ANTENNA OF DIPTERA. 329 



these distal joints have been highly specialized. The scape is 

 perhaps never entirely reduced. The genus Chionea, a wingless 

 tipulid, has the conical third antennal joint terminating in a slen- 

 der, three-jointed style, a structure very much like that of the 

 Nemestrinidae, for instance. Do these joints represent the first 

 four of the flagellum ? Osten Sacken thinks that the reduced 

 number of twelve joints in Toxorhina, belonging in a group hav- 

 ing the normal number of sixteen, is due to the coalescence of 

 the basal joints of the flagellum. The stratiomyid genus Chryso- 

 chlora, as one of numerous instances, with the normal number of 

 flagellar joints, has the last or eighth specialized into a slender 

 arista. Is this arista homologous with the arista of the housefly, 

 for instance, where it is serially the fifth or sixth ? The defect 

 of the Comstock-Needham system of venation nomenclature is 

 the assumption that the disappearance of veins has always been 

 due to coalescence, whereas we positively know that in many 

 cases it has been due to their loss without coalescence. Has the 

 reduction of the flagellum been the result of the close fusion of 

 segments, or the absolute loss of proximal ones ; or has the ar- 

 ista been variously and repeatedly produced by the attenuation 

 of the last segment or segments, whichever they happen to be ? 

 I believe that the arista has usually resulted from the former 

 method and that it generally is homologous. It is quite clear, 

 however, that the diminution in the number of homologous or 

 homonymous joints has often been due to the loss of distal seg- 

 ments, whatever may have been the case with heteronymous 

 forms ; and the proof of this is apparent in the oftentimes vestigial 

 condition of the terminal joints. The subject, however, is one 

 worthy of investigation, and may throw light on the relationships 

 of many of the diptera. 



Sixteen antennal joints seem to be the primitive normal num- 

 ber of the modern Nemocera, a number acquired so long ago 

 that very few examples yet remain of the more primitive condi- 

 tion. Do they indicate a single phylum ? It seems doubtful. 

 The forms with multiarticulate antennae not only belong in the 

 three chief subdivisions of the Tipulidae, but are also found among 

 the Cecidomyidae, which would seem to indicate that the number 

 sixteen had been acquired independently in different lines of de- 



