484 
Fishery Bulletin 99(3) 
intervals and ending as near shore as possible. Bottom 
depths averaged approximately 120 m (range: 25-220 m). 
The grid was surveyed twice in summer and once, par- 
tially, each autumn. Hydroacoustic and hydrographic pro- 
file data were collected but are presented elsewhere. 3 4 ' 3 4 5 
Where fish were detected with hydroacoustic equipment, 
we either interrupted the survey or returned after the tran- 
sect was completed to fish with a mid-water beam trawl. 
The net was generally fished 20-35 minutes each trawl. 3 4 
The trawl’s effective mouth opening was 50 m 2 , and net 
mesh sizes diminished from 5 cm in the wings to 1 cm in 
the codend. A 0.3-cm mesh liner was sewn into the codend, 
which terminated in a plankton bucket having 500-pm 
nytex mesh. In summer, beach-seine and dip-net samples 
occasionally supplemented the trawl catches. Subsamples 
of forage species (72 = 10 to 15 per species) were preserved 
in 10% buffered formalin-seawater solution on the vessels 
for later stomach analysis in the laboratory. We classified 
samples collected between 08:00 and 20:00 as “day” and 
those between 20:01 and 07:59 as “night.” 
In 1995, the zooplankton prey spectrum was assessed 
from dual vertical hauls taken at each station within two 
hours of fish catches by using conical nets that were 0.5 m 
in diameter and equipped with 303-pm mesh in summer 
and 243-pm mesh in autumn. We towed the nets from a 
standard depth of 20 m or to the depth at which fish were 
caught (or using a combination of both depths). Depth of 
hauls were categorized as “shallow” (<25 m) or “deep” ( >25 
to <100 m). Samples were collected at both depths at seven 
stations in summer (from 95-1-53 to 95-1-62 and 95-1-112) 
and one station in autumn (95-2-7, Table 1). 
Laboratory methods 
After a minimum of six weeks in formalin solution, fish 
samples were transferred to a solution of 50% isopropa- 
nol for at least 10 days before stomach analysis was per- 
formed. Ten specimens of each species were measured 
(mm fork length, FL; mg wet weight), and size was used to 
develop age-class categories for diet samples (Smith, 1981; 
Paul et al., 1998a). Walleye pollock (20 to 120 mm FL) and 
3 Haldorson, L. 1995. Fish net sampling. In Forage fish study 
in Prince William Sound, Alaska, p. 55-83. Exxon Valdez Oil 
Spill Restoration Project 94163A Annual Report, Juneau Center 
School of Fisheries and Ocean Science, Univ. Alaska Fairbanks, 
11120 Glacier Hwy., Juneau, AK 99801. 
4 Haldorson, L. J., T. C. Shirley, and K. O. Coyle. 1996. Bio- 
mass and distribution of forage species in Prince William 
Sound. In APEX project: Alaska predator ecosystem experi- 
ment in Prince William Sound and the Gulf of Alaska (D. C. 
Duffy, compiler), Exxon Valdez Oil Spill Restoration Project 
Annual Report (Restoration Project 95163 A-Q), Alaska Nat- 
ural Heritage Program, Department of Biology, Univ. Alaska 
Anchorage, 707 A Street, Anchorage, AK 99501 
5 Haldorson, L. T. Shirley, K. Coyle, and R. Thorne. 1997. For- 
age species studies in Prince William Sound. In Alaska preda- 
tor ecosystem experiment in Prince William Sound and the Gulf 
of Alaska (D. C. Duffy, compiler). Exxon Valdez Oil Spill Resto- 
ration Project Annual Report (Restoration Project 96163 A-Q), 
Alaska Natural Heritage Program, Department of Biology, Univ. 
Alaska Anchorage, 707 A Street, Anchorage, AK 99501. 
Pacific herring (60 to 120 mm FL) were classified as YOY 
(age-class 0). Stomachs were excised, weighed, and their 
contents were removed. The weight of prey contents was 
recorded as the difference between full and empty stom- 
ach weights. Fish were considered to have been feeding if 
their stomachs contained more than a trace of food. Rela- 
tive stomach fullness was recorded as integers represent- 
ing empty stomachs (1), stomach containing trace contents 
(2), stomachs that were 25%, 50%, 75%, or 100% full (3-6), 
or stomachs that were distended (7). State of digestion was 
recorded as partially digested contents (1), mostly digested 
contents (2), and empty stomachs (3). 
Stomach contents and zooplankton samples were identi- 
fied with a binocular microscope to the highest taxonomic 
resolution possible and enumerated. The prey category of 
calanoid copepods was also segregated into “large” (>2.5 
mm total length, TL) and “small” individuals (>2.5 mm 
TL). We pooled the common pelagic cyclopoid copepod Oi- 
thona similis with small calanoids. We subsampled all zoo- 
plankton samples and stomach samples when practical, 
using a Folsom splitter to achieve a minimum count of 200 
of the predominant taxon. Counts were expanded and to- 
tal prey weights were determined by multiplying the ex- 
panded number observed by the mean weight per taxon. 
Weights per taxon were obtained from data on file (speci- 
mens from zooplankton samples or fish stomachs) collect- 
ed from spring through autumn of various years in south- 
eastern Alaska or PWS (Coyle et ah, 1990; Stark 6 ; senior 
author, unpubl. data). 
Analytical methods 
Forage fish were considered to occur in allopatric aggre- 
gations if only one species and one age class were caught 
in a net haul. They were considered to be sympatric if 
at least two species or age classes (>10 fish each) were 
caught together. For this study, we restricted analyses to 
YOY pollock and herring that were allopatric or that co- 
occurred only with each other to limit the complexity of 
trophic interactions; we excluded pollock and herring that 
were caught in other types of aggregations, such as with 
other species or older conspecifics. We examined the size 
of forage fish and their feeding attributes. Size included 
FL and wet weight. Feeding attributes included measures 
of the quantity of food consumed, measures of feeding fre- 
quency, and measures of prey composition. Food quantity 
was expressed as means of the total number and weight 
of prey (ln-transformed), stomach fullness index (rounded 
to nearest 25%), and prey percent body weight (%BW; 
ratio of wet stomach-content weight to fish body weight). 
Feeding frequency was measured as the percentages of 
feeding fish and the percentages of fish with partially or 
mostly digested stomach contents. Prey composition was 
expressed as the percent number and percent biomass of 
prey categories. Zooplankton density per cubic meter and 
numerical percent composition was calculated for species, 
6 Stark, C. 1995. Personal commun. Institute of Marine Sci- 
ence, Univ. Alaska Fairbanks, P. O. Box 757220, Fairbanks, AK 
99775-7220. 
