(W; ( XLiii. coMMELiNACE.t: (clarke). [Aiieileiiha. 



Buchanan, 286! Mount Sochi, Scott-Mliot, 8527! Manganja Hills, Waller! 

 Mount Cliiradziiln, Meller ! and witliout precise locality, Buchanan^ 459 ! 510 ! 528 ! 

 1401 ! • Matiibclcland, Elliott ! Salisbury to Buluwayo, Hon. Mrs. JEvelyn Cecil, 81 ! 



Also in South Africa, the Mascarene Islands, and Arabia. 



Schimjier, 1660, contains, by admixture, both A. cequinoetlale and A. tacaz- 

 zeanum. A, Kichard describes therefrom his A. tacazzeanum as having six seeds to 

 each cell of the capsule, so that his plant was certainly A. cequinoctiale, Kunth. 

 Lamprodithyros tacazzeanus, Hassk. in Peters, Reise Mossamb. Bot, 531, is said to 

 be Aneilema tacazzeanum^ Walp. Ann. iii. 658, which is merely copied, Avith a little 

 confusion, from A. Richard ; but Hasskarl describes his Lamprodithyros tacazzeanus 

 at great length, and says the ovules were only 2-1 in each cell ; so that Hasskarl's 

 ])lant was really the tacazzeanus of Hochstetter, excluding the citation of Walpers. 

 There is ample evidence of the vari-ation in the colour of the petals ; the different- 

 coloured flowers have been esteemed different species by many collectors. 



?>. A. tacazzeanum, Hochst. ex C. B. Clarke in DC. Monogr. Phan. 

 iii. 222, not of A. Rich. Thinly hairy. Stems 1-2 ft. long, weak, de- 

 cumbent at the base and there sometimes divided, rooting often at the 

 lower nodes. Leaves :] by 1 in. (in Hochstetter's two " types "), elliptic, 

 i»cute, base contracted, often subpetioled. Inflorescence terminal, pe- 

 duncled, 1|- by Ih in., very lax, obscurely pubescent; cyme-peduncles 

 alternate ; bracts less than ^ in. long, ovate ; bracteoles less than J in. 

 long, obovate. Sepals about \ in. long, elliptic, concave. Corolla 

 .smaller than in A. cequbiocticde, appears to have been yellow. Ovules 

 :5-2 in each cell. Capsule nearly J by J in., white, truncate, with the 

 two upper angles acute, mostly 2-celled, with 3-1 seeds in each cell. — 

 Durand & Schinz, Conspect. Fl. Afr. v. 432; K. Schum. in Engl. Pfl. 

 Ost-Af]'. C. 130. A. Forskalcei, Schweinf. in Bull. Herb. Boiss. ii., 

 Append, ii. 5S, not of Kunth. Lamprodithyros tacazzeanus, Hassk. 

 in Peters, Reise Mossamb. Bot. 531 (excl. the syn. of Walpers), and 

 in Schweinf. Beitr. Fl. Aethiop. 295. Cfr. DC. Monogr. Phan. iii. 

 '22-2. 



wile Iiand. Eritrea : Gorge of Ailet, Schtoeinfurth Sf Riva, 427 ! below 

 Geleb, 5000 ft., Schtoeinfurth, 1416, 1788 ; Ginda, 3250 ft., Schweinfnrth, 128, 392. 

 Abyssinia : Valley of the River Tacazze, Schimper, 1660, 1729 ! Quartin-Dillon cf 

 Petit, 449 ! 



This is a much weaker plant than A. cequinoctiale, the capsule smaller, with often 

 only 2-1 seeds in each cell. The example of Schweinfurth, 427, is much stouter 

 than Schimper' s, with larger inflorescence and 3 seeds to each cell; this is separated 

 from the smaller examples of A. cequinoetlale by the smaller, more concave elliptic 

 sepals. Hua (in Bull. Mus. Hist. Nat. Par. i. 121) appears to draw the line between 

 the two species at a different point. 



Schweinfurth reduces here A. Ehrenbergii, C. B. Clarke (which see No. 27 below), 

 and A. Forskalcei, Kunth. The latter is fully described as Commelina paniculata by 

 Vahl (Eninn. ii. 179), who appears to have had Forskal's plant; and who describes 

 inter alia " peduncles lateral and terminal, paired, with a sheath \ in. long at the base 

 of each" ; so that this plant must have been very remote from A. tacazzeanum, and 

 was almost certainly a species of Commelina. 



X. A. nyasense, C. B. Clarke. Pubescent. Portion of stems seen 

 1 ft. long, unbranched. Leaves 3 by 1 in., elliptic, acute. Panicle 

 terminal, 5 by 2J in., loose; cyme-peduncles mostly in distant false- 



