173 The Effect of Small Traces of Water Vapour on the Velocities of Ions produced by Rontge Rays Air .
By Robert Tabor Lattey .
( Communicated by Prof. J. S. Townsend , F.R.S. Received May 25 , \#151 ; Read June 9 , 1910 .
) Some experiments by Prof. J. S. Townsend* on the lateral diffusion of a narrow stream of ions moving in an electric field led to the conclusion that negative ions in perfectly dry air are much smaller than those in air .containing a small quantity of moisture .
It was consequently to be expected that the complete removal of water vapour would cause an increase in the velocity with which negative ions move under the influence of an electric field of force .
At his suggestion the following investigation of the velocities of ions in air at low pressures was undertaken , and it was found that , while the complete removal of water vapour had only a small effect on the velocities of positive ions , yet the same cause increased the velocities of the negative ions by a factor varying between 2 and 30 for the range of pressures and electric forces used in the experiments .
The method adopted was to let the ions travel between two gauzes under a known electric force for a time t and then to reverse the field .
If v is the velocity of the ions and d is the distance between the gauzes , then ions starting from one gauze will reach the other if If t is gradually decreased , it is possible to find , by means of an electrometer , when ions cease to reach the second gauze ; when this happens v \#151 ; djt .
The ions were generated by X-rays passing through an annular gap between a guard-ring ( A , fig. 1 ) and a thick lead disc ( A ' ) , and falling on an annular brass plate ( C ) .
By connecting A and C to the poles of an insulated battery { Bi ) , the ions of one kind can be driven downwards to the gauze which is fastened to C , or driven upwards and discharged by A ' .
Below C is a series of concentric rings insulated by ebonite blocks , and between each ring is a coil of a certain high resistance .
The lowest of these ( D ) has a gauze similar to that attached to C. Below this gauze is a brass disc ( E ) , insulated and surrounded by a guard ring in metallic connexion with the base of the apparatus .
E is connected to an electrometer .
The gratings , or gauzes , attached to C and to D were carefully constructed of fine brass wire , 0T mm. in diameter , and stretched at intervals of l'O mm. * ' Roy .
Soc. Proc. , ' 1908 , A , vol. 81 , p. 464 .
174 Mr. E. T. Lattey .
Effect of Small Traces of [ May 25 , across the opening , so that they did not rise appreciably above the plane of the lower surface of C or the upper surface of D. A small batteiy ( t\gt ; 2 ) attached to I ) and to the base maintains a P.X ) .
between I ) and the guard-ring of the same sign as that between A and C. The ring C is attached to a rotating commutator , by means of which connexion is made for equal periods of time to alternate ends of a large battery of cells ( B3 ) .
An intermediate point of this battery is connected to D. Suppose that a determination of the velocity of negative ions is being-made , then A is made negative with respect to C , and D is negative with respect to E and the guard-ring .
In this way negative ions are sent downwards into the part of the apparatus between C and D , and any of these which reach D are sent on to E and detected by the electrometer .
The equal resistances between the rings and the presence of gauzes at C and D ensure uniformity in the potential gradient between these two surfaces .
The gauze at D , being at a constant potential , also prevents effects due to induction on E , due to changes in the field between C and E. The alternating E.M.F. between C and E is made larger in one direction than in the other , so that any ions starting from C which do not reach I\gt ; while C is negative with respect to E are completely removed during the period in which the field is reversed .
1910 .
] Water Vapour on the Velocities of Ions , etc. Negative ions coming from the upper grating will therefore move downwards while the first field is on .
If they reach the lower grating before the field is reversed their presence will be detected by the electrometer , but if the field is reversed before this has happened the electrometer will be unaffected by them .
If , therefore , the commutator is run sufficiently fast , the electrometer should show no deflection ; but at a certain speed of rotation , depending on the velocity of the ions and the distance between C and D , a small deflection will be observed .
A brass cylinder , having in its upper surface an aluminium window , was placed over the apparatus .
The joints were made air-tight , and a side tube in the brass cover was attached to a two-way tap communicating either with a small bulb F , containing water , or with a tube G , packed with phosphorus pentoxide ; this in turn communicated with a T-piece , on one limb of which was a tap by which air could be let into the apparatus , while the third limb was sealed on to a small bulb ( H ) containing soda-lime and on to an arrangement by which small changes could be made in the pressure of the air .
A gauge and a Toepler pump were also connected to the apparatus .
The Commutator.\#151 ; A heavy brass disc had four equal segments cut away symmetrically from one side , and was mounted on an axle .
A similar disc was fastened to this by bolts with insulating washers in such a way that the central part of the rim of the compound .
disc was formed by alternate portions of the two components .
Brushes touching this portion of the rim would therefore make contacts with the right and left halves of the disc alternately ( fig. 2 ) .
To ensure good contact , two graphite brushes were used at opposite ends of a diameter .
The framework of the apparatus was connected to one pole of the battery and a brush touching the insulated disc to the other pole .
To ensure smooth running , the small gaps between segments were filled with vulcanised fibre ( see fig. 2 ) .
The commutator was tested by connecting one set of segments in series with a battery and voltmeter .
When the brush of the commutator rested on one of these segments the voltmeter gave a reading of Y , which was reduced to V/ 2 on rotating the commutator , so that the brush must have been in good contact with the segments all the time the commutator was revolving .
The number of revolutions per minute wTas taken either by a speed indicator or by a counter .
In this way two values for the time could 176 Mr. R. T. Lattey .
Effect of Small Traces of [ May 25 , be obtained at most speeds .
They were found to be in satisfactory agreement .
J The commutator was driven by an electric motor and shafting , and its speed could be altered by changing either the resistance of the motor circuit or the ratios of the diameters of the pulley wheels .
In this way the time of one revolution could be varied between about 0*65 and 002 second .
Method of Experimenting.\#151 ; The apparatus was exhausted as completely as possible and left overnight .
Air was then let in through long tubes packed with soda-lime and phosphorus pentoxide until a suitable pressure was reached .
By raising and lowering the pressure regulator the air contained in the apparatus could be drawn backwards and forwards over the pentoxide and so completely dried .
In order to obtain accurate results it is necessary to make two experiments for each speed of rotation of the commutator , in order to eliminate from the charge reaching E the effect due to the few ions which are generated by stray radiations between the gauzes .
The deflection Si is found when the force from A to C is in the same direction as that between D and E. A second experiment is then made with the field between A and C reversed , and a smaller deflection 82 is obtained , which is due to ions generated below the gauze C. The quantity Si \#151 ; S2then represents the deflection due to ions coming through the upper gauze .
Each observation was made by allowing the rays to pass into the apparatus for 30 seconds , and noting the deflection of the electrometer and the speed of rotation of the commutator .
The deflections were then plotted on squared paper against the time of one revolution , and by varying the speed of revolution a time could be found for which no additional deflection was obtained due to ions originally starting from C , i.e. , a value for T for which 8i\#151 ; 82 = 0 ( see fig. 3 ) .
Observations were made both with positive and with negative ions ; the positive ions appeared to be but little affected by extreme drying , but with negative ions very large effects were observed .
Positive Ions.\#151 ; Although the experiments were made alternately with each kind of ion with the same sample of air , yet it seems more convenient to discuss the two kinds of ions separately .
10 20 Deflection Fig.3 .
1910 .
] Water Vapour on the Velocities of Ions , etc. When the gas was very dry it was found that the velocity of the positive ions was directly proportional to the force ( X ) and inversely proportional to the pressure ( p ) .
In fact , up / X was constant , where is the velocity .
p ( mm. ) .
X ( volts per cm .
) .
X/ j\gt ; .
v ( cm .
per sec. ) .
up / X 14-3 1 -66 0 T16 129 -6 1116 1 -785 0 T25 144-3 1153 2-04 0-143 161 -0 1127 18 -6 1 -94 0-104 116 -0 1110 2 27 0-122 136-5 1116 2 -53 0-136 151 -5 1112 24 5 2 -53 0-103 112 -5 1095 2-70 o-iio 140-0 1305* 28 '8 2-70 0 -0937 107 -0 1140 Mean .
# No. 8 was omitted in calculating the mean .
This is in accordance with the determinations of other observers who did not take especial precautions to dry the air in their apparatus .
The following are the principal values of up / X available for comparison :\#151 ; Langevin . .
1105\#151 ; 1076* Chattock f 996 ( wet)S Zeleny f 1030 ( wet)f LI 071 ( dry ) c 1040 ( dry ) Phillips ... 1060|| Wellisch ... .
1170+ Przibram f 908 ( dry)IT L 841 ( wet ) The values of up / X found by different observers seem to depend in a small measure on the value of X/ p used in their experiments .
It is a little difficult to find out from Langevin 's papers the actual values of X used by him , but apparently they were such that X/ p varied between 003 and 1*3 ; Zeleny used only very small values of X/ p ( about O'Ol or less ) .
After admitting moisture to the apparatus the following result was obtained:\#151 ; p ( mm. ) .
X ( volts per cm .
) .
X/ p. V. up / X. 18 -0 2-315 0 -1285 101 -3 780 * ' Ann. Chim .
Phys. , ' 1903 , vol. 28 , p. 514 .
t ' Phil. Trans. , ' 1900 , A , vol. 195 , p. 193 .
J ' Phil. Trans. , ' 1909 , A , vol. 209 , p. 249 .
S ' Phil. Mag. , ' 1899 , vol. 48 , p. 401 , and 1901 , vol. 1 , p. 79 .
|| 'Roy .
Soc. Proc. , ' 1906 , A , vol. 78 , p. 167 .
IF ' Wien .
Ber .
, ' Ila , 1908 , vol. 117 , p. 665 , and 1909 , vol. 118 , p. 331 and p. 1419 .
178 Mr. R. T. Lattey .
Effect of Small Traces of [ May 25 , The comparatively large effect observed here compared with those observed by Zeleny , Przibram , and Chattock is probably due to the much larger percentage of water vapour admitted in this experiment , since the total pressure was small .
After drawing the air backwards and forwards over phosphorus pentoxide until the gauge had ceased to alter , the pressure had fallen by 1*4 mm. As water was only admitted to part of the apparatus its actual pressure will be greater than this .
The ratio of the volume of the whole apparatus to that of the wet part was 1*495 .
Hence the actual pressure of water vapour will be 2*1 mm. , or about 11*5 per cent , of the whole pressure .
How air saturated with moisture at 20 ' and 760 mm. only contains 2*3 per cent , of moisture .
The above experiment was therefore made with about five times the percentage of water vapour used by Przibram .
Negative Ions.\#151 ; The velocities observed in the case of negative ions in dried air were greatly in excess of those observed by former experimenters .
The velocity is apparently a function of which is independent of the actual values of either X or p ; this is shown by the following table of values read off from the smoothed curves drawn for observations made at each of the pressures mentioned :\#151 ; p ( mm. ) .
X/ p = 0-04 .
0-05 .
0-06 .
0-07 .
0-08 .
J 0 -09 .
0*1 .
14 -3 107 175 -5 310 580 1126 ( 2200 ) ( 4210 ) 18 -4 103 163 279 514 -5 1006 ( 2050 ) ( 4120 ) 24*5 ( 119 -5 ) 172 -5 286 509 936 -5 1799 ( 3480 ) 28-8 116 180 298 -5 519 -5 926 ( 1652 ) Mean 112 5 173 -2 287 -8 510 -5 953 -5 1845 3635 The values here given as means were obtained from the smoothed curve drawn after plotting all values of v against X/ p. The values in brackets were obtained by extrapolation .
No weight can be ascribed to the apparent slight rise in the value of v when p decreases , since it must be remembered that when p = 14*3 mm. and X/ p = 0*08 , then an alteration of 1 mm. in the value of will change the value of v by about 60 cm .
per second , while an error of 0*001 secoud in determining the time of revolution of the commutator will affect the value found for v by about 40 cm .
per second .
Por velocities over 1000 cm .
per second it is necessary to run the commutator at more than 1100 revolutions per minute .
There is a possibility that at these speeds contact may not be perfect , and though the contact at high speeds was frequently 1910 .
] Water Vapour on the Velocities of Ions , etc. 179 checked with a voltmeter , and any results which appeared doubtful rejected , yet there remains a fair margin of possible error in the higher velocities .
It should be noted , however , that the sources of probable error indicate that the values given for v are , if anything , too low rather than too high .
The actual valueS obtained are given in the following table:\#151 ; No. p ( mm. ) .
X ( volts per cm .
) .
X/ p. v ( cm .
per sec. ) .
1 14-3 0-60 0 -0421 118 2 0-888 0 -0621 352 3 1 -165 0 -0815 1244 4 1 -415 0 -0990 2670 5 18 -2 1 -185 0 -0651 389 6 1 -42 0 -0809 1075 7 18 -6 0-875 0 -0471 137 8 1 -16 0-0622 292 9 1 -375 0 -0740 645 10 24 -5 1 -385 0 -0566 239 11 1 -385 0 -0566 218 12 1 -54 0 -0630 339 13 1 -76 0 -0717 451 14 2-03 0 -0829 1120 15 2-03 0 -0829 1340 16 2 -33 0 -0952 2550 17 28 -8 1 -oo 0 -0347 90 -5 18 1 -24 0 -0432 134 19 1-715 0-0596 294 20 1 -94 0 0677 412 21 2*20 0 -0764 800 22 2-30 0 -0801 932 23 2-38 0 -0826 920 After completing these observations a small quantity of water vapour was admitted , and an observation of the velocity obtained , p = 18-2 , X = 214 , X/ p = 01107 , v = 1302 .
The value of v corresponding to this value of X/ p from the extrapolated curve for dry air is about 7500 cm .
per second .
A small further quantity of water reduced the value of v to 875 cm .
per second .
The two-way tap was shut so that no further quantity of moisture could get in and so that communication with the rest of the apparatus was also stopped .
The next morning the value of v was observed with the same values of X andp ; it had now risen to 7000 cm .
per second , owing , no doubt , to the condensation of moisture on the walls of the apparatus and on the silk covering of the resistances .
A fairly large quantity of moisture was now admitted ( about 2T mm. pressure , see p. 178 ) , and the following values obtained:\#151 ; 180 Effect of Small Traces of Water Vapour , etc. X. \#166 ; x/ * .
V. up / X. 18 -5 1 -715 0 -0927 99 -4 1070 19 -3 2 -30 0-1191 110 924 18 -0 2-29 0-127 113 890 The first value was obtained immediately after admitting water , and the other values some hours later .
The more complete diffusion of the water vapour through the gas may account for the relatively lower values of these later experiments .
All these values , together with those of other observers , , are shown on the accompanying graph .
X Observations Dry Air o Observations in Moist Air OOI 002 003 004 005 0 06 007 008 0 09 0-10 Variation with Temperature of Viscosities of , etc. 181 The only values available for comparison with these for ions in moist air are those of Zeleny , who obtained values for ranging from 960 to 1080 , hut his work was done with values of X and of very different from those considered here , and the experiments of Przibram , who obtained for up / X the values 1005 in moist air and 1160 in air partially dried by sulphuric acid .
On the Variation with Temperature of the Viscosities of the Gases of the Argon Group .
By A. O. Bankine , D.Sc .
, Assistant in the Department of Physics , University College , London .
( Communicated by Prof. F. T. Trouton , F.R.S. Received May 25 , \#151 ; Read June 9 , 1910 .
) In continuation of the determinations of viscosity already recorded , * measurements of the temperature variation for the five gases of the argon group have now been obtained .
The apparatus used was of the same type as that described in the above papers , with certain modifications to render it suitable for working at different temperatures .
Unfortunately , the method is only practicable for a somewhat limited range of temperature .
On the one hand , the temperature must be lower than that at which the vapour pressure of mercury becomes appreciable , on account of the very high viscosity of the latter ; on the other hand , it must not approach the solidifying point of mercury .
At 100 ' C. the vapour-pressure of mercury is about 1/ 5 mm. This was judged to be as much as could be safely neglected , and observations were therefore taken at two temperatures only , viz. , atmospheric temperature and the temperature of steam .
However , although these limitations render observations over a large range impossible , they still allow interesting comparisons to be made of the temperature variations of the different gases .
Incidentally , redeterminations of the ratios of the viscosities of the gases to that of air have also been obtained with an entirely new apparatus .
Apparatus .
Fig. 1 represents the apparatus in its modified form .
The capillary tube was bent twice as shown in order to reduce the total length of the apparatus * ' Roy .
Soc. Proc. , ' A , vol. 83 , pp. 265 and 516 .
VOL. LXXXIV.\#151 ; A. O

