1907.] Wheeler, Fungus-growing Ants of North America. 785 



moulding it into the form of a sponge containing numerous habitable cham- 

 bers and galleries. This substance is, of course, much harder and more 

 compact than the comminuted leaves, etc., employed bv the Attii. Second, 

 the fungus grown on this substratum forms bromatia (the spherules or oidial 

 heads) of a very different type from those found in the gardens of the Attii. 

 And third, the termites that are in the habit of growing fungi are not exclu- 

 sively mycetophagous like the Attii, but subsist also and probably very 

 largely on dead wood, twigs and leaves. If it be true as Holtermann and 

 Doflein believe, that the termites are not instrumental in maintaining the 

 purity of the fungus culture, we should have another striking difference, but 

 it is quite conceivable that both in the termites and the ants some effluvise 

 emanating from the myriads of insect bodies may be responsible not only 

 for the suppression of alien fungi but also for the aberrant gro^^i;h of the 

 food-plant. 



I have already called attention to the fact that Holtermann cannot be 

 said to have demonstrated that the Acjaricus rajap is the fruiting form of the 

 fungus which grows in the gardens as a mycelium with oidial spherules. 

 And Doflein's and Fetch's observations are open to similar doubts. Not 

 only is there no satisfactory proof that the termite fungus is a basidiomycete, 

 but the same is true also of Moeller's statement that the South American AttcB 

 cultivate the mycelium of a fungus (Rozifes gongylophora) belonging to the 

 same group. A careful perusal of Moeller's observations shows an important 

 lacuna at this point. That his Attop ate portions of the pileus and stem of the 

 Rozites does not prove that it is the fruiting form belonging to the fungus they 

 habitually cultivate and eat. Nor is ^Mceller on much surer ground when he 

 assumes that the mycelia cultivated by different genera of Attii belong to 

 different species of fungi, for it is very probable that the ants of one species 

 would avoid fungus taken from the nest of another on account of the alien 

 nest-aura. Certainly, to the human olfactories the fungus gardens of Atta 

 texana have a very striking odor which is altogether lacking in the gardens of 

 Trachymyrmex, and it would be strange if these differences did not affect 

 the appetites of such sensitive insects as the ants. In my opinion, it is not 

 improbable that the fungi cultivated both by the termites and ants may be 

 more closely related to the moulds (Ascomycetes) than to the mushrooms 

 (Basidiomycetes). Moeller does in fact, call attention to certain ascomycete 

 peculiarities in the mycelium cultivated by Acromyrmex discigera. This is 

 a matter, however, to be settled by the mycologist, and I merely call attention 

 to it in this connection, because INIoeller's somewhat guarded statements 

 have assumed an unduly positive form in the writings of subsequent reviewers 

 of his work. 



