218        M.  E.  Edlund's  Researches  on  the  Electromotive 
Exp.  23.  Intensity  =  tan  25°  45'. 
Deviations. 
5-70 
8-12 
Mean     .     .     6*91 
The  calculation  of  these  three  experiments  gives  /3=0  and 
«  =  14*5.  The  circumstance  that  j3  is  here  =0,  and  conse- 
quently the  deviations  are  proportional  to  the  intensity  of  the 
current,  proceeds  doubtless  from  the  very  slight  elevation  of  the 
temperature  within  the  cylinders,  this  resulting  from  the  mini- 
mum development  of  heat  in  the  two  wires,  which  are  good  con- 
ductors. The  cooling  of  the  cylinders  becomes  then  proportional 
to  the  excess  of  temperature  of  their  sides,  and  this  excess,  in  its 
turn,  proportional  to  the  excess  of  temperature  of  the  enclosed  air. 
If  we  calculate  the  three  experiments  supposing  the  propor- 
tionality of  the  deviations  to  the  intensity  of  the  current,  we  ob- 
tain the  following  comparison  between  the  observed  and  the 
calculated  numbers : — 
Observed. 
Calculated 
6-91 
6-99 
10-89 
10-80 
12-49 
12-49 
Another  series  of  observations,  made  with  the  unsilvered  cy- 
linders before  mentioned,  gave,  for  the  copper-gold  combination, 
a  =  12*56.  The  series  effected  with  the  same  cylinders  upon 
the  iron-copper  combination  gave  a  =  115" 73.  Both  these  values 
are  inferior  to  those  before  indicated,  viz.  14*5  and  132*73. 
Dividing  these  by  the  preceding,  we  obtain 
14*50     ,  .„.        ,  132-73     -4AK. 
1^  =  1-164  and  __  =1-147, 
quotients  which  may  be  regarded  as  equal.     These  four  series, 
therefore,  confirm  one  another. 
The  following  experiments,  on  the  same  copper-gold  combi- 
nation, were  made  with  15 -minute  intervals  between  the  obser- 
vations. 
Exp.  24.  Intensity  =  tan  33°. 
Exp.25.  Intensity  =  tan  43°  15' 
Deviations. 
Deviations. 
6*33 
9*20 
7*90 
10*33 
6*98 
9-38 
7*58 
9-45 
6-70 
9-70 
Mean  .  *  MO 
Mean  .  .  9*61 
