enthusiastic author, whose scientific acumen, wide knowledge, and judicious treatment of the varied subjects of his compilation are certainly further proved by this volume. As yet, he informs us in the Preface, the enterprise has not been profitable; but he hopes that it will be successful by the co-operation of more and more Subscribers; and indeed we sincerely hope that this handy and valuable Annual will be worthily supported. XXI. Intelligence and Miscellaneous Articles. ON THE NATURE OF THE REFLEXION OF ELECTRICAL WAVES AT THE END OF A CONDUCTING WIRE. BY MM. KR. BIRKELAND AND E. SARASIN. ́ IN Na communication of April 17, 1893, one of us attempted, starting from the theory of the motion of electromagnetic energy in space, to make a hypothesis on what takes place in the vicinity of the end of a metal wire along which electrical waves are passing. We have examined the question together, exploring the electrical field about the end of the wire, with small resonators of 10 to 25 centim. in diameter, and although the point of view explained in the note in question has not been confirmed in all points by our results, they form none the less an interesting experimental contribution to the remarkable theories of Prof. Poynting. The electrical waves were furnished by a small plate-exciter, the sparks of which, about 3 millim., passed in oil. Opposite one of these primary disks was a similar one from which proceeded a copper tube 1 centim. in diameter and 9 metres in length. This tube was supported on thin wooden rods 1.5 metre high. The resonators were circular and fixed vertically with the spark uppermost; they had a double motion-they could be rotated about a vertical axis passing through the spark, and they could also be moved parallel to the conducting tube. We arranged so that even in darkness we could mark the distance from the centre of our resonator to the end of the conducting wire, measured parallel to the wire, and also the angle which the plane of this formed with the wire. The observations were made by means of a telescope mounted horizontally at a distance of a metre from the circle, which is necessary owing to the considerable disturbance produced by the body of the observer on the rapid oscillations. Our researches were directed to two points principally: we first determined the four first nodes in seven different distances from the wire; the plane of the circle being always normal to this. These nodes are each determined by at least ten measurements; the numerical results thus obtained for the circle of 10 centim. are given in the subjoined table. The different distances of the resonator from the conducting wire are counted from the axis of this latter to the nearest point of the circle. The numbers in each column give the distance of each node to the normal at the end of the wire. In fig. 1 we have represented by small dots the position of the centre of the resonator of 10 centim. in the 27 nodes given numerically above. The small circles on the same figure give the positions of the nodes due to a circle 25 centim. in diameter in four different distances from the conducting tube. The impression which directly results from the distribution of all these nodes is that, if the first shock arrives at the resonator almost parallel to the conducting wire, the second must arrive there by a direct radiation proceeding from the neighbourhood of the end of the wire. In fact all the nodes are situated virtually as if the energy producing the second shock in the resonator travelled quite close to the wire until it arrived at the end, and thence moved directly on the circle. It must, however, be observed that this mode of viewing the reflexion does not justify the considerable recession of the first node when the resonator is close to the conducting wire; a recession which, according to the experiments of Sarasin and De la Rive, is greater as the circle itself is greater. We believe, however, we have now well established that this recession is due to the geometrical form of the resonator; the electrical shocks, arriving along two rectilinear conductors, tend each to charge the nearest point of the resonator, so that the first oscillation is produced between the diametrically opposite parts of the circle. As, however, the electricity tends to oscillate in all the amplitude of the circle, this takes then its normal period, and the nodes are arranged along the wire. We have in the second place investigated how the plane of the resonator must be turned about its vertical axis in order that the total effect of the two "shocks" which excite the oscillation may be as great as possible. These directions are indicated in fig. 1. They have for the most part been determined where the maxima should be according to the measurement of the nodes. Among other orientations figured are two which have been taken in the nodes (dotted lines). These directions of maximum effect pronounce strongly, it seems to us, for a direct radiation from the end of the wire. Of the three series of observations made in the centre of the loops it follows that the perpendicular to the circle is almost along the bisectrix of the angle which the line going directly to the end of the wire forms with the parallel to the conducting wire. The two observations made in the nodes themselves show that to have the maximum effect the circle must be orientated so that the electrical undulations coming from the end of the wire arrive perpendicularly to its plane, and that the action of the second shock is by this fact annulled. The results which we have enunciated enable us to give a certain development to the conception of the moving electrical tube devised by some English men of science. Assume that the electrical tube of the direct wave, which moves at each point at right angles to its momentary direction, is almost rectilinear, and at right angles to the conducting wire: this will no longer be the case with the reflected wave. Now measurements made by one of us of the interferences on the surface itself of the conducting wire show that in this case there is no appreciable recession of the first node, and that therefore the part of the electrical tube immediately near the conductor turns about the end of this almost instantly. But the distant parts of the tube cannot traverse simultaneously the same angular distance; they remain behind and the electrical tube curves then almost like the tail of a comet about the end of the wire (fig. 2). This would be the origin of this characteristic radiation proceeding from the end of the conductor, the existence of which we may admit. It would follow from this that as the elements of the tube continue to move at right angles to their direction for the time being, the energy escapes from the end of the wire and is lost in the surrounding space*.-Comptes Rendus, Nov. 6, 1893. DENSITIES IN THE EARTH'S CRUST. To the Editors of the Philosophical Magazine and Journal. GENTLEMEN, Probably many of your readers will have read the Introductory Review in Mr. Blake's Annals of British Geology' for 1892, and will, like myself, have been surprised at the very direct manner in which he has impugned the accuracy of Mr. O. Fisher's conclusions regarding the comparative densities and thicknesses of those parts of the earth's crust which underlie oceans and continents respectively. Mr. Blake undertakes to show that Mr. Fisher's method "is wholly fallacious." After stating the method referred to, he gives three reasons from which "it is easily seen that this method cannot lead to correct results." The manner in which these reasons are expressed does not make it clear to the ordinary reader that Mr. Blake has entirely upset Mr. Fisher's reasoning; but when he narrows the issue to "the deånite point where the fallacy comes in," the non-mathematical geologist must wish to know how far Mr. Blake's mathematics are correct. I think that all who are interested in terrestrial physics must hope that Mr. Fisher will either admit or deny the value of Mr. Blake's criticism, and as he cannot reply in the Annals of British Geology,' perhaps he will favour your readers with a note on the subject, the issues being of considerable importance. A. J. JUKES-BROWNE. There should be an appreciable loss of energy on reflexion. We hoped to complete our research in this direction by measurements on the wire itself (conf. Birkeland, Wied. Ann. vol. xlvii. p. 583). For three wave-lengths A1 = 6m, λ= 2·7 m, and λ= 1.2 m, we found the reflected wave respectively 0.6, 0.45, and 0.35 of the direct wave. we have devised another method of directly measuring the loss and we have not again found these values, so that we only give them under reserve, not having succeeded in explaining these contradictory results. But THE LONDON, EDINBURGH, AND DUBLIN PHILOSOPHICAL MAGAZINE AND JOURNAL OF [FIFTH SERIES.] MARCH 1894. IN XXII. The Luminosity of Gases. PART I.-The Luminosity of Flames free from Solid N a brief note communicated to the Chemical Society in 1892 (Proc. Chem. Soc. no. 105) I described some experiments on flame-coloration made with my flame-cone separating apparatus, and in order to explain the results obtained I adopted provisionally the view that flame-spectra in some cases must be regarded as the direct effect of chemical action. Almost simultaneously E. Pringsheim read before the Akad. d. Wissenschaften, Berlin, a paper on "Kirchhoff's Law and the Radiation of Gases," which was subsequently printed in Wied. Ann. xlv. p. 428, 1892. In this paper Pringsheim records an elaborate series of experiments on the radiation and absorption of sodium vapour, and arrives at the conclusion that the spectrum of sodium salts in a flame and of sodium vapour heated in neutral gases is due directly to chemical action. In a subsequent paper ‡ Pringsheim has extended his experiments to other substances and has obtained similar results. * Communicated by the Author. † A summary of this part of the paper was read before the Chemical Section at the Nottingham meeting of the British Association, Sept. 19, 1893. Wied. Ann. xlix. p. 347 (1893). Phil. Mag. S. 5. Vol. 37. Nọ. 226. March 1894. S |