lOO 
ANNE M. LUTZ 
It may be asked, Why should we assume that 0. semilata de Vrles 
had 1 6 chromosomes rather than 14, or even 15? 
We know that most 14-chromosome forms breed true and there is 
evidence to indicate that a 15-chromosome form may breed true. The 
semilata de Vries whose constancy de Vries tested, appeared among 
the offspring of 0. lata (15) X 0. Lamar ckiana (14).^^ As is well 
known, this cross produces among others, 14-chromosome 0. nanella 
and 0. Lamarckiana, 15-chromosome 0. albida and 14- or 15-chromo- 
some 0. ohlonga, all of which are said to breed true. The three 16- 
chromosome mutants described in this communication are the only 
i6-chromosome forms which have been reported thus far ajid these 
three plants exposed no pollen whatever, hence no offspring have been 
obtained from them and we have no evidence to indicate that i6-chro- 
mosome forms are more likely than 15- to breed true. While it is 
probable we shall find that most forms having an odd number of chromo- 
somes are inconstant, whether the number be 15 or 15+, it does not 
follow that forms having an even number of chromosomes may be expected 
to breed true when the number is in excess of 14, except, perhaps, when 
this number is twice 14. It is probable that a Lamarckiana offspring or 
descendant having a double set of the original, parental 14 chromo- 
somes, as is possible in the case of 0. gigas de Vries, might be more 
stable than a i6-chromosome offspring, yet it is quite certain that the 
28-chromosome mutant offspring of 14-chromosome forms with which 
we are familiar do not breed true in the same sense as O. Lamarckiana. 
The progeny of 0. gigas de Vries do not all duplicate the parental 
individual in every character, as in the case of 0. Lamarckiana. As 
is well known, a number of types are found among the offspring, yet 
since they have many characters in common and since one type, when 
selfed, seems to produce the same lot of offspring as any other, we 
speak of this form as constant. Gates's 28-chromosome Palermo 
strain of 0. gigas also "showed a considerable range of variation, 
though not so great as in the Amsterdam race" (p. 121). Heribert- 
Nilsson's race (Heribert-Nilsson, '12), representatives of which Gates 
('15a, p. 124) has recently shown to have 28 chromosomes, also pro- 
duces a variable lot of offspring. Furthermore, Bartlett ('i5<:) has 
announced that the 28-chromosome mutant offspring of 0. pratincola 
which appeared in his cultures, namely 0. pratincola mut. gigas gave 
De Vries recognized three semilata mutants in his cultures. All had lata 
mothers. 
