Page:Proceedings of the Royal Society of London Vol 69.djvu/118

110

In the course of the work numerous determinations of the refracting angle were made at different temperatures ; of these the following pair may be cited, each value being the mean of several de- terminations : This difference is quite insignificant. It is possible that differences in the refracting angle at different temperatures, noticed by previous observers, may be due not to any alteration in the prism by itself, but to the expansion of the wax or other cementing material, as to the thick- ness of which and the method of spreading it details are not given.

In the more recent experimental work upon refractive indices the prism has generally been enclosed in some form of jacketing arrangement through which a stream of water at about the required temperature was allowed to flow. Some of the arrangements suggested increase the difficulty of manipulating the prism in any direction, while for every case it appears probable that, if observations are required at a temperature considerably above that of an ordinary temperature, an optical strain might result from convection currents from the hot jacket, and thus observations at a higher might not be comparable directly with those at a lower temperature.

Some observers have remarked upon the number of constants required for an empirical equation

to fit the observations ; possibly the explanation may be found in such a strain as described above.

As in this particular investigation the substance under examination may begin to decompose at a temperature of about 30, the plan was adopted of altering the whole temperature of the room so that not only the prism, but also the lenses, telescope fittings, micrometer scales, &c., should be affected pro rata as regards their individual co- efficients of expansion.

It was judged that, when the thermometer in the prism contents and a second of the same order of delicacy placed on the rotating table registered the same temperature for a time longer than that