Resumen
Ancient collagen is used as archive for multiple pre-mortem traits. Testing the quality of the collagen extract is a common concern of those who engage in the reconstruction of ancient diets. The aim of this study is to improve our understanding of the pyrolysis fingerprints of human bone collagen especially in relation with diagenetic alteration. Pyrolysis-GC-MS was applied to 28 collagen samples extracted from archaeological human bone, corresponding to different chronological periods (Bronze Age to post-Medieval period; 1900 BC–1800 AD) and different types of burial environment (acidic and alkaline) from NW Spain. Collagen was extracted following the common methodology used in paleodiet analysis, and a commercial gelatin sample was included for comparison. Data evaluation was based on 58 pyrolysis products using Principal Components Analysis (PCA). Principal component 1 (PC1, 45% of total variance) was related to the relative abundances of pyrolysis products of specific amino acids, with relatively degraded samples having larger proportions of the pyrolysis products of Pro/Hyp, Phe and Ala, while more intact samples showed larger proportions of Tyr, Trp and pyrolysis products of unspecific amino acid origin. PC1 scores were related to the period to which the samples corresponded, which reflects differences in diagenetic impact, probably controlled by a combination of age and burial deposit characteristics. PC2 (15%) probably reflects the well-known effects of disruption of the amino acid sequence (depolymerization), causing a decline in dimerization products (diketopiperazines) upon pyrolysis. This process was more intense in the collagen samples from acidic deposits than in the samples from alkaline deposits (a calcareous cave and coastal sand deposits with biogenic carbonates). The relationships between the PCA and individual pyrolysis products with known parameters of collagen quality (% C, % N, C/N ratio, % extractable collagen) were generally insignificant or weak. This might be explained by the rather narrow C/N range (3.19–3.36) of the samples, which had to meet the criteria for suitability for paleodiet analysis. Moreover, there was no significant relation between the isotopic composition of the extracted collagen (δ13C, δ15N) and pyrolyzate composition, suggesting that diagenesis has little effect on the isotopic fingerprints used in palaeodietary studies. Finally, no substantial contamination of microbial or exogenous tissue from the deposition environment to the osteological collagen extracts was identified. It is concluded that the δ13C and δ15N as proxies of palaeodiet from these diverse necropoleis in NW Spain is sustained.