Theory  of  Magnetism.  417 
raent  of  the  atoms  or  molecules  of  the  conducting  substance. 
On  these  points  it  will  now  be  proper  to  make  some  remarks. 
Galvanic  currents  may  be  conceived  to  be  generated  in  the 
same  manner  as  magnetic,  so  far  as  regards  the  condition  of  a 
gradation  of  atomic  density,  the  gradation  being  in  their  case 
produced  and  maintained  by  chemical  action  in  such  manner  as 
to  exist  permanently  in  the  neighbourhood  of  the  surfaces  of 
contact  of  the  substances  between  which  the  action  takes  place. 
The  currents  thus  generated  must,  from  what  has  already  been 
argued,  fulfil  the  condition  of  flowing  in  a  complete  circuit  in 
order  that  they  may  be  permanent  \  for  which  reason  it  is  neces- 
sary to  connect  the  poles  of  a  galvanic  battery  by  some  material 
(as  copper  wire)  capable,  in  respect  both  to  form  and  quality,  of 
conducting  galvanic  currents.  But  from  the  foregoing  mathe- 
matical argument  it  may  be  inferred  that  currents  so  conducted 
can  proceed  only  in  spiral  courses.  Now,  supposing  the  course 
to  be  of  this  kind,  it  appears  from  experiment  that  the  turn  of 
the  spiral  is  always  in  the  same  direction  relative  to  the  course 
.of  the  current  along  the  conductor.  Hence  it  would  seem  that 
such  courses  are  impressed  by  the  wire  itself  or  other  conducting 
substance,  because,  as  far  as  regards  hydrodynamical  conditions, 
the  course  might  either  be  dextrorsum  or  sinistrorsum.  Suppo- 
sing the  mean  direction  of  the  current  to  be  in  the  positive  di- 
rection of  the  ordinates  z,  if  the  arbitrary  constant  b  in  the  fore- 
going equation  (k)  be  positive,  the  value  of  z  is  greater  as  the 
arc  tan-1-  is  less,  and  therefore  the  spiral  course  is  dextrorsum-, 
x 
that  is,  to  a  person  looking  along  the  axis  in  the  positive  direc- 
tion the  turn  of  the  spiral  above  the  axis  is  from  the  left  hand 
to  the  right.     But  hydrodynamically  it  is  equally  possible  that 
y 
b  may  be  negative,  in  which  case  z  is  greater  as  tan-1  -  is  greater, 
and  the  turn  of  the  spiral  is  sinistrorsum. 
28.  Assuming,  for  the  reasons  above  given,  that  the  direction 
of  the  spiral  course  is  determined  by  the  atomic  or  molecular 
constitution  of  the  conductor,  it  is  conceivable  that  this  effect 
may  be  attributable  to  a  particular  arrangement  of  the  constituent 
atoms,  causing  the  path  of  least  resistance,  instead  of  being  rec- 
tilinear, to  be  continuously  maintained  in  a  spiral  form.  For 
instance,  such  a  modification  of  the  patli  might  be  produced  if 
the  mean  retardation,  due  both  to  the  reaction  and  the  arrange- 
ment of  the  atoms,  operated  in  a  direction  not  exactly  opposed 
to  that  of  the  stream. 
29.  The  quantity  of  fluid  which  in  a  unit  of  time  passes  a 
plane  perpendicular  to  the  axis  of  the  wire  is  proportional  to  the 
integral  of f(r)dr  between  certain  limits,  and  is  therefore  greater 
Phil.  Mag.  S.  4.  Vol.  43.  No,  2*8,  June  1872.  2  E 
