The stable isotope analysis became a major tool in soil food weds studies. A shift in δ15N value of 3.4‰ is generally suggested as a mean enrichment in 15N per trophic level. However trophic fractionation is not uniform across different animal taxa and different tissues, and varies with a range of factors, including quality and isotopic composition of the diet. We performed a laboratory experiment aiming to compare the trophic fractionation of nitrogen isotopes (Δ15N) in different Collembola species fed with three fungi of different palatability. Collembolans were fed with single-species fungal diet during 75 days. There was a positive correlation between the whole body C:N ratio and reproduction rates of collembolans. We therefore used the C:N ratio as a proxy of nutritional status. In all species of collembolans, the trophic fractionation decreased strongly with increased whole body C:N ratio, and this factor explained most of the within-species variation in Δ15N. In addition, the Δ15N decreased with increased δ15N in the diet. The analysis of published data on the trophic fractionation of collembolans in laboratory experiments confirmed these conclusions. The mean trophic fractionation in most collembolan species studied in the laboratory was fairly similar. In contrast, field studies have documented a consistent difference in δ15N among different families of collembolans. In particular, in nearly all published cases Onychiuridae had δ15N higher than Isotomidae or Entomobryidae. A sharp contrast between laboratory-based estimates of the trophic fractionation (generally uniform in different species and families) and field data (which show a consistent difference in δ15N among different species and even families of collembolans) confirms that δ15N values of field-collected animals bear important information on the trophic position occupied by a particular species, and is not heavily affected by the species-specific differences in trophic fractionation. As the Δ15N in collembolans depends on their nutritional status, we suggest that the C:N values should be reported along with isotopic data to allow cross-studies comparison.
doi:10.1016/j.pedobi.2010.10.004