122 
Dividing this series into two, the first containing the 
minima numbered 1 to 18, and the second those numbered 
from 19 to 35, and determining the mean epoch for each series, 
we obtain from the interval between the two epochs and 
the corresponding number of periods, the value of the mean 
period on the assumption that the period is uniform and 
only subject to accidental changes. This value is 70T54 
days, and the mean epoch for the entire series of minima is 
1866, July 25-815. 
Calculating the times of minima from these elements and 
comparing the observed with the calculated times we obtain 
the following difierences : — 
No. 
1. - 8-44 days. 
No. 
19. + 4-11 days. 
5) 
2. -6-59 
yy 
yy 
20.+ 5-96 
yy 
JJ 
3. -1-90 
yy 
yy 
21. + 4-80 
yy 
JJ 
4. -4-05 
yy 
yy 
22.+ 8-34 
yy 
5) 
5. -1-20 
yy 
yy 
23.+ 7-19 
yy 
» 
6. -2-35 
yy 
yy 
24.+ 8-03 
yy 
?) 
7. -1-66 
yy 
yy 
25.+ 8-88 
yy 
)? 
8. -2-81 
yy 
yy 
26.+ 7-26 
yy 
9. + 2-04 
yy 
yy 
27.+ 5-11 
yy 
J? 
10. + 1-88 
yy 
yy 
28.+ 7-65 
yy 
11. + 1-42 
yy 
yy 
29.+ 6-49 
yy 
jy 
12.-2-74 
yy 
yy 
30.-13-51 
yy 
yy 
13. + 4-65 
yy 
yy 
31.-11-66 
yy 
yy 
14. + 3-50 
yy 
yy 
32.- 9-82 
yy 
yy 
15. + 1-34 
yy 
yy 
33.-13-28 
yy 
yy 
16. + 4-88 
yy 
yy 
34.-12-43 
yy 
yy 
17. + 6-73 
yy 
yy 
35. - 12-57 
yy 
„ 18. + 4-57 
The distribution of 
yy 
the plus and 
minus errors shows at 
once that the period has not been uniform, but that it was 
longer in the earlier than in the later years. I therefore 
grouped the minima observed in the years 1859 to 1868 into 
four groui>s, and on forming the usual equations, and treating 
