that I have found the pressure of
the _Atmosphere_ lighter at the top of St. _Paul_'s Steeple in _London_
(which is about two hundred foot high) then at the bottom by a sixtieth or
fiftieth part, and the expansion at the top greater then that at the bottom
by neer about so much also; for the _Mercurial Cylinder_ at the bottom was
about 39. inches, and at the top half an inch lower; the Air also included
in the Weather-glass, that at the bottom fill'd only 155. spaces, at the
top fill'd 158. though the heat at the top and bottom was found exactly the
same with a scal'd _Thermometer_: I think it very rational to suppose, that
the greatest Curvature of the Rays is made nearest the Earth, and that the
inflection of the Rays, above 3. or 4. miles upwards, is very
inconsiderable, and therefore that by this means such calculations of the
height of Mountains, as are made from the distance they are visible in the
Horizon, from the supposal that that Ray is a straight Line (that from the
top of the Mountain is, as 'twere, a Tangent to the Horizon whence it is
seen) which really is a _Curve_, is very erroneous. Whence, I suppose,
proceeds the reason of the exceedingly differing Opinions and Assertions of
several Authors, about the height of several very high Hills.

8. Whether this Inflection of the Air will not very much alter the supposed
distances of the Planets, which seem to have a very great dependence upon
the Hypothetical refraction or inflection of the Air, and that refraction
upon the hypothetical height and density of the Air: For since (as I hope)
I have here shewn the Air to be quite otherwise then has been hitherto
suppos'd, by manifesting it to be, both of a vast, at least an uncertain,
height, and of an unconstant and irregular density; It must necessarily
follow, that its inflection must be varied accordingly: And therefore we
may hence learn, upon what sure grounds all the Astronomers hitherto have
built, who have calculated the distance of the Planets from their
Horizontal _Parallax_; for since the Refraction and _Parallax_ are so
nearly ally'd, that the one cannot be known without the other, especially
by any wayes that have been yet attempted, how uncertain must the
_Parallax_ be, when the Refraction is unknown? And how easie is it for
Astronomers to assign what distance they please to the Planets, and defend
them, when they have such a curious _subterfuge_ as that of Refraction,
wherein a very little variation