192 



KATUIiAL mSTORY OF PLANTS. 



always sessile or nearly so, more rarely stipitate (fig. 229). They 

 vary i^reatlv in form, being ovoidal, obovoidal, cylindroidal or club- 

 shaped. The seeds, which contain a ruminated albumen and a 

 small embryo near the apex, are separated from one another by false 

 transverse dissepiments, to which answer circular external 

 contractions, usually but little marked and sometimes 

 quite wanting. The Uvarias are shrubs, often creeping 

 and climbing. Their alternate leaves are usually covered 

 with a more or less abundant down, as are the 3'oung 

 branches, peduncles, calyces, receptacles, and fruits.' 

 The flowers are axillary or terminal, often leaf-opposed 

 in the latter case, sometimes solitary, sometimes united 

 into few-flowered cymes. About two-score species are 

 admitted,- of which the number will probably have to 

 be reduced. 



In the flowers of U. sjjhenocarpd^ (Ceylon), the petals 

 are united for a certain height into a corolla, which 

 flills ofl" in a single piece. This character, which in 

 certain other groups is considered of capital importance, can have 

 none in the genus Uvaria, for in it we find every intermediate stage 

 between the gamopetalous U. spJie7iocarpa, the species where the 

 union of the petals is scarcely indicated, and those which are com- 

 pletely polypetalous. 



Even besides Asimina there are other Uvarias of American origin ; 

 viz., the Poj'celias* of Peru, which possess all the essential characters 

 of this genus, the sexual organs and perianth being exactly the 

 same. The petals of both corollas are imbricated, especially those 

 of the inner one.* The carpels are indefinite, and occupy the centre 

 of a convex receptacle. Each contains an indefinite number of 



Uvaria rufa. 



Fig. 229. 



Berry. 



' The hairs of which it consists are often 

 stellate, and arc whitish, t:iwny, or rust-coloured. 



=" See p. li)5, notes -t-6. 



3 Hook. & Thoms., Fl. Ind., i, 99, n. 7.— 

 TnwAiT., Enum. PI. ZeyL, 6, n. 3.— Walp., 

 Ann., iv. 46. We may make a section of this 

 genus under the title of Sj/vta-aria to include 

 those species in whii-h the corolla thus falls ofi' in 

 a single piece ; but we must recognise the fact 

 that the exact limits between this section and 

 those containing quite polypetalous species arc 



often very artificial ; a proof of the little value 

 of the genera Hexalobtis, &c. 



* Rriz & Pat., Piodr. Fl. Pei-uv. et Chil., 

 Si, t. 10; Si/st., i. 144.— DC, Prodr., \. 88.— 

 Dr>'., 3Ion., So. — Exdl., Gen., n. 4/17. a. — 

 U. H., Gen., 23, 956, n. 4.— H. B.v., Adansonia, 

 viii. 303. 



■^ They grow for a long time, even after the 

 expansion of the flower, and their bases become 

 gradually contracted, especially iu the inner 

 petals. 



