934 Gates.—Pollen Formation in Oenothera gigas . 
among plants Drosera rotundifolia x D. longifolia (thirty chromosomes), 
and Oenothera lata x 0 . gigas (twenty-one chromosomes). The former 
hybrid has, I believe, only been observed in the wild condition where both 
parent species occur in the same locality, so that it is probably a hybrid, 
although there is no experimental proof of its hybrid origin. Its characters 
agree with this supposition, and it has never been suggested that this form 
might be a mutant instead of a hybrid, nor would the suggestion have any 
probability. Stomps (TO, p. 59) cites an observation of Geerts (' 09 , p. 52), 
who found a single case of a megaspore mother-cell in O. Lamarckiana 
having twenty-eight instead of fourteen chromosomes. Stomps makes the 
assumption that such a mother-cell would undergo reduction, and that 
fertilization with a normal male cell would occur, giving a form having 
twenty-one chromosomes. It is probable that the tetraploid number of 
chromosomes originated in this cell through its failure to complete a 
mitosis. It does not follow that such a megaspore mother-cell would 
undergo reduction to fourteen chromosomes followed by embryo-sac 
formation and fertilization of the egg. From what we know of the close 
relation between the occurrence of the tetraploid condition and apogamy, 
it is at least equally probable that this tetraploid megaspore mother-cell 
would develop apogamously, without undergoing reduction or fertilization, 
and would therefore give rise to a form having twenty-eight chromosomes. 
Indeed, this may have been the manner of origin of O. gigas , instead of 
the failure to complete a mitosis in the fertilized egg as I previously sug¬ 
gested. A decision as to whether this suspended mitosis occurred in the 
megaspore mother-cell or in the fertilized egg or young embryo will only 
be possible with further observations, but either of these hypotheses seems 
at present more probable, and more in accord with related facts, than an 
origin from the union of two unreduced germ cells. 
It further seems probable, though not certain, that whenever in the 
life cycle the duplication in the number of chromosomes occurred in 
O. gigas, other simultaneous changes took place in the germ-plasm. My 
previous paper (’09 a) on this subject was an attempt to explain as far 
as possible the peculiar characters of O. gigas on the basis of these changes 
in the dimensions of the cells. It is probable that such an explanation 
is inadequate to account for all the new characters of O. gigas , and that the 
appearance of the tetraploid number of chromosomes was accompanied by 
other changes. The further application of this method of analysis to 
O. gigas and other forms having a tetraploid number of chromosomes will 
aid in determining the nature and extent of any such change. It is to 
be hoped that an analysis of other cases from this point of view will be 
attempted. 
