Sample preparation effects on stable C and N isotope values: a comparison of methods in Arctic marine food web studies

We compared effects of different sample preparation techniques on the stable carbon and nitrogen isotope values of fish muscle, whole crustaceans and particulate organic matter (POM). Comparisons were also made to untreated (i.e. dried and homogenised) samples. Relatively carbonate-rich samples treated with weak acid (0.1 N HCl), either by quickly wetting or acid dampening, were on average 1.3‰ more enriched in 13C than duplicates soaked in 2 N HCl for 5 min, indicating incomplete carbonate removal with the weaker acid. In comparison, no differences in δ15N values were found between acid treatments, and a following water rinse had no effect on the δ13C or δ15N values. Chloroform-methanol (2:1 by volume) extraction overnight removed less lipids than Soxhlet extraction with 7% methanol in dichloromethane for 2 h, resulting in ~1.2‰ difference in δ13C values between treatments of lipid-rich duplicates. Different lipid-extraction methods did not lead to consistent differences in δ15N values, however. Depending on the lipid and carbonate content, untreated samples were depleted in 13C by 0.8 to 4.4‰ and in 15N by 0.6 to 1.4‰ compared to treated duplicates. We conclude that δ13C and δ15N values of samples with low lipid and carbonate content are highly comparable among studies regardless of pre-treatment methods, whereas the δ13C values of relatively lipidand/or carbonate-rich samples must be carefully considered based on the pretreatment applied to samples. In comparison, δ15N values are relatively robust to differences in carbonate and lipid-removal methods, and δ15N values of untreated vs. carbonateand lipid-treated samples are comparable within ±1.0‰.