286 
Fishery Bulletin 96(2), 1 998 
treated as a subset of adult diets (e.g. Groot and 
Margolis, 1991). Therefore, as a first step to un- 
derstanding trophic interactions during their 
first marine summer, our research focused on the 
feeding habits and diet overlap of juvenile 
chinook (O. tshawytscha), coho (O. kisutch), sock- 
eye (O. nerka), chum (O. keta ), and pink (O. 
gorbuscha) salmon. We compared diet overlap 
between species, habitats, months, and years and 
also examined the similarity of salmon diets with 
samples of prey collected concurrently. 
Methods 
Fish coliection 
Juvenile salmon were collected in 120 purse-seine 
sets made along transects with two vessels and 
two seines (Table 1; see also Jaenicke and 
Celewycz, 1994). The 28-m NOAA RV John N. 
Cobb fished a table seine in the marine waters of 
southeastern Alaska during August 1983 and 
July and August 1984 (Fig. 1). The 24-m FV 
Bering Sea fished a drum seine in the open 
coastal waters of northern British Columbia dur- 
ing July 1984. Purse-seine sets at all locations 
were standardized to compensate for different 
sizes, meshes, and areas enclosed; both nets had 
25-mm mesh in the bunt (Jaenicke and Celewycz, 
1994). Each station was fished only once during 
a period, except on a few occasions when a set 
was repeated following an empty haul. Fishing 
was conducted almost exclusively between 0600 
and 1800 hours. All sets were round hauls (i.e. 
the seine was not towed or held open to increase 
catches). The distribution, size, and abundance 
of juvenile salmon examined for our study have been 
summarized in a companion paper by Jaenicke and 
Celewycz (1994). 
The waters sampled were partitioned into discrete 
habitats: 1) outside waters: the open waters adjacent 
to the Gulf of Alaska; 2) inside waters: the enclosed 
marine waters of the Alexander Archipelago; and 3) 
protected outer-coast inlets (Fig. 1). Outside waters 
were further partitioned into nearshore (0-37 km 
offshore) and offshore (46-74 km offshore) during 
August 1984; seining was restricted to within 37 km 
of shore during other sampling periods. Inside wa- 
ters were partitioned into bays and passages (con- 
necting to the Gulf of Alaska). Sampling in outside 
1 Brodeur, R. D. 1990. A synthesis of the food habits and feed- 
ing ecology of salmonids in marine waters of the North Pacific 
(INPFC Doc.) FRI-UW-9016. Fish. Res. Inst., Univ. Washing- 
ton, Seattle, 38 p. 
Figure 1 
Location of purse-seine sets in southeastern Alaska in 1983 and 
1984 and in British Columbia in 1984. 
waters was along transects about 72 km apart in 
southeastern Alaska and 108 km apart in British 
Columbia. Transects began as close to shore as sub- 
marine topography and seine depth permitted and 
continued up to 74 km offshore, depending on fish 
abundance and weather. 
Fish from each purse-seine haul were anesthetized 
with MS 222 (tricaine methanesulfonate), sorted, 
identified, and measured on board the vessel. A 
subsample of each salmon species (n< 25) was pre- 
served in 10% formalin-seawater solution for later 
stomach analysis (Table 1). Nonsalmonids caught 
incidentally were identified and enumerated, catch 
per unit of effort (CPUE) was estimated, and a 
subsample was measured for length as part of prey 
assemblage assessments (see below). Sea surface 
temperature (SST) to the nearest 0.2°C was recorded 
at every purse-seine site. 
