348         Dr.  G-undry  on  the  Asymmetrical  Action  of  an 
frequency  the    greater  is  the  variation   of  — -9    with  i0,  and 
Table  VIII.  shows  that  by  making  the  current  smaller,  the 
variation  is  greatly  reduced.  Were  it  possible  to  work  with 
still  smaller  currents,  it  may  well  be   that  a  much  nearer 
approach  to  constancy  of  7-5  at  these  low  frequencies  might 
be  found.  l° 
Finally,  Table  IX.  shows  a  test  for  an  after-effect.  A 
comparatively  strong  alternating  current  (3'05  x  10~ i  amp.) 
was  run  for  half  an  hour  through  the  cell,  and  readings  of 
the  galvanometer  deflexion  were  made  every  3  minutes.  The 
alternating  current  was  then  shunted  out  and  readings  taken 
every  \  minute.     There  was  40,000  ohms  in  the  circuit. 
Table  IX. 
Frequency  680.     Altera,  current  3*05  x  10  ~4  amp. 
Time, 
rain. 
0  
3  
6  
9  
12  
15  
18  
Readings  of  Galvanometer. 
Deflexion. 
Without 
Alt.  Cur. 
With 
Alt.  Cur. 
677-3 


...... 
;;;■" 

662-5 
672-4 
675-0 
675  6 
676-1 
676-4 
675-2 
475-5 
485-5 
478-0 
482-9 
4887 
497-6 
507-0 
507-1 
513-9 
514-7 

201-8 
181-8 
1993 
194-4 
188-6 
179-7 
170-3 
170-2 
163-4 
162-6 
14-8 
4-9 
2-3 
1-7 
1-2 
0-9 
2-1 
21  
24  
27  
30  
30i 
3Gi 
30J 
31   
m 
31* 
31* 
Even  after  this  long   subjection   to  a   strong  alternating- 
current,  very  slight  after-effect  is  to  be  observed. 
By  using  a  capillary  of  wider  bore  I   found   the  value 
of  ^2  ^°  varJ  inversely  as  the  fourth  power  of  the  diameter^ 
which  is  also  in  agreement  with  the  theoretical  result. 
