1916] DEV RIES—DIMORPHIC MUTANTS 269 
The percentage of typical individuals in the second generation 
is about the same as for O. cana (25 and 53 per cent), for O. pal- 
lescens (35 per cent), and for O. Lactuca (43 per cent). From this 
it may be concluded that the 4 races have the same hereditary 
constitution which, moreover, is the same as in O. scintillans. 
The next year (1915) I cultivated a third generation of the 
second mutant of the table (mutant no. 1 from Jata). The har- 
vest had been small, as in the previous generation, and only 33 
seeds germinated. Of these 9 were liguida, 1 was pallescens, 1 
oblonga, and the others Lamarckiana. All of them have flowered. 
The percentage for liguida was 27, or about the same as in the first 
generation. Moreover, I have sown for each of the 3 cultures of 
1914 the seeds of one or two typical individuals, and also for each 
of them the seeds of two of the atavistic or Lamarckiana type. 
These 6 last sowings contained 150-300 seedlings each, together 
1311, of which 8 were mutants (3 oblonga, 4 lata, 1 cana); the 
remainder were all of the Lamarckiana type, no liguida occurring 
among them. The seedlings of the 4 liguida specimens gave the 
results indicated in table VIII. 
TABLE VIII 
THIRD GENERATION OF QO. MUT. liquida 
: : Percentage of 
Race issued from Total of seedlings euida 
pallescens (5 0 foe 84 25 
pallestens 06.4, <. 47 28 
SBIR OG, Eee 80 41 
Fate NO: 26s el 26 35 
TOES ee 237 32 
The countings were made in June and July in the boxes in 
which the seeds had been sown; the plants were all young rosettes 
with leaves 15 cm. long in the Lamarckiana type, and 6-10 cm. 
long in the liguida specimens. The differences were clear and sharp. 
The table shows that the splitting was almost exactly the same in 
the third as in the second generation. 
Dimorphic races do not seem to be rare among the mutants 
of O. Lamarckiana, and have been observed to spring also from its 
