S ince Bowlby's seminal contributions (1), attachment and separation have become familiar theoretical components of ego psychology, developmental psychology, and psychodynamic theory. In spite of the important roles these concepts have played in studies of psychological development, relatively little is known about their neurobiological basis. If one considers that abnormal social attachments characterize virtually every form of psychopathology, it seems especially remarkable that the chemistry, the anatomy, and the physiology of social bond formation remain largely unexplored. To be sure, there is a considerable literature on the neurobiology and psychobiology of separation (2), but this work may not prove instructive for understanding social bond formation. Attachment is not simply the absence of separation. Accordingly, there is no obvious reason why attachment and separation should be subserved by the same neural pathways.Attachment, of course, is not a thing but a process (3). More specifically, it is a social process and thus not likely to be defined by a single neurochemical pathway nucleus or represented in a single anatomic nucleus. As Harlow noted throughout his insightful writings on this subject (4), attachment includes several quite different processes depending on the social context: parent-infant, filial, and pair (male-female) bond formation are all forms of attachment. All of these forms involve seeking proximity and all involve a response to separation, but the strategy for and the consequences of achieving proximity vary depending on the relationship. None of these forms of attachment is uniquely human, suggesting that the neural basis can be investigated in animal models. It is also possible (and experimentally testable) that these various forms of attachment use similar neural pathways, suggesting that, biologically, attachment is a singular process that is manifested with different behaviors depending on the external (e.g., social) or internal (e.g., endocrine) context. This review will describe research on two neuropeptides, oxytocin and vasopressin, that have been implicated in the central mediation of attachment. I will de-