Dietary Mg intake has been linked to osteoporosis. Previous studies have demonstrated that severe Mg deficiency [0.04% of nutrient requirement (NR)] results in osteoporosis in rodent models. We assessed the effects of more moderate dietary Mg restriction (10% of NR) on bone and mineral metabolism over a 6-mo experimental period in rats. At 2, 4 and 6 mo, serum Mg, Ca, parathyroid hormone (PTH), 1,25-dihydroxy-vitamin D, alkaline phosphatase, osteocalcin and urine pyridinoline were measured. Femurs and tibiae were collected for measurement of mineral content, microcomputerized tomography, histomorphometry, and immunocytochemical localization. By 2 mo, profound Mg deficiency had developed as assessed by marked hypomagnesemia and up to a 51% reduction in bone Mg content. These features continued through 6 mo of study. Serum Ca was slightly but significantly higher in Mg-deficient rats than in controls at all time points. At 2 mo, serum PTH was elevated in Mg-deficient rats but was significantly decreased at 6 mo in contrast to control rats in which PTH rose. Serum 1,25-dihydroxy-vitamin D was significantly lower than in controls at 4 and 6 mo. A significant fall in both serum alkaline phosphatase and osteocalcin suggested decreased osteoblast activity. Histomorphometry demonstrated decreased bone volume and trabecular thickness. This was confirmed by microcomputerized tomography analysis, which also showed that trabecular volume, thickness and number were significantly lower in Mg-deficient rats. Increased bone resorption was suggested by an increase in osteoclast number over time compared with controls as well as surface of bone covered by osteoclasts and eroded surface, but there was no difference in osteoblast numbers. The increased bone resorption may be due to an increase in TNF-alpha because immunocytochemical localization of TNF-alpha in osteoclasts was 199% greater than in controls at 2 mo, 75% at 4 mo and 194% at 6 mo. The difference in TNF-alpha may be due to substance P, which was 250% greater than in controls in mononuclear cells at 2 mo and 266% at 4 mo. These data demonstrated that a Mg intake of 10% of NR in rats causes bone loss that may be secondary to the increased release of substance P and TNF-alpha.