218 ELATINACE.E. 



stipules, the absence of pellucid dots in the leaves, and by the perfect sym- 

 metry of the distinct stamens ; and on the other with Crassulaceee, through 

 I)i;imorpha, as has already been suggested by the acute Fenzl. 



The two Linnaean genera rested on the number of stamens and styles; 

 Elatine being octandrous and tetragynous, and Bergia, decandrous and pen- 

 tagynous. But it afterwards appeared that the parts of the flower were 

 occasionally ternary, and the stamens sometimes of the same number as the 

 petals only, namely 3, 4, or 5 : and in 1817, Nultall introduced a third ge- 

 nus, Crypta, with a dimerous and diandrous, or rarely trimerous and trian- 

 drous flower, and few seeds. This genus, as well as Bergia, Fischer and 

 Meyer soon afterwards referred to Elatine. In 1829, Cambessedes found- 

 ed on a Brazilian plant a fourth genus, Merimea; which he distinguished 

 from Bergia by its smooth (instead of ribbed) seeds, and strictly septicidal 

 dehiscence, the dissepiments coming aw-ay attached to the valves ; while in 

 Bergia he supposed them to remain adherent to the axis as in Elatine; 

 — leaving the pentamerous instead of 3 - 4-merous flowers to distinguish 

 Bergia from Elatine. The next year, Mr. Arnott, or at least in 1834, 

 in conjunction with Dr. Wight, also referred all the plants of the family 

 (except apparently Lancretia of Delile) to Elatine ; at the same time errone- 

 ously characterizing the dehiscence as loculicidal. Endlicher retained the 

 genera Elatine, Bergia, and Merimea as left by Cambessedes, distinguishing 

 Bergia by the valves of the capsule being introflexed but separating from the 

 persistent dissepiments ; and appending Lancretia to Hypericineae. In 

 1810, Hooker (in Irones Planlarum) figured a Texan pentamerous species, 

 with just the habit of the Indian Bergias, under the name of Merimea (or 

 Bergia?) Texana ; which, not having ascertained its dehiscence, he referred 

 to Merimea chiefly because of its being a native of the New World. The 

 dehiscence of this species having been ascertained to be septifragal, as in 

 Elatine proper, it was referred to that genus in the Supplement to the 

 Flora of North America, Vol. I. More recently, Fenzl has successfully 

 reduced the whole to the two Linnaean genera ; Elatine comprising all those 

 with septifragal dehiscence, and Bergia (including Lancretia and Merimea) 

 those with septicidal dehiscence ; giving to the former, however, the origi- 

 nal Linnajan species of Bergia. But Seubert has since observed that this 

 species is septicidal, and has restored it to Bergia ; wrongly taking with it, 

 however, our E. Texana, the dehiscence of which, he says, is not clearly de- 

 scribed, although, in the work referred to, it is explicitly said to be septifragal. 



As this last plant is now the only pentamerous Elatine, and presents just 

 the aspect of Bergia, I have allowed it to form a distinct section. It should 

 be remarked that the specimens gathered by Lindheimer are decandrous ! 



Finally, as it appears that our Crypta differs from its allies by the eva- 

 nescent partitions of the delicate pod (if we mistake not), and the basilar 

 placenta, as well as by the few seeds and prevailingly dimerous flowers, I 

 have separated it from the gerontogeous species, as a subgenus merely ; 

 proposing, however, its entire reestablishmcnt, in ease of its corroboration 

 through the New Zealand E. gratioloitks of Cunningham, — a trimerous 

 species of similar habit, which is described as having a unilocular capsule. 



