May 2 1, 1908] 



NA TURE 



55 



Radio-activity of Potassium and other Alkali Metals. 



Prof. McI.ennan's .lelter in Nature of May 14 (p. 29) 

 iiialcL-s it desirable that I should record that my experi- 

 ments continue to be in contradiction to those he describes. 

 Whether the activity of potassium is due to an extraneous 

 impurity or to the' element itself, it is to be expected that 

 some separation of the activity should be possible. For 

 the last year and a half I have been trying to effect such 

 a separation without result. All samples of the same pure 

 salt which I have procured or prepared are identical in 

 activity. In one experiment a sample of the sulphate was 

 crystallised twenty-two times, in another a sample of the 

 nitrate was crystallised eighteen times without introducing 

 any difference between the final products ; in a third a 

 sample of potassium sulphate was prepared direct from 

 wood ashes, and found to agree with a commercial sample 

 prepared from the Stassfurt deposits. I propose now to 

 prepare a third sample direct from felspar. 



The only difference which I can detect between Prof. 

 McLennan 's experiments and my own lies in the fact that 

 he places the active material inside the ionisation vessel, 

 while I place it outside an aluminium window. If he were 

 measuring the effect of very soft rays, the difference 

 between our results might bo explained. 



Norman R. Campbell. 



Trinity College, Cambridge, May 14. 



On Dispersion and Spectrum Series. 



In reply to Mr. Campbell's letter of April 30, it must 

 suffice to point out that my letter of March 5 was limited 

 to infinite spectrum series and to luminous gases. This 

 seemed to me at the time of writing obvious from the con- 

 text, as well as from the express reference to the finiteness 

 of the refractive index of luminous hydrogen. Apparently 

 I was mistaken ; anyhow, Mr. Campbell's suggestion that 

 I confused the emission lines of luminous with the absorp- 

 tion bands of dark hydrogen rests on a misconception of 

 my meaning. 



To avoid further misunderstanding, I will add that by 

 " electrical theory of absorption and dispersion, of mag- 

 netic rotation and Zeeman effect," I mean, of course, the 

 theory of Drude, together with its extension by Voigt to 

 all magneto-optic phenomena ; I do not mean theories such 

 as those of Lorentz and Ritz, which deal with the Zeeman 

 effect alone. I know no reason for confining that theory 

 to one member only of a series, e.g. the D lines of sodium, 

 and contend that it leads to contradictions when applied 

 to all the members of an infinite series of similar lines. 



The remaining points raised by Mr. Campbell seem to 

 me foreign to my argument. I cannot discuss them in the 

 space at my disposal here, but hope to do so elsewhere. 



Bonn, Mav 8. G. A. Sciiott. 



Secondary Waves of Light. 



It has hitherto been held that, so long as the diffrac- 

 tion apertures used (cut in perfectly opaque or perfectly 

 reflecting screens) are large compared with the wave-length 

 of light, Fresnel's expression for the amplitude of the dis- 

 turbance due to a surface-element gives us a close 

 approximation to the observed diffraction effects, and that 

 the exact value for the obliquity factor is of little import- 

 ance (e.g. see Schuster's "Optics," sec. 48). That this 

 is true only in the special case in which the apertures are 

 held normal to the waves of light, and not in other cases, 

 is shown by some new diffraction phenomena that I have 

 made the subject of study. 



The only experiment so far known which might seem 

 to show effects due to the obliquity factor is the well- 

 known one with the circular disc, but it is really in- 

 conclusive. The observed fact, that the illumination along 

 the axis of the disc decreases as the disc is approached, is 

 more or less entirely due to minute irregularities in the 

 rim of the disc, and not, as is sometimes stated, to the 

 increasing obliquity of the secondary waves producing the 

 illuniination. 



The theoretical grounds on which my experiments were 

 based were these : — if diffraction bands are produced and 

 observed in a direction in which the amplitude of the 

 disturbance in the secondary waves varies rapidly from 



NO. 2012, VOL. 78] 



point to point, we might expect effects due to varying 

 obliquity. Such effects would obviously not occur if the 

 diffraction aperture or mirror is, as is usual, held normal 

 to the waves of light incident on it, but might if it be 

 held obliquely. 



In the Philosophical Maga:)ine for November, 1906, I 

 showed that the diffraction bands due to a rectangular 

 aperture held very obliquely are not equidistant, that the 

 band-width increases progressively from one side of the 

 pattern to the other, and that the number of bands on 

 one side of the pattern is limited. They are most easily 

 seen on the spectrometer if the image of the slit of the 

 instrument formed by light reflected very obliquely from 

 the face of a prism is observed. The positions of the 

 minima of illumination, actually observed, are closely in 

 agreement with those calculated from the usual formula 

 (cos 9 — cos 0= +ttA/a), e, being the complements of the 

 angles of incidence and diffraction. Further observations 

 have elicited the following : — from the expression for the 

 intensity of the illumination in the pattern deduced by the 

 ordinary method 



= rZ- Sin- - - (cos S - cos <p) -^ „ (cos 6 - COS <f )-, 

 A \~ 



it would appear that the maxima of illumination in corre- 

 sponding bands on either side of the middle one should be 

 of equal brightness. This is flatly contradicted by observa- 

 tion, both visual and photographic. It is found that the 

 bands on one side are considerably fainter than those on 

 the other, and this difference becomes very large as the 

 light approaches grazing incidence. The illumination in 

 the diffraction pattern (with a given angle of incidence) 

 decreases and dies away as we approach the limiting plane 

 of the fringes, which is the plane of the reflecting surface 

 (ff in the diagram). 



This effect is inexplicable if the question of the variation 

 of the amplitude in different directions of the secondary 

 waves, supposed to be sent out by the elements of the 

 reflecting surface ff, is not taken into account. It must 

 be remembered that we are not dealing with apertures 

 small compared with the wave-length ; both the aperture 

 and its projection are large compared with A, and there 

 are no polarisation effects observed. The question may be 

 attacked analytically, and it can be shown that an element 

 of the surface of a reflecting body is equivalent in its 

 effect to a double source of appropriate intensity which, it 

 is known, produces zero effect in its equatorial plane and 

 a maximum along its axis. The effect of an element of 

 the surface FF is therefore zero along the line ff, and in 

 other directions increases as we move away from the line 

 FF. Remembering that the elements are not in the same 

 phase, and integrating their separate effects, we get an 

 explanation of the phenomenon observed. 



A fuller discussion and a mathematical investigation will 

 be published in due course. I found that similar effects 

 are observed when the transmitting aperture is used. Some 

 experiments with coarsely ruled gratings are in progress 

 which seem to point in the same direction. 



C. V. Raman. 



Science .'\ssociation Laboratory, Calcutta, April 2. 



The Corrosion of Iron and Steel. 



In Nature of May 14, Dr. J. Newton Friend alludes to 

 GalUonella ferruginea as obtaining its life's energy by 

 oxidising ferrous carbonate and organic ferrous salts, 

 causing the precipitation of rust, or ferric hydroxide. May 

 I point out that GalUonella ferruginea can live and grow 

 well without anv iron at all. and so cannot be a vital factor 

 in the metabolism of the bacterium, using the term 

 ** bacterium" in its widest sense? The oxidation which 

 takes place can be simply explained by the fact that 

 ferrous carbonate in solution is very unstable, becoming 

 very rapidly oxidised. W. F. Macfadyen. 



54 Dunard Street, Glasgow, May 16. 



