( 440 ) 



coiistaiil : il li;i<l iuci-cascd ahoiil O, '2 m.iii. diiriiii;- the time iVom 



hofbro lill aflcr llic oltsci-xalioiis. A (Iclcriiiiiialioii oC a pari ot' llic 



leii.üili, wliicli was pcrlnriiKMl hclwccii the oWscrxalioiis, coiiviiuHMl us 



ihal the cliaiiuc liad laken place ui-adiiallv. Tli(> mean \aliio of llie 



len.uili hetore and al'ler llie (>!»sei-\ alions is used as llie lenutli of llie 



scale for all ol»s(M'\"alioiis. 



This \\ ill liax'e l)iil a vei-y small inllnence on llie resull ; it Avill 



cause llié tii'st of the values found for the e(|ui\aleul to he somewhat 



too lari:e, the last to he somevvlial too small. The dilference how (>\-er 



1 

 (k)es not amount to more than ;, ; . The distance from the scales 



1()00<) 



to the fjah'anonietei's was ahout ;^14,'2 cm., Ilial to the hifilar mau'- 



iieionieter about 317,5 cm. 



We ha^'e made 24 determinations of th(M'leclro(diemical e(|uivaleiil 

 of sih'er. The annexed table (p. 44(S) shows the i-esnlts. 



The values tbr the elcH'ti-ocdiemical eipnvalent a deduced from the 



1 



observations with the diUerent L^ahaiiometers diHer less than . 



2U()U() 



As the mean \alue of all delei'ininalions we lind : 



r/ zzzO.OIllSl.S d= ().()()()()( )()4. [},H',i,t i'rn>r]. 



In conneclion with the aii'reement between the ditfereni obsei" 



valions, we are of opinion that this innuher is accui-ale to . 



'J'he obsei"vations will be ])ul)lished later in extenso. 



(January 21, 1904). 



