488     M.  F.  Zollner  on  the  Origin  of  the  Earth's  Magnetism, 
From  the  observations  of  temperature  made  at  the  Cape  of 
Good  Hope  from  1841  to  1870  with  very  good  instruments, 
Stone*  calculated  the  mean  yearly  values  and  represented  them 
graphically  by  a  curve.  He  compared  with  this  a  second  curve 
indicating  the  frequency  of  sun-spots  as  given  by  the  observations 
of  Wolf,  and  found  that,  if  he  reversed  the  latter  curve,  it  coin- 
cided so  exactly  with  that  of  the  mean  yearly  temperature,  that 
he  expresses  the  opinion  that  "  the  same  cause  which  leads  to 
an  excess  of  mean  annual  temperature  leads  equally  to  a  dissipa- 
tion of  solar  spots." 
In  a  letter  to  Sabine  he  says  expressly  :  "  I  may  mention  that 
I  had  not  the  slightest  expectation,  on  first  laying  down  the 
curves,  of  any  sensible  agreement  resulting,  but  that  I  now  con- 
sider the  agreement  too  close  to  be  a  matter  of  chance.  I  should, 
however,  rather  lean  to  the  opinion  that  the  connexion  between 
the  variation  of  mean  temperature  and  the  appearance  of  solar 
spots  is  indirect  rather  than  direct,  that  each  results  from  some 
general  change  of  solar  energy." 
Piazzi  Smyth  had  already  made  a  communication  to  the  Royal 
Society  in  April  7,  1870,  in  which  he  got  the  same  result  by 
other  observations.  He  discussed  and  reduced  the  observations 
made  in  the  years  1837  to  1869  with  the  four  large  earth-ther- 
mometers which  had  been  sunk  into  the  rocks  of  Calton  Hill  in 
Edinburgh  by  Principal  Forbes,  in  consequence  of  a  decision  of 
the  British  Association.  Amongst  other  periodical  variations  in 
the  temperature  of  the  earth,  Piazzi  Smyth  found  a  period  of  11*1 
years,  or  the  sun-spot  period.  Some  circumstances  discussed  by 
Smyth  in  detail  lead  him  also  to  the  result  that  the  sun-spots 
cannot  be  the  true  cause  of  the  variation  of  temperature  in  the 
earth. 
I  think  it  very  probable,  after  these  results,  that  if  thermome- 
ters could  be  sunk  still  deeper  into  the  earth  the  influences 
would  be  much  more  decided,  as  the  retarding  influence  of  the 
incomplete  conduction  of  the  earth's  crust  would  be  less.  It 
depends  upon  the  degree  in  which  these  retarding  influences 
may  be  annihilated  whether  the  yearly  period  of  magnetic  dis- 
turbances and  auroras  will  be  found  in  these  observations. 
23. 
Although  in  §  9  I  have  deduced  theoretically  the  necessary 
connexion  between  mechanical  and  magnetical  changes  on  the 
surface  of  the  earth,  we  must  yet  inquire  whether  the  frequency 
of  mechanical  reactions  of  the  glowing  liquid  nucleus  of  the  earth 
shows  also  periodical  changes  which  coincide  with  the  changes 
of  those  causes  which  can  produce  the  mechanical  changes. 
*  Proceedings  of  the  Royal  Society,  No.  127  (1871). 
