A prospective study was designed to investigate Bowlby's (1958, 1969) theory that the development of the affectional bond between mother and infant - generally termed "attachment" - is the result of certain species-specific stimuli being prepotent as elicitors of instinctual responses in mothers and infants, and to contrast this approach with a reformulation by the author which attempts to include recent evidence pertaining to the receptor capabilities of neonates. In brief the author's formulation is that because of an evolutionary process the human infant discriminates certain visual and auditory stimulus dimensions more readily, these stimulus dimensions being particularly well represented by the caretaker's face and voice and thus once orientated to these stimulus sources selective attention will occur at a high rate. Initial orientation is seen as the result of the caretaker's proximity occurring because of response to infant signals and nutritional requirements. To test the appropriateness of the two approaches thirty primiparous women between the ages of 20 and 32 were obtained at Nelson Hospital, Nelson, New Zealand, during the first week after having given birth to a healthy infant, the group being subjected to a 17 minute film designed to teach the mothers to emit certain behaviours. These behaviours were selected as those which would provide either the infant with stimuli Bowlby (1958, 1969) suggests will elicit instinctive responses (mother's face, voice and ventral surface) or the mother with stimuli which elicit instinctive responding (infant crying, smiling and vocalizing). A second group of thirty mothers matched with the first on a number of relevant variables acted as a control group. Specific hypotheses were made which would enable the attachment relationship of the two groups to be compared, and differentiation between the two formulations to be made. The experimental manipulation was successful in producing desired infant and mother behaviours, the outcome in terms of the quality of interaction of mother and infant, and infant and stranger clearly favouring the experimental group. Specifically the mothers and infants of the experimental group engaged in more reciprocal interaction in which each was responding in a manner complementary to that of the other, such interaction beginning early during the observation, being unbroken and relatively enduring. The mutual orientation of mother and infant which brought this about was the result of infant response to the mother's presenting of her smiling, moving, talking face within the infant's visual field, this stimulus complex rather than maternal responses to signals from the infant operating. While maternal response to signals was not a significant factor in relation to maternal orientation to the infant, none-the-less it was the mother who initiated and maintained the continuing exchange and thus maternal or infant variables other than those measured must have been responsible for maintaining maternal responding. The infants in the experimental group were initially significantly more sociable to a male stranger, but by six months of age this positive response had changed to marked negative reaction for most of the infants, with some even having passed through this stage. This was in contrast to the control group whose responsiveness to the stranger was minimal at both three and six months and only rarely negative, none of the infants having passed through the stage of negative reactions to strangers. The appropriateness of current models of the attachment relationship in describing attachment was tested. All were able to describe the relationship of mother and infant in a manner which differentiated the two groups, this result being considered to give evidence of their basic similarity. The results were found to support the author's reformulation that the mother's face and voice have special stimulus characteristics in attracting high rates of attention from the infant, the resulting mutual orientation of mother and infant giving rise to attachment interaction. Thus the mother's face and voice and infant orientation towards them can be described as "precursors of attachment". Support for the mother's face, voice and ventral surface as elicitors of instinctive responses from the infant and infant signals eliciting instinctive responses from the mother was not forthcoming.