^o8 



NATURE 



[January 5, 191 1 



difficult. But this Prof. Birkeland says was likewise true 

 of Toronto, yet Dr. Krogness retains this station ; why he 

 rejects Honolulu Dr. Krogness does not say. Again, he 

 overlooks the fact that when he corrects Birkeland 's 

 scaling for San Fernando he has improved the easterly 

 progression — Prof. Birkeland 's value was nearly two 

 minutes too high. In view of the uncertainties in Prof. 

 Birkeland's scalings revealed by Dr. Krogness, and as 

 Prof. Birkeland fails to specify the particular element con- 

 sidered, not full weight could be attached to this disturb- 

 ance in the above table. It should also be stated here 

 that Prof. Birkeland considered, in all, six characteristic 

 points of the disturbance curve, and my result was based 

 on all the scalings — seventy-two in number — and not 

 merely the half-dozen taken by Dr. Krogness. Did I 

 myself consider such limited data as Dr. Krogness uses 

 adequate for the purpose, I might point out that his own 

 figures show an easterly progression of the times on the 

 order of what is to be expected, which would have been 

 still further accentuated had he not rejected Honolulu. 



Just as I am preparing this reply, I am in receipt of a 

 letter from Dr. Chree, dated December 6, 1910, accom.- 

 panying a copy of the proof-sheets, which he kindly let me 

 see, of his paper before the Physical Society, November 1 1 ; 

 he had also given a paper on the same subject at the 

 British Association meeting. He is not in agreement with 

 my general deductions or with those of Mr. Faris. His 

 criticisms are in part covered by the foregoing reply to 

 Dr. Krogness, and in part \iy my article in the December 

 (1910) issue (loc. cit.). I can only say here that I cannot 

 agree with Dr. Chree in several of his own deductions, 

 and especially with regard to the possible inaccuracy of 

 Mr. Faris's time scalings: I beg to refer him to pp. 213-4 

 (loc. cit.). Nor can I enter here into a discussion with 

 regard to his criticisms of my hypothesis of ionic currents, 

 for it would seem that he has unintentionally put into his 

 discussion ideas which are new to me. I will only re- 

 mark that nowhere in my papers have I supposed such a 

 simple overhead electric current in the plane of the geo- 

 graphical equator as postulated by him ; this is best shown 

 by my mathematical analysis. 



In conclusion, I would like to state my position once 

 more, viz. even our most sudden magnetic storms begin 

 at measurably different times for various stations dis- 

 tributed over the globe. The data thus far available would 

 show that the Greemvich mean times of beginning increase 

 more often in an eastward direction than in a westward 

 one. 



Our explanations as to the cause may differ, but I 

 believe what I have just stated to be an actual fact. 



L. A. Bauer. 



Washington, D.C., 



December 19, 1910. 



Tribe Luminescence of Uranium. 



MoissAN first directed attention to the pyrophoric 

 properties of metallic uranium. The luminosity shown on 

 shaking a bottle containing metallic uranium is due to the 

 oxidation of small particles of the metal. Uranium is a 

 hard but brittle metal ; when pieces of it rub together 

 small particles are knocked off, and if these are neither 

 too small nor too large the friction may be sufficient to 

 heat them above 170° C, at which temperature uranium 

 inflames in air. The presence of smaller particles, which 

 do not inflame visibly in air, is shown by their incan- 

 descence in a gas flame lit by the " spark " from the 

 metal. The luminosity obtained by rubbing metallic 

 uranium is not the same class of phenomenon as the 

 luminescence produced by shaking a tube containing 

 uranium nitrate ; the latter has been described as tribo 

 luminescence (Wiedemann). If the tube containing 

 metallic uranium is filled with hydrogen no luminosity is 

 obtained, whereas the luminescence of the uranium nitrate 

 is unabated in such an atmosphere. The sparks obtained 

 from uranium are hot enough to kindle a gas flame or 

 explode a mixture of hydrogen and oxygen ; in fact, I 

 have been able to work a petrol engine by igniting the 

 gas charger by means of such sparks. The luminescence 



NO. 2149, VOL. 85] 



of the uranium nitrate crystals, on the other hand, is 

 unaccompanied by any considerable rise in temperature. 

 Pyrophoric properties similar to uranium are shown to a 

 remarkable extent by Welsbach's alloys of rare earth 

 metals and iron. 



Tribo luminescence is shown by a large number of 

 organic and inorganic compounds, e.g. arsenic trioxide, 

 uranium nitrate, potassium sulphate, zinc sulphide, quinine 

 valerate, aniline hjdrochloride, benzoyl j8-naphthylamine. 

 Crystallo luminescence, or the luminosity produced during 

 crystallisation, is practically the same phenomenon, being 

 caused by the fracture of crystals after formation ; it is 

 well shown by mixtures of sodium and potassium sulphate. 

 Tschugaefi' found a connection between the optical activity 

 and the tribo luminescence of organic substances, but 

 Gernez has disputed the existence of any relation between 

 them. Substances that phosphoresce readily under X-rays 

 generally show tribo luminescence, and the connection 

 between the two phenomena is accentuated by the observa- 

 tions of Karl, which show that quite pure inorganic sub- 

 stances do not show tribo luminescence. It is remarkable 

 in view of the radio-activity of uranium that salts of this 

 metal should show phosphorescence and tribo luminescence 

 to such a degree; Karl has found, though, that quite 

 " pure " uranyl acetate does not show tribo luminescence, 

 while Tschugaefi' mentions that the chloride and sulphate 

 also do not exhibit this property, though they are all 

 phosphorescent. The tribo luminescence of crystals may be 

 likened — though analogies are dangerous guides to theories 

 — to the bursting of an elastic band with a snap ; when 

 the cohesive forces between the molecules of the crystal 

 are overcome the electrons are disturbed, and light waves 

 result, while substances which easily phosphoresce or are 

 radio-active would the more readily have their electrons 

 disturbed. 



Mr. Rudge mentions that the yellow oxide of uranium 

 shows slight tribo luminescence ; I could only obtain the 

 effect by fairh' vigorous rubbing in a mortar, and as the 

 oxide changes to a dark colour with this treatment, the 

 luminosity may be due to oxidation. 



Mr. Rudge's letter directs attention to two interesting 

 but distinct phenomena. 



Alfred C. G. Egerton. 



R.M.A., Woolwich. 



The Clarification of Liquids by the Process of 

 Tanking. 



I SHALL be glad if any of your readers can give me 

 information upon the following problem. In the clarifi- 

 cation of liquids by the process of tanking, the settled 

 clear liquid is drawn off from a tap situated on the side 

 of the tank above the muddy layer. When the tap is 

 turned on, does only the liquid above the tap run out or 

 does some of the liquid below the tap run out also? In 

 the special case of tanking oils, there is very little differ- 

 ence in specific gravity between the upper clear lajer and 

 the lower muddy layer. Further, how should the outlet 

 be fitted so that on running out the upper layer the lower 

 should remain least disturbed? 



Rowland A. Earp. 



Preston Brook, near Warrington, 

 December 22, 1910. 



The Conduct and Song of Birds. 



This morning, Thursday, is clear and frosty, but until 

 now we have had constant rain. In spite of this the 

 birds, for three or four days, have been singing as in 

 early spring. The rooks have been visiting their old 

 nests in the elms, and, our gardener assures me posi- 

 tively, have been carrying sticks and repairing their nests ; 

 this he has seen himself, and marked as exceptional. I 

 suspect that this (unusual?) conduct and song herald a 

 period of fine dry weather. 



F. C. COXST.IBLE. 



Wick Court, near Bristol, 

 December 22, 1910. 



P.S. — Fine weather here since December 22 until to-day, 

 January 2 ! 



