( 559 ) 



centre, by which meaus the proboscis-groove is formed. The bristles placed on this 

 central area, presnmably as a j)rotection for the proboscis, vary much according to 

 the species, and the central portion of the hind margin exhibits also considerable 

 differences in outline in the various genera. The sides of the under surface are 

 deeply hollowed out, the anterior and lateral margins of the cavity being armed 

 with a comb of ilat, obtuse, spines. The comb varies especially in the development 

 of the lateral spines. In some species these spines are small and do not project 

 beyond the edge of the head (PI. XIII. fig. 5. 0. 8), in other forms they are much 

 longer than the anterior spines and are visilile from above (PI. XII. fig. 1. 2. 3. 4). 

 The spines on each side of the proboscis-groove are usually much nearer together 

 at their tips than at their bases, projecting considerably over the groove. In 

 order to allow the proboscis to ])ass through the gap between these spines the 

 second segment of the proboscis is narrowed, which is especially noticeable in fig. 2 

 and 4 of PI. XII. The comb is present in all the specimens, adult and immature, 

 which we have seen, with the exception of one very young individual o^ Hesperoctenes, 

 in which a similarly arched conil) is situated on the first segment of the antenna 

 instead. 



The antennae are always four-segmented. Their length and armature and 

 especially also the shape of the first segment afford characters of considerable 

 taxonomic value. The variability within the limits of a species is slight as regards 

 mature specimens. In the drawings hitherto published the insertion of the antenna 

 in the head has not been clearly indicated. When moving the antenna of specimens 

 preserved in alcohol, one distinctly notices that it is joined to the head with its 

 posterior corner marked A in our text-fig. 1 (p. 557). The first segment can easily be 

 moved horizontally to a limited extent, but the jnvot on which it works allows 

 also some downward movement of the anterior margin {i.e. a slight turning of the 

 antenna) and of the apical margin {i.e. a downward flapping of the antenna). The 

 mature Old-World species have the first segment provided with a median or a 

 subaj)ical row of blunt spines and an anterior row of smaller ones. The adult 

 American specimens bear pointed, and more or less slender, bristles instead. Im- 

 mature specimens of the Eastern Hemisphere may have bristles only (PI. XII. tig. 4) 

 or spines as well (text-tig. 1), while the only young American specimen which we 

 have seen has a regular curved comb of spines on this antenual segment, replacing 

 the gular comb of the head. The second segment is always longer than either of 

 the following two, and is more or less cylindrical in those species which carry the 

 antenna on the upper side of the i)rothorax, probably when moving through the fur. 

 In the American species and one of the Old-World forms (PI. XIV. tig. 10. 11) the 

 inner surface of the segment is longitudinally impressed, and there is a comb of spines 

 at the npper edge of this groove and some long bristles near the lower. In these 

 species the antenna is laid along the side of the pronotum, whose edge fits in 

 between the comb and bristles. The apical segment is alwaj's jjalc at the tip and 

 bears here some pale hairs, which presumably have a sensory function. 



The ujiper surface of the head, thorax and elytra is always flattened in the 

 centre, never markedly convex, excepting the sides, and the thorax and elytra often 

 have a distinctly impressed mesial line. The sides of the prothorax and elytra, 

 however, are always convex and flexed downwards. The prothorax, being laterally 

 hollow underneath, varies in outline to a certain extent according as the margin 

 is more or less curved downward, which often depends on the state of preserva- 

 tion of the individual. One of the specimens of nycteridis, for instance, has an 



