794 Bulletin American Museum of Natural History. [Vol. XXIII, 



3. The Phylogeny of the Attii and of the Fungus-groivine/ Habit. 



The Attii belong to a complex of Myrmicine genera once grouped to- 

 gether as Cryptocerides on account of their superficial resemblance to the 

 ants of the genus Cryptocerus. Forel in 1S92 was the first to split up this 

 artificial group. He divided the genera into four tribes, the first including 

 the Attini, the second the Dacetonini, again divisible into three subgroups: 



a, Strumigenys, Oredognathus, Epitritus and possibly Hypopomyrmex; 



b, Daceton and Acanthognathus, and c, Rhopalothrix, Ceratobasis and 

 Cataulacus. To a third tribe he assigned Meranoplus and Calyptomyrmex, 

 which were recognized as having affinities with the Tetramorii, and to a 

 fourth tribe he assigned Cryptocerus and Procryptocerus. In 1893 he said: 

 "Taxonomy has proved to me that the Attini are intimately related to the 

 Dacetonini (Strumigenys, etc.) and has led me to suppose that the Attini 

 are of secondary derivation. This is all the more probable, because they are 

 confined to the American continent, whereas the Dacetonini are distributed 

 over the whole world, even to New Zealand." 



Emery, writing in the same year (1893), expresses himself somewhat 

 more explicitly. "If we separate from the ensemble of the ancient Crypto- 

 cerides, on the one hand Cryptocerus and Procryptocervs (group Cryptocerini), 

 on the other hand Cataulacus (forming by itself a distinct group), and if 

 furthermore, Meranoplus and Calyptomyrmex be attached to Teiramoriuvi 

 and its allies, all that remains of M. Ford's Attini may be divided into two 

 groups according to the venation of the wings. In the genera Atta, Seri- 

 comyrmex, Cyphomyrmex, Glyptomyrmex [Myrmicocrypta], Apterostigma, 

 the radial cell is closed and there is no trace of a discal cell nor of a recurrent 

 nervure, the trunk of the cubital nervure being straight or feebly sinuous. 

 In the genera Rhopalothrix, Strumigenys and Epitritus the radial cell is 

 open; in the female Rhopalothrix petiolata Mayr I find a vestige of a recur- 

 rent nervure, and in the male Strumigenys imitator Mayr the trunk of the 

 cubital vein is strongly arcuate behind at the base, indicating the point of 

 insertion of a recurrent nervure that has disappeared. According to Smith's 

 figures, Daceton, which has a discal cell, belongs to this latter group; 

 probably the same is true of Acanthognathus, Ceratobasis and Oredognathus, 

 whose wings are still unknown. The former of these two groups, which we 

 may call the Attini genuini is exclusively American, whereas the latter, 

 which may bear the name Dacetini, is represented in all the zoological regions 

 except the Ethiopian. These two groups are, however, very closely allied, 

 and the fossil genus Hypopomyrmex, which undoubtedly approaches the 

 ancestors of Strumigenys very closely, has a discal and a closed radial 

 cell. The closed radial cell is an archaic character and is found only in a 



