SEA-FISHERIES LABORATORY. Ni) 
“gie-zaging” up and down in quite a violent manner. 
This will be seen on consulting figs. 11 and 12, which 
show these smaller temperature variations at the More- 
cambe Bay Light Ship. ‘hree-daily averages have been 
calculated for every day during the first six months of 
1907—that is, the temperature record for each day 1s 
allowed to be influenced by those of the preceding and 
succeeding days. If the actual daily temperatures were 
plotted the changes would be still more violent. ‘Then 
the curve obtained by plotting thirty-daily averages every 
ten days is superposed on this three-daily curve. In the 
figures this latter curve is represented by the horizontal 
straight line, and the three-daily temperatures are shown 
as variations above and below what may be regarded as 
the mean for the whole period considered. 
Now, it will be seen that during the months January 
to March, inclusive, the sea temperature fell at intervals 
about 0°5° C. to about 195° C. below the mean, and rose, 
on the average, about 0°5° to 0°73° C. above the mean. 
Further, during the months April to June, inclusive, the 
temperature varied between about 15°C. below, and 
30° C. above the mean. Similar variations are observed 
when we consider the third quarter of the year, and 
there are smaller variations during the fourth quarter. 
It is inconceivable that these smaller variations can 
be due to variations in the amount of heat directly trans- 
mitted to the sea by varying solar radiation, or by heat 
yielded up to the surface layers by contact with warmer 
or colder wind currents, or by differences in the amount 
of evaporation occasioned by atmospheric changes. The 
specific heat of sea-water 1s so great that enormous 
quantities of heat would have to be exchanged between 
sea and atmosphere in order to account for a variation of 
3°C. in the period of a few days. It is, of course, 
