electric quantities by the canal and cathode rays exceeds the exciting current. The net therefore appears to absorb less than corresponds to the area occupied by the meshes.

Moreover, it may be concluded that although apparently the values of the cathode and canal rays are somewhat different; yet it seems fairly probable that when absorption &c. are fully considered, these values become equal, so that the production of the cathode and canal rays is nothing else than a splitting up into positive and negative ions, which travel in opposite directions from the electrodes, and derive their charges from the current. The different degrees of absorption are in any case connected with the differences in the velocity and size of the particles forming the cathode and canal rays.

In conclusion, my hearty thanks are due to Prof. Dr. W. Wien for his many-sided suggestions and friendly help, as well as to Prof. Dr. Des Coudres, Privatdozent Dr. W. Sitz, and Dr. Fr. Harms, to whom I am indebted for much advice. Würzburg, Physical Institute. June 1903.

XVIII. On a Norel Instrument for Drawing Parabolas. By KARL PEARSON, F.R.S., University College, London*. [Plate XV.]

FOR

NOR a number of years I have much desired a really effective instrument for rapidly constructing parabola on the drawing-board. As far as my experience goes existing mechanisms for this purpose are occasionally ingenious, but always ineffectual. Yet in ordinary drawing-office practice the construction of parabolas in both graphical statics and graphical dynamics is an almost daily necessity. One has only to think of the solution of continuous girder problems, of speed-curves from energy-curves, of stability of dams, and a variety of other matters to realize how much energy is wasted in the drawing of these curves which ought to be done rapidly and effectively by aid of a mechanism.

The grant made by the Drapers' Company to my department has placed me in a position to carry out, in a practical form, some of the needs we have long felt for labour-saving mechanisms in our drawing-office, and the following brief description of the new "parabolograph" may be of interest to those who have felt the like want.

The principle made use of in the new mechanism is the fundamental metrical property of the parabola. Let PVP/ (Pl. XV. fig. 1) be a parabola, V the vertex, and VX the axis. If P be a point on the parabola PN2=cx VN, where PN is the perpendicular on the axis. Draw PFT parallel to the axis. Join VP and take VT perpendicular to VP to meet PFT in T. Drop VF perpendicular on TP. Then clearly FV2= FT× FP since the angle at V is right. It follows, therefore, that FT=c the parameter of the parabola. Hence, if a bar TP slide so as always to remain parallel to the axis, with a definite point F slipping along the tangent FV at the vertex, then a bar TVP, bent at right-angles at V, round which point it pivots, and passing through a fixed point T on FT, will give points on the parabola by its intersection P with the same line.

This simple property was used by my former assistant Mr. H. Payne, now Professor of Engineering in the South African College, to design a parabolograph. This with his consent I forwarded to Herr Coradi, of Zürich, as the man most likely to make a really practical machine, the step from perfect theory to effective practice being, as I know from experience, a rather long one.

The machine as designed by Payne and executed by Coradi is represented in the accompanying fig. 2 (Pl. XV.), which almost explains the method of working. Herr Coradi has introduced *Communicated by the Author.

By

XIX. Heating Effect of the Radium Emanation*. E. RUTHERFORD, F.R.S., and H. T. BARNES, D.Sc., Professors of Physics, McGill University, Montreal †.

P. CURIE and Laborde first observed the rapid rate of heat emission of radium, and deduced that 1 gram of radium emitted heat at the rate of about 100 gram-calories per hour. In a later paper P. Curie § found that the rate of emission of heat depended upon the age of the radium preparation. The heating effect for freshly prepared radium compound was small at first, but gradually increased to a maximum after a month's interval, and remained constant over a further interval of two months.

The present experiments were undertaken with the view of seeing how the heat emission of radium is connected with its radioactivity. It has been shown by Rutherford, and Soddy || that the radiation emitted from a radium compound in a state of radioactive equilibrium may be divided into three parts :

(1) A non-separable radiation consisting entirely of a rays and constituting about 25 per cent. of the total radiation.

(2) The radiation from the emanation occluded in the radium, also consisting entirely of a rays.

(3) The excited radiation produced by the emanation in the mass of the radium, and consisting of a, B, and y rays. (2) and (3) together constitute about 75 per cent. of the total radiation.

Some experiments have been recently made to find how much of the activity of radium is supplied directly by the emanation occluded in it. The saturation-current, between parallel plates, due to a radium preparation spread uniformly over a platinum plate, was determined by means of an electrometer. The platinum plate was then heated rapidly to a temperature sufficient to completely drive off the emanation and the saturation-current due to the radium immediately measured. There was a decrease observed corresponding to 18 per cent. of the total. The gradual decay of the excited activity left behind in the radium after the removal of the emanation is shown graphically in fig. 5, curve A (p. 213).

We may thus conclude that the emanation supplies

A short account of the preliminary results was published in 'Nature' (Oct. 29, p. 622, 1903). Read before the American Physical Society, St. Louis, Dec. 29, 1903.

+ Communicated by the Authors.

↑ Comptes Rendus, cxxxvi. p. 673 (1903).

$ Société de Physique, 1903.

Phil. Mag. April 1903.

18 per cent., the non-separable activity 25 per cent., and the excited activity 57 per cent. of the total activity of radium.

The excited activity produced on bodies has been shown to be due to a deposit of radioactive matter on their surface. The term "excited activity" refers only to the radiations from this active matter. It is convenient to have a definite name for the matter itself. It is suggested that the name emanation X" be given to it, since the matter which causes excited activity is produced directly from the emanation. This name is given from analogy to the products UrX and and ThX, which are produced directly from uranium and thorium respectively. On this nomenclature, the radium produces the emanation at a constant rate, and this in turn is transformed into the emanation X. The matter of emanation X of radium itself undergoes at least three and probably four successive changes. The nature of these changes and their connexion with the radioactivity will be discussed later.

On heating or dissolving a radium compound in an open vessel, the emanation is released and can be entirely removed by a current of air. The emanation X, which is non-volatile, is left behind with the radium, and it at once commences to lose its activity. In the course of a few hours the activity due to it has practically disappeared. The 6 and 7 rays which are produced only by emanation X disappear from the radium at the same time, and there then remains a non-separable activity of radium consisting entirely of a rays.

At the same time that the emanation X, left behind in the radium, is undergoing change, fresh emanation X is being produced by the separated emanation, and at such a rate that the activity at any time due to the emanation X left in the radium, together with that due to the emanation X formed afresh by the emanation, is equal to the original activity of the emanation X stored up in the radium.

Since fresh emanation is being continually produced by the radium and occluded in it, the activity of the radium after falling to its minimum gradually rises again, and in the course of about a month has nearly reached its original constant value.

The experiments which will now be described were undertaken to see if the heat emission of radium varied in the same way as its activity when the emanation was removed. For this purpose, the heating effect of the radium was first determined. The emanation was then removed from it and collected by condensation in a small glass tube, and the distribution of the heating effect between the emanation and emanation X and the radium was determined, and also the