Stable isotope proxies measured in the proteinaceous fraction of archaeological mollusc shell represents an increasingly important archive for reconstructing past ecological and biogeochemical conditions of nearshore environments. A major issue, however, is understanding the impact of diagenetic alteration in sub-fossil shell isotope values. “Bulk” stable isotope values of nitrogen (δ15N), and especially carbon (δ13C) often shift strongly with increasing C/N ratios in degraded shell, resulting in unreliable data. Here, we examine preservation of an entirely new set of shell paleo-proxies, compound-specific isotopes of amino acids (CSI-AA). We examine carbon (δ13CAA) and nitrogen (δ15NAA) patterns and values from the organic fraction of California mussel (Mytilus californianus) shells from the California Channel Islands. Archaeological shell samples ranging in age from ca. 6,100 to 250 cal BP exhibiting a wide range of degradation states were collected from varied depositional environments (e.g., exposed coastal bluff, buried strata, etc.), and were directly compared to modern shells of the same species and region.Our results indicate organic matter C/N ratios as the best bulk diagnostic indicator of the relative degradation state of shell organic fraction, including changes at the molecular level. Modern shell organic C/N ratios ranged from 2.8 to 3.5, while those in archaeological shell were substantially elevated (3.4–9.5), exhibiting strong and significant negative correlations with bulk δ13C values, weight %C, and weight %N, and a significant but weaker correlation with δ15 N values. An additional “cleaning” step using weak NaOH helped to remove possible exogenous contaminants and improved bulk values of some samples. However, relative molar AA abundances revealed that some AAs, especially the two most abundant, Glycine and Alanine, progressively decreased with increasing C/N ratio. The loss of these amino acids permanently alters bulk isotope values regardless of removal of contaminants. Modeling the bulk isotope change expected due to amino acid molar composition showed major and predictable shifts in bulk δ13C values from selected AA loss, and similarly large but far more variable impacts from exogenous contaminants.In contrast to bulk data, key CSI-AA values and patterns remained almost entirely unaltered, even in the most degraded shell samples, closely matching expected biosynthetic isotope patterns in modern mussel shell. AA isotope proxies for “baseline” (δ15N-Phenylalanine and average δ13C-Essential AAs) and planktonic trophic structure (δ15N-Glutamic Acid and δ15N-Phenylalanine) were not statistically altered with degradation in any sample. Overall, we conclude that while bulk isotopes, particularly δ13C, are very likely to be unreliable in archaeological or subfossil shell with C/N ratios higher than ∼4.0, CSI-AA proxies can still be used to reconstruct past climatic and ecological conditions of the nearshore marine environment.