The buffer theory postulates that social support moderates the power of psychosocial adversity to precipitate episodes of illness. In this paper, we review the theory as applied to minor affective disturbances. Research in this area suffers because of the many disparate conceptualizations of social support and the resulting difficulty of deciding on the content of measures. Moreover, the meaning of the term buffering is itself unclear. These problems have not, however, inhibited research, and many cross-sectional and longitudinal studies have now been carried out. Our review leads to the conclusion that evidence for a buffering role of social support is inconsistent, reflecting methodological differences between studies but probably also indicating that buffering effects are not of dramatic proportions. Moreover, it is possible that the observed relationships are the result of spurious association or contamination of measures.