784 Bulletin American JShiseum of Natural History. [Vol. XXIIl, 



'spheres.' These consist of branching hypha? bearing either spherical or 

 oval cells. The spherical cells do not germinate. The oval cells germinate 

 readily, but it has not been possible to reproduce the 'spheres' from them. 

 When the comb is old an agaric grows from it. This agaric appears in two 

 fornis, one of which has been assigned by various mycologists to Lentinus, 

 CoUyhia, Phiteus, Pholiofa and Flammida, and the other to Armillaria. It 

 develops in a cartilaginous, almost gelatinous, universal veil and is a modified 

 Volvaria. Sclerenchymatous cells occur at the base of the agaric stalk and 

 in aborted agarics. It has not been possible to germinate the spores of the 

 agaric or to grow the sphere-producing mycelium from its tissues. When 

 the comb is enclosed in a bell jar, Xylaria stromata are produced. Sole- 

 rotia may also be formed: the same stromata grow from these. This 

 Xylaria is probably A", nigripes. The shape of the stroma and conidio- 

 phore depend on the age of, and amount of moisture in, the comb. When 

 sown on agar the spores of these reproduce the Xylaria stromata. These 

 stromata occur most abundantly in combs which have produced an agaric. 

 After continued rain Xylaria nigripes grows from deserted termite nests. 

 Other fungi which grow on combs removed from the nest include Mucor, 

 Thamnidium, Cephalosporium, Peziza. As these are not found in the nest 

 though some of them are capable of development under ground, it is prob- 

 able that the termites 'weed out' foreign fungi from the cultivation of the 

 comb. The comb material is probably sterilized by its passage through the 

 alimentary canal. That the 'spheres' form the food of the termites is 

 probable, as in the case of the leaf-cutting ants: neither case can be con- 

 sidered definitely proved. Termes redemanni and T. obscuriceps undoubt- 

 edly prefer fungi, or wood which has been attacked by fungi. Whether a 

 difference in food causes the differentiation of termites into workers, 

 soldiers, and sexed insects, is not decided. A Ceylon agaric, Entolovia 

 microcarpmn, possesses a mycelivun composed of spheres of swollen cells: 

 the details of these spheres resemble the parts of the termite spheres, but 

 are not so highly developed. It is most probable that the 'spheres' in the 

 termite comb and the 'Kohlrabihaufchen' of the leaf-cutting ants investi- 

 gated by Moeller are parts of a normal mycelium, and that their shape is 

 modified by the insects only in a very slight degree, if at all. The available 

 evidence appears to show that the 'spheres' are part of the mycelium of the 

 Volvaria, but it has not been possible to connect these forms experimentally." 

 A review covering some other features of Fetch's work has just been published 

 by Harris in the American Naturalist (1907). 



The foregoing accounts from several observers show that the fungus- 

 growing termites differ from the Attiine ants in several important particulars. 

 In the first place the termites use their own excrement as a substratum, 



