1907.] Whedcr, Fungus-growing Ants of North America. 773 



of the bodies is undoubtedly due to this method of feecHng. It is equally 

 certain that these bodies keep growing in size and regenerating the consumed 

 portions by a rapid proliferation of their component cells. Caterpillar 

 excrement freshly introduced into the nest was "seeded" by the workers 

 either with entire bromatia brought from older portions of the garden or 

 with small pieces bitten off from the bromatia and sprinkled over the new 

 substratum. In the artificial nests the ants were unable to raise sufficient 

 fungi for their consumption, so that in the course of a few weeks they de- 

 voured all of the bromatia and eventually died of starvation. As a rule the 

 substratum employed by C. comalensis and the other forms of rimosus, that 

 have come under my observation, consists exclusively of caterpillar droppings, 

 but in several of the nests of the subspecies minutus in the island of Culebra, 

 I also found small pieces of plant substances which I was unable to identify 

 and a few small decomposing insect larvae. These Avere mingled with the 

 caterpillar excrement and also dotted with flourishing bromatia. 



On one of my artificial nests of comalensis I made an observation which 

 proves that this ant can also eat animal food. Several of the larvse and pupse 

 that had been injured while the colony was being captured were eaten with 

 avidity not only by the workers but also by the males and winged females. 

 They did not, however, eat other insects, such as flies and small beetles, 

 which I placed in their nest. The remains of the larvae and pupse were 

 eventually inserted among the caterpillar excrement and carefully seeded 

 with pieces of bromatia. This would seem to indicate that the beetle frag- 

 ments seen in the nests of C. icheeleri at Fort Davis may have been similarly 

 employed as a portion of the substratum. 



Both in the natural and artificial nests of C. comalensis and minutus the 

 brood was carefully kept to one side of the damp fungus garden, which 

 would certainly be a very unwholesome and inappropriate nursery com- 

 pared with the flocculent gardens of other Attii. The larvae of comalensis 

 were fed by the workers with small pieces of the bromatia. I have seen a 

 few virgin females in the nests of this variety as early as May 10, but these 

 and the males were not found in numbers till June 10 to 21. In the more 

 southern countries, such as Culebra and Porto Rico, the winged phases 

 appear as early as March and April. They "feign death" like the workers, 

 but the males less readily than the females. 



9. Atta (Mycocepunis) smithi Forel. 



This species, originally described from the island of St. Vincent, seems 

 to be widely distributed through the West Indies and Mexico, but I have 

 seen it only in Porto Rico, where it is rej)resented by the variety borinquencn- 



