The aim of the thesis is to examine relations between pattems of attachment with mother and subsequent behaviour with peers and perceptions of self in young children. The sample consisted of 39 five year-old children (22 girls, 17 boys). Attachment classifications had been detelmined when the children were 4 1/2 years old, as part of a longitudinal study, using procedures and coding systems originally developed for infants by Ainsworth (Ainsworth, Blehar, Waters, & Wall, 1978) that were modified for 3 - 4 year-old children by Cassidy and Marvin (1988). These were based on behaviour shown in the lab to mother on reunion after a brief separation. Children were classified as: Secure, Insecure-avoidant and Insecureambivalent. When each child was five years old, behaviour with peers was assessed through direct observation on the school playground for five 15-minute periods. A continuous commentary of interactions was made into a hand-held microphone, while a radio microphone concealed on the child picked up the child's speech and speech directed toward him/her. A 15-minute video recording was also made. Tapes were transcribed using a coding system based on that used by Hinde, Easton, Meller and Tamplin (1983). Analysis revealed meaningful patterns of relations between patterns of attachment and subsequent behaviour with peers. Insecure-ambivalent children exhibited more negative behaviour toward peers and sought the attention of peers more than did Secure and Insecure-avoidant children, and they complied to controls less than did Secure children. Insecure-avoidant children tended to engage in more neutral, less involved behaviour (neither 'positive' nor 'negative' with peers (e.g., just listening as a response to peers). Secure children tended to show more playful behaviours (play aggression, play noises, playful teasing and imitating) than did Insecure-avoidant children and tended to exhibit less negative behaviour than did Insecure-ambivalent children. These results are consistent with previous evidence (Arend, Gove & Sroufe, 1979; Sroufe, 1983) characterizing Insecure-avoidant, Secure, and Insecure-ambivalent children on a dimension ranging from over-controlled to under-controlled (Block & Block, 1980). In addition, ratings of security and avoidance upon reunion with the mother in the lab predicted behaviour with peers. Security ratings were positively correlated with playing games alone on the playground and negatively correlated with listening as a response and neutral speaking. Security ratings were also correlated with peer behaviour directed toward the child. Security was positively correlated with peers speaking boastfully and making play noises to the child and negatively related to peers asking the child questions. Avoidance ratings were positively correlated with listening as a response to peers but negatively correlated with neutral activity (doing nothing). Analysis of girls and boys separately revealed further significant relations. For example, for boys, avoidance ratings were positively correlated with speaking with hostility, seeking entry into games and automanipulating, and negatively correlated with positive expressive behaviours and engaging in large muscle play. Perceptions concerning perceived competence and social acceptance, self-efficacy, perceived popularity with, and liking of, peers and interpersonal problem-solving ability were assessed through a series of four separate interview sessions with each child. Insecure-avoidant children generally reported relatively negative self-perceptions while Insecure-ambivalent children reported very positive (perhaps idealized) perceptions concerning competence, social acceptance, and peer friendships. Results showing different relations for girls and boys indicate a need to consider this potentially important variable when studying links between attachment, behaviour and perceptions. The results provide support for the predictive validity of the attachment classifications and for Bowlby's (1969/82, 1973, 1980) proposition that the child's attachment relationship with mother forms the basis for behaviour in relationships with others and relates to perceptions concerning the self and others in the absence of mother.