Alteration  of  Aberration  of  Light  through  a  Refracting  Medium .  313 
Remarking  that  the  mean  results  for  Geographical  Latitude  of  the 
Instrument  (determined  from  observations  made  when  the  Aberration 
of  the  star  had  respectively  its  largest  +  value  and  its  largest  —  value) 
agree  within  a  fraction  of  a  second,  I  think  myself  justified  in  con- 
cluding that  the  hypothesis  of  Professor  Klinkerfues  is  untenable. 
Had  it  been  retained,  the  Aberrations  to  be  employed  in  the  cor- 
rections would  have  been  increased  by +15"  and  — 15"  respectively, 
and  the  two  mean  results  would  have  disagreed  by  30". 
The  latitude  of  the  instrument  from  these  observations  is  about 
51°28'34"'0.  The  position  of  the  instrument,  as  measured  on  the 
Observatory  Map,  is  340  feet  south  of  the  Transit-circle,  a  spatial 
distance  corresponding  to  about  3"'35.  The  latitude  of  the  Transit- 
circle  being  taken  at  51°  28'  38"*4,  the  geodetic  latitude  of  the  in- 
strument is  51°  28'  35"*05 — an  agreement  closer  than  I  expected, 
consideration  being  given  to  the  form  of  the  ground.  It  appears  very 
probable  that  at  the  place  of  the  Transit-circle,  on  the  north  brow 
of  the  hill,  the  zenithal  direction  is  disturbed  towards  the  north,  and 
the  astronomical  latitude  is  too  great. 
There  is  only  one  point  in  this  investigation  upon  which  a  doubt 
can  be  suggested  as  possible,  namely  the  evaluation  of  the  micro- 
meter-scale. It  was  thus  conducted  :—  The  micrometer-plate  con- 
tains 26  wires,  and  the  fixed  part  of  the  instrument  contains  25 
crosses,  each  interval  being  nearly  256".  With  this  arrangement 
every  wire-interval  is  measured  with  great  ease,  and  the  whole  series 
of  25  intervals  is  accurately  obtained  in  terms  of  the  micrometer.  By 
placing  the  instrument  in  a  proper  position,  the  same  intervals  are 
obtained  in  time  of  the  star's  transit,  which  is  easily  converted  into 
arc.  The  comparison  of  these  gives  the  value  of  micrometer-divisions 
which  has  been  employed. 
The  following  verification,  of  somewhat  inferior  accuracy,  has  been 
made  by  measures  of  the  instrument.  It  appears  that  the  ray  of 
light  passes  through  0'9  inch  of  glass,  35'3  inches  of  water,  and  0*8 
inch  of  air  (nearly,  the  measure  of  the  last  being  slightly  uncertain). 
Remarking  that  the  dividing  surfaces  are  horizontal  and  plane,  it  is 
easily  seen  that  the  micrometer-scale  ought  to  be  such  as  is  due  to 
an  air-telescope  whose  length  in  inches  =  J|  +  -fUs  +  0*8  =  27'8 
inches.  And  from  this,  with  observation  of  transit  of  the  star,  it  was 
found  that  the  measure  of  25  intervals  of  wires  ought  to  be  0*8693 
inch :  as  measured  with  a  pair  of  compasses,  it  was  found  some- 
times 0-8/l,  sometimes  0-8/5.  The  agreement  is  fully  as  close  as 
can  be  expected  from  the  rudeness  of  the  operation,  and  shows  di- 
stinctly that  there  can  be  no  error  of  principle  in  the  method  of 
evaluing  the  micrometer-scale. 
