ON OUR KNOWLEDGE OF SPECTRUM ANALYSIS. 323 



then be adjusted till some lines in the green — say h in the solar spectram — 

 are in accurate focus. The instrument is then in adjustment.' When 

 used on the red or blue portion of the spectrum, the focus may be adjusted 

 with the observing telescope, but the collimator should not be altered. 



It is necessary that the ray to be measured should be in exact focus 

 together with the cross-wires. If this is not the case, the ray will alter 

 its position slightly with reference to the cross-wires, if the eye be 

 slightly moved. The adjustment may therefore be tested by moving the eye 

 slightly and observing whether the ray and the cross- wires move together. 

 There is also a slight movement of the rays consequent on lateral shifting 

 of the source of light ; this is less the narrower the slit is, and the more 

 distant the source of light is. 



Some instruments are provided with a tangent-sci-ew micrometer, — • 

 that is, a long screw, the head of which is divided into a hundred 

 equal parts, by means of which a slow motion can be given to the observ- 

 ing telescope, and the number of turns of the screws, and parts of a turn 

 necessary to carry the cross-wires from one line to another, is noted. 



In the bright-line micrometer^ the image of a fine slit in a brass 

 plate is seen by reflexion at the first surface of the prism, and so is super- 

 posed upon the spectrum ; the plate and slit have a slow motion given by 

 a micrometer screw. This form of micrometer is specially useful with 

 very faint spectra, when cross-wires would be useless. In observing 

 with cross-wires a luminous spectrum the lines of which are faint, it is 

 necessary to admit a certain amount of light into the observing telescope, 

 sufficient to illuminate the wires (conveniently by raising an edge of the 

 cloth used to cover up the prisms). This general light renders very faint 

 lines invisible. In all these methods of consecutive coincidences it is neces- 

 sary that no shifting of the parts of the instrument by bending or shaking, 

 nor any disturbance of the position of the source of light, nor of the exact 

 position of the eye, should take place during the passage of the cross- 

 wires from one line to the next. In the methods of ' simultaneous coin- 

 cidences ' all these sources of error are avoided by observing at the same 

 instant two lines — one a known line, used as a reference line, and the 

 other the line to be measured. 



The method of the reflected photographed scale, already described at 

 some length, may be employed as a method of simultaneous coincidences, 

 and so made more exact if, when the reading of any line is noted, care 

 be taken to observe that the sodium-line is still exactly at 50 ; or if the 

 sodium-line is not in the field, then that some other line used as reference 

 line is exactly in its right position at the moment of observation. 



The most accurate measuring instrument for use with large spectro- 

 sco])es is the bifilar micrometer eyepiece. This is an eyepiece similar 

 to those employed for astronomical purposes, provided with two crosses of 

 fine spider-lines in the focus of the eyepiece, which must therefore be of 

 the Ramsden construction. One of these cross-wires remains fixed ; the 

 other is moved by means of a micrometer screw. The interval between 

 the line to be measured and a line of known wave-length can thus be 

 determined with great precision. In taking an observation, a slight 

 motion is given to the fixed cross-wires by means of the slow motion or 

 tangent screw of the observing telescope, the micrometer screw of the 



> For a different method of adjiisting the collimator of a spectroscope, see a paper 

 by Dr. Schuster, Phil. Mag. [5] vii. 95. 

 ■ Microscopical Journal, January 1870. 



