General relativity passes every local and strong-field test, yet on galactic scales we routinely add invisible components – dark matter halos – to explain flat rotation curves, the mass–discrepancy–acceleration relation (MDAR), and the baryonic Tully–Fisher relation (BTFR). This work explores a different possibility: that some of these regularities reflect a universal boundary response rather than galaxy-by-galaxy dark halos. Using the public SPARC sample of disk galaxies, I test a deliberately minimal “boundary gravity” law that maps the Newtonian acceleration from baryons, gbar, to a total radial acceleration g via a single global shape parameter β, with the acceleration scale a0 fixed to the canonical Milgrom value. Galaxies are split at the galaxy level into training and test sets: β is fitted only on the training galaxies and then held fixed when predicting the test galaxies. A like-for-like analysis is performed with a standard one-parameter MOND interpolating function on the same train/test split. The boundary law reproduces the SPARC MDAR with train/test scatter and chi-squared metrics comparable to MOND, yields a narrow deep-regime acceleration floor clustered around the fixed a0, and generates a tight BTFR with a physically reasonable slope, all without assigning individual dark halos. The accompanying Python scripts and derived tables provide a complete, reproducible pipeline so that this boundary-first interpretation can be checked, extended, or ruled out with independent implementations.
Grant Mark (Mon,) studied this question.