138 THE FACTORS [Part i' 



days, in the appearance of a very rich siibatrial mycelium, on which two 

 kinds of conidia are produced. This luxuriant growth occasions not 

 only the rapid exhaustion of the substratum, but also the emptying of 

 the ' kohlrabi-clumps ' or the cessation of their production. 



Moller was able, in cultures in nutritive solutions, to induce the fungus 

 to form ' kohlrabi-clumps,' which were identical with those in the fungus- 

 gardens and were eaten just as greedily by the ants. These peculiar 

 structures are therefore by no means ant-galls, but a product of cultivation 

 comparable with kohlrabi. The phylogenetic starting-point of their evolu- 

 tion is to be sought in the tendency of the fungus to produce all kindsl 

 of swellings. 



The parasol-ants are not the only species that cultivate fungi. Frau 

 Brockes, one of the daughters of Fritz Mtiller, discovered the same custom 

 in another genus of ant, Apterostigma, near Bkimenau, and Alf Molleri 

 has minutely investigated the fungus-gardens of the remarkably hairyl 

 little species of this genus, which are therefore termed Jiairy ants. They 

 belong to four different species: Apterostigma Molleri, Forel, A. pilosum, 

 Mayr, A. Wasmanni, Forel, and another species not yet described and 

 here referred to as A. IV. They live in much smaller communities than 

 do the species of Atta, and construct correspondingly smaller gardens 

 for which purpose they chiefly employ wood-dust, produced by the 

 activity of insect-larvae, and the excrements of the latter. 



Finally, Moller also recognized the hump-backed ants — species of tht 

 genus Cyphomyrmex (C. auritus, Mayr, and C. strigatus, Mayr) — a; 

 fungus-cultivators. Their fungus-gardens resemble those of the hair) 

 ants. 



The fungi of the gardens of the hairy and the hump-backed ants diffe 

 specifically from one another as well as from those of the species 

 Atta, but the different species of each genus of ant cultivate the same 

 species of fungus. The fungus of Apterostigma and of Cyphomyrmex, liki 

 that of Atta, produces kohlrabi, but of a somewhat different structure, and! 

 after the ants are removed, also produces a luxuriant subaeriai myceliun 

 from which conidia are abstricted. Unfortunately the highest form 

 sporophore has not been observed, so that the systematic position 

 the fungi associated with the hairy and hump-backed ants is not ye 

 precisely determined, but they undoubtedly belong to the Basidiomy 

 cetes and probably to the Agaricinae. The 'kohlrabi-clumps' of th. 

 different species of ants are highly instructive, as they represent structure 

 that have remained at different stages of selective evolution. Tiie specie 

 of Atta have elaborated the most highly developed product (Fig. 74' 

 Somewhat less developed are the 'kohlrabi-clumps' of Apterostigm: 

 Wasnp.anni, as the apices of the individual kohlrabi-hyphae assume no 

 a globular, but a swollen clublike form, and are arranged in less dcfinit 



