SOME CONTEMPORARY .mr.lNCnS IN PHYSICS— III 273 



question that matters; and radio-frequencies penetrate great tliick- 

 nesses of rock or brick more readily than the thinnest slieet of metal 

 foil. 



To explore the region of the spectrum in wliich the absorbing- 

 power of matter is at its greatest, it is necessary to make a high 

 vacuum over the entire path of the rays from their source to the 

 receiver (photographic plate, ionization-chamber, or electrode for 

 photoelectric emission). This necessity can be escaped only if the 

 obligation of measuring wave-lengths is evadcfl, for then the path may 

 be very short; the receiver may be brought quite close to the piece 

 of solid substance or the stratum of gas in which the rays are excited. 

 If the wave-lengths are measured, it must be done with a ruled or 

 crystalline diffraction-grating, which enforces a lengthy path (often 

 as much as two metres). No solid windows can be interposed in it 

 to confine a diffusing gas to the region where the ra\s are excited (the 

 only exceptions yet developed are Holweck's .0001-mm. celluloid 

 windows, which when stretched over and sustained by a fine-meshed 

 gauze are said to be able to support a o-cm. pressure-difference l)etween 

 their two faces). The excitation must therefore take place, whenever 

 f)ossibIe, in vacuo. This is simple enough when dealing with the rays 

 excited from solids by electron-bombardment, and originating from 

 displacements of electrons deeper down in the atomic system than the 

 valence-electron; for the bombardment can be carried on in vacuo. 

 But the arcs and sparks which are commonly used to displace the 

 valence-electrons of free atoms or molecules, and so produce the fre- 

 quencies for which these are responsible, are usually operated in an 

 atmosphere composed of a comparatively few of the atoms being 

 studied, mingled with a large amount of air or some other permanent 

 gas. \'et it has been found possible to operate both arc and spark 

 discharges "in vacuo,", that is, without the atmosphere of permanent 

 gas; though they differ in various ways from the like-named and 

 familiar discharges in air, and do not display quite the same spectra. 



Vaniiim arcs, when once ignited, can be maintained with a moderate 

 voltage between electrodes of various metals; the mercury vapor 

 lamp is the familiar example, but arcs of such metals as magnesium, 

 aluminium, and lead were developed as early as 190.5. The name 

 "vacuum arc" is, of course, a misnomer; the discharge occurs in an 

 atmosphere of the vapor of the metal, but this congeals as soon as it 

 starts to diffuse away from the discharge, and does not impair the 

 vacuum in the light-path. The condition for an easily-maintained 

 vacuum arc is that the vapor-pressure of the metal involved be com- 

 paratively high. Yet arcs between carbon electrodes in vacuo seem 



