The formation of attachments is an important phenomenon occurring in the realm of socioemotional development. This study examined the impact of infant massage on infants' subsequent attachment security. Fifty-seven mother-infant dyads (48 dyads from Head Start, 9 from the community at large) were randomly assigned to a treatment or control group when infants were less than 8 months of age. The treatment group received training in infant massage and education about infant development during four 1-hour sessions. The control group received education only about infant development during four 1-hour sessions. Mothers completed questionnaires prior to the intervention and when their infants were 12 months old. Attachment security was assessed using the Attachment Q-set at the 12-month follow-up. There were no pretest or demographic differences between the two groups. Twelve subjects were lost at the 12-month follow-up, with equal loss occurring in both groups. The results indicated that mothers who massaged their 12-month-old infants more than once per week had infants who were statistically significantly more securely attached than infants of mothers who massaged their infants less than once per week, and were more securely attached than infants in the control group. Surprisingly, seven mothers in the control group reported that they massaged their infants at least twice weekly. Attachment security was also related to infant's gender, number of siblings, and maternal age. Regression analyses revealed that maternal age accounted for 20 percent of the variance in attachment security, and massage frequency accounted for 18 percent of the variance. (Author/KB)