A series of small fine-grained and single-crystal bars, with strain from 0% (recrystallized) to 50%, were given different amounts of chemical polishing. Four-point resistivity () data was used to characterize the electron scattering from dislocations, hydrogen, and any other trace contaminants. As noted by previous studies, annealed Nb displayed a weak linear increase of (11 K) with polishing time due to hydrogen absorption, and bulk hydrogen concentration did not exceed 15% for 200 µm metal removed. Cold-worked samples displayed steeper slopes with polishing time (after subtracting resistivity due to strain alone), suggesting that dislocations assist the absorption of hydrogen during polishing. Absorption accelerated above 30% strain and 100 µm material removal, with room-temperature hydrogen concentration rising rapidly from 2% up to 5%. This threshold is significant, since superconducting radio-frequency (SRF) cavities are usually polished as-formed, with >35% strain, and polishing removes >150 µm of metal. Resistance jumps between 40 and 150 K, which signal the formation of hydride precipitates, were stronger in cold-worked samples, suggesting that dislocations also assist precipitate nucleation. High-vacuum anneals at 800 °C for 2 hours, which are known to fully recrystallize cavity-grade niobium and de-gas hydrogen, removed the 40-150 K jumps and recovered the resistivity increase due to chemical polishing entirely. But, about 30% of the resistivity increase due to cold work remained, possibly due to residual dislocation clusters. Continued annealing only facilitated the diffusion of surface impurities into the bulk and did not recover the initial 0% state. Strain, polishing, and annealing thus appear to combine as irreversible paths that change the material. Bearing this in mind, the significant difference in hydrogen uptake between annealed and coldworked samples suggests that annealing SRF cavities prior to chemical polishing could greatly reduce hydrogen uptake and storage in the metal, reducing risk of quality-factor loss. This inverts key steps of the present widely-used cavity processing sequence.