Figure 3.—Distribution of autumn  bottom-water 
temperatures, 1963-75. (Dotted areas represent Georges 
Bank temperatures greater than 14°C; gridded areas 
represent Gulf of Maine temperatures greater than 8°C.) 
Autumn 1973 
TCI by the percentage area within that interval and 
dividing the total by 100. This method was chosen as it 
was not possible to make direct year-to-year compari- 
sons of station data because randomly located stations 
were sampled during each cruise. The stratified random 
sampling scheme did, however, result in representative 
sampling by depth and geographic subareas. This 
method also eliminates bias associated with depth 
variability as the indices derived are based solely on an 
areal basis and the same geographic areas are used to 
compare relative annual differences in bottom-water 
temperatures. Since no attempt is made to determine 
absolute temperature indices, the method chosen for 
data analysis seems justified. For illustrative purposes, 
portions of the histograms of TCI distributions have been 
shaded to emphasize temperature extremes. For Georges 
Bank TCI’s<4° and>8°C in spring (Figs. 7, 9) and<6° 
and >10°C in autumn (Figs. 15, 17) are shaded. For the 
Gulf of Maine TCI’s<4° and>6°C in spring (Figs. 11, 13) 
and<8° and >14°C in autumn (Figs. 19, 21) are shaded. 
The shaded portions of the histograms generally lie out- 
side expected mean bottom temperature conditions. 
Dates for the collection of temperature data are shown 
in Figure 4. The effects of irregular seasonal sampling are 
difficult to interpret, especially for shoal areas like 
Georges Bank where seasonal cycles of bottom-water 
