published in the Calcutta Journal of Natural 

 Science. 



Annual, erect, 1 to 2 feet high. Stems succu- 

 lent, very watery, a row of villi on one side, par- 

 ticularly near the ends of the branches. Leaves 

 alternate, petioled, membranaceous, nearly entire, 

 semicordate or oblique at the base, acuminate; 

 above sprinkled with adpressed hairs, glabrous be- 

 neath. Racemes opposite the leaves, floriferous 

 towards the apex, each pedicel with a small sub- 

 ulate bract at its base. Calyx o-cleft: tube 5- 

 angled, the upper one with a broad wing-like 

 crest near the base. Corolla monopetalous, un- 

 equally bilabiate; tube white; upper lip much 

 smaller, dentately truncate: lower one broadly el- 

 liptical, obtuse, entire, with two cavities at the 

 base; colour deep blue, with a yellow blotch at 

 the base; about 15 lines long. Stamens 4, an- 

 thers all perfect, 2-celled, cohering. Ovary free, 

 1-celled, with 2 parietal placent®, each dilating 

 into a broad ovuliferous lamina: ovules very nu- 

 merous: stigma cup-shaped. Capsule enclosed with- 

 in the persistent calyx; 1-celled, 2-valved; dehis- 

 cence locuIicidaL Seeds minute, pendulous, testa 

 reticulate- 



1354 Epithema Cetlanica (Gardner), pilosely 

 hispid all over: inferior leaves opposite, or solitary 

 by abortion, petioled, broad ovate cordate, doubly 

 serrato-dentate, the upper ones opposite, sessile: 

 peduncles terminal 1-3, elongated, spicate at the 

 3*pex: spikes dense, secund, circinate, bractiate at 

 the base : bracts cordate, cuculate, obtuse; dentate. 



Neilgherries, on moist shaded rocks^ on the banks 

 of the stream at Burlear, abundant. Courtallum on 

 similarly situated rocks* Ceylon, "in clefts of moist 

 rocks in forests." 



This is a curious and 

 peculiarities of which I 



genus, the 

 had not studied when 



interesting 



Western 



the drawing was made which is therefore defec- 

 tive in one or two minute points. 



1355. IsAiYTHERA PERMOLLis. (Nces. CyHaudra 

 lanuginosa, Br. in Wall. List. Is. foribunda^ Gard- 

 ner, CaL Journal.) 



Courtallum, in shady moist jungles, 

 slopes of Shevagherry mountains in similar situa- 

 tions, in great profusion, flowering in August 



SufFruticose, erect, simple, glabrous below, woolly 

 tomentose above: leaves congested towards the 

 apex, short petioled, alternate, obovato-cuniforin, 

 acute qr shortly acuminate, minutly serrated, pmni- 

 veined!?^ubescent above, tomentose beneath; to- 

 mentum in the dried specimen rusty or tawny 

 coloured. Peduncles axillary, about the length at 

 the petiols, slender drooping: cymes many-flower- 

 ed: calyx 5-parted, lobes lanceolate acute, woolly, 

 nearly as long as the corolla. Corolla 5-lobed, 

 lobes suborbicular. Stamens 4, didynamous, insert- 

 ed on the bottom of the tube, filaments recurved. 

 Anthers subglobose, 1-celled, dehiscing transverse- 

 ly. Ovary free attenuated upwards, 1-celled, or im- 

 perfectly 2-celled: placentae 2, parietal, expanding 

 within into a flat ovuliferous lamina, revolute on 

 the margins : style stort, stigma simple, blunt. Cap- 

 sule 2-valved, 1-celled, dehiscence loculicidal. Seed 

 minute, elliptical, compressed or angular, from mu- 



So far as can be learned from description, un- 

 aided by figures or specimens, the Indian and 

 Ceylon specimens seem not to be distinct. On 

 which account I have quoted, but with a doubt, 

 Mr. Gardiier's L Jlonbunda as a synonym to Nees' 



1. permoUts. 



tual pressure, testa brown reticulate. 



CONVOLVULACE^. 



Under No. 839 I offered some remarks on the 

 genera Rivea and Argyreia. At that time I had 

 not been enabled to enter critically on the ex- 

 amination of the order and could suggest no remedy 

 for what I then considered erroneoiis, viz. the loose 

 way in which M. Choisy had referred species to 

 his newly constructed genera, Rivea and Argyreia. 

 Since then I have been enabled to examine the 

 subject with more care and have published the 

 result in a brief article in the Calcutta Journal 

 of Science for July, 1847, which I shall reproduce 

 here for those who may not have seen the original 



RivEA, Argtreia and Lkttsomia. 



Mr. Choisy, in his Memoir on Indian Convolvula- 

 cece, in taking up Loureiro's genus, Argyreia^ has 

 changed its character so essentially, that every 

 one of Loureiro's genuine species must now be 

 excluded. I say genuine^ because if Choisy is 

 correct in referring Argyreia /estiva^ Wall, to A. 

 acuta, Lour., which I doubt, then that is not a 

 true species of his genus, which, as defined by 

 himself, has a 4-celled ovary, while A. /estiva has 



it 2-celled. 



Loureiro's character of the firuit of Argyreia is 

 "bacca subrotunda exsucca i4ocularis f^ Choisy's, 

 "ovarium 2-loculare 4-spermum." If the berries 

 in Loureiro's plants have four cells, it is obvious 

 the ovary must have had at least an equal num- 

 ber: hence, in assigning a 2-celled ovary to Ar- 

 gyreia, Choisy has altogether suppressed the original 

 genus, and set up a most distinct one in its place, 

 while at the same time he has added to the confu- 

 sion by placing in his new genus, numerous species 

 with 4-celled ovaries and fruit. In fact, nearly 

 the whole genus, as it now stands in DeCandoUe's 

 Prodroraus, will, I apprehend, be found not to come 

 within his generic character. 



It is a curious fact, that Roxburgh fell into a 

 similar error in regard to his genus Leltsomia^ 

 which, according to his definition, has 2-celled 

 ovaries, while nearly all his species have them 

 4-celled. When both he and Loureiro wrote, the 

 same importance was not attached to that point 

 of structure that Mr. Choisy has shown it de- 

 served, and their enor is easily traced to too 

 rapid generalization. Loureiro must have exarn- 

 ined a species with a 4-celled fruit, and took it 

 for granted all the others had the same structure. 

 Roxburgh on the other hand, when drawing up 

 the character of his genus Lettsomia, seems to 

 have had a species before him with a 2-celled 

 ovary, and assumed that all the other species with 

 baccate fruit had likewise only two cells- He 

 consequently associated under that character many 

 species with 4-celled ovaries, and only two or 

 three having them 2-celled. M. Choisy, m the 

 course of his examinations, met with some species 

 having four cells, others having two cells: of the 

 former he has constituted the genus Rivea, of the 



latter his genus Argyreia 



But falling into the 



( 11 ) 



