342 BOTANICAL GAZETTE [NOVEMBER 
Tried in the same way, they gave the percentages of dwarfs shown 
in table III. 
TABLE III 
B. Dwarrs or O. gigas 
Generation et Total of seedlings Dwarfs Bhai eon o 
6th generation........ I 132 ° 2.0 
Wey Fiat teeth mee ued 2 105 ° 0.0 
Tg SCO ae Ray 4 155 I 0.6 
PG heme oe Diy 4 161 ° 0.0 
Re aes aes ee 5 159 25 15.7 
Whar Cpe cg soe 6 76 ° -0.0 
ee ee ies 7 151 re) 0.0 
Se 8 130 ° 0.0 
De ae res eae 9 124 19 15.0 
All in all, 19 specimens were studied. Among them three gave 
a percentage of 15-15.7-17.8, but the others gave only 1-2 per 
cent or no dwarfs at all. The dwarfs produced by this latter group 
were evidently due to ordinary mutability, but the figures for the 
former group differed too widely from these to be looked at in the 
same way. I consider them to be due to Mendelian segregation, 
and assume that the fact that they fall short of the expected 25 per 
cent is due to the difficulties of cultivation and to a less viability 
of the dwarfs as compared with the normal specimens.’ I chose 
no. 3 of the first group (17.8 per cent dwarfs) for continuing the 
experiment. 
If the segregation in this second generation followed the law 
of MENDEL, then among the plants of normal stature one-third 
must be constant in their progeny and the remainder must split 
up according to the same law. I succeeded in having a dozen of 
plants flower and ripen their seeds as annuals, fertilized them purely, 
and sowed the harvest in the spring of 1914. The result is given 
in table IV. 
Three of the individuals yielded no more dwarfs than in ordinary 
mutation, and the seven others showed figures which approach 
the Mendelian law as nearly as might be expected. If we combine 
these figures with the 17.8 per cent of dwarfs of the former genera- 
tion, we find for this about 18 per cent dwarfs, 57 per cent hybrids 
7 See GaTEs, op. cit., p. 89. 
