•22»J NATURAL HISTORY OF PLANTS. 



the projections from the petals (which also cover in the sexual organs 

 like a dome) are Hattened in the radial direction of the flower, instead 

 of beino- of the same breadth in every direction as in A. suaveolens, 

 or laterally compressed as in most species of RolUnia. The genus 

 Artaf)ofri/s contains about fifteen species, of which three or four come 

 from Africa,' and the rest from tropical and eastern Asia," or the 

 Indian Archipelago.^ They are shrubs, often climbers, with alternate 

 usually smooth leaves, and flowers grouped into clusters of often 

 few-flowered cymes. The chief axis of each cluster is flattened and 

 dilated into a sort of recurved fasciated hook, which bears, chiefly on 

 its convexity, groups of pedicellate flowers, whose development is 

 often partly arrested (fig. 278). f 



The tree from Ceylon called Ci/athocalyx^ has a corolla like that of 

 y/. inicnfa, with erect blades of even more membranous texture, and 

 only touching by the edges in the bud. But the calyx is like a deep 

 cup, whose edges alone are incised into three teeth, and the flattened 

 summit of the receptacle only supports a single carpel. The uni- 

 locular ovary only supports one parietal placenta, on which are 

 inserted two rows of anatropous ovules,' and the style rapidly dilates 

 into a large flattened stigmatiferous head. The fruit is a many-seeded 

 berry. The leaves are glabrous alternate, and the flowers are solitary 

 or grouped into few-flowered cymes, terminal or leaf-opposed." 



In }[cxalohiis^ (Ags- 279, 280), the perianth is like that of a 

 Ci/afhocnl'i/Xy or any of the species oi Artabofn/s analogous to A. uncata ; 

 l)ut the six petals are united into a tubular corolla in the whole of 

 the part enveloping the sexual organs. The membranous flat- 



came out. We may therefore accord priority to a unicarpellary Artdbotrys % and as we have 



the specific name which he proposed. said, there would be no doubt no reason for 



* Hook. F., Isiger, 207. — Benth., lAnn. hesitating to suppress the genus Cyathocalyx if 



7Va«*., xxiii. 466. — 'S\iq.,Ann. Mtis. Liigd.Bat., its flowers were borne on fasciate hooked axes, 



ii. 13. — Olit. Fl. of Trop. Afr., 27. since unicarpellary species are admitted in the 



''■ IIuoK. Y. & TiiOMS., Fl. Inch, i. 127. — genera Bocagea, JJmna, &c. As for the cha- 



Tnw., J?»MTO. PL Zeyl., 9. — Benth., Fl. Hong- racter derived from the depth of the calyx, which 



hong., 10. has given its name to the genus, it is of no great 



' Hl., (yp. oil., 59, t. 28-31. — MiQ., Fl. Ind.- value, since it may vary as much in several genera 



Bat., i. p. ii. 38 ; Suppl., i. 154 j Ann. Mus. otherwise perfectly natural, such as Unona, &c. 



Lttgd. Eat., ii. 38, 43. — Walp., Reji., i. 80 ; Moreover, we should not forget that in certain 



Ann., n.l^; iv. 63 ; vii. 53. species of ^r^aio^r^/* the axis of the inflorescence 



•• C. zeylank'un Champ., ex Hook. F. & is not, or at least is not constantly, hooked and 



TuoMS., Fl. Ind., i. 126.— B. H., Gen., 24, n. flattened. 



9. — Walp., Ann., iv. 63. — H. Bn., Adansonia, ' A. DC, Mem., 36, t. 5, A. — Endl., Gen., 



viii. 312, 341. n. 4718.— B. H , Gen., 24, 956, n. 11.— H. Bn., 



^ 'lliere are usually five or six in each row. Adansonia, viii. 312, 332, 341 (nee A. S. H. & 



•' The flower is on the whole quite tliat of '1 fl., Ann. Sc. Nat., ser. 2, xvii. 133). 



Ii 



