454 BOTANICAL GAZETTE [DECEMBER 
Several mutations of each class have been studied by the writer 
in more or less detail, and the results will soon be published. As 
already announced,” the interesting mut. nummularia belongs to 
class I, as do also all of the mutations characteristic of Lexington E. 
Mut. latifolia is a typical member of class II. There are mutations, 
of course, which show neither type of behavior, but they need not 
be involved in the present discussion. 
HERIBERT-NItsson’s hypothesis demands the recessiveness of 
mutations of class I, regardless of which way they are crossed with 
the parent. This condition is not fulfilled. It demands that the 
female gametes of the mutations of class II should be of one kind, 
and the pollen of two kinds. Neither is this condition fulfilled. 
His hypothesis makes no provision for the appearance of mutations 
in excess of one-third of the progeny. In this respect it is quite 
inadequate. On Mendelian grounds it is as difficult to account 
for too many mutations as for too few. His assumption is that after 
a homozygous and recessive condition has been attained in O. 
Lamarckiana, except for one of the plural factors which produce 
the Lamarckiana phaenotype, monohybrid splitting will take place. 
The one-fourth of dominant homozygotes will be eliminated, and 
therefore the progeny will consist of heterozygotes and recessives 
(mutations) in a 2:1 ratio. He has not attempted to explain how 
more than one-third of a progeny can consist of mutations, although 
he states in a vague and general way that the discovery of highly 
mutable strains is an argument in favor of his thesis. Nothing, he 
says, has made the mutation phenomena appear so exceptional as 
the low frequency of mutations. In his opinion, the high mutability 
of O. Reynoldsii has rendered the mutation fiction an absurdity. 
Further comment on this opinion is rendered unnecessary by 
the serious discrepancies between HERIBERT-NILSSON’s hypothesis 
and the facts. It can do no harm to point out, however, that even 
if mutations appeared through the operation of Mendelian segre- 
gation, as no one denies may sometimes be the case, it is still neces- 
sary to account for the origin of heterozygosis in the parent strain. 
The writer believes that mutations may often appear as a result 
of segregation, but that the antecedent heterozygosis has its origin 
in a mutative change. To attempt to account for the hetero- 
™ Amer. Jour. Bot. 2:146. 1915. 
