MENI8PEEMACEJE. 11 



with but one well developed cotyledon, 1 but this is closely folded on 

 itself, as in Hyperbaric!., and Chondodendron. Rameya consists of 

 Lianas from Madagascar, with entire leaves, three-ribbed at the base, 

 and with its flowers collected in small numbers on the wood of the 

 branches. 



III. CHASMANTHEBA SEKIES. 



Chasmantherd (figs. 16, 17) has nearly the male flowers of Cocculus ; 

 but its stamens are monadelphous over a variable extent ; 3 in the 

 female flower they are represented by six sterile rods. The three 

 carpels are formed as in Cocculus, with a reflexed style of variable 

 form. The fruit and seed present peculiarities which alone justify 

 the formation of this very artificial series. The three drupes are 

 nearly ovoid, flattened on the face towards the centre of the gynae- 

 ceum. Here the stone is depressed by a deep hemispherical or 

 vertically elongated pit. The walls are thus pushed into the 

 true cell, which is moulded on their convexity to form a meniscus, 

 concave internally, convex outside. The seed is similar in form, 

 contained in the ovary cell ; within its very thin coat is contained 

 a fleshy ruminated albumen of no great thickness, which may 

 divide into two la} r ers, one thrust into the other. Between them 

 is the incurved embryo. This is very peculiar in form ; its radicle 

 is superior and cylindrical, and the two very thin cotyledons are 

 parallel to the two faces of the seed, and are divaricated, or widely 

 separated at the base to form an angle with its apex superior ; 

 here it is that the two layers of albumen come in contact. 4 



The first Chasmanthera described was C. dependens Hochst., from 

 Tropical Africa, possessing broad digitiveined leaves, cordate at the 

 base. Its flowers form axillary or supra-axillary racemes, grouped 

 in cymes in the axil of each bract in the male plants, but usually 

 solitary in the females. 



1 This is at least the case in one species, which 3 The anthers are formed as in Cocculus, and 

 we can only doubtfully refer to this genus under after dehiscence show four low chambers, broadly 

 the name of R. 1 loucoubensis. We have also sug- open above, and separated from one another by 

 gested, not unhesitatingly, that Tinospora ? four low septa intersecting in a cross. 



funifera Oliv. (Fl. Trop. Afr., i. 44) might 

 belong to this genus. 4 All these characters of the seed and embryo 



2 Hochst., in Flora (1844), 21. — EkdIi., are not figured here because they are the same 

 Gen., n. 46% 1 .— B. H„ Gen., 34, 960, n. 3.— in Anamirta Cocculus (figs. 20, 21). 



H. Bn., in Adansonia, is. 305. 



