ON  THE  PEAK  OF  TENEEIFFE. 
511 
The  next  step  is  a more  delicate  one,  namely,  to  compare  the  intensity  in  declinations 
towards,  and  from,  the  sun.  For  this  purpose,  of  the  three  normal  Edinburgh  days, 
October  30  may  be  taken  by  itself,  and  the  two  shorter  series  on  November  20  and  24 
may  be  taken  together ; we  then  have  two  series,  both  indicating  the  maximum  of  polari- 
zation in  about  98°  of  sun  polar  distance,  thus — 
October  30. 
November  20  and  24. 
Sun. 
Polar  dist. 
No.  of 
observations. 
Inverse  inten- 
sity of  sky 
polarization. 
Sun. 
Polar  dist. 
No.  of 
observations. 
Inverse  inten- 
sity of  sky 
polarization. 
O 
75 
1 
26-0 
0 
O 
80 
1 
24-5 
80 
3 
23-5 
85 
1 
20-5 
85 
3 
21-6 
90 
2 
20-5 
90 
6 
21-1 
95 
3 
19*8 
95 
3 
21-7 
100 
1 
20-5 
100 
3 
21*7 
105 
1 
20-5 
105 
3 
28*1 
110 
1 
25-0 
115 
1 
26-0 
At  Alta  Vista  the  observations  of  September  10  appear  most  suitable  for  this  purpose, 
and  give  a maximum  intensity  about  82°  distance  from  the  sun,  thus — 
Sun. 
Polar  dist. 
No.  of 
observations. 
Inverse  inten- 
sity of  sky 
polarization. 
64-0 
26 
30-1 
70-0 
27 
26-1 
80-5 
27 
26-2 
85-0 
21 
25-1 
91*7 
25 
28-8 
100-8 
25 
29*6 
A series  was  also  taken  at  Alta  Vista  on  September  10  and  11,  to  try  the  intensity 
in  different  planes,  the  normal  plane  employed,  designated  on  the  instrument  by  90°, 
being  when  the  eye-piece  moved  in  a plane  at  right  angles  to  the  sun-shade  plane. 
Plane 
employed. 
No.  of 
observations. 
Inverse  inten- 
sity of  sky 
polarization. 
— 82-5 
18 
28-5 
-85-0 
18 
27-3 
— 87-5 
18 
26*3 
90-0 
17 
26-7 
+ 87-5 
19 
26-8 
-f  85-0 
19 
27*1 
+ 82-5 
18 
28*4 
These  observations  may  be  taken  as  establishing  90°  as  the  true  plane ; nine  observations 
in  Edinburgh  on  October  30,  gave  no  effect  from  +80°  to  — 80°  at  all  equal  to  the 
error  of  observation,  which  appeared  to  increase  with  the  distance  from  90°. 
The  plane  was  also  attempted  to  be  ascertained  from  the  estimated  amount  of  rotation 
MDCCCLVIII.  3 Y 
