SECTION V: G. KIRKII 317 



ones that are ovate rotund, deeply aurioled (almost peltately attached), thick, 

 coarsely and conspicuously toothed, tomentose on both surfaces, especially 

 when young, accrescent and submembranous on the fruit, which they com- 

 pletely enclose. Corolla small, tube very short, petals obovate, almost 

 uniform, that is not oblique, with a well-marked darker coloured spot near 

 the claw, outer surface looking striated, owing to the assortment, in trans- 

 verse lines, of very short compact stellate hairs. Calyx small, campanulate, 

 truncate, minutely toothed, thick woolly, veins not visible, but a ring of glands 

 placed near the base on the outside and a corresponding tuft of fairly long 

 hairs on the inside. Fruit small, rotund ovate, 3- or 4-celled ; seeds few, 

 ovate, smooth, polished, free from each other, not angled through being com- 

 pacted together, as in most wild species, but striated with grey bands that 

 correspond with the lines along which the short, dark, rust-coloured wool is 

 attached (a striation that recalls the markings on the seeds of G. brasiliense) ; 

 wool so readily separable as to usually show only the tip of the seed with a 

 minute tuft of hairs adhering (ff. 5 & 6). For an account of the pollen-grains 

 see p. 348. 



Habitat. East Tropical Africa, purely wild and never apparently Tropical 

 cultivated. Africa - 



Citation of Specimens. The following are examples seen in the Kew Speci- 

 Herbarium : Dar-es-Salaam (1869), Sir J. Kirk; N. Bote on Umba Eiver, mens. 

 by Ksessner, n. 28 ; Dodori, East Africa Protectorate, Tanaland (top of low 

 hill, Mundane Eange), by Brand. Eecently a further supply has come from 

 Malindi, East African Protectorate, where it is said to be found in remote 

 forests truly wild. In the British Museum, North of Mombassa to Lamu 

 and Witu, collected by A. Whyte (1902). 



Nomenclature. This plant has, in what might be called its exter- p ogs ible 

 nal appearances, a strong resemblance to G. brasiliense. In fact, hybrids, 

 although I can discover no record of any practical tests that would 

 support the suggestion, I am strongly tempted to speculate that it 

 may have contributed to the production of some of the special African 

 cultivated races attributed to that species. The leaves of G. brasi- 

 liense in shape, texture, comparative absence of hairs, and the colour 

 into which they dry in the herbarium, are distinctly suggestive of the 

 wild G. Kirkii, The stipules of G. brasiliense (and some of the other 

 plants associated with it) do not, it is true, become so pronounced as 

 those of G. Kirkii s but they are nevertheless the most conspicuous 

 stipules of all known cultivated cottons. Lastly, the undoubted 

 abundance of G. brasiliense in Africa, often met with in a state of 

 complete acclimatisation, and the claims sometimes made for its 

 being even indigenous, may be accounted for by this purely African 

 species having been mistaken for Brazilian cotton. But there is 

 perhaps no other wild species more worthy of careful study than the 



