he most important
of this class of errors arise from the non-existence of
natural indications other than those afforded by astronomical
observations themselves, whether an instrument has, or has
not, the exact position with respect to the horizon, and the
cardinal points, etc., which it ought to have, properly to
fulfill its object.

"Now, with regard to the first two classes of error, it must
be observed, that in so far as they can not be reduced to
known laws, and thereby become the subjects of calculation and
due allowance, _they actually vitiate in their full extent the
results of any observations in which they subsist_. With
regard to errors of adjustment, not only the possibility, _but
the certainty of their existence in every imaginable form, in
all instruments_, must be contemplated. _Human hands or
machines never formed a circle, drew a straight line, or
executed a perpendicular, nor ever placed an instrument in
perfect adjustment, unless accidentally, and then only during
an instant of time._"

The bearing of these important and candid admissions of error in
astronomical observations upon all kinds of other observations made by
mortal eyes, and with instruments framed by human hands, in every
department of science, is obvious. No philosophical observation or
experiment is absolutely accurate, or can possibly be more than
tolerably near the truth. The error of a thousandth part of an inch in
an instrument will multiply itself into thousands, and millions of
miles, according to the distance of the object, or the profundity of the
calculation. Our faith in the absolute infallibility of scientific
observers, and consequently in the absolute certainty of science, being
thus rudely upheaved from its very foundations by Sir John Herschel's
crowbar, we are prepared to learn that scientific men have made errors
great and numerous.

To begin at home, with our own little globe, where certainty is much
more attainable than among distant stars, we have seen that astronomers
of the very highest rank are by no means agreed as to its diameter. Its
precise form is equally difficult to determine. Newton showed that an
ellipsoid of revolution should differ from a sphere by a compression of
1/230. The mean of a number of varying measurements of arcs, in five
different places, would give 1/299. The pendulum measurement differs
very consi