214        M.  E.  Edlund's  Researches  on  the  Electromotive 
by  the  same  process.  This  artifice  was  necessary  on  account  of 
the  proper  movement  of  the  index  independently  of  the  direction 
of  the  current.  As  was  said  above,  in  these  numbers  unity  cor- 
responds to  about  0-002°  Celsius. 
Exp.  2.  Intensity = tan  16°  38'. 
Exp.  3.  Intensity=tanl6°55' 
Deviations. 
Deviations. 
36-9  ] 
30-5    32-10 
30-5  ^ 
39-7  133-48    Mean  32' 79 
24-0  J 
32-41 
30-8  U3-42 
39-7J 
Exp.  4.  Intensity  =  tan  5°  53'. 
Exp.  5.  Intensity  =  tan  5°  35'. 
Deviations. 
Deviations. 
14-9i 
10-2    13-68 
19-4  J 
7-7  113-20     Mean  13-44 
18-0  J 
7-21 
16-2  ^13*33 
13-7J 
Reducing  the  deviation  in  experiment  2  to  the  same  intensity 
of  current  as  in  experiment  3,  and,  in  the  same  manner,  the 
deviation  in  experiment  5  to  the  intensity  of  the  current  in 
exp.  4,  we  obtain  respectively  33*38  and  14-05.  Taking  the 
means  of  these  numbers  and  those  directly  observed  for  the  same 
intensities  of  the  current,  for  tan  16°  55'  the  deviation  33§40  is 
obtained,  and  for  tan  5°  53'  the  deviation  13*74. 
By  calculating  these  observations  in  the  manner  above  indi- 
cated, we  obtain,  as  mean  value,  /3  =  3*1825,  and, 
from  the  observation  with  the  current- 
intensity  =  tan  16°  55', 
«=125-0 
=  tan    5   33,     a  =  135*6 
tan  23  38,     *  =  136*9 
Mean  =132*5 
Several  months  after  the  preceding  experiments,  and  after 
other  metallic  combinations  had  been  investigated,  the  iron- 
copper  combination  was  again  put  into  the  apparatus  and  sub- 
mitted to  fresh  researches  with  the  view  of  ascertaining,  among 
other  things,  whether  during  that  long  time  the  apparatus  had 
undergone  any  alteration.  The  following  results  were  ob- 
tained : — 
