46 



ical to be unsatisfactory > when one considers that the 

 raried combinations of the characters of the generally 

 recognized species here presented strongly suggest inter- 

 crossing and the possible presence of "unfixt" (heter- 

 ozzgous) forms* 



Plainly nothing but pedigree cultures can solve 

 the problem, and these, under satisfactorily controlleft 

 conditions, would no doubt be extremely difficult, be-- 



* 



cause of the perennial nature and slow growth of the 

 plants, their peculiar habitat -requirements, and the 

 rarity of seed-production. 6ray*s Manual says of D. 

 diphylla "pods rarely maturing", but seed-produetion 

 seems to be only occasional with the other two species- 

 types* Seed of D » laciniata and D* maxima has been 

 found, but apparently the capsules dehisce while the 

 seeds are still green in color and unhardened, and 

 special treatment, with avoidance of drying, may be nec- 

 essary to secure germination under controlled conditions. 

 It seems best, on the whole, to group the variant 

 types, provisionally, around the 3 accepted species of 

 this region, and to describe the characters of each in 

 relation, primarily, to the type of the species appar- 

 ently nearest, with reference also, where necessary, to 

 the intermediate types of Gray*s Manual. This procedure 



