i 9 i6] SHERFF—BIDENS 497 



more diverse characters. The twining habit is to be met with in B. rubijolia 

 and its allies; the large number of achene awns (very rarely so numerous in 

 Bidens) suggests a certain African species of Bidens, to be treated later; the 

 shape of the achenes is rather that of B. tripartita and its allies; but (he thick, 

 necklike constriction at the apex of the achenes and below the awns is so pro- 

 nounced as to suggest a slight affiliation with Heterospermum, 2 in which genus 

 the outer, winged achenes arc frequently so constricted. 



Bidens pilosa L. Sp. Plant. 832. 1753. — Bidens hispida H.B.K. 

 Nov. Gen. 4: 237. 1820; Bidens andicola H.B.K. var. B, Wedd. 

 Chlor. And. 1:70. 1855. 



For many years, Bidens hispida H.B.K. appears to have 

 remained a comparatively unknown form to botanists dealing 

 with South American plants. Weddell's treatment of it is short 

 and arbitrary. No reasons are given by him for its reduction to 

 Bidens andicola H.B.K. 



The original type plant of Bidens hispida, as also that of Bidens 

 andicola, remains in good preservation today (in Mus. Hist. Nat. 

 Paris). In a very general way the two plants, both of them hispid, 

 show a resemblance; but Bidens hispida is seen, on closer inspec- 

 tion, to have small heads and these, moreover, are discoid. Also, 

 the lower part of the plant displays a wholly different foliage aspect 

 from that found in more recent specimens of the yellow-rayed 

 B. andicola (the lower part of the type for the latter species is 

 lacking). The plant is seen to be merely a very hispid form of 

 B. pilosa L. It is matched more or less exactly by many speci- 

 mens in herbaria, particularly by some from Venezuela, Colombia, 

 and Peru (for example, Mr. and Mrs. J. N. Rose 18740, Santa 

 Clara, Peru, July 18, 1914, in U.S. Nat. Herb.). 



Bidens anthriscoides DC. Prodr. 5:600. 1836. 



DeCandolle's description of this species, based upon Bcrlan- 

 dier 1010 and 1152, is much too narrow not to mislead. Excellent 

 specimens of Bcrlandier 1010 are in London (in Herb. Brit. Mus.), 



2 A genus very closely connected with Bidens through Heterospcrmiim Xanli Gray, 

 the type of which (in Herb. Gray) is matched by that of Bidens Xantiana Rose (in 

 U.S. Nat. Herb.); a species best retained in Heterospermum, a view in accord with 

 the treatment by Brandegee (Zoe 1:309. 1890), who, moreover, has since positively 

 corroborated (in litt. 1913) my equation of Bidens Xantiana Rose with Heterospermum 

 Xanti Gray. 



