There have been numerous arguments from the domains of phonotactics and stress for including the rhyme as an internal constituent of the syllable (Halle and Vergnaud, 1980). The rhyme has also been proposed as a production unit on the basis of slips of the tongue and word games (Fudge, 1987; Treiman, 1985). However, like the syllable itself, this constituent is rarely subject to error in spontaneous slips. In the experiment reported on here, the rhyme is experimentally tested using the error elicitation technique (SLIP) described in Motley, Baars, and Camden (1983). Subjects (n=44) were tested on VC (rhyme), CV, and C1C2 sequences where C1 was either a sonorant (SC) or an obstruent (OC). As predicted CV and VC were indistinguishable from one another: both sequences rarely moved as a unit and they were split apart equally often. On the other hand they were both significantly different from SC and OC sequences (Wilcoxon p<0.01). Besides lending more support to the ecological validity of the SLIP technique (Stemberger and Treiman) these results suggest that when sublexical errors occur, the rhyme is not a unit of syllabic structure on a par with other syllabic constituents. Other syllabic constituents like the onset, peak, and coda are subject to error.