Volatiles from potato plants (Solanum tuberosum L.) infected with Potato leaf roll virus (PLRV) attract and arrest the principal vector of PLRV, the green peach aphid, Myzus persicae (Sulzer), more strongly than volatiles from non-infected plants. The total concentration of volatiles detectable in the headspace of PLRV-infected plants is greater than that in the headspace of non-infected controls, and the relative composition is altered. To determine the basis of the aphid response to PLRV-infection-induced volatiles from potato, behavioral bioassays were conducted. We measured arrestment of aphids by individual components, by synthetic blends of these, and by a naturally occurring blend by using an emigration rate bioassay, and quantified observations of the behavior of individual aphids. The components tested were those elevated at least twofold in response to PLRV infection. Before conducting the behavioral bioassays, electroantennograms confirmed the electrophysiological responses of aphids to the components of the blend. For bioassays, individual compounds or blends were tested by applying them in solution to paper strips at concentrations designed to mimic those present in the headspace of the plants. All bioassays were conducted by placing aphids on fine-mesh screening positioned above treated paper strips. Arrestment was measured by placing groups of 30 aphids directly over the treated strips and counting the number moving away at 10-min intervals for 50 min. Among the individual compounds tested, only beta-pinene was a mild arrestant. The other compounds did not elicit significant changes in arrestment or behavior at a range of physiologically relevant concentrations. In contrast, synthetic blends that mimicked the concentration and composition present in headspace of PLRV-infected potato plants arrested aphids significantly more strongly than blends mimicking volatiles from the headspace of non-infected plants. The naturally occurring blend collected from headspace of PLRV-infected plants also arrested M. persicae more strongly than the blend collected from headspace of non-infected plants. Aphid behavior was quantified by directly observing individual aphids and recording their activities during a 5-min period on screening above strips treated with test materials. Few differences in time budgets were observed among aphids exposed to individual components, but synthetic blends and trapped headspace volatiles from PLRV-infected plants resulted in significantly less time spent walking by aphids than synthetic blends and trapped headspace from non-infected controls. Our results indicate that arrestment of M. persicae by PLRV-infected plants requires the blend of volatile organic compounds released by these plants and is not produced in response to a single compound.