Kingsford et al.: The influence of elemental chemistry on the widths of otolith increments in Pomacentrus coelest/s 
139 
3.6 
3.4 
3.2 
3.0 
2.8 
2.6 
2.4 
2.2 
2.0 
1.8 
* 
* 
2000 Ocean 
2000 Lagoon 
2001 Ocean 
2001 Lagoon 
2002 Ocean 
2002 Lagoon 
2 3 
Ba/Ca (pmol/mol) 
Figure 4 
Relationship (mean +1 standard error) between increment width 
(pm) in otoliths of neon damselfish ( Pomacentrus coelestis ) 
and elemental ratios for all experimental years for each treat- 
ment group (n = 9 per treatment per year) for (A) Ba/Ca and 
(B) Sr/Ca. 
the pattern observed for wild fish. Additionally, 
these experiments demonstrated that increment 
widths could be altered in the absence of food and 
temperature differences. There are a number of 
factors that can potentially influence the incre- 
ment widths of otoliths, such as feeding regime 
(McCormick and Molony, 1992), lipid reserves 
(Paperno et al., 1997; Molony and Sheaves, 1998), 
the onset of metamorphosis (Wilson and McCor- 
mick, 1999), water temperature (Powell et al., 
2004), and metabolic rate (Mosegaard et al., 
1988; Wright et al., 2001). Temperature and food, 
the most well-studied factors, were not relevant 
here because both were held constant during the 
experiments. Additionally, metabolic rates were 
expected to be consistent among individuals, 
given the similar size, age, and the constant tem- 
perature and feeding regimes of these individu- 
als. It is also unlikely that metamorphosis played 
a role; experimental fish immediately “settled” in 
the tanks. Settlement marks were noted to form 
with narrower post-settlement increments, consis- 
tent with what is known for this species (Wilson 
and McCormick, 1999). However, these narrower 
increments were unlikely to have affected the 
results as they occurred in all fish in the experi- 
mental treatments over all years. 
It is possible that the differences noted in in- 
crement widths were related to initial size or the 
condition of the fish during the presettlement 
phase. Although the standard length of the fish 
was measured after the study and no differences 
were found, it was not possible to measure fish 
before the study (see Methods) and more explicit 
measures of condition (e.g., Fulton’s K\ Hoey and 
McCormick, 2004) were not applied before the 
experiment. Therefore, pre-existing differences 
in size or condition may have influenced the re- 
sults to some degree (e.g., some fish may have 
started the experiment with higher lipid reserves 
which could have influenced their growth rate 
and increment widths irrespective of experimen- 
tal conditions). The randomized allocation of fish 
between treatments on three separate occasions, 
however, makes this scenario highly unlikely. In 
addition, any size or conditional differences were 
more likely to have introduced variation within the 
treatment groups rather than the between treatment 
differences observed in this study. 
Water chemistry may have influenced increment 
width in an indirect manner. Late-stage reef fish have 
well-developed olfactory systems and are capable of 
detecting chemical differences of ocean and lagoon wa- 
ters found around OTI (Atema et al., 2002). Some taxa 
prefer lagoon water (e.g., apogonids), whereas it has 
been shown that P. coelestis actively avoid settling in 
the lagoon and prefer the coral rubble habitat found on 
reef slopes (Doherty et al., 1996). Indeed, P. coelestis 
were relatively rare in the lagoon at OTI, and this 
finding confirms the previous report of Doherty et al. 
(1996). Although not well-studied in reef fish (but see 
Arvedlund et al., 1999 and Gerlach et al., 2007), natal 
imprinting and homing has been documented in fish 
species, most notably in the olfactory-driven homing of 
salmonids (Dittman et al., 1996). Because P. coelestis 
are likely spawned and undergo their presettlement 
life outside of lagoon habitats, they may have imprinted 
on nonlagoon waters. Lagoon water, therefore, may be 
an indication of substandard habitat that P. coelestis 
actively avoid. Holding fish in lagoon water may have 
triggered a physiological response (e.g., stress), result- 
ing in reduced growth and narrower otolith increment 
widths for fish held in nonlagoon waters (e.g., Marchand 
et al., 2003). However, this seems unlikely to have been 
