Tyndall has modified it by filling the upper vessel with mercury, which is a better conductor of heat than water.
For relative measurements, as for instance a comparison of the amounts of heat received from the sun at different hours of the day, Crova employs a slightly different instrument, of which Fig. 2, copied from his paper in the "Annales de Chimie" for February, 1880, is a representation.
An exceedingly sensitive alcohol thermometer, shown separately at T, with a large bulb carefully blackened, is inclosed in a double-walled sphere B, nickel-plated on the outside. An opening in the walls of the sphere, carefully aligned with a similar opening in a double screen E, allows a beam of light to fall upon the thermometer-bulb, the beam being about two thirds the diameter of the bulb. Theis constructed with a supplementary reservoir, r, at the lower end, by means of which the end of the indicating column can be made to fall near the middle of the scale at any temperature, the object being to measure only changes of temperature, not absolute temperatures. The bulb and tube are so proportioned that a degree on the scale is nearly half an inch long, thus permitting great accuracy of reading.
In order, however, to determine just how much heat is required to raise the thermometer of this instrument 1º, it is necessary to compare it with one of the standard instruments, by exposing it to the sun at the same time.
This method of procedure, by which we determine the rate at which a sunbeam of given dimensions communicates heat to a measured mass of matter, is known as the dynamic method; it is somewhat inconvenient in requiring considerable time and a number of readings.
There is a different process for deducing the same results, which has been employed by Waterston, Ericsson, Secchi, Violle, and others, and may be called the statical method. It consists essentially in observing how much the sun will raise the temperature of a body, exposed to its rays, above that of the inclosure in which it is placed, this inclosure being kept at a fixed and known temperature by the circulation of water, or some such means.
Instruments based on this principle are called actinometers. Of these probably the most complete in its arrangements is that of Violle, described in his paper upon the mean temperature of the sun's surface, published in the "Annales de Chimie" in 1877. We give a diagram of the instrument (Fig. 3). It consists of two concentric spheres of thin metal; the outer, twenty-three centimetres in diameter, the inner, fifteen centimetres. The outer is polished on the outside, the inner is blackened on the inside. The space between the two spheres is filled with water, which is kept at a uniform temperature, either by mixing snow or ice with it, or else by a current circulated through it by means of the stopcocks tt. A sensitive thermometer, T, has its blackened bulb