422 



Fishery Bulletin 90(2), 1992 



the otolith formed under natural con- 

 ditions (inside the OTC band) were 

 analyzed. 



An otolith from a randomly selected 

 fish was mounted on a slide with a tolu- 

 ene-based medium. It was ground using 

 600-grit paper along the saggital plane 

 to a level near the central primordium. 

 The mounting medium was then melted, 

 the otolith was removed, washed, and 

 remounted on its opposite side with 

 heat-setting epoxy. The second side was 

 then ground to the central primordium 

 and polished with a series of diamond 

 and alumina grits, ending with 0.05^m 

 alumina. The specimen was cleaned 

 ultrasonically in detergent and water 

 between grit changes and given final 

 rinses in water and methanol. Prior to 

 microprobe analysis, the specimen was 

 carbon coated. 



A second purpose was to determine the effect of 

 beam exposure on the constancy of Sr/Ca ratios. This 

 was necessary because analytical techniques, such as 

 increasing the counting time, will improve the preci- 

 sion of an analysis but may reduce its accuracy through 

 beam damage to the specimen (e.g.. Smith 1986, Potts 

 1987). This problem is encountered in the analysis of 

 other carbonates, but is particularly severe for otoliths, 

 which contain organic material in addition to CaCOs 

 (Degens et al. 1969). CO2 is lost during electron beam 

 exposure and, because it is not actually measured by 

 the microprobe but assumed to occur on a 1:1 basis 

 with cations such as Sr and Ca, concentrations of 

 those elements wall increase with increasing beam 

 damage. However, if Sr and Ca are not fractionated 

 from one another by beam damage, their ratio should 

 remain unchanged. Absence of change would indicate 

 that methods which improve precision can be imple- 

 mented without affecting the accuracy of Sr/Ca ratio 

 determinations. 



Methods 



Dover sole Microstomus pacificus is a common Pacific 

 coast flatfish. Juvenile Dover sole 54-104 mmSL were 

 captured by trawling off the Oregon coast on 17 March 

 1990 and immediately injected with oxytetracycline 

 (OTC). Within 12 hours, fish were transferred to aqua- 

 ria in Corvallis, Oregon, where they were held for up 

 to 48 days. The OTC produced a fluorescent band which 

 delineated growth prior to capture from subsequent 

 grovrth under laboratory conditions. Only portions of 



Beam power density and precision 



Wavelength-dispersive electron microprobe analysis 

 was performed with a Cameca SX-50 microprobe with 

 a 40° beam angle. Three levels of beam-power density 

 were obtained by varjang the beam diameter while 

 holding accelerating voltage and beam current constant 

 at 15 KV and 20 nA, respectively. These voltage and 

 current settings are common to most of the previous 

 studies (R. Radtke, Hawaii Inst. Geophys., Univ. 

 Hawaii, Honolulu 96822, pers. commun. 1990), al- 

 though Kalish (1989) used a lOnA current. Defocused 

 beam diameters of 5, 7, and lO^^m resulted in beam- 

 power densities of 1.019, 0.520, and 0.255 nA/fim. The 

 most common beam diameter used in previous studies 

 was 5^im (R. Radtke, pers. commun. 1990), although 

 Kalish (1989) rastered a 12.5fimx 12.5jim square. 



Sr and Ca concentrations were calculated as nor- 

 malized mole fractions (equivalent to the atomic ratios 

 of Kalish 1989). Mole fractions are more informative 

 than weight percentages for examination of Sr/Ca 

 ratios, since the substitution of Sr for Ca in otolith 

 aragonite theoretically occurs on a per-atom basis (e.g., 

 Radtke 1989). Normalization also reduces effects of 

 beam damage on concentrations. 



Precision of elemental measurements was deter- 

 mined as the coefficient of variation (CV) (Williams 

 1987), 



CV = Ok.ratio^k-ratio 



where the k-ratio is the ratio of x-ray counts from the 

 otolith to those of the standard (i.e., the calibrated frac- 

 tion of that element in the otolith) and o^.ratio is the 



