218 Mr. J. Miers on the Affinities of the Icacinacese. 



XX. — Observations on the Affinities of the Icacinacese. 

 By John Miers, Esq., F.R.S., F.L.S. 



[Continued from p. 132.] 



One of the foremost of the conclusions last alluded to, is the 

 necessity of removing the tribe of the Icacinea of Mr. Bentham 

 from the family of the Olacacea, for which many cogent reasons 

 have already been offered (hiij, op. viii. p. 173), and this group I 

 l)ropose to establish at once as suggested, either as a suborder, or 

 more deservedly as a distinct family under the name above 

 stated. In his excellent memoir on the Olacinece before quoted, 

 Mr. Bentham enumerated only eleven genera of that order, two 

 of which were then first described by him. Of these five belonged 

 to his tribe of the Icacineaj and one I have shown has no rela- 

 tion there; hence only five genera out of that list were then 

 known as really appertaining to the former family. I have now 

 enumerated thirteen genera of the Olacacea {loc. cit. p. 172), 

 and an equal number of the Icacinaceae {idem, p. 174) *. This last- 

 mentioned group exhibits characters so widely difi^erent from 

 those of the former, that it becomes necessary to place the two 

 at a considerable distance in the system. These characters con- 

 sist in their frequently polygamous, almost dioecious flowers ; the 

 alternate position of their stamens with respect to the petals; 

 the evidently normal polycarpellary structure of the ovarium (for 

 where developed with more than one cell, the dissepiments are 

 always found complete at their summit) ; the pendulous position 

 of their ovules, generally two in each cell, near the summit, 

 where they are suspended, one always more or less above the 

 other, from a peculiar cupshaped podosperm ; their constantly 

 indutive seeds ; the existence of a distinct testa, of inner integu- 

 ments, of a chalaza and of a raphe : all these are important dif- 

 ferences, quite at variance with the very peculiar points of struc- 

 ture that characterize the OlacacecB. 



With such completely irreconcileable characters, there is only 

 one proper course to pursue, and that is to separate them ; but 

 it then remains to be considered, where in such case they should 

 be placed in the system, for it is evident that they cannot even 

 remain in juxtaposition. In the face of the difficulty of com- 

 bating long-established opinions, it becomes essential, at the 

 risk of the charge of prolixity, to recapitulate here a few of the 

 arguments before assigned, in order to guide us to a right deci- 



* One of these, Ptychopetalum, was placed here, because its ovules were 

 stated to be attached to a parietal placenta ; but having since had an oppor- 

 tunity of examining it, I find it to possess the general characters of the 

 Olacace<B. 



