OF CERTAIN TERMITE NESTS. 253 



are at first terminal, but subsequent budding beneath makes 

 them lateral ; the new branches do not necessarily produce other 

 spherical cells. Others form a succession of cells and after a 

 short distance grow on again as a hypha (fig. 35 ). In rarer 

 cases branching occurs in the chain of spore-like cells in the 

 same way as in the spheres of the termite comb (fig . 36). Many 

 of the radiating hyphae with swollen ends grow out without 

 producing special cells, and these appear to form the foundation 

 of the next ball. 



When the agaric develops, the spheres turn yellow and 

 collapse. The large spherical cells are then indistinguishable 

 and the ball appears to consist chiefly of hyphae. The agaric 

 is not formed in the interior of a sphere as in Goprinv* stereor- 

 arius, but grows on the top of a cluster. 



Though at first sight dissimilar, a close resemblance can be 

 traced between the parts of the sphere of Entoloma microcar- 

 pum and those of the sphere on the termite comb. In both cases 

 there is an outer covering of comparatively large spherical cells : 

 these are borne in branched chains in the termite sphere but 

 singly in Entoloma; in both cases also the same hypha may 

 produce different cells on a side branch, and in neither do the 

 spherical conidia give rise to hyphae. The oval conidia of the 

 termite sphere are matched as far as their branching is concern- 

 ed by those of fig. 36 from the sphere of Entoloma. A more 

 striking resemblance is seen when the termite sphere is placed 

 in water or when it grows for a short time after the insects 

 have been removed, for then radiating hypha? with swollen 

 tips are produced from the ends of the oval conidial chains 

 just as they are from the chains of cells in the Entoloma sphere. 

 Similar production of a chain of swollen cells which terminate 

 again in a hypha has been shown to occur in the mycelium 

 developed from these oval conidia of the termite sphere on 



agar. 



The Entoloma spheres constitute a perfectly normal myce- 

 lium. Moller's "Kohlrabi" heads have been regarded as 

 special structures evolved under the influence of ants. Kiister 



