It was however reserved for Humboldt to determine (1817) from the registers of observed temperatures in Europe, and from the numerous observations made by himself and other travellers in different regions of the world, the constancy of the mean annual temperature of places, and to "connect graphically by lines those points where accurate observations indicated equality of mean periodic temperature." Those lines (or curves supposed to be traced on the surface of the earth) which connect the places where the mean annual temperature is the same he celled Isothermal lines. In order to ascertain with the utmost possible precision the mean temperature of any place from the tables there kept, Humboldt divided the sum of all the temperatures observed in each day at Intervals of one hour by the number of observations ; and the sum of all these mean daily temperatures being divided by 365, gave the mean annual temperature. And in determining the series of points for his lines of equal temperature, when there existed no observations on which he could depend, he interpolated the tempe rature and geographical position between the values of those elements at two or more places where they were well known.
The diagram on next page represents an orthographical projection, on the plane of the equator, of the principal meridian:, and parallels of latitude in the northern hemisphere of the earth ; and the strongly marked curves represent the nine isothermal lines whose forms were determined by Humboldt. Their distances from one another are such as correspond on the earth to a change of mean annual temperature equal to 2.5 degrees of the centigrade thermometer (4.5° of Fahr.), and the most northern curve is that on which the mean temperature is ex pressed by zero on the former, or 32° on the latter scale. The number on each curve in the diagram expresses, according to Fahrenheit's ther mometer, the mean annual temperature, at the level of the sea, of all the places through which the curve purees. The centre P represents the pole of the earth, and the longitudes of the meridian lines are numbered eastward and westward from the meridian of Greenwich.
The isothermal line of 32° passes about 4° southward of Nain, a Moravian settlement on the coast of Labrador ; and under the influence of the gulf stream, makes a remarkable Inflexion, ascending as high as North Cape in Lapland, where it abruptly returns southward, and attains its lowest limit in the eastern parts of Asia, about 50' N. lat. Proceeding westward from Labrador the curve crosses the lower I extremity of Hudaon'a Bay, from whence it again tends northwards to the Great Slave Lake, reaching its northern maximum in about 70° N. kit. The positions of the other curves seem to be affected in a greater or less measure by the same influences as act upon the curve just mentioned. In their progress from the western coast of Europe to the
eastern coast of America they incline towards the terrestrial equator, yet so that the southern curves approach near to parallelism with that great circle of the earth. Within the territory of the United States they assume a form which is convex to the equator, and farther west they appear to reaacend towards the north. In the isothermal lines of 50° and 60° westward, and those of 40' to 10° eastward of tho meridian of Greenwich the curves have their convexities turned north wards ; and farther eastward they descend towards the equator. The isothermal line of 54.5° is one that has been traced nearly round the earth : commencing at the mouth of the Columbia, on the western coast of North America, it passes south of Council Bluffs, and near the city of Washington with its convexity towards the south ; and after crossing the Atlantic it rune between Paris and Bordeaux, from whence it continues to a point a little north of the city of Pekin, where it is again convex towards the equator.
But it must be remembered that in every country the mean tem perature varies with the height of the place above the level of the sea; and Humboldt, from observations made as well on the Cordilleras as in Europe, having determined that at every 343 feet the mean temperature of the air is diminished by a quantity equal to that dimi nution which is consequent on an augmentation of latitude equal to one degree, calculated a table of the corrections which elicitild be made in the curvatures of the isothermal lines at the level of the sea, in order to obtain the forms of those which appertain to points at any given elevation, but this calculation, as pointed out under CLIMATE (col. 968) does not hold with strict accuracy for places situated without the tropics.
The differences between the mean summer and mean winter tempera tures Humboldt found to be very considerable at places whose mean annual temperature is the same; and these differences are not equal in the Old and New Continent. On the isothermal line of 32' in Europe, that difference proved to be equal to 39.6', and in America to 54'; and on the isothermal line of 68° the differences were respec tively 21.6' and 27°. He also remarked that the differences between summer and winter are least near the northern, and greatest near the southern bends of the curves. To indicate these variations, Humboldt laid down other lines. The curves formed by connecting, on the isothermal lines, points at which the mean temperature of summer is the same he called Isotheral lima; and those formed by connecting points at which the mean winter temperature is the same Isocheimal lines : both these systems of lines he found to deviate more than the isothermal lines from the parallels of terrestrial latitude.