In all accurate work certain corrections have to be applied to barometer readings to make them strictly comparable. These are: (1) _correction for altitude_; (2) _correction for temperature_; and (3) _correction for latitude_. The first is necessary because of the fact that the weight of the air decreases upwards, and a barometer reading on a hill or a mountain is not comparable with one at sea level unless the former has been corrected by the addition of the weight of the column of air between the hill or mountain and sea level. The correction for temperature is rendered necessary by the fact that with increasing temperature the mercury in the barometer tube expands more than the metallic scale, because mercury is more sensitive to heat, and unless some allowance is made for this fact, barometer readings made at high temperatures will show somewhat too high a pressure. The readings of the attached thermometer give the temperature of the mercury and are used in making the corrections for temperature. As gravity varies from a maximum value at the poles to a minimum value at the equator, barometer readings made at different latitudes are _corrected for latitude_, which means that they are reduced to latitude 45°, midway between 0° and 90°. The correction is +0.08″ at the poles and -0.08″ at the equator. Tables for use in correcting barometer readings for altitude and for temperature are given in Chapter XXVI. [Illustration: FIG. 11.] =Thermograph and Barograph.=——Two instruments of much interest are the self-recording thermometer, or _thermograph_, and the self-recording barometer, or _barograph_, manufactured by Richard Brothers of Paris. In the thermograph (Fig. 11) there is a brass cylinder around which a sheet of paper is wound, this paper being divided into two-hour intervals of time and into spaces representing differences of 5° or 10° of temperature. The cylinder revolves once in a week, being driven by clock-work contained within it. The thermometer consists of a flat, bent, hollow brass tube containing alcohol, one end of the tube being fastened to the metallic frame seen at the right of the figure, and the other end being free to move. With rising temperature, the liquid in the tube expanding more than the metallic casing, by reason of its greater sensitiveness to heat, tends to straighten the tube, while with falling temperature the elasticity of the tube turns it into a sharper curve. These movements of the free end of the tube are carried through a train of levers and thus magnified. At the end of the last lever is a metallic pen filled with ink, which rests lightly against the paper on the revolving drum. A rise or fall in temperature is thus recorded by a rise or fall of the pen on the record sheet, and a continuous curve of temperature is secured. The pen of the thermograph should be frequently adjusted to make the reading of the instrument accord with that of a standard mercurial thermometer, and care