408 CLAUDE FULLER. 



the tunnel, ai*e always an elongate oval. Their purpose is 

 for the temporary accommodation of the night's harvest. 



A peculiar feature of the galleries is that at every minor 

 branch the connection of the branch with the main gallery is 

 an indirect one, and occurs between two and three inches deep 

 in the soil ; this connection can only be compared to a knot. 

 Galleries which approach a feeding ground suddenly taper 

 and branch in a palmate form, each branch running out to 

 the surface. The surface exits are small round holes of a 

 fairly permanent character which have been noticed to remain 

 open for over a week. There is some evidence that feeding- 

 grounds are more or less permanent, and that continued 

 feeding about one point results in barren or bare patches. 



The accompanying sketch illustrates one of a series of 

 observations upon galleries and nest-connections. It is drawn 

 approximately to scale, and shows, as nearly as possible, the 

 directions taken by the various galleries. Tlie mounds 

 A, B, c, D, F and K were inhabited by active colonies, and in 

 each a queen was recovered. Unfortunately this series was 

 rather near to the homestead, and some of the nests had 

 been removed for chicken food. The sites of these are indi- 

 cated at G and h ; the galleries ran through them, descended 

 below, and rose again on the other side of the site ; from this 

 it is concluded that the nests when standing formed part of 

 the series. It will be obvious that only a few of the galleries 

 have been exposed, but the galleries are not more abundant 

 than indicated in the rough ti'iangle formed by nests b, k, g, 

 which were all exposed. If the evidence revealed is followed 

 to its natural conclusion it would appear that all the nests 

 except E formed a complete series. Nest e appears to belong 

 to a different series. Point is given to this by the obvious 

 avoidance of the galleries of the nest d by that from e. Again 

 the partly exposed gallery x x, obviously leading to the 

 gallery connecting nests c and d, was deliberately under- 

 mined by the second gallery from nest e. 



It is to be noticed in the foregoing account that all the 

 nests explored harboured colonies ; this was fortunate, as it is 



