NOD mice were treated with silica (which is selectively toxic to macrophages) from 4 or 20.5 wk of age. Syngeneic neonatal pancreases were transplanted into the renal subcapsular space of the NOD mice at 21 wk of age. Silica treatment was continued until 24 wk of age, and then the mice were killed for examination of islet morphology. Neither the islets in transplanted pancreases nor the host pancreatic islets from the early long-term silica-treated animals revealed 1insulitis. In contrast, most of the islets in transplanted pancreases from the late short-term silica-treated animals showed severe insulitis and β-cell necrosis, as did the host islets. A further experiment was performed to compare the effect of late short-term silica treatment with that of anti-L3T4-antibody treatment of the same time and duration. In contrast to the late short-term silica-treated animals, the transplanted pancreases in the anti-L3T4-antibodytreated animals revealed intact islets, although most of the host islets showed insulitis. The control group, which received no treatment but did receive neonatal pancreases, revealed severe insulitis and β-cell necrosis of both transplanted and host islets. These results suggest that early macrophage depletion can abolish the development of β-cell-specific immunologic effectors but that late macrophage depletion, after the development of insulitis, does not affect the destruction of β-cells by preexisting effectors other than macrophages. We conclude that macrophages are essential for the development of β-cell-specific cytotoxic effectors in the initial phase of insulitis in NOD mice.