1955] 



REVISION OF THE GENUS SAGITTARIA 



181 



triflora, with ternate leaves, gives all indications of being either a Ranunculus 

 or a Potentilla. One other Linnaean species of 1753, Alisma subulatum t has been 

 since transferred to Sagittaria e 



Adanson (1763, 2: 459) revived a pre-Linnaean name, Sagitta, for the genus, 

 which of course cannot be maintained. For the balance of the 18th century and the 

 earlier part of the 19th century, numerous species were proposed by various au- 

 thors, some of them new, but for the most part they were found, even in those 

 days, to be synonymous with previously described ones. 



Rafinesque (1825, p. 3) proposed two new genera, Diphorea and Drepachnia, 

 for some American Sagittarias. His basis for separation (stamen numbers) is so 

 tenuous and so variable a character, that it is useless for segregating even forms. 



Kunth (1841, 3: 155-162) attempted to arrange in some orderly fashion the 27 

 species he included in the genus. He set up three sections, two of which were 

 unnamed, the third, Lophiocarpus, was described as being closer to Echinodorus 

 than to Sagittaria, In this section were several species which today have all been 

 united with the pantropical S, guyanensis. 



Lophiocarpus was raised to generic rank by Miquel (1870, p. 49-5 l)o It in- 

 cluded only the Asiatic species of those listed by Kunth. Micheli (1881) enlarged 

 Lophiocarpus to include S. seubertiana (here treated as a form of S. guyanensis 

 with entire-winged smooth-faced achenes) and the North American S, calycina 

 (here included as a subspecies of S. montevidensis). 



Bentham and Hooker (1883, 3(3): 1006) reduced Lophiocarpus to a section of 

 Sagittaria and proposed Eusagittaria as a second section. 



Durand (1888, X, 504), without passing on its merits as a genus, discovered 

 that Lophiocarpus was a later homonymn of a chenopodiaceous genus, Lophio- 

 carpus Turcz., and proposed the name Michelia as a substitute (he thought that 

 Micheli was the original author of Lophiocarpus) , Durand then learned that 

 Michelia already had been used by Linnaeus for a genus of Magnoliaceae, and so 

 later, in the same publication, he decided on the name Lophotocarpus, 



Baillon (1894, 12: 70-81) not only did not recognize Lophotocarpus but reduced 

 Echinodorus to a section of Sagittaria. Kuntze (1898, 3(2): 325-329), following 

 Baillon's example, lumped the South American species of Echinodorus under 

 Sagittaria, and also reduced all the previously recognized South American species 

 of Sagittaria to varieties of S, pugioniformis, 



J. G. Smith (1894) attempted the first thorough revision of the North American 

 species of Sagittaria, He recognized 22 species, besides two species of Lophoto~ 

 carpus B Later (1899) he revised the Lophotocarpi north of Mexico, proposing seven 

 species, most of which have since been considered as edaphic variations of one 

 or possibly two taxa. 



Buchenau (1889) maintained the two genera in much the same manner as 

 Micheli. Later (1903) he narrowed the gap between the two considerably by leav- 

 ing in Lophotocarpus only the pantropical S, guyanensis, Buchenau *s later work 

 is the only thorough treatment of the Sagittarias of the world. He recognized 31 

 species and listed as doubtful five others, as well as maintaining Lophotocarpus 

 with two species as discussed above. 



Small (1909) last revised the North American members of the genus. He fol- 

 lowed Smiths concepts and recognized 35 species of Sagittaria and seven of 

 Lophotocarpus, 



The present writer, for reasons discussed above, cannot find any basis for 

 maintaining Lophotocarpus on a generic level; however, he does separate Sagit- 

 taria into two subgenera, Lophotocarpus and Sagittaria, This is justified on the 



