OBSERVATIONS ON SOME SOUTH AFRICAN TERMITES. 863 



ave not cemented together in any way. All were entirely 

 composed of subsoil, and presumably represent the soil removed 

 by the insects in making large hive cavities and other low- 

 level excavations. But very few galleries traverse the mounds, 

 and such as do are purely temporary, being made only for the 

 purpose of throwing out the excavated earth. This is discarded 

 in a very similar manner to that adopted by Hodotermes, 

 but there are no hardened tubes made through the centre of 

 the earth piles. The pellets are brought up and pushed out 

 loosely without the insects exposing themselves. In this way 

 many perfectly conical heaps of soil are formed, and those 

 seen were from 6 to 10 in. in height and 12 to 18 in. in dia- 

 meter. They were found anywhere upon the surface of the 

 larger weathered mounds and also round about them, some 

 being even a little outside the fringe of the nest-site. 



The mound examined was oval in outline, 5 to 6 ft. long, 

 4 ft. across, and 24 to 30 in. in height (PI. XXVII, fig. 1). It 

 was so recently made that no roots had as yet penetrated it 

 (fig. 2). Immediately under the centre of this mound was 

 found a large sub-globular cavity with a flattened floor and a 

 somewhat conical dome (fig. o). A layer of undisturbed soil 

 of some 18 to 20 in. in thickness overlaid the apex of the 

 dome. The height from the floor of the cavity to the roof 

 of the. dome was between 24 and 30 in. and the diameter 

 at the floor 30 to 36 in. The walls of the cavity were 

 smoothed, but not plastered over Avith a cemented clay 

 lining, as is the case with most species, and in the somewhat 

 similar cavities made by T. badius, T. vulgaris, T. late- 

 ricius, and T. incertus. Access to this cavity was gained 

 by the termites through a series of more or less triangular 

 and fairly large openings to devious galleries leading away 

 from it. 



The cavity was completely filled by a large sponge-like 

 formation of claj^-laminte or shelving, remarkable for the 

 fact that at no point was it attached to the Avails of the 

 cavity as in the case of T. natalensis and T. latericius, 

 and also for its simulation of certain fungus-gardens, such as 



VOL. 3, PART 2. 25 



