It has been claimed that spoken utterances consist of isochronous rhythmic units (metric feet), lasting from one major stress to the next. A study of 17 specially constructed sentences showed that productions of the same metric foot types have remarkably similar durations. However, there are clear differences in the average duration of different foot types in the same position, and there is neutralization of durational differences in final position. The partial isochrony found in the study leads to two further questions: the influence of syntax on the production of rhythmic units, and the perceptual nature of isochrony. If rhythmic units are independent of syntactic structure, sentences consisting of the same lexical items should have the same temporal structure regardless of syntax. A study of syntactically ambiguous sentences showed that they can be successfully disambiguated by adjusting the rhythmic units to reflect syntactic structure. The hypothesis that isochrony is primarily perceptual was tested by having 30 listeners first judge the duration of metric feet in the 17 analyzed sentences and then perform the same task with nonspeech stimuli consisting of sequences of noise bursts and noise-filled intervals duplicating the rhythmic structure of the sentences. Listener performance was significantly better in the second test, suggesting that isochrony may indeed be at least partly a perceptual characteristic of spoken English.