154 DR HUGH ROBERT MILL ON THE 



one whole, and the averages may, accordingly, be accepted as approximately true for the 

 Area as a whole. The mass temperatures for each year are not meaned, as the difference 

 in the depth of the divisions deprives them of comparative value, and the same holds 

 good for the retardation of temperature at great depths. The mean of the 1886-87 

 average is given as a rough datum for comparison in all cases, even where it would 

 require to be modified by weighting for difference in depth and volume in order to be 

 truly comparable. 



Looking first at the general results for the two years 1886 and 1887, we find that 

 the air-temperature in 1887 was 0°'8 higher than the previous year, while the temperature 

 of the superficial 5 fathoms was 1°*3 higher ; and, taking account of the relation of surface 

 temperature to local air-temperature for each division, the excess of surface water over 

 air temperature was twice as great in 1887 as in 1886, i.e., 2° as compared with l°'l. 

 The contrast between the two years was greatest in summer, and is best shown in the 

 rapidity with which the surface temperature rose in 1887, the maximum being reached 

 only 41 days after the air maximum, as compared with 51 days in 1886. Since it has 

 been shown that the curve of surface water temperature rises until it is cut by the 

 descending curve of air-temperature, rapid heating means a high surface maximum, and 

 explains the somewhat curious fact that in 1887, the year of high water-temperature, the 

 air was warmer than the water for 113 days only, as compared with 160 in 1886 ; and 

 that in 1887 the period of mass-cooling of the water was 21 per cent, longer than that 

 of heating, whereas in 1886 it was 3 per cent, shorter. Put briefly, the contrasted 

 temperature results of the two years indicate that the longer the duration of heating 

 in surface water the lower is the ultimate maximum, or that the rate of heating is 

 proportional to the amount of heating. 



The range of temperature between the coolest and warmest months was, on the 

 average of the two years, 20° '7 for the air, 12°*2 for the superficial 5 fathoms of water, 

 and 10° "2 for the whole mass of water, thus indicating that sea-water in layers, averaging 

 30 fathoms deep, is just half as responsive to temperature changes as the air at the 

 Earth's surface ; while layers 5 fathoms deep, when resting on deeper water, are 60 per 

 cent, as responsive as the surface air. The same result as to mass-heating is brought out 

 by comparing the mean daily rate of seasonal heating and cooling, this being practically 

 half as great in the water as in the air. 



Arranging the various divisions with reference to the difference of their data from 

 those of the Channel, we are able to get a hint as to the modifying influence of depth 

 and isolation. When this is done for the seven different elements of surface temperature, 

 mass temperature, excess of surface over air temperature, surface retardation of maximum, 

 retardation of maximum at 35 fathoms, rate of change of temperature in the mass, and 

 proportion of time of cooling to that of heating, and the numeral expressing the position 

 of the division in the series, the Channel being taken as zero, is written after its name : 

 the sum of the numbers in each case would tend to approach equality if the order were 

 merely accidental. If a well-defined order appears it cannot be accidental, but must 



