Animals eavesdrop on signals and cues generated by prey, predators, hosts, parasites, competing species, and conspecifics, and the conspicuousness of sexual signals makes them particularly susceptible. Yet, when sexual signals evolve, most attention is paid to impacts on intended receivers (potential mates) rather than fitness consequences for eavesdroppers. Using the rapidly evolving interaction between the Pacific field cricket, Teleogryllus oceanicus , and the parasitoid fly, Ormia ochracea , we asked how parasitoids initially respond to novel changes in host signals. We recently discovered a novel sexual signal, purring song, in Hawaiian populations of T. oceanicus that appears to have evolved because it protects the cricket from the parasitoid while still allowing males to attract female crickets for mating. In Hawaii, there are no known alternative hosts for the parasitoid, so we would expect flies to be under selection to detect and attend to the new purring song. We used complementary field and laboratory phonotaxis experiments to test fly responses to purring songs that varied in many dimensions, as well as to ancestral song. We found that flies strongly prefer ancestral song over purring songs in both the field and the lab, but we caught more flies to purring songs in the field than reported in previous work, indicating that flies may be exerting some selective pressure on the novel song. When played at realistic amplitudes, we found no preferences–flies responded equally to all purrs that varied in frequency, broadbandedness, and temporal measures. However, our lab experiment did reveal the first evidence of preference for purring song amplitude, as flies were more attracted to purrs played at amplitudes greater than naturally occurring purring songs. As purring becomes more common throughout Hawaii, flies that can use purring song to locate hosts should be favored by selection and increase in frequency.