110 
BRITISH ASSOCIATION.- 1835 . 
of the barometer at the successive hours of the day; and thus a 
curve representing the actual diurnal oscillations of the barometer 
can be placed before the eye, and a registry kept from day to day 
on separate sheets of paper. The mean curve can also be had by 
making the pencil traverse, day after day, for a long period, the 
same sheet of paper; for the pencil-marks will at length become 
blackest and heaviest upon the parts corresponding to the mean 
curve; and thus all the labour of actual observation, registry, &c. 
will be avoided, and thus, too, much of the trouble of reduction, if 
not all, will be saved. 
Many mechanical methods of suspending the cistern will readily 
suggest themselves to persons conversant 'with practical matters ; but 
the method that is preferred by the author is by a mercurial hy¬ 
drometer, the cistern, for the sake of stability, being suspended 
underneath the hydrometer, as in Ronchetti’s modification of Xichol- 
son’s hydrometer. The accompanying drawing will give an idea of 
the form of the instrument; the following is the description of it. 
The guide-wheels and supports are omitted. 
B the barometer tube (it may be of 
iron) firmly fixed in its place, and dip¬ 
ping below into 
C, the cistern, which is suspended by 
F, a frame, supported by 
S, the pillar or stem of 
H, the hydrometer ball, which floats 
in 
A, a vessel firmly fixed, and contain¬ 
ing the mercury (or other fluid) in 
which the hvdrometer floats. 
In the description of this instrument 
given to the .Subsection, it was sup¬ 
posed that the surfaces of the mercury 
in the cistern and in the vessel A were 
so large that the rising or falling of 
the fluid in these vessels might be 
neglected; also, since the instrument 
is very sensible, it was supposed that 
the lower part of the barometer tube 
which dips into the cistern should be 
rendered very small, in order to di¬ 
minish unsteady oscillation. Also the 
internal part of the barometer tube 
B at the upper part, the external 
part where it dips into the mercury in 
the cistern, as well as the cistern and 
the vessel A, at the surfaces of the 
mercury in them, were all supposed to 
be cvlindrical. And it was then shown 
in a popular manner, that if the internal cross section of the baro- 
