ANTENNAE OF DIPTERA. 325 



to teach that there are not. A fly with two-jointed palpi, for 

 instance, could not have been ancestral to one with four joints in 

 these organs. 



Of all the organs of diptera, the antennae, it seems to me, have 

 received less critical comparative study than any others, and I 

 believe that there is a fertile field here for fruitful phylogenetic 

 studies. The differentiation between the nematocerous and 

 brachycerous flies was, for a long while, based almost exclusively 

 upon the structure of these organs ; until it was conclusively 

 shown that, by themselves, they have little classificational value. 

 The antennae, for instance, of Bibio are so nearly identical with 

 those of Xylophagus that they might be interchanged without 

 affecting generic characters, even as the wings and palpi of the 

 phorids might be interchanged with some of the scatopsines with- 

 out affecting generic characters. It was doubtless because of 

 this primary division long ago by Latreille and Macquart into 

 " many-jointed" and "three-jointed" antennae that a misconcep- 

 tion still exists among many as to the real structure of these 

 organs. There are very few three -jointed antennae among dip- 

 tera, and even these will usually show, under high magnification, 

 vestiges of additional joints. The great majority of existing dip- 

 tera have five or six joints in their antennae, and this majority 

 includes the whole of the Cyclorrhapha, with but few exceptions. 

 The fallacy has been in considering the antennal style or " arista " 

 as an outgrowth or addition to the real antenna, whereas it is of 

 course merely the specialized and more or less attenuated distal 

 (?) part of the flagellum, showing all stages of attenuation and 

 abbreviation ; and it is yet to be shown that the arista is quite 

 homologous in all diptera. 



In the following table I have condensed the results of con- 

 siderable observation and research on the structure of the antennae 

 in the different families of flies. It is of course impossible for 

 one to examine critically all the genera of diptera, and the pub- 

 lished data are yet, in many cases, inexact, and this inexactness 

 is especially apparent when it comes to the detection of vestigial 

 or minute joints. I have found under critical examination not a 

 few instances of minute joints which have been neglected by 

 systematists in general. Such a study as the present one must 



