364 CLAUD K FULLER. 



the supplementary gardens of T.badius, latericius, and 

 more particularly vulgaris. When sectioned it Avas found 

 that all the shelves were close together — seldom more than an 

 inch apart — those at the bottom were level, but from one- 

 third of the height the shelving sloped upwards from the 

 sides to the centre of the construction, the inclines becoming 

 more acute towards its apex. 



On one of the lower shelves fragments of dried grass were 

 found. Dispersed about other parts were flattened separate 

 fungus-gardens, none of which were more than four inches 

 across. These fungus-gardens resemble those of T. natal- 

 en sis; their upper surfaces are flattened and the lower 

 coronetted like the surfaces of the gardens of natalensis, 

 but the perforations are arranged more like the cells of a 

 honey-comb. The lower surfaces of nearly every garden had 

 been recently gnawed away by the termites, but the few places 

 that had not been mutilated showed that this surface some- 

 what approached the coronetted surface of the garden of 

 natalensis. 



The hive was inhabited by major and minor soldiers, 

 workers, and by young in many stages. The minor soldiers 

 were not as numerous as majors, and were not at all aggres- 

 sive ; on the other hand, the major soldiers poured out and 

 were aggressive, drawing blood with each bite they managed 

 to inflict. The workers retired; but, if one ceased disturb- 

 ing the nest for a minute, they came forward with pellets of 

 soil in their jaws prepared to repair the breaches. 



Termes natalensis, Havilaud. PI. XXVI, figs. 8-19 ; 

 PI. XXVIII, figs. 1-12. 



Mounds and Nest-Sites. 



The mounds of the natalensis series do not agree with 

 the historic illustration given by Smeathman, nor are they 

 begun and built up after the manner which he describes. In 



