OBSERVATIONS ON SOME SOUTH AFRICAN TERMITKS. 361 



material mixed with sand, such as E. parvus has, in excep- 

 tional cases, been found building. As a rule only one gallery 

 is built up the main stem, and a nearly perpendicular pathway 

 is made from the level of the soil. If a branch is made in the 

 gallery upon the trunk of the tree, it is only in the form of a 

 loop. From these galleries major and minor workers, a few 

 minor soldiers, and a rare major soldier may be recovered. 



Termes waterbergi sy. n. PI. XXVII, figs. 1-3. 



My attention was fii-st drawn to this insect by Mr. C. A. 

 Simmonds, of the Division of Horticulture, who sent to me 

 material from a mound found below some thorn trees at 

 Warrabaths, in the Waterberg district, during October, 1911. 



Mr. David Gunn, of the Division of Entomology, subse- 

 quently obtained for me a large series of major soldiers from 

 a single gallery which he found in this mound. Time did 

 not, however, permit him to make any particular examination 

 of the nest. 



More recently (March, 1915) I have been fortunate in 

 finding this species at De Wilt on the northern side of the 

 Magalies range. The country here is what is generally called 

 middle veld. It is hot and somewhat dry and covered by 

 various acacias and a large variety of other bush-veld trees. 

 Throughout, at very short intervals, which are studded mostly 

 with acacias, are more or less circular clumps of dense set 

 trees. The trees in these clumps are of a number of kinds, 

 and all are more umbrageous than those between the clumps. 

 In a fair number of clumps stands a solitary and large 

 Euphorbia, that characteristic feature of the park-formations 

 of Zululand. Without exception the soil-level beneath the 

 tree groups is higher by 2 to 4 ft. than that around them, and 

 has obviously been raised around the stems of all the older 

 plants. The better part of three days was spent in examining 

 this particular aspect of park-formation, and, except in the 

 case of some tree groups inhabited by T. waterbergi, there 

 was ample evidence to show that all were the homes of 



