Jan. 2 2, 1 8 74 J 



NATURE 



225 



true order of their wave-lengths, and partly from the 

 nature of our ordinary photographic substances. He 

 proved that actinism, or the power of chemical decompo- 

 sition, does not particularly belonjj to the violet end of 

 the spectrum, but is found throughout its whole length. 

 But bromide and iodide of silver, as used in collodion 

 photography, are more readily decomposed by vibrations 

 of certain lengths and periods th.m by others, and hence 

 the excess of action seen at the violet end is a function of 

 certain silver compounds, and not of the spectrum. Other 

 substances, as carbonic acid, show maxima elsewhere, as 

 in the yellow region. The solar beam is therefore not 

 compounded of three forces, light, heat and actinism, but 

 it is a series of ethereal vibrations, which give rise to one 

 or other of these manifestations of force, depending on 

 the surface upon which it falls. 



In order to provide against this excess of action in certain 

 parts of the spectrum, I introduced a system of diaphragms 

 placed in the vicinity of the sensitive plate, and removed at 

 suitable times during the exposure. The region from wave- 

 length 4000 to 4350 only recjuires about one-tenth of the 

 time demanded by that from 3440 in 3510. In the nega- 

 tive which produced the accompanying plate, the line O 

 had 15 minutes and G i\ minutes, and the former is 

 under-exposed. These exposures seem at first sight un- 

 usually long for a wet collodion surface, but it must be 

 remembered that the slit used was only yJu of an inch 

 wide, and that the diffraction grating gives an almost 

 complete circle of spectra round itself, amongst which 

 this thin band of light is divided. A beam ,]jj of an 

 inch (■00023 metre) wide is spread out in this case into a 

 streak about 7S ft. (2377 metres) long. 



After the production of spectra that were in focus from 

 end to end, it was next necessary to attach a scale to 

 them by which the wave-lengths might be read. At first 

 I tried, by reducing Angstrom's maps to the proper di- 

 mensions, to accomplish this object, but the undertaking 

 proveel to be difficult, and was unsuccessful, because, 

 though the original drawing on the stone was undoubtedly 

 correct, the paper proof of it which 1 had, had stretclied 

 unequally in printing, and on applying a photo- 

 graphic reduction to my spectra, coincidence could 

 not be obtained. As, however, the subject of dividing a 

 scale for these difi'raction spectra is cf prime importance 

 in giving value and precision to the wave-lengths pre- 

 sented in this memoir, I propose to describe fully the 

 method eventually employed in fitting a scale to the 

 photograph. 



The wave-lengths of the ultra-violet rays have never 

 as far as I know, been either determined or published 

 except by J. W. Draper in 1844, Mascart in 1866, and 

 Cornu in 1872. J. W. Draper's memoir has a steel en- 

 graving of some of the principal lines, from which the 

 wave-lengths may be approximatelv read. 



The large plate which accompanies Mascart's long and 

 valuable memoir is of the prismatic spectrum, but he fur- 

 nishes in addition the following table of wave-lengths : — 



38i9'o 

 3728-8 

 3580-2 

 3440 • I 

 3360-2 

 3285-6 

 3177-5 



These numbers do not entirely coincide in all cases 

 with my photograph, as I will show farther on. 



The detailed' results of M. Cornu have not appeared in 

 any publication that has reached me. 



I have used as a basis the numbers given by Angstrom 

 for the rays D.,, b^ and G, and if there should be any 

 sm.all error in his determination, my scale will require a 



proportionate correction, which can easily be effected. At 

 first sight it seemed better to take G and H as fi.xed 

 points, but the line H is so broad, and has so many com- 

 ponent lines, that its position is uncertain, and moreover, 

 being almost at the limit of visibility in Angstrom's appa- 

 ratus, it was more open to errors of measurement. Tht fe 

 reasons led me to take advantage of the fact that the 

 second spectrum overlaps the third, the ray D of the 

 second being near H of the third, and b of the second 

 near O of the third. It is obvious that we have thus the 

 means of ascertaining the wave-lengths of three points, 

 one at each end, and one in the middle of my photograph. 

 As the rays D and b cannot impress themselves on collo- 

 dion by any length of exposure that it is convenient to 

 give, and as in my method of working the ultra-violet 

 rays could not be seen simultaneously with them, it was 

 necessary to resort to the following device : — I placed in 

 front of the sensitive plate and close to it two fine steel 

 points, one of which was carefully adjusted to the position 

 of D.2 of the second order, and the other to b^ of the 

 second order. When, therefore, after a suitable exposure 

 to the ultra-violet spectrum of the third order, the collo- 

 dion picture was developed, there were two sharply-de- 

 fined images of the steel points superposed on the spec- 

 trum. The point which had been coincident with Dj cf 

 the second order was then found to have cast its shadow 

 on Hj of the third order, and the point at b^ of the second 

 order had impressed itself near O of the third order. 



By a simple calculation it was thus rendered evident 

 that a given ray in the compound Hne Hj was of the wave- 

 length 39301 ten miUionths of a millimetre, and that 

 another line near O had the wave-length 3444-6. By 

 looking at the photograph, the reader will see that 3930 

 falls upon a fine division in H,, which is beautifully shown 

 in both the spectrum with the scale and the enlarged 

 proof below. Of course, the ray G of the third order, the 

 wave-length of which is known, had impressed itself pho- 

 tographically on the collodion. 



Having thus ascertained the wave-lengths of three fixed 

 points in the photograph, the next step was to apply a 

 scale reading to a single ten-millionth of amillimetre,and,if 

 possible, fractions thereof. After many abortive attempts 

 to use that pait of Angstrom's map which lies between G 

 and H, and to attach thereto an aclditional length of scale 

 sufficient to extend to the end of the ultra-violet region, 

 1 was compelled to resort to a linear dividing engine, and 

 rule a scale which was about twice the length of the pho- 

 tographic reduction shown in the accompanying plate. 

 Of course this necessitated drawing in by hand the same 

 systems of lines and lettering as are shown on Ang- 

 strom's chart, and this I did as carefully and faithfully as 

 I could. 



It only remained to reduce this divided scale to the 

 proper size to fit the spectrum photograph ; after many 

 trials it was accomplished. 



It is proper in this place to make a criticism on my 

 scale, and to point out a small error, which may be due, 

 however, to an incorrect determin ition of the wave- 

 lengths that I have used as fixed points. Taking the dis- 

 tance from G (wave-length 4307) in the photograph to 

 the fixed line 3930 in H„ and dividing it into 377 parts, 

 and then prolonging these divisions toward O, it was 

 found that the third fixed point was not attained, but that 

 there was an error of about two divisions. But if the 

 position of Do in Angstrom's determinations sho-jld be 

 incorrect to the extent of one ten-millionth of a milli- 

 metre, or if this small error should be partly attributed to 

 Do, and partly to G, my scale would be correct. Future 

 measures of the wave- lengths of these rays, and of b^, can 

 alone settle this delicate point, for the determinations of 

 Mascart and Angstrom and Thalen differ nearly to the ex- 

 tent mentioned above. The same remark is true of Ang- 

 strom compared with Ditschcincr, while the difference 

 between .\ngstrom and van der Willigcn is mire than 



