Again, the isotherm is necessarily constructed from average temperatures which have been corrected so as to be applicable to the same level. On the equator, for instance, the summit of a lofty mountain is seen by the snow on it to have a temperature not exceeding 32° F., while at sea-level the temperature may be 90°. But observations have been made showing the rate at which the temperature of the air diminishes as the height increases, and although the rate varies in different places and at different seasons, it may be taken roughly as one Fahrenheit degree in 300 feet. Now if the mountain top with a temperature of say 30° F. is known to be 18,000 feet above the sea, the addition of 1° for every 300 feet, or 60° altogether, would give the temperature of 90° as that corresponding to sea-level. By applying such corrections, the isothermal maps have been constructed to show the distribution of temperature at the level of the sea. In order to compare the temperature he has observed with that on the map the observer must calculate the average of his daily observations for the month in question, and then make the correction for the altitude of his station. Similarly, in ascertaining from an isothermal map the mean temperature of a particular place, care must be taken to subtract from the number of degrees of the isotherm passing through the place one degree for every 300 feet of elevation. Of course it will usually happen that no isotherm as shown on the map runs through the point the mean temperature of which it is desired to obtain. In that case the temperature at the point will be found by considering its relative position between the two nearest isotherms. Thus, if it lie half-way between the lines of 60° and 70°—measured perpendicularly to the isotherms—the temperature of 65° may be assumed; if it lies one-tenth of the distance from 60° and nine-tenths from 70°, it is safe to assume 61°; if three-tenths from 70° and seven-tenths from 60°, then assume 67°. If the point lie in a loop of a single isotherm, _e.g._, Cape St. Roque, the eastern point of South America in the map for January, lying within the 80° isotherm, one can only guess that the temperature is above 80° and it may be assumed to be below 85°. The method of representation is unsatisfactory in such a case. These facts being borne in mind, the study of isotherm maps will be found to give an excellent general idea of the distribution of climate at sea-level, and if the contour lines of 600 and 6000 feet are traced on the maps the areas within which corrections of over -2° and -20° have to be applied to the isothermal values to get the temperature at the place will be easily recognised.