XCYI. The Doppler Effect in Positive Rays in Hydrogen. By T. Royds, M.Sc., 1851 Exhibition Science Scholar [Communicated by the Author.]. Little is known of the cause of luminescence of the canal rays and of the origin of the discrepancy between their actual velocity and that calculated from the potential drop through which they have fallen. Some explanation of these would probably be obtained if the velocity of the positive rays at different parts of the discharge were compared with the distribution of potential, of electric force, and of the rate of change of the latter. Unfortunately there is no satisfactory method of isolating the different portions between the cathode and the negative glow for the observation of the Doppler effect. At one point of the discharge, however, a comparison is possible, namely, at the point where the cathode glow ([citation redacted]) commences. Since the canal rays mostly start from the negative glow, we can assume that the commencement of the cathode glow corresponds to the minimum Doppler effect when the cathode is viewed from the anode side. That the cathode glow near its commencement shows the Doppler effect was proved by placing the discharge-tube inclined to the axis of the collimator until light from only the first portions of the cathode glow fell on the slit, the portions nearer the cathode being screened off by the cylinder carrying the cathode. The 896 [header] length of the dark space was 5*0 cm., the distance between the negative glow and the cathode glow 2'2 cm., and light from the first centimetre of the cathode glow was focussed on the slit. After 20 hours' exposure with a cathode fall of 2800 volts, a marked Doppler effect was seen corresponding to the positive rays approaching the cathode. When the gas (hydrogen) is pure and the pressure not too low the cathode glow commences at a rather sharply defined surface which can be fixed upon with reasonable accuracy. The potential fall from the negative glow to this surface has been compared with the minimum Doppler effect, i. e., the distance between the stationary line and the nearest edge of the shifted line, the cathode (without holes) being viewed from the anode side. The discharge-tubes had a diameter of from 4*5 to 4*8 cm. The potential in the discharge was measured by means of sondes of thin platinum wire inserted at right angles to the axis of the tube. A length of 2 mm. was exposed near the axis and the rest insulated by means of fine glass tubing. [figure redacted] The sondes were fixed in the tube, and the aluminium cathode A, fitting closely in the tube, could be made to approach the sondes by means of a magnet outside the tube attracting the iron piece B attached to the cathode. The sonde C gave the potential in the dark space, the sonde D remaining in the negative glow throughout which the potential is practically constant, This second sonde was introduced in order to show that the cathode fall was not altered by bringing up the cathode. The anode was placed in a side tube E in order to allow the cathode to be seen through the end F of the tube. The current was taken from a battery of accumulators with a maximum potential of 3000 volts and the potential measured by means of a calibrated Braun [header] 897 electrometer. A telephone was inserted in the circuit and measurements only made when it was absolutely silent. Measurements were only possible before disintegration of the cathode set in, for as soon as a metallic film forms on the walls of the tube the discharge is altered by bringing up the cathode. The values of the electric force and its rate of change along the discharge were obtained from the potential changes when the cathode was displaced a few millimetres on each side of the point where the value was required. It is found that the commencement of the luminescence of the cathode glow seems to depend on the potential fall from the negative glow rather than on the values of the electric force and its rate of change at this point. The value of the potential fall to the commencement of the cathode glow is independent of the pressure but rises slowing according to a straight line law with the cathode fall. The electric force and its rate of change vary between wider limits. Some of the values obtained are given in the accompanying table. Table I. Values at the commencement of the cathode glow Potential fall Electric force Rate of change of electric force [table redacted] The Doppler effect was observed from the direction F (fig. 1) by focussing the cathode glow on the slit of a grating spectrograph. The grating was mounted in the manner described by Paschen [citation redacted]. The camera included from H[alpha] to H^j the dispersion varying from 9-90 to 9*15 A.U. per mm. The density of the photographs was not very different for H[alpha], H[beta], and H[gamma]. H[beta] was the strongest, H[alpha] being sometimes weaker sometimes stronger with different batches of plates than H[gamma]. The Doppler effect when the cathode without holes is viewed from the anode side has already been observed by Paschen [citation redacted], 898 [header] and for the cathode with holes by Stark [citation redacted] and Trowbridge [citation redacted]. In the latter cases the Doppler effect is chiefly due to the canal rays behind the cathode which are seen through the holes. It is of interest to note that the general appearance of the Doppler effect in front of the cathode is similar to that observed in the canal rays behind the cathode in that with values of the cathode fall between certain limits it consists of two strips. The extreme edge of the shifted line is not sharp as might have been expected. As noted by Paschen, the shifted line lies nearer the stationary one than in the case behind the cathode. It was observed by Paschen and by Trowbridge that there is, besides the Doppler effect of rays approaching the cathode, also one on the opposite side of the stationary line denoting a motion towards the anode. Paschen found that the latter effect was more intense than the former. These rays probably consist of the particles reflected on impact at the cathode surface, for when a holed cathode is focussed on the slit the Doppler effect of rays approaching the negative glow is weaker at the holes. The reflexion is probably diffuse, for the Doppler effect is smaller and the space between the stationary and shifted lines is not so dark ; this effect may also be due to light reflected diffusely at the mattened surface of the cathode. On account of the feebleness of the light in front of the cathode, experiments were only made with high potentials and as strong a current as consistent with the safety of the discharge-tube. With a cathode fall of 2650 volts an exposure of 10 hours was found necessary to avoid underexposure^ which would cause the measurement of tlie minimum Doppler effect to be too large. The photographs were measured by means of a Zeiss comparator. The cross-wires were set upon the edge of the stationary and shifted lines by eye. The densest portion was estimated photometrically with the arrangement described by Paschen [citation redacted]. Although it does not follow that the homogeneity of the light in the canal rays is the same as that in the stationary line, the minimum Doppler effect was measured from the farther edge of the stationary line, since the spectral width of the latter was small compared with the width of tie slit-image. Even when the minimum was measured from the centre of the stationary line the relative differences in [header] 899 the values for different spectral lines were made but a little smaller. The observed values of the velocities are given in the following table. Table II. Cathode Falls 2650 volts. Length of dark space 3*0 cm. Velocity of rays approaching the cathode, Velocity of reflected rays in cm./sec. [table redacted] It is seen that the minimum velocity is not constant for different wavelengths, the differences being rather larger than the errors of measurement. Since H[alpha] generally appears on the photographic plates weaker than H[beta] and stronger than H[gamma], the result cannot be explained as a photographic effect. The minimum velocity is approximately inversely proportional to the square root of the wavelength, the product of these two quantities having the following values. rays approaching the cathode. Reflected rays H[alpha] 97-1 85-0 H[beta] 99-7 93-4 H[gamma] 98-1 Stark and Steubing [citation redacted] have found this law to hold for the minimum velocity behind the cathode. In front of the cathode, although the light is fainter, there is the advantage that the edge of the shifted line is sharper. The variations in the measurements of the velocity corresponding to the densest portion of the shifted line do not lie outside the limits of the errors of measurement. 900 [header] In the following table are given the ratios between the observed and calculated velocities assuming [citation redacted]. In column 1 the minimum velocity in front of the cathode is compared with the potential fall from the negative glow to the commencement of the cathode glow (400 volts for cathode fall of 2650 volts). Columns 2 and 3 refer to the canal rays behind the cathode with the same cathode fall of potential, these being less complicated than in front of the cathode since all rays have passed through the cathode fall. Table III. Ratio observed velocity calculated velocity. Minimum in front of cathode. Densest portion behind the cathode. Maximum behind the cathode. [table redacted] It is seen that the ratio for the minimum in front of the cathode is about the same as for the maximum behind the cathode. Since the minimum velocity in front of the cathode was found to be not constant for different wavelengths, measurements were also made for the canal rays behind the cathode. It was observed in agreement with Stark and Steubing that the minimum velocity was smaller for greater wavelengths. For the point of maximum density, however, the variations were within the limits of error, even for the outer strip observed with a cathode fall of 550 volts. I wish to express my thanks to Prof. Dr. Paschen for his interest in these experiments and the numerous suggestions he has made. Physikalisches Institut, Tübingen, Aug. 4th, 1909.