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



have to be made in order to account for the dehcate and intricate adaptations 

 shown by existing Attii in the cultivation of their fungi, that further specu- 

 lation seems idle till we are in possession of a greater body of careful observa- 

 vations. 



Less hypothetical and worthier of confidence are the views of Forel and 

 von Ihering concerning phylogenetic development within the narrow con- 

 fines of the Attiine tribe itself. But here, too, we must proceed with caution. 

 The ants of the genera and subgenera Cyphomijrmex, Myrmicocrijpta, 

 Sericomi/rmex, Apterosfigma, Mycocepurus and Mycetosoritis on the one 

 hand, are obviously primitive, for they form small colonies and have mono- 

 morphic workers and proportionally small males and females. On the 

 other hand, Atta s. str. would seem to be the most recent and highly special- 

 ized genus of the tribe, because the colonies are very populous, the workers 

 are polymorphic Avith marked division of labor, and the males and females 

 are very large. Between these two groups, Trachymyrmex, Acromyrmex 

 and Moillerius occupy an intermediate position. Moeller and subsequent 

 writers have been inclined to find a parallel development in the instincts, 

 but this is not so clear as the morphological sequence and relations of the 

 various genera and subgenera, for we find Atta s. str. and Acromyrmex 

 building gardens on the floors of their chambers like Cyphoviyrmcx, whereas 

 Apterostigma has highly specialized gardens, suspended and enveloped in a 

 mycelial web not known to occur in any other Attii. Moreover, at least one 

 species of Cyphomyrmex (;rimosus) and a species of Atta s. lat. {Mycocepurus 

 smithi) cultivate a very different fungus from that knoAvn to occur in the nests 

 of any other species; C. wheeleri does not, at least as a rule, use caterpillar 

 excrement as a substratum but only small plant slivers ; Mycetosoritis special- 

 izes to the extent of using only the anthers of flowers, and Sericomyrviex 

 opacus has a predilection for fruit pulp. All of these species are therefore 

 aberrant in their habits, though belonging to primitiA^e genera. Moeller 

 has certainly OA^erestimated the primitiA^e nature of the treatment bestowed 

 on the fungi in the nests of Cyphomyrmex as a group, and although the 

 bromatia of the Apterostigma gardens may be of a generalized type, this 

 genus is in many other respects more highly specialized than Atia s. str. 



Granting the cogency of these considerations, it still remains true that 

 the Attii in general present a series of increasingly specialized forms as we 

 pass from the species of Cyphomyrmex through the subgenera Mycetosoritis, 

 Trachymyrmex, Acromyrmex and MceUerius to Atta s. str. in which we see 

 the culmination of a Avonderful progress in adaptation. These insects in 

 the fierce struggle for existence, CA^erywhere apparent in the tropics, haA^e de- 

 veloped a complex of instinctiA-e activities which enables them to draw upon an 

 ever-present, inexhaustible food-supply through utilizing the foliage of plants 



