Ap X Ap In addition, the values of are also given. AQ ' p' p A Q Each table refers to a fresh mixture of the gases. The tables are illustrated by the curve (fig. 4), and the points lying near it or on it. 50x5 1 Ap PAQ' and the ordinates the The numbers attached to the points indicate the table from which these values are taken. These results show that 1 Ap PAQ We conclude that, for the range of X and p indicated by Before proceeding, it is necessary to find an expression for the number of molecules of water, the formation of which is associated with the passage to the electrodes of each pair of ions. Let N be the number of molecules in a c.c. of gas at the fall of pressure per dp 760 mm. pressure. Let Xp with p. Ap varying AQ' Then the number of molecules of hydrogen and oxygen lost per coulomb is, since the capacity of the apparatus was 570 c.c., Nx 570 3N Let e be the charge in electrostatic units on an ion. Then from experiments on the diffusion of ions (J. S. Townsend, Phil. Trans. 1899), Ne 1.2 x 1010. Now one coulomb is equal to 3x 109 electrostatic units. Therefore when one coulomb passes round the circuit, the number of negative ions which reach the anode (or of positive ions which reach the cathode) is Therefore, when the pressure is p mm. it follows that the N arrival of negative ions at the anode involves the dis4 appearance of 3N A molecules of hydrogen and oxygen. That is, the passage of each negative ion to the anode is accompanied by the formation of On referring to the Table VII., by way of illustration, it will be found that- =2.2 when p=1.8, this being one of ΔΡ the lowest pressures recorded in the tables. Therefore, as each pair of ions reached the electrodes at this pressure, 44 molecules of water were formed. Next, on taking the highest recorded pressure 6.5 (Table V.), it will be found that 88 molecules of water were formed as each pair of ions reached the electrodes. And on all occasions recorded above, more molecules of water were formed than pairs of ions. Now let W, denote the number of molecules of water formed for the passage of each pair of ions to the electrodes. where ƒ) is the ordinate corresponding to the abscissa f of the curve given above. (2) Χ p If the force were uniform between the plates, X might be taken as the force, since the plates were 97 cm. apart. And then equation (2) would strongly support the view that the formation of water is attributable to the collisions with the neutral molecules of the mixed gas of the ions as they move to the electrodes. The simplest hypothesis would be that the ions by collision separate the neutral molecules of oxygen into uncharged atoms, and that the latter then combine with the neutral molecules of hydrogen. But one is prevented from basing these views on equation (2), by the large and uncertain departure from uniform intensity which the electric force makes near the cathode. Moreover, the range of variation of the potential X was not large enough to prove the validity of equation (2) except for a very restricted variation of X (see the tables above). But further experiments will shortly be made with the view of deciding this and other points. During these experiments the phenomena do not appear to be complicated by the effects of the heat set free by the chemical union of hydrogen and oxygen, although this seems to play such an important part at pressures of a higher order. The elimination of these effects at low pressure may be explained by the hypothesis that the energy set free by the formation of a molecule of water is radiated as well as conducted through the gas. For the intensity of such radiation would be proportional to the inverse square of the distance from the molecule of water, and the radiation would have no effect on the neighbouring molecules of oxygen and hydrogen unless they were within a certain distance of the molecule of water which was formed. When the pressure of the gases is reduced below some critical value, the probability of the reactions taking place would be extremely small. I wish to express my thanks to Professor Townsend who suggested this investigation. I am greatly indebted to him for his suggestions and criticisms. I XXII. EDITORIAL NOTE. HAVE received from Lord Blythswood a letter of date Jan. 23, with a specimen of cambric rendered thoroughly brittle or rotten by exposure for about three days to radium bromide. He had put a little circle of cambric in place of the circular sheet of mica which is commonly used to cover the cavity containing radium bromide in the little receptacle in which it is usually sold. The cambric is quite broken away, leaving an irregularly shaped hole of about 3 mm. greatest diameter in the place which was directly exposed to the radium. This is certainly a very interesting` and, I believe, important discovery. Lord Blythswood found the same result in several other trials with exposures of two or three days. Largs, Jan. 26, 1904. XXIII. Notices respecting New Books. KELVIN. A Handbook for the Electrical Laboratory and Testing Room. Vol. II. By J. A. FLEMING, M.A., D.Sc., F.R.S. London: "The Electrician" Printing and Publishing Company, Limited. 1903. Pp. viii+622. NOTHING could better illustrate Dr. Fleming's tremendous activity as a writer of text-books than the appearance of this second volume of his 'Handbook' within something like a year from the date of publication of Vol. I. The compilation of this second volume must have involved a very large amount of labour, and the very full references to original sources show that the author has spared no pains to make his treatise of as encyclopædic a nature as possible. The book is divided into five chapters. In Chapter I. we have an account of the measurement of electric quantity and energy. We are heartily glad to see Dr. Fleming's strong condemnation of ballistic galvanometers of the needle type. In connexion with the damping correction of a ballistic galvanometer, attention may be drawn to the extremely clumsy form, sanctioned, no doubt, by long usage, in which this correction is applied. We are sorry to see the time-honoured treatment reproduced by Dr. Fleming. The student is told, by teachers and text-books alike, to find the "logarithmic decrement" of the galvanometer, and then to correct the "throws" by multiplying them by 1+ Now this 2' ** λ 2 λ λ mysterious 1+ is simply an approximation to e2, being, in fact, R |