Gallium(III) is a new therapeutic agent for hypercalcemia. Ga3+ reduces osteoclast action, but how it inhibits the cell's physiology is unknown. In vivo, 7–12 μM Ga(III) reduces calcium release from bone, but surprisingly, 10–100 μM Ga3+; added to isolated avian osteoclasts did not reduce their degradation of L-(5-3H)-proline bone. 3H-proline labels bone collagen specifically, and collagenolysis is an excellent indicator of bone dissolution because collagen is the least soluble component of bone. Ga(III) > 100 μM inhibited osteoclasts in vitro, but also killed the cells. To resolve this apparent conflict, we measured 67 Ga distribution between bone, cells, and media. Gallium binds avidly but slowly to bone fragments. One hundred micrograms of bone clears 60% of 1 μM gallium from 500 μI of tissue culture medium, with steady state at > 24 h. Osteoclasts on bone inhibited gallium binding capacity ∼ 40%, indicating a difference in available binding area and suggesting that osteoclasts protect their substrate from Ga binding. Less gallium binds to bone in serum-containing medium than in phosphate-buffered saline; 30% reduction of the affinity constant suggests that the serum containing medium competes with bone binding. Consequently, the effect of [Ga] on bone degradation was studied using accurately controlled amounts of Ga(III) pre-bound to the bone. Under these conditions, gallium sensitivity of osteoclasts is striking. At 2 days, 100 μg of bone pre-incubated with 1 ml of 1 μM Ga3+, with 10 pmoles Ga3+/μg bone, was degraded at 50% the rate of control bone; over 50 pM Ga3+/μg bone, resorption was essentially zero. In contrast, pre-treatment of bone with [Ga3+] as high as 15 μM had no significant effect on bone resorption rate beyond 3 days, indicating that gallium below ∼150 pg/μg bone acts for a limited time and does not permanently damage the cells. We conclude that bone-bound Ga(III) from medium concentrations < 15 μM inhibits osteoclasts reversibly, while irreversible toxicity occurs at solution [Ga3+] > 50 μM.