Eating disorders are prevalent and complicated disorders which are difficult to treat. Unicausal and main effects models are not likely to do justice to the complexity of psychopathology encountered, as one considers etiology and pathogenesis. Risk and protection can arise out of several domains: biological, psychological and social. Risk and protective factors aggregate in specific developmental phases and interact to produce adverse outcomes. Temperamental factors, eating dysregulation, attachment, deficient self regulation and sociocultural ideals of health and beauty all contribute to pathogenesis. Applying the insights of developmental psychopathology to these disorders has considerable potential to lead to early and preventive interventions. Reviewing the current literature from this perspective and updating a similar discussion from 8 years ago, we witness a continued accumulation of quality empirical data. Compared to previous reviews, the field's attention has shifted to psychosocial/cultural domains relevant to eating, away from biological risk. In the aggregate, these data make possible the increasing differentiation of eating disorders from other psychopathology, and the specific pathways in which anorexia and bulimia may develop. Understanding of risk and vulnerability still outweighs our knowledge of protective factors and resilience. While an ideal study would be longitudinal, such studies are still extremely difficult to conduct and costly, thus, forcing us to further our understanding from lagged designs, cross-sectional data and case control studies. While these have many limitations, they do seem to produce an increasingly coherent account of the development of these disorders and prepare us for more targeted and longitudinal study of high risk populations.