Marginal copper (Cu) deficiency is difficult to study, in part because its effects may be small, but also because feeding of a deficient diet may not cause a discernable change in Cu status. The key to resolution of effects may be in the choice of Cu status index. In this study, liver Cu concentration, a commonly used index of Cu status, was compared with activity of ceruloplasmin (CP), a circulating Cu-dependent enzyme, and kidney Cu concentration for their utility in resolving effects of marginal Cu deficiency. Seventy male, weanling rats were fed diets containing, nominally, 0, 1.5, 3, 4.5, or 6 mg Cu/kg diet for 5 wk. All three indices showed strong depression with severe deficiency (dietary Cu=0), but were relatively weak in their ability to distinguish between animals fed marginally deficient diets when compared by group statistics (ANOVA). Further, group statistics revealed no effect of marginal deficiency on six other variables known to change with severe Cu deficiency: heart weight/body weight, hematocrit, red cell distribution width, neutrophil count, glycated hemoglobin, and platelet count. To take into account interanimal variation, the three putative indices were plotted against these six variables and linear regression was performed on points representing marginally deficient rats. None of the variables showed significant regression with liver Cu or serum ceruloplasmin, but three showed significant regression with kidney Cu. These findings indicate that kidney Cu is preferable to liver Cu or ceruloplasmin as an index of Cu status in marginal deficiency and that linear regression is a possible way of testing for effects of marginal Cu deficiency, especially when effects are subtle.