definite composition by suitable heat treatment produced no apparent change in the crystal form of the alloy. 7. The superposition of a magnetic field of 3500 gauss did not alter the crystal form of the alloys or make any noticeable change in the relative intensities of reflexion from the different crystal planes, indicating no change in orientation of the crystal unit. 8. Arguments have been presented tending to show that the ultimate magnetic unit is not the molecule or atom or any complex group of these, but that the magnetic effects are associated with the valence electrons. In conclusion, the writer wishes to express his appreciation and thanks to Professor J. C. McLennan, who suggested the problem and has materially helped with the research, and also to the Honorary Advisory Council for Scientific and Industrial Research, Ottawa, Canada, who made this work possible by the award of a Fellowship. University of Toronto, XXIX. Disintegration in Discharge Tubes. By H. P. WARAN, M.A., Ph.D. (Cantab.), F.Inst.P.* THE [Plate V.] 1. Historical. HE disintegration of discharge tubes may be divided broadly into two classes (a) the disintegration of the electrodes known as the sputtering, (b) the disintegration of the glass walls. Of these, the disintegration of the electrodes has been noted from very early times. Thus, as early as 1861, it had formed the subject matter of a paper by Gassiot † to the British Association. However, Sir William Crookes was the first to conduct a systematic investigation of the phenomenon using different metals for his electrodes. Recently Kohlschütter § and the research staff of the General Electric Co. have investigated the phenomenon in great detail. On the other hand, very little is known about the Communicated by Prof. A. W. Porter, D.Sc., F.R.S., F.Inst.P. Proc. Roy. Soc. 1. p. 88 (1891). $ Jahrb. d. Rad. U. Elek. ix. p. 355 (1912). || Phil. Mag. xlv. p. 98 (1923). Phil. Mag. S. 6. Vol. 46. No. 272. Aug. 1923. X * disintegration suffered by the glass walls, probably for the reason that its effects are not so very marked. But, as Dr. Coolidge has pointed out, in X-ray tubes that have failed from the cracks resulting from the chipping off of the glass in the neighbourhood of the cathode, such disintegration is prominently noticed. In special cases where the effects get accentuated by a focussing of the ions, the intense local heating melts the glass, and the pressure of the external atmosphere blows in a hole of the type shown at X in fig. 1. In the present case a cylinder of aluminium formed the electrode, sending out a concentrated axial pencil of ions that hit the glass at X, the tube A not having been blown quite axial with the cylindrical electrode. In an earlier paper† on the subject I drew attention to the curious type of channels cut on the glass walls when a discharge is deflected on to it by a transverse magnetic field. 2. Effect of the positive ions. As in the case of cathodic sputtering the agency responsible for the observed effect is mainly the ionic bombardment of the walls. Such a bombardment may bring about the disintegration, either mechanically or thermally, or by a combination of both actions. The apparent characteristics of the convergent set of markings first obtained led me to the suggestion that the markings are of the nature of tracks left by the motion of the particles along the glass walls. In that case it is the kinetic energy of the ion that is responsible for it. The lightest positive ion of hydrogen has a velocity about a tenth of that of an electron, a mass nearly 2000 *Phys. Rev. xi. p. 409 (1913). + Phil. Mag. xliii. p. 226 (1922). times as great, and a kinetic energy about twenty times as great. Hence there was no doubt that the disintegration was produced mainly by the positive ions. 3. A Mechanical Theory of Origin. The evidence in support of such a mechanical orgin might be summarized briefly as follows. Firstly, the short light scratches on the glass could be imagined as arising from the sand-blast action of the swiftly moving positive ions moving towards the cathode, an equality in the energy of the particles accounting for the nearly equal length of the markings. Secondly, the peculiar shape and convergency of the tracks as shown in fig. 2, may be due to the deflecting Let such a positive action of the transverse magnetic field. particle strike the cylindrical wall at A. So long as the particle has a sufficient forward velocity the magnetic field exerts a deflecting force at right angles to the axis of the tube. The cylindrical walls constraining such a motion, the particle can reach a higher level B only by travelling as shown in (a) along the cylindrical walls, and the track whose projection is AB is thus inclined to the axis of the tube. The bend BC towards the axis as shown at (b), may be due to the loss of velocity suffered by the particle after it has travelled the length AB, the particle of slower velocity being forced to a higher level C by motion along BC which is shorter than the straight path BC'. Such an explanation, though fitting some of the facts observed, is obviously defective. For instance, in the absence of the magnetic field there is a uniform distribution of the positive particles across the cross-section of the discharge, and when they get deflected by the magnetic field |