MISSING CONTRIBUTIONS AND PRINTED ERRORS IN THE VMS FRAMEWORK: TWENTY-FIVE RESULTS WITH MACHINE-CHECKED PROOFS

This paper reports twenty-five new mathematical results on the VMS working-paper record by V. Waters (Zenodo, doi:10.5281/zenodo.17239587). It does not repeat the findings of my two earlier releases (doi:10.5281/zenodo.23624777 and doi:10.5281/zenodo.23656880). Ten results are printed errors, where a table entry, formula or limit contradicts a definition or another equation in the same record. For example: the Chemical Elements Table’s mass-excess column disagrees with its mass column in 117 of 118 rows; the molecular Coulomb term has the wrong sign; the large-A binding limit drops a bulk asymmetry term; the single-ray optical invariant is not conserved by the record’s own transfer matrices. The other fifteen are conditional obstructions. Each is shown with an explicit stable model, and the paper states which repairs still work. The paper is self-contained. The accompanying Lean 4 file has 523 theorems that compile with no errors or warnings and use only standard axioms, and every result was also re-derived with exact arithmetic. None of the results claims that every possible completion of VMS is inconsistent.