"Throughout this work the relativity equation of Sommerfeld is assumed to apply exactly. This does not represent the neglect of the spinning electron theory or the new wave mechanics, because these theories together seem to give exactly the equation of Sommerfeld (Richter, Proc. Nat. Acad., 1927, XIII, 426). Furthermore, the agreement of the Rydberg constant, computed by this equation, from the first three lines of the Balmer series, shows the very exact experimental applicability of this equation, apart from the various theoretical ways of deriving it (Houston, Phys. Rev., 1927, XXIX, 748).

"To apply this equation, it is necessary to know the wave-lengths of the individual components of the fine structure. For this purpose the hydrogen line 6563 was examined with the compound interferometer (Houston, Phys. Rev., 1927, XXIX, 478). Fig. 1 [not given here] shows a microphotometer curve from a plate taken with this instrument. The assymetry in the shorter wave-length component is so pro

nounced that it is possible to determine the position of the third component which is causing it. This is the component predicted by the theory of Sommerfeld and Unsold, and which has previously been inferred from the displacement of the maximum (Houston, Astrophys. Journ., 1926, LXIV, 81). The assymetry has also been observed by Hausen (Annalen, 1925, LXXVIII, 558), and by Kent, Taylor, and Pearson (Phys. Rev., 1927, xxx, 266). Because of this assymetry, only the long wave-length component is used in the computation of the Rydberg

constant.

"However, although this long wavelength component appears symmetrical in the figure, the theory requires that it shall be multiple, and consist of one strong and two weak components. If we assume the distribution of intensities predicted, we can correct the position of the observed centre of gravity to the position of the strongest component. Table III gives the results of this correction and the values of RH.

COR. + .0056

+ .0024

The uncertainty in the individual values is equivalent to the mean deviation of the observed wave-lengths, while the uncertainty of the means is merely the mean deviation.

=

"In He 4686 the two strong components were identified with the components la and IIb. With this identification the values of RH. are given in Table IV.

λ (AIR). 4685.7030 4685.8030

v (VAC.). 21335.5622 21335.1070

"These mean derivations can hardly be taken as representing the precision of the values of R, and yet it seems reasonable to believe that, relative to each other, these values are correct within 0.020. Of course, the actual error from the correct value based on the cadmium standard is possibly larger, perhaps of the order of 0.050.

"The difference between this value of RH. and that given by Paschen is due to the fact that his value gave more weight to his component 4685.809 than to his component 4685.703 [this wave length is not of course to be confused with another one of

a conse

In the above quotation the writer has made a few trivial changes in the text which in no way alter the meaning. For example, in setting up equations by machine composition there appears to be no stock matrix for the inferior infinity sign and as This quence a figure eight is laid down. takes time and adds to the cost of setting up technical matter which renders scientific publication so expensive, and is a source of serious concern with societies and publishers generally. Now this can be obviated by using inferior letters if the printers have matrices for them, as in the case in point. Another disadvantage of built-up equations by hand setting is that the signs, letters, etc., are apt to get displaced, and the infinity sign might appear anywhere. These remarks would not apply if the cost of printing were not so great relative to the profits (if any) with scientific publications. The equations have been numbered for reference later.

It should be explained here that it is tacitly understood that the mass value of the electron given by Houston would not give a correct answer in the case of equations 1 and 2 on account of the relativity effect.

Since it has been shown that Michelson's new determination of c (in his 66 Studies in Optics," 1927, p. 137, the vacuum value is given as 2.99796 × 101 cms. per second, with a possible error of 1 in the last digit given), of much greater precision than hitherto, can be calculated by the equation (see Loring, Chemical News, 1928 exxxvI, 33) :Rint/(24 x 10-33/h) × 1010

infs

=

= c = 2.9979640...

when Rint 109737.37 and h 6.556 × 10-27, this suggests trying this value for Rin as it is in close agreement with Houston's value :- 109737.424; and, moreover, near to Birge's value (Phys. Rev., 1926, XXVIII, 848) :-- 109737.3 ± 0.3.

Now F. W. Aston has quite recently redetermined a large number of atomic masses with a high degree of precision (see Aston : Bakerian Lecture, Roy. Soc. Proc., 1927, A, cxv, 487; and also Nature, Dec. 31, 1927, p. 956), and he has arrived at the following values for H and He on the O= 16 basis: H = 1.0077. He 4.0021.

=

The last digit is uncertain according to Aston, which is indicated here by the inferior figures. It does not seem likely that Aston would have made the values any higher than necessary on account of the previous work by experienced experimentalists. The helium value (atomic weight) does not agree with the one by Baxter and Starkweather which was used by Houston (see citation above).

says:

It should be noted, however, that Aston "The new instrument [mass spectrograph] has five times the resolving power of the old one, far more than sufficient to separate the mass lines of the heaviest element known. Its accuracy is 1 in 10,000, which is just enough to give rough first order values of the divergences from whole numbers."

It will, nevertheless, be of interest to try Aston's values in connection with Houston's spectroscopic findings. At the same time an adjustment is made in m to justify the slightly modified Rint constant and

Houston's RH constant. Then

The value for m was adjusted by trial and error, but no claim for absolute accuracy can be made. In fact it is absurd to attempt 6th and 7th decimal places with precision, but the trial figure came out with a tendency to repeat. Probably the true value lies between 0.000545 and 0.000547. 0.0005457 was used by the present writer (see Chemical News, 1927), but this value. seems a little low when adopting Aston's atomic weight values, especially in the case of He.

Now since the above equation for evaluating c seems as accurate as could be expected, it becomes possible to evaluate the writer's A with apparent precision, and it comes out at 0.015251652262... See Chemical News referred to above during the year 1927. Ot was shown that

Curtis. It will be remembered that W. E. Curtis (Roy. Soc. Proc., 1919, A, xevi, 147) gave a range of calculations for RH as follows: 109678.28; -109678.73; 109678.10; 109677.58; 109678.76; 109677.79, as worked out by the methods of different physicists. The last one due to H. S. Allen is nearest to the one above except for Bell's value which is in agreement. The first value in this set is the one due to Curtis, and this value has been taken by Fowler in his "Report on Series in Line Spectra," the edition being

1922.

These values are, however, narrowing down to a variation in the decimal region, and the above one by the writer as a result of these calculations cannot be said to be in serious error. There are certain features about the hydrogen lines that perhaps require further investigation, but this is more a matter for experienced workers to consider. Paschen's value was 109677.691 from which Rint was calculated as 109737.11. Now if the latter increases in the decimal region it is to be expected that the former might do so also, and this appears to be the

case.

Now turning to equation 4, and using these new values, in which mH/m 1843.8635, the result here is

e/m = (mH/m)(9647/mн) = 17659995. This is, of course, practically

1.7660 x 107 E.M U. per gram.

This value agrees fairly well with the European continental physicists' value given in the first part. In fact it is a somewhat better comparison to make as Babcock's value was a mean of determinations that showed a very wide range of variation, though this fact of variation might apply more or less generally but the word somewhat qualifies the statement sufficiently.

In the concluding lines of Houston's important paper, mention is made of evaluating AvH. Now Avн is in one sense a variable quantity, but it appears to the writer that in the equation, as in Sommerfeld's book, the evaluation of this constant from e, h, etc., really means that a limiting value is obtained. Such a value would, of course, be higher than those observed in the case of the usual hydrogen doublets, and the writer evaluated a numerical quantity designated AH, which by virtue of the multiplier 24 (see equation for the velocity of light above)

This led to another method of determining e/m. The writer found that certain agreements in the various relations described in the Chemical News, 1927, involving Aн, could be brought into harmony by introducing a correcting term in certain equations in which occurred, and it was concluded that it would be worth while considering a modified called . This obviated the use of a higher Rint than seemed possible. In other words Rint as below could be used, and it was finally found that e/m could apparently be calculated thus:

T

17659995

17665515

17671155

17674000 17674961

It does not seem possible to arrive at any definite conclusion, but 1.767 x 107 would seem a fair value to take, pending further information, especially for the e/m value.

IMPERIAL CHEMICAL INDUSTRIES LIMITED.

The offer made on February 15 to the shareholders of the Tees Salt Company Limited to exchange their shares for shares of Imperial Chemical Industries Limited has been accepted by holders of over 75 per cent. of each class of capital of the former Company, and transfers for the carrying out of the exchange are being issued.

:

TRADE FROM FAIRS. Lord Strathspey writes: "British industry has achieved so noteworthy a victory at our record Fair that I may be pardoned for suggesting that we might look to Europe and overseas for other fields to conquer.

If business men from abroad have found it profitable to visit our Fair, might our own business men not find it equally profitable to mingle in the Fairs of other lands? Yet I find that last year the number of British business men who thought it worth while to see the Swiss Industries Fair at Basle was just-eleven! We were the sixth country in their visiting list, France leading with 679 commercial ambassadors and Germany coming a close second with 666. His Lordship thinks that sorry record ought to be improved upon when the Basle Fair once again opens its doors next month from April 14-24. Switzerland is a prosperous country with a large demand for the textiles, machinery and so on that we could give her; nevertheless I find that, while she does purchase more cotton goods from us than from any of her neighbours, we are beaten in woollen goods by France and Germany; by these countries and by the United States and Italy in machinery."

THE ELECTRON.

On Saturday, March 3, Prof. Sir J. J. Thomson, Master of Trinity College, Cambridge, delivered the first of the Founders' Memorial Lectures at Girton College. He described how recent experiments, especially those made by Professor G. P. Thomson on the passage of electrons through very thin films of metal had thrown quite a new light on the nature of the electron. They had proved that a uniformly moving electron was accompanied by a train of waves which preceded and guided it, the electron following in the wake of the waves. The electron had thus, like light, a dual structure; one part, that containing the energy, was corpuscular; the other, the part controlling and guiding the path, was undulatory. The waves which accompanied the electron vibrated far more rapidly than the hardest Röntgen rays; the only rays approaching them in this respect were the y rays from radio-active substances. The vibrations were quicker when the electron moved fast than when it moved slowly, but however slowly it might move, the waves never made fewer than a very large number of vibrations per second. Short as these waves were, their lengths were more than 10,000 times the

diameter usually ascribed to the electron.

These results showed that the early conception of the electron as a point-charge of negative electricity surrounded by a struetureless medium could not be adequate; the electron or its surroundings must be much more complex; there must be something beyond the electron. If, however, they supposed that the electron did not represent the final stage in the structure of matter, but that it was itself composed of, or surrounded by, something which might be regarded as a mixture of sub-electrons and sub-protons, it might be shown that its properties would resemble those which it had lately been shown to possess. Through such surroundings energy would be accompanied and guided by waves, and the connection between the speed with which the energy moves and the wave-length of the waves would be the same as that found in Professor G. P. Thomson's experiments.

An example on a cosmical scale of a structure analogous to that assumed for the electron and its surroundings was the Heaviside layer in the upper region of the atmosphere, where free electrons and protons were mixed with the ether. Through that layer energy could travel at any speed less than that of light if accompanied and guided by suitable waves, all of which were of high pitch. The consequences of that theory, which was founded on purely physical considerations, agreed in a remarkable way with the consequences of the very interesting wave dynamics due originally to Louis de Broglie and extended by Schrödinger and others.

CANADIAN TALC AND SOAPSTONE INDUSTRY SHOWS PROGRESS. The production of talc and soapstone in Canada in 1926 totalled 15.767 tons, valued at $217,195. These figures represent an increase of 1,293 tons and $11,360 over 1925. The first six months of 1927 show a slight decline of 159 tons and $3,707 over the corresponding period of 1926.

The bulk of the ground talc produced in Canada continues to come from the Madee district, Hastings County, Ontario, where important deposits of superfine, foliated, white tale have been worked for over twenty years. The Madoc output is ground in local mills, and finds employment in the talcum powder, paper, textile, soap, and rubber industries. Domestic consumption is small; the greater part of the production being exported.