Spontaneous recovery was found in four-fifths of all stuttering cases. Data were obtained from 5138 students at the University of California, Berkeley and Los Angeles, of whom 147 had at some time been definitely categorized as stutterers. Severity was the most important factor related to probability of recovery. For those who had never been worse than mild, seven-eights recovered, while for those who had ever been severe, only one-half recovered. When severity was held constant, enrollment in public school therapy had no effect, positive or negative, upon the probability of eventual recovery. The fact that fewer of those who received public school therapy had recovered was primarily attributable to the finding that the severe were more likely to be enrolled. Parent, teacher, or speech clinician diagnosis of stuttering was significantly related to assignment to public school therapy, while self-diagnosis was inversely related. Those who began with blocking and became severe were more likely to have continued as stutterers, while those who began with syllable repetition were more likely to have recovered. Although familial incidence was higher in stutterers than in normal controls, it did not distinguish those who continued to stutter from those in whom the problem had disappeared.