458 CLAUDE FULLER. 



of winged adults made in Pretoria, and one nest collection 

 from Alberton, Transvaal. 



With regard to T. natalensis Haviland, it is necessary to 

 point out that the Haviland material ^ contains major (and 

 minor) soldiers of several forms. One series (a) possessing 

 large and thick heads and flattened frons, and of which 

 certain members of a nest-series display a very aberrant 

 meso-thorax ; aiiother (b) having a rather flattened fi-ons, the 

 sides of the meso-thorax rounded and the heads large and 

 thick," and (c) a third series in which the frons is deeply 

 impressed and grooved. The prevalent South African form 

 is T. natalensis Haviland, and it seems to be T. belli- 

 cosus Smeathman ; but, being unable to decide this point, I 

 have grouped, for the present purpose, the whole collection of 

 nest-series studied as T. natalensis Hav. 



I have relied upon the constant features of the mandibles 

 and the antennse of the Haviland material. 



I have come to the conclusion that the aberrant forms 

 which happen to predominate in the portion of Haviland's 

 material before me represent a locality variation. The pre- 

 vailing form, in my collection, comes from the coast, from the 

 thorn country at Kimberley, and from the thorn country of 

 the tablelands of the Transvaal. The several nest-sei-ies from 

 the midlands of Natal illustrate both the typical and aberrant 

 forms of T. natalensis. The latter form is at once recog- 

 nisable by the presence of certain soldiers (major or minor, or 

 both), of which the meso-thorax presents caudo-lateral lobes, 

 and the sides, instead of being rounded, are oblique and 

 divergent. That Haviland regarded this as only an aberra- 

 tion is obvious from the fact that he does not mention it in 

 his description. As to whether natalensis is synonymous 



' The Haviland material in the Natal Museum comprises a number 

 of insects from different nest-series collected and labelled by Haviland 

 himself ; some of those before me were actually measured l^y liim and, 

 no doubt, for the purpose of his original description. 



^ Like the types, as far as one can jiidge from Professor Sjostedt's 

 discussion. 



