We used experimental manipulations to test adaptive explanations for the courtship display of the male widow spider, Latrodectus pallidus O. Pickard-Cambridge 1872. Two hypotheses have been suggested to explain a long and complex male display: a) Cooperation of males and females in the effort to physically stimulate the female. As the time of male arrival is not predictable, females may delay sexual readiness until the appearance of a courting male. b) Conflict between males and females regarding the display cost. Females impose on the males an energetically costly display that may last several hours as a test of their quality. To test both hypotheses, we manipulated the previous experience of either the male or the female. We presented naive or experienced males (males that had courted and were accepted by females but were prevented from copulating) to females that were either naive or experienced (had been courted by a male but prevented from copulating). We also presented naive males to mated females. Following the stimulation hypothesis, courted females were presumed to have been stimulated to mate and thus were expected to accept non-courting males as mates. Both naive and mated females, however, were expected to await male stimulation before allowing copulation. In contrast, the conflict of interest hypothesis predicts that the female tests each male for quality indicators and therefore a non-courting male should not be accepted as a mate. Mated females, however, should apply a less stringent test to courting males. Our results show that 1) naive females prevented males that did not perform a full courtship display from entering the nest and mounting; 2) naive males courted virgin females with the full display, independent of the female previous courting history; and 3) naive males shortened their courtship when presented with mated females. The results are consistent with the conflict of interest hypothesis.