Developmental theorists have made strong claims about the fundamental prosocial or aggressive nature of the human infant. However, only rarely have prosocial behavior and aggression been studied together in the same sample. We charted the parallel development of both behaviors from infancy to childhood in a British community sample, using a two-construct, multimethod longitudinal design.Data were drawn from the Cardiff Child Development Study (CCDS), a prospective longitudinal study of a volunteer sample of parents and their firstborn children. A sample of 332 mothers was recruited from National Health Service (NHS) prenatal clinics and general practice clinics in Wales, UK, between Fall of 2005 and Summer of 2007. Potential participants represented the full range of sociodemographic classifications of neighborhoods. Participating families were divided about equally between middle-and working-class families, were somewhat more likely to have sons than daughters, and the majority (90%) were in a stable partnership. In response to standard categories recommended for use in Wales at the time, the majority (93%) of mothers reported themselves as Welsh, Scottish, English, or Irish; most others named a European or South Asian nationality.Of the 332 families agreeing to participate, 321 mothers (M age = 28 years) and 285 partners (M age = 31 years) were interviewed during the pregnancy and 321 of the families contributed data at least once after the child's birth. After an initial home visit at 6 months, data collection occurred in four additional waves of testing when children's mean ages were approximately 1, 1.5, 2.5, and 7 years. Data collection alternated between family homes and Cardiff Uni-

It has long been hypothesized that pretend play is beneficial to social and cognitive development. However, there is little evidence regarding the neural regions that are active while children engage in pretend play. We examined the activation of prefrontal and posterior superior temporal sulcus (pSTS) regions using near-infrared spectroscopy while 42 4- to 8-year-old children freely played with dolls or tablet games with a social partner or by themselves. Social play activated right prefrontal regions more than solo play. Children engaged the pSTS during solo doll play but not during solo tablet play, suggesting they were rehearsing social cognitive skills more with dolls. These findings suggest social play utilizes multiple neural regions and highlight how doll play can achieve similar patterns of activation, even when children play by themselves. Doll play may provide a unique opportunity for children to practice social interactions important for developing social-emotional skills, such as empathy.

Shared humor is a central feature of children's close relationships and a universal, integral part of human experience that is characterized by playful incongruity or, a benign conflict between what is expected and experienced

Although it is well established that features of maternal speech are associated with children's social understanding in the preschool years, few studies explore this relationship in middle childhood. Within the context of a prospective longitudinal study of a representative community sample of families (subsample n = 207, mean age = 82.88 months), we investigated concurrent associations between mothers' internal state language and aspects of 7-year-olds' social understanding, including children's understanding of belief and spontaneous references to internal states during free play. When sociodemographic, maternal, and child characteristics were controlled, mothers' references to their own cognitions were associated with dimensions of children's social understanding. Our findings suggest that exposure to others' perspectives contributes to children's advanced understanding of minds, which has implications for interventions that foster social understanding. K E Y W O R D S community sample, family relationships, internal state language, longitudinal study, social understanding