The purpose of this investigation was to quantitatively and qualitatively characterize speech disfluencies exhibited by 3- and 4-year-old children who do (CWS, N = 36) and do not (CWNS, N = 36) stutter. Five measures of speech disfluency (e.g., percentage of total, other, and stuttering-like disfluencies, mean number of repetition units, and weighted SLD measure) were used in attempts to differentiate CWS from CWNS. Similar measures of stuttering (e.g., percentage of stuttering-like disfluencies consisting of disrhythmic phonations) were used to characterize speech disfluencies in 3- and 4-year-old CWS in relation to time since stuttering onset (TSO). It was hypothesized that such measures of speech disfluency should significantly differ between CWS and CWNS, as well as 3- versus 4-year-old CWS in relation to TSO. Results indicated that 4 out of the 5 dependent measures significantly differed between CWS and CWNS, and within the CWS group there was a significant relationship between TSO and the percentage of stuttering-like disfluencies when the effects of chronological age were partialled out of the regression analyses. Furthermore, 4-year-old CWS exhibited a moderate correlation between TSO and the percentage of stuttering-like disfluencies consisting of disrhythmic phonations, whereas 3-year-old CWS exhibited no such relationship between these two variables. Findings were taken to suggest that certain measures of speech disfluency appreciably differentiate CWS from CWNS and that 4-year-old CWS exhibit changes in nonreiterative forms of stuttering as a function of time since stuttering onset.