Gates.—Pollen Formation in Oenothera gigas . 933 
pollination. This was presumably due to a defect of the ovules, but it is not 
stated whether the pollen was good. In 1909 I observed one plant of 
Oenothera which was wholly sterile in its anthers, but produced plenty of 
large seed capsules, and therefore must have borne normal ovules and was 
presumably pollinated from adjacent plants. Regarding this plant, I need 
only say that it was one of a culture derived from near Liverpool, England, 
and resembling O. grandiflora, Ait., in many of its characters. Material was 
collected for a cytological study, but has not yet been examined, so it~can« 
not be stated what was the nature of the sterility in this case. The plant 
differed in no other particular from other plants in the same culture and it 
bloomed abundantly, but the anthers in every flower examined at different 
times were dry and empty. 
The observations on this type of sterility indicate the correctness of 
Tischler’s (’ 08 ) conclusion, from his study of sterile hybrids, that the 
sterility does not depend upon any form of chromatin repulsion. It does 
not seem possible to define the general cause of sterility more definitely 
than ‘ lack of nutrition \ 
In my previous paper on Oenothera gigas (’09 a ), dealing with the 
size relationships of cells and nuclei, the point in the life cycle of 
O. Lamarckiana at which this mutant originated was discussed, and the 
evidence which is brought together in that paper pointed to the proba¬ 
bility that the tetraploid chromosome number originated in the early 
divisions of the fertilized egg, through the failure of a nucleus to complete 
its division after the fission of the chromosomes had taken place. This 
gives two sets of fourteen chromosomes which are identical two by two. 
Strasburger (TO) agrees with this view as to the manner of origin of the 
O. gigas mutant, but Stomps (TO) thinks it more probable that the tetra¬ 
ploid number of chromosomes originated through the union of two germ 
cells in each of which reduction had failed to take place, which would be 
in accord with de Vries’ theory regarding the origin of a mutation. As 
I have pointed out elsewhere (Gates, ’09 a, p. 544), there are several reasons 
why this manner of origin is less probable than the one I have suggested. 
If Stomps’ theory were correct, we should have a mutant occurring with 
twenty-one chromosomes, and it would be much more frequent in occurrence 
than in O. gigas. Such a mutant has never yet been found, and all the 
other mutants which are known have fourteen chromosomes, as in O. Lamarc- 
kiana. In the paper referred to (Gates, ’09 p. 545) I made a list of cases 
of related species one of which has the diploid and the other the tetraploid 
number of chromosomes, the inference being that, as in the case of O.gigas , 
the tetraploid species had originated from a related diploid species. 
Although a number of such cases are now on record, there is not, so far as 
I am aware, a single case of a species whose sporophyte has the triploid 
number of chromosomes. But several such hybrids are known, notably 
