( 445 ) 



from the deviations (if ihe iiinu-iicloiiieler-necdles caused hv the lunfi'iiet 

 when placed luiniial lo llic iiiauiielic iiiei-idiaii in Iwu din'creiil 

 ]n)sili(tiis sviimictrical willi rcspcel lo lliesc iici'dics. Two dcler- 

 iiiiiialioiis ill w hieli iIm' dislaiices were cIioscmi: <S() uihI 100 cm., and 

 75 and 105 cm. \ icidcd tlic \aliies i;),40 and {'.\,TA cm. The dilfe- 



.1 

 rence ina\ he (hie io expei'inieiilal en-ors. A Nariation of in.ni. 



of one of die (hslances. causes — ceteris parihiis - a- \ai'iation in the 

 \ahie of die [xtiar (hsianee lar,uer than the (hUerence h(Mween the 

 two \aliies foiiiMh We haxc lakeji foi- die polar dislancc the mean 

 \alne I'A^W cm. This aurcf's fairlv well w idi '/,. of the lenu'lh. 



In order to ascertain die dislancc hctwcen the cenlei's of the 

 needles, \\-c lirsl delermiiicil lli<> dirierence of their distanee from 

 the cocoon silk snspcnsion-lihres. This dilference was deri\ed from the 

 deviations of the needles caused Im" die nuiunel, as an ell in one of die 

 posidons of die maiiiielometers as a\ licji thev had ehan^ed place, care 

 heinu' taken that the silk lihi'es had the same positions both times. In two 

 ohserAatioiis this dilference amounted lo 0,007 c.m. and 0,001) cm.; 

 average \aliie O.OOS cm. A ditference of distanee of 0,001 cm. has 



an iidlnenee on die resiill of 



1 



120000 



The distance of the silk fibres was measured hv [trojectini^' the 

 fibres from two telescopes, i<SO cm. apart and at a distanee of about 

 5 m. from the magnetometers, on a horizontal scale placed behind 

 them. 



It apptNired that the walls of the tube of the magnetometer lia<I 

 ]io intlneiice on the conrse (»f the ravs. 



We delermined Ihe \alne of // before and after the lime of the pas- 

 sage of the cnrrent, in order to arrixe at the mean \alne of // (hiring 

 that time. During all this time the indications of the local-xario- 

 nieler of Koiii.Jx'.vrscn, erected in a room \\ itli nearly constant tempe- 

 ratnre, were read, during the determinations of // e\ery 2 minntes, 

 in the time bet\veeii those determinations e\ery 5 minntes. 



The \aliie of // dni'ing the |)assage of the current was derived 

 from Ihe constants of the \ariomeler, the mean iH'ading diiiMng the 

 first delei'nnnation of // and during the [>assing of the cnrrejit and 

 the valne of // found in the lii'sl delermination of //. 'fhe same 

 calculation Nvas performed w itli the second delermination of //. The 

 two \aliies found in this \n ay for // during the |)assage of the 



ciii-rejit dilfered in most cases onh' or not e\"en so iiiiich. Oiih 



• 9000 



