Measurements in the Extreme Infra-red Spectrum. 765 



variation often attaining several millimetres. That this varia- 

 tion was not due to air-currents or to disturbances o£ a 

 mechanical nature was demonstrated in the following 

 manner : — A Hefner-Alteneck pressure variometer and the 

 radiomicrometer were simultaneously observed, and an 

 absolute parallelism in the disturbances o£ the two in- 

 struments noted ; this indicated that small deflexions were 

 caused by the spontaneous pressure variations of the air. 

 Apparently the adiabatic compression and dilatation of the 

 air, producing s/nall temperature differences, cause the 

 observed effect, in view o£ the difference in the heat capa- 

 cities of the two junctions of the thermo-element. To 

 obviate this a vacuum-tight bell-jar, with a quartz window, 

 was constructed : under this the radiomicrometer gave a 

 very quiet zero-point. Fig. 6, curve y (PI. XIII.). shows a 

 series observed with this instrument, provided with the 

 bell-jar, the latter having been used only on the longest- 

 waved " Reststrahlen " of potassium iodide. 



Production of Rest str olden. 



With the aid of this apparatus the wave-lengths and 

 energy distribution of the Reststrahlen of rock-salt, sylvine, 

 potassium bromide, and potassium iodide have been investi- 

 gated. In every instance we limited ourselves to the use 

 of four reflecting surfaces and increased the purity of the 

 rays by means of a rock-salt shutter instead of the usual 

 metal screen * (D, fig. 2). In spite of these precautions 

 the Reststrahlen did not prove to be entirely pure ; they 

 still contained an additional fraction of shorter infra-red 

 rays, which, in the case of Reststrahlen of rock-salt and 

 sylvine, amounted to 2 per cent., and in the case of Rest- 

 strahlen of potassium bromide and potassium iodide about 

 8 and 20 per cent, respectively. In the interferometer 

 measurements of the wave-lengths this impurity was insig- 

 nificant in its effect ; a uniformly increased intensity of 

 radiation being the only observed result. On the other 

 hand, in all measurements concerned with the reflectivity 

 and absorbing power of various substances, the impurity 

 was determined and taken into consideration. 



To produce the Reststrahlen of rock-salt four plates 

 2 cm. thick and 10 cm. square were used. These were 



* As is well known, such a screen allows nearly the entire radiation of 

 the source to pass, excepting the small fraction which is reflected on its 

 surface, while the rays which we here seek to isolate are completely 

 absorbed. The deflexion we therefore obtain on raising- the shutter is 

 caused almost entirely by the "Reststrahlen" (see H. Rubens, Ber. d. 

 Phys. Ges. Nov. 1896). 



