1915] DE VRIES—~A MENDELIAN MUTANT 339 
From time to time I have noticed the same deviating percent- 
ages in my own cultures. Thus, for instance, I fertilized in r910 a 
specimen of O. gigas by its own pollen, and among 50 seedlings 
of its offspring 10 were dwarfs, pointing to a percentage of about 20 
per cent.4 Similar facts have since occurred more than once in my 
cultures. 
SCHOUTEN and Gates have interpreted these figures as indicat- 
ing a Mendelian proportion of dwarfs, and on this assumption the 
parent plant would have been a mutant hybrid in the same sense 
as explained above for O. brevistylis. Mutant hybrids would then 
occur in a race which produces dwarf mutants also, and the latter 
would then, of course, have to be considered as the products of the 
combination in fertilization of two sexual cells, both of which had 
mutated into nanella.° The production of dwarfs from O. gigas 
would then follow the same process which is to be assumed for the 
origin of O. gigas itself from O. Lamarckiana; and the copulation 
of two similarly mutated cells would then more easily be accessible 
to experimental investigation. 
In order to verify the exactness of this conception I have fol- 
lowed up the progeny of such a presumed mutant hybrid, and on the 
other hand have made crosses between O. gigas and O. gigas nanella. 
In both cases the truth of the assumption was easily ascertained. 
Mutations of single gametes may be discovered by different 
means in other instances also, the production of potential nanella 
gametes by O. Lamarckiana being the most likely to be betrayed 
in this way.’ I have observed such cases in crosses between 
O. Lamarckiana and O. rubrinervis. From these ordinarily two 
types arise in the first generation, one of which resembles the 
mother and the other the father. In my book on Gruppenweise 
Artbildung I have called them “‘Lamarckiana”’ and “ subrobusta.” 
Both types are usually constant after self-fertilization. But, 
from time to time, individuals appear which in their progeny 
produce an unexpected number of dwarfs. The following cases 
may be adduced. 
4 DeVries, Huco, Gruppenweise Artbildung. p. 340. 1913. 
5 Besides the production of gametes for gigas by O. Lamarckiana, as shown by 
the occurrence of specimens of semigigas in self-fertilized strains of the parent — 
or by the eoluniee of the Hero-type in crosses of O. Lamarckiana with allied spe 
