THE POXERIXE ANTS. 245 



subfamilies, notwithstanding the convergent character presented by the 

 pedicel of certain workers (Eciton, ^Enictus and the Myrmicinae). 



" The Myrmicinse have no direct connection either with the Campo- 

 notinse or with the Dolichoderinae. This is evident. On the contrary, 

 the connection of the Myrmicinse and the Ponerinse through groups 

 like M \rmecia, the Cerapachyii, and perhaps Pseudoinynna, admits of 

 no doubt. The structure of the Myrmicine gizzard and poison appa- 

 ratus remains the same. 



" The Dolichoderinse were derived directly from the Ponerinae 

 through" a gradual transformation of the gizzard and a shortening and 

 reduction of the poison apparatus, which has become rudimental and 

 almost replaced by the anal glands. Nevertheless, the fundamental 

 plan of the poison apparatus remains the same as in the Ponerinse. . . . 

 The discovery of the genus Aneuretus Emery (Fig. 140) has enabled us 

 to put our finger on the direct derivation of the Dolichoderinse from the 

 subfamily Ponerinae. Indeed the genus Aneuretus constitutes a true 

 connecting link between these two subfamilies, as is shown by the fact 

 that Emery first placed it among the Ponerinse, but later came to accept 

 my opinion that it should be attributed to the Dolichoderinae. 



" There remain the Camponotinse, of which we have been speaking. 

 The transformation of the gizzard is explained by the intermediate 

 forms in the inferior genera of this subfamily (Myrmoteras, Diinor- 

 t>lioin\nnc.v, etc.). But it is very difficult to understand the complete 

 transformation of their poison apparatus. Here the series is incom- 

 plete. Let us hope, nevertheless, that the future discovery of some 

 extant relict of paleontological times will give us the key to this 

 enigma, as has been done by the genus Aneuretus for the Dolichoderinse." 



