A meta-analysis indicated that clinical depression was associated with several large alterations in cellular immunity. Analyzing only methodologically sound studies, reliable immune alterations included lowered proliferative response of lymphocytes to mitogens (effect size rs = .24-.45), lowered natural killer cell activity (r = .28), and alterations in numbers of several white blood cell populations (rs = .11-.77). Immune alterations were greater in both older and hospitalized samples. There was also evidence of a linear relation between intensity of depressive affect and indicators of cellular immunity. Estimates of sample sizes needed to detect reliable effects for each immune outcome are provided. How neuroendocrine mechanisms or health practices might link depression to immunity is discussed, and design features needed to better understand these pathways are specified.