Paperno et a I.: Daily growth increments in otoliths of juvenile Cynoscion regalis 
523 
weight of mysids consumed (weight of mysids offered 
minus weight of mysids not eaten) and estimated fish 
weights (assuming exponential growth between 
weighing). Mean daily ration was calculated for each 
feeding level treatment (Table 1). Henceforth, feed- 
ing will refer to daily ration, whereas treatment lev- 
els will continue to be referred to as percentage of 
maximum ration (MR). 
Because measurements of feeding depended upon 
the reliability with which uneaten mysids were col- 
lected, retrieval efficiency was determined. Live 
mysids, in amounts comparable to the feeding levels 
described above, were weighed to the nearest mg and 
placed in all ten containers with recirculating sea- 
water. After 24 hours, the mysids were retrieved and 
reweighed. Mean weight of retrieved mysids was 89% 
(0.017 SE) of initial weight. Therefore, differences 
between the weight of mysids provided and the 
weight retrieved was considered to be a useful esti- 
mate of feeding. 
Otolith preparation and analysis 
Otoliths were ground by hand following modified 
procedures of Neilson and Geen (1981) and Volk et 
al. (1984). Both sagittal otoliths were embedded in 
EPON resin. Otoliths were attached to a glass slide 
with thermoplastic and ground to half their thick- 
ness across a transverse plane by using a series of 
400-600 grit carborundum paper. Otoliths were pol- 
ished with 0.3 pm alumina oxide paste, reattached 
to a glass slide with the polished side down, then 
ground and polished to produce a thin section through 
the nucleus. All counts and measures were made from 
the origin along the dorsal edge of the neural groove to 
the otolith margin. All other transects lacked precision. 
Tetracycline-marked otoliths were examined with 
UV light at 400x magnification. Increment counts 
were made from the fluorescent mark to the edge of 
the otolith. Each otolith was counted twice, without 
knowledge of the previous measurement, and con- 
firmed by an independent counter. Otoliths from ex- 
periment 2 were examined under 400x magnifica- 
tion with transmitted light. Mean IW was calculated 
from three “blind” measurements. We made all counts 
and measurements with an Olympus Cue 2 Image 
Analysis System. One pair of otoliths from the 66% 
ration treatment was not readable and was subse- 
quently discarded. 
Statistical analyses 
Experiment 1 — daily otolith increment valida- 
tion To validate the daily nature of otolith incre- 
ment, the regression slope of increment count on day 
was tested to determine if it differed from one (Stu- 
dents’ Atest). Outliers were detected by calculating 
the leverage coefficients and by computing the stan- 
dard residuals from the regression equation line. 
Only 2% of the otoliths were reexamined because the 
leverage coefficient was greater than 4 In and the 
standard residual was greater than the Avalue for a 
sample size of n (Sokal and Rohlf, 1981). 
Experiment 2 — effect of ration level on increment 
width and specific growth rate Mean IW among 
ration treatments for increments formed during the 
pretreatment period was compared with ANOVA 
(a =0.05), followed by Tukey’s multiple comparison 
tests (Zar, 1984). Mean IW among ration treatments 
for weeks 1 and 2 and between weeks within ration 
treatments was analyzed with two-way ANOVA 
(a =0.05). Mean specific growth rate for each treat- 
ment was regressed against mean increment width 
following confirmation of normality ( Kolmogorov - 
Smirnov test) and homogeneity of variances (Coch- 
ran’s C test) with a=Q.Q5 for all treatment levels. 
Increment width was compared with SGR, daily ra- 
tion, and Fulton’s K at the end of the experiment for 
each fish (Pearson product-moment). Regression 
analysis was used to determine the relation between 
IW and SGR Oog{G+l}) for each fish. Regression lines 
were fitted by using a second-order regression against 
daily ration (Zar, 1984). 
Results 
Experiment 1 — daily otolith increment 
validation 
The slope of the regression increments on days after 
injection was not significantly different from one 
(y=0.975x + 0.825, P>0.05), thus supporting the daily 
periodicity of otolith increment formation in this spe- 
cies (Fig. 1). 
Experiment 2 — effect of ration level on 
increment width and specific growth rate 
No differences were found in mean IW among treat- 
ments during the 2-d pretreatment period. Initially, 
IW narrowed in the lower feeding levels: 17%, 32%, 
and 46% maximum MR (Fig. 2). Mean daily IW 
ranged from a low of 2.3 pm (17% MR or 4.1% body 
weight/d) during week 2 to a high of 4.5 pm (66% 
MR or 15.6% body weight/d) during week 1 (Table 2). 
Mean IW was significantly lower for the 17%, 32%, 
and 46% MR treatments during the entire 14-d pe- 
riod as compared with the higher ration treatment 
