298 Annals of the South African Museum. 



well be termed chaotic, and it is to be hoped that a paper on its 24 

 species prepared by the writer will contribute to some extent to bring 

 light into this tediously obscure matter. The following table and 

 descriptions of nine regional species ai'e an extract from that more 

 extensive study. 



The original generic definition cannot (as in many other cases) be 

 literally applied, unless unnatural and undesirable sub-divisions be 

 the consequence. The origin of A* at the Cuq (and not proximal), 

 although an important feature on the whole, must be interpreted with 

 some allowance to a more proximinal origin, even for as much as nearly 

 the length of Cuq itself ; variation between nearly allied species and 

 even individuals of the same species makes such an allowance necessary. 

 The same must be said of the most important structural character of 

 the female sex, the styloid processes of the posterior prothoracic 

 border (P. acaciae). Absence of light-coloured postocular spots would 

 exclude the two species furcigerum and caff rum, otherwise clearly 

 congeneric, had we to apply the definition literally. No writer on 

 systematics will fail to observe that giving definitions in a few words 

 and applying them to the letter will very often be satisfactory as long 

 as a limited faunal district is under observation, or a small proportion 

 of existing forms only known ; but with an increasing knowledge, and 

 the extension of the area under discussion, the matter becomes 

 rapidly more difficult, and with a clear insight into natural affinities 

 it may be impossible for the moment to give any short and compre- 

 hensive definitions to do justice to such affinities. Of course minute 

 subdivision of genera Avill apparently overcome such difficulties, but 

 the advantage is much more apparent than real ; it is often better to 

 wait for the right word to be found at some later date for a new 

 definition of an old genus than splitting it up into fragments for the 

 apparent needs of the moment. 



The following table is given for the males only. 



Females may be identified by consulting the single descriptions, and 

 it must not be forgotten that isolated females will sometimes present 

 great difficulties and even be found impossible of identification. 

 The reason is obvious: definition of species is largely based on the 

 structural characters of the males, and even the colour schemes are 

 much more precisely differentiated in this sex. As a rule, a very 

 intimate knowledge of a group a knowledge accessible in many 

 cases only to the naturalist observing the species in nature is needed 

 for ready identification of closely-allied females that lack structural 

 characters. This is true for many of the Agrionidae and obvious for 

 Pseudagrion. 



