Forty-seven of 145 selected SPARC curves fail the simple algebraic MOND law within the tested bounds for fixed mass profiles

We test the simple algebraic MOND law on 145 selected SPARC galaxy rotation curves with 2,696 retained radii. The model uses fixed mass profiles, constant stellar mass-to-light ratios, no external field, and distance and inclination shifts of up to three listed errors. Each galaxy may use its own acceleration scale. A change of variables reduces the fit at a fixed transformed acceleration scale to a linear problem. Interval bounds cover the full scale range. We save fit examples and rejection weights, then check them with separate 80-digit code that uses no fit solver. With primary mass bounds, an acceleration scale from 0.5 to 2.0 times 10^-10 m/s^2, and speed shifts of up to three listed errors, 47 curves have no allowed fit. A 5 km/s error floor added in quadrature reduces this count to nine. Wider mass bounds and a scale range from 0.1 to 10 reduce it to seven with that floor. All 725 galaxy-setting cases are resolved. The separate check passes 3,134 rejection intervals and all saved fit examples, including seven error-floor brackets narrower than 0.05 km/s. Five of the seven remaining cases have listed inclinations of at least 80 degrees. NGC5985 is the only quality-1 case below 80 degrees; its required added floor lies between 6.977 and 7.022 km/s under the wider bounds. These bounded fit results show the effect of the error model. They do not establish a new physical anomaly or exclude MOND as a whole. The package contains numerical inputs, source hashes, code, results, and figures.

SPARC · galaxy rotation curves · MOND · interval bounds · linear programming · numerical certificates · velocity errors · reproducible research