherited from gene- 

 ration to generation. But it is wholly inexplicable on the Mendelian 

 hypothesis. Still more so is the ratio of practically 2 dwarfs to i tall, 

 obtained in Cult. 54. Mendelianä who cling to the conception of fixed 

 and imalterable units can only fall back upon "reversed dominance" 

 or some equally fantastic hypothesis. But if, as we have shown to 

 be the case with the character R, different individuals exhibit different 

 degrees of prepotency, then even a wide fluctuation in the percentage 



(') The ratios in cults. 53 and 54 are significantly different, such divergence 

 occuping only once in 3000000 times. But cults. 60 and 66 are not significantly 

 different. 



(') In Table III the total number of plants catalogued in this culture is 50. 

 The remaining 10 include the 9 dwarfs and i intermediate in pigmentation. 



