Recent studies of auditory streaming have suggested that repeated synchronous onsets and offsets over time, referred to as "temporal coherence," provide a strong grouping cue between acoustic components, even when they are spectrally remote. This study uses a measure of auditory stream formation, based on comodulation masking release (CMR), to assess the conditions under which a loss of temporal coherence across frequency can lead to auditory stream segregation. The measure relies on the assumption that the CMR, produced by flanking bands remote from the masker and target frequency, only occurs if the masking and flanking bands form part of the same perceptual stream. The masking and flanking bands consisted of sequences of narrowband noise bursts, and the temporal coherence between the masking and flanking bursts was manipulated in two ways: (a) By introducing a fixed temporal offset between the flanking and masking bands that varied from zero to 60 ms and (b) by presenting the flanking and masking bursts at different temporal rates, so that the asynchronies varied from burst to burst. The results showed reduced CMR in all conditions where the flanking and masking bands were temporally incoherent, in line with expectations of the temporal coherence hypothesis.