cannot be raised
beyond the theoretic limit of 32 feet: a limit made much lower in
practice by imperfections in the apparatus. For if, simplifying the
conception, he supposes the tube of the pump to be a square inch in
section, then the atmospheric pressure of 15 pounds per square inch on
the water in the well, can raise the water in the tube to such height
only that the entire column of it weighs 15 pounds. Having been thus
enlightened about the pump's action, the action of a barometer becomes
intelligible. He perceives how, under the conditions established, the
weight of the column of mercury balances that of an atmospheric column
of equal diameter; and how, as the weight of the atmospheric column
varies, there is a corresponding variation in the weight of the
mercurial column,--shown by change of height. Moreover, having
previously supposed that he understood the ascent of a balloon when he
ascribed it to relative lightness, he now sees that he did not truly
understand it. For he did not recognize it as a result of that upward
pressure caused by the difference between the weight of the mass formed
by the gas in the balloon _plus_ the cylindrical column of air extending
above it to the limit of the atmosphere, and the weight of a similar
cylindrical column of air extending down to the under surface of the
balloon: this difference of weight causing an equivalent upward pressure
on the under surface.

Why do I introduce these familiar truths so entirely irrelevant to my
subject? I do it to show, in the first place, the contrast between a
vague conception of a cause and a distinct conception of it; or rather,
the contrast between that conception of a cause which results when it is
simply classed with some other or others which familiarity makes us
think we understand, and that conception of a cause which results when
it is represented in terms of definite physical forces admitting of
measurement. And I do it to show, in the second place, that when we
insist on resolving a verbally-intelligible cause into its actual
factors, we get not only a clear solution of the problem before us, but
we find that the way is opened to solutions of sundry other problems.
While we rest satisfied with unanalyzed causes, we may be sure both that
we do not rightly comprehend the production of the particular effects
ascribed to them, and that we overlook other effects which would be
revealed to us by contemplation of the causes as analyzed.