This study examined the convergent validity of maternal reports of child emotion in a sample of 190 children between the ages of 3 and 6. Children completed a battery of 10 emotion-eliciting laboratory tasks; their mothers and untrained naive observers rated child emotions (happiness, surprise, fear, sadness, and anger) following each task, and trained coders rated videotapes of each task for the same emotions. Consistent with studies using other designs, maternal reports demonstrated weak to moderate convergence with the other rating methods. Extending prior research, a number of maternal characteristics (particularly lifetime psychiatric diagnoses and personality traits) were associated with their reports of child emotions in the lab, above and beyond the effects of objective coding and observer reports of child emotions. For some emotions, mothers' mental health and dispositional variables were more strongly related to their reports of the child's emotions than were objective indices of the child's observable emotional behavior. (Contains 4 tables.)