SPECIAL FEATURE: BIOMARKERS IN TROPHIC ECOLOGY Lake zooplankton δ13C values are strongly correlated with the δ13C values of distinct phytoplankton taxa

Analyses of carbon stable isotopes are often used to estimate the contributions of allochthonous and autochthonous dietary resources to aquatic consumers. Most pelagic food web studies assume that all phytoplankton taxa have a similar δ13C value. We studied pelagic food web compartments (dissolved inorganic carbon [DIC], phytoplankton, bacteria, seston, cladoceran zooplankton) in 12 small (< 0.1 km2) lakes in southern Finland. These lakes were classified as oligotrophic, mesotrophic, eutrophic, and dystrophic based on their concentrations of total phosphorus and dissolved organic carbon. Additionally, we studied phytoplankton photosynthetic carbon fractionation (εp) in laboratory conditions. The photosynthetic fractionation in 28 phytoplankton cultures from nine different phytoplankton classes varied significantly at the class level, and fractionation correlated significantly with the DIC concentration of the growth media. In small boreal lakes, the δ13C values of different phytoplankton taxa, as directly measured or estimated from the δ13C values of biomarker fatty acids, varied greatly (−18‰ to −44.5‰). Phytoplankton δ13C values varied significantly by lake type and were most depleted in dystrophic lakes even though the δ13C values of the DIC was similar to mesotrophic lakes. Further withintaxa variation was found between lakes and between different depths within a lake. Vertical samples from dystrophic lakes also showed lower εp in the phytoplankton from metaand hypolimnion, possibly as a result of reduced light intensity. Altogether, in nine of the 10 sampled lakes, the δ13C values of cladoceran zooplankton were between the minimum and the maximum phytoplankton δ13C value of each lake, and thus, phytoplankton alone could explain zooplankton δ13C values. We conclude that stable isotope mixing models should take into account carbon variation among different