74 



INSECTS OF SAMOA. 



while the $ of the former is considerably larger than the male. Among the 

 material collected, there is a slight variation in size and coloration. The 

 specimens taken in Savaii are much darker than those obtained in Upolu. In 

 the latter material the brown margins of the pronotum are sometimes much 

 more widened at the apex. 



First segment of antennae much the longest, second and fourth subequal, 

 or the former a little shorter than the latter ; third, the shortest ; fourth, unlike 

 that in L. huxtoni, always much shorter than the first. 



Genital segments of both sexes very similar to those of L. huxtoni as shown 

 in the figures. 



Length of body : ^ 5-6 mm., $ 9-10 mm. ; breadth of body : ^ 2 mm., 

 $ 3 mm. 



Upolu : Malololelei, 2,000 feet, type ^, type 3 and 2 $ paratypes, 

 4.V.1924 ; 1 ^, 1. v. 1924 ; 6 6 $$, 14-30.vi.l924. Savaii : Safune, rain 

 forest, 2,000-4,000 feet, 4 3 9$, 5.viii.l924 (Bryan). 



It appears probable that this species is confined to the mountain streams ; 

 L. huxtoni occurs at lower levels. 



6. Halobates princeps B. White. 



Hulohates princeps B. White, Rept. Voij. " Cliallciiger,'' Zool., vii, pt. 19, pp. 39, 44, pi. i, fig. 3, 1883. 

 Halobates alluaudi Bergrotli, Rev. (TEnt., xii, p. 204, 1893 ; Kirkaldy, Ann. Soc. Ent. France, 



Ixviii, p. 103, 1899 ; Bsaki, Ann. Mus. Nat. Hung., xxiii, p. 133, 1926 (Syn. nov.). 

 Halobates sericeus Matsumura {nec Bschsclioltz), Thous. Ins. Japan, Addit., i, p. 97, pi. xi, fig. 8, 



1913. 



Halobates niatsuiimrai Bsaki, PsycJie, xxxi, p. 117, pi. v, fig. D, 1924 {Syn. nov.). 

 Upolu : Apia, 1 c^, 1 2.iii.l924, on surface of sea. 



Bergroth, who described Halobates alluaudi from material from the 

 Seychelles, wrote (lac. cit., p. 205) : H. pri?icipi B. White afiinis, sed tarsis 

 omnibus, praesertim posticis, multo brevioribus et trochanteribus mediis iner- 

 mibus bene distinctus." Recently I have examined the type of H. princeps 

 B. White (a single female, in the British Museum), and have found an important 

 particular in which B. White's statement and fig-ure are incorrect. He writes 

 (lac. cit., p. 45) : " [the hind] tarsus about five-eighths of the length of tibia," 

 but, in reality, the measurements afforded by the unique type are : — hind tibia : 

 tarsus = 43 : 13, i.e. roughly 3:1. This agrees quite well with the statement 

 of BergToth {loc. cit., p. 205), in the description of H. alluaudi : " [tibia postica] 

 tarso saltem triplo longiore," which is the only difference between H. princeps 



