14 REPORT—1848. 
eyelet-hole made in a small plate attached for this purpose. The reflector is pro- 
perly adjusted when a star is seen in coincidence with the left-hand angular point 
of the plate at the opposite end of the bar, at the same time that it is seen by direct 
vision through the other eyelet-hole in coincidence with the right-hand angular 
point. There are two clamps, one for clamping the bar to the vertical arc, and the 
other for clamping the vertical arc to the azimuthal circle. The latter may be held 
by the right-hand to give the azimuthal movement, and at the same time to be 
in readiness to clamp, while the bar is held by the left-hand for aiming. When 
the bar is not clamped to the vertical arc, it is prevented slipping partly by a spring 
and partly by a counterpoise. There are verniers to read off azimuths and altitudes 
to single minutes, and the vertical arc carries a small spirit-level for the horizontal 
adjustment. The instrument has a tripod support, with three screws for adjusting 
horizontally, and when in use is placed on a wooden stand, to the upper surface of 
which are fastened three brass Ys. The feet of the screws are placed in these Ys, 
and thus the instrument is put expeditiously in a given position. When not in use, 
it is kept under cover near the stand. 
On very dark nights the edges of the plate at the object-end of the bar were seen 
with difficulty. To remedy this inconvenience the face of the plate turned towards 
the eye was painted white, after which the light from a lamp at a considerable 
distance made it sufficiently visible. In general the luminosity of the sky makes 
the plate appear dark on a light ground. 
It is proposed to employ the meteoroscope in measuring the positions of arches 
and coronz of the Aurora Borealis, the dimensions and position at different times 
of the year of the Zodiacal Light, and the points of first appearance and disappear- 
ance of meteors and shooting stars. Several of these observations, to be of any 
value, require to be made simultaneously at different localities, and with the same 
degree of precision. It seems to me surprising that meteorologists have not hitherto 
provided themselves with instruments like that [ have been describing ; many obser- 
vations, of meteors having been comparatively useless on account of want of accu- 
racy. 1 consider that with care the altitude of a star may be measured by this 
instrument with a probable error of two minutes, and that it is abundantly accurate 
for the purposes to which it is proposed to apply it. In any case in which it is 
employed it is advisable to take the altitude and azimuth of a known star at a noted 
time near the place of the meteor, in order to eliminate index errors and errors of 
adjustment. 
Cambridge Observatory, August 9, 1848. 
On a Self-Registering Thermometer. 
By MansrFie_p Harrison (in a letter to Professor Puriiies). 
The principle on which the instrument is constructed is the difference in the ex- 
pansion and contraction of two metals, from the effects of heat and cold, and 
it acts by the direct pull of the contracting metal, when it is kept in a perfectly 
straight line. It is made sufficiently powerful to overcome any resistance which 
the fulcra of the levers or the tracing-pencil may cause. I have selected cast iron 
and hard rolled copper as the best suited for the purpose. I find from tables pub- 
lished by Smeaton and others, that copper expands ;3,th of its length, while cast 
iron only expands 535th, with a variation of 180° of Fahrenheit’s thermometer, 
which leaves a difference of about the + >;5th of its length; and as the range of the 
thermometer in the shade in this climate is about 90°, or half of 180°, I have the 
seazoth part of the length of the copper bar employed as a moving power. I fixed 
upon a bar 10 feet long as being a convenient length; the two metals will then vary 
nearly the one-and-twentieth part of an inch between the hottest day in summer 
and the coldest day in winter. This variation I multiply by means of a compound 
lever, so as to get a sufficient scale to divide. The end of the last lever carries a 
pencil which traces upon a revolving cylinder the variations that take place. In 
order to divide the scale accurately, I procured a standard thermometer by Trough- 
ton and Simms; I placed it in the same situation, and made several observations in 
the day, for some weeks, in the spring of the year, when the range of the thermo- 
meter is the greatest. After I had got the scale properly divided, I engraved it on 
