The temperature of sea water had already been observed by Ellis, in 1749, in the Atlantic, and subsequent expeditions have furnished a great number of temperature observations in various seas and for various depths. The diversity of instruments and of methods employed by the earlier observers, and the faulty methods of recording, have made the uniform reduction of many of these observations difficult or impossible. The most complete and valuable collection of these older observations up to 1868, with an account of the instruments and methods used by each observer, was published by Prestwich, in 1876, in the Philosophical Transactions, Vol. 165. With the advent of the great scientific expeditions, which were supplied with modern and refined instruments, our knowledge of the thermal conditions of the sea has progressed immensely, and we are now able to construct charts of all the oceans, showing the distribution of the isotherms with considerable accuracy. The annual average surface temperature has been found higher in the Indian Ocean than in either the Atlantic or Pacific; the North Atlantic is slightly warmer than the North Pacific, but the South Pacific is warmer than the South Atlantic; this holds generally good also for the temperatures between surface and bottom. The temperature generally decreases more or less rapidly from the surface down to about 500 fathoms, at which depth it is quite uniformly between 39° and 40° F. From that depth it decreases slowly towards the bottom: in the Polar seas to between 27° and 28° F.; in the middle and higher latitudes of the northern hemisphere and at depths of 2,000 to 3,000 fathoms, to between 34° and 36° F.; at the equator and in southern latitudes it remains in the neighborhood of 32° F. The low temperatures at the bottom are thought to be due to a steady but slow circulation of water from the Polar seas towards the equator, and, where the circulation is most free and unobstructed, as in the South Atlantic, South Pacific and Indian Ocean, the bottom temperature is slightly lower than in the North Atlantic and North Pacific, both of which are connected with the Polar Sea by comparatively narrow and shallow straits. The theory of this circulation from the Polar seas is greatly strengthened by the facts appearing from the investigation of the bathymetric isotherms in inclosed seas, i.e., seas which are separated from the deep oceans by submarine barriers. In such seas the temperature decreases slowly from the surface down to the depth of the barrier, and from there on remains constant to the bottom. The influence of currents on the surface temperature is very marked, cold currents bending the isothermal lines towards the equator, and warm currents bending them towards the poles. The seasonal changes in surface temperatures are considerable, being the least in the tropical zones.