Examined the role of child temperament (T) in regulating adult-child cognitive interaction, in 2 studies. In study 1, 26 children (aged 2.5 yrs) and their mothers (Ms) participated in a joint problem solving task (JPST), followed by a child-only posttest involving related tasks. In study 2, another group of 26 Ss (aged 2.5 yrs) and their Ms participated in a JPST with different tasks. T ratings of Ss were completed by their Ms, using the Toddler Temperament Scale (W. Fullard, S. C. McDevitt, & W. B. Carey, 1984). Ss rated as having more difficult T received more cognitive assistance and disapproval from their Ms, as well as greater involvement in challenging aspects of JPST. T was not related to involvement during JPST nor to task performance during interaction or on posttest. Study 2 replicated the findings of Study 1. It is concluded that child T may mediate the cognitive opportunities that arise during parent-child interactions.