82 
ANNE M. LUTZ 
70-80 percent of the total number of offspring reproduce the char- 
acters of the mutant parent. The remainder are, for the most part, 
0. Lamarckiana, but with a considerable number of 0. oblonga ("oft 
bis 20%") and a few other mutants. 
0. semilata Gates (15), is an inconstant form, as Gates and Miss 
Thomas ('14, p. 532) and Gates ('15a, pp. 114-115) have shown, pro- 
ducing 0. Lamarckiana, 0. semilata Gates, a few 0. lata which may- 
be classed as mutants and (p. 114) others "forming a continuous 
series running to Lamarckiana." 
0. elliptica, having I5(?), chromosomes, reverts almost entirely to 
Lamarckiana, according to de Vries ('09, Vol. I., pp. 397-398). From 
one 1895 mutant, selfed, he obtained "some hundred of seedlings," 
all of which proved to be ordinary Lamarckiana. From a second 
mutant of the same year 500 offspring were secured, I of which was 
elliptica, and the remainder Lamarckiana. A third 1895 mutant 
"gave rise to 27 seedlings not one of which was an elliptica.'' From 
an 1896 mutant he obtained 32 offspring, 5 of which were elliptica and 
the remainder Lamarckiana; from an 1899 mutant he secured about 
100 offspring, all of which were 0. Lamarckiana. 
0. lata rubricalyx, in which Gates and Miss Thomas counted 
15 chromosomes, when selfed, according to Gates {'15a, p.- 288), 
produced a nearly uniform lot of offspring (44 plants), " all having the 
red pigmentation of rubricalyx, but were intermediate between rubri- 
calyx and grandiflora in foliage and buds. . . . The plants which 
were examined had 14 chromosomes, as was doubtless the case with 
all of them." No lata rubricalyx plants were found among the off- 
spring. 
While all of the above forms are clearly inconstant, de Vries's 
researches indicate that a 15-chromosome form may breed perfectly 
true. He selected 5 biennial albida plants ('09, Vol. I, p. 229) in 
1897 and grew a second generation consisting of 86 individuals in 1898 
and a third, consisting of 36, in 1899. "Both generations," he adds, 
"were absolutely constant and exhibited no signs of reversion." 
If 0. oblonga be a 15-chromosome form, it indicates even more 
strongly (because of the larger number of offspring obtained) that a 
15-chromosome form may be constant. During a period of over 13 
years, de Vries (pp. 346-348; also, '13, p. 315) selfed a number of 
oblonga mutants and obtained a total of 2,919 offspring, all of which, 
with the exception of ii mutants (7 rubrinervis, 3 albida, i elliptica) 
were oblonga. 
