Abstract. — The chemistry of 

 calcified tissues has been suggested 

 as a source of useful information 

 on the population structure and 

 environmental histories of fishes. 

 We evaluated this possibility by 

 examining in detail regional and 

 ontogenetic variability in the 

 chemical composition of sagittae of 

 juveniles and adults of the temper- 

 ate marine groundfish Nemadac- 

 tylus macropterus . Six elements (in 

 order of decreasing abundance, Ca, 

 Na, Sr, K, S, and CI) were consis- 

 tently detected in the sagittae at 

 concentrations greater than 200 

 ppm; all exhibited levels of indi- 

 vidual, ontogenetic, and regional 

 variability well in excess of their 

 respective scales of measurement 

 error. Comparisons of juveniles and 

 adults from different sites indicate 

 that composition of the otolith is 

 most alike in fish from adjacent 

 sites, that most juveniles are simi- 

 lar to adults collected from the 

 same site, and that the differences 

 in composition that characterize 

 sites are manifest during most, if 

 not all, of the fish's ontogeny. These 

 results are consistent with the hy- 

 potheses that otolith composition 

 reflects population structure and 

 that these populations are largely 

 self-recruiting. However, the re- 

 sults also suggest that the chemi- 

 cal composition of otoliths is much 

 less sensitive to environmental con- 

 ditions than previously thought. 

 Rather, it appears that regional dif- 

 ferences in composition either have 

 a genetic basis or are set by envi- 

 ronmental influences early in life 

 and are then maintained throughout 

 subsequent life history. 



An evaluation of electron-probe 

 microanalysis of otoliths for stock 

 delineation and identification of 

 nursery areas in a southern 

 temperate groundfish, 

 Nemadactylus macropterus 

 (Cheilodactylidae) 



Ronald E. Thresher 

 Craig H. Proctor 

 John S. Gunn 



CSIRO Division of Fisheries 

 POBox 1538 

 Hobart, Tasmania 

 Australia 7001 



Ian R. Harrowfield 



CSIRO Division of Mineral Products 

 ROBox 124 



Port Melbourne, Victoria 

 Australia 3207 



Manuscript accepted 11 April 1994. 

 Fishery Bulletin 92:817-840. 



Knowledge of geographic structure 

 is fundamental to understanding 

 the dynamics of marine fish popu- 

 lations (e.g. Sinclair, 1987). None- 

 theless, even for the small number 

 of species thus far investigated, 

 there remains considerable uncer- 

 tainty regarding population struc- 

 ture. This is due to the lack of a 

 widely applicable, direct means of 

 mapping how far and in what direc- 

 tions larvae disperse. A variety of 

 indirect techniques have been ap- 

 plied to the problem, e.g. modelling 

 larval advection from oceano- 

 graphic features, analyzing parasite 

 loads, mapping phenotypic charac- 

 ters, and locating and enumerating 

 discrete spawning areas. However, 

 all are limited in scope and in the 

 strength of the inferences that can 

 be drawn from them. For this rea- 

 son, the geographic structure of 

 marine populations is usually in- 



ferred from genetic studies (e.g. 

 Avise et al., 1987; Waples and Ros- 

 enblatt, 1987; Smith et al., 1990). 

 But genetic techniques are also far 

 from ideal for this task: they will not 

 detect differences in the face of even 

 low levels of larval or adult mixing 

 among populations (Hartl and Clark, 

 1989); they cannot directly measure 

 rates of individual exchange among 

 sites or, usually, specify the origin 

 of individuals; and results from 

 even a successful study can be am- 

 biguous, i.e. a genetic difference 

 between sites suggests little dis- 

 persal, but the lack of any difference 

 is largely uninformative. Such dif- 

 ficulties have prompted continuing 

 research into alternative, and per- 

 haps more definitive, techniques for 

 evaluating population structure. 



One alternative technique is the 

 analysis of the chemical composi- 

 tion of calcified structures. As early 



817 



