Experimentation with the subjective location of clicks heard during speech supports the following conclusions: (a) Clicks are attracted towards the nearest major syntactic boundaries in sentential material. (b) The number of correct responses is significantly higher in the case of clicks located at major segment boundaries than in the case of clicks located within segments. (c) These results are consistent with the view that the segments marked by formal constituent structure analysis in fact function as perceptual units and that the click displacement is an effect which insures the integrity of these units. (d) The distribution of acoustic pauses in the sentential material does not account for the observed distribution of errors. (e) There is a slight tendency to prepose responses to clicks in sentences. This tendency is reversed during later stages of the experimental session. Both these effects are asymmetrical for the two ears.