462 
COAST AND GEODETIC SURVEY REPORT, 1904. 
On each night the observations were commenced as early as possible. Two 
observers, J. F. H. and one other, took alternate stars in the list, J. F. H. taking the 
odd numbers on about half of the nights and on the remaining nights the even numbers. 
The observers were forced to depart from the rule of observing alternate stars on some 
occasions on account of missed stars, or for the purpose of balancing the azimuth factors 
in a half set. When each observer had secured observations on five or six stars, the 
horizontal axis was reversed and another half set taken and thus a complete time set 
secured for each observer. The stride level was read four times during the double set 
of stars. Each observer thus secured a time set of ten to twelve stars, from which the 
the errors of azimuth, collimation, and level could be eliminated with considerable 
accuracy by computation. The mean epochs of the two time sets were nearly the same 
and they were affected in nearly the same manner by any irregularity in the running of 
the chronometer. The difference between the two derived chronometer errors from 
such a pair of time sets was evidently due almost entirely to errors of observation and 
to relative personal equation. 
Each of such a pair of complete time sets was computed independently of the 
other, except that the stride level readings were common to the two. In a few cases 
in which incomplete sets were secured, the collimation constant (and sometimes, also, 
the azimuth constant) was adopted from some complete set. 
The computations were made by a modified form of the field method of reduction 
stated in Appendix 9 of the Report for 1896. 
Principal Direct Results of Observation. 
In the following tables, c is the error of collimation of the telescope as derived from 
the observations and expressed in equatorial seconds of time. The quantities a w and a e 
are the azimuth errors of the instrument with band west and east, respectively, expressed 
in seconds of time. The time sets are grouped in pairs in the table to indicate which 
ones were taken simultaneously. 
Table i. 
Date 
Observers 
Number of 
stars 
Chronometer 
error 
c 
a w 
a e 
1904 
s. 
Mar. 
15 
E. G. F. 
II 
+0. 957 
+ 0.016 
+0. 288 
+0. 483 
J. F. H. 
5 
+1-053 
* 
+0. 034 
Mar. 
15 
fE. G. F. 
II 
+0.957 
+0.016 
+0. 288 
+0. 483 
O. B. F. 
IO 
+ 1.070 
* 
+0. 046 
Mar. 
16 
E. G. F. 
IO 
+ 1. 192 
+0.075 
+0. 286 
—0. 060 
J. F. H. 
12 
+ 1-233 
+0.093 
+0. 223 
—0. 008 
Mar. 
16 
E. G. F. 
IO 
+ 1-356 
+0.118 
+0. 041 
—0. 040 
J. F. H. 
II 
+ I-34I 
+0.113 
+0. 022 
+0.232 
Mar. 
16 
O. B. F. 
12 
+ 1. 445 
+0. 070 
+0. 158 
+0.003 
J. F. H. 
II 
+ I-532 
+0.097 
+0. 084 
—0. 104 
* c adopted from simultaneous set by E. G. F. No reversal of telescope, 
f This line of table repeated from above.