296 Analysts of Boohs, [Oct. 



appears that neither the number of the stamina, nor of the 

 nerves of the leaves, afford subdivisions consonant to the natural 

 series. I met with this plant on the very summit of the Sugar- 

 loaf, along with Rhododendra and Vaccinia. ^^ 



A plate accompanies this communication, exhibiting the parts 

 of fructification and the fruit of Melastoma Maiabatnrica, 

 M, exi^ua, and M, aipestris, 



II. Un Cj/rtandracecc, a new Natural Order of Plants. By 

 William Jack, MD. Communicated by Aylmer Bourke Lam- 

 bert, Esq. FRS. VPLS. 



Dr. Jack's introductory remarks in this paper are as follows : 



" In examining some of the numerous Jumatran species of 

 Ci/rtandra, I was lately led to observe the great inaccuracy of 

 Forster's description and figure of the fruit, which has been the 

 cause of deception in regard to its natural affinities. His error 

 consists in representing the septum as complete, with adnate 

 placentjB similar to what obtains in some genera belonging to 

 ScropliularincR ; whereas, in reality, it is bipartite through the 

 axis of the fruit, and the placentae are no other than the revolute 

 lobes of the septa. This peculiar structure is more distinct in 

 the nearly related genus of^ Didi/mocarpus (Mai. Misc. vol. i.), 

 which has capsular fruit, and where the lobes of the contrary 

 dissepiment so completely bipart the cells as to give it the 

 appearance of being quajdrilocular. It is obvious that this 

 character is totally inconsistent with that of Scrophularincc, and 

 it does not accord exactly with any of the Jussisean orders. 

 Didi/mocarpus is related to Bignoniacecz through Incarvillea, but 

 it is not admissible into that family as defined by Mr. Brown in 

 his Prod. Fl. Nov. Holl. I am therefore inclined to think that 

 Cjt/rtandra, Didumocarpus, and another genus, which I shall 

 here present under the name of Loxonia, which agree remarka- 

 bly in general habit as well as in carpological structure, may 

 properly form a small and distinct family near to Bignoniacea. 

 The two first genera are numerous in the Malay islands ; and I 

 may remark that, as far as my present observations extend, the 

 Cj/rlandrcE appear to prevail principally to the south of the 

 equator, and the Didt/mocarpi on the north, where it has even 

 been found, according to the observations of Dr. Wallich, to 

 extend to the alpine regions of Nepal. I shall proceed to give 

 the characters by which this family and its genera are distin- 

 guished, and shall add descriptions of all the species that I have 

 as yet had an opportunity of examining." 



CvRTANDRACEiE. 



Calyx monophyllus, divisus. Corolla monopetala, hypo- 

 gyna, sajpius irregularis, 5 loba. Stamina. Filamenta 4, quo 

 plerumque, nunc quatuor antherifera. Anfhercc biloculares, per 

 paria connexae. Ovarium disco glanduloso cinctum, bilocu- 



are vel pseudo 4 loculare, polysporum. Stylus simplex. Stigma 



