394 CLAUDE FULLER. 



plan which, whilst difficult to expose, has very constant 

 features. 



As Haviland states, it is not possible even in the field to 

 define the limits of a colony. In the account given of the 

 behaviour of the winged imagos it is indicated that at Pretoria 

 the earth is thoroughly occupied by the species, and it is no 

 exaggeration to say that to a depth of 4 ft. the soil is pregnant 

 with its fungus-gardens. 



In order to obtain some idea of the main features of the 

 underground occupations of T. incertus numerous galleries 

 were traced through the soil for considerable lengths, and it 

 is upon the result of this examination that the foregoing con- 

 clusions have been arrived at. 



Termes incertus feeds upon dead grass and will often 

 attack plants which have died and may damage newly set out 

 seedlings, especially when planted in virgin soil that has 

 recently been broken up. It is also known to destroy fruit 

 ti-ees and rose bushes. In this work, however, it retains all 

 its subterranean characters and is very rarely found above 

 soil-level, unless in a dead post. 



In grass lands it makes a network of thin tubular galleries 

 through the soil amongst the roots, and secures its food under 

 cover of inconspicuous clay canopies. From this upper net- 

 work of galleries equally tenuous shafts are sunk deep into 

 the soil. The tracing of these is very laborious owing to the 

 difficulty of following so small a gallery, but many interesting- 

 features are presented every inch or so along the length 

 followed. These features are illustrated in a sketch (PI. 

 XXIX, fig. 9) made of a portion of such a gallery whilst 

 being examined. Before summarising these, however, it is 

 necessai'y to say that the long lateral galleries subsequently 

 referred to generall}^ wander through the soil for several feet, 

 and connect with other descending shafts without displaying 

 any such features. 



The descending shaft figured Avent downward at a fairly 

 regular angle of 45°. Here and there, for an inch at most, 

 it would fall almost perpendicularity, but this feature was 



