E T H I ( ) I ' I A N X AS U T 1 T E K M I T I X A E 



129 



Minor Soldiers : 



Head length to tip of nose 

 Head width . 

 Depth of head capsule 

 Width of head capsule 

 Length of pronotum 

 Length of hind tibia 



Range Mean 



173-204 .... 1 -go 



0-78-0-94 .... o-86 



0-58-0-72 .... 065 



Q-43-0-55 .... 050 



0-20-0-29 .... 024 



1 39-1 -68 . . . . 154 



I 'ariation. 1 n the imago, there are sometimes small diverging depressions on the frons. The 

 large variation in eye size results in specimens of very different general appearance, and there is 

 some indication of a cline in this character from moister to drier areas, the largest eyes occurring 

 in the moister southern vegetation types. 



The " major " and " minor " soldier castes are not always distinctly separable. It appears 

 to be characteristic of the largest colonies that a complete gradation of intermediate forms 

 develops. 



The imago of T. trinervius closely resembles that of T. geminatus, but is usually 

 darker in colour, with larger eyes, a slightly more inflated postclypeus and longer 

 $ cerci ; there is however an overlap in all these characters, and since the two 

 species are sympatric, they constitute a further example of near-siblings among 

 termites. The soldiers are usually distinguishable by the more hairy nose tip and 

 darker coloured abdominal tergites of T. geminatus. T. oeconomus is also closely 

 similar in the imago, but is distinguished by the vertex pilosity. 



Among allopatric species, only the darker soldiers of T. gratiosus closely resemble 

 T. trinervius, and are in some cases virtually indistinguishable, though usually 

 much shorter-nosed. 



The close similarity of T. trinervius to T. geminatus and T. oeconomus has in the 

 past led to confusion which has arisen mainly because the unique holotype <$ imago 

 lacks the head. Rambur (1842) omitted mentioning the head or its absence, while 

 including it in descriptions of other species ; it can thus be presumed that the 

 holotype was headless when he described it. Hagen (1858) noted the deficiency. 



Walker (1853) described as T. trinervius a specimen of T. oeconomus labelled 

 ' Tripoli ", but this must have come from tropical Africa, probably the Sudan. 

 The reference of Sjostedt (1900) is included in the synonymy because of the change 

 of genus involved ; the specimens described were probably again T. oeconomus, 

 though some from South Africa were included. The specimens described by Silvestri 

 have not been examined, but would appear to have been T. geminatus ; some of 

 the material collected by Silvestri in Senegal was however T. trinervius. 



I have found no reliable specific characters in the pronotum shape, colour, or 

 pilosity, nor are the outlines of meso- and metanota of any value in Trinervitermes. 

 The legs are not distinctive in closely related species, and the abdominal characters 

 occur in the $. I therefore think it unlikely that the use of Rambur's name will 

 ever be finally settled by examination of the holotype. The general appearance, 

 the colouration, and the shape of the $ cerci of this specimen, though not conclusive, 

 appear to me to be most consistent with the species hitherto known as T. posselensis 

 of which T. carbonarius is a junior synonym. In spite of usage by some authors 

 subsequent to Rambur, I consider it inadvisable to use T. trinervius as senior 

 synonym to T. oeconomus, since that species is well known and established under its 

 present name, with adequate type material and several junior synonyms. 



