88 BULLKl'liN UK i/hKKBIKK BOISSIKK (i^^e SIÎR.). 1907 



Described from the unique type-specimen, now in Herb. Mus. Brit., 

 galhered by Don near the mouth oflhe Gambia, in 1823. Tiiere is no other 

 spécimen known, as tlie plant has not been collectée! since llie time of 

 Don's expédition. In his monograpli of tlie Andropogonese, Hackel simply 

 copies out Sir J. D. Hooker's fragmentary description, as he had not the 

 opporlunity of examining the plant. ïhe description given above had lo be 

 drawn up with some caution, for fear of damaging the by no means satis- 

 factory spécimen. In the Index Kewensis and in oiher citations the wrong 

 Don is given as theauthority for the species, for the reason probably thaï 

 Hooker does not indicate which of the three members of Ihe family is 

 isinlended. ïliere is neitherdate norlocality, nor note ofanykind, allached 

 to Ihe spécimen. 



3. Andropogon sorghum Brot. — Under the name of kous ex- 

 tensively cullivated ihroughoul the colony for native consumption, like 

 maize and rice. Large tracts are cleared round the villages on boLh banks 

 of R. Gambia for growing this African millet, which is one of the staple 

 industries of the natives. In his Origin of Cultivated Plants, p. 383, Al- 

 phonse de Gandolle says, « We must probably seek its origin in tropical 

 « Africa, where the species is generally cultivated. Sir William Hooker 

 « mentions spécimens from the banks of the river Nun, which were per- 

 « haps wild. The approaching publication of the Gramina in the Flora of 

 « Ti'opical Africa will probably throw some light on the question. » This 

 pious wish, expressed 24 years ago, will shortly be fulfilled. The final 

 volume of the Flora of Tropical Africa, devoted to the Graminacese, is 

 nearly ready for pubhcation. There are three, if not more, varieties of 

 African millet cultivated in British Gambia; as there are at least three 

 native names for différent forms of the cereal. None of thèse are indicaled 

 in Hackel's monograph of the Andropogonese, but more explicit détails are 

 given by Körnicke ' in his exhaustive work published four years previously. 

 So as to distinguish the three usual forms which are cultivated in the 

 colony. in the following descriptions Hackel's names and characters are 

 given. with the critical détails of Körnicke and others, — omitting the 

 specific characters common toall the varieties. 



subsp. sativus var. ovulifer Hack.l. c. 510. — Panicula ovali-oblonga 

 laxiuscula, ramispatentibuscum rhachi scabris, inferioribus dimidia pani- 

 cula brevioribus, ad 4-5 cm. inferne nudis. Spiculaehermaphroditse 5 mm. 

 long., 3 mm. lai., demum pulchre purpureo-spadiceae, praeter gluraœ primée 

 margines glaberrinicB. Arista 12 mm., genu longe exserlo. Caryopsis ovato- 

 subglobosa. glumis parum brevior. nivea, glumis valde hiantibus denudata. 

 Pedicelli spicularum sterilium quam dimidia spicula hermaphrodita parum 

 breviores. {descr. ex typo m herb. Nées). 



Known only from Gambia (Dur. et Schinz, Consp. fl. Afr. Y. 725). 



subsp. sativus var. obovatus subvar. niger Hack. 1. c. 514. — Panicula 

 ovala laxiuscula vel taxa, ramis inferne Isevibus, spiculee hermaphroditse 

 demum splendide nigerrimœ. Arista 6 mm. (native name, « BassiWulima «). 



Syn. Holms niger P. Arduino, in Saggi scientif. e lelterar. Accad. Padova, 

 i. p. 134, l. 5 (1786); A. sorghum subsp. contractns var. niger Körn. 

 I.e. 311. 



* Arten und Varietäten Getreides (188o), being the first volume of Körnicke & 

 Werner's Handb. Getreideb. 



