144 
H. H. BARTLETT 
and will make it possible to determine the hereditary qualities of the 
latter much more satisfactorily than has thus far been possible. The 
method, of course, will be to compare the hybrids between the parent 
forms with those between the gigas-muta.tions. The crosses in both 
cases will be between gametes with the same number of chromosomes, 
so we need not expect the high degree of sterility which prevents the 
study of hybrids between the forms characterized respectively by 14 
and 28 chromosomes. It would give increased significance to the 
discovery of the new gi gas -mutations if tetraploid species of Oenothera 
should be found in nature. I mention this fact in the hope that every- 
one who is interested in the problem of mutation may be on the look- 
out for them. I already have one wild species in my garden which 
appears to possess certain traits of the gigas-forms but Mr. Arzberger 
was unable to get his preparations ready in time to make a chromo- 
some count before this meeting took place. 
It has already been mentioned that in the genus Primula two types 
of gigantism occur. One type is characterized by a doubling of the 
chromosome number and seems to represent the same type of mutation 
as that which in Oenothera gives rise to the gigas-iorm. Gregory has 
found that not only are the chromosomes doubled in the tetraploid 
Primulas, but also the Mendelian factors for numerous characters. 
The hereditary behavior of these mutations is, therefore, entirely 
different from that of the diploid races. 
The differences in chromosome number which occur among species 
of the same genus represent changes which must necessarily have 
taken place abruptly. We can not imagine the origin of a gigas-ra.ce 
by gradual selection or by MendeHan segregation. Heribert-Nilsson 
has indeed offered a Mendelian explanation of Oe. gigas, but he has 
wholly neglected the cytological facts in the case. In Oe. stenomeres 
mut. gigas the chromosome count was made in the generation following 
the first appearance of the mutation. In the corresponding mutation 
from Oe. pratincola the count was made in the original mutation. 
There is the best of evidence, then, that the new chromosome number is 
acquired simultaneously with the new morphological characters. It is 
more reasonable to believe in a causal relationship between the cytologi- 
cal and morphological changes than to believe that the latter result from 
the hypothetical influence of hypothetical crossing in the indefinite past. 
It cannot be assumed that the modification of the chromosome number 
is itself due to Mendelian segregation, for there are too many facts 
