456     M.  F.  Zollner  on  the  Origin  of  the  Earth's  Magnetism, 
It  has  been  shown  (§  15)  that  the  position  of  the  sun's  axis 
with  respect  to  the  plane  of  the  ecliptic  causes  theoretically  a 
stronger  magnetic  induction  near  the  two  days  of  the  year  during 
which  the  earth  is  most  exposed  to  the  inductive  action  of  one  of 
the  sun's  poles,  leaving  out  of  consideration  the  influence  which 
the  inclination  of  the  earth's  axis  must  have  on  this  induction. 
It  is  therefore  supposed  that  on  these  two  days  the  radius  vector 
of  the  earth  is,  as  in  the  time  of  the  equinoxes,  perpendicular  to 
the  earth's  axis.  As  those  two  days  have  been  found  to  be  the 
6th  of  September  and  7th  of  March,  the  supposition  may  be  con- 
sidered approximately  true  for  a  rough  comparison  with  the  mag- 
netic disturbances.  According  to  our  theory,  at  those  times  all 
those  phenomena  must  have  a  maximum  which  are  caused  by 
an  increased  magnetic  induction  of  the  earth ;  such  are  the  mag- 
netic disturbances  and  the  aurorse  boreales. 
I  take  as  material  for  the  examination  of  these  consequences 
the  Resultate  aus  den  Beobachtungen  des  magnetischen  Vereins 
1836,  III.,  which  are  discussed  by  Gauss  in  the  fifth  volume 
of  his  collected  works. 
Under  the  title  "  Mean  Variation  of  the  Magnetic  Declination 
during  1834-1837,"  these  values  are  given  as  monthly  means 
for  the  three  years  mentioned.  Although  the  number  of  ob- 
servations is  much  too  small  to  allow  a  decisive  comparison  with 
my  conclusions,  I  still  give  them,  expressed  in  seconds  of  arc 
(/.  c.  p.  567). 
Mean  Variation  of  Declination  at  Gottingen  from  1834  to  1837. 
Month.  Variation. 
January      .     . 
.     .     189 
February    .     . 
.     .     155 
March  .     .     . 
.     206 
April     .     .     . 
.     164 
May      .     .     . 
.     196 
June     .     .     . 
.     172 
July      .     .     . 
.     223 
August      .     .     , 
.     244 
September      .     . 
.     204 
October     .     . 
.     216 
November .     . 
.     191 
December  .     .     . 
.     195 
Although  these  numbers  show  a  maximum  in  March  and  be- 
tween August  and  September,  agreeing  with  the  times  men- 
tioned above,  a  much  greater  number  of  observations  is  required 
to  be  discussed  from  this  point  of  view. 
The  frequency  of  the  aurora  borealis  is  also  subjected  to  a 
yearly  period.     Loomis  has  deduced  such  a  period  in  the  paper 
