1907.] Wheeler, Fungus-growmg Ants of North America. lib 



turret about a centimeter high. Under the palms the .sand of the craters 

 was often of a deep red color, unlike that of the surrounding surface, so that 

 the galleries must have been rather deep. Unfortunately my stay in this 

 locality was so brief , that I could not examine the nests at my leisure. Al- 

 though I subsequently collected in many localities on the island, I never 

 again encountered i/. britfoni. Santurce is, however, easily accessible from 

 San Juan, and the future observer Avill have no difficulty in finding the nests 

 and of learning much more concerning the habits of this interesting ant. 



Part IV. The Attii and the Other Fungus-growing Insects. 



Many insects, especially of the orders Coleoptera and Diptera, either in 

 the larval or imaginal stages, are knoAvn to feed on fungi, but the ability to 

 cultivate or to control the growth of these food plants is, so far as known, 

 restricted to certain termites, Scolytid beetles and ants. The taxonomic 

 relationships of these three gi'oups to one another are so remote that we are 

 compelled to regard this control as the result of convergent development. 

 In other words, the fungus-growing habit must have arisen independently 

 on three separate occasions in the phyletic history of the Insecta. In order 

 to secure a broader comparative basis for a discussion of the fungus-grow- 

 ing habits of the Attii it will be necessary to summarize our knowledge of 

 the similar habits in the termites and ambrosia beetles. 



1. The Fungus-growmg Termites. 



Several observers have undoubtedly seen and described the fungus 

 gardens of termites without being aware of the full significance of their 

 observations. As these gardens are perforated sponge-like masses filled 

 with the insects and their brood and lying on the floors of sul:)terranean 

 chambers, they have often been regarded as the true nests of the termites. 

 The earliest author to call attention to these structures seems to have been 

 Konig (1779). After describing the vaulted, smooth-walled earthen cham- 

 bers of Termes Jatalis at Tanjore, he mentions the gardens full of holes and 

 lying on the floors as being "covered with little knots on their outer and 

 inner surfaces, like chagrin skin. This texture is most clearly seen at their 

 margins near the openings and entrances. Under a magnifying glass they 

 appear fibrous or woolly." In the light of our present knowledge it is 

 evident that this fibrous or woolly appearance was caused by the fungus 

 mycelium. 



