A Monograph of the Formicidae of Sotdh Africa. 483 



9 . Antennae 10- or 11-jointed. Thorax elongate, pronotmn 

 not exposed dorsally. Epinotum with the spines or teeth usually much 

 shorter than in the £ , or reduced to mere angles. Abdomen oblong- 

 ovate, narrowed apically. Wings with 1 cubital and 1 discoidal cell ; 

 the radial cell usually closed. Petiole as in the $ , but the first 

 segment thicker. 



$ . Considerably smaller than the 9 ■ Antennae 12-jointed, 

 the scape short, rarely longer than the 1st and 2nd joints of the 

 flagellum taken together ; usually all the joints of the flagellum except 

 the 1st longer than wide. Pronotum not exposed above ; mesonotum 

 and scutellum high ; Mayrian furrows obsolete or indistinct. Epinotum 

 unarmed. Segments of the petiole shorter and flatter above than in 

 the 9 • Abdomen subcordate or ovate. Legs thin. 



Distribution both hemispheres. 



This very large genus contains very many closely allied forms, and to 

 judge by its tendency to form minute variations, appears still to be in 

 a very plastic condition. Owing perhaps to this condition, which does 

 not appear to have been sufficiently realised by some myrmecologists, 

 and also to the lack of abundant material, the taxonomy of the genus 

 is hopelessly confused, and its study presents us with greater difficulties 

 than that of any other genus represented in our region. Many species 

 have a very extended distribution, and exhibit numerous geographical 

 subspecies and varieties. The variation may apply to one or several 

 anatomical features, in such a way that a species may possess several 

 extreme forms, at least some of which are connected with the central 

 type-form by a series of more or less gradual and transitional links. 

 Where such links are numerous, no doubt is left as to the relationship 

 of the terminal forms to the central type. But on the other hand, such 

 links are frequently missing, and where the gaps appear to be rather 

 wide, the tendency has been to accord specific rank to each form. If 

 at a later period the annectant forms are discovered, the natural 

 result is to render the boundaries between the so-called species 

 exceedingly vague, in which case it becomes very difficult to place the 

 latter in any analytical key unless their status is revised and altered. 

 With the yearly growth of our knowledge of the genus and the increase 

 of material brought about by labours of collectors over the whole con- 

 tinent, the time will come when we shall be able to unite many so-called 

 species as forms of one type, and thereby reduce the number of 

 apparently distinct species which are recognised at the present day. 



Apart from the differences manifested in geographical varieties, 

 minor distinctions may be shown to occur in the members of the 

 different nests of one species, and it is therefore desirable that large 



31 



