Our understanding of deceit in animal signalling is limited by our conceptual tools as well as by the available data. Mate choice in polyterritorial birds, such as pied flycatchers, Ficedula hypoleuca, is a particularly contentious example. Dale & Slagsvold (1994, Anim. Behav., 48, 1207-1217) developed a quantitative model of discrimination of male quality (mating status) based on repeated sampling of territories by searching females. The model emphasizes the trade-off between correctly rejecting already mated males and incorrectly rejecting unmated males. In the present study, their model was incorporated into the framework of signal detection theory and the relationship between search costs, search tactics and the resulting rates of discrimination errors was examined. Over a range of reasonable assumptions about probabilities and payoffs, the scenario suggested by Dale & Slagsvold implies that female pied flycatchers should inspect each territory only a few times and accept only males that are present at every inspection. Consequently, male mating status should be moderately discriminable. The quantitative tools offered by signal detection theory are consistent with the conceptualization of deceit established by Dawkins & Krebs (1978, Behavioural Ecology: An Evolutionary Approach) and with the recent evolutionarily stable strategy (ESS) analyses of signalling developed by Johnstone & Grafen (1992, Proc. R. Sec. Lend. Ser. B, 248, 229-233). (C) 1996 The Association for the Study of Animal Behaviour