Clear et al : Validation of annual increments in otoliths of Thunnus moccoyil 



29 



intensity of the band depends on the magnitude 

 of the difference in Z between the two portions of 

 the otohth. 



This kind of analysis requires a flat, polished 

 surface; therefore we sectioned the sagittal oto- 

 liths either along the postrostral axis to produce 

 an oblique longitudinal section (LS), or along 

 a transverse axis. The rostral axis often shows 

 clear increments, but we did not find distinct Sr 

 marks in this part of the otolith. The sections 

 were ground and polished following Gunn et al.'s 

 (1992) methods and an evaporated carbon coat 

 (25-30 nm thick) was applied to the sections to 

 minimize charging in the SEM. The position of 

 the Sr mark along the axis was measured with 

 the vernier attached to the SEM stage drives 

 (Fig. 2B). 



Energy-dispersive spectroscopy (EDS) In the 



later stages, an EDS x-ray microanalysis system 

 became available and we used it to confirm 

 the presence of Sr in the bright bands, and 

 also to detect Sr marks in unsectioned otoliths 

 (the use of unsectioned otoliths decreased the 

 preparation time required for SEM analysis ). The 

 system consisted of a Link 133 eV Si(Li) detector 

 with light element capability and a Thomson 

 Scientific "WinEDS" PC-based analyzer attached 

 to a Philips 515 SEM. Before x-ray analysis, 

 whole (unsectioned) otoliths were acid-etched 

 along the postrostral axis from the surface with 

 1 N and 3.5 N HCl, to expose the growth 

 plane, then rinsed in bleach and distilled water, 

 and dried. To minimize charging in the SEM, 

 the otoliths were dipped in a dilute carbon 

 DAG solution ( approximately 1:50 with dichloro- 

 ethane) immediately before analysis. Strontium marks 

 were detected by operating the SEM in "spot" mode 

 and searching for a point or zone where a significant 

 Sr signal was detected on the x-ray microanalyzer 

 (typically, a strong peak at 1.81 keV in the spectrum, 

 corresponding to emission of Sr La x-rays). When a 

 strontium mark was detected, its position along the 

 PR axis was measured and the mark photographed 

 either with conventional SEM photography (Fig. 4) 

 or with a rapid, low-grade video print, which also 

 showed the features of interest. To confirm that the 

 suspected mark was strontium-rich, two plots of the 

 x-ray spectra from the otolith were taken: one on 

 the strontium mark and one just before the mark. 

 Acceptable evidence of the correct identification of a 

 strontium mark was considered to be the presence of 

 an enhanced Sr level in the area analyzed, together 

 with an absence of Sr (except for background levels) 

 immediately before the area (Fig. 5). 



Figure 1 



The drilling technique used to extract otoliths from southern bluefin 

 tuna destined to be sold as "whole" fish. 



The measurements of increments on the whole 

 otoliths, and the strontium mark in sections or whole 

 otoliths, were made along the same axes without 

 reference to the other. This procedure enabled us to 

 compare the number of increments observed after the 

 strontium mark with the number expected, calculated 

 from the known time at liberty after tagging. 



Quantitative EDS analyses for linescans were car- 

 ried out on carbon-coated polished sections in the Phil- 

 ips 515 SEM operated at an accelerating voltage of 20 

 kV, by using a focused electron beam of 0.15 pm diam- 

 eter and analysis times of 60-200 seconds. The effec- 

 tive area analyzed by the beam was larger than the 

 diameter of the beam itself because the beam spreads 

 within the specimen after entry; examination of su- 

 perficial beam damage to specimens after analysis 

 suggested that the area analyzed by the beam is in 

 the order of 2 pm diameter. Elemental concentrations 

 were calculated by reference to appropriate standards 



