612 
THE EAEL OF EOSSE OX THE EADIATION 
Observations with glass plate ; reductions to zenith. 
I. 
II. 
III. 
IY. 
Y. 
VI. 
VII. 
VIII. 
1S70. 
Apparent 
zenith- 
distance. 
Obsei-Ted 
G. corr. 
Gr. calc. 
c-o. 
p.e. 
z' 
p.e x . 
Aug. 8. 
O 
77-4 
11-0 
20*4 
Aug. 10. 
73-3 
36-0 
56-7 
Aug. 11. 
69-8 
34-7 
33-4 
—'1-3 
+ 1*3 
47-5 
+ 1-3 
69*5 
32-3 
33-6 
+ 1*3 
Aug. 12. 
66-8 
27-9 
29-7 
+ 1*8 
+ 1-7 
39*5 
+ 1*6 
65*8 
32-0 
30-3 
-1-7 
Aug. 13. 
75*0 
15-0 
15-1 
+ 0*1 
+ 0-3 
25*2 
+ 0-2 
71-6 
16-5 
16-9 
+ 0*4 
61-5 
21*1 
20-7 
— 0*4 
Aug. 14. 
68*6 
14-5 
14-4 
-0*1 
+ 0*1 
20*0 
+ 0*1 
65*4 
15-4 
15-4 
0-0 
' Aug. 15. 
62-9 
7*1 
9-4 
+ 2-3 
+ 1*4 
11-7 
+ 0*9 
57-8 
10-5 
10-0 
-0*5 
54-0 
12-0 
10-3 
-1*7 
! Aug. 16. 
58-7 
9-5 
9-3 
-0*2 
+ 0*1 
11-0 
+ 0*1 
| 
54*3 
9-7 
9*7 
0*0 
51*6 
9-7 
9-9 
+ 0*2 
Aug. 1.9. 
57-5 
3-1 
2*8 
— 0-3 
+ 0*3 
3*2 
+ 0*2 
51-4 
2-6 
2-9 
+ 0-3 
Oct. 4. 
72-1 
13*8 
13-4 
-0*4 
+ 0*4 
20*3 
+ 0-4 
72-4 
12-9 
13-3 
+ 0*4 
Oct. 9- 
58-2 
62-7 
60-5 
-2*2 
+ 1*3 
70*9 
+ 0-8 
54*6 
61*4 
62-3 
+ 0*9 
52*8 
62-0 
63-2 
+ 1*2 
Oct. 10. 
59-0 
47*4 
45-8 
-1*6 
+ 1*5 
54*2 
+ 1-2 
54-3 
46-2 
47-7 
+ 1*5 
Before these results can be compared with each other, it is necessary to examine into 
another source of uncertainty, which, however, as will be seen, can not only be got rid 
of, but even made to yield a most interesting result. During the month of August an 
old and consequently tarnished speculum was used ; this was replaced by one of great 
brilliancy before the observations of October were commenced. Hence the readings 
taken in August are below, and those in October above what they would have been with 
a mirror of ordinary working quality. Taking these facts into consideration, the obser- 
vations were divided into two groups, according as they were made with the old or new 
speculum; and by comparing these groups with the Phase Table for 1871-72, at p. 605, 
the following values of speculum factors for reducing the observations to the standard 
mirror of those years were obtained. The weight which has been attached to the factor 
