82 Dr. W. M. Watts on the Spectrum 



mainly due to manganese oxide (not, as has been stated, to man- 

 ganese , no single line of whose spectrum coincides with a Bes- 

 semer line) . These observations were made at the works of the 

 Barrow Hsematite Steel Company at Barrow ; and I am glad to 

 have this opportunity of expressing my sense of the liberality 

 with which the Directors of the Company assisted a purely sci- 

 entific investigation, not only by affording every facility for the 

 carrying out of experiments, but even by contributing to the 

 expense of the investigation. 



The instrument employed was Browning's automatic spectro- 

 scope of six prisms. Most of the measurements were made by 

 means of the micrometer tangent-screw of the instrument, which 

 requires 2 - 94 turns to carry the cross-wires of the telescope from 

 the lithium- o?*ange line (wave-length 6101) to the least-refran- 

 gible D line (wave-length 5895). In some of the later measure- 

 ments much more exact results were obtained by the use of a 

 micrometer eyepiece furnished with two pairs of cross-wires. 

 With this instrument the interval between the same two lines 

 is represented by 12*49 turns of the micrometer- screw. The 

 spectrum was mapped throughout on the scale of wave-lengths, 

 the wave-lengths of the lines of the Bessemer-spectrum being 

 obtained by interpolation from the wave-lengths of the known 

 lines of some metal, whose spectrum was arranged so as to be 

 visible together with the Bessemer-spectrum. 



It was found, however, that (in consequence, probably, of the 

 complicated motion of the prisms in the automatic arrangement) 

 the observing-telescope did not always travel at exactly the same 

 rate, so that the interval between two given lines was not always 

 represented by the same reading. For example, in twelve suc- 

 cessive measurements of the interval between the most refran- 

 gible D line and the least refrangible b line the following num- 

 bers were obtained :— 



14-11, 13-71, 13-73, 13-68, 13-69, 13-77, 13-77, 

 13-74, 13-77, 13-78, 13-75, 13-78. 



But although the rate of motion varies, yet in any one passage 

 of the telescope (and prisms) from the red end to the blue end 

 of the spectrum the motion is uniform, so that the wave-length of 

 any line in the Bessemer-spectrum is correctly obtained by inter- 

 polation from the wave-lengths of two known lines between which 

 it lies. Accordingly in ea< h " blow " the spectrum was mapped 

 steadily from the red end to the blue end, the tangent-screw 

 being kept firmly clanped and the telescope never being allowed 

 to return upon itself. The readings obtained in each blow were 

 then reduced to wave-lengths by means of one interpolation- 



