The same is done for noting the variation in temperature by measuring the space between the ends of the iron and zinc rods in each case. In order to leave no cause for error, the two ends of each rod are cut to sharp edges and made one vertical and the other horizontal, and the horizontal end of one rod faces the vertical end of the other. The same is done for the zinc and iron. When one of the cases is moved forward and placed in position ahead of the others, all the measurements between any two rods are repeated in order to detect whether any of the other rods has moved while the other was being placed in position. Spirit-levels are on each case, and their variation is carefully measured with microscopes. All the data are registered, and have afterward to be carefully gone over, and all compensations for error, temperature, etc., duly allowed. This apparatus is known as Bessel's base-measuring apparatus, the chief feature of which is the metallic thermometer with iron and zinc.
The length of a geodetical base varies according to the area to be covered by the net and to the possibility of finding a good ground for laying such a base. From five to ten miles would be considered a good geodetical base, one longer than the higher figure not being advisable on account of the possible increase in the inevitable error of such measurement. Even bases of about two miles have been measured, but these could not be used for very large triangulations, and are more intended as a check on other measurement and as a means of compensating errors. The ground has to be carefully leveled before the base can be measured. The two end-points of the base are marked by stone monuments which can be seen at a distance, and the whole length is divided into so many sections, each of which constitutes a day's work, stone and metallic tablets being laid and the line marked in advance before the actual measurement can be undertaken.
It is not necessary that the base should be actually measured before the real triangulation work—that is, the measuring of the angles at each of the stations of the net—is begun. The work of measuring a base is necessarily very slow. All apparatuses have to be carefully tested before being actually used in the field. Each of the rods has to be subjected to a series of experiments at different temperatures, in order to determine the actual expansion of the iron by comparison with that of the zinc. The actual measuring in the field occupies one or two months' time, and may be longer if the weather is not favorable, the whole distance, divided into sections, each of which constitutes a day's work, being measured at least twice, once in each sense. Five or six experienced operators are required, besides a number of assistants to do the menial work—such as carrying the rods forward, etc. The calculations are, later on, done in the office, and are in themselves a very slow and exhaustive process. Taking all together, the time required is little short of one year, including the preparatory experiments and the calculations.
