The Astrophysical journal . taken away, and the smallscreen turned in such a mannerthat the radiation falls on bothstri])s, and then the jjosition ofrest of the galvanometer can bedetermined. The screen is thenturned toward one side, and atthe same time the current is closed through the shaded stripand regulated in such a manner that the galvanometer takes thesame position of rest as before. After the strength of thecurrent is read off, the screen and commutator are reversed, andthen the determination repeated. In this way in a few minutesmany pairs of determinations can be obtained. The ther-
The Astrophysical journal . taken away, and the smallscreen turned in such a mannerthat the radiation falls on bothstri])s, and then the jjosition ofrest of the galvanometer can bedetermined. The screen is thenturned toward one side, and atthe same time the current is closed through the shaded stripand regulated in such a manner that the galvanometer takes thesame position of rest as before. After the strength of thecurrent is read off, the screen and commutator are reversed, andthen the determination repeated. In this way in a few minutesmany pairs of determinations can be obtained. The ther-mometer projecting into the instrument tube gives the tempera-ture of the air close to the strips. 5. The errors in the resulting determination with this instru-ment depend first, on the accuracy with which the constants ofthe instrument are determined, and, second, on the accuracywith which the settings for the same temperature and the obser-vations of the strength of the current can be carried out, Since(iQ dr lib da 2 di. Fig. 6, ELECTRIC COM PENS A T/OX P1 RHELIOME TER 3 4 I we find, in agreement with the above, that the error which willbe caused through dr, db, and d(7, will amount at most to percent. The error in di dej^ends, naturall}, in a great degree, onthe special conditions under which the measures are taken. Thestrength of the current can be determined without difficulty, towithin per cent. The error in Q resulting from di will, there-fore, at most, not exceed j)er cent. The total error of asingle determination of Q will, therefore, amount at most toabout 2 per cent., of which about per cent, is constant, per cent, is accidental error. 6. As proof of the reliability of the method, the followingobservations may be given, viz.: first, the determination of theradiation of a lamp with two instruments, one of which I mademore than four years ago, the other having been lately con-structed ; second, the comparison of the observations in solarradiati
Size: 1428px × 1750px
Photo credit: © The Reading Room / Alamy / Afripics
License: Licensed
Model Released: No
Keywords: ., bookcentury1800, bookdecade1890, booksubjectspectru, bookyear1895
