INSECTS AND FUNGI 237 



continued into all the surrounding spheres. There is a fairly 

 close resemblance between the termite sphere and the sphere 

 of Entoloma microcarpum. Both show two kinds of cells, the 

 spherical ones being confined to the exterior; and there is 

 the same kind of branching in the oval cells of both. When 

 the termite sphere is placed in a culture solution the parallel 

 becomes closer : radiating hyphae with swollen tips grow out 

 from it, and on teasing out the sphere it is found that these 

 proceed only from the ends of the internal chains of oval cells — 

 the spherical cells do not produce hyphae. The same holds 

 good for the Entoloma sphere : the hyphae which grow out 

 to the next sphere arise from the internal chains of cells — the 

 spherical cells do not germinate. A fringe of similar hyphae 

 grows from the termite sphere on the comb before it collapses, 

 and it is conceivable that, if the requisite conditions could be 

 realised, a mycelium similar to that of Entoloma microcarpum 

 would be produced. The intervention of the ants need only 

 be demanded to prevent this formation of adjacent spheres, and 

 if the first spheres are eaten as soon as they are full grown, 

 this condition is secured. 



Application to Accepted Theories 



But the mycelium of Entoloma microcarpum carries us 

 further than this. So long as the only recognised type of 

 agaric mycelium consisted of fine hyphae sometimes united 

 into strands, it was perhaps reasonable to regard the kohlrabi 

 heads of the leaf-cutting ants as special structures evolved under 

 cultivation by the insects. Kuster regards them as galls, and 

 Schimper states that they represent a new structure which has 

 arisen as a result of artificial selection exercised by the ants. 

 " These peculiar structures are a product of cultivation com- 

 parable with kohlrabi. The phylogenetic starting-point of 

 their evolution is to be found in the tendency of the fungus 

 to produce all kinds of swellings." It is not explained how 

 this evolution operates, nor why the fungus should acquire 

 a character which is being continually eradicated. 



But the discovery of a normal mycelium consisting of a mass 

 of small spheres built up of swollen cells affords another 

 explanation. It seems most probable that Moller's kohlrabi 

 heads are part of the normal mycelium of an agaric, and that 

 their form is due to cultivation by the ants only in a small 



