THE ROYAL ARTILLERY INSTITUTION. 
93 
zenith. Five altitudes of a star as near as possible to the prime vertical 
east of the meridian, and a similar number of a star nearly in the same 
position west of the meridian. 
In north latitude where the pole star is at a convenient height for 
observing; ten observations on it at any time may be substituted for the 
circummeridian observation north of the zenith. The advantage of so doing 
as regards economy of time is very great, as it will be perceived that the only 
star out of the four which requires to be observed at a fixed time is the one 
culminating star south of the zenith. The other stars may be observed before 
or after this one according to the will or convenience of the observer. The 
star south of the zenith will be convenient for observing for two or three 
consecutive weeks, and a mere glance at the observation book for two places 
on consecutive nights gives immediate and tolerably accurate information 
with regard to the northing or southing made during the day. Continuous 
observations of the same stars obviate the necessity of removing the telescope 
from the sextant in order to bring the star down, as it is called, or to make 
both images of the star appear in the field of the telescope. 
An hour is amply sufficient to make the suggested observations, including 
those for index error. 
In observing stars near the meridian for latitude, it is a good plan to give 
the slow motion screw a turn or two in either direction after each observation 
before making another, in order that the judgment may not be biased by 
knowledge of which way the star ought to have moved with reference to its 
last observed position. The previous observation may have been a bad one, 
but however good, it should have nothing whatever to do with the one 
following it. Every observation should in fact be truly independent. The 
most convenient method of observing the time star is to set the instrument 
at a particular reading, wait for the two star images to coincide, and note the 
time of coincidence. In observing with the artificial horizon, the observer 
should sit down either on the ground, or on anything not higher than a 
sextant box ; he will then find that his right knee affords a firm support to 
his right arm near the wrist. 
If the system of observing stars uniformly and of not depending on the sun 
be adopted, it will be advisable to have the pocket chronometer or chro¬ 
nometers regulated to sidereal time. 
Box chronometers, such as are used on board ship, cannot be carried on 
shore so as to preserve anything like a uniform rate, but they are very 
useful in a field observatory where a transit or other telescope is the 
instrument observed with. 
Having objected to observations of the sun as compared with those of 
stars, it may be well to show what increase of accuracy may be expected 
from the latter, always excepting equal altitudes of the sun for time. These 
involve the principle now advocated and give unexceptionable results, but 
occupy a whole day, and are therefore generally impracticable. 
Besides index error sextants usually and almost invariably have other 
large errors arising from various causes, which it is not my intention to enter 
into here. They have been fully discussed by Mr W. H. Simms,'* and 
* The Sextant ancl its Applications, 1858.—Taylor and Francis. 
