Although Bowlby's original formulations on attachment emerged from observations of clinical populations (Bowlby, 1944, 1958, 1977a, 1977b; Rutter, 1979), until recently we possessed only rudimentary empirical information regarding attachment relationships in clinically disordered populations. However, during the past decade, as research in the discipline of developmental psychopathology has burgeoned (cf. Cicchetti, 1984a, 1984b; Rolf, Masten, Cicchetti, Neuchterlein, and Weintraub, in press; Rutter & Garmezy, 1983), increased attention has been paid toward understanding the organization of the attachment system in handicapped, “high-risk” and clinically disordered groups of youngsters (see, for example, Cicchetti and Serafica, 1981; Crittenden, 1988; Egeland and Sroufe, 1981a, 1981b; Radke-Yarrow, Cummings, Kuczynski, and Chapman, 1985; Sigman and Ungerer, 1984).

Proponents of attachment theory have argued that many psychopathological disorders may be brought about by deviations in the development of the attachment behavioral system or, much less commonly, by failure of its ontogenesis (Ainsworth, 1973, 1980; Bowlby, 1969/1982, 1977a, 1977b; Cic-chetti, 1987; Guidano and Liotti, 1983). Additionally, advocates of the attachment theory perspective have stated that strong causal relationships exist between children's experiences with their caregivers and their later capacities to develop secure affectional bonds (Bowlby, 1944, 1977a, 1977b; Sroufe and Fleeson, 1986). Because the study of risk conditions from a developmental perspective is believed to augment our understanding of normal and abnormal forms of ontogenesis (Cicchetti, in press; Cicchetti, Cummings, Greenberg, and Marvin, in press), we think that the study of attachment in maltreated infants can make significant contributions to our knowledge about relationship formation and dissolution.