NOTE ON VAV.5:A AND RHTTIDANDRA. 335 



The only observed discrepancies are the valvular aestivation of the corolla in Rhyti- 

 dandra, and its bifid style; — neither of which characters is likely in the present case 

 to indicate more than a generic distinction. For the flattened divisions of the style, 

 themselves more or less bifid at their summit, would by a further union produce nearly 

 such a four-lobed stigma as that of 31arlea and of AJaugium. And if the narrow 

 petals are really convolute in aestivation in the former as well as the latter genus, their 

 margins can but slightly overlap,* while the strictly valvate mode would be no un- 

 expected character in a new genus of a small group, which — following Mr. BroAvn's 

 suggestion made thirty-six years ago — it is now conceded must be merged in the 

 Cornacc(i!.\ 



In its unilocular ovary, Hhytidandra accords with Ahiiifjium, as also with an occa- 

 sional state of Marlea ; % while the stamens correspond with those of Marlca in 

 number and position, and have even shorter filaments. The anthers are distinct, not 

 connate into a tube, as those of Marlca are said to be by Lindley and by Endlicher 

 (but not by De Candolle) ; nor are the stamens united by pairs, as those of Marlea are 

 characterized and represented by Lindley; unless, indeed, wliat I had taken for a 

 quadrilocellate anther should consist, as it possibly may, of a pair of closely coalescent 

 anthers. Their dehiscence, if known, would determine this point. In respect to it I 

 can only say that, if the anthers of Rhytidandra really open longitudinally at all, they 

 must do so by the lateral grooves, one on each side, which correspond with an inter- 

 nal partition, longitudinally dividing each half of the organ into two locelli ; and in 

 that case the whole must constitute a single stamen, as I had supposed it to do ; and 

 I suspect this is the case in 3Iarlea also. 



However this prove to be, Rhytidandra is sufficiently distinguished from Marlea by its 

 mouiliform and chambered anther-cells, its one-celled ovary, and its bifid style with 

 elongated and slender but flattened lobes. 



This peculiarity of the style is of considerable interest ; for the lobes may be justly 

 compared with the style of Nyssa ; the affinity of which to the Alangieee was happily 

 suggested (though with doubt) by Brongniart,§ while its relationship to the Cornea; was 

 practically recognized by Blume, who referred his genus Mastixia first (and justly) 



* Wight and Arnott's authority (Prodr. Fl. Ind. Or. 1. p. 325) should settle the point, at least for Alangium. 

 But the figure of Marlca hegonimfolia in Bot. Reg. 21, t. 61, appears as if the petals were valvate. 



t Bennett, Plantse Javanicce Eariores, p. 194. In collating Marlea with the Cornea, no difference in 

 jEstivation is here mentioned ; from which it may be inferred that the petals of the former genus are valvate. 



J Lindley, Bot. Reg. 1. c. Clarke, in Kew Jour. Bot. 2. p. 129. 



§ Enum. Pi. Hort. Mus. Par. p. \xx. note. 



