XLVIII. The Mobility of the Negative Ion. By S. Ratner, of the University of Petrograd [Communicated by Sir J. J. Thomson, O.M., F.R.S.]. IT is well known that in general the mobility (k) of a gaseous ion varies inversely as the pressure (p), so that the product kp is constant for a given gas, but at low pressures the mobility of the negative ion becomes abnormally great, and kp increases rapidly with diminution of pressure. The phenomenon of the abnormal mobility of the negative ion has been studied by a large number of experimenters. Langevin [citation redacted], Kovarik [citation redacted], Lattey [citation redacted], and others could observe the abnormal increase of the mobility only in the case when the pressure was reduced below a certain value (75-200 mm.) and therefore assumed the existence of a " critical " pressure at which the abnormality sets in. Frank [citation redacted], and later Haines [citation redacted], have shown that in carefully purified nitrogen, argon, helium, and hydrogen, the mobility of the negative ion is abnormally great, even at atmospheric pressure. Kovarik and Lattey (loc. cit.) drew attention to the fact that at low pressure the mobility of the negative ion varies with the electric force. J. S. Townsend [citation redacted] has shown that the mobility of the negative ion may in general be expressed as a function [footer] 442 [header] Of [formula redacted] where X is the electric force and p is the pressure, and that for large values of [formula redacted] the mobility begins to diminish with increase of the force. Experiments were carried out in order to determine whether the mobility of the positive ion also departs from the inverse pressure law. Todd [citation redacted] has found that at pressures below 1mm. the mobility of the positive ion also becomes abnormally great ; a comparatively small increase of the mobility of the positive ion was observed also by J. S. Townsend [citation redacted], at higher pressures, but these results are contradicted by other experimenters. The study of the mobility of ions leads us to results which may throw considerable light upon the nature of an ion and should be thoroughly carried out in all possible directions. Unfortunately the methods usually employed are but little adapted for this purpose when the mobility becomes abnormally great. Rutherford's, Langevin's, and Zeleny's methods and their modifications serve only in the case when the velocity of the ion does not exceed the order of 10 3 cm/ sec. and therefore can hardly be used for a complete study of the mobility of the negative ion. Chattock's method involves other difficulties, as in this case the mobility has to be investigated under very unfavourable conditions, i. e. when the electric force acting on the ion is very strong and not uniform. The only suitable method for measuring large ionic velocities is that given by J. S. Townsend, depending on the action of a magnetic force on the motion of a stream of ions. In the present paper a new method of measuring ionic mobilities is described and some results are given. Method and Apparatus. The method may be considered as a modification of that given by Chattock [citation redacted], and is based on the production of a wind in an ionized gas when acted on by an electric field. In the case of the discharge from a point to a plate, Chattock deduced the following expression for the mobility (k) of an ion in terms of the discharge current (c), the corresponding wind-pressure (P), and the distance between the plates (d) : [formula redacted] [header] 443 In a previous note the author [citation redacted] has shown that if two parallel plates are immersed in an ionized gas, a small potential difference established between the plates produces a wind-pressure large enough to be measured by a special gauge [I wish to thank Prof. Debierne for suggesting to me the idea of this gauge.] , which will be described later on. When a surface ionization of one sign close to one of the plates is produced, the density of electrification, as well as the strength of the field, is uniform throughout the volume between the plates (if the disturbing effect of the ions is neglected), and in this case Chattock's formula maybe deduced in a very elementary way. Let P be the total surface electrification on the plate, X the strength of the field, w the velocity of the ions, c the total ionization current, P the corresponding wind-pressure, and d the distance between the plates. Then Pw is the quantity of electricity streaming through a plane parallel to the plates in unit time and [formula redacted]. When the ions move with a uniform velocity their drag on the gas is equal to the product of the total electrification between the plates and the strength of the field [This is experimentally proved by the fact, that using the above formula Chattock deduced for the mobility of an ion the same value as given by other methods. ]. Hence [formula redacted]. These two equations give [formula redacted] where [formula redacted] is the mobility of an ion. In the case when the velocity in of an ion is not proportional to the electric force X, [formula redacted] is what is called the abnormal mobility of an ion. The experiments may be carried out at any pressure, the potential difference between the plates varying over a very wide range, viz. from zero up to that required for a spark- discharge between parallel-plate electrodes, and, as will be shown later, by this method ionic velocities may be measured of the order of from 10 to 10 7 cm/sec . The apparatus used is shown in figs. 1, 2, and 3. Two brass plates A and B (fig. 1) are mounted on a wooden base ST. The circular plate A, which consists of two connected metal sheets 8 cm. in diam., is supported by a rod f sliding in a brass collar e and is insulated by an [footer] 444 [header] ebonite column L fixed in an earthed guard-ring R. The plate is cut out in the middle in the way shown in figs. 1 and 3, and in this gap a strip of thin platinum foil mn is stretched in the plane of the plate. One of the extremities of the strip is insulated from the plate by the ebonite plug p, [figure redacted] while the other extremity is brought into contact with the plate and is fixed on a sliding metal support n pulled by a weak spring s. The strip can be heated to any desired temperature by passing a current through it from an insulated battery of accumulators kept at the same potential as the plate A. The plate B, 8 cm. wide and 14 cm. high, has [header] 445 cut out of its centre a circular hole 3 cm. in diam. opposite to the strip of platinum and covered with metal gauze gh. The gauge employed is shown in figs. 1 and 2. A vane abed is cut out in the form shown in fig. 2 from a sheet of aluminium 0*1 mm. thick, and is suspended by a fine brass ribbon z attached to a torsion head X which may slide on a rod v projecting from the plate B. On its lower part the [formula redacted] vane is provided with a mirror r, a small weight g, and an arrangement for damping. The earthed cage k forms an electrostatic protection for the vane. The apparatus is placed on a brass plate pierced by insulated electrodes and is covered with a bell-jar provided with a window of plane glass, so that the image formed by the mirror may be undistorted. Before being admitted into the bell-jar the 446 [header] air passes through calcium chloride, concentrated sulphuric acid, and glass-wool. The plate A may be earthed or charged to a desired potential by means of a battery of small cells, one terminal of which is earthed, while the plate B is connected to a galvanometer provided with a series of shunts, by means of which any current between 10~ 9 and 10~ 5 ampere could be measured. The strip of platinum, coated with a mixture of barium oxide and aluminium phosphate, emits when heated a copious supply of negative or positive ions, according to the direction of the electric field between the plates. The motion of these ions, as shown above, produces between the plates a stream of gas which passes by inertia through the grating gh and a small tube t, and imparts to the vane a deviation measured in the ordinary way by means of the mirror r. When the platinum strip is heated, no electric force being applied between the plates, a convection current due to the heating of the strip is produced in the gas, and the image from the mirror changes its zero position on the scale. For each temperature of the strip a correction of the zero-point of the apparatus is therefore necessary, and precautions must be taken in order to diminish, as far as possible, these displacements of the zero. The suspension of the vane should not be too sensitive, the platinum strip has to be very narrow and placed in a vertical line, as shown in the figure, and the vane should not be placed close to the edge of the tube t, but about 5 millimetres from it. The ionization current, as well as the corresponding wind-pressure, may be varied over a wide range by changing the temperature of the strip, or the electric force between the plates. It is seen from equation (1) that for a given distance d between the plates the mobility of an ion [formula redacted] is proportional to p, the ratio between the ionization current and the corresponding wind-pressure. A great number of preliminary experiments were carried out in order to verify the above equation. As at high pressures the mobility of an ion is known to be constant over a wide range of the strength of the field, the ratio p, according to equation (1), must also be constant under the same conditions. Further, supposing [formula redacted] for negative ions and [formula redacted] for positive, [formula redacted] must be the ratio between the mobilities of the negative and the positive ion. [header] 447 A sample of the results of these experiments is shown in Table I., where p is the pressure of air under the bell-jar, v the potential in volts applied to the plate A, and c and P are given in arbitrary units. [table redacted] It is seen from the Table that ^ is constant within the limits of experimental error and that the value of [formula redacted] is in good agreement with the well-known ratio between the mobilities of the negative and the positive ion in air. Experiments of this kind, together with some other control experiments, which will be described later, prove the efficiency of the method and show that the results are reliable. It is easy to see that the wind-pressure produced by the stream of gas can be measured by the gauge only in arbitrary units and, consequently, the absolute value of the mobility cannot be deduced from equation (1). This, however, is not necessary for the purpose of this work, as in the results given below the abnormal mobility of the negative ion is expressed in terms of its normal mobility — a constant determined to a high degree of accuracy by other methods. Results. 1. The mobility of an ion [formula redacted], as measured in arbitrary units by the ratio v, is constant over a certain range of the 448 [header] applied electric force X, which is in agreement with the well-known experimental facts. But in the case of negative ionization, when X is gradually increased the experiments show that at a certain value of X = X1 depending upon the pressure of the gas a lack of proportionality between c and P suddenly comes in in such a way that [formula redacted] begins to increase. Equation (1) shows that in this case the velocity of an ion increases more rapidly than the electric force, i. e., that its mobility becomes abnormally great. The value of this critical electric force, X 1 , decreases rapidly with diminution of the pressure, so that at pressures below 200 mm. the same effect is reached with comparatively small forces when the ionic velocities are small enough to be measured by Langevin's or Rutherford's method. In the curves I., II., III., IV., V., and VI. (fig. 4), the [figure redacted] mobilities [formula redacted] of the negative ion at pressures of 748, 600, 500, 400, 300, and 200 mm. respectively, are plotted against the electric force X, and the curve [formula redacted], which is a straight line parallel to the axis of X, shows the mobility of the [header] 449 positive ion at 200 mm. pressure. The mobilities are given in arbitrary units, the mobility of the negative ion at atmospheric pressure being equal to 0'72 in these units. The Table II., corresponding to curve I., is given to show the procedure of the experiments. Table II. [table redacted] The curves show that when the critical value of the electric force is reached the mobility of the negative ion begins to increase, at first rapidly. At atmospheric pressure the mobility is doubled when the electric force is increased from 1800 to 2400 volt/cm; at 400 mm. pressure the mobilities at 450 and 2000 volt/cm are as 1 to 5. With larger forces the mobility increases less rapidly, and it may be noticed from the curves that it apparently tends to reach a constant value. The curves show also that at high pressures the velocity of the negative ion cannot be expressed as a function of [formula redacted]. The determination of the critical electric force X 1 at a given pressure involves great difficulties. At high pressures the result of experiments carried out under the same conditions varies from day to day to the extent of 20 per cent. The amount of moisture present in air does not apparently affect the results, although in these experiments the air was never dried with special care. The effect may be due to 450 [header] some other impurities in the air, and this supposition is supported by the fact that, as shown below, even imperceptible traces o£ chloroform vapour change completely the aspect of the mobility curves. At lower pressures, when the critical force X1 is small, another difficulty arises, as the ionization-current and the corresponding wind-pressure decrease with the potential difference between the plates and become, in this case, too small to be measured with accuracy. In the Table III. the values of X1 and [formula redacted] at different pressures are given, X x being the mean approximate value from a great number of experiments. [table redacted] It is seen from the Table that [formula redacted] decreases rapidly with diminution of pressure, so that at high pressures the negative ion may attain comparatively large velocities before it begins to assume an electronic state, and that is the reason why the abnormal mobility could not be observed at high pressures by other methods. It is, however, worth noticing that even at atmospheric pressure the largest velocity attained by the normal negative ion, viz. 3200 cm/sec , is small compared with its velocity of thermal agitation. 2. Interesting mobility curves are obtained with larger values of — . It has been already noticed from the curves p J (fig. 4) that at large electric forces the mobility tends to remain constant. When the force is still further increased [header] 451 the mobility of the negative ion attains a maximum value and begins to diminish. This phenomenon could be predicted from Langevin's formula for the mobility of a charged particle whose mass is small compared with that of a molecule : [formula redacted], where e and m are the charge and the mass of the particle, l its free path, and u its velocity of agitation. The mass associated with the negative ion gradually diminishes with increase of electric force, and at this stage of evolution of an ion its mobility increases with the force. When, finally, the electronic state is reached by the ion, its mobility follows the above equation, and since the velocity of agitation of an electron increases with the electric force [citation redacted], its mobility begins to decrease [citation redacted]. [figure redacted] This decrease in the mobility of the negative ion is shown in curves I., II., III., and IV. (fig. 5), for pressures of 200, 100, 50, and 10 mm. respectively. The curves III.(+) and IV.( + ) show the mobility of the positive ions at 50 and 10 mm. 452 [header] pressure. In this figure the mobilities at different pressures are given in different units in order to avoid using too small a scale. In the curves I. and II. the mobility at any point may be given in terms of the normal mobility represented as straight lines in the beginning of the curves, while in the curves III. and IV. it may be calculated from the mobility of the positive ion at the same pressures. The curves show that the velocity of the ions always increases with increasing forces, even in the case when their mobility diminishes. By their maximum points the curves are divided into two parts, the first part corresponding to the gradual dissociation of the ion and the second to the pure electronic state of an ion. The value of [formula redacted] corresponding to the maximum mobility is constant at pressures up to 200 mm. and is equal to 6' 2, which is very large compared with 0*2 — the value given by J. S. Townsend in case of carefully dried air. The rate of decrease of the mobility, as seen from the curves, diminishes with the increase of the force, and at sufficiently large values of [formula redacted] the mobility tends to reach a minimum value. At low pressures this bend in the curve is well marked, as shown by curve IV. Experiments with still larger values of [formula redacted] are made impossible by the luminous discharge which takes place under these conditions between the heated strip and the grating. It is possible that the luminous discharge is preceded by a feeble ionization by collision, which might be responsible for this bend of the curves. Further experiments in this direction are now in progress. With regard to positive ions, the experiments show that at pressures down to 5 mm. and with very large forces the mobility remains constant. 3. Experiments in hydrogen. — The same experiments were carried out in hydrogen — a gas in which the negative ion is known to exist in the electronic state under ordinary conditions. The hydrogen was generated in a Kipp apparatus from hydrochloric acid and zinc and purified by passing slowly through NaOH, KMn0 4 , and H 2 S0 4 . It was always contaminated with small quantities of impurities, since the brass-plate and the bell-jar enclosing the apparatus contained a great number of different joints which allowed a constant small leakage of air. The mobility curves in hydrogen show the same characteristic features as in the case of air. The [header] 453 critical value X1 of the electric force, at which the mobility of the negative ion becomes abnormal, is very small in hydrogen and increases with the amount of impurities to such an extent that it may serve as a good criterion for testing the purity of hydrogen. The curves I., II., and, III. (fig. 6) represent the mobility [figure redacted] of the negative ion at pressures 750, 375, and 100 mm. respectively, in the purest hydrogen obtained for these experiments. The critical value X 1 of the electric force is about 40 at atmospheric pressure, and at lower pressures it becomes too small to be measured by this method. The value of [formula redacted] corresponding to the maximum point of the mobility curves is very large, varying from 1*8 at atmospheric pressure to 1*1 at 100 mm., and with further increase of the electric force the mobility tends, as in air, to attain a minimum value. These high values of [formula redacted] are probably due to the impurities contained in the hydrogen. Experiments 454 [header] in purer hydrogen and nitrogen, as well as in carefully dried air, are now in progress. 4. Experiments in heavy gases. — Experiments were also made in order to test the mobility of the negative ion in heavy gases, such as the vapours of chloroForm, carbon tetrachloride, and methyl iodide [In these experiments the silvered surface of the mirror was attacked by the halogens. I wish to thank Mr. E. Everett for suggesting to me the idea of platinizing the mirror and for setting up the cathode-ray apparatus necessary for this operation.]. The mobility in these gases was found to be constant for the negative as well as for the positive ion, the mobility curves being straight lines parallel to the X axis even at low pressures (down to 5 mm.) and with large electric forces. A striking effect illustrating the influence of impurities on the mobility of the negative ion was observed during these experiments. Even after the chloroform was removed and the apparatus, as well as the bell-jar and the pump, was thoroughly cleaned and freed from traces of this gas, no further experiments in other gases were possible for many days. It was found that chloroform vapour present in air to an amount as small as 1 part in 1,000,000 changes completely the aspect of the mobility curves. Control Experiments and Sources of Error. 1. It is important for the accuracy of the experiments, that the total wind-pressure produced by the ions should be given by the gauge, and this is secured if all the ions reach the plate B within the area of the grating. In order to ascertain this the grating was insulated from the plate by a narrow air-gap, and the galvanometer connected by means of a suitable key, either with the plate or with the grating. It was found that even at the largest distance between the plates, and under conditions in which the lateral diffusion of ions is abnormally great, the total current is received by the grating, the charge picked up by the plate being imperceptible. 2. The temperature of the gas close to the heated strip must be very high, and therefore the mobility of the ion at the moment when it leaves the strip must be large. In order to ascertain the magnitude of this source of error, curves representing the mobility of an ion as a function of the electric force were drawn for distances between the plates varying from 5 to 45 mm. No difference between them could be observed, which shows that the path traversed by [header] 455 the ion with a larger velocity is small compared with the distance between the plates and may be neglected. 3. It was supposed at first that the electric field between the plates was uniform. This, however, is not the case, the electrification between the plates being large enough in these experiments to disturb considerably the uniformity of the field. Unfortunately, this source of error cannot be eliminated or even diminished, since the ionization-current must be considerable in order to produce a perceptible wind-effect. It is useless to increase the sensitiveness of the gauge beyond certain limits, as the electric wind produced must be large compared with the convection current due to the heated strip. There is, however, sufficient evidence to show that the results are but little affected by this source of error. In almost all the experiments described the ionic velocities were very large, so that the density of electrification between the plates was considerably reduced. The equation [formula redacted], as shown above, was found to be true over a wide range of X, c, and P. The mobility of the negative ion was shown to be a function of the electric force only, and not to change with the distance between the plates, which would be the case if this source of error was great. 4. The platinum strip freshly coated with salts gives off, when heated for the first time, a considerable amount of smoke (consisting probably of charged particles), which make the results inconsistent. In this case the strip has to be strongly heated at reduced pressure and in a strong electric field before measurements are taken. Electric Wind in case of Ionization of both signs. It seemed to be of interest to study the pressure of the electric wind in the case of ionization of both signs, when the wind is produced by positive and negative ions moving in opposite directions. For this purpose a small cell containing 25 mgr. of radium bromide and provided with a thin mica window was placed in the centre of the plate A in place of the platinum strip, and the gas between the plates strongly ionized by the [alpha] rays. The plate A could be positively or negatively charged, the plate B being earthed, and the wind-pressure measured in the usual way by the gauge. In this case the pressure of the wind at any point between the plates is a resultant of two opposite forces produced by the motion of positive and negative ions. 456 [header] In the curves ( + ) and (— ) (fig. 7) the wind-pressure is plotted against the electric force at 450 mm. pressure, in the case when the ions moving towards the grating are positive and negative respective!}'. These curves are in full [figure redacted] agreement with the results relating to the mobility of the negative ion given above. For small forces the wind-pressure is small, as the ions in this case recombine on their way and traverse only a small part of the distance between the plates. The wind-pressure increases with the force until the saturation current is reached, and with further increase of the force a striking difference between the two curves is observed. Under these conditions the mobility of the negative ions, as measured by the ratio [formula redacted], begins rapidly to increase, and since c is constant in these experiments, the wind-pressure P produced by them must decrease at the same rate ; while the mobility of the positive ions, and consequently the wind-pressure due to them, is not affected by the Increase of the force. When the plate A is negatively Charged and the negative ions are moving towards the grating, the wind-pressure produced by them diminishes [header] 457 with the strength of the field, and when the opposite force due to the positive ions begins to prevail, the pressure is reversed and assumes a negative value. In the case when the positive ions are moving toward the grating the wind-pressure produced by them remains constant, but since the opposite force due to the negative ions gradually diminishes, the resultant pressure measured by the gauge increases with the electric force. At atmospheric pressure the negative direction of the electric wind can be produced only with large electric forces exceeding 3000 volt/cm at 50 mm. pressure a force of 80 volt/cm is required for the same effect. It is of interest to note that with small electric forces applied within the plates the wind-pressure increases with diminution of pressure of the gas down to a certain limit. This is due to the fact that the ionization-current, which is in this case far from saturation at atmospheric pressure, gradually approaches it when the pressure is reduced. It would not be out of place to consider these results in connexion with the observations made by Joly [citation redacted] on the motion of radium in an electric field. Almost all the peculiar properties of radium described in that paper are easily explained, in the light of these results, by the electric wind produced by radium. On the Nature of the Negative Carrier. In the discussion of the results given above the existence was assumed of a transition stage between a negative ion and an electron, at which the average mass associated with the ion gradually diminishes with increasing electric forces. Whilst these experiments were in progress a new paper on the mobility of the negative ion appeared [citation redacted], in which a radically different point of view as to the nature of the ion is put forward. According to E. M. Wellisch there are two distinctly different kinds of negative carriers in a gas : normal ions and free electrons, the proportion of the latter increasing with diminution of pressure ; the conception of an intermediate stage between an ion and an electron is erroneous and is due merely to the attempt at averaging the different properties of these two kinds of ions. [footer] 458 [header] P X Equation (1) may be written in the form [formula redacted] , which shows that the wind-effect produced by the ions is inversely proportional to their mobility. According to Wellisch's theory the wind-pressure measured by the gauge in these experiments is mostly due to the slow normal ions, the effect produced by the free electrons being very small. The ratio [formula redacted] , as shown above, diminishes with increasing electric forces, and reaches in some experiments 5 per cent, of its initial value. It is necessary therefore to suppose that the proportion of free electrons increases with the electric force, so that, finally, the amount of normal ions becomes vanishingly small. Wellisch, however, states that the proportion of free electrons is independent of the strength of the field, and is always less than 50 per cent, of the total number of the negative carriers. In order to verify the results of Wellisch, experiments were carried out which were based on the principle of separating, by means of an alternating electric field, the normal ions from the free electrons, and measuring the wind-effect due to one of these two kinds of ions separately. It appeared necessary for this purpose to study at first the wind-effect produced by the negative ions in an alternating field under conditions in which the mobility is normal, i. e. when, according to Wellisch, all the ions are normal ions. The alternating field between the plates A and B was established by means of a high-speed commutator which could produce up to 2000 alternations per second [The commutator was that used by G. Todd, Phil. Mag. xxii. p. 791 (1911).], the potential difference between the plates being kept constant. The experiments show that when the number of alternations is gradually increased, the wind-pressure at first slowly diminishes and then rapidly falls down to a very small value. If the wind-pressure is plotted against the number of alternations per unit time, the curves obtained are all of the type of curve ( + ) (fig. 8), and simple calculations show that the sudden drop in the wind-pressure corresponds to the number of alternations at which the ions emitted by the heated strip begin not to reach the grating. Now, if we repeat the same experiments under the conditions, in which the mobility of the negative ion is abnormal, i. e. when, according to Wellisch, the wind-pressure is mostly produced by the normal ions and only to a small extent by the free electrons, the following observations are to be [header] 459 expected : The wind-pressure should at first slowly diminish with increasing rate of alternations, and when the number of alternations per second is large enough to prevent the normal ions from reaching the grating, a sudden drop in the wind-pressure should be observed. [formula redacted] The experimental results are, however, altogether different from what might be expected if Wellisch's theory were true. No drop can be observed in the wind-pressure when the frequency of alternations reaches the critical value, which decisively proves that the wind is produced only by one kind is of ions whose mobility ( as measured by the ratio [formula redacted] ) is intermediate between that of a normal ion and of a free electron. The curve (— ) (fig. 8) was drawn at 100 mm. pressure, the distance between the plates being 3 cm. and the p.d. between them 200 v. ; 900 alternations per second is enough under these conditions to separate the normal ions from the electrons ; the curve, however, shows that the wind-pressure continues to diminish slowly with further increase of the rate of alternation. The curve ( + ) shows the wind-pressure under the same conditions, but in the case when it is produced by positive ions. [footer] 460 [header] The experiments of Wellisch were carried out on the supposition that the mobility of the negative ion is independent of the electric force applied ; this, however, is shown to be true only for comparatively small electric forces. The study of Wellisch's paper leaves but little doubt that the sudden increase in the mobility of the negative ion with increasing forces is responsible for the characteristic bend in the curves, which have led him to the erroneous conclusion of the existence of two different kinds of negative ions. Thus, for a satisfactory explanation of the results given above, we must assume the gradual diminution of the average mass of the negative ion with increasing electric forces. The conception of "the average mass associated with an electron - ''' is, however, complicated, owing to the uncertainty as to the nature of an ion, and may be interpreted in a way different from that adopted in this paper. According to the interpretation recently given by Sir J. J. Thomson [citation redacted] the electron may easily escape from the system of molecules constituting the ion and travel a certain distance in a free state, until it is once more attached to a molecule, and so on, so that the negative carrier makes its way through the gas, partly as a free electron and partly as a normal ion. From this point of view there is no " transition stage " between an ion and an electron, and the increase in the mobility of the negative ion, as well as the apparent diminution of the mass associated with it, show only that under certain conditions the proportion of time during which the ion is moving as a free electron increases with the electric force. Summary. 1. A new method of measuring ionic mobilities in eases is described. 2. At a given pressure the mobility of the negative ion is shown to be constant only over a certain range of electric forces applied. With increasing electric Forces X a certain value of X = X 1 is reached when the mobility of the ion begins to increase rapidly. The value of the critical force X1 increases rapidly with the pressure, being equal to about 1800 volt/cm. at atmospheric pressure. 3. The mobility of free electrons was measured and shown to diminish with increase of the electric force. 4. The mobility of the negative ion in hydrogen and in some heavy gases was measured. [header] 461 5. No abnormality in the mobility of the positive ion could be observed at pressures down to 5 mm. and with very large electric forces. 6. The pressure of the electric wind in case of ionization of both signs was studied. 7. The recent theory of E. M. Wellisch with regard to the nature of the negative ion was investigated and shown to be erroneous. These experiments were to a great extent carried out in the Physical Laboratory of the Polytechnic Institute in Petrograd. I am indebted to Prof. A. F. Joffe for allowing me to work in this laboratory and to Mr. J. S. Shcheglaeff for putting the necessary apparatus at my disposal. In its final stages the work was completed at the Cavendish Laboratory, Cambridge, and I wish to take this opportunity of thanking Prof. Sir J. J. Thomson for permission to use the laboratory and for his interest of the progress of this work. Cavendish Laboratory, August 1916.