Carcelia evolans (Wied.) is a common parasite of Diparopsis watersi (Roths.) in Northern Nigeria. The egg and early larval stages have not been observed, but the latter are thought to be passed in the larva of the host. The fully grown parasite larva appears to leave its host after the latter has entered the soil and formed its pupal cell, and either immediately before, or after, the host pupates. The parasite puparium is thus formed inside the pupal cell of D. watersi, but outside the host. Before forming its puparium, the parasite larva excavates a pit in the wall of the cell, thus facilitating the subsequent exit of the adult fly, which can push through the weakened cell at this point.
The life-cycle of C. evolans is closely adapted to that of its host, with short-term generations during the cotton season alternating with a long-term, or diapause, generation during the dry season. It is believed that C. evolans goes into diapause as a minute first-stage larva, but the factors that induce diapause are unknown.
The short-term pupal period of C. evolans (12–14 days) was less variable than that of D. watersi (11–31 days) at 27±5°C. In dry conditions, inside and outside the laboratory, the mean duration of diapause for C. euolans (27·3–38·8 weeks) was longer than that of D. watersi (22·9–36·2 weeks). In pupation troughs, approximating to field conditions, the mean duration of diapause in D. watersi (43·2 weeks) was greater than that of its parasite (36·1 weeks). The spreadover of emergence of the parasite from pupation troughs was more regular than that of moths, with no peak in October.
Earlier estimates of the incidence of parasitism at Samaru are probably inaccurate because adult flies have now been shown to be capable of escaping through the apertures of the perforated zinc of the cages then in use. Estimates of the rate of parasitism, made in southern Katsina, showed that 23·1 per cent, of pupal cells obtained at Daudawa between 13th November and 24th December 1959 from cotton that had earlier been treated with insecticidal sprays were parasitised, compared with 30·7 per cent. of those obtained in the same period from unsprayed cotton four miles away. The rate of parasitism was particularly low (16 per cent.) amongst the larvae collected at Daudawa in November, due possibly to the insecticide applications that had been made earlier in the season.
These results imply that the parasite is easier to kill than its host, and an increasing use of insecticides on cotton in Northern Nigeria may therefore adversely affect the degree of control achieved by the parasite.