17 



account of liis results tliat I kiiuw nr. wliicli !■< tu Im- lunml in tlic Andiiidn 

 Joiinial of 'Scti'iici; tVir July INT'.I. I iil'urtnnatfly, tlii; il;il;i ^ivi'ii tlii'ic iuc 

 iiut snHicii.Mit t'> ciKiljle iiiu tn (IcttTiniiio with ;ili.si)liite certainty tin; liim set ii|ii)ii. 

 Mr. I V'ircc luadi.' two series ut'iihservatioiis; in the hist he measured tlie ile\ iatinn ,,{' 

 iis|H'ctral lini'. "van tier WilÜLCeu's No Hi,' "iisiri^ a certain fitter of o40ö liiu;s to 

 tlie milliiiiefer'. 1 liave not been able to refer directly to van der Wiiligen's meas- 

 urenieiitstoat;i:ertrtin theinisition ol'liis "Nolo" but as l\[r. IVirco's «uniting space 

 was double mine I have endeavored to locate it from his measures. I have a-ssmueil 

 that the work \viu> done in the spectrum of the i'ourth order. Althouj;h nothiuL; is 

 Siiid ut' the order of the si>'jctrum in the article referred to, it is clear thai no other 

 snp|)ositioii would be admissible. This line should, therefore, have the .same <levia- 

 liiin in t lie spectrum of the second order I'f the L^ratiuL; used iu these uieasiu'enieiifs, 

 Mr.-l'eirce says iiothinj;' about the temperature at which his nieasurcmenls were 

 made so that there is room for some variation on this account. 



Upon exainining the sjiectrum at this particular deviation there ap[ieare(| 

 to \n: no line at tliat point which seemed to uio to possess sutHcient distinct ne.ss 

 und |iromiuence to have been selected for this important duly. A well deliued 

 and sutliciently jirominent line very close to the indicated point, a little nearer i,, 

 seemed to Ikj most ])robubly the line upon wliicli the settings had been made. 

 This line was accordingly selected and a considerable series of careful measiue- 

 ments were made upon it. Assuming this to be the same line used by Air. 

 I'eirce, the dillereuce is not greater than cui easily be accouuterl for. Jlis 

 result is 



/, = 44-' 57' '.V.T.") 



and my own, the mean of Ii4 mcivsures whieh will In- found in detail in IIm.' tol- 

 Ifiwing pages, was; — 



/, = 44' -,r,' rnr.-i-) 



the ilillerelice being I'.l .4 



Mr. I'eirce does not stiite the tenii« latnre towhirh his results wi're redureil. 

 With my own grating the increment of arc lor this line was 'X'.H'.i for \^C, so 

 that it is clear that there is a |iossible variation in temperature easily suflieieid 

 to account for the «lilference iu the angular measiu'emeiit-s, esp(;fially when it is 

 rememl)Cred that the gratings may have been ruled atdill'ereut temiieratiues and 

 that tliev are composed of substances widely dill'ering in their co-ellicients of 

 expatision. 



The second series of measurements maile by Mr. I'eiree was upon another 

 line near the first an<l wilh "another nmrh liner gitter. " As lliis is a region 

 crowded with lines and as tlie fineness of the grating, was not delinilely slated, 

 it was impossible for me to decide certainly as to the line made u.se of, s(j it was 

 detennined to rely n|K)n tlie first results, wiiich are undoubtedly of sulHcieut 

 jiccumcy, assuming that the right line lias been fnund. 



