Vision often dominates audition when attentive processes are involved (e.g., the ventriloquist effect), yet little is known about the relative potential of the two modalities to initiate a "break through of the unattended". The present study was designed to systematically compare the capacity of task-irrelevant auditory and visual events to withdraw attention from the other modality. Sequences of auditory and visual stimuli were presented with different amounts of temporal offset to determine the presence, strength, and time-course of attentional orienting and reorienting as well as their impact on task-related processing. One of the streams was task-relevant, while crossmodal distraction caused by unexpected events in the other stream was measured by impairments of behavioral task performance and by the N2, P3a, and reorienting negativity (RON) components of the event-related potential (ERP). Unexpected events in the visual modality proved to be somewhat more salient than those in the auditory modality, yet crossmodal interference caused by auditory stimuli was more pronounced. The visual modality was relatively constrained in terms of a critical time-range within which distraction effects could be elicited, while the impact of auditory stimuli on task-related processing extended over a longer time-range. These results are discussed in terms of functional differences between the auditory and visual modalities. Further applications of the new crossmodal protocol are deemed promising in view of the considerable size of the obtained distraction effects.