150 THE JOU11NAL OF BOTANY 



account of the previously existing L. hernsleyana Maxim. It may 

 be named L.frcmchetii R. Knuth. Carex cavaleriensis [sic] should 

 be rejected on account of the earlier Carex at nth riei. The emploj - 

 ment of such pairs of names is bound to lead to confusion, especially 

 in the case of cultivated species. 



5. Accidental binomials occurring in works published since 1753, 

 but which did not adopt the binary system of nomenclature consis- 

 tently, should be rejected. 



Examples : : — Those contained in Garsault's works (1764-7), 

 revived by Thellung in Bull. Herb. Boiss. ser. 2, viii. 713, 778 

 et sea.; those in Hill's British Herbal (1756), revived by Druce 

 in Rep. Bot. Exch. CI. Brit. Isles, 1913, iii. 43(3 (1914). 



6. Specific names should be rejected when they are homonyms, 

 irrespective of the validity of the earlier name. Under Art. 28 and 

 51 of the International Rules the validity of a specific homonym is 

 dependent on the non-validity of the earlier name. As it may be a 

 matter of opinion whether the earlier name is valid or not, the 

 validity under these Articles of many homonyms will always be more 

 or less' doubtful. It is therefore desirable, in the interests of " fixity 

 of nomenclature," that all specific homonyms should be rejected.. 



7. If a new combination is associated by its author in the original 

 place of publication with specimens belonging to a different species, 

 it should be treated as a " nomen delendum " ; and the combination 

 may be made de novo. If the original combination were treated 

 as valid, it would become " a permanent source of confusion " 

 (Art. 51, 4°). 



Examples : — The new combination Mania nervosa was applied 

 by Oliver ( Fl. Trop. Afr. i. 84 ; 1868) to a Zambesiland species 

 which he erroneously identified with Niebuhria nervosa Hochst., 

 a native of Natal. Grilg and Benedict have shown that the Zambesi- 

 land and Natal plants are not conspecific. They have re-named the 

 former Moerua cylindricarpa Grilg et Benedict, and have proposed 

 the combination Moerua nervosa (Hochst.) Grilg et Benedict for the 

 latter (Engl. Jahrb. liii. 241, 244; 1915). Even if the name 

 Mceraa nervosa Oliv. is retained, it is uncertain to which species it 

 should be applied; whether to the Zambesiland species on which 

 Oliver's description was mainly based, or to Niebuhria nervosa 

 Hochst. which he cited as a synonym and of which he retained the 

 trivial. Botanists are not agreed as to the correct incidence of the 

 name in such a case. 



Schumann proposed the new combination Bolichandrone atro- 

 virens expresslv to replace B.falcata Seem. (Engl, et Prantl, Nat. 

 Pflanzenfam. iv. 3 B, 240; 1894), being misled by C. B. Clarke's 

 citation of Bignonia atrovirens Heyne as a synonym of the latter 

 (PI. Brit. Ind. iv. 380; 1884). But as I have elsewhere pointed out 

 ( Kew Bull. 1919, 307), Bignonia atrovirens is conspecific with 

 Bolichandrone crispa Seem., and the combination D. atrovirens 

 (Heyne) Sprague should accordingly be used in place of B. crispa. 

 Schumann cannot be quoted as the authority, inasmuch as he applied 

 the name to a different species. 



