232 
SECOND REPORT-1832. 
erly point in Europe at which any observations of long con¬ 
tinuance on this subject have been made. The results have 
been published at length in the Edinburgh Transactions* ; and 
I have also entered into an analysis of all the existing informa¬ 
tion on the subject. The following are the general results at 
which I have arrived. 
1 st, That near Edinburgh, in lat 56°, the mean annual oscil¬ 
lation between 10 a.m. and 4 p.m. is *0106 inch or 0 mra, 27. 
2nd, That the hours of maxima are further from noon in 
spring and summer than in autumn and winter; and that the 
amount of oscillation of both the diurnal periods diminishes 
regularly through the seasons from spring to winter. These 
conclusions, derived directly from my own observations, I have 
shown to be the most probable for all parts of the globe, as far 
as existing observations guide us. 
3rd, That the St. Bernard observations, and those of Cap¬ 
tain Parry in the arctic regions, both indicate a true negative 
oscillation, though the second result has been overlooked by 
M. Bouvard. 
4th, That M. Bouvard’s hypothesis and formula mentioned 
above, are founded upon too hasty generalization. This 1 
have shown upon various grounds, but especially from his own 
quotation of the St. Bernard observations, where, as the mean 
temperature is much above 0° cent, in summer and below it in 
winter, the oscillation should be distinctly positive in the former 
case, and negative in the latter. This I have shown to be pre¬ 
cisely the reverse of the fact. 
5th, Availing myself of M. Bouvard’s excellent Table, with 
such additions as I could make to it, I proceeded to investigate, 
from observations made near the level of the sea alone, the in¬ 
fluence of latitude in modifying the oscillation; and from a 
careful combination of the best results, by reducing the squares 
of the errors to a minimum, I obtained the following equation, 
which represents wonderfully well the existing observations : 
% = 3*031 cos^ $ — ‘38 1 
for millimetres, z being the oscillation in latitude d. This gives 
for the equatorial oscillation 2 mm '650, and for the poles —*381. 
The latitude where the oscillation changes its sign, or is — 0, 
is 64° 85 
•* Vol. xii. The title of the paper is, “ On the Horary Oscillations of the 
Barometer near Edinburgh, deduced from 4410 Observations ; with an Inquiry 
into the Law of Geographical Distribution of the Phenomenon.” An abstract 
has been printed in the Edinburgh Journal of Science for April 1832. 
