664 



Fishery Bulletin 88(4), 1990 



keV) photons by the siHcon detector of an ED spec- 

 trometer (Reed and Ware 1972). The energy of the 

 photon producing the escape peak is E^ - Esi, where 

 Ex is the incident photon energy and Egj is the energy 

 of a Si Ko photon (1.739 keV). The silicon escape peak 

 due to Ca K^ would have an energy of 1.951 keV 

 (3.690 keV - 1.739 keV) and an intensity 0.78% that 

 of the Ca K„ peak (Reed 1975). Such a peak, within 

 less than 150 eV of the Sr peak, would create signif- 

 icant errors in the estimation of low levels of Sr. Mea- 

 surements of trace levels of phosphorus with the K„ 

 line (2.013 keV) would be even more vulnerable to 

 errors because of the closer proximity to the Ca K„ 

 derived silicon escape peak. Other errors can result 

 from a silicon internal fluorescence peak (Reed and 

 Ware 1972) and the silicon absorption edge (Goldstein 

 et al. 1981, Statham 1981), both artifacts of the ED 

 spectrometer and associated silicon detector, and of 

 particular interest to the determination of trace levels 

 of Sr. These sources of error are not present when 

 using a WD electron microprobe. 



There are numerous other potential sources of error 

 in otolith trace-element studies carried out with ED 

 electron microprobes. These errors can occur due to 

 the relatively poor energy resolution of the ED spec- 

 trometer and the resultant inability to discriminate 

 between X-ray lines separated by less than approx- 

 imately 150 ev (Statham 1981). Furthermore, the pres- 

 ence of a greater proportion of continuum background 

 radiation or "bremsstrahlung" in the ED spectrometer, 

 due to decreased energy resolution, results in a five- 

 fold increase in the peak to background ratio in a WD 

 spectrometer over an ED spectrometer (Reed 1975). 

 These factors, and others, make it necessary to view 

 with caution studies with ED microprobes that inves- 

 tigate the distribution of large numbers of trace (<0.5% 

 wt) elements, particularly without reference to criteria 

 for determining detection levels or precision. In an in- 

 vestigation of otolith composition using X-ray fluores- 

 cence spectroscopy (XRF) Cu, Cd, Cr, and V were 

 below the detection limits of the instrument (<3.0 

 ppm), Ni was found at levels averaging 2.0 ppm, and 

 Fe, Zn, and Ba levels were below 10 ppm in the four 

 species studied (J.M. Kalish, unpubl. data). Edmonds 

 et al. (1989) used inductively coupled plasma atomic- 

 emission spectroscopy (ICP— AES) to study otolith 

 composition of the pink snapper Chrysophrys auratus 

 for stock discrimination, and their data indicate that 

 Mg, Si, and Fe are at levels below the detection limits 

 achievable using either ED or WD microprobe analysis. 



With the above results in mind it appears that with 

 a WD electron microprobe the only elements that can 

 be reliably quantified in otoliths are Ca, Na, Sr, K, S 

 and, in some cases, CI. However, in the case of CI, the 

 presence of this element in the most frequently used 



mounting medium, epoxy, makes the quantitative 

 determination of this element difficult. If it is desired 

 to detect CI, an alternative mounting medium should 

 be considered. The utility of ED microprobe analysis 

 to studies of the quantitative composition of fish oto- 

 liths seems to be limited to Ca and, in some rare in- 

 stances, Na and Sr. 



Although WD microprobe analysis is generally cap- 

 able of detecting the levels of Sr found in freshwater 

 and marine fish otoliths, there are some limitations to 

 the method in studies of anadromous fishes. Most im- 

 portant is the limited spatial resolution of the micro- 

 probe. In this study the probe size was maintained at 

 10 X 10 ^m to minimize beam damage of the specimen. 

 A spot of this size on the otolith of an adult fish would 

 encompass more than a single day of otolith growth 

 and in slow-growing individuals could encompass a 

 month or more. In slow-growing fish, this would limit 

 the ability to determine the temporal scale of migra- 

 tory events. Determination of the temporal resolution 

 of the microprobe data would be dependent on esti- 

 mates of fish age based on otolith annuli and micro- 

 increment data. In many instances, the utility of otolith 

 Sr data is dependent on the accurate determination of 

 age using the same otolith. 



The elemental analysis of otolith primordia should 

 provide an objective criterion for assessing if an indi- 

 vidual is the progeny of an anadromous or non-anad- 

 romous female. This information alone, or in combina- 

 tion with data from life-history transects, should make 

 it possible to investigate the relationships between 

 genetics and environment on diadromous behavior and 

 aid in the management of species that display facula- 

 tive diadromy. However, as with any new method, it 

 is important to confirm the validity of these results for 

 the particular species and habitat in question. 



Acknowledgments 



I thank Jane Andrew (Inland Fisheries Commission, 

 Tasmania) for rearing the rainbow trout and supply- 

 ing trout ova, and Stephanie Kalish, Peter Davies, and 

 Laurie Cook, also of the Inland Fisheries Commission, 

 for providing salmonids and ova from various localities. 



Citations 



Alderdice, D.F. 



1988 Osmotic and ionic regulation in teleost eggs and larvae. 



Ill Hoar. W.S., and D.J. Randall (eds.l, Fish physiology, vol. 



XI. part A. p. I(i3-li.")l. Academic F'ress. San Diego. 

 Bagenal, T.B.. F.J.H. MacKereth, and J. Heron 



1973 The distinction between brown trout and sea trout by the 



strontium content of their scales. .1. Fish. Biol. .5:.5,55-.'5.57. 



