CHARACTERISTIC'S OF LARV-E OF ANOPHELINA. 119 



iiistauees. On the first occasion it was observed in a living- 

 larva. Six others collected at the same time were closely 

 examined, but no such hair was detected. It is translucent, 

 and overlies the palmate hair in a different plane, and in 

 hurried examination of a lively larva might readily be over- 

 looked, unless the observer were especially alert for it, when 

 it cannot be missed. Subsequently, in looking through some 

 mounted specimens, this hair was found in a second instance. 

 A large collection was made from the pool from which the 

 latter had been taken some months earlier. Thirty individuals 

 were carefully examined, and a similar hair was found in one 

 only. Both larvae developed into imagines. No difference was 

 observed between the markings of these two and of others, 

 greater than between any of the combinations given on page 

 149. Both proved to be males, but male insects were obtained 

 from larvae on which this dendrifoi-m hair was not present. 



The ])rinciple of grouping into a genus different species, of 

 which the points of resemblance in imago are no stronger 

 than the features of difference in larva is unsound, but it 

 must be readily admitted that any attempt to give equal con- 

 sideration to larval characters would make all attempts at 

 generic classification hopeless. 



James and Liston propose to divide Anophelina into 

 groups, taking into consideration the characters of egg, larva 

 and imago, together with habitat, habits and pathological 

 affinities, particularly the last. This is doubtless satisfactory 

 for a country in which these particulars are known, but may 

 be misleading if applied to the sub-family generally. In 

 such an arrangement the South African species funesta, 

 ardensis, cinereus, pretoriensis and natalensis would 

 be grouped with the malaria-carriers listoni (or Christo- 

 pher si, according to Giles) and culicifacies, but cos talis, 

 as the authors themselves state, with Stephen si and rossi. 

 Funesta is a known nuilaria-carrier, but there is no evidence 

 against the other four, whereas costal is, although placed 

 with rossi, which ajjparently does not act as host to the 

 malaria parasite in natural conditions, is, as we have shown 



