﻿I 



( 



i 



i 



i 



J 



\ 



■ 



i 



i 



MR. G. BENTHAM ON THE MIMOSE^. 337 



EumimosecB, with definite stamens; Acaciece, or the genus Acacia, with indefinite 

 stamens, free or very shortly and irregularly united at the hase in the centre of the 

 flower ; and Ingece, with indefinite stamens united, at least at the base, in a tube sur- 

 rounding the pistil. Grisebach, it is true (El. Brit. W. Ind. 220), rejects the cbaracter 

 because "decandrous species still exist among Bentham's Acacice and CalliandrcB, e. g. 

 A, tamarmdifolia, W., and 0. moUissima, Benth." But, as to one of them, he has mis- 

 matched the flowers of Mimosa with the fruiting specimens of the true Acacia tamarindi- 

 folia, which has always above 50 stamens ; and Calliandra moUissima^ although figured 

 by Kunth as having 10 stamens, appeared to me to have them few, indeed, but variable 

 in number ; and they are accordingly described by Kunth as being about (not exactly) 

 10 ; they are also monadelphous at the base, not free as in the truly dccand 



A secondary character, however, by which I had separated Adcnanthercaj from Eumi- 

 mose^, the presence of a glandular appendage to the anthers, has failed in a few instances. 

 The genera Farlcia, Piptadenia, and Frosopis, in which the anthers arc usually pro- 

 vided with that appendage, have each of them one or two species in which it is wholly 

 wanting, even in the unexpanded flower ; and Fesmanthus, where it is usually deficient, 

 has a Madagascar species with a minute gland sometimes, if not always, present. Xylia 

 also, which I had inadvertently placed among the eglandular genera, has, on the con- 



trary, very prominent glands to the anthers. 



I have now been able to introduce as a new generic character, apparently absolute, 

 and especially useful for distributing the Eumimose^ into two subordinate groups, but 

 which I had formerly neglected, the presence or absence of albumen in the seeds. I was, 

 indeed, aware that Schleiden and Yogel (Nov. Act. Nat. Cur. xix. part ii. 68) had a^cer- 

 tained the existence of albumen in some seeds only of Mimoseae ; but as they also sta ed 

 that they found no constancy in this respect in the same genus, as I had not then the 

 opportu4y of verifying the circumstance in a sufficient number of cases and as I was 

 ^^ ^ _ this respect in some natural genera, such as Folygala, I did 



notleeljuliifielin making any use of it in my generic characters ' ^ ' 



examination of a large number of seeds, that the albumen is^ ^^^--'1^1^^' 



of the variability 



I now find, from the 



in each j_ 



.nu«, .« modified in my synopsis, although variable in sorne Willde 

 'Schldden cites, as examples of inconstancy, three specics of ^^^^^^ 



genera. Schleiden cites, as exampit.=> u. ^^^^^^^ „,u,„^ * :^ riresont whilst in th 

 ?7 . r • „^;i 4 -Rprtpriana, in which albumen is present, wnusi in in 



thocarpa, A, farnesiana , and A. Bertenana 



But of those three the first is, in it 



n^ajority of species the seeds are exalbu— ^^^ ^ J^, ^ / ^ /; ,„^ 

 stamens and otlier characters, a Mmosa, and not an Acac / 



Acacia \ but I 

 examined half 

 localities, and found 

 than in most 



ovoid shape, not at «" A^"^"^^ V^ Jl'';;' intermediate layer of it is somewhat cartila- 

 GummifertB, and the inner or, ra , albumen of Mimoaa ; but it is continuous, 



ginous, haying almost the appe^^""^ " , ^^^ tapering towards the margins of the 



and of equal thickness t''-"^^;;'^''^;;",^; i,,er lining as from the outer coating of 

 cotyledons, and is inseparable as wen uum ^ ^ ^ 



1 



i 





