498 Prof. H. F. Weber's Researches on 



at the close of every three hours ; so that in this series of 

 observations 72 data were obtained. 



As was mentioned above, the compensation method was 

 used for measuring the electromotive forces. The compensa- 

 ting Daniell element possessed an almost perfectly constant 

 electromotive force : the ascertained variations of it never 

 exceeded y^y of its mean value. Besides the Daniell element 

 there was another resistance in the first circuit, of 130 mercury 

 units. The resistance of the measuring wire, which was 1 

 metre in length and perfectly homogeneous, amounted to 

 2*221 mercury units. The galvanometer of the second circuit 

 was rendered so sensitive that the rectification could be made 

 upon the measuring-wire exactly to 0*1 millim. Immediately 

 after the preparation of the experiment, X was found to be 

 =976*1 millims. ; twelve hours later it had already fallen to 

 951-6. During the next twelve hours it diminished 80' 1 

 millims., and in twelve hours more almost the same amount. 

 The high degree of delicacy and exactness of the method 

 employed can be perceived from these numbers. A change 

 in the difference of the concentrations of the two boundary 

 layers equal to the 800th part of that alteration of the concen- 

 tration-difference which took place during half a day could be 

 perceived with absolute certainty ; the value of X diminished 

 during the first days from minute to minute about 0*1 millim. ; 

 the progress of the diffusion could thus be as it were 

 tangibly traced from minute to minute. The initial difference 

 of the concentrations of the boundary layers, z 2 — £ 1 = 0*1562, 

 corresponded to X = 976*1 millims. On account of the very 

 close proportionality between the concentration-difference and 

 the electromotive force, an alteration of the concentration- 

 difference amounting to 0*016 milligram corresponds to an 

 alteration of the value of \=0*1 millim. Therefore, if the 

 lower boundary layer lost by diffusion only ^} 2 q milligram of salt 

 per cubic centimetre, and at the same time the upper boundary 

 layer gained the same amount, this effect of diffusion could still 

 be distinctly perceived. 



The following Table gives all the particulars of the fourth 

 series of observations. In the first column stands the date ; 

 the second contains the temperature of the apartment at the 

 time of each observation ; the third gives the rectifications X 

 upon the measuring-wire ; the last column contains, from the 

 second day onwards, the differences of the ordinary logarithms 

 of each two succesive values of X. 



