When two masked, attended targets (T1 and T2) are presented within approximately half a second of each other, report of T2 is poor, compared with when the targets are presented farther apart in time--a phenomenon known as the attentional blink (AB; Raymond, Shapiro, & Arnell, 1992). Some researchers have suggested that an amodal bottleneck on working memory consolidation underlies the AB (see, e.g., Arnell & Jolicoeur, 1999). In the present work, T1 was masked, whereas T2 was unmasked. The modality of T1 (visual or auditory) and the modality of T2 (visual or auditory) were factorially manipulated across four experiments. For all modality combinations, T2's P3 event-related brain potential component was found to be delayed when T2 was presented soon after T1 (lag 3), compared with when T1 and T2 were presented farther apart (lag 8). Results suggest that the working memory consolidation bottleneck is amodal in nature, and provide evidence that visual, auditory, and cross-modality ABs all result from a bottleneck on consolidation operations.