273 



The following vegetables and salads thrive well at M'bali 

 Sana : — 



Bean (Broad Windsor, Dwarf Bean and Haricot varieties), 

 Pea (Stratagem and others), Cabbage, Carrot, Potatoes, 

 Pumpkins, Vegetable Marrow, Parsnip, Tomato, Onion, 

 Leek, Celery*, Cho-Cho, or Christophine {Sechium edule), 

 Lettuce, and Hadish. Ehubarb also thrives well. 



LIIL— NOTES ON THE GENERA CORDYLINE, 



DRACAENA, PLEOMELE, SANSEVIERIA AND 



TAETSIA, 



N. E- Brown, 



Whilst preparing a monograph of the genus Sansevieria it 

 became evident that the place assigned to it in Haemodoraceae by 

 Bentham in Bentham ^ Hooker^ Genera Plantarumy vol. iii* p. 



679, cannot be sustained, and that its ivueiposition is mLiliaceae, 

 next to Dracaena. Engler, in Engler ^ Prantl, Pflanzen- 

 faTnilien, vol* ii. pt. 5, p. 84, has already transferred it to 

 Liliaceaey though he has not laid stress upon its manifest affinity 

 with the genus Dracaena. Apart from the form and fleshiness of 

 its leaves, there is absolutely no technical character by which 

 Sansevieria can be separated from Dracaena as that genus is at 

 present constituted. There are shrubby and stemless species in 

 both genera, and some species of Dracaena have thick coriaceous 

 or sub-fleshy leaves, whilst the inflorescence, articulation of the 

 pedicel, flowers, fruit and seed of Sansevieria are in no way 

 different from those of many species of Dracaena. After having 

 examined the whole of the Kew material and tabulated all the 

 vegetative and structural characters in an endeavour to discover 

 some definite distinctive character whereby Sansevieria might be 

 distinguished from Dracaena, the writer is satisfied that although 

 Sansevieria may be recognised by its appearance, yet when the 

 details are put into words, there is no single character or combina- 

 tion of characters that will really distinguish all the species of 

 Sansevieria from all the species of Dracaena as the latter genus is 

 at present constituted. If, however, Dracaena be divided into 

 two genera, in accordance with the presence or absence of a dis- 

 tinct tube to the perianth^ then definite distinctive characters can 

 be stated- It is, therefore, here proposed to readjust the species 

 of Dracaena into two genera characterised by a structural differ- 

 ence in their flowers. 



The genus Dracaena was founded in 1768 by Vandelli upon the 

 well-known D, Draco, Linn., of the Canary Islands, in a very 

 rare work entitled Dissertatio arbore Draconis, which is re- 

 printed in Eoemer, Scriptores, pp. 39-46 and 58, t. 2^, 2^. In 

 this plant the perianth is divided into six segments nearly to the 

 base, without any very evident tube, except such as may be formed 

 bv the mere overlapping of the margins of the segments. Subse- 

 quently other plants have been added to this genus in which the 



