Theory  of  Magnetism.  425 
43.  Accordingly  the  inoon,  since  it  completes  relatively  to  the 
aether  at  rest  a  revolution  about  its  axis  in  a  month,  will  gene- 
rate gyrations,  the  effect  of  which  might  possibly  extend  to  the 
earth,  and  be  perceived  as  a  very  small  variation  of  terrestrial 
magnetism.  But  as  these  gyrations  are  in  the  direction  of  the 
inoon' s  revolution  about  her  axis,  it  will  be  found  on  trial  that 
the  disturbances  of  the  magnetic  needle  which  they  produce  are 
not  in  accordance  with  the  observed  law  of  the  lunar-diurnal 
variation  of  magnetism.  "When,  however,  the  effect  of  the  dis- 
turbance of  the  aether  by  the  moon's  orbital  motion  is  considered, 
it  will  be  seen  that  as  the  impressed  motion  is  tangential  to  the 
orbit  and  contrary  in  direction  to  that  of  the  moon's  motion,  the 
generated  circulating  motion  that  reaches  the  earth  will  he,  from 
the  right  hand  to  the  left  of  a  spectator  on  the  earth  looking 
towards  the  moon,  and  therefore  in  the  direction  opposite  to  that 
of  gyrations  resulting  from  the  moon's  rotation  about  her  axis. 
Also  their  effect,  it  may  be  presumed,  would  be  much  greater 
than  that  of  these  gyrations.  Now  the  circulating  streams  pro- 
duced as  above  stated  by  the  moon's  orbital  motion  would  dis- 
turb the  magnetic  needle  in  a  manner  conformable  with  the  law 
of  the  lunar-diurnal  variation  ascertained  by  observation.  For 
when  the  magnet  is  on  the  meridian  and  under  the  moon,  these 
streams,  flowing  eastward,  would  oppose  the  needle's  entering 
streams  at  its  north  end  on  the  east  side,  and  the  issuing  streams 
at  its  south  end  on  the  west  side,  so  that  by  both  actions  the 
western  declination  would  be  increased.  The  variation  of  the 
declination  in  any  other  position  of  the  needle  relative  to  the 
moon  might  on  these  principles  be  readily  investigated ;  and  it 
is  easy  to  see  that  according  to  this  theory  there  would  be  two 
maxima  and  two  minima  of  declination  in  the  course  of  twenty- 
four  lunar  hours.  These  deductions  agree  with  the  results  of 
observation  given  in  p.  (ccxliii)  of  the  Greenwich  Observations  of 
1867. 
44.  The  solar-diurnal  variations  of  magnetism  follow  a  differ- 
ent law,  there  being  but  one  principal  maximum  in  the  course  of 
twenty-four  hours.  This  fact  is  accounted  for  by  the  circum- 
stance that  the  sun's  heat  produces,  cceteris  paribus,  a  gradual 
diminution  of  the  density  of  the  atmosphere  in  all  directions 
converging  to  the  position  of  greatest  heat,  it  being  assumed, 
conformably  with  the  argument  in  article  4,  that  such  grada- 
tion of  density,  in  consequence  of  its  extending  over  a  large 
space,  is  capable  of  generating  magnetic  streams  of  sensible  mag- 
nitude. So  far  as  this  cause  operates,  we  might  expect  that 
there  would  be  magnetic  effects  due  to  the  rapid  changes  of  tem- 
perature and  of  atmospheric  density  in  the  day-time  of  a  more 
decided  character  than  any  due  to  the  slower  changes  in  the 
