Controls on the nitrogen isotopic composition of fish otolith organic matter: Lessons from a controlled diet switch experiment

Abstract

The nitrogen isotopes (δ15N) in the organic fraction of accretionary hard part structures, such as fish otoliths, may provide life histories of dietary change. We performed controlled experiments to validate the dynamics of the isotopic signal incorporation into biominerals following dietary shifts and also compared whole-otolith and serial sampling approaches for diet reconstruction. Laboratory-reared Atlantic croaker (Micropogonias undulatus) were switched from a high quality, high δ15N diet (δ15N = 10.7‰) to one of two lower quality, lower δ15N diets (δ15N = 5.9 or 3.7‰). Using the oxidation-denitrifier method, including cleaning protocols required for fossil otoliths, we measured both otolith-bound δ15N (δ15Noto) of whole-otolith subsamples (δ15Nwhole) and sequentially micromilled otolith powders (δ15Nmicro) and compared these results to white muscle tissue and liver δ15N obtained through traditional techniques. Both δ15Nwhole and δ15Nmicro recorded the diet switch, although sampling limitations muted the signal in δ15Nmicro, especially in the slowly growing otoliths of the fish on the lowest quality diet, in which otolith accretion was particularly slow. The timescales at which otoliths and tissues approached the new δ15N after the diet switch varied, but the slowest was for muscle and the fastest was for liver. For δ15Nwhole, there were two factors: (1) the turnover time of the N provided to the otolith and (2) that the otolith is accreting and thus integrating over the entire life history of the fish. Using a model to account for the accretionary growth, turnover time for N supplying otolith growth ranged from 12 to 16 days. In both soft tissues and otoliths, the trophic discrimination factor (the δ15N elevation relative to diet) appears to have been lower in the larval fish, prior to the onset of the diet switches. This raises questions about the interpretation of the δ15Noto of otolith cores; however, the core is typically a minute portion of the total otolith. We conclude that (1) δ15Nwhole is useful for tracking diet despite the whole-life integration of the δ15N signal and (2) δ15Nmicro also records diet, although higher resolution tracking will require further optimization of δ15Nmicro sampling.