Changes in the Sun's apparent Diameter. 271 
The double integral therefore becomes 
8 
; + 00 
Fe [etn 00 des Fa ag aot. 
- © £ 
during the years 1862-1870 inclusive. The difference of me- 
ridian, five hours, will prevent the detection of any inequality of 
which the period is less than a day, while one with a period of 
six months or a year will be confounded with errors of observa- 
tion having that same period, which probably arises from at- 
mospheric condition. But, an inequality, either regular or 
irregular, of which the period ranges between a day and a half 
year, will admit of complete detection by the proposed com- 
parison. 
All the Greenwich observations which we have used were 
made with the transit circle: from January, 1862, to January, 
1866, the Washington observations were made with the Erte 
transit and the Troughton mural circle; after this date the 
Pistor and Martin’s meridian circle was alone used for this 
urpose. 
The method of observation at each place is well known, and 
it only remains to be said that all transits were registered by 
chronograph. The observations are distributed as follows: 
No. of Observations, 
H.D. V.D. 
Greenwich 1862-1870: 832 905 
Washington 1862-1865: 491 430 
te 1866-1870: 490 491 
1813 1826 
Many observers were employed in this work, and it was first 
necessary to make the observations homogeneous by subtract- 
ing from each separate “apparent error of Ephemeris diam- 
eter” the “personal error of the observer.” These last errors 
Were assumed to be constant throughout a year, and were deter- 
e 
given by his observations, which was called the “adopted per- 
sonal error” of that observer for that year. In some cases 
Some slight changes from this rule have been allowed, but the 
following tables are believed to represent each observer's habit 
as well as possible from the data. 
