84 
R. Ruggles Gates. 
about this conclusion if control counts of chromosomes had been 
made in the germ cells. 
At least two of these reputed 15-chromosome forms, rubrinervis 
de Vries and albida, “ breed true.” Miss Lutz points out that this 
can only take place (1) by apogamy, of which there is practically no 
evidence, or (2) when two gametes having different chromosome 
numbers only (one odd and one even) unite to produce viable seeds. 
Moreover, it is possible as regards rubrinervis that this individual 
derived from lata X lamarckiana and having a small extra 
chromosome may differ in some respects as yet unobserved from 
other rubrinervis mutants derived from lamarckiana direct and 
possibly not having the extra fragment. More critical study of 
albida may lead to a similar result. 
A number of other mutants recently described, such as cana , 
pallescens , lactuca , and liquida from CE. lamarckiana (De Vries 
1916b), which split in their offspring and are closely related to each 
other in several features, may contain an extra chromosome. The 
same may be true of (E. stenomeres mut. lasiopetala (Bartlett 1915c). 
These forms will be considered later. 
One of the most interesting results obtained by Miss Lutz 
(1916), is the existence of two types with a small extra chro¬ 
mosome—a fragment. In rubrinervis (one plant) 125 mitotic 
figures showed the small chromosome, and in 52 of them all 
the chromosomes, 14 +1 , could be counted. The small chromosome 
is constantly larger than in the new type aberrans, of which there 
were two individuals. In one plant 30 figures and in the other 
8 figures all showed 14 + 1 chromosomes. 
In a former paper (Gates and Thomas, 1914) it was shown that 
a variety of irregularities occur in the meiotic divisions in pollen 
formation of CE. lata , semilata and other forms which are more or 
less sterile in pollen. This included cases of fragmentation or 
pulling apart of chromosomes, particularly the odd one, on the 
heterotypic spindle. It might he expected that such fragments 
would occasionally be included in the daughter nuclei, and if they 
afterwards persisted they might be perpetuated by the mitotic 
mechanism. This is apparently what has happened in rubrinervis 
de Vries and aberrans , and it is significant that these three 
individuals all appeared in the offspring of lata x lamarckiana. 
The writer does not agree with Hance (1918) that they probably 
represent merely a temporary fragmentation of certain somatic 
chromosomes. Obviously if new forms can arise having gained a 
