THE ANTENNAE OF DIPTERA ; A STUDY IN 

 PHYLOGENY. 



S. W. WILLI STON, 

 Chicago, III. 



No classification of the order Diptera is, on the whole, satis- 

 factory. Four or five more or less elaborate schemes have been 

 proposed by various writers in the past, but none has received 

 general endorsement by students of the order. Scarcely any two 

 writers agree as to the relationships of the larger part of the 

 families, nor as to the value of many of their distinguishing 

 characters. And this unsatisfactory condition is doubtless 

 largely due to our failure to differentiate between homoplastic or 

 convergent resemblances and genetic characters. As in all other 

 large groups of animals there have been many phyletic lines of 

 descent, many parallel adaptations to like environments, and these 

 adaptive characters, here as elsewhere, have been, too often, used 

 as fundamental classificational characters. Of past writers Osten 

 Sacken was, I believe, most appreciative of such accidental re- 

 semblances ; but even he often mistook adaptive for hereditary 

 characters I am convinced. 



The attainment of like characters by evolution by no means 

 necessarily implies common ancestry. One would not think of 

 uniting all flies having two-jointed palpi in one group, nor all 

 those having a club-shaped abdomen in one suborder. But there 

 are many other characters, less conspicuous ones, which have 

 been used for such purposes, whose origins have been due to 

 adaptations ; and it will be long before we have thoroughly 

 learned to distinguish them. New characters acquired in different 

 phyla are seldom, perhaps never, exactly alike, though there may 

 often exist the most curious resemblances, due doubtless to 

 similar determining causes, or to orthogenesis, if there be such a 

 thing. To paleontology we are indebted for the formulation, at 

 least, of the apparent law that evolution is irreversible — that 

 organs once functionally lost are never regained. There may be 

 exceptions to this rule, but, so far, the history of past life seems 



324 



