1897.] D. Prain — Some additional Leguminosa?. 385 



Burma with all tho characters of the South Indian plant originally differentinted 

 by Dr. Wallich, makes it more satisfactory to give the form varietal rank because 

 its inclusion in var. styracifolius somewhat mars the symmetry of an otherwise very 

 well characterised form ; the establishment of this variety obviously involves the 

 deletion of the synonym A. pilifer under Mr. Baker's var. styracifolius. It should 

 be added moreover that Dr. King's Collectors note the corollas as yellow in this 

 variety, all the others are noted as having them purple. 



* * * Desmodiastrum. Calyx much longer than the first joint of 

 the pod, its teeth not imbricated in the fruiting stage. Pods as in 

 Desmodium. 



9. Alysicarpus belgaumensis WigJit. 



10. Alysicarpus racemosus Benth. 



This is - reduced, in the F. B. I., to A. belgaumensis. It is, however, impossible - 

 to confound the two plants, their pods being remarkably different and no inter- 

 mediates occurring; and there is hardly a doubt that Mr. Bentham was justified inv 

 giving the present one specific rank. There is, however, very considerable difficulty 

 in separating this species from the two plants named by Dalzell Alysicarpus parvi" 

 fiorus and A. rotundifolius, both of which the jP. B. I. has transferred to Desmodium, 

 That these two are congeneric with Alysicarpus racemosus admits of no doubt ; the- 

 question whether, with Mr. Baker, we are to treat A. parvifiorus and A. rotundifolius 

 as Desmodia or, with Mr. Dalzell, to treat them as Alysicarpi is one that may be 

 answered with much reason either way. But wherever these two are placed, A. 

 belgaumensis and A. racemosus must accompany them. Compromises in taxonomy 

 are necessary, indeed the systematic arrangement of species is essentially the art of 

 happy compromise. But an arrangement which places one half of a natural group 

 of forms in one genus, the other half in a second, strains unduly the privileges 

 that the art of compromise allows. 



11. Alysicarpus parviplorus Dalz. in Hook. Kew Journ. iii. 211. 

 Desmodium parviflorum Bak. in Flor. Brit. Ind. ii. 172. 



Only distinguished from A. racemosus by its further-exserted pods and its 

 rather longer pedicels ; its leaves are occasionally 3-foliolate as in A. belgaumensis , 

 the leaflets being oblong or lanceolate as in that species. 



12. Alysicarpus rotdndifolius Dalz. Desmodium rotundifolium 

 Bah. in Flor. Brit. Ind. ii. 172. 



Fruits exactly as in A. parvifiorus, from which it is distinguished by its rather 

 larger, slightly exserted corollas and its obovate-oblong to orbicular leaves which 

 are hardly distinguishable from those of A. racemosus. The writer's reason for 

 proposing a new subgenus for this natural group is that the calyx-teeth in none of 

 them become imbricated : hence all of them violate the limits of the section Macro- 

 calycinse as defined in the F. B. I., which includes two of them therein. His reason 

 moreover for retaining the group in Alysicarpus rather than for transferring all four 

 species to Desmodium, of which all have the pods, is that it seems better to locate 

 the group in a small manageable genus like Alysicarpus, than to transfer them to 

 one, like Desmodium, already of unwieldy bulk. As a matter of fact the group 

 stands intermediate between these two genera, and indicates that probably they 

 are not naturally separable. 



