distances and positions of 380 double and triple stars , &c. 19 
When indeed a measure (on looking down the list, without 
reference to the observations of former nights) was found 
to differ considerably from the rest, the micrometer was 
usually set to the suspicious reading off, and the measure 
re-examined by both observers. If declared erroneous (and 
the contrary would occasionally happen), it was corrected 
by him whose measure it originally was, and the result 
set down in the place of that rejected. In general, the de- 
gree of discordance in the measures of any particular star, 
may be taken as a pretty fair criterion of the difficulty which 
attended the observation. 
The instrument with which each set of measures was taken, 
is mentioned. In the north preceding and south following 
quadrants, the micrometers show angles of position comple- 
mentary to the true ones. These are, however, (except in 
one or two instances ) set down as read off, and the mean 
afterwards subtracted from 90°. In the measures of distance, 
the index error is applied to the mean of the micrometer parts 
in each set, and the result reduced into seconds is stated. 
The index error was at first only taken at pretty consider- 
able intervals; but, being soon found liable to a trifling 
change, it was afterwards regularly taken on the morning 
after each night's observation, or at least as soon as circum- 
stances would permit. The zeros applied are means of at 
least five, but frequently of ten separate determinations. 
In order to make this paper more complete, and to save 
trouble to those who may wish to consult our measures, or 
prosecute farther this interesting department of astronomy, 
we have presented at the end of our observations of each star, 
1st, the mean result of our own measures, reduced to a mean 
