74  Royal  Society. 
perature  upon  the  direction  of  the  currents  were  the  same.  The 
molecular  movements,  therefore,  produced  by  friction  are  not  in  all 
cases  similar  to  those  produced  by  heat. 
Influence  of  Magne-optic  rotating-power  of  the  Liquids, 
Being  desirous  of  determining  whether  the  thermo-electric  pro- 
perties of  liquids  were  dependent  on  the  molecular  structure  by  virtue 
of  which  liquids  under  the  influence  of  magnetism  polarize  light  cir- 
cularly, I  made  the  following  apparatus  and  experiment : — 
A  and  B  (fig.   11)  are  two  straight  glass  tubes,  about  f  inch  in 
diameter  and  10  inches  long,  with  two      Fig-  11. 
similar  (but  bent)  tubes,  C  and  D,  at- 
tached  to  their  free  ends  by  india-rubber 
tubing.  The  sloping  ends  of  the  straight 
tubes  are  ground  flat,  and  are  joined 
together  securely  at  their  edges  by  melted 
shellac,  with  a  thin  and  projecting  sheet 
of  platinum  between  them  to  separate 
the  liquids.  E  and  F  are  two  strong 
electro-helices  wound  upon  stout  tubes  of  soft  iron  which  enclose 
the  glass  tubes.  The  apparatus  is  secured  upon  a  board  in  an 
inclined  position  with  the  sloping  ends  of  the  tubes  uppermost; 
and  the  two  helices  are  held  together  at  their  upper  ends  by  an 
india-rubber  band,  G.  I  filled  one  of  the  tubes  with  a  clear  and 
strong  solution  of  perchloride  of  iron  (of  negative  magne-optic  ro- 
tatory power,  see  Verdet,  Phil.  Mag.,  June  1858),  and  the  other 
with  a  similar  solution  of  chloride  of  nickel  (of  positive  magne-optic 
rotatory  power),  and  connected  the  liquids  in  the  bent  tubes  with  a 
galvanometer  16  feet  distant  by  means  of  the  platinum  wires  H 
and  I.  I  now  excited  the  helices  in  various  ways  by  means  of  12 
strong  Grrove's  cells ;  no  current  was  induced  in  the  liquid.  I  next 
heated  the  junction  of  the  tubes  gradually ;  the  solution  of  iron 
became  thermo-electro-positive,  and  a  steady  but  feeble  deflection  of 
the  needles  took  place ;  and  during  the  continuance  of  this  current  I 
again  excited  the  helices  in  various  ways  as  before ;  again  no  elec- 
trical effects  were  produced.  The  results  of  this  experiment  strongly 
support  the  conclusion  that  the  thermo-electric  properties  of  liquids 
are  not  dependent  upon  the  magne-optic  polarizing  power  of  the 
liquids,  nor  upon  the  properties  of  their  mass. 
On  examining  the  thermo-electric  properties  of  the  solution  of 
ferric  chloride  with  platinum  plates  in  the  apparatus  described  in 
the  'Philosophical  Magazine,*  1857,  vol.  xiii.  p.  1,  the  hot  plati- 
num was  strongly  negative,  value  of  temporary  deflection  *8475. 
With  the  nickel  solution,  similarly  examined,  the  hot  plate  was  also 
negative,  value  of  deflection  *0409.  These  results  agree  with  that 
obtained  with  the  two  tubes  in  the  last  experiment,  the  more  posi- 
tive condition  of  the  iron  solution  than  that  of  the  nickel  one  de- 
termining the  direction  of  the  current  in  that  experiment. 
General  Conclusion. — The  electric  currents  produced  by  the  direct 
