actuated by a weight slipped down the line to cause the reversal, as in the patterns of Rung, Mill and others.
All thermometers sunk into deep water must be protected against the enormous pressure to which they are exposed. This may be done by the method suggested by Arago in 1828, introduced by Aimé in 1841 and again suggested by Glaisher in 1858, of sealing up the whole instrument in a glass tube exhausted of air; or, less effectively, by surrounding the bulb alone with a strong outer sheath of glass. In both forms it is usual to have the space between the bulb and the protecting sheath partly filled with mercury or alcohol to act as a conductor and reduce the time necessary for the thermometer to acquire the temperature of its surroundings.
The warming of the ocean is due practically to solar radiation alone; such heat as may be received from the interior of the earth can only produce a small effect and is fairly uniformly distributed. On account of the high specific heat of sea-water the diurnal range of temperature at the surface is very small. According to A. Buchan's discussion of the two-hourly observations on the “Challenger” the total range between the daily maximum and minimum in the warmer seas is between 0.7° and 0.8° F., and for the colder seas still less (0.2° F.), compared with 3.2° F in the overlying air. The maximum usually occurs between 1 and 2.30 P.M., the minimum shortly before sunrise. The temperature of the surface water is generally a little higher than that of the overlying air, the daily average difference being about 0.6° F., varying from 0.9° lower at 1 P.M. to 1.6° higher at 1 A.M. There are few observations available for ascertaining the depth to which warmth from the sun penetrates in the ocean. The investigations of Aimé in 1845 and Hensen in 1889 indicate that the amount of cloud has a great effect. Aimé showed that on a calm bright day in the Mediterranean the temperature rose 0.1° C. between the early morning and noon at a depth of about 12 fathoms. Luksch deduced a much greater penetration of solar warmth from the comparison of observations at different hours at neighbouring stations in the eastern Mediterranean, but his methods were not exact enough to give confidence in the result. The penetration of warmth from the surface is effected by direct radiation, and by convection by particles rendered dense by evaporation increasing salinity. Conduction has practically no effect, for the coefficient of thermal conductivity in sea-water is so small that if a mass of sea-water were cooled to 0° C. and the surface kept at a temperature of 30° C., 6 months would elapse before a temperature of 15° C. was reached at the depth of 1.3 metres, 1 year at 1.85 metres, and 10 years at 5.8 metres. Great irregular variations in radiation and convection sometimes produce a remarkably abrupt change of temperature at a certain depth in calm water; the layer in which this sudden change occurs has been termed the Sprungschicht. How closely two bodies of water at different temperatures may come together is shown by the fact that in the Baltic in August between 10 and 11 fathoms there is sometimes a fall of temperature from 57° to 46.5° F. Such a condition of things is only possible in very calm weather, the action of waves having the effect of mixing the water to a considerable depth. After a storm the whole of the water in the North Sea assumes a homothermic condition, i.e. the temperature is the same from surface to bottom, and this occurs not only south of the Dogger Bank, where the condition is normal, but also, though less frequently, in the deeper water farther north. Similar effects are produced in narrow waters by the action of tidal currents, and the influence of a steady wind blowing on- or off-shore has a powerful effect in mixing the water.
The warmest parts of the Indian Ocean and Western Pacific have a mean annual temperature of 82° to 84° F., but such high temperatures are not found in the tropical Atlantic. In the Indian Ocean between 15° N. and 5° S. the surface temperature in May averages 84° to 86° F., and in the Bay of Bengal the temperature is 86°, and no part of the Atlantic has so high a monthly mean temperature at any season. G. Schott's investigations show that the annual range of surface temperature in the open ocean is greatest in 40° N., with 18.4° F., and in 30° S., with 9.2° F.; on the contrary, near the equator it is less, only 4° F. in 10° N., and in high latitudes it is also small, 5.2° F. in 50° S. The figures quoted above are differences between the average surface temperatures of the warmest and of the coldest month. As to the absolute extremes of surface temperature, Sir John Murray points out that 90° F. frequently occurs in the western part of the tropical Pacific, while among seas the Persian Gulf reaches 96° F., only 2° under blood-heat, and the Red Sea follows closely with a maximum of 94°. The greatest change of temperature at any place has been recorded to the east of Nova Scotia, a minimum of 28° F. and a maximum of 80°, and to the north-east of Japan with a minimum of 27° F. and a maximum of 83°. In those localities, however, it is not the same water which varies in temperature with the season, but the water of different warm and cold currents which periodically occupy the same locality as they advance and retreat. The zones of surface temperature are arranged roughly parallel to the equator, especially in the southern hemisphere. Between 40° N. and 40° S. the currents produce a considerable rearrangement of this simple order, the belts of warm water being wider on the western sides of the oceans and narrower on the eastern.
The arrangement of the isotherms thus affords a basis for valuable deductions as to the direction of ocean currents. The surface temperature of the Atlantic is relatively lower than that of the other oceans when the whole area is considered. According to Krümmel's calculation the proportional areas at a high temperature are as follows:—
Percentage of Ocean Surface with Temperature.
| Atlantic. | Indian. | Pacific. | |
| Over 77° F. (25° C) | 22.4 | 38.0 | 40.1 |
| Over 68° F. (20° C) | 50.1 | 51.7 | 58.4 |
This disparity results in some degree at least from the comparative narrowness of the inter-tropical Atlantic, and the absence of a cool northern area in the Indian Ocean. Krümmel calculates that the mean temperature of the whole ocean surface is 63.3° F., while the mean sea-level temperature of the whole layer of air at the surface of the earth is given by Hann as 57.8° F.
We are still ignorant of the depth to which the annual temperature wave penetrates in the open ocean, but observations in the Mediterranean enable us to form some opinion on the matter. The observations of Aimé in 1845 and of Semmola in the Gulf of Naples in 1881 show that the surface water in winter cools until the whole mass of water from the surface to the bottom, in 1600 fathoms or more, assumes the same temperature. Towards the end of summer the upper layers have been warmed to a depth which indicates how far the influence of solar radiation and convection have reached. Aimé estimated this depth at 150-200 fathoms, while the observations of the Austrian expedition in the eastern Mediterranean found it to be from 200 to nearly 400 fathoms. In the Red Sea, where a similar seasonal change occurs, the depth to which the surface layer warms up is about 275 fathoms. The great difference in salinity between the surface and the deep water excludes the possibility of effective convection in the seas of northern Europe, and in the open ocean the currents which are felt everywhere, and especially those with a vertical component, must exercise a very disturbing influence on convection.
The vertical distribution of temperature in the open ocean is much better known than that of salinity. The regional differences of temperature at like depths become less as the depth increases. Thus at 300 fathoms greater differences than 9° F. hardly ever occur between 50° N. and 50° S., in 800 fathoms the differences are less than 5.5° and in 1500 fathoms less than 2°. Even in the tropics the high temperature of the surface is confined to a very shallow layer; thus in the Central Pacific where the surface temperature is 82° F. the temperature at 100
