Calibration by two independent methods (dipole-sample superposition and long thin solenoid) agreeing within 10 percent; magnetometer response linear and charge-symmetric over 0.3-300 g_D. The trapping null itself is model-independent: it assumes only that a slow (beta ⇐ 10^-3) monopole binds strongly to aluminum nuclei.