8021

• UCREFRP Catalog Number: 8021
• Author: Bestgen, K. R. and J. M. Bundy
• Title: Environmental Factors Affect Daily Increment Deposition and Otolith Growth in Young Colorado Squawfish
• USFW Year: 1998
• USFW - Doc Type: Transactions of the American Fisheries Society
• Copyright Material: YES

108

BESTGEN AND BUNDY

each of five 3.8-L aerated jars representing 0, 5,
10, 15, and l7.5-d-long starvation treatments.
Temperature was constant at 2loC (Table 1). Up
to 10 specimens were sampled from each treatment
0, 5, 10, 15, 20, and 28 d after the experiment
began. Mean survival, measured in three replicate
beakers for each treatment, was 87-95% for all
treatments except 17.5 d, where survival was 57%
(Bestgen 1996). Because survival was high in all
but one treatment, otolith and somatic growth ef-
fects should not have been confounded by differ-
ential survival or density. Mean otolith size and
TL were measured for five specimens per treatment
per sampling occasion to determine if otolith
growth rates reflected changes in fish length as
influenced by food availability.
Experiment 4 was conducted to determine ef-
fects of a temperature increase and of the time of
first feeding on otolith growth in early life stages
of Colorado squawfish. Five unfed 6-d-old larvae
from a stock held at 250C were placed in each of
four 2-L aquaria and maintained without feeding
in an l80C water bath until fish were 14 d old.
Two aquaria were then randomly assigned to a
constant temperature treatment of 18 or 250C (Ta-
ble 1). Fish in one randomly chosen aquarium in
each temperature treatment were fed for the 7-d
treatment period; the others were unfed. The 21-
d-old fish were sacrificed and measured to nearest
0.1 mm TL, and diameters of their lapilli were
measured. Differences in fish otolith growth be-
tween treatments were compared via two-factor
analysis of variance (PROC GLM) that had tem-
perature (18 or 250C) and food (present or absent)
as main effects and the interaction.
All statistical analyses were conducted with
SAS statistical software (SAS Institute 1988).
Plots of data and residuals were used to evaluate
normality and heteroscedasticity, and data were
transformed when necessary to meet assumptions.

Results

Early Otolith Development

Lapillar and sagittal otoliths were present in
Colorado squaw fish embryos 1 d before hatch (4
d postfertilization) at incubation temperatures of
22 and 260C and 1-2 d before hatch (4 d postfer-
tilization) at 180C. Four to eight small, round nu-
clei (primordia) formed the otolith core. Subse-
quent material was deposited around fused pri-
mordia to form round to slightly oval structures.
Asterisci (the third otolith pair) were first found
in larvae 15 d posthatch. At hatching, the lapillus

and sagitta were similar in size to each other and
similar in size at all temperatures, averaging 37.5
f.Lm (range, 32.4-44.3 f.Lm) in diameter. Left and
right otoliths were similar in size so the left was
used exclusively unless it was lost. A clear, high-
contrast otolith increment was deposited the day
of hatch in fish reared in all temperature regimes.
Embryos incubated at l80C hatched about 1 d later
than those incubated at 22 or 260C, and in some
cases they had an additional low-contrast incre-
ment inside the high-contrast hatch increment.

Age Validation

Increments were deposited daily in the lapillus
and sagittae of Colorado squawfish larvae and ju-
veniles beginning at hatch; the asteriscus was not
present at hatch. Slopes of regression lines for in-
crement count versus known age did not differ
significantly from 1.0 for constant or fluctuating
temperature treatments, for all fluctuating temper-
ature treatments combined, or for all treatments
combined (Table 2). Intercepts for all relationships
were not significantly different from zero, which
was the expected value if daily increments were
first deposited at hatching. High r2 values and re-
sidual plots indicated that all relationships were
linear. For all fluctuating temperature treatments,
confidence intervals about slope and intercept es-
timates were relatively small. For example, a 5%
deviation from a hypothesized slope of 1.0 usually
would be detected with at least 0.95 pr.obability
because confidence intervals did not overlap hy-
pothetical slope values of 0.95 and 1.05. Confi-
dence intervals (CI) were much wider for the con-
stant 220C treatment data than for others.
Inverse prediction intervals for estimated age as
a function of known age were much narrower for
fluctuating temperature treatments (CI, :t7-8 d)
than for the constant 220C treatment (:t18 d). Pre-
diction intervals for known age calculated for fish
less than 25 d old (:S12.5 mm TL) were :t2.5 d
for fish reared with fluctuating temperatures but
:t 5 d for fish reared at constant temperature. Oto-
liths growing at constant temperature had more
subdaily increments and lower overall increment
contrast, leading to less accurate enumeration, than
otoliths subjected to fluctuating temperatures.
The ANCOV A did not detect a significant dif-
ference between readers in the slopes (age X read-
er term) or intercepts (reader term) for regressions
of estimated versus true age for larvae reared at
220C (Table 3). Mean ages estimated by reader 1
(41.5 d; SE, 0.845) and reader 2 (40.4 d; SE, 0.840)
was 1.1 d (CI, :t2.35). Data from the two treat-