A series of phonetic production and perception experiments were designed to describe the phonological or phonetic domains of two effects in spoken English: final lengthening, generally interpreted as a mark for the edge of some linguistically-defined unit of speech production, and stress-timed shortening, generally interpreted as evidence for rhythmic forces that work toward equalizing the distance between stressed syllables. The results suggest that intonational phrase-final lengthening and a smaller durational effect, seeming word-final lengthening, are very different types of effects. The latter does not appear to be a phonologic boundary mark because of highly inconsistent production and perception, nor does final lengthening appear to be related to the edges of any prosodic constituent below the level of the intonational phrase, either phonologically or phonetically. It is suggested that the smaller effect is a result of articulatory timing relationships within the stress foot. (MSE)