]\gt ; received by the three insulated electrodes , and .
These were portions of a disc 7 cm .
in diameter , the central section being mm. wide and separated from the two equal side plates and by air gaps mm. a wide .
The narrow gaps between the electrodes were parallel to the direction of the slit in B. In the calculations it will be supposed that the electrode is 5 mm. wide , and that the side plates and come within mm. of the central line .
When no magnetic force is the charges and received by the electrodes and were equal and the centre of the stream fell on the centre of .
If be the charge received by the ratio depends on the electric force and the velocity of agitation of the ions .
In the previous experiments the factor by which the energy of agitation of the ions exceeded that of the surrounding molecules was deduced from observations of the lateral diffusion of a stream of circular section after traversing a distance of 7 cm .
under the action of the electric force .
The stream was received on a small circular electrode in the centre of a large metal ring , so that the proportion of the charge received by the small disc diminishes rapidly as the motion of agitation increases .
The apparatus was therefore suitable for determining the smaller values of the quantity The apparatus described above was better suited for the determination of large increases in the velocity of agitation and the values of may easily be found from the ratio , as is seen from the following investigation .
3 .
Let the origin of co-ordinates be taken in the centre of the slit , the axis of being normal to the plate and the axis of parallel to the length of the slit .
Since the central electrode is much than the slit the diffusion in the direction does not affect the number of ions that are received on the electrode , and it will only be necessary to find the motion in the directions and Considering the ions that pass at a uniform rate through a narrow section in the centre of the slit parallel to the axis of , the distribution in the electric field when the motion becomes steady is given by the equation* No When the electric force is measured in volts per centimetre , and the values of the constants No and are subsbituted , the equation becomes * J. S. Townsend , ' Boy .
Soc. Proc 1908 , , vol. 81 , p. 469 .
by neglecting the term is probably smaller than the experimental error .
It thus appears that the distribution of the quantity as expressed in terms of and is the same as the distribution of temperature , in terms of and , in an infinite solid initially at zero temperature throughout , except at the plane where the temperature has a constant value when The temperature is obtained from the equation .
and the solution given by Fourier is In the problem of the distribution of ions in the space between the plates $* and the sulface conditions are when for all values of except , so that in terms of and is given by the equation !
: $ The distance of the electrodes from the origin was 4 cm .
and the central } electrode was 5 mm. wide , so that the ratio , of the charge received by ] the central electrode to the total , is 1913 .
The Motion of Electrons in Gases .
339 $ where When various values are given to the quantity the corresponding ratios may be obtained from the tables of the values of the integrals .
It is necessary to take into consideration the width of the slit , since with the larger values of the proportion of the ions coming through a section of the slit near the edge that arrive on the central electrode is somewhat less than the proportion of those that come through at the centre .
The exact proportion for any section of the slit is easily calculated and the ratio when the ions come through all sections of the slit equally may be found in terms of The curve , , gives the values of in terms of 4 When the ratio is determined experimentally with an electric force acting between and may be found from the curve , fig. 2 , and the corresponding value of may be determined .
The curves , fig. 3 , give the values of found for air at various pressures .
The ratio corresponding to a given force and pressure depends on the dryness of the gas .
When the pressure is reduced to the required value , may be observed at intervals while the small quantity of moisture in the apparatus is removed slowly by the phosphorus pentoxide .
As the air dries the value of diminishes and after some weeks a constant minimum value is attained .
The continuous curves , fig. 3 , represent the minimum values obtained for dry air with different forces and pressures .
Several experiments were made with air which had not been dried very completely .
The values of are then practically the same as those obtained the amount of water vapour present , provided the amount exceeds a certain $ small value .
The dotted curve , fig. 3 , illustrateH the behaviour of air containing a small quantity of moisture , the pressule being mm. With a force of 1 volt per centimetre , which is the value of in the theoretical : curve at the point .
The two curves continue to coincide for some distance as increases , but when becomes 2 volts per centimetre , , as determined experimentally , is , the theoretical curve giving when .
The value of corresponding to the ratio is , so that when .
The value of reaches a maximum at and an increase in the force is accompanied by a remarkable increase in the divergence of the stream .
The ratio attains a minimum value 03 at the point where , the value of being .
Further increases in the force cause the divergence to diminish , and gradually the dotted curve takes a place between the two curves for dry air at and mm. pressure .
For values of exceeding 30 the values of are practically the same as for dry air at pressure .
' small .
For large forces the effect of the water vapour , and the .
electrons move freely with akinetic energy of agitation exceeding that of the surrounding molecules by the factor .
In the particular example illustrated by the dotted curve the effect of the water vapour disappears rapidly between the points A and and when the force exceeds 30 volts per centimetre the electrons move as in dry air .
When the amount of water vapour increases , the ratios and the forces corresponding to the maximum and minimum points A and , increase .
For different pressures , when the amount of moisture is approximately : proportional to the total pressure , the forces corresponding to the points .
and are also approximate]y proportional to the pressure .
This is interesting , : ; from a theoretlcal point of since these properties of the electrons depend on the velocity acquired between collisions with molecules , so that any particular effect should be obtained with the same value of if the proportion of the different gases that are present is not altered when the total pressure is altered .
5 .
The values of for dry air were obtained from the curves , fig. 3 , and it was found that depends on the ratio .
This is shown by the following examples of the numbers obtained from experiments at different pressures when is approximately constant:\mdash ; The agreement .
between the different determinations for values of between and was not as accurate as with the higher values .
This is probably due to the fact that small quantities of impurities have more effect when the lower forces are acting .
The results of the experiments on the lateral diffusion may be represented by a single curve giving as a function of .
The smaller values of from to are given in the curve , fig. 4 , and the larger values from to are given in the curve , fig. 5 .
Prof. J. S. Townsend and Mr. ] FIG. 4 .
FIG. 5 .
values obtained by Haselfoot for ions generated by ultra-violet light .
The conclusions to be drawn from these experiments depend on certain formulae that have been obtained for cases in which the velocity of agitation of the ions is large compared with the velocity under an electric force .
It will be seen from these investigations that when exc , eeds the ions in dry air are in the electronic state , so that the value of may be taken as .
The value of for a molecule of air being , the masses in the proportion , so that the velocity of agitation of an electron in thermal equilibrium with air is 230 times the velociby of agitation of a lectric forces acting exceed talue bactor rder thatmolecule oecond.ctual velocities ogitation w the ordinary formulae derived from the kinetic theory of gases should apply to the motion of the electrons it is necessary that the velocity in the direction of the electric force should be considerably less than the quantity .
: ' 6 .
The velocities were with the same apparatus for the same forces and pressures A transverse maguetic force was produced by a current in a pair of large circular coils outside the apparatus , the direction of the force being parallel to the slits .
The stream was then deflected so that the centre no longer fell on the centre of the electrode and the charge acquired by the electrode inc'reased with the force G. In these experiments the electrodes and were joined and the current in the coils was adjusted until the charge received by the electrodes and was equal to the charge acquired by the electrode .
The centre of the stream was thus deflected through mm. ( half the width of the central plate ) while the electrons travelled 4 cm .
in the direction of the electric force .
The magnetic forces required to deflect the centre of the stream through the angle when the ions are moving under the electric force X are given by the curves ( fig. 6 ) .
The different curves correspond to the different pressures of the air , which varied from to mm. The velocity under the electric force is given by the equation which expresses the condition that the direction of motion of the centre of the stream is along the resultant of the forces and We .
The latter force being small , the velocities may be assumed to be the same as those in the experiments on the lateral diffusion when the forces were acting . .
It follows that is a function of and the velocities may be expressed by means of a single curve .
The portion of the curve corresponding to values FIG. 8 .
7 .
The following table gives a comparison between the velocity of agitation and the velocity in the direction of the electric force , in terms of The velocity of a charged particle moving under an electric force is given proximately by the formula , when the mass of the particle is small compared with that of a molecule of the gas , and the velocity of gitation u is large compared with being the mean free path of the particle .
When the pressure is constant is also constant , and the velocity is proportional to when and are independent of .
In these cases is constant , being the mass of an electron , but increases with so that does not increase in proportion to the force .
The results of the experiments are thus in accordance with the theory , since increases less rapidly than when is constant ; in fact , the where is the mass of the charged particle , that of a molecule of the gas , the sum of the radii of the particle and a molecule , the number of molecules of the gas per cubic centimetre , and three-fourths of the reciprocal of the kinetic energy of agitation of the particle or a molecule of the gas .
Since the mass of the electron is small compared with a molecule the ratio reduces to , and is the mean free path of the electron .
The velocity of an electron thus becomes being the velocity of agitation of the electron .
The velocity is therefore independent of the mass of the molecules and is proportional to , the mean interval between collisions .
Langevin 's formula is deduced on the hypothesis that the mean energy of agitation of the particle is the same as that of a molecule of the gas .
When the masses are of the same order of magnitude the interval between P. Langevin , ' Ann de Chim .
et de Phys 1906 , , p. 246 . ! .
; must baken actual velocity ogitation , which exceeds that of agitation oarticle mHence iealing wlectrons telocity corresponding to a particle in thermal equilibrium with the molecules of tha .
gas by the factor a- : The value of for the electrffied particle may be obtained from the above .
equation by eliminating the quantity .
Thus mean faken awhere Nlimetre pressure t that tlectronand eiven bormula is cm .
Hence .
The values of thus obtained from the determinations of and corresponding to the different values of are given in the above table ( p. 345 ) .
The numbers obtained for do not differ very much from the value found by the more acourate methods uuder conditions in which the effects of collisions between electrons and molecules may be neglected .
The differences between the values found for for different values of are not entirely due to experimental errors , although the probable error in the final results is greater than that in the direct measurements , since the square of the velocity is involved in the formula for .
The nature of the collisions between electrons and molecules may not resemble the collisions between elastic spheres to the same extent for different values of the velocity * Mr. F. B. Pidduck ( p. 296 , supra ) has made a theoretical investigation of the motion of ions in gases ; he shows that when the negative ions are in the eIectronic state their velocity of agitation may become abnormally large , and the above values of and may be explained on the hypothesis that the collisions between electrons and molecules are such as would occur between imperfectly elastic spheres .

