The efficiency of electrocuting devices currently used for sampling tsetse flies (Glossina spp.) and similar insects, was studied in Zimbabwe by recording approaches, kills, and escapes, with video. The kill rate of an electrified netting screen increased with the discharge frequency of the device up to 200 Hz (ca the highest practicable frequency) reaching ca 90% at best. The same kill rate was achieved by an electrified black cloth target. However, 'two-choice' comparisons of electric nets and their components showed avoidance by the tsetse of the black mosquito netting between the electric wires, and even of the electric wires on their own, though probably not of the black metal frame that supported them. The proportion of tsetse avoiding a standard electric net was ca 27% in full sun, ca 40% in shade, implying an overall sampling efficiency of, at best, ca 65% at the optimum 200 Hz discharge rate in sunshine, and ca 40-50% with the 67 Hz nets used currently in Africa. Potential for improvements therefore lies mainly in reducing the visibility of the nets; suggestions are offered.