ANON ACE JE. 237 



most AnonacecB. However, in certain species, especially the one 

 which served to found the genus, at the lower part of each of the 

 inner petals is a sort of sac or obtuse spur projecting below the 

 insertion of the petal (fig. 292). The designation Saccopetalum^ has 

 been applied to plants in which this gibbosity is more marked, so 

 as to form a sort of purse or boat-shaped hollow. But as there is 

 every transition between those species of Saccopetahim in which this 

 is well developed, and those of Miliusa in which it hardly exists, it 

 has appeared impossible to us to retain the two genera as absolutely 

 distinct. Moreover, all other characters are the same in both, and 

 Saccopetalum has the gynseceum of the multiovulate Miliuaas. In the 

 former the leaves are caducous, and the fiowers spring from the axils 

 of the last year's leaves ; they are solitary, or in small clusters, often 

 supported on long slender peduncles. The young leaves of their 

 year appear with them, and are covered with rather copious down. 

 These characters allow us to make of the six or seven known 

 Oceanian and x\siatic'" species a special section in the genus Miliusa 

 as we understand it. 



The flower of PJt(saiithm^ is, as regards its perianth, exactly that 

 of Milium, but the form of its stamens is different ; they are formed 

 like those of Unoiia and Uoaria, with a short dilatation of the con- 

 nective, more or less depressed or rounded at the summit.^ The 

 ovaries and fruit are those of the uni- or biovulate Miliusas ; but the 

 ovules of Fhceanihus, instead of being near the base of the ovary, are 

 inserted a little higher up on the inner angle, and are slightly 

 ascending, with the micropyle downwards and outwards. Only five 

 species of Phaaufhus proper are known, natives of India and the 

 Indian Archipelago.' To this genus we think we may add Hetero- 

 petaJum brasiliense Benth.'' (fig. 295) as the type of a separate section 



* Benn., pi. Jav. Rarior., 165, t. 35. — Endl., extrorse, or nearly marginal. The filament is 



Gen., Suppl. i. n. 4712'.— B. H., Gen. 28, 958, articulated at the base, and falls very early. 



n. 35. — H. Bn., Adansonia, vii. 343. •'' Walp., Ann., iv. 73 ; vii. 57. — Zoll., Lin- 



- Walp., Rep., i. 74; Ann., iv. 76 ; vii. 59. naa, xxix. 324. — Miq., J"/. Ind.-Bnt., i. j). ii. 51; 



—Hook. & TnoMS., Fl. Ind., i. 151.— MiQ., Fl. Ann. Mus. Lugd. Bat., ii. 40. 



Ind.-Bat., i. p. ii. 52. — Zoll., LinncEa, xxxi. ^ Jonrn. Linn. Soc, v. 69. — B. II., Gen., 



325. — Bentii., Fl. Austral., \. 53. 27, n. 27. — H. Hn.. Adansonia, viii. 313. — 



^ Hook. &THOMS., Fl. Ind., i. 116. — B. II., Gnafleria /leteropetala Ue'STU., Hook: Joiirn., 



Gew., 27, 957, n. 25. — U..MN.,Adansonia,\iu.24:3. ii. 360. If this species is referred to the geuns 



■* This summit is like a long lozenge, with a Phaanthm it must take the name of P. hetero- 



large transverse axis in P. nutans Hook. & 2)etalus. The large petals are not so thick as in 



Thoms., and its superior surface more or less the Asiatic sjiccies, and the thick connective is ho- 



concave. The anther-cells are either markedly rizontally truncated above the anther-ceils. The 



