Morphometric studies of five allopatric parasitoid populations (genus Psyttalia Walker) from coffee plantations in Cameroon (Nkolbisson), Ghana (Tafo) and Kenya (Rurima, Ruiru and Shimba Hills) and one non-coffee population (from Muhaka, Kenya) were compared with individuals of Psyttalia concolor (Szépligeti), a species released in several biological control programmes in the Mediterranean Region since the 20th Century. Analyses of wing vein measurements showed the second submarginal cell of the fore wing and its adjoining veins had the heaviest principal component weights and served as the main contributing variables in the diagnostic differentiation of the populations. Two populations (Rurima and Ruiru) were found to be the closest to each other and with the strongest phenetic affinity toward P. concolor (and forming one cluster). Populations from Shimba Hills (of unknown identity), Nkolbisson (P. perproximus (Silvestri)) and Tafo formed a second cluster and were separated from P. concolor. Comparison using amplified fragment length polymorphism (AFLP) also showed the Shimba, Nkolbisson and Tafo populations forming a cluster in a dendrogram generated from their genetic distances, with the Shimba and Tafo populations placed as the most closely related species. Based on consistent morphological similarities, morphometric and ecological data coupled with the genetic evidence from AFLP data, the Shimba population is suggested as belonging to the P. perproximus group and, thus, represents a new occurrence record in Kenya. Our results also support earlier conclusion from cross mating data that populations from Rurima and Ruiru belong to the Psyttalia concolor species-group.

Three species of fruit fly parasitoids, Psyttalia concolor (Szépligeti), Psyttalia cosyrae (Wilkinson) and Psyttalia lounsburyi (Silvestri) (Hymenoptera: Braconidae) were reared on different host larvae to determine the effects of host species, size and colour on parasitoid development. Ovipositor, ovipositor sheath and hind tibia lengths were found to be different for progenies from preferred and non-preferred host larvae. Ratios of ovipositor–tibia and ovipositor sheath–tibia showed no difference when P. concolor was reared on a bigger host (Ceratitis cosyra) (Walker) (Diptera: Tephritidae), whereas rearing P. cosyrae on a smaller host (C. capitata) (Wiedemann) showed significant differences. Comparison of the linear measurements, ratios and number of flagellomeres of parasitoids reared on preferred hosts, separated the species into their natural groupings. When reared on a different-coloured host, P. lounsburyi lost its dark markings but other characters remained unchanged. Morphometric analysis also indicated differences between parasitoids reared on preferred host larvae and those reared on non-preferred host larvae. Rearing parasitoids on larvae other than their preferred hosts resulted in colour and size changes, and thus, these characters alone were not reliable for the identification of parasitoids. This is especially important in post-release sampling surveys to ascertain establishment of the parasitoids in new environments where they may adapt to new host species.