( ciii ) 



du Bresil aux iles du Midi de FAsie par les Portugais, depuis 

 la fin du XV e siecle." A. De Candolle, " Origine des plantes 

 cultivees," p. 333. 



" II n'est pas douteux que l'introduction sur la cote occi- 

 dentale d'Afrique se soit faite par l'intermediaire des negriers 

 portugais, des le XVI e siecle. La culture de l'Arachide sue 

 le continent noir prit rapidement une grande extension, car, 

 au dire de Sloane, les negriers chargeaient leurs navires de 

 cette graine pour nourrir les esclaves pendant la traversee." 

 Dubard in Bull. Mus. Hist. Nat., vol. xii (1906), p. 340. 



Mr. Farquharson does not state whether the beetles feed 

 on the large conspicuous sterile flowers on the upper part of 

 the plant, or on the small inconspicuous imperfect flowers 

 producing hypogaeal pods. [Mr. Farquharson has since stated 

 that the former are eaten.] 



Dr. G. A. K. Marshall has described and figured South 

 African Cantharid associations with their Mullerian mimics 

 belonging to the Longicornia, the Phytophaga and the Heini- 

 ptera (Trans. Ent. Soc, 1902, pp. 518, 519, plate XIX, figs. 

 1-17). The patterns are similar to those here described. 

 He has also shown that certain South African Cantharids 

 have adopted the prevalent and very powerful Lycoid aposeme 

 (ibid., p. 517, plate XVIII, figs. 20-23). 



Another interesting feature of Mr. Farquharson's collection 

 is the possible indication of sexual preference, although the 

 evidence, to become convincing, would require to be immensely 

 increased. The table on p. cii shows that only 5 examples 

 (3 <J, 2 $) of the type form of M. hermannioides were 

 captured, as compared with 40 of the var. lambomi (10 £, 

 9 $ in Mr. Farquharson's series). Two of the <$ type were 

 paired with $ vars. ; the third $ type was paired with a £ 

 type; the second $ type was unpaired. Thus, considering 

 the small number of <$ and $ types, there is a very slight 

 indication that type prefers to pair with type — an indication 

 sufficient, I hope, to stimulate the collection of evidence on 

 a large scale. The absence of any pairs of the var. <$ and $ 

 is surprising. Four of the type form and six of the var. are 

 shown in the lowest row of figures on Plate B. 



Nothing can be made of C. hermanniae or D. affinis, from 



