REVISION OF LEUCOSPIDAE 117 



all, different from L. ornata. Most puzzling is one rather slender and large female 

 (12 mm) from Rabai, Kenya, with widespread red colour and unusually long 

 and dense pilosity, subdecumbent on the gaster. Both thorax and gaster are 

 very densely and relatively finely punctured, as well as hind coxa, with the pilosity 

 in places about twice as long as in the average L. ornata. Although the ovipositor 

 reaches the propodeum, I do not think now that this specimen is specifically different 

 from another female of the same origin. The latter specimen belongs to another, 

 often extensively red, form represented also by females from Nigeria, Central 

 African Republic, Sudan and South West Africa, possibly from semi-desert 

 conditions. In these the relatively large and slender body has the pubescence 

 of medium length, puncturation of thorax and gaster rather dense though less so 

 than the mentioned aberrant specimen, in many respects more like the typical 

 tricarinata, ovipositor reaching anterior half or third of first tergite. The next 

 form, to which also the type of tricarinata belongs, has the pubescence still shorter, 

 hind femur more densely and about as coarsely punctured as sides of fifth tergite, 

 body often smaller; it is represented by specimens from Portuguese Guinea, 

 Equatorial Guinea, Gabon, Zaire, and Durban in South Africa. The most sparsely 

 punctured specimens come mostly from South Africa (including the type of ornata), 

 with the extremely sparsely ones, at least as to hind femora, from the Springbok 

 district (Hester Malan). In these 'typical' ornata the pronotal carinae usually are 

 more angulate, the premarginal one nearer to the discal than to the carinate hind 

 margin of pronotum, but on the whole this is not confined to South African specimens 

 and even among them some have relatively weaker carinae. 



The types of variation discussed above are difficult to associate with any 

 geographical or ecological information and very little is known about the hosts. As 

 all gaps seem to be linked by intermediate forms I regard the mentioned variation 

 as intraspecific and classify all those specimens as L. ornata. 



I hope to have separated correctly L. ornata from the closely related species, 

 also in the males, unless the very similar L. varicollis is only a subspecies of ornata, 

 which I doubt. There is no difficulty in separating the females, but the males are 

 much more troublesome. In ornata they usually show the following set of characters. 



(J. T5-^o-^ mm. First tergite relatively broad, 1 "29-1 -47 times as long as broad; hind 

 femur less broad basally, often rather densely punctured (at least those from Central Africa), 

 bordered with yellow or white mainly along whole dorsal and ventro-basal edges. Gaster 

 mostly with two cross-bands on carapace, apart from apical transverse macula. Wings relatively 

 more, often strongly infuscate, uncus of stigmal vein relatively shorter than in varicollis (for 

 that species see under varicollis). The two species are recorded from the same host bee and 

 have a considerable overlap in distribution areas. 



Biology. Parasite in nests of Megachiline bees, e.g. of Lithurge capensis 

 Friese, Lithurge sp. and Megachile willowmorensis Brauns. Some of the specimens 

 from Alice, bearing otherwise the same data as those reared from the last named 

 host, are labelled as coming from a 'carpenter bee' (Pmisidentification of the host). 



Distribution. Portuguese Guinea, Sierra Leone, Liberia, Ghana, Nigeria, 



