246 



Fishery Bulletin 100(2) 



islands at the confluence of the Sacramento and San Joa- 

 quin Rivers, forms the eastern boundary. Measurable salin- 

 ity (>1 psu) occurs on the western side of the delta in Suisun 

 Bay. Water flows through Suisun Bay into San Pablo Bay, 

 then into the Central Bay of San Francisco Bay before exit- 

 ing the estuary at the Golden Gate. The estuaiy has a sur- 

 face area of about 1100 km- and is fringed on the northern 

 shore with marshes, which have shrunk by more than 90%, 

 to 125 km-, since 1850 (Conomos, 1979). Most of the estuary 

 shoreline is urban, suburban, and industrial development, 

 however. The Gulf of the Farallones, a relatively broad 

 expanse of the continental shelf extending from Pt. Reyes 

 (38°00'N, 123°01'W) to Pillar Pt. (37°30'N, 122°30'W) to 

 the Farallon Islands (Southeast Farallon Island— 37°42'N, 

 123°00'W) on the edge of the continental shelf ranges 

 to 90 m but is mostly 20 to 50 m deep. This hydrodynami- 

 cally complex area is influenced by the cool southerly flow- 

 ing California Current: freshwater discharge from the San 

 Francisco Estuary; and seasonally strong coastal upwelling 

 (spring and summer) and a northerly flowing countercur- 

 rent, the Davidson Cuirent (winter). 



Field sampling 



Juvenile chinook salmon in this study were collected from 

 the fall run, as determined by the daily length criteria 

 used to discriminate juveniles among the four runs (John- 

 son et al., 1992). Because salmon from the four runs are 

 phenotypically indistinguishable, to target fall-run juve- 

 niles we used daily length criteria and collected fish during 

 the period when fall-run chinook salmon dominate the 

 migration toward the ocean. 



We collected juvenile chinook salmon at four locations 

 spanning the San Francisco Estuary (Fig. 1): at km 68, the 

 west side of the delta at the confluence of the Sacramento 

 and San Joaquin Rivers; km 46, the exit from Suisun Bay; 

 km 26 in San Pablo Bay; and several sites within a 1-km 

 radius of km 3 for greater coverage of the estuary exit at 

 the Golden Gate and to avoid ship traffic. Two multiday 

 surveys of the estuary were completed starting at km 68 

 and proceeding through successive downstream locations 

 to the Golden Gate. The first sampling date was 30 April 

 1997 at km 68; the last was 15 July 1997 at km 3. 



Collections were made in the estuary with a midwater 

 trawl towed at 2-3 knots for 15-30 min. The trawl was 

 made of nylon mesh with a 10-m headrope and footrope. 

 10-m height at the mouth, and 20-m length. Mesh size was 

 1.6 cm at the headrope and decreased to 0.4 cm before 

 the codend. The codend was fitted with a 1.27-cm knotless 

 mesh liner. The net was kept open by aluminum spread- 

 ers and depressors. For most tows, the spreaders were vis- 

 ible at the surface, confirming that the net fished the up- 

 per layer of the water column. 



Juvenile chinook salmon were also obtained from the 

 coastal ocean in the Gulf of the Farallones. Stations where 

 salmon were caught were within 15 km of the Golden Gate 

 in waters 18 to 36 m deep. A high-speed midwater rope 

 trawl was towed at 3-4 knots for 15-30 min at each site. 

 The net had a 53-m headrope and footrope with a 1.27-cm 

 mesh codend liner (Dotson and Griffith, 1996). Tempera- 



ture-depth recorders attached to the footrope and head- 

 rope indicated that the net fished 5-10 m below the sur- 

 face with a vertical opening of 13 m. 



Fish were removed from the net as soon as practicable 

 and placed in labeled plastic zip-top bags. Juveniles cap- 

 tured within the estuary were kept under ice: those from 

 the ocean were stored at -80°C, until returned to the 

 laboratory 



We collected hydrologic data during all field trips. With- 

 in the estuary, a Hydrolab H20' Multiprobe connected by 

 cable to a Surveyor 3- data logger (Hydrolab Corp., Aus- 

 tin, TX) was used to record vertical profiles of tempera- 

 ture, salinity, dissolved oxygen, pH, and turbidity. In the 

 Gulf of the Farallones, a "SeaCat SBE 19-03" (Sea-Bird 

 Electronics, Inc., Bellevue, WA) conductivity-temperature- 

 depth sonde (CTD) recorded profiles of temperature and 

 salinity in a grid of stations spaced at 2' latitude and lon- 

 gitude intervals encompassing the salmon fishing sites. 



Laboratory analyses 



In the laboratory, we examined, measured, and dissected 

 fish, usually within 24 h of capture. Fork length and total 

 body weight were recorded. The peritoneal cavity was 

 opened by incision along the ventral side from vent to 

 opercular isthmus. The stomach and its contents were 

 removed and stored in 10% buffered formalin for subse- 

 quent analysis of prey Sagittal otoliths were removed, 

 cleaned of membranes, rinsed in deionized water, and 

 stored for subsequent aging. The remaining portion of the 

 fish was placed in a Whirl" bag. purged with N,^, and 

 stored at -80°C for lipid and protein analyses. 



Heads of juvenile salmon containing coded-wire tags, 

 evident by a missing adipose fin, were removed and sent to 

 the U.S. Fish and Wildlife Service, Stockton, CA, to obtain 

 data on the location and time of release of each fish. 



Concentrations of lipid classes and total protein were 

 determined in 15 fish randomly chosen from each location. 

 Heads, fins, and stomachs were removed to limit analyses 

 to body constituents. Frozen bodies were minced with a 

 knife, then homogenized in a blender for about 30 sec to 

 a uniform paste. We extracted lipids from a Ig- to 3g-ali- 

 quot by using the method of Bligh and Dyer (1959). Total 

 lipid was quantified by thin-layer chromatography with 

 flame-ionization detection with an latroscan TH-10 Mark 

 V" (latron Laboratories. Inc., Tokyo, Japan) according to 

 procedures published previously (MacFarlane et al., 1990, 

 1993). Total lipid was separated into steryl or wax ester, 

 triacylglycerols, nonesterified fatty acids, cholesterol, and 

 polar lipid classes and quantified according to methods in 

 MacFarlane and Norton ( 1999). We estimated total protein 

 concentration by the Lowry method, using bovine serum 

 albumin as a standard (Lowry et al., 1951). All lipid and 

 protein values are expressed as wet weight (mg/g). 



Fish ages were estimated from otolith analysis (Broth- 

 ers, 1987 1. Sagittae were embedded in epoxy, then ground 

 and polished on the distal surface to a mid-sagittal sec- 

 tion. Otolith concentric bands were counted under oil im- 

 mersion with transmitted polarized light illumination in- 

 to a video microscopy system at a monitor magnification 



