Date of Award

Spring 2004

Document Type


Degree Name

Doctor of Philosophy (PhD)


Biological Sciences


Ecological Sciences

Committee Director

Cynthia M. Jones

Committee Member

Robyn Hannigan

Committee Member

Kent E. Carpenter


A quantitative understanding of habitat use of estuarine-dependent fishes is critical to the conservation of their most essential habitats. Because recruitment and fitness may be influenced by the quality of juvenile habitats, developing methods to quantify habitat-specific survivorship is pivotal to such understanding. An initial step to quantify survivorship is to validate the habitat-specific natural tags contained in otoliths. To this aim I investigated the variability in the chemistry of surface waters and otoliths of juvenile spotted seatrout, Cynoscion nebulosus, in five seagrass habitats of Chesapeake Bay, namely: Potomac, Rappahannock, York, Island, and Eastern Shore. I measured Mg, Ca, Mn, Sr, Ba, and La in water and otoliths by inductively coupled plasma-mass spectrometry, and δ13C and δ18O in otoliths using an automated carbonate analyzer. Multivariate analyses of variance and regressions were used to test the hypothesis that otoliths accurately record the chemistry of natal habitats of juveniles, whereas cross-validated k-nearest neighbor functions were derived to discriminate habitats based on water and otolith chemistry. Concentration of Mg, Mn, Sr, and Ba in water was significantly different between habitats independent of temporal variation. Classification accuracy of water samples was low in Rappahannock (37%), moderate in Potomac and Eastern Shore (60–70%), and high in York and Island (81–82%) habitats. Weighted regressions showed that salinity could predict accurately [Ba/Ca]otolith and [La/Ca]otolith. There was a positive correlation between [Ba/Ca]otolith and [Ba/Ca]water, but the relation was not linear as previously found in laboratory experiments. Contrary to expectation, [Sr/Ca]otolith did not correlate with water chemistry, however there was a predictive relation between [δ18O]otolith and [Sr/Ca]water. Otolith microchemistry of juveniles collected in 1998 and 2001 was significantly different among habitats within and between years, but the ability of trace elements to allocate individual fish to natal habitats was variable (0–82%). However, the combination of [Ba/Ca], [Mn/Ca], with δ13C and δ18O in 2001 otoliths significantly improved classification rates, allowing 82–100% accuracy. These results showed that otolith microchemistry might be useful in identifying specific seagrass beds in Chesapeake Bay, with the potential of being used as natural tag to quantify survivorship and to determine essential habitats for juvenile spotted seatrout.