in Theory and Practice. 381 



Before usinor a solution, an observer should always see 

 Mhat its eifect is by a preliminary experiment. 



Methods of Work. 



A spectroscope is used for two purposes — to measure the 



lines in solar or metallic spectra, or to establish coincidences 



simply. For both of these, the concave grating is far 



superior to any other on account of the overlapping spectra. 



The micrometer-eyepiece, of course, can be used only in 

 the visible spectrum, while the methods of photography give 

 us this and the invisible too. Rowland's micrometer-eye- 

 piece, as noted above, has a run of 5 inches, and so can in- 

 clude a great number of hues. When a metallic spectrum is 

 to be measured, the solar spectrum is turned on, a series of 

 measurements is taken, then the metallic spectrum replaces 

 the solar, another series is taken, then finally a series of solar 

 lines. All this is done in one run of the screw, and without 

 the observer leaving the eyepiece. The solar lines are found 

 on Rowland's map, and then the wave-lengths of the metallic 

 ones are deduced by interpolation. This same method of 

 interpolation will also give the relative wave-lengths of the 

 solar lines, using the overlapping spectra. The probable error 

 of a wave-length determined this way is ±0'01 Angstrom 

 unit. 



Now that we have Rowland's map and his list of solar 

 lines, the photographic process for the measurement of 

 metallic spectra is generally used as far as the erythrosin 

 plates extend or to the D line, although those expert in the 

 use of cyanine plates may photograph below C or even A, as 

 Mr. Burbank has shown in the Phil. Mag. for Oct. 1888. 



Owing to the astigmatism of the grating, it is not possible 

 to adopt the usual method of illuminating part of the slit with 

 the solar image and part with the spark or arc ; and so a 

 different and far better plan is adopted. A compound photo- 

 graph of the two spectra is taken in the following manner : — 

 The brass plate on the back of the camera-box (see fig. XIV.) 

 is placed vertical, the solar spectrum is photographed along 

 the middle of the sensitive plate, the sunlight is turned off, 

 the brass plate is revolved through 90°, and the metallic 

 spectrum is allowed to fall along the upper and lower parts of 

 the photographic plate. Then, finally, the sunlight is turned 

 on again along the middle of the plate. If there has been 

 any gradual displacement of the camera during the operation 

 the error is eliminated by this process, if the two times of 

 exposure to the solar spectrum are the same. 



It is important to notice that record must in all cases be 



