392 CLAUDE FULLER. 



fungMis-gardeus of the other species under discussion, and con- 

 sisted of a series of layers 1^ in. thick enwrapping a central 

 core. Fragile as the garden is these layers separate readily. 

 The chamber in which this mass lay conformed roughly to its 

 outline ; the floor was overlaid with a clay mesh-work, bone- 

 like in structure, which formed a raised skeleton platfoi-m. 

 From this platform arose upright columns of clay — most 

 reminiscent of arm and leg bones — which converged and 

 connected indirectly with the roof of the cavity. This 

 peculiar frame-work of clay, both in horizontal and perpen- 

 dicular directions, was the scaffolding or skeleton supporting- 

 the fungus-garden, although the latter was not in any way 

 cemented to it. 



Immediately above the fungus-garden there was a narrow 

 conical dome, hollow except that from it depended numerous 

 thin clay arms ; these being the indirect connections of the 

 thicker vertical columns which penetrate the fungus-garden. 

 Some of these miiior arms had living grass roots at their 

 ■ cores, the roots penetrating the matrix of the mound from its 

 periphery. As many hang rigid, like stalactites, it was 

 conjectured that either the grass root was preserved by the 

 application of a layer of clay, when the dome was excavated 

 and during an enlargement of the nest cavity, or the roots 

 were encouraged, by applications of clay to their growing- 

 points, to descend through the hollow dome into the main 

 upright and so strengthen them ; a conjecture supported by 

 the discovery of roots in some of these latter. From the 

 dome itself several wide shafts (1^ in. in diameter) led 

 upwards into the mound, there broadening out, an inch 

 below its surface, into conspicuous cavities. Apart from these, 

 other shafts led up from the base of the nest around the cavity 

 to the crust of the mound there tei-minating in a similar 

 manner. Means of access to the cavity and other galleries from 

 these shafts were provided by apertures at their bases and 

 along their length, the openings being both large and small. 



The mound of vulgaris appears purely protective, and 

 although repaired in the same manner as in natal en sis it 



