■ AXONACEJE. 203 



every warm climate, wliose corolla is much elongated, wliile the sta- 

 mens are surmounted by an acute prolongation of the connective. 

 The berries are slightly contracted between the seeds. The other 

 differentiating characters cited to separate this plant from U?io??a 

 having no real existence, we have thought it right not to retain it in 

 a distinct genus. 



In certain other Unonas which form the section DaspnascJialon,^ 

 the outer petals become far more elongated still ; the inner ones 

 remain small, or are even altogether wanting. This is usually the 

 case in the flowers of U. lougijiora^- in which the corolla is as much 

 as from 2J to 3 inches [ 6 to 8 centimetres] in length. The number of 

 ovules is often not great in the carpels of this species ; certain 

 ovaries contain no more than two, inserted at variable heights on the 

 inner angle. ^ This number is constant in certain species from tropical 

 Asia ; and it has been impossible to separate them generically from 

 the other Unonas, for there is no dissimilarity in any other character. 



It has been proposed to establish a fresh genus Meio(/pie* for a 

 Javanese Uvaria, whose llowers have rather short buds, and only 

 five, four, or even three carpels. This generic section has not been 

 adopted. It has been truly remarked with reference to this, that 

 there is an African species of TJnona^ which we cannot exclude from 

 the genus, with a corolla differing in form from that of Meiogijne, 

 but, like that plant, with but few carpels. Now this species has all 

 the essential characters of the American Unonas described under the 

 name of Trigyneia.^ The perianth of the latter may have exactly 

 the same conformation, and the generic name shows that the same 

 reduction" in the number of carpels has been observed in the proto- 



' See p. 201, note 6. distinct genus, simply on account of the small 



- RoxB., Plant. Coromand., iii. 87, t. 2yU ; number of its carpels. 



jFZ. Jw(Z., ii. 668. — Hook. & Thoms., JF'Z. Zhc^., i. ^ B. H., Gen., loc. cit. — V. Oliveriana 



134. — Walp., Ann., iv. 67. — H. Bk., Adau- H. Bn., Adansonia, viii. 307. In this species 



sonia, viii. 176. The single whorl that rcprc- n)ost of the organs are covered by very tine scaly 



sents the corolla may be even reduced to two hairs. There are from three to live carpels. In 



eqival or unequal pieces, free, or united for a very the former case they are superposed to the outer 



variable height. petals. In the bud the corolla has exactly the 



^ They are sometimes almost superposed. same form as that of certain species of Jhlo- 



Nevertheless each is placed (m one lij) of the durum, a triangular pyramid, with somewhat 



placenta. Sometimes the one is almost basilar : blunted angles. 



by this means we cannot distinguish this plant ^ Schltl., Z/rt»ir«, ix. 328. — Bexth., Jbiovi. 



from Po/_(/«/^/i/a, properly so called. Linn. Soc, v. 69. — B. H., Gen., 25, u. 15. — • 



■• MiQ., Ann. Mu.f. Liiffd. Bal., ii. 12. — II. Bx., .•if/(7«so7(/«, viii. 178, 337. 



V>. H., Gen., 956, n. 13. — H. Bk., Adansonia, '' The number of carpels is also, as we liavi' 



viii. 337. Ancana, too, has been proposed as a seen, much reduced in Ancana F. MuJiLL., wliich 



