324 EEPOET — 1881. 



eyepiece being at. the same time adjusted by the other hand, till the 

 observer is satisfied that each line is truly coincident with the intersec- 

 tion of the corresponding spider-lines. 



Another device for measuring the interval between two lines, quite 

 equal in accuracy to the bifilar micrometer, is that of the divided-lens 

 micrometer.^ In this instrument the micrometer screw moves one- 

 half of a lens placed just in front of the prisms, and divided along a 

 horizontal diameter. The effect is to cause one-half of the spectrum to 

 move along under the other half, and the sodium or any other conve- 

 nient line is used as a substitute for the cross-wires, and is brought into 

 coincidence with each of the lines to be measured. It will be seen that 

 the necessity of admitting extraneous light to illuminate cross- wires is 

 avoided, and this instrument can therefore be used in faint spectra with 

 precision. 



The photographic method is, of course, a method of simultaneous 

 coincidences, inasmuch as the positions of the known lines which are 

 employed as reference lines are recorded at the same instant as those of 

 the unknown lines. 



The bifilar or the divided-lens micrometer may have fitted to it a 

 device for mapping the spectrum at the same time that the positions of 

 the lines are measured. For this purpose the steel rod on which the 

 screw of the micrometer is cut is made about three times as long, and the 

 extra length has cut on it a much coarser thread. , On this there travels 

 a little brass piece carrying a steel point, with which a trace can be made 

 on a slip of blackened glass. We thus obtain a mark on the blackened 

 strip of glass corresponding to each line of the spectrum. The map so 

 made has the defect of repi'esentiug all lines, whether intense or weak, 

 exactly alike ; but it would be easy to alter it, so as to limit at pleasure 

 'the length of stroke of the tracing point. A bright line would then be 

 denoted by a long trace, and a weak line by a short one. The same 

 instrument might easily be made available for measuring the positions of 

 'the lines in the photograph of a spectrum, since, of course, to take a 

 photograph of a mass of lines in a spectrum is not to have measured the 

 wave-length of these lines, or to have determined their chemical origin. 



Another instrument — very useful in measuring photographed spectra, 

 •or in drawing maps of spectra from measurements — is Beckley's spectro- 

 graph. This consists of a brass cylinder, on which the photograph is 

 stretched, and the edge of the cylinder is graduated and provided Avith a 

 vernier. There is also a straight edge, which can be brought down upon 

 the photograph parallel to the lines of the spectrum. Each line in suc- 

 cession is brought up to the straight edge, and the position of the cylinder 

 is read off by means of the vernier. The instrument is generally gra- 

 duated into degrees and minutes, but it is desirable that it should carry 

 also (on the other edge) a division into millimetres, the vernier reading 

 to the tenth of a millimetre. The accuracy of reading is increased by 

 substituting for the straight edge a small microscope with a 3-inch ob- 

 jective, and with cross-wires in the eyepiece. 



We have already remarked the necessity of reducing the numbers — 

 by whatever instrument obtained — to a uniform scale. 



The scale to be employed must be applicable to all spectroscopes 

 alike, and must be independent of the peculiar construction of the instru- 



' Phil. Mag. August 1875. Proc. Physical Society, vol. i. p. 160. 



