Thresher et al.: Otolith analysis of Nemadactylus macropterus 



837 



Although the data are obviously preliminary, these 

 mismatched individuals could be direct evidence of 

 an exchange of individuals among populations dur- 

 ing the larval stage. The apparent exchange rate 

 varies depending upon the criterion selected and may 

 well differ with site. Even a conservative estimate 

 for the Tasmanian samples (i.e. defining migrants 

 as individuals with a probability >90% of not being 

 derived from the Tasmanian adults) nonetheless sug- 

 gests an exchange rate of about 7-8%, which is high 

 enough to prevent genetic divergence among samples 

 from the NSW, Victorian, and Tasmanian sites 

 (Elliott and Ward, 1994). 



Determinants of otolith composition 



The working hypothesis underlying our approach is 

 that otolith composition is largely determined by 

 environmental factors that presumably differ at rela- 

 tively fine space scales. The data to support this en- 

 vironmental sensitivity, however, are not abundant 

 and to a large extent are drawn from the inverte- 

 brate literature (e.g. Rosenberg, 1980; Schneider and 

 Smith, 1982). Studies on teleosts are ambiguous. To 

 date, all reported effects have involved Sr, which has 

 been reported as sensitive to changes in salinity 

 (Radtke et al., 1988; Kalish, 1990; Secor, 1992) and 

 temperature (Radtke et al., 1990; Townsend et al., 

 1992; however, see Kalish, 1989; and Gallahar and 

 Kingsford, 1992). 



There are two reasons to suspect that most of the 

 elements detected in our study are less responsive 

 to the environment than is widely assumed. First, 

 most are physiologically important and their concen- 

 trations tightly regulated in plasma and hence pre- 

 sumably in endolymph (Kalish, 1991). For example, 

 an expectation that relatively slight changes in sa- 

 linity significantly affect the incorporation of Na and 

 CI in otoliths is unrealistic in an animal with well- 

 developed osmoregulatory mechanisms. Of the six 

 elements detected, only Sr is likely to be relatively 

 unaffected by such physiological controls, though it 

 is presumably affected by many of the same factors 

 that constrain variation in Ca concentrations and 

 may well be subject to a suite of other physiological 

 constraints (see Kalish, 1991). 



Second, our data are not consistent with a strong 

 and direct effect of the environment on composition. 

 Two observations are particularly relevant: 1) settle- 

 ment into nursery areas had no apparent effect on 

 otolith chemistry, other than a slight effect on Sr, 

 and 2) differences among regional groupings are 

 manifest from the primordium to nearly the otolith 

 margin and hence were apparently unaffected by life 

 history stage, irrespective of habitat occupied. Re- 



garding the transition to the nursery areas, for ele- 

 ments other than Sr there was no indication of con- 

 vergence on a common chemical phenotype by indi- 

 viduals in a given nursery area, nor evidence of di- 

 vergence among nursery areas in response to local 

 conditions. This suggests that the concentrations of 

 5 of the 6 elements we measured do not vary in re- 

 sponse to environmental conditions in the nursery 

 areas in any direct way. The nursery areas sampled 

 ranged from mid-shelf to shallow coastal embay- 

 ments and differed markedly in temperature and 

 salinity histories, water-column chemistry, depth, 

 substratum, turbidity, and in invertebrate composi- 

 tion (and hence presumably in the diets of the juve- 

 niles). The apparent lack of an impact of any of these 

 on otolith composition suggests their effects at the 

 >100 ppm level are weak or indirect (or both), except 

 possibly for effects on Sr. Similarly, the consistency 

 of regional differences in concentrations through life 

 suggests these differences are largely unaffected by 

 changes in habitats that range from high seas nekton 

 to coastal embayments. Although the concentrations 

 of several elements ( Sr, Na, K, and S ) clearly vary on- 

 togenetically in otoliths, this variation is superimposed 

 on, and apparently separate from, whatever determines 

 regional differences in composition. 



The causes of the regional 'base' differences in 

 composition are not clear. There are several broad 

 possibilities: 



• The chemical phenotype is modified retrospec- 

 tively, based on the adult habitat or sample prepa- 

 ration; for reasons discussed above, we think this 

 mechanism unlikely; 



• Life cycles for each region are closed within areas 

 of a uniquely diagnostic environment. This seems 

 unlikely given the diversity of habitats occupied 

 by the species during its life history, but cannot be 

 rejected until the factors that affect otolith compo- 

 sition are determined; 



• The base composition is ontogenetically set by en- 

 vironmental influences early in the larva's life, and 

 then maintained, although overlayed by ontoge- 

 netic modification, throughout its subsequent life 

 and environmental history; 



• The base composition is determined genetically. 



The information currently available is not suffi- 

 cient to discriminate between a 'locked phenotypic 

 effect' (#3) and a genetic hypothesis (#4). A key da- 

 tum that would permit such discrimination is a mea- 

 sure of year-class effects on otolith composition. Re- 

 gional differences in otolith composition that vary 

 among year classes argue against a genetic basis and 

 for an environmental influence early in larval devel- 



