1954] 



PLANTS COLLECTED IN NY AS ALAND 



491 



C. schimperi as they have been interpreted in India, but in addition there are 

 probably as many specimens which break down this correlation in a most unco- 

 operative way. Brass 16342 is one such, having the seeds not at all trigonous 

 but a short bushy dichotomous habit. My own present inclination would be to 

 treat the two seed-shapes as of subspecific importance, and for that reason I have 

 taken the species in a wide sense here. But it is possible that the difficulty 

 found in Africa is due to free crossing between two distinct species. Until this 

 can be studied carefully in the field, I feel that the taxonomic course adopted 

 here is at least not going to lead to more confusion than there may be already. 

 This genus is certainly one to be commended to the attention of botanists in 

 Africa. 



Hutchinson and Dalziel in 1931 made the new combination Cephalostigma 

 peroti folium and this has since been generally used. However there is no doubt 

 that Dentella erecta Roth provides an earlier epithet, as was recognised long ago 

 by Vatke. Indeed Wight and Arnott seem deliberately to have suppressed the 

 epithet erecta in favour of their own, perhaps because they felt that Roth had 

 made such a bad shot at placing the plant. Roth's description of the habit, the 

 mention of subsessile lanceolate leaves, and the glabrous calyx-lobes suggest 

 that his plant was what has been called C. schimperi and not C. hirsutum. If C. 

 hirsutum were to be considered as a distinct species it would apparently keep its 

 name. 



Lightfootia abyssinica Hochst. ex A. Rich. Tent. Fl. .Abyss. 2: 1. 1851, sensu 

 lato. 



Kota-kota District: Chintembwe, in rocky grassland, perennial herb 50-70 cm. 

 high,- flowers yellow-green, 1400 m., Sept. 9, 1946, 17587. Widespread from Abys- 

 sinia and the Anglo-Egyptian Sudan to S. Rhodesia and Angola. 



This genus requires critical revision and the naming should therefore be ac° 

 cepted with some caution. 



Lightfootia glomerata Engl. Bot. Jahrb. 19 (Beibl. 47): 52. 1894. 

 Lightfootia capitata Bak. Kew Bull. 1898: 158. 1898. 



Zomba District: Zomba Plateau, frequent in Brachystegia woodlands, herb up 

 to 1 m. high, flowers pale blue, 1500 m., June 2, 1946, 16156. Blantyre District: 

 Blantyre, in Brachystegia woodlands, herb 40-60 cm. high, with several to many 

 stems ascending from a stout taproot to form a bushy crown, flowers blue, 1100 

 m., June 18, 1946, 16354. North Nyasa District: Nyika Plateau, occasional in 

 open grasslands, perennial herb 30-40 cm. high, flowers blue, 2250 m., Aug. 16, 

 1946, 17257*; ibid., Nchena-chena Spur, common in open grasslands, perennial 

 herb 40-70 cm, high, rootstock fleshy, whitish, flowers blue, 2000 m., Aug. 20, 

 1946, 17348. Tanganyika Territory, N. and S. Rhodesia, and Nyasaland. 



For the present I consider that these capitate-flowered plants should be 

 treated as one species, but this opinion must be taken with caution until field ob- 

 servations are available on how the inflorescence varies, e.g. whether there is a 

 gradual passage from capitate to spicate. 



Brass 16354 closely resembles the type of L. capitata, and like that seems to 

 me to be a state in which the main stem has died or been decapitated, many lat- 

 eral shoots having thus been encouraged to grow out. Brass 17257 and 17348 ap- 

 pear to be the normally grown state of the same thing. It is perhaps worth noting 

 that the calyx-lobes of Brass 16354 are shorter (about 2 mm.) compared with those 

 of the other two sheets (about 3-4.5 mm.). Brass 17257 and 17348 are certainly 

 what has been called L. glomerata Engl. var. subspicata Engl, Pflanzenw, Ost- 

 Afr. C: 400 (1895), and also L. densa M. B. Scott, an unpublished name, L, 



