Eighty girls and 64 boys (M= 6 years; 8 months, SD= .65) narrated a wordless picture book in mixed- or same-gender dyads. In mixed-gender as well as same-gender dyads, girls used more emotion explanations than did boys. Combined across dyad type, girls used more emotion labels than did boys. Girls used a higher proportion of collaborative speech acts than did boys in same-gender dyads, but girls and boys used the same amount in mixed-gender dyads. Whereas girls used a higher proportion of informing acts in mixed-gender dyads than did boys, boys used more than did girls in same-gender dyads. The findings support contextual models of gender and suggest that speaker as well as partner gender influence emotion expression and conversational style.