Mr. J. Miers on the Affinities of the Olacacefe. 160 



they are evidently analogous to the sterile stamens of Agonandra, 

 a new genus of Olacaceee: the depressed ovarium, partly im- 

 mersed in the disk, is unilocular, with two to five o^niles 

 suspended from a cionosperm, or free central placenta. lodina 

 from its habit, with its spinous leaves more resembling those of 

 the Holly, might well be supposed to belong to AqiiifoliacecB, but 

 the sestivation of its coroUa, and the peculiar structure of its 

 ovarium, refer it, without doubt, to Olacacece. The genus lodina, 

 at first sight, ofiers a close resemblance to Cervantesia, which 

 has in like manner five large petaloid scales, alternating with as 

 many fertile stamens, and all originating in one common whorl, 

 from the margin of a cupuliform disk ; but in this genus the 

 disk is not free, as in lodina, but is entirely adnate with the tube 

 of the floral envelope, so that when the fruit ripens, the drupe 

 exhibits on its sides the persistent lobes of the corolla, and the 

 petaloid stamens ; but as the principal floral envelope must be 

 regarded as a perigonium, having no calyx at its base, and as 

 the disk is adnate with this perigonium, this genus must be 

 referred to Santalacea, while lodina and Agonandra must belong 

 to OlacacetE. There is one very unusual point of structiu-e in 

 Cervantesia, which appears to me without example ; the floral 

 envelope, deeply cleft above into five equal segments, is adnate 

 to the disk, a httle below the level of its free margin, but at this 

 point it descends again below the same line of attachment, in 

 the form of five other reverse segments, equal in size and con- 

 tinuous with the upper ones, and quite free from the disk and 

 pedicel, which they enclose, so that it appears to consist of five 

 elhptical segments, pointed and free, both above and below, and 

 confluent only with each other and vriXh the margin of the 

 disk by a naiTOw transverse zone running across their middle : 

 these inferior free processes must be spurhke extensions of the 

 perigonium. 



We have still another striking instance of the consimilitude in 

 the external characters of the Olacacece and AquifoUacece, which 

 has led to a confusion of reference, in an opposite direction : this 

 occurs in the genus Bursinopetalum of Wight, who assigned it 

 to the former family, but which appears to me clearly belonging 

 to the latter, as it agi'ees with it in the imbricate aestivation of 

 its corolla ; the petals, though distinct, and somewhat valvate at 

 base, are decidedly imbricated for at least two-thirds of their 

 length, two alternate petals being exterior to the others, and 

 their margins overlapping to a considerable extent; they have 

 the same prominent internal keel, and the apex is deeply inflected 

 by long processes, which are torsively complicated together, 

 as in Villaresia ; the ovarium (probably from a similar cause) 

 is unilocular, with an ovule (oi two?) suspended on one side 



