222 Mr. J. Miers on the Affinities of the Icacinaceae. 



nearly equal to them in length; filaments erect, often im- 

 plicated, fleshy, compressed, subulate, sometimes invested with 

 glandular hairs ; anthers introrse, 2-lobed, 2- or 4- celled, cells 

 dehiscing lengthways by a cleft along one of their margins, 

 rarely otherwise. Ovarium entirely free, supported upon a cup- 

 shaped disk (which is either quite free or partially adnate with it), 

 fleshy, oblong-conical, often surmounted with a conspicuous 

 fleshy epigynous gland, which is sometimes lateral, generally by 

 abortion 1-locular, sometimes 3-locular, and then the cells are 

 excentrically disposed, showing the normal number of united 

 carpels to be 5 : ovules geminate, somewhat collaterally sus- 

 pended, one a little higher than the other, from near the sum- 

 mit of the cell, by a short fleshy podosperm, which is generally 

 expanded in the form of an inverted cup, anatropal, and some- 

 times resupinate. Style erect, or incurved, somewhat excentri- 

 cal, as long as the stamens, sometimes wanting. Stigma generally 

 clavate or obsoletely lobed. Drupe baccate, containing a single 

 1 -celled, in dehiscent putamen ; seed single, filling the cavity of 

 the cell, and apparently resupinate; testa thin and somewhat 

 membranaceous, raphe arising from the nearly basal chalaza and 

 extending to the summit along the dorsal face. Embryo in the 

 axis and summit of fleshy and copious albumen, sometimes much 

 shorter, and almost terete, with small oval cotyledons scarcely 

 longer and broader than the superior radicle ; but often nearly 

 the length of the albumen, with large, ovate, foliaceous cotyledons, 

 much longer and broader than the short terete superior radicle. 



The affinity of the Icacinacece is evidently nearest to the Aqui- 

 foliacece and the Celastracea, differing from both in the aestiva- 

 tion of the corolla and the tenuity or frequent obliteration of the 

 hypogynous disk. From the latter family they are distinguish- 

 able by their suspended ovules and their generally unilocular 

 ovarium, and the absence of an arillus about the seed. The seminal 

 characters have been derived from examinations of the seed of 

 several species of Mappia, and confirmed by the few details fur- 

 nished by the 'Icones ' of Dr. Wight in tab. 1153 of ApodyteSj 

 and tab. 934 of Stemonurus {Gomphandra) . I have little hesita- 

 tion in concluding that the genus Pennantia, which has been 

 placed by botanists in many different positions, belongs to this 

 group, and from this source I am enabled to add many additional 

 features which will probably be found to exist in the structure 

 of other genera of this family. 



I propose to divide the order into three tribes. 



1. Icacinecs. In all the genera composing this tribe, the 

 ovarium is constantly unilocular in consequence of the complete 

 abortion of the other cells, hence it is always somewhat gibbous 

 and the style is distinct and in some degree lateral : the anthers 



