preterm-infant studies, Swedish massage or heavy stroking (stroking with pressure) was used. Although infants, particularly premature infants, may seem to be fragile, some pressure is needed for the massage to be effective. A review of the infant massage literature suggests that those who used December 1998 • American Psychologist 1271 light stroking did not report weight gain, for example, whereas those who used stroking with pressure reported weight gain (Scafidi et al., 1986). Preterm Infants In a series of studies I recently reviewed (Field, 1998), preterm infants in the neonatal intensive care unit were given 15-minute massages three times a day for 10 days while they were still in the incubator (massaged through the incubator portholes; Field et al., 1986; Scafidi et al., 1990). The treated infants compared with control infants gained 47% more weight and were hospitalized for six days less at a hospital cost savings of $10,000 per infant. Norepinephrine and epinephrine levels increased in the massaged infants relative to the control infants (who did not receive massage therapy) across the treatment period. Because these neurotransmitters normally increase during the neonatal period, this finding was interpreted as massage therapy facilitating the normal developmental increase in these catecholamines in preterm infants during their newborn period (Kuhn et al., 1991). Finally, the treated infants performed better on the Brazelton Neonatal Behavior Assessment Scale (Brazelton, 1973). At one year the treated infants were still showing a weight advantage, and they also performed better on the Bayley Scales of Infant Development (Field, Scafidi, & Schanberg, 1987). Their scores averaged 12 points higher on the mental scale and 13 points higher on the motor scale than the control group. The infants' more responsive behavior on the Brazelton Neonatal Behavior Assessment Scale apparently elicited more stimulation from their parents, which led to the later gains in growth and development. Replication studies have been conducted in Israel (Goldstein-Ferber, 1998) and the Philippines (Jinon, 1996). In the Philippines study (Jinon, 1996) which was an exact replication of the Field et al. (1986) methodology, the preterm infants who were massaged gained 45% more weight than the nonmassaged infants. In the Israeli study (Goldstein-Ferber, 1998), which used mothers as therapists, 31% greater weight gain was reported for the massaged versus control preterm infants. In addition, the mothers who provided the massage experienced a decrease in depression. The Philippine and the Israeli studies respectively approximated the weight gain data (45% and 31%) published a decade ago (47%) by Field et al. (1986) and Scafidi et al. (1990; 31%) respectively. Finally, a recent study by Dieter, Field, and Hernandez-Reif (1998) suggested that a 46% greater weight gain can be achieved in preterm infants following only one week of massage. Cocaine-Exposed Infants Similar weight gains were noted in a study on cocaineexposed infants who were massaged versus those who were not (Wheeden et al., 1993). In addition, these infants showed superior motor behavior. HIV-Exposed Infants In a study on HIV-exposed infants, the mothers of the infants were used as the massage therapists. The mothers' treatment compliance rates were very high, perhaps because of the guilt they expressed for having exposed their infants to HIV and their own high anxiety levels (Scafidi & Field, 1996). Teaching parents to massage their infants often lowers their anxiety levels that are related to their feelings of helplessness about their infant's or child's condition. Helping with their children's treatment might be expected to decrease their anxiety levels and make them feel that they are contributing to the treatment. In addition, daily massages are economically feasible when the parents are used as therapists. In the study on HIV-exposed infants, the massaged infants' weight gain was significantly greater than the control group who did not receive massage therapy, and massaged infants showed significantly fewer stress behaviors. The parents' involvement and their own reduced stress can be thought of as a confounding or a contributing variable in all studies using parents as the massage therapists. Full-Term Infants In a study of 40 full-term one- to three-month-old infants born to adolescent mothers, infants were given 15 minutes of either massage or rocking for 12 days over a six-week period. During the massage sessions, the massaged versus the rocked infants (a) spent more time in active alert and active awake states; (b) cried less; (c) had lower salivary cortisol levels, suggesting lower stress levels; and (d) spent less time in an active awake state after the massage session (as opposed to the rocking session), suggesting that massage may be more effective than rocking for inducing sleep (Field, Grizzle, Scafidi, Abrams, et al., 1996). Analyses of the longer term effects suggested that by the end of the six-week treatment period, the massage group infants (a) gained more weight; (b) improved on emotionality, sociability, and soothability temperament dimensions; (c) showed better face-to-face interaction behaviors; (d) had decreased urinary stress hormones (cortisol) and catecholamines (norepinephrine, epinephrine); and (e) had increased serotonin