1909] GATES—CHROMOSOMES IN OENOTHERA 193 
in different individuals from 21 to 28 according to the number of 
supernumerary chromosomes present. The distribution of the super- 
numeraries during spermatogenesis is also irregular, so that the num- 
ber varies in the different sperm nuclei of an individual. Occasional 
irregularities in the distribution of idiochromosomes during reduction 
were observed in the form having 22 chromosomes, both idiochromo- 
somes passing to the same cell. This is believed to be the origin of 
_ Ue variation in different individuals, the supernumeraries being 
| merely duplicates of the idiochromosomes. 
Similarly, I have observed occasional irregularities in the chromo- 
some distribution during reduction in nearly all the mutants of 
_ Oenothera examined, and this doubtless accounts for the different 
_ humbers of chromosomes found in different individuals, the extra 
chromosomes being duplicates of others already present, and the 
presence of 20 chromosomes in one individual of O. lataXO. gigas 
being due to the absence of one chromosome from one of the germ cells 
Which produced that individual. 
_ Montcomery (27) first proposed the theory that in synapsis 
homologous chromosomes of maternal and paternal origin pair with 
each other, and much subsequent work, especially with animal 
chromosomes, sustains that view. As I have already shown, pairing 
of Oenothera chromosomes frequently fails to take place, and this 
allows a chance for the irregularities in chromosome distribution 
4 Which occur. Boverr (3, p- 54) has suggested that when there is an 
_ absence of pairing in synapsis in any organism, its chromosomes are 
hereditarily equivalent. In Oenothera there is clearly a tendency to 
pair, but it is only partly carried out. I have already compared (11, 
_ P- 25) the behavior of the chromosomes in this hybrid with the con- 
7 dition described by ROSENBERG (31) in the Drosera hybrid having 
o+20 chromosomes. The method of segregation of chromosomes 
_ In the heterotypic mitosis is evidently different in the two hybrids, 
4nd this was used as an argument in favor of a different method of 
Teduction in the two cases. In the Oenothera hybrid the 10-11 
Segregation shows that the segregation cannot be between chromo- 
Somes of maternal and paternal origin, but it must be remembered 
_ im this connection that the 14 paternal chromosomes are very prob- 
a ably a double set of 7 O. Lamarckiana chromosomes. The regu- 
