INSECTS AND FUNGI 231 



ground, and it may then have from six to twelve chimneys ; 

 but as the nest is extended and the hill is built, the number 

 of chimneys diminishes until (as a rule) the largest nests have 

 only one. It seems probable that the hill was originally only 

 a means of getting rid of the earth excavated in making sub- 

 terranean chambers. This would then agree with the similar 

 habit of many true ants, though these do not cement the particles 

 together. At present, the chimneys serve as a permanent 

 scaffold to direct the further additions to the hill, and as a 

 means of exit for the winged (i.e. male and female) insects. 

 The workers and soldiers do not leave the nest via the chimneys. 

 A series of observations has shown that the internal temperature 

 of the hill varies only by a few degrees. The hill is, in fact, a 

 natural hot-house, not a structure scientifically ventilated by 

 means of the chimneys. 



In the hill, and in the ground beneath it, are excavated 

 numerous chambers connected with one another and with the 

 chimney shafts by minute galleries, which are only large enough 

 to admit the passage of two or three termites at once ; and each 

 chamber (except the small one inhabited by the king and queen) 

 is almost filled by a brown structure which resembles in form a 

 coarse bath sponge. This structure, which may be called the 

 comb, serves the insects at once as a nursery and a living room. 

 The eggs are stored in the lower portion, while the larvae, 

 workers, and soldiers crowd the upper galleries. The comb 

 consists entirely of wood which has been digested by the 

 termites. It is built up of small balls of excreta cemented 

 together in the same way as the particles of earth which form 

 the hill. 



The Fungus Garden 



It is at once seen that the comb is hoary with a covering 

 of short fungus hyphae, and that scattered everywhere over 

 its surface are minute white spheres, each about the size of 

 a pinhead. These spheres are more abundant on the roof 

 and sides of a passage than on the floor. Here we have, as has 

 been always recognised, the termite food which corresponds 

 to the kohlrabi heads of Moller's leaf-cutting ants. This has 

 been established by feeding experiments and dissections, 

 the latter being, as Holtermann pointed out, the only reliable 

 method of determining this point, either in the case of termites 



