Two experiments explored the influence of neonate characteristics on maternal behavior. Experiment 1 investigated the influence of maternal attribution to the neonate of certain behavioral characteristics on maternal behavior in interaction. After videotaped mother-neonate pair interaction, neonates were removed for a nominal examination. Mothers were informed only that their neonates had been either alert or not alert in the exam. Mothers were then videotaped in interaction and afterwards informed that the neonate was now in the opposite state. Another interaction was then videotaped. Findings revealed that mothers of normally alert neonates jiggled and rocked their infants' bodies less than when they received minimal instruction. Mothers' pre- and post-delivery medication level was related to several behaviors in complex patterns. Experiment 2 compared maternal response patterns that occurred after mothers were instructed to follow gross themes with their neonates. A total of 52 mother and neonate pairs were observed under three videotape instruction conditions. Under "comfort instruction," mothers held neonates by the torso more, repositioned them less, and jiggled and rocked them more, than under minimal instruction. Race and parity determined maternal behaviors in complex patterns. Findings are discussed in terms of maternal perception of, and attachment to, neonates. (RH)