Focal Isolation of Long Heat-Waves. 251 



Details regarding the micro-radiometer will be found in 

 previous papers *. 



In the present case the instrument was protected by an 

 air-tight metal helmet provided with a quartz window. The 

 sensitiveness was such that a candle at a distance of 2 metres 

 gave a deflexion of 700 scale-divisions, in spite of the quartz 

 window. 



The great advantage of this method of focal- isolation is 

 that large angular apertures can be used (in the present case 

 F 3*5), and the radiation is weakened only by reflexion from 

 and absorption by the two quartz lenses. Its only disad- 

 vantage is the circumstance that a rather wide spectral 

 range is transmitted, for all radiations for which the re- 

 fractive index is in the neighbourhood of 2*14 are passed 

 through the apertures. 



For wave-length 63 fju the refractive index is 2*19, decreas- 

 ing to 2*14 as a limiting value, with increasing wave-length f. 

 The very powerful absorption of quartz for radiations 

 ranging from 60-80 fju prevents the shorter waves from 

 getting through the system. Beyond 80 //, the quartz begins 

 to show stronger transparency, 17 per cent, of the residual- 

 rays from potassium iodide passing through a quartz plate 

 of 18 cm. thickness, corresponding to the amount of quartz 

 in the path of our rays %. 



This circumstance gives to our transmitted radiation an 

 energy curve very steep on the side towards the shorter 

 wave-lengths, while on the other side it slopes down gradually, 

 at the rate determined by the energy curve of the source of 

 light, decreasing with the 4th power of the wave-length, 

 if the Welsbach mantle has no selective properties in this 

 region, as is probably the case. The increasing transparency 

 of the quartz with increasing wave-length will make the 

 face of the energy curve still more gradual, and we should 

 expect our isolated radiation to have an unsymmetrical 

 energy curve with a maximum shifted towards the shorter 

 wave-lengths with respect to its centre of gravity, and to 

 extend from 80 towards the longer wave-lengths over a 

 range of more than an octave. This expectation was ful- 

 filled, as the results of our experiments will show. 



The 40 mm. deflexion of the micro-radiometer was reduced 

 to zero by interposing a plate of rock-salt, 3 mm. thick, 

 into the path of the rays, while a quartz plate of 4 mm. 



* II. Schmidt, Ann. der Flips, xxix. p. 1003; H. Rubens and H. Holl- 

 nagel, Phil. Mag. [6 J xix. p. 764. 

 f H. Rubens and E. F. Nichols, Wied. Ann. lx. p. 418. 

 \ Rubens and Hollnagel, he. eit. 



