8136

• UCREFRP Catalog Number: 8136
• Author: Bestgen, K. R.
• Title: Growth, survival, and starvation resistance of Colorado squawfish larvae.
• USFW Year: 1995.
• USFW - Doc Type: \
• Copyright Material: YES

Results

Growth response

Growth of Colorado squawfish larvae was not sig-
nificantly different in constant and fluctuating re-
gimes (p = 0.93), so data for the same temperatures
were pooled in subsequent analyses. Response sur-
face analysis indicated Colorado squawfish larvae
achieved maximum growth at the highest food
abundance and temperature (Fig. 1, Table 1). Mal-
lows Cp selected a model that included the main ef-
fects of temperature and food abundance, their re-
spective squared terms, and their interaction (Table
2). The final model had an R2 = 0.976 and RMSE =
0.0125. Food abundance explained most growth
rate variation (partial R2 for food abundance, its
squared term and food x temperature interaction =
0.916). The LOF test was highly significant (p
< 0.0001), but residual analysis did not indicate
model LOP. The regression growth model F-value
of 933.7 far exceeded the Box-Wetz criterion of
31.8, which suggested adequate model fit.
Optimization suggested a maximum growth rate
of 0.305 mm TL d-I at 310 C and food abundance of

201

345 nauplii fish-I day- 1, both of which are just out-
side of the upper range of values tested in this ex-
periment (300 C and 320 nauplii fish-I day-I). Thus,
optimization and the data showed growth of Col-
orado squawfish larvae was highest between 26 and
310 C and at 320 to 342 nauplii fish-I day-I. Model
solutions showed that growth of Colorado squaw-
fish larvae declined 4.5-7% for each degree drop in
temperature from 22 to 180 C at maximum food
abundance.
A significant food x temperature interaction (p
< 0.0001) was likely due to reduced growth rates of
Colorado squawfish larvae in treatments at 180 C
and food abundance of 142 and 320 nauplii fish-I
day-\ when compared to growth rates of larvae in
the same food abundance levels for 22, 26, and
300 C (Fig. 2). Growth rates at 180 C with food
abundance of 142 and 320 nauplii fish-I day-I were
20 and 31 % less than growth rates averaged for the
three higher temperatures with the same food
abundance. Growth rates at 12.5, 28, and 64 nauplii
fish-I day-I at 180 C were comparable to those
achieved at higher temperatures. Significant inter-
actions may invalidate interpretation of model
main effects unless interactions are orderly (au

Table 2, Least squares (growth model) and maximum-likelihood (survival model) estimates and significance probabilities for coefficients
from quadratic response surface regression models relating growth and survival responses to water temperature and food abundance,
Response variable or df Coefficient estimate SE
interaction p R'
Growth'
Intercept 1 - 0,2627170 0,041603 < 0.0001 0.976
Temperature (T) 1 0.0286770 0.003491 < 0.0001
Food (F) 1 0,0007160 0,000071 < 0,0001
T' 1 - 0,0006370 0.000072 < 0.0001
F' 1 - 0,0000024 0,000001 < 0.0001
TxF 1 0,0000313 0,000002 < 0.0001
Survivalb
Intercept 1 - 6.5160000 3,527200 0,0647
Temperature (T) 1 0.2953000 0.256300 0.2511
In (Food) (InF) 1 2.0144000 0,749800 0.0086
T' 1 - 0,0169000 0.005300 0,0012
lnF' 1 - 0.4452000 0,074900 < 0,0001
TxlnF 1 0,1084000 0.018800 < 0,0001

, Growth model statistics: df = 5, 114, model F = 933,7 (F = 3.18 at p '" 0,01), lack of fit p < 0,0001, root mean square error = 0.012553;
response in mm d-1.
b Survival response is logit (survival). Food abundance is natural log (food), Deviance/df (n = 114) = 1.9194, inflation factor for standard
errors was the square root of deviance/df = 1.3854,