140 Altitudes of Mountains, Sc. visible from 
A spirit level firmly fastened to the upper, and another to the 
under surface of a firm brass bar being substituted for the former 
level of the large Sector, the mean of the readings gave 30” as 
the error of the lesser instrument. A plane piece of glass with 
a very delicate mark in the middle superseded the cioss wires, 
and being fixed upon the object with the index at zero, the 
bubble of the upper glass tube was adjusted to its mark. ‘Phe 
telescope was next inverted, replaced upon the object, and the 
index levelled by the other glass tube, now uppermost. 
The double of the angle was thus obtained, and the whole 
operation repeated, with the bar carrying the levels reversed. 
One fourth of the two double angles is of course the correct one. 
Finally, the eye tube and the one containing the object glass 
were taken out of the smaller Sector, and reversed. ‘Ihe eleva- 
tions were in conseyuence increased 52”, half of which, or 26”, 
is the error thereof. This is perhaps the most satisfactory test of 
the three, the other methods not being perfectly unobjectionable. 
With this correction the mean refraction will be found to be 
about 45. The following remarks may render the more marked 
deyiations from this quantity more intelligible. 
Ist. When the are is but small, an error of a few seconds in 
the observations, or in the reduction of the height of the instru- 
ment to the ground, will cause a material alteration in the de- 
termined value of the refraction. 
2nd. Some few of the arigles at Rumbles Moor were only 
taken at the time of the diurnal variation, and the extremes 
were not always observed. Rumbles Moor and Jack Hill come 
under this class. 
od. The refractions at Great Almias are unusually small, but 
the station is on a group of huge rocks, which were probably 
heated to such an excess at the time of the observations by the 
previous intolerable heat of the sun’s rays, as to render the lower 
strata of the air rarer than those immediately above. 
Athly. Stations on isothermal curves will have refractious dif- 
fering from those at right angles to them. 
Excluding the journal kept at the Observatory, the mean of 
the heights of the barometer at the different stations would be 
28°50, and the temperature 54. According to the below obser- 
vations, the thermometer falls one degree tor everv ascent of 
224 feet. 
With these data, the computed will not be found to exceed 
the observed refraction very materially. 
Remarks. The greatest difference of temperature was observed 
when the thermometer was 10 degrees lower at the Moor than 
at the Observatory, It was but very rarely that the air proved 
warmer at the more elevated station, rT 
ne 
