28: 



A' A JURE 



\_Jai!. 20, 1 88 1 



Examhiation of Lines in the Solar Spectrum which are given in 

 the Maps as common to Twi or more Substances. — For this purpose 

 a spectroscope of high dispersion has been constructed by com- 

 bining the grating mentioned above, which has about 4 square 

 inches of ruled surface, with a collimator and observing telescope 

 each of 3 inches aperture and about 42 inches focal length, 

 using magnifying powers ranging from 50 to 300. The apparatus 

 is arranged upon a wooden fnime-work, and when in use is 

 strapped to the tube of the 12-feet equatorial of our observatory, 

 so that it is kept by the driving-clock directed to the sun. An 

 image of the sun is farmed on the slit by an achromatic object- 

 glass of 3 inches aperture, in order to increase the light and 

 to avoid the widening of the lines due to the sun's rotation. \ 

 large pri^m of about 20° angle was sometimes placed in front of 

 this object-glass (between it and the sun) to separate the colours 

 before reaching the slit ; and in examining the darker portions 

 of the spectrum a concave cylindrical lens was sometimes used 

 next the eye, like a i-hade glass, to reduce the apparent width of 

 the spectrum and thus increase its brightness. 



The grating is an admirable one, on the whole the best I have 

 ever seen. But I have been greatly surprised at its excessive 

 sensitiveness to distortion by pressure or inequahties of tempera- 

 ture. Although the plate is fully § of an inch thick, and only 

 3^ inches square, an abnormal pressure of less than a single 

 ounce at one corner will materially modify its behaviour, and 

 a quarter of a pound destroys the definition entirely. In fact 

 the plate is not naturally exactly flat, and to get its best perforiu- 

 ance it is necessary to crowd a little wedge gently under one 

 corner. When it is in good humour and condition, however, 

 the performance is admirable ; one could wish for nothing better, 

 unless for a little more light in the violet portions of the 

 spectrum. 



With this instrument I have examined the 70 lines given on 

 Angstrom's map as common to two or more substances. Of the 

 70 lines, 56 are distinctly double or triple ; 7 appear to be 

 single ; and as to the remaining 7, I am uncertain ; in mist 

 cases, because I was unable to identify the lines satisfactorily on 

 account of their falling upon spaces thickly covered with groups 

 of fine lines, none of which are specially prominent. 



As a general rule the double lines are pretty close, the dis- 

 tance being less than that of the components of the 1474 line. 

 Generally also the components are unequal in width ordarkne-s. 

 or both, though in perhaps a quarter of the cases they are alike 

 in appearance. The doubtful lines are the following, designated 

 by their wave length on Angstrom's map: 5489"2, 5425 'o, 

 5396'!, 5265 '8, 427 1 '5, 4253^9 and 4226"S. I strongly suspect 

 5396'! and 5265'8 (which present no diiihculty in identification) 

 of being double, but could never fairly split either of them, and 

 therefore leave them among the doubtfuls. 



Those which show no signs of doubling, so far as could be 

 seen, were: 6121-2, 6064-5, 5019'4> 4585"3i 4S78'3, 4249'S, 

 and 4237 'S- 



In respect to the lines 5019-4, 4585-3 and 4237-5 it is quite 

 possible there may be some mistake as to the coincidence, since 

 in his tables Thalen gives neither of them as due to iron. An 

 accidental strengthening of the dotted line, which, on the map. 

 leads up from the symbol of the element concerned, through the 

 iron spectrum, would account for the matter, by making the 

 line appear on the map as belonging to iron also. 



As the facts stand, therefore, it is obvious that arguments 

 which have been based upon the coincidence of lines in the 

 spectra of different elements lose much of their force ; it ap- 

 pears likely that the coincidences are in all cases only near 

 approximations. At the same time this is certainly not yet 

 demonstrated. The complete investigation of the matter requires 

 that the bright line spectra of the metals in question should be 

 confronted with each other and with the solar spectrum under 

 enormous dispersive po« er, in order that we may be able to 

 determine which of the components of each double hue belongs 

 to one, and which to the other element. If in this research it 

 should be found that both of the components of a double line 

 w ere represented in the spectra of two different metals, and the 

 suspicion of impurity were excluded, we should then indeed 

 have a most powerful argument in favour of some identity of 

 material or architecture in the molecules of the two substances 

 involved. 



Distortion of Solar Prominences by a Diffraction Spectroscope. — 

 Generally, in such an instrument, the forms seen through the 

 opened slit are either disproportionately extended, or compressed 

 along the line of dispersion. The reason is this : if the slit be 



illuminated by monochromatic light, the image of the slit, 

 formed on each side of the simple reflected image in the focus 

 of the view-telescope (which is supposed to have the same focal 

 length as the collimator), will have the same width as the slit 

 itself only in one special case, not usually realised with a 

 reflecting grating. 



If the angle, between the normal to the grating andj^the 

 view -telescope, is less than that between the normal and the 

 collimator, the slit image will be narro~,.oer \h.^n the slit, and a 

 prominence seen through it will be compressed in the plane of 

 dispersion. If the relation of the angles be reversed, then of 

 cour.'e the distortion will also be reversed, and we shall have 

 extension instead of compression. 



The mathematical theory is very simple. Suppose the colli- 

 mator and telescope to be fixed at a constant angle, as in the 

 now usual arrangement. 



Let an_le between telescope and collimator — a. 



Angle between tele-cope and normal to grating =■ r. 

 Then angle between collimator and normal = « = a - t. 

 Also, let space between adjacent lines of grating = s. 

 And the order of spectrum observed ~ ;/. 



Then, by principles of spectrum formation, we have 



I 



(' . , 



A being the wave-length of the ray which is in the centre of the 



field of view : 



whence sin t = -f sin k. 



s 



Difl^erentialing, we have at once 



sm T - sm I 



dT-- 



Uk; 



■ die, or 



cos T COS T 



+ sin o tan t) dK. 



which reduces to, dr = (cos o + sin o tan t) dK. Distortion can 

 only disappear in cases when this coefficient of dK reduces to 

 unity. Special cases — 



1. If T — K there is no distortion — but also no dispersion : 

 it is the case of simple reflection. 



2. If K = o, the grating being kept normal to the collimator, 

 then dr = sec a dK. 



3. If T =0, the grating being kept normal to the telescope 

 (which in this case must be movable), then dr = cos o dK. 



4. If = 90°, i/r^tan t dK. 



5. If a = o, dT = dK, and there is no distortion. 



This is possible only by u-ing the same tube and object-glass 

 both for collimator and view -telescope, the grating being slightly 

 inclined at right angles to the plane of dispersion. The prin- 

 cipal difficulty in this form of instrument lies in the diffuse 

 light reflected by the surfaces of the object-glass. It is hoped 

 that this may be nearly obviated by a special construction of the 

 lens which will throw the reflected light outside of the eyepiece. 

 An instrument on this plan is being made for Prof. Brackett 

 by the Clarks, for use in the physical laboratory at Princeton, 

 and is now nearly completed. 



Princeton, September 27, 1880 C. A. Young 



UNIVERSITY AND EDUCATIONAL 

 INTELLIGENCE 



Dr. J. E. Harris (D.Sc. Lond.) has been appointed to the 

 vacant Professorship of Natural Philosophy at Trinity College, 

 London. 



From the new Calendar of the University College of Wales 

 we learn that the present number of students is fifty-seven. We 

 see there are classes for most of the branches of science, only 

 unfr rtunately they are all taught by one professor, which, to say 

 the least, must be rather hard on him. We hope the college 

 will soon be able to have separate teachers, at any rate for the 

 physical and biological sciences. 



The new University Library at Halle has just been opened. 

 It is built entirely on the French system, and special precautions 

 have been taken with regard to fire. It now numbers some 

 200,000 volumes, but there is room for half a million. The cost 

 of the building amounts to 400,000 marks (20,000/.). 



SCIENTIFIC SERIALS 



The American Naturalist for December, 1880, contains: — 



D. Cope, on the extinct cats of America. — F. V. Hayden, Twin 



