ANTENN/E. 



29 



of four closely united segments. By the comparison of a 

 horsefly's antennae with that of a Bibio and that of a mos- 

 quito it will be easily understood that the complex third 

 joint is merely the closer fusion of the real joints of the 

 flagellum, though I know of no instance of a complex 

 joint having more than eight segments. Nor is the horse- 

 fly's so-called third antennal joint homologous with that 

 of the housefly, but rather with the joint and the 'arista' 

 combined. Either the complex or the simple third joint 

 may terminate in a more or less slender, and more or less 

 freely differentiated portion called the style, or in a bris- 

 tle-like, elongated portion, called an arista. It is very 

 evident, however, that the style or arista represents 

 merely attenuated distal joints of the flagellum, since a 

 close examination of them will invariably, or almost in- 

 variably, disclose from two to five segments (fig. 5, 7), 

 though some may be very small or almost completely 

 atrophied. The arista therefore as might be supposed, 

 is not sharply distinguishable from a slender style; in- 

 deed the arista always, or almost always, has its basal 

 portion thickened more or less, in some cases so much so 

 as to form a real though short style provided with a long 

 bristly extremity; and the style is often provided with a 

 short bristly end. An examination of the antennae of the 

 Leptidae and Stratiomyidae, figured further on, will make 

 these statements clear. The arista or style is of course 

 not at all apparent in the Nemocera, since the antennal 

 structure is here generalized, and it maybe entirely want- 

 ing among the Brachycera, either because the distal 

 flagellar joints have not become at all differentiated, as 

 may be observed in Xylophagus of the L,eptidae, or because 

 there has been an actual atrophy of the distal part of the 

 antennae; but the cases are rare where some of the flag- 

 ellar joints beyond the basal one are not observable in 

 flies, albeit very rudimentary. Whenever the style or 



