Abstract. A study was carried out in the Peruvian Andes to test the suitability of CDC light traps for monitoring changes in the human‐landing rate of endophagic phlebotomine sandflies, following house‐spraying with pyrethroid insecticide. On four pairs of consecutive nights, sandflies were caught inside eight sprayed and eight unsprayed houses, either by human bait or by CDC light traps. The sandflies collected were Lutzomyia verrucarum (97%) and Lu. peruensis (3%), both probable vectors of Leishmania peruviana, and the species composition was unaffected by house‐spraying. A non‐linear relationship was detected between light‐trap and human‐bait catches, but the relationship did not diverge significantly from linearity within the range of sandfly abundance found in most houses in the endemic area (i.e. between 3 and 200 sandflies/house‐night), and did not differ significantly between sprayed and unsprayed houses. However, light trap catches had a significantly lower proportion of blood‐fed females in sprayed than in unsprayed houses, probably due to an insecticidal effect on post‐blood‐feeding behaviour. The proportion of Lu. verrucarum was significantly higher in light trap than in human bait catches, indicating that Lu. peruensis is either more anthropophilic or less phototropic than Lu. verrucarum. Copyright © 1995, Wiley Blackwell. All rights reserved