330 GARCINIA AFFORDING GAMBOGE IN SIAM. 



1864. which Professor Oliver has favoured me, and which is as 

 follows : 



"The specimens of the gamboge Garcinia from Messrs. 

 D'Almeida have afforded me excellent material for the examina- 

 tion of the anthers of this species, which exhibit an unexpected 

 and curious structure, which structure, however, is no doubt 

 common to all the species of the section Hebradendron. 

 Structure of " Professor Graham, in his paper upon Hebradendron, in the 

 auther. Companion to the Botanical Magazine (ii. 193), quotes an ex- 

 tract from a letter which he had received from the late Robert 

 Brown, in which Mr. Brown pointed out to him ' that approaches 

 to this structure (referring to the eircumscissile anthers), and 

 which serve to explain its analogy with the ordinary structure 

 of the family, exist in Garcinia! Messrs. Planchon and Triana, 

 in their excellent memoir on Guttiferce, in the description of 

 their sixth section of Garcinia ( Hebradendron) , refer thus to 

 the peculiar structure of the anthers : ' anthene peltatae rima 

 circulari dehjscentes, ideoque quasi circumscissce.' Mr. Brown's 

 observation as to the existence of intermediate forms, connect- 

 ing the structure of the Hebradendron-anthei with that of other 

 Garcinice, is a true one ; but in the genus Garcinia there occur 

 two distinct forms or types of anther, and he does not indicate 

 of which form he regarded the ffebradendron-Q.uih.QT as a modi- 

 fication. In some Garcinice, as in G. paniculata, the anthers are 

 truly peltate, the comparatively slender apex of the filament 

 being attached near the middle of the back of the anther 

 (Fig. 1). In these species the anther-cells are right and left on 

 the upper surface, and the dehiscence is longitudinal, as usual. 

 In some other Garcinice the filament is continued directly into 

 the base of the anther, passing from the base to the apex, the 

 anther being technically adnate. A priori the explanation of 

 the Hebradendron-an.t'hQT would be simply this : that the con- 

 nective of a peltate anther had become dilated transversely to 

 such a degree that the lines of dehiscence were marginal, and 

 the dehiscence ^^-circumscissile, as Messrs. Planchon and 

 Triana term it. I believe, however, that the anthers of the 

 gamboge Garcinia are not peltate, that they are truly circum- 

 scissile, and that they are a singular adaptation of the adnate 

 type of anther. It would appear as though in this species thick, 

 nearly sessile, and very densely packed anthers, of the adnate 

 type, have their lateral normal polliniferous lobes wholly, or 

 almost wholly, obliterated, and the pollen, lodged in 'cells of 

 irregular form and number towards and around the apex of the 



