Vasslides and Able: Importance of shoreface sand ridges as habitat for fishes 
103 
Table 5 
Eigenvalues and cumulative percentage variance for each axis analyzed by season for the 1991-95 beam trawl data and 2001-06 
otter trawl data by using canonical correspondence Analysis (CCA). The eigenvalues are a relative measure of the importance 
of each axis. 
1 
CCA axis 
2 3 
4 
1991-95 beam trawl data 
Midsummer 
Eigenvalue 
Cumulative percentage variance 
0.296 
0.176 
0.145 
0.117 
of species data 
21.3 
34.0 
44.4 
52.8 
of species environment 
39.1 
62.3 
81.4 
96.8 
Late summer 
Eigenvalue 
Cumulative percentage variance 
0.459 
0.209 
0.153 
0.132 
of species data 
17.2 
25.0 
30.8 
35.7 
of species environment 
46.8 
68.1 
83.8 
97.2 
2001-06 otter trawl data 
Midsummer 
Eigenvalue 
Cumulative percentage variance 
0.572 
0.356 
0.277 
0.176 
of species data 
10.5 
17.0 
22.0 
25.3 
of species environment 
33.0 
53.6 
69.6 
79.8 
Late summer 
Eigenvalue 
Cumulative percentage variance 
0.343 
0.236 
0.128 
0.067 
of species data 
10.7 
18.1 
22.1 
24.3 
of species environment 
39.8 
67.2 
82.0 
89.9 
Table 6 
The P values from Monte Carlo permutation tests on the significance of the environmental variables in each data set. Significant 
P values (P< 0.05) are shown in bold. A dash indicates that the variable was not measured. 
Variable 
1991-95 beam trawl 
2001- 
-06 otter trawl 
Midsummer 
Late summer 
Midsummer 
Late summer 
Salinity 
0.036 
0.874 
0.200 
0.996 
Temperature 
0.212 
0.002 
0.002 
0.360 
Depth 
0.120 
0.266 
0.004 
0.090 
Distance from ridge 
0.062 
0.094 
0.406 
0.002 
Habitat complexity 
0.796 
0.028 
0.132 
0.112 
Dissolved oxygen 
— 
— 
0.022 
0.654 
Water clarity 
— 
— 
0.218 
0.824 
pH 
— 
— 
0.212 
0.172 
Otway, 1994), these gradients were all at substantially 
larger spatial scales. Few studies of either juveniles 
or adults have been conducted at a resolution similar 
to that of our study in inner continental shelf waters. 
Juvenile fish assemblages of the nearshore (<40 km) 
coast of Georgia exhibited a cross-shelf pattern in win- 
ter and spring, and a shallow group (8 m in depth) was 
separated from the other stations (12-18 m in depth) 
(Walsh et ah, 2006). Fish assemblages in northern Ar- 
gentina in the spring were identified as either those of 
the inner, central, or middle regions of the coastal shelf 
(Jaureguizar et al., 2006). 
Analyses of fish assemblages across continental 
shelves often point to depth as the primary environ- 
mental variable correlated with the changes in fish-as- 
semblage structure (Mahon et al., 1998; Walsh et ah, 
2006), whereas studies focused on shorter distances 
indicate a combination of environmental and physi- 
