SIMULIIDAE OF AFRICA 5 



investigated as yet for the fauna of Africa, and from work elsewhere on such cyto- 

 taxonomic characters there is no evidence to suggest that they can be of value above 

 the species level in the Simuliidae : similarly the completely phenetic approach of 

 numerical taxonomy has not been attempted in the present work, although the time 

 is approaching in Simuliid taxonomy when the computer might be used with advan- 

 tage. At present there is still much refinement of the supraspecific classification 

 that needs to be done by conventional means, and one of the main objects here has 

 been to consider the various segregates (or from another viewpoint, aggregates) 

 found in the fauna of the Ethiopian and Malagasy Regions and to equate them in 

 rank as species-groups or named genus-groups taxa with similar categories found in 

 other regions. 



The fauna of the Ethiopian Region presents some unusual difficulties in the 

 satisfactory delimitation of segregates within Simnlium Latreille, the preponderant 

 genus in the region, which I am convinced should be maintained in its old broad 

 sense with subdivision below the generic level, because of the occurrence of many 

 species in which one or another stage is aberrant (so that there is a lack of congruence 

 between adult and immature forms). This phenomenon occurs especially in the 

 curious phoretic forms which live as larvae and pupae in obligatory association with 

 crabs or mayflies, for in these the immature stages (presumably in adaptation to the 

 unique ecology) are often atypical in many characters, although the adults are 

 normal, but several free-living species also are unexpectedly disjunct between the 

 stages : as examples, Simnlium wellmanni has the immature stages normal for the 

 group to which it clearly belongs but has an exceptional male hypopygium with 

 multiple style spinules quite unlike that of its congeners, and Simnlium albivirgu- 

 latum has adults and pupae essentially similar to those of its obvious allies but has a 

 completely different and unique larval stage. Because of such incongruence 

 between the developmental stages it is not always easy to define subgenera or 

 species-groups in a way which is completely satisfactory for both sexes, pupae and 

 larvae, and some segregates are more easily recognized in one stage than another ; 

 but so far as possible the diagnoses given have been framed to cover all variants that 

 may occur in all stages of the included species. Subgenera are defined by combina- 

 tions of characters from all stages, but I have tended to attach greatest weight to 

 differences or resemblances among adults and to consider aggregates of species as 

 consubgeneric if no adult differences exist, for if completely equal weight is attached 

 to characters of all stages (as Davies, 1965 : 167, considers best for erection of black- 

 fly genera) it results in too much taxonomic value being attached to the occasional 

 aberrant larval feature (such as the short flat mouth-brush of Simnlium copleyi or 

 unique hypostomium of S. berneri), to bizarre forms of pupal gill or to an aberrant 

 pupal abdominal onchotaxy (such as that of 5. lumbwanum). Within subgenera, 

 however, there are species-groups that are best distinguished on features of the 

 immature stages, often the pupal gill. 



Although this paper is concerned primarily with classification above the species 

 level, and not with the identity of species among themselves, it is necessary to allude 

 briefly to my treatment of the pupal forms as recognized by Freeman & de Meillon 



