32 



MEMOIRS OF THE NEW YORK BOTANICAL GARDEN 



[Vol. 9, No. 1 



[2b. Selago thyrsoidea Bak. var. austrorhodesica Brenan, vaf. nov. 



Caules ramulos numerosos laterales breves foliatos ex axillis emittentes. 

 Nodi minus conferti, 3-5 mm. inter se distantes. Folia caulina plerumque circiter 

 1.2-2 cm. longa. Calyx usque ad vel ultra medium divisus. Lobus maximus corollae 

 circiter 2 mm. longus. 



S. RHODESIA: Inyanga District: Troutbeck, all over the veld, a very common 

 herb, mauve flowers, 2130 m., Mar. 21, 1948, /. M. Rattray 1405 (Govt. Herb. 

 No. 20641) (TYPUS varietatis in Herb. Kew.). The Downs, Jan. 13, 1944, C. P. 

 Piers in Govt. Herb. No. 11658 (Herb. Kew.).] 



[2c. Selago thyrsoidea Bak. var. nyikensis (Rolfe) Brenan, comb. nov. 

 Selago nyasae Rolfe in Thiselton-Dyer, Fl. Trop. Afr. 5: 270. 1900. 

 Selago nyikensis Rolfe, Kew Bull. 1908: 261. 1908. 



Caules vel rami laterales primarii aut subsimplices aut ramulos numerosos 

 laterales breves foliatos emittentes. Folia caulina plerumque 0.6-1.2 cm. longa. 

 Calyx usque ad medium vel ultra divisus. Lobus maximus corollae 1-2 mm. longus. 



Tanganyika Territory, Portuguese East Africa, and Nyasaland. I have adopted 

 the epithet nyikensis rather than nyasae as the cotypes of the latter are both 

 scrappy, but nyikensis may be no more than an ecad of var. thyrsoidea. 



S. thyrsoidea in its wide sense is very easily distinguished from 5. thomsoni 

 by its paniculate not corymbose inflorescence, coupled with linear glabrous 

 leaves. The calyx-teeth of S. thyrsoidea are usually more densely ciliate than 

 in S. thomsoni, and very often more deeply divided; in addition the bracteoles 

 are often longer. 



S. thyrsoidea is apparently related to the S. African S. schlechteri Rolfe, 

 which may possibly prove synonymous, but with short straight or slightly curved 

 not crisped puberulence and much more ample panicles; also perhaps to 5". meyeri 

 Choisy (5. ramulosa E. Mey., non Link), but with more ciliate calyx-lobes.] 



[3a. Selago welwitschii Rolfe var. welwitschii. 



Selago welwitschii Rolfe, Jour. Bot. 24: 175. 1886, sensu stricto. 



Selago hoepfneri Rolfe in Thiselton-Dyer, Fl. Trop. Afr. 5: 271. 1900, sensu stricto, 



quoad Hoepfner 42, lectotypus speciei. 

 Selago holubii sensu Rolfe in Thiselton-Dyer, Fl. Trop. Afr. 5: 271. 1900, pro parte, 



quoad McCabe 28] non S. holubii Rolfe. 



Caules et folia pilis minutis decurvis appressis sparse usque plus minusve 

 dense vestiti. 



S. Rhodesia, Angola, Bechuanaland Protectorate, and South Africa.] 



[3b. Selago welwitschii Rolfe var. holubii (Rolfe) Brenan, comb. nov. 



Selago holubii Rolfe in Thiselton-Dyer, Fl. Trop. Afr. 5: 271. 1900, sensu stricto, 

 excl. McCabe 28. 



Selago hoepfneri sensu Rolfe in Thiselton-Dyer, Fl. Trop. Afr. 5: 271. 1900, excl. 

 Hoepfner 42; non S. hoepfneri Rolfe. 



Caules et saepius etiam folia pilis longis patentibus rectis vel flexuosis plus 

 minusve dense vestiti. 



S. Rhodesia, Angola, Bechuanaland Protectorate, Southwest Africa, and ? the 

 Transvaal. The various numbers collected by Holub on which Selago holubii was 

 based are all rather uniform, and all little bits. I choose Holub 1090 as the 

 lectotype. 



Schinz s.n. (Omulonga) and Schinz 8, both cited by Rolfe in the Flora of 

 tropical Africa under S. hoepfneri, are in my opinion var. holubii, but they are 

 shorter-haired than usual and somewhat intermediate between var. holubii and 

 var. welwitschii. 



