3^2 



NATURE 



[January i i . 191 2 



in'CVSr PROGRESS IN SPFCTROSCOPIC I 

 METHODS.' 



A 



. , M.N.I w, vvIk) for lhi» first tiiii= ilu lijjht of 



ihp »«un through a prism cannot fail to fxprfss his 

 \vond«T and delight nl the gorgrous display of colours into 

 -. hich the white light is separated ; and if the observation 

 - made under the same conditions as in the celebrated 

 \periment of Newton, iWWi, there is, in truth, nothing 

 Isc which he could observe. You will rememb<'r that he 

 allowed a beam of sunlight to stream through a round 

 opening in » shutter of his window, falling on a glass 

 prism, which b»nt the sun-rays by diflen-nt amounts 

 depending on their colour, thus spreading out the white 

 round sunlit spot on the opposite wall into a coloured band 

 — the spectrum — which he rather arbitrarily divided into 

 seven colours — red, orange, yellow, green, blue, indigo, Jind 

 violet. (If the division were made to-day I doubt if indigo 

 would be included.) There is, in fact, no definite demarca- 

 tion between these, and they .shade insensibly into each 

 other, and if the solar spectrum were always produced 

 under these conditions we should say it was continuous ; 

 indef-d, if it were not the sun, but an argand burner or 

 an incandescent lamp, which served as source, it would 

 really be so. 



But even if the source consisted of isolated (but suffici- 

 ently numerous) separate colours, the fact would be dis- 

 guised by the overlapping of the successive images. In 

 other words, the spectrum is not pure. In order to prevent 

 this overlapping, two important modifications must be made 

 in Newton's arrangement. First, the light must be 

 allowed to pass through a very narrow aperture, and, 

 secondly, a sharp image of this aperture must be formed 

 by a lens or mirror. 



The first improvement was introduced by Wollaston in 

 1802, who writes : — " If a beam of daylight be admitted 

 into a dark rt)om by a crevice one-twentieth of an inch 

 broad, and received by the eye at a distance of lo or 12 feet 

 through a prism of flint glass held near the eye, the beam 

 is seen to be separated into the four colours only — red, 

 yellowish-green, blue, and violet. . . . The line that bounds 

 the red side of the spectrum is somewhat confused. . . . 

 The line between the red and green ... is perfectly dis- 

 tinct ; so also are the two limits of the violet. There are 

 other distinct lines (in the green and blue . . . )." 



The second improvement was effected by Fraunhofer, 

 1814, and by observing the light which fell from such a 

 narrow aperture upon a prism by means of a telescope he 

 discovered upwards of 750 dark lines in the solar spectrum, 

 and mapped their position and general character. 



In recognition of the enormous importance of this dis- 

 covery, these lines are always known as the Fraunhofer 

 lines. 



A minor inconvenience in Fraunhofer's arrangement lay 

 in the fact that the slit source had to be at a considerable 

 distance from the telescope ; and this was obviated in the 

 apparatus of Bunsen and KirchhofT, i860, which is essenti- 

 ally the same as the modern spectroscope of to-day, con- 

 sisting of a slit and collimator, prism, and observing (or 

 photographic) telescope. 



On this beautifully simple device rests practically the 

 whole science of spectroscopy, with all its wonderful appli- 

 cations and all the astonishing revelations of the structure 

 and motions of the sidereal universe and of the constitution 

 of the atoms of matter of which it consists — nay, even of 

 the electrons of which these atoms are built ! 



Without the telescope it is evident that the science of 

 ■spectroscopy would be as limited in its field as was the 

 science of astronomy without the telescope. It is interest- 

 ing, indeed, to compare the progress of the two sciences 

 as dependent on the successive improvements in the two 

 instruments. 



Without the telescope nothing could be discovered con- 

 cerning the heavenly bodies (with the exception of a few 

 of the more evident features of the sun, the moon, and the 

 comets) except the brightness and places of the stars and 

 the motion of the planets, and even these could, at best, 

 be very roughly determined (say, to within one part in five 

 thousand, or something over a half-minute of arc). With- 



• Address of Dr. A. A. Michelson, retiring president of the American 

 Afisociation for the Advancement of Pcience, delivered at the Washington 

 meeting of the .\ssociation on December 37, 191 1. 



out the telescope *p*>rfrovr»py would al«o have b^en limited 

 to ob«ervationii > ' ' ~ of radia- 



tions and abfor 1 of the 



position of the -j -. ...." .. ,,....,...... .;iur of this 



sam** order of magnitud*-. 



In fact, the resolving power of the eye is Rieasured by 

 the number of light waves in its diameter, about 5000, and 

 if a double star (or a double spectral line) presents a 

 smaller angle than 1/5000 it is not ** resolved." The 

 resolving power of a telescope with a i-inch objective 

 would be alxjut 100,000, so that details of the solar and 

 lunar surfaces, and of planets, nebula:, and of double stars 

 and star groups can be distinguished the angular distance 

 of which is of the order of 1/100,000. The discs of the 

 planets, the rings of Saturn, the moons of Jupit*^, and 

 some star groups and clusters, begin to be di ible. 



Our largest telescopes have a resolving pow- as 



2,000,000, corresponding to a limit of scparauoti <>i one- 

 tenth of a second. 



But in order to realise the full benefit of the telescope 

 when used with a prism, the latter must be so large that 

 the light which falls upon it entirely fills the object glass. 

 The efficiency of the prism then depends on its size and on 

 its dispersive power. 



In order to form an idea of the separating or resolving 

 power in spectroscopic observations it will be convenient 

 to consider the Fraunhofer line D of the solar spectrum 

 or the brilliant yellow line corresponding to the radiation 

 given out by a salted alcohol flame. This Fraunhofer 

 recognised as a double line, and the length of the light- 

 waves of the components are approximately 0-0005890 mm. 

 and 0-0005896 mm. respectively. The difference is, then, 

 6/5893 of the whole, or about i/iooo, requiring a prism of 

 resolving power of 1000 to separate them. If the prism 

 were made of flint glass with a base of 25 mm. it would 

 just suffice to show that the line was double. 



Now we know of groups of spectral lines the com- 

 ponents of which are much closer than those of sodium. 

 For instance, the green radiation emitted by incandescent 

 mercury vapour consists of at least six components, some 

 of which are only a hundredth of this distance apart, and 

 requiring, therefore, a resolving power of 100,000 to 

 separate them. This means a glass prism of 100 inches, 

 the construction of which would present formidable difficul- 

 ties. These may be partially obviated by using twenty 

 prisms of 5 inches each ; but owing to optical imperfections 

 of surfaces and of the glass, as well as the necessary loss 

 of light by the twenty transmissions and forty reflections, 

 such a high resolving |>ower has not yet been realised. 



The parallelism of the problems which are attacked in 

 astronomy and in spectroscopy is illustrated in the. follow- 

 ing table. It is interesting to observe how intimately 

 these are connected and how their solution depends on 

 almost exactly the same kind of improvement in the observ- 

 ing instruments, particularly on their resolving power ; so 

 that not only are the older problems facilitated and their 

 solution correspondingly accurate, but new problems, before 

 thought to be utterly beyond reach, are now the subject of 

 daily investigation. 



Astronomical. Spectroscopic, 



(i) Discovery of new stars. Discovery of 



nebulae, and comets. 

 (2) Star positions. 



new 

 elements. 



Wave-length of spec- 

 tral lines. 



(3) Double stars and star Double lines, groups, 

 clusters. and bands. 



(4) Shape and size of Distribution of light in 

 planets and nebulje. spectral " lines." 



(?) Star discs. 



(5) Star motions (normal Star motions (parallel 

 to line of sight). with line of sight). 



Resolution of doubles. Resolution of doubles. 



Solar vortices. Solar vortices. 



Protuberances, &-c. Protuberances, &c. 



(6) Changes of character 



and position of lines with 

 temperature, pressure, and 

 magnetic field. 



(7) Spectroheliograph. 

 (Combination of telescope and spectroscope.) 



XO. 2202, VOL. 88] 



