84 cxLiii. ooMMELiNACiLE (clarke). [Cyanotis. 



exserted. The sect. Ochrecejlora (DC. Monogr. Phan. iii. 244) is not a section — but 

 a group of species with undefined limits. Scott-Elliot's examples are incomplete, and 

 are less sliaggy than the typo specimens ; but they must be eithei' C. somaliensis, or 

 a species clos(?ly allied thereto. 



14. C. flexuosa, C. B. Clarke. Glabrescent. Roots fibrous. 

 Stem procumbent, rootin^^ at the lower nodes, with very numerous stout 

 branches attaining 20 in. long, with flexuose internodes. Leaves oblong- 

 lanceolate, attaining 2 in. long, the upper by degrees shortened into the 

 floral leaves, which are ovate abbreviated into a short narrow recurved 

 tip. Inflorescences axillary, very numerous, nearly included in the leaf 

 sheaths ; bracts not appearing 2-ranked. — C. nodijiora, var. ft madagas- 

 carica, C. B. Clarke in DC. Monogr. Phan. iii. 258 partly ; Durand & 

 Schinz, Conspect. FI. Afr. v. 435; Rendle, Cat. Afr. PI. Welw. ii. 80. 

 Commelina Jlexuosa, Welw. ex C. B. Clarke in DC. Monogr. Phan. iii. 

 258. 



ZiO\irer Guinea. Angola : Huilla ; Monino, in damp rocky places, very rare, 

 3800-5500 ft., Wehoitsch, 6584 ! 



This is more remote from C. nodijiora than is the Madagascar plant with which 

 it was united in DC. Monogr. Phan. I.e. 



Imperfectly known sjyecies. 



15. C. Dybowskii, Hua in Bull. Mus. Hist. Nat. Par. i. 122. 

 Hirsute. Stem simple, rather large, erect. Leaves linear, the upper 

 surface glabrous, the lower hairy. Cymes at the apex of the stem or of 

 axillary branches, several, approximate, nearly sessile ; bracts similar 

 to the leaves, or shorter ; bracteoles up to 20, falcate, hairy. 



XTpper Guinea. French Congo ; Kemo Station, DyhowsJci, 760. 



Not seen. Hua compares it with C. djurensis ; he does not describe, and pro- 

 bably did not possess, the root. There is a Cyanotis from Mount Elphinston 

 Flemming, on the Kworra (Niger), Barter, 492, wrongly referred by me to C. lanata, 

 Benth. (in DC. Monogr. Phan. iii. 258). The root of this is not known, but the 

 portion seen of the robust stems is 18 in. long, little divided, the stem-leaves 4^ by 

 ^ in. It is certainly not C. lanata, Benth., and had been removed to the neighbour- 

 hood of C.djurensis. So far as the description goes, it. might be C. Dyboivskii, Hua. 



16. C. cephalotes, Fenzl ex Schweinf. Beitr. Fl. Aethiop^ 205 ; 

 Durand & tSchinz, Conspect. Fl. Afr. v. 435. 



DTile- Ziand. Sennaar (ex Fenzl), 



This appears to be a bare name ; at least, Durand cannot find where it is described. 



10. FLOSCOPA, Lour. ; Benth. et Hook. f. Gen. PI. iii. 852. 



Sepals 3, oblong to obovate. Petals 3, obovate, sessile or scarcely 

 clawed, distinct. Stamens 6 (rarely 5), perfect ; filaments without 

 hairs. Ovary 2-celled, with 1 ovule in each cell. Capsule obovoid, 

 compressed, membranous, loculicidally 2-valved, 2-seeded. Seeds soli- 

 tary, attached laterally, hemispheric or depressed conic. — Stems all 

 bearing leaves. Inflorescence hairy and often glandular, of many 

 flowers in a terminal more or less leafy panicle ; upper part of the 



