184 
Fishery Bulletin 11 7(3) 
Table 2 
Comparison of the arithmetic mean and standard deviation (SD) of environmental and demo¬ 
graphic variables between single- and multiple-brooding rockfish species (Sebastes spp.): lati¬ 
tude (decimal degrees), temperature (°C), dissolved oxygen (DO) concentration (mL/L), depth 
(m), maximum fork length (L„, cm), observed maximum age (Amax, years), and natural mor¬ 
tality rate (M). Data are provided for fish collected along the West Coast of the United States. 
Single brooding Multiple brooding 
Variable 
Mean 
SD 
Mean 
SD 
Latitude 
41.24 
4.83 
35.72 
2.49 
Temperature 
8.11 
1.29 
9.09 
0.50 
DO 
1.94 
0.44 
1.91 
0.46 
Depth 
157.44 
28.18 
161.35 
34.33 
L 
46.62 
17.96 
46.89 
19.70 
^MAX 
55.36 
23.96 
40.69 
14.19 
M 
0.12 
0.08 
0.14 
0.08 
Table 3 
Top 5 generalized linear models as defined by the Akaike information criterion (AIC) analysis. Models were used 
to evaluate the probability of each rockfish species (Sebastes spp.) being a multiple brooder; all variables were con¬ 
tinuous. The number of parameters in the model (K), maximum log-likelihood (LL), AIC, AIC corrected for small 
sample sizes (AIC c ), difference between the AIC c of model i and the smallest AIC c among the considered models 
(AAIC c) , and AIC weights (AIC W ) are provided for each model (models 1-5 [M1-M5]). Variables include those for 
latitude (Lat), dissolved oxygen (DO) concentration, maximum fork length (L„, cm), and temperature (Temp). 
Model name 
Model (log(y) = ) 
K 
LL 
AIC 
AIC c 
AAIC c 
AIC W 
Ml 
25.27 - 0.51X Lat - 2.78X DO 
3 
-9.95 
25.91 
27.11 
0.00 
0.39 
M2 
13.32 - 0.35X Lat 
2 
-11.61 
27.22 
27.79 
0.68 
0.28 
M3 
25.14 - 0.52Xj, at - 2.85X nn + 0.01X, 
4 
-9.87 
27.73 
29.84 
2.73 
0.10 
M4 
-10.55 + 1.98A Xemp - 3.25Z DO 
3 
-11.54 
29.07 
30.27 
3.16 
0.08 
M5 
-9.02 + 1.06Z Xemp 
2 
-13.68 
31.36 
31.93 
4.82 
0.04 
the estimate for the temperature coefficient indicates that 
when DO is fixed and temperature increases by 1°C, the 
odds of a shelf rockfish species being a multiple brooder is 
7.2 times that of being a single brooder. Finally, the results 
of evaluation of relative importance of influence of each 
predictor indicate that there was substantial evidence that 
species in lower latitudes had a higher probability of being 
a multiple brooder (LER for latitude=0.63; Table 4). There 
was minimal evidence indicating that any of the other vari¬ 
ables affected the probability of a species being a multiple 
brooder (all other LERs were <0.5; Table 4). 
Discussion 
Our results indicate that species with the capacity for 
multiple brooding were more likely to occur in lower lat¬ 
itudes (32-36°N) or in warmer waters (>9°C at depth). 
This finding agrees with the established knowledge 
on differing distributions and community assemblages 
of rockfishes regionally (e.g., by latitude) (Love et al., 
2002; Williams and Ralston, 2002; Gunderson and Vet¬ 
ter, 2006). Oceanographic conditions in waters of South¬ 
ern California and Central California are characterized 
by warmer conditions and more variable, less seasonal 
upwelling patterns relative to conditions in the northern 
part of the California Current (e.g., north of Cape Men¬ 
docino), where ocean temperatures tend to be cooler and 
upwelling is both stronger and more strongly seasonal, 
conditions that result in higher primary and secondary 
productivity (Parrish et al., 1981; Ware and Thomson, 
2005; Checkley and Barth, 2009). Consequently, rock¬ 
fish species in waters of Southern California, where 
primary and secondary productivity and therefore pre¬ 
sumably forage availability are both lower and charac¬ 
terized by greater interannual variability, may have an 
