Desmodilm. LEGUMINOS.'E. 3(;i5 



The species of this extremely difTicult genus present such diversities, and ofkon 

 appear to approach each other by such gradations, tliat it is by no moans easy to 

 define them properly. Tlio most constant characters may coinnionly bo dLrivud 

 from the form of the joints of the legume ; but these are frequently slight and not 

 easy to be expressed in words. Having obtained additional materials and informa- 

 tion since our account of the earlier portion of this genus was published (Oct. 

 1833 ; to spec. no. 8. p. 3f)0), we append a few notes on the subject. 



The genus Ototropis, N. ah E. (delect, sem. hort. Vratisl. 1838, p. 3.) would 

 include many of our species, so far as the caly.x is concerned, and several agree 

 with it in the form of the legumes ; but in this respect evory gr.ul.ition may be ob. 

 served, and other species which agree with Ototropis in fruit h.ivc a verj' diff-rcnt 

 calyx ; so that we see not how it can bo adopted even as a section. It sh'uuld have 

 been stated, perhaps, in the generic character of Dcsmodiuin, that the upper lip of 

 the calyx is often entire, and sometimes the calyx is nearly equally 4- or S-toothed. 

 The joints of the pod are reticulated and more or less hi-spid with uncinate hairs in 

 all our species. 



D. Canadense. DC! (fledysarum Canadonse, Linn, liort. Cliff.! «J- spec..') 

 To the character should be added : Petioles very much shorter than the leaflets ; 

 stipules rather persistent ; legumes nearly sessile (much smaller than in D. canes, 

 cens), of about 5 roundish rather oblique joint.-*. 



0. longijolium : leaflets lanceolate-ovate, membranaceous ; stipules lanceolate ; 

 racemes mora lax. — D. longifoliuin, Niitt..' ?nss. — Arkansa.s, Nuttall! Dr. 

 Vitcher ! — We have specimens intermediate between this and the ordinary form of 

 the species, collected in Illinois by Mr. Buckley. The petiolos in D. Canadense 

 are ordinarily scarcely longer than the potiolulo of the terminal leaflets; but this 

 variety somewhat approaches D. canescens. 



D. caneicens, DC. (ex syn. & descr.) Hedysarum canescens, Linn. hort. 

 Cliff. ; (1737) ^ ex syn. Gronov. .' not of hort. L'ps. ! (1748), which is a Jamaica 

 plant; excluding also the syn. H. triphyllum, &.c. Sloane .' and Onobrychis 

 Americana, floribus spicatis. &,c. Pluk. .' which both relate to another West Indian 

 species. H. viridifloruin, Willd. .' Purs/i .' ^c. D. viridiflorum, DC. .' ^c. — The 

 H. canescens, Willd. .' is the Linnaean plant of the Ilorlus Upsalensis. Our plant 

 is accordingly, as we supposed, the original H. canetcens, according to the speci- 

 men in the herbarium of the Ilortus Cliffortianus, now belonging to the British Mu- 

 seum. The proper herbarium of Linutcus contains the plant of the Hortus Upsa- 

 lensis only. It is the former which LinniEus compares with H. Canadense (also 

 established in Hort. Cliff".) To distinguish it more perfectly from this as well as 

 the two species with which Linnceus confounded it, we add the following 

 characters : Petioles about the length of the broadly ovate leaflets, which 

 are reticulated beneath, and scabrous on both sides, but especially beneath, with 

 long and stitt'appressed hairs and a minute uncinate pubescence ; legume with a 

 short stipe. 



13. villosissimum : panicle and upper part of the stem very villous ; leaflets ob- 

 long-ovate ; upper lip of the calyx sometunes very deeply cletl. — D. Canadense, 

 var. Hook. ^ Am. compan. to hot. mag. 1. p. 22. — St. Louis, Missouri, Drum, 

 mond ! 



D. Dillenii, Darlingt. (D. Boottii, Torr. .' in Curtis, cat. Wilmingt.pl. Hedy- 

 sarum Maiilandicum, Willd. I) — The joints of the legume were accidentally 

 described as oblong instead of rhomboid. It varies greatly in tho size of the 

 leaves, which are sometimes obtuse at both ends, and sometimes acute. 



D. viridiflorum. Beck. — Stem scabrous towards the summit ; stipules rather per- 

 sistent. — In specimens from Florida, recently received, with mature fruit, we find 

 the legumes somewhat oven on the back ; the joints slightly convex on that side, 

 and much dilated and rounded below, so as to assume a rather semi-orbicular out. 

 line. The old leaves are less tomentose beneath. We liave another form from 

 Louisiana (Dr. Hale) with ovate-oblong leaves, and the joints of the young le. 

 gumes somewhat rhomboidal. Possibly two or more species are confounded under 

 this name. 



Hedysarum volubile, Z«inn. (founded on Dill. Elth. t. 143,/. 170.) is Galactia 

 pilosa or G. mollis. 



