274 



BELL SYSTEM TECHNICAL JOURNAL 



to be easy to maintain, though the vapor pressure of carbon is im- 

 measurably small ; one is led to suspect the gases inevitably occluded 

 in this element.' Saunders produced waves as short as 978A with an 

 arc in calcium \apor, and Simeon waves down to 375A with a "carlxin 

 vacuum arc." 



These vacuum arcs are started either by heatini; the electrodes to 

 produce a momentary high vapor-density, and appKing a transient 

 high \()ltage between them; or by touching them together and drawing 

 them apart while the moderate voltage is applied. If the latter 

 method is tried when the voltage is too low to maintain an arc, there 

 is a transitory flash, the breakspark; its spectrum in the \isible region 

 has been noticed by von Welsbach, who finds the relati\e intensities 

 of certain lines strangely altered from what they are in the ordinary 

 spark; but according to McLennan and Lang, it yields no rays of 

 wave-length inferior to 2()00A. 



The vacuum spark or hot spark cnipUixcd 1)\' Miliikan and his 

 associates is an altogether different affair; it is a brilliant spark which 

 occurs between electrodes a millimetre or so apart (the limits 0.1 

 mm. and 2 mm. ha\e been assigned) in an extremely high vacuum, 

 when a transient potential-difference of the order of several hundreds 

 of thousands of volts is laid across them. This is a mysterious phenom- 

 enon, which has been studied by several scientists, without satisfactor\' 

 conclusions. Whatever the vacuum spark really is, there is no doubt 

 that it exists, and that wave-lengths are found in its spectrum which 

 are shorter than any hitherto observed in any spectrum of arc or 

 spark; and it is likely that the.se high-frequency rays are not excited 

 at all in the ordinary electrical discharges of relatively low \oltage, 

 -SO that the high vacuum provides the conditions for stimulating as 

 well as for transmitting them. The least wa\'e-length yet measured 

 with an optical method (ruled grating), which is LSGA, occurs in 

 the sjiectra of some of these sparks. 



Most difficult of all is oln-iousK' llic iiniblriii lA deli'cting the rays 



emitted by the atoms or molecules of a permanent gas, which must 



of necessity occupy the entire path of the light from the place where 



it is excited to the place where it is recei\"ed, unless intercepted by 



a solid partition which would intercept the desired waves also. If 



the dischariii-tiibc coiilaining the luminous gas communicates only 



li\ a n.irrciw >lii wiiii the chamber containing the diffracting and re- 



ceixing ai)paratus, it is practicable to connect a ]K)werful pump to 



" The iiiininuim niainlaiiiinK voltage for arcs in vacuo is given l)y Simeon as 

 follows, for electrodes of the following materials: C.W to 40 volts, A'n M) to 40, .1/ SI) 

 to 100, .SV 9.5 to ID.S. The distance l)etwccn the electrodes is described as "slight," 

 the degree of vacuum before arcing is not stated. 



