]\gt ; 1903 .
] ctrical of 415 " " The Electrical Conductivity arted to a Vacuum by Hot Conductors By O. W. , B. , Fellow of Trinity .
Communicated by Professor J. J. .
Received Iarch , 1903 .
( Abstract .
) The experimental of is atio of the 'electrical collductivity of the space hot of carl ) , and sodium , low .
In addition , the first portion of the paper is occupied in dedncing a theory by which the experimental results are .
Some of the that have been obtained with platinnm were described in a preliminary before the Cambridge Philosophical Society on ) , 1901 .
The present paper is ided as follows:\mdash ; A. Theoretical investigation .
1 .
Calculation of the }ttion clrrent .
Equilibrinm of hot conducting } ) of infinite B. investigation .
1 .
xperiments with platinnm .
2 .
riments n 3 .
Lxperiments with innl .
C. Conclusio ] ] .
The periments show that the negative leak from hot lvires at low pressures is a definite ) ctio of the temperature of the wire , increases rapidly as the temperature is raised .
Professor lIcCIelland* had previously found this leak independent of the pressure at pressures less than mm. , whiist Professor J. J. Thomhad shown in addition that current was carried by corpuscles or electrons .
The theory here put to xplaiu these facts , and those to described later , is 1 , ased on the corpuscular theory of condnctio in metals .
On that view metal contains a very great number of free corpuscles whose free path is comparable with that of molecule in air at atmospheric pressure .
The corpuscles must , therefore , be moving with a distributionl of velocity given by the Boltzlaw .
Since the corpuscles do not escape from the metal at ordinary temperatnres , it is evident that there must be discontinuity of potential at the surface which prevents their escape .
* Camb . .
Soc. ' vol. 10 , p. 241 , and , p. 296 .
Phil. Mag , p. 547 .
41 Mr. O. W. ardson .
'le On the temperatnrc of the , the of the coris increased , and at high enough temperatures some of them -ill to shoot the into the mding space .
The of corpnscles which escape any temperalure calcnlated on this vicw .
The tion Crent , hich c spouds to the ) ) emitted per , is ) equation he is the number of frec corpusclcs in 1 .
of the the charge on an the of the hot-metal surface , the gas constant for a single hose m is ?
) , the temperature , and the work by an ion in ssing throngh the surface layer .
The which energy is emitted , to , is to the part which the Qation from hot odies played in certain recent cosmological theories , the tions hich determine the eqnilibrinnl of corpuscles a plane of hot nletal of infinite extent are also given and The chief which has been attacked experimentalIy is the itl lvhich the saturation ctlrrent from a } surface to ries 1 the turc of the In the case of platiIltlm , the hot metal con sisted of file wire passing alorg the axis of a stlrromding trod .
The temperature obtained from the resistance of the wire .
In the of rbon , the leak was meastlred from a small lamp meant to a ulding cylinder .
The tnre was estimated in tn stening a platinum and thermocotlple of very fine wire round the meant , and from resistance of the filament .
sodinm this method could 1lot bs adopted .
The nlctal distributed on the inner snrface of steel the ctlrrent from it to an insulated wire the lvas .
The temperature was obtained by a thermo-couple of nickel .
Owing , doubtless , to the peculiar of the eleclrodes and the some- high of the gas , the lrrent soditlm nevel .
For reason the uudeloltage 1 measured of curlent .
, it to examinc the tion I ) etweetl the the pplied e foIce .
Curcent arc givcn thrcc ) tances , an the case of rbon , to for a ttble range of pressures .
To account for the thts ( the higher , it is necessary to that ions by collisions .
The riation of lrrcrt is mined ovel tho ving range:\mdash ; For platinum from to per .
cm .
of , , , , 3 , , , 10 , , rent .
The corresponding ranges of platiuum and sodiun ) roughly from 1000o C. to 1600o C. , and from C. to C. speetively .
The small current ; from were measured with quadrant electl'ometer , but general rule , oalvanometer with snital ) shunts used .
Perhaps the most result of the investigation is the ) tively currcnts have been obtained .
The filament to an electrode placed neal it ; this corresponded to a current of 2 amperes per sq .
of the carbon surface , the potcntial on the filament being volts .
The pressnre iu this experirllent only mm. This and some of the rrents f sodium were on In all these tho vcre too smitll to maintain a discharge between the odes .
Throughout the above , the tion -een the satnration current turo was to be lted by a formla of the type A and constants for conductor .
The fues -hich been for these Fo 9 .
11 .
oditUll , of A very ) idly with the vahle f ) , so only its ordor of magnitude is given .
The number of in a .
of the metal is from A. For platinum this gives ]0-1 , founrl 10 In the of the other condnctors , mbel .
fonnd ) surdly gompared }-ith * ' Phil. lIag .
' , 1 ) .
418 ctivity of [ Feb. 28 , The can ) made to disappear by assuming a small peratnre riation of This assumption is shown to be consistent the general nature of The work required to drive an ion through the face layer is ated , in each , from the value of , to which it is proportional .
Dividing by the charge on an ion this yields the discontinuity of potential at the surface of the conductor .
The values found for this are : for sodium , voIts , for platinum ) for rbon Solts .
These numl ) are inversely proportiollal to the cube roots of the respective atomic volumes .
This leads to the concIusion that the work required to force a corpusole out of a metal varies , approximately , inversely as the cube root of the fttomic volume of the metal .
In all these experiments , the curtent when the hot is charged positively is small with that obtained with the metal tively charged .
Only in the case of sodium was the positive current large enough to deflect a sensitive galvaIlometer .
The results which have been obtained are shown to furnish a complete explanation of the phenomenon known as the Edison effect .
The fact that such enormous currents are obtained at such very low pressures confirms the conclusion that the are not produced from the gas ) its interaction with the metal .
Calculation shows that to tain the currents with carbon , each gas molecule would hnve to give rise to each time it oollided with the hot metal surface .
The energy lost to the escape of the puscles is compared ith the energy emitted in the form of electromagnetic tion .
The former is shown to ) maller than the latter at the temperatures at which measurements been made , but it increases more rpidly with the temperatnrc .

