92 
Fishery Bulletin 95( 1 ), 1997 
were significant (P<0.05) during certain months in 
each of the subareas (Table 2). Significant relation- 
ships persisted between C. hamatus density and a 
temperature variable for several extended periods. 
Surface temperature was negatively correlated with 
abundance from November to March in the MAB 
subarea and also from June to February in SNE 
Month 
Figure 3 
Bar graph of the monthly log mean abundance of 
Centropages hamatus with a line graph of the monthly 
mean surface water temperature for each subarea 
(MAB=Middle Atlantic Bight; SNE=Southern New En- 
gland; GBK=Georges Bank; and GOM=Gulf of Maine) . 
waters. Abundance in the GBK subarea was posi- 
tively correlated to average water-column tempera- 
tures from May to July and with bottom tempera- 
tures from September to December. In the GOM sub- 
Month 
Figure 4 
Time series monthly log mean abundance (solid line) of 
Centropages hamatus and the 95% confidence interval 
(dashed line) of the mean for each subarea. Single points 
are the log mean abundance of individual surveys. Sur- 
veys that departed substantially from the time series mean 
are labeled and the 95% confidence interval of the mean 
indicated with a error bar (MAB=Middle Atlantic Bight; 
SNE=Southern New England; GBK=Georges Bank; and 
GOM=Gulf of Maine). 
