2(H 



NATURAL HISTOliY OF PLANTS. 



type species.' Many other species have since been observed in 

 GuiiuKi, Brazil, and Peru,'- in which the carpels become as numerous 

 as in the true U/i07ias of the Old World. In some of the latter 

 the carpels are, as we have seen, few in number. In Monocar- 

 pia^ there is only a single one, and yet we cannot retain it as a distinct 

 genus, for many other perfectly natural genera of the same order 

 contain species with several ovaries as well as species with only 



one. 



Ifclodorum" has been considered by some as a section of the genus 

 Vnona, by others as a perfectly autonomous genus. It has, however, 

 the indefinite multiovulate carpels and the valvate perianth of the 

 true JJnonas. The corolla, it is true, may undergo very great modi- 

 fications in the form and thickness of its parts. In certain species 

 it is all together globular in the bud, while in others it has exactly 

 the same form that is common in XyJop'ia, a form even exaggerated 

 in Pi/ranndn/ifheJ' Only this character cannot be considered of any 

 absfdute value, as there are species of Melodorum with conical buds, 

 whose petals are exactly those of some Unonas, and, on the other 

 hand, we may have globular buds in the latter genus. 



The same observation applies to Kentia^ whose general floral 



Bentham & Hooker (Gen., 956, n. 13) refer to 

 the genus Unona (see p. 203, note 4). 



^ Trijjyneia ohlong'ifolia Schltl., loc, cit. — 

 Uoaria triyyna Makt., Fl. Bras., Anonac, 40. 



■ H. Bx., Adan.sonia, viii. 179-181. We 

 liave referred to this group the following species : 

 1. Anona peduncularis Steudj 2. Ucaria 

 guatterioides A . DC ; 3. Anona Perrottetii 

 A. DC. In a plant that we can only consider 

 as a form {lanceolata) of this last species, we 

 liave only found one ascending ovule in each 

 carpel, just as in our Trigyneia rvfescens {loc. 

 cit., 180, note 1), while in T. Matheivsii Bexth. 

 and the species above enumerated, the number of 

 ovules is greater. In the former case, the seed 

 is ellipsoidal, with a longitudinal ridge along its 

 edge. When the steds are numerous, they are 

 reduced to tlattened disks, j)iled up like a rouleau 

 of coins, and the j)rojecting circular rim is hori- 

 zontal, occupying the only jiortion of the surface 

 of the seed that is not in contact with its 

 neighbours. 



^ AIlQ., Ann. Mus. Lugd. Bat., ii. 12. — B. H., 

 Oen., 956. n. 13, a, — II. Bn., Adan.sonia, viii. 

 338. 



* Dry., Mon., 115 (sect. fMonrc).— Bl., Fl. 

 Jai\, Anonac, 13, t. 15 (sect. Uvaria;). — DC, 

 Syst., i. 497; Prodr., i. 91. — Endl., Gen.,n. 



4717, «.— B. H., Oen., 28, 958, u. 31.— IMiq., 

 Fl. Ind. Bat., i. p. ii. 34 j Ann. Mus. Liigd. 

 Bat., ii. 37.— Hook. & TnoMS., FL Ltd., i. 115. 

 — Thw., Fnum. PI. Zeyl., 6. — ZoLL., Linnaa, 

 xxix. 317. — Benth., Linn. Trans., xxiii. 477 ; 

 Fl. Austral., i. 5^. — Walp., Ann., iv. 57. — 

 H. Bn., Adansonia, viii. 296, 306, 328. — (an 

 Lour., Fl. Cochinch., cd. Ulyssip., 351 ?). — 

 Cyatliostemma Gkiff., Notul., 707, t. 650. 



* MiQ., Ann. Mus. Lugd. Bat., ii. 39. — 

 H. Bn., Adansonia, viii. 329. — Several species 

 of Melodorum, properly so called, present this 

 elongated form of the Hower-bud, as Bexthaji & 

 HoOKEE very justly remark {Oen., 958). Others 

 have a conical bud, with the petals of a uniform, 

 or nearly uniform, thickness all over, as in certain 

 species of Unona and Polyalthia. Others again 

 have a globular bud, like that of Anona globi- 

 Jtora. The stamens are often surmounted by an 

 acute prolongation of the connective j but this 

 character has no absolute value, being wanting in 

 several species of Melodorum proper. In 

 M. africanum Benth. (Linn. Trans., xxiii. 

 477), we have observed that the outer stamens are 

 often transformed into petaloid scales, as occurs 

 in several species of Aberemoa, Xylopia, &c. 



^ Bl., fl Jan., Anonac., 71, t. 58 (sect. 

 Polyalthia).— K H., Gen., 28, n. 31 (2).— 



