The paper explores the question whether resonants preceding and following a vowel can be considered as part of the syllable nucleus. A set of monosyllabic words was studied, beginning and ending with voiced and voiceless obstruent consonants, and containing syllable nuclei of the following types: V (long and short), V+R, R+V, and R+V+R. Three speakers produced each of the 156 words five times. An analysis of the durations of the segments yields (a) that syllable nuclei consisting of V+R function as a unit with respect to the timing patterns conditioned by the voicing of the following consonant, and (b) that resonants combine with long vowels in such a way that the durational differences between syllable nuclei of the types V, V+R, and R+V are below the perceptual level (i.e., below established difference limens for duration). The evidence is considered as supportive of the assumption that timing in speech is organized with reference to major changes in manner of articulation.