measuring the rate of leak between two metal electrodes, or sets of electrodes, when the gas between them has been ionized. For the purposes of the present experiments this had to be done at various temperatures. The gas to be experimented upon had therefore to be inclosed in a vessel which could be heated to a fairly high temperature. When investigating air the experiments may be made in either of two ways. The first way is to inclose the air in a vessel which is not air-tight, and which will allow the air to expand freely when heated, so that the measurements may be taken in air at a constant pressure but in which the density changes as the temperature is varied. The second way is to inclose the air in an air-tight vessel so that the volume and density of the air experimented upon are kept constant at the various temperatures. In the present investigation both of these methods have been employed. A description of each, with the results obtained, will be given in turn. Experiments on Air at Constant Pressure. As already mentioned these experiments arose out of the investigation on the rate of recombination of ions, and the apparatus used was originally designed for the experiments on the recombination. As it was suitable for the present experiments it was therefore employed, although it was a little more elaborate than would have been really necessary for the present instance. In fact, in the course of investigating the rate of recombination one of the quantities measured was the amount of ionization, and therefore some of the results given in this paper were obtained concurrently with those on recombination. A full description of this apparatus has already been given in the paper on recombination of ions, and diagrams showing its general arrangement and construction also appeared in that paper. It will therefore not be necessary to give a full description of the apparatus here, but a less detailed diagram may be given for reference to show the general arrangement of the apparatus. This is shown in fig. 1. The Röntgen-ray bulb and induction-coil were as usual inclosed in a lead-covered box, and the rays emerged through a well-defined circular orifice, A, in the lead, and passed into the brass cylinder BC, where they ionized the air. The bulb used was one with an automatic vacuum regulator attached. The brass cylinder BC was surrounded by a sheet-iron cylinder, as shown in the diagram, so that there was a uniform air-space of about 10 cms. between the two cylinders. The inclosed air was heated by means of a long Bunsen burner placed underneath the outside cylinder, and which ran almost the whole length of the iron cylinder. The air in the brass cylinder was thus surrounded by a jacket of heated air, and by regulating the supply of gas to the burner the temperature of the air in the cylinder BC could be kept fairly constant for a considerable time. The temperature of the air was measured by the two mercury thermometers T and T' shown in the diagram, and the mean of the temperatures indicated by these thermometers was taken as the average temperature of the air under investigation. The electrode E' was connected in parallel with a condenser to one pair of quadrants of an electrometer, while the other pair was to earth. The electrometer used throughout the experiments was one of the Dolezalek type which gave about 2000 scale-divisions for a difference of potential of one volt between the quadrants when the needle was charged to 120 volts. The other electrode E was connected through a large liquid resistance R to one pole of a battery of accumulators, while the other pole was to earth. Now if the gas between two insulated electrodes be ionized by a constant source of ionization, and if to one of these electrodes a steady voltage be applied sufficiently large to extract all the ions from the gas before they have time to recombine, the other electrode will charge up at a rate proportional to the number of ions produced in the gas per second. The rate at which this electrode charges up will therefore be a measure of the amount of ionization produced in the given volume of gas per second. The deflexion of the electrometer-needle per second, which is proportional to the rate at which E' charges up, will be proportional to the amount of ionization produced in the gas, and will therefore be a measure of the amount of ionization. To measure the ionization at different temperatures of the gas the following method was adopted. The rays were started and allowed to pass into the cylinder for an interval of five or ten seconds, so that the ionization might reach a steady state. During this interval the electrode E' and the quadrants of the electrometer were connected to earth. At the end of this time the quadrants connected to E' were insulated, by a key worked at a distance by means of a cord, and were allowed to charge up for a given number of seconds, and at the end of the given time the rays were stopped and the deflexion of the electrometer-needle observed. Several readings like this were observed at the ordinary temperature of the room and the mean of these readings taken. cylinder and inclosed air were then heated up to a given temperature, and when the temperature became steady the deflexions were observed as before. This having been done the air was once again heated to a still higher temperature and the deflexions again observed. This was done for several temperatures up to the highest one investigated. The amount of ionization at the different temperatures could thus be compared by comparing the deflexions obtained at these temperatures. The Instead of starting the series of readings at the lower temperature and gradually heating the air up to the higher temperatures, the order of procedure was in some cases reversed, and the air was heated up to the highest temperature to start with, and the deflexion corresponding to the ionization observed. The gas was then gradually cooled down from point to point, observations being taken at each temperature. Similar results were obtained in both cases. In making this comparison it was very essential that the source of ionization should remain constant in intensity throughout the series of observations, otherwise no comparison could be made unless the amount of variation in the intensity was known. Even when using an automatic regulating bulb it is almost impossible to obtain perfect regularity in the intensity of the rays. When taking the observations the bulb was run as far as possible at regular intervals so as to keep it steady, and also several readings were taken in each case and the mean taken. However, even with these precautions, one cannot be certain as to whether the rays remain constant in intensity throughout the experiment unless some independent check is employed in order to test their constancy. For the purpose therefore of testing the constancy of the rays a small standard apparatus was introduced between the source of the rays and the large cylinder as shown in fig. 1. It is shown in detail in fig. 2. It consisted simply of a rectangular cylinder Fig. 2. made of sheet-lead about 12.5 cms. square and 10.4 cms. in length, and the ends were covered with paper. It contained two parallel zinc plates, about 9 cms. square, acting as electrodes. One of these plates was connected to the same pole of the battery as E. From the central part of the other plate was cut a circular disk 35 cms. in diameter, and this was insulated from the other part of the plate. The outer part of the plate acted as a guard-ring and was connected to earth, while the central disk could be connected to the electrometer by the key K when desired. The rays passed between these plates on the way to the brass cylinder, and ionized the air between them. If the intensity of the rays remained constant the saturation-current between these plates should be constant, and any variation in the intensity of the rays should be shown by a corresponding variation in the current. After the readings had been taken on the air in the cylinder BC, the connexion at K was transferred to the electrode F, and the rate of leak in the standard apparatus measured. By this means it could be determined whether any variation, which might occur in the rate of leak between the electrodes in the cylinder BC, was due to a variation in the intensity of the rays or to some other cause. This therefore served as a test of the rays. In observing the amount of ionization, as indicated by the deflexion of the electrometer-needle, it was found that as the temperature of the air increased the deflexions decreased, and that they varied in the inverse ratio to the absolute temperature. Now we must take into account here the fact that as the temperature of the air is increased its density decreases in the inverse ratio, since the gas is quite free to expand into the outside air. It was shown by Perrin* that the amount of ionization produced in a gas is proportional to the pressure of the gas, and this result was later confirmed by Rutherford and McClungt. Therefore, in the present instance, when the density of the gas decreases, there would be a corresponding decrease in the amount of ionization produced, due entirely to the change in density. To determine the effect on the ionization due to a change of temperature alone, a correction must be made for the change of density of the gas. In making these experiments then it was found that the decrease in the amount of ionization which took place as the temperature rose was just the amount of decrease which would occur on account of the decrease in the density of the gas, and when the necessary correction was made for the change of density no alteration in the amount of ionization was produced by the variation of the temperature itself. In other words, if the density of the gas were kept constant the amount of ionization produced by rays of given intensity would be independent of the temperature of the gas. A series of readings is given in Table I. as a sample of the results which were obtained. *Annales de Chimie et de Physique, xi. p. 496 (1897). |