ANTENN.E OF DIPTERA. 327 



Cecidomyidse, do I believe that we shall often, if ever, find ves- 

 tigial joints additional to the maximum — simply for the reason 

 that the loss of additional joints has been so far back in geologi- 

 cal history that vestiges have wholly disappeared. 



If the law of irreversibility in evolution be true, then it is ap- 

 parent that no fly has regained the use of a joint of the antennae 

 once functionally lost. The question at once becomes important : 

 What was the original number of antennal joints in the Diptera ? 

 If we could only be assured of the origin of the order from the 

 main insect stem, we might, perhaps, answer this question with 

 satisfaction. But, since we cannot we are forced to depend upon 

 the internal evidence presented by the diptera themselves. May 

 we assume that this primitive number was sixteen, the number so 

 conspicuous in the table ? Or was it thirty-nine (or more) a 

 number known in a single species of diptera? I have assumed 

 that the evolution of the dipterous antenna has been by the re- 

 duction of the number of segments, and never by accretion. 

 And I believe that this assumption is justified, though the mat- 

 ter is perhaps open to debate. There are very few forms of dip- 

 tera known possessing more than sixteen antennal joints. Some 

 species of Pachyrhina and of the nearly allied Nephrotoma among 

 the Tipulidae have nineteen joints in the male, fifteen or sixteen in 

 the female. The genus Ctcdonia, of the same family, from Chile, 

 has twenty-two or twenty-four joints in the flagellum of the 

 females of two species, fifteen in that of a third. . . . The very 

 closely related Cerozodia, from Australia and New Zealand, with 

 two species, known only in the males, has thirty-two and thirty- 

 seven flagellar joints respectively, the largest number hitherto 

 discovered in any dipteron. As Osten Sacken truly said : "The 

 close affinity between Cerozodia and Ctedonia affords a new in- 

 stance of the curious relationships between the Australian and 

 New Zealand fauna and that of Ghile [South America] ; a rela- 

 tionship exemplified in abnormal forms, apparent survivals of 

 past ages, of which we already have" many other equally re- 

 markable instances in all branches of animal life, recent and fos- 

 sil. And precisely similar is the relationship between Tanyderus 

 pictus, of the same family, from Chile, with twenty-five antennal 

 joints, and T. ornatissimus from Amboina, with twenty-two joints. 



