ASTRONOMICAL OBSERVATIONS. 
541 
with, the tangent screw, the reflected star coincide with the one seen 
by direct vision. The reading on the arc will be the index error, minus 
if on the arc, plus if off the arc. 
A very handy instrument, by Staff Commander George, R.N., is in Artificial 
general use by the Royal 'Geographical Society, and furnished by it horizon - 
to travellers. It consists of two small cisterns, of which one has a 
glass cover screwing on and off, the other holds the mercury used for 
the reflection. They are connected by a neck in which is a tap for 
letting the mercury out of the latter cistern into the former. When 
this is done the tap is closed, the glass unscrewed, and a carefully 
ground disc of glass is dropped on to the quicksilver on which it floats, 
making a good steady surface on which to take an observation. The 
diameter is 3*14 ins. A great advantage is that you have never to 
touch the mercury, and without great carelessness it is impossible to 
lose it. In still weather the glass may be dispensed with. 
The card should be half black, half white, to make it readable in a Compass, 
dim twilight. The graduations should range from 0° to 360° right 
round, not twice over from 0° to 180°. The needle should work 
steadily and quickly, not make long, slow oscillations. It may be 
adjusted to obviate depression from dip, by a sliding collar. A line 
for True North temporarily marked on the card in a position most 
appropriate to the magnetic variation of the country to be visited, is 
convenient. 
As the accuracy of an observation depends on the precision of the Watch, 
recorded time, the watch used should be a half-chronometer—that is to 
say, a lever watch with compensated balance, and carefully tempered 
pendulum spring: on these two considerations its ability to keep a 
steady rate chiefly depends. Great care should be taken to see that 
there is perfect coincidence in the pointing of the seconds hand to zero 
when the minute hand points to a division, otherwise the observer can¬ 
not tell to which minute the seconds belong. The case should have 
norther hinge nor spring. 
This should be capable of working without fracture over the highest Aneroid, 
mountain pass that is expected. It must be borne in mind that the 
aneroid, though excellent for differential observations, is unreliable for 
absolute heights, unless checked by boiling point barometer, as its index 
error is apt to vary without the traveller being able to ascertain or 
allow for the change. It is also uncertain at high altitudes, and cannot 
be depended on for giving correct measurements when descending from 
elevations. It has been simply alluded to as a useful instrument, and 
space does not admit of any remarks of its use for obtaining the 
differences in height between two stations. 
Maximum and minimum thermometers—these should be graduated Thermo 
on the stems, and it would be advisable to have them originally meters ‘ 
tested at Kew, and their errors there determinated and tabulated. 
Such a certificate will be sufficient for all practical purposes, as good 
thermometers are very slow to change their errors. 
Raper's navigation Tables, 12th edition, contain almost every Reduc- tables, 
tion. The chief recommendation of this book is the handiness of the 
tables, affording at a single reference results which would otherwise 
67 
