262 
Fishery Bulletin 113(3) 
3.5 
3 0 
2.5 
re 
5 - 2.0 
>, 
1.5 
c 
(D 
13 1.0 
c 
re 
® 0.5 
0.0 
10 20 30 40 50 60 70 
Depth (m) 
Figure S 
Mean densities of age-0 southern Tanner crabs 
( Chionoecetes bairdi ) from beam trawl hauls, by 
depth, conducted at Pillar Creek Cove, Kodiak 
Island, Alaska, in July and August 2011. Error 
bars indicate standard error of the mean. 
July to August (Wald % 2 =23.4, df=l, P<0.001), the ef- 
fect of depth on density was consistent across months 
(monthxdepth: Wald % 2 =7.1, df=3, P=0.055). 
At Holiday and Pillar, the depth distribution of poly- 
chaete worms in 2011 (Fig. 6) was similar to that ob- 
served in 2010 (Fig. 3), albeit, overall worm abundance 
was lower. Worms at Kalsin displayed a distribution 
similar to that of worms at Pillar, in distinct contrast 
to Womens, where worms were most abundant at the 
2.0 r 
5 10 15 20 25 
Depth (m) 
Figure 6 
Mean worm index scores (averaged over months), 
which provide a measure of relative abundance of 
worm tubes on a 5-point scale (0-4), plotted against 
depth for each of the 4 sites surveyed in 2011 around 
Kodiak Island, Alaska, at Holiday Beach, Pillar Creek 
Cove, Womens Bay, and Kalsin Bay. Error bars indi- 
cate standard error of the mean. 
Figure 7 
From quadrat surveys conducted by divers in 2010, 
mean densities (A) of age-0 southern Tanner crabs 
( Chionoecetes bairdi) and (B) of worm tubes, plotted 
against depth for the study site at Pillar Creek Cove, 
Kodiak Island, Alaska (y-axis variables are averaged 
over months). Error bars indicate standard error of the 
mean 
shallowest depth and decreased in abundance as depth 
increased. On first examination, worm abundance 
might be construed as influencing the distribution of 
crabs at Womens because crabs also were more abun- 
dant in shallow water. However, worm abundance did 
not have a significant influence on crab abundance in 
the GLM (Wald * 2 =3.0, df=l, P=0.083). 
Diver estimation of crab density 
The quadrats used by divers to assess crab density 
during 2010 provided an examination of habitat asso- 
ciations on a micro scale (<1 m); the crab scrape, on 
the other hand, provided an examination at a larger 
scale (>10 m). Although there was a tendency for crab 
density from quadrats (Fig. 7A) to decline through the 
summer, this effect was not significant (Wald x 2 =5.3, 
