186 



MEMOIRS OF THE NEW YORK BOTANICAL GARDEN 



[Vol. 9, No. 2 



On the basis of the morphological (perfect vs. imperfect flowers) and distribu- 

 tional (tropical vs. temperate) evidence presented above, it is reasonable to 

 assume that the subgenus Lophotocarpus is more primitive than the subgenus 

 Sagittaria. A diagrammatic chart has been prepared (fig. 1) that suggests the sup- 

 posed relationships of the species of Sagittaria as well as their correlated geo- 

 graphic distribution. It is the writer's belief that Sagittaria and Echinodorus are 

 closely related, with Echinodorus the more primitive genus, as demonstrated by 

 its perfect flowers and more tropical distribution. 5. guyanensis, a pantropical 



FIGURE 1. 



non-plastic primitive species, is probably somewhat intermediate between these 

 genera,. The widespread plastic and primitive New World S, montevidensis has 

 probably changed but little from the ancestral prototype that gave rise to the re- 

 mainder of the genus. This species and all of the purely tropical Lopbotocarpi 

 have appressed sepals and mostly perfect flowers in the lower whorls. The warm- 

 north-temperate Lopbotocarpi have the sepals spread and only rarely have perfect 

 flowers. Apparently not all of the Sagittariae have the same origin. Those char- 

 acterized by pubescent filaments and non-sagittate leaves (S. graminea and allies) 

 appear to be more directly related to the Lopbotocarpi, They are generally more 

 tropical in distribution than the other Sagittariae and S a graminea is sufficiently 

 close to S a montevidensis to have probably hybridized with that species. S. pap- 

 ulosa and S, ambigua are allied to the S. graminea group, and are separated from 

 it only by the loss of pubescence on the filaments. The anomalous S. macrophylla 

 occupies a position intermediate between the balance of the Sagittariae and the 

 Lopbotocarpi. The remainder of the subgenus with glabrous filaments is further 



