138 
Fishery Bulletin 106(2) 
6.0 
5.8 
5.6 
5.4 
5.2 
5.0 
^ elssss Ocean 
r — i Lagoon 
2000 2001 
3.5 
3.0 
2.5 
2.0 
B 
i 
2001 
Year 
I 
2002 
Figure 2 
Mean (+1 standard error) increment width (pm) in oto- 
liths of neon damselfish ( Pomacentrus coelestis) (n = 9 
per treatment per year) for all years from ocean and 
lagoon treatments (A) before the start of the experi- 
ment and (B) during the experiment for each year of 
the experiment. Tukey’s HSD contrast groupings are 
indicated as letters (i.e., A, B) where overall differences 
were significant (P<0.05). Note the difference of the 
y-axis scale in A and B. 
3 0 A Ocean 
r — -i Lagoon 
2.5 ' 
2 000 2001 2002 
A B A B 
Year 
Figure 3 
Ba/Ca ratios from otoliths of neon damselfish ( Pomacen- 
trus coelestis) (n = 9 per treatment per year) for all years 
from ocean and lagoon treatments (A) before the start of 
the experiment and (B) during the experiment. Tukey’s 
HSD contrast groupings are indicated as letters (i.e., A, 
B) where overall differences were significant (P<0.05). 
Note the difference of the y-axis scale in A and B. 
had a higher mean increment width than the fish in the 
other two tanks (2.8 ,um vs. 2.4 pm and 2.4 am), but this 
did not obscure major differences between treatments. 
No significant difference in increment widths between 
treatment groups was detected for 2001. Finally, a sig- 
nificant difference in increment widths was detected 
among years for the ocean treatment group, but not for 
the lagoon treatment group (ANOVA: ocean: F 2 24 =16.56, 
P<0.05; lagoon: F 2 2 4=2.28, P>0.05). Fish in the 2002 
ocean treatment group had the widest increments of any 
fish in the experiments. Individual paired /-tests indi- 
cated that for both ocean and lagoon treatment groups 
there was a significant difference in increment width 
between the pre-experimental and the experimental 
increments (paired /-test: ocean: T 24 ,,=90.45, P <0.05; 
lagoon: T 242 = 3 2 2.05, P<0.05). 
There were significant relationships between mean 
increment widths and Ba/Ca ratios for two of the three 
experimental years (2000: Pearson correlation coef- 
ficient, r-0.644, P<0.01; 2001: r=0.052, P>0.05; 2002: 
r=0.671, P<0.01; Fig. 4). There was no significant cor- 
relation between increment widths and Sr/Ca ratios 
for any of the experimental years (2000: Pearson coef- 
ficient, r=-0.417, P>0.05; 2001: r=0.033, P> 0.05; 2002: 
r=0.266, P>0.05; Fig. 4). 
SEM indicated that the aragonite crystal-lattice with- 
in experimental otoliths was orientated perpendicularly 
to daily increments (Fig. 5). This pattern was consistent 
for ocean and lagoon treatments (angle from daily incre- 
ments: ocean: 90° ±0°; lagoon: 90° ±0°) and control fish 
(90° ±0°). Variation in increment width could only be 
explained by variation in the length of the intra-incre- 
mental lattice rather than by a change in orientation 
or crystal packing. 
Discussion 
The pattern of wider increments observed for experimen- 
tal P. coelestis held in ocean waters was consistent with 
