Electron probe microanalysis of fish otoliths — evaluation of techniques for studying age and stock discrimination

Abstract

Electron probe microanalyzers fitted with energy-dispersive (ED) and wavelength-dispersive (WD) X-ray spectrometers are the standard instruments for measuring microscale differences in the chemical composition of the calcified tissues of fishes. We review the principles of using these spectrometers, detail procedures for preparing otoliths for microanalysis, and compare the accuracy and sensitivity of the two instruments. ED systems on electron probes are prone to produce spectral artifacts, peak overlaps and difficulties in modelling nonlinear backgrounds. Current generation ED systems do not, therefore, appear suitable for quantitative analysis of trace elements (<5000 ppm) in otoliths. For the species we examined, WD systems could measure six elements (Ca, Na, Sr, K, S and Cl) accurately and precisely to levels of several hundred ppm. However, electron probes operating at the beam powers required for WD analysis damage otoliths, causing pitting and chemical change, which increases measurement errors. This damage and the consequent loss in data quality vary directly with the power density of the electron beam, which, in turn, determines in part the rate of data acquisition. Use of a WD spectrometer to analyze otoliths, therefore, requires making a trade-off between acquisition time and data quality, which has to be evaluated against the magnitude of natural variability in the composition of the otoliths of each species studied. In the species we examined, we concluded that beam power densities >3 μW · μm −2 resulted in unacceptable levels of specimen damage.