Pearcy and Fisher: Ocean distribution of Alosa sapidissima along the Pacific coast of North America. 
443 
LoglO (catch +1) 
+ 
0 
o 
0 001 
- 0.25 
o 
0.25 - 
0.50 
o 
0.50 - 
1.00 
O 
1 00 - 
1.50 
Figure 1 
Average catch of American shad (Alosa sapidissima) in 10' latitude x 10' longitude sampling 
areas during (A) 30-minute bottom tows conducted by the Alaska Fisheries Science Center 
for the years 1977-2004 combined, and (B) during 15-minute bottom tows conducted by the 
Northwest Fisheries Science Center for the years 2003-08 combined in early season sampling 
(May-July), and (C) late season sampling ( August-October). Isobaths are 200 m and 500 m. 
Catch is transformed by log 10 (catc /2 +1). 
Table 1). On the shelf shad were found in 28% of AFSC 
tows (1132 of 4010), whereas, off the shelf in deeper 
water they were found in only 2% of tows (36 of 1586). 
Similarly, shad were found in 24% of NWFSC tows 
(394 of 1669) on the shelf, but in <1% of off shelf tows 
(9 of 2093). 
The average weight of individual shad caught in AF- 
SC trawls increased with both latitude and day of year 
(Fig. 6, A and B). However, because latitude and day of 
year were highly correlated during each AFSC cruise 
(correlation coefficient [f?] = 0.90-0.99), it was impos- 
sible to separate the effects of these two variables on 
size. Average weight consistently increased with depth 
during both surveys and often with day of year during 
the AFSC sampling. An effect of sea surface tempera- 
ture on size was evident in only a few years (Table 1). 
Similar increases in weight were obtained when latitude 
was substituted for day of year (not shown in Table 1). 
The increase in size of shad with depth was consistent 
in all sampling collections that we examined. Besides 
the increase in average weight of shad with depth in 
the AFSC and NWFSC tows (Table 1), shad weight in 
purse seines was also positively correlated with depth 
(72 = 0.40, n = 43 hauls, P=0.007), and in the limited- 
entry trawl fisheries off Washington, Oregon, and Cali- 
fornia during both summer and winter seasons (linear 
regressions of average weight of shad by 5-m tow depth 
intervals: summer, n- 36, slope =0.0026 kg/m, coeffi- 
cient of determination [r 2 ] = 0.53, P<0.001; winter, n = 33, 
slope=0.0016 kg/m, r 2 =0.56, PcO.001). 
The weight of shad caught in later years, uncorrected 
for date of sampling or depth of tow, was also signifi- 
cantly less than that in early years (Fig. 7A). To correct 
weight for depth and date of sampling in the different 
years a general linear model was applied to the weight 
data from both the AFSC and NWFSC demersal sur- 
veys combined (Fig. 7B). Even when adjusted for bottom 
depth and date of capture (much earlier for 1995—2004), 
