A NEW EVENING PRIMROSE. — GATES. 145 



There is no evidence that the 12.7% of empty or nearly 

 empty seeds represented a class of zygotes vvhi^^h was being 

 eliminated, although it is possible this may have been so. 

 The pollen was also examined at the end of the blooming 

 season, and three flowers from as many different plants gave 

 respectively 41.7%, 25.9% and 29.5% of bad pollen. Occa- 

 sional 1-lobed grains (as in Oe. gigas) were also observed. 

 Thus we find again a wild species producing a considerable 

 amount of sterile pollen and non-viable seeds. Yet the 

 flowers are close-pollinated and there is no reasonable basis 

 for assuming this sterility to be the result of crossing. 



The distribution of this species remains to be determined. 

 It probably occurs widely in western Nova Scotia, but many 

 species of Oenothera are found to be quite local in distribu- 

 tion. Another form in the muricata series, from Middleton, 

 has been figured elsewhere^. 



In conclusion, the writer would be grateful for Oenothera 

 seeds sent him from any part of Canada. Seeds should 

 always be collected from individual plants separately, and 

 should be accompanied by notes on the exact location and 

 habitat. It is desirable to collect seeds separately from 

 several individuals in a locality, to obtain a knowledge of 

 the range of variation. By this means it will be possible 

 ultimately to reach a thorough understanding of the variabil- 

 ity and evolution of this remarkable genus from the systematic 

 point of view. 



A portion of the expenses of these cultures was defrayed 

 by a grant from the Elizabeth Thompson Science Fund. 



New York Botanical Garden. 



»0p. oit.,p. 23, figs. 4, .5. 



