34 
Fishery Bulletin 108(1 ) 
ro 
X 
4 - 
3 - 
2 - 
1 - 
0 - 
Area 1 
□ Lightly trawled 
■ Heavily trawled 
Area 2 
I 
Area 1 Area 2 
Figure 3 
(A) Numbers of hagfish ( Eptatretus spp.) burrows 
(thousands/ha, ±1 standard error [SE ] ) and (B) the 
Shannon-Wiener diversity index for all invertebrate 
species and species groups at two lightly and heav- 
ily trawled sites at Nehalem Bank, Oregon (see 
Fig. 1). 
occasionally observed hagfish using the burrows. Counts 
of trawl door tracks, by site, corresponded generally with 
the average index of trawling history from logbook data 
(Fig. 2), however the trawl track data indicated more 
trawling at site IB than at 2B and the logbook data 
indicated more trawling at 2B. The trawl door tracks that 
we observed were relatively large features in this very 
flat mud habitat, consisting of linear furrows about 0.2- 
0.5 m deep. The habitat at all four sites appeared very 
similar, but a few very small, low-relief, rocky outcrops 
were noted within site 1A. The most abundant benthic 
macroinvertebrates that we observed were sea whips and 
orange sea pens (Ptilosarcus gurneyi) at the two northern 
sites and sea whips and heart urchins ( Brisaster spp.) 
at the two southern sites (Table 2). Unidentified flatfish, 
eelpouts (Zoarcidae), Pacific hake ( Merluccius pf'oductus), 
and rockfishes ( Sebastes spp.) were the most abundant 
fish species observed at the four sites (Table 2). 
Analysis of variance showed substantial heteroge- 
neity between the northern and southern site pairs, 
confirming the appropriateness of using the blocking 
variable. Sea whips and orange sea pens were more 
abundant at the northern sites (Table 2, blocking vari- 
able PcO.OOl), whereas heart urchins, squat lobsters 
(Galathoidea spp.), Pacific hake, eelpouts, juvenile rock- 
fish, and skates (Rajidae) were more abundant at the 
southern sites (blocking variable P<0.05). The density 
of hagfish burrows was higher at the northern sites 
(Fig. 3A, blocking variable PcO.OOl). The density of 
trawl tracks was not significantly different between 
northern and southern site pairs (Fig. 3B, blocking 
variable P> 0.05). 
Statistically, trawling history (LT or HT) accounted 
for some of the heterogeneity between sites. The HT 
sites had more hagfish burrows (Fig. 3A, P<0.05) and 
trawl tracks (Fig. 2B, P<0.01) contributing to physi- 
cal complexity, but had reduced densities of sea whips 
(Table 2, PcO.Ol). Densities of the orange sea pen were 
also lower at HT sites, but the difference was not sta- 
tistically significant. The apparent influence of trawling 
history on the density of sea whips was much greater at 
the southern sites (interaction P <0.01), indicating that 
the effect of ocean shrimp trawling on sea whips could 
be variable between areas and possibly density depen- 
dent, having greater impacts at sites that are less favor- 
able habitat for sea whips. It is also possible that this 
interaction indicates differences in sea whip abundance 
between sites that are unrelated to trawling history. 
Several other benthic macroinvertebrates were ob- 
served at lower densities at the HT sites (Table 2) (in- 
dicating negative effects from chronic trawling) such as 
the flat mud star ( Luidia foliolata, PcO.OOl), unidenti- 
fied sea stars (Asteroidea, P<0.05), and squat lobsters 
(unidentified Galathoidea, PcO.OOl). The reduction in 
squat lobster density from trawling was greatest at the 
southern sites, where squat lobster densities were also 
the highest (interaction P<0.05), again indicating vari- 
ability in how trawling influenced macroinvertebrate 
density between sites. Although their numbers were 
not tested statistically, sea cucumbers (unidentified 
Holothuroidea) and unidentified corals (Hydrocoralia) 
were observed at both LT sites (present in <50% of tran- 
sects) but were not observed at either of the HT sites. 
Three taxonomic groups of fish were also observed at 
lower densities at HT sites, including juvenile rockfish 
(PcO.Ol), eelpouts (PcO.Ol), and skates (P<0.05). 
Wilcoxon rank sum tests showed that for the southern 
sites, mean sea whip length was greater at the HT site 
(PcO.OOl) because of a relative lack of smaller individu- 
als (Fig. 4). The difference in mean length between the 
northern HT and LT sites was not significant (P>0.05). 
Although not tested statistically, sea whips were much 
larger at the northern than at the southern sites where 
their densities were lower (Fig. 4). 
Invertebrate diversity based on the Shannon-Wiener 
diversity index was higher at the LT sites (Fig. 3B) 
and generally higher at the southern sites because of 
reduced dominance of sea whips (Table 2). Species rich- 
ness per unit of area surveyed was greatest at the two 
southern sites and was not consistently reduced at HT 
sites (13.8 and 17.1 taxonomic groups/ha at sites 2A and 
2B, 9.3 and 10.6 taxonomic groups/ha at 1A and IB, 
respectively). The absolute number of taxonomic groups 
was highest at the LT sites. Fourteen and 16 of the 17 
