772 Bulletin American Museum of Natural Hidory. \\o\. XXIII, 



unable to find it in my preparations. The cell wall is always very thin and 

 transparent. These cells closely resemble those of the common yeast {Sac- 

 charomyccs) except that they are considerably larger. Like the yeast cells 

 they may often be found in the act of budding or dividing. In this manner 

 probably arise the minute cells scattered about among those of much larger 

 dimensions. All the cells are held together in the bromatial mass merely 

 by cohesion of their surfaces without assuming polyhedral shapes from 

 mutual pressure, and there is no perceptible intercellular substance nor 

 any trace of an envelope enclosing the mass as a whole. 



Neither the mycologists with whom I am acquainted nor the botanical 

 works to which I have access, have given me any satisfactory information 

 concerning the natural affinities of this singular fungus. That it must be 

 in a purely vegetative stage of growth will probably be admitted, since there 

 is nothing to suggest sporulation in the structure of the bromatia or the cells 

 of which they consist. It is also evident that this plant must represent 

 an entirely different fungus from any of those described by Mceller. Its 

 cultivation on some artificial medium, such as agar mixed with sterilized 

 extract of caterpillar excrement, may be expected to throw light on its affini- 

 ties and to show that it belongs to some well known genus or species, but 

 this can be undertaken only by a trained mycologist. It will be a long time, 

 however, before we are in possession of any information in regard to these 

 matters, if botanists continue to manifest as little interest in the fungi cul- 

 tivated by ants as has been the case during the past fifteen years. In the 

 meantime the singular fungus cultivated by C. comalensis and the other 

 forms of rimosus over such an extensive area of the American tropics cer- 

 tainly deserves a name, and even at the risk of creating a synonym, I propose 

 to call it Tyridioviyces formicanim gen. et sp. nov. and to assign it provision- 

 ally to the order Exoacese, a group which also includes the well-known yeast 

 fungi. 



I have proved that the ants eat the Tyridioviyces, by observing their 

 behavior in artificial nests. On several occasions colonies were brought 

 from New Braunfels to Austin, where they were kept in Petri dishes for 

 periods of from one to four weeks and provided with the excrement of cater- 

 pillars (Hyperchiria io) which feed on the leaves of the southern hackberry 

 (Celtis viississippiensis). The captive ants were as careful of the bromatia 

 as of their brood. When the garden was disturbed they rearranged the 

 pellets of excrement and deftly replaced the scattered and detached fungus 

 bodies. Workers, females and males were frequently seen holding these 

 bodies between their forelegs and eagerly rasping off portions of them with 

 their tongues. Sometimes an ant would consume a whole bromatium, but 

 more frequently only a portion was eaten. The irregular polygonal shape 



