, and
out of these to select those only which agree, or very nearly agree.
If a hundred observations are made, the cook must be very unlucky if he
cannot pick out fifteen or twenty which will do for serving up.

Another approved receipt, when the observations to be used will not
come within the limit of accuracy, which it has been resolved they shall
possess, is to calculate them by two different formulae. The difference
in the constants employed in those formulae has sometimes a most happy
effect in promoting unanimity amongst discordant measures. If still
greater accuracy is required, three or more formulae can be used.

It must be admitted that this receipt is in some instances rather
hazardous: but in cases where the positions of stars, as given in
different catalogues, occur, or different tables of specific gravities,
specific heats, &c. &c., it may safely be employed. As no catalogue
contains all stars, the computer must have recourse to several; and if
he is obliged to use his judgment in the selection, it would be cruel
to deny him any little advantage which might result from it. It may,
however, be necessary to guard against one mistake into which persons
might fall.

If an observer calculate particular stars from a catalogue which makes
them accord precisely with the rest of his results, whereas, had they
been computed from other catalogues the difference would have been
considerable, it is very unfair to accuse him of COOKING; for--those
catalogues may have been notoriously inaccurate; or--they may have
been superseded by others more recent, or made with better instruments;
or--the observer may have been totally ignorant of their existence.

It sometimes happens that the constant quantities in formulae given by
the highest authorities, although they differ amongst themselves, yet
they will not suit the materials. This is precisely the point in which
the skill of the artist is shown; and an accomplished cook will carry
himself triumphantly through it, provided happily some mean value of
such constants will fit his observations. He will discuss the relative
merits of formulae he has just knowledge enough to use; and, with
admirable candour assigning their proper share of applause to Bessel, to
Gauss, and to Laplace, he will take THAT mean value of the constant used
by three such philosophers, which will make his own observations accord
to a miracle.

There are some few reflections which I would venture to