In a previous publication, this group reported a dissociation between multiple sleep complaints and polysomnographic abnormalities in prepubertal major depressive children. Spindle activity has been thought by some to be a correlate of restorative sleep. On this basis, it was hypothesized that prepubertal depressives might exhibit differences in sleep spindle characteristics when compared with normal and psychiatric control groups, and with their former state during sustained recovery in a drug-free state. Further, it was thought that clinical complaints of disrupted sleep might be related to differences in spindle characteristics. Spindle frequency (Hz) was the only significant finding among the groups, but it was not specific to depressive disorder, since all psychiatric groups presented higher frequencies than did normal children. As expected, significant differences across halves of the sleep period time (SPT) were also found for the number of spindles, frequency (Hz), duration (ms), spindle density (rate), and the amount of stage 2 sleep (min). The presence or absence of sleep complaints had no bearing on spindle characteristics, thus indicating a lack of apparent relationship between the two. Upon recovery the depressives exhibited no changes across clinical state, and a cross-sectional comparison of the recovered data with the original control data gave results essentially identical to the first (active illness) comparisons. It is suggested that high spindle frequency may be a nonspecific correlate of emotional disorders in prepuberty, which, at least in depressives, does not change with clinical state.