BENTHAM ON MIMOSE.E. 329 



leaflets is vei-y uncertain. On comparing the Senegambian 

 specimens of Parhia Africana, Br. (fully described by 

 Guiilemin and Perrottet), with those of P. higlandulosa^ 

 W. and Arn., from the East Indian Peninsula, of P. Bni- 

 nonis, Grab, in Wall. Cat. n. 52S8, and P. Roxhurghii, 

 O. Don, from the Calcutta Garden, and with the description 

 of the East Indian specimens by Arnott, Roxburgh, &c., of 

 the Timor specimens^ by Decaisne, (which he refers to P. 

 Moxhiirg?m, and associates to it as a synonym Lifja Tirno- 

 riana, DC); and of the Philippine Islands tree described by 

 Blanco under the name of Mimosa peregrina ; they all agree 

 in essential particulars. The only discrepancies are in the 

 colour of the flower, which is said by Roxburgh to be yellow 

 in the Sylhet plant, and by Guiilemin to be purple in the 

 African one, and in the breadth of the pod, which is variously 

 given, from half an inch to an inch and a half. All agree in 

 its general form, length, number of seeds, &c. Unable 

 therefore to discover any differences that may not be refer- 

 rible to mere varieties, I have included them all under the 

 original specific name. 



2. P. platycephala, (sp. n.), pinnis 6 — 12 jugis, recepta- 

 culo florum depresso-globoso, late capitato — Brazil, Prov. 

 of Ceara, Gardner, n. 1582, Prov. Bahia, Blanchet, n. 2868. 



This is really a well-marked species, with much the foliage 

 of P. biglobosa; but the receptacle of the flowers is broader 

 than it is long, and the pod is not four times as long as 

 broad. 



Inga pendula, and I. fiUcina, Willd., Spec. 4. p. 1025, 

 "lay perhaps be species of Parkia. 



The genus Parkia is in many respects allied to Inga, and to 

 the genera separated from it, especially in the monadelphous 

 stamens, and the pollen masses, which, though numerous, 

 are biseriate, and perhaps definite in number, whilst the form 

 of the calyx and the estivation is very different from all 

 Mimosece. There is however no one genus of Ccesalpiniece 

 which can be said to be nearly allied to it. 



Vol. IV._No. 30. 2 T 



