Zinc deficiency is prevalent throughout the world, including the USA. Severe and moderate deficiency of zinc is associated with hypogonadism in men. However, the effect of marginal zinc deficiency on serum testosterone concentration is not known. We studied the relationship between cellular zinc concentrations and serum testosterone cross-sectionally in 40 normal men, 20 to 80 y of age. In four normal young men (27.5 +/- 0.5 y), we measured serum testosterone before and during marginal zinc deficiency induced by restricting dietary zinc intake. We also measured serum testosterone in nine elderly men (64 +/- 9 y) who were marginally zinc deficient before and after 3 to 6 mo of supplementation with 459 mumol/ d oral zinc administered as zinc gluconate. Serum testosterone concentrations were significantly correlated with cellular zinc concentrations in the cross-sectional study (lymphocyte zinc versus serum testosterone, r = 0.43, p = 0.006; granulocyte zinc versus serum testosterone, r = 0.30, p = 0.03). Dietary zinc restriction in normal young men was associated with a significant decrease in serum testosterone concentrations after 20 weeks of zinc restriction (baseline versus post-zinc restriction mean +/- SD, 39.9 +/- 7.1 versus 10.6 +/- 3.6 nmol/L, respectively; p = 0.005). Zinc supplementation of marginally zinc-deficient normal elderly men for six months resulted in an increase in serum testosterone from 8.3 +/- 6.3 to 16.0 +/- 4.4 nmol/L (p = 0.02). We conclude that zinc may play an important role in modulating serum testosterone levels in normal men.