FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 8 1 
certain cases but one small culture of offspring from these plants, selfed, 
has been grown, and have reason to believe that only one or two, con- 
taining but very few offspring, have been grown from others, and since 
it is probable that the offspring derived from these forms, as recorded 
by various workers, do not represent the whole of the possible progeny 
in any case, i. e., do not represent all that would have been obtained 
had means been employed to secure the germination of every seed 
capable of germinating — it is clear that the evidence upon which we 
would like to base our conclusions is not wholly reliable. This fact 
should be borne in mind throughout the discussions which follow. 
Nevertheless, the evidence as it now stands is not devoid of signif- 
icance. 
O. lata (15) produces, as we shall see, 0. Lamarckiana, 0. lata"^^ and 
a certain percentage of mutants, — the number of Lamarckianas 
greatly exceeding the number of latas in the cultures grown by Mac- 
Dougal and de Vries. The behavior of sublinearis, if a 15-chromosome 
form, appears to be similar to that of 0. lata, since the 31 offspring 
which de Vries obtained from sublinearis, selfed ('09, Vol. I., p. 401) 
were classified by him as follows: 19 Lamarckiana, 3 sublinearis, 
I lata, I nanella, 1 albida, 3 subovata, 2 oblonga and i gigas.^^ Here 
again we see that the number of Lamarckianas greatly exceeded the 
number of forms which reproduced the characters of the mutant 
parent. 0. bipartita (15), selfed, produced 0. Lamarckiana, 0. bi- 
partita, a few forms resembling the parent in most ways but having 
fewer flowers with cleft petals and extra-lobed stigmas than is common, 
and a few mutants. Here, also, a higher percentage of Lamarckianas 
than of bipartitas was obtained in the time allowed for the germination 
of the seeds, probably about 4 months. 
As previously stated, the number of chromosomes present in 0. 
scintillans is unknown, but it is probable that it is 15. This form, when 
selfed, according to de Vries ('13, p. 257), produces a variable number 
of scintillans; sometimes 35-40 percent or less and again as high as 
23 Bartlett ('15a, p. 103) calls attention to the similarity in the behavior of 0. lata 
and 0. stenomeres mut. lasiopetala. From the latter form, selfed, he obtained 60 
percent stenomeres and 40 percent lasiopetala. Mr. Arzberger has counted 14 
chromosomes in 0. stenomeres; the chromosome number of 0. lasiopetala has not yet 
been announced, but it is probably 15. 
De Vries has since concluded ('12, p. 34) that this plant and the other identi- 
fied as O. gigas in 1899 ('09, Vol. I., p. 327) were probably triploid, and not tetraploid, 
forms. 
