twentieth edition, revised, augmented, and above a thousand faults
amended, by Henry Mose, late servant and successor to the author.
Boston: printed by J. Franklin, for S. Phillips, N. Buttolph, B.
Elliot, D. Henchman, G. Phillips, J. Elliot, and E. Negus, booksellers
in Boston, and sold at their shops. 1719."

The book is a very small octavo, the type and execution are creditable, the
woodcut at the beginning is clumsy. It is a perfect copy, page for page, of
the English editions of Mose's Hodder, of which the one called seventeenth
is of London, 1690. There is not a syllable to show that the edition above
described might not be of Boston in England. Presumptions, but not very
strong ones, might be derived from the name of _Franklin_, and from the
large number of booksellers who combined in the undertaking. It chanced,
however, that a former owner had made the following note in my copy:

"Wednessday, July y^e 14, 1796, att ten in y^e forenoon we sail^d from
Boston, came too twice, once in King Rode, and once in y^e Narrows.
Sail^d by y^e lighthouse in y^e even^g."

{267}

No ordinary map would decide these points: so I had to apply to my friend
Sir Francis Beaufort,[422] and the charts at the Admiralty decided
immediately for Massachusetts.

PARADOXES OF ORTHOGRAPHY AND COMPUTATION.

The French are able paradoxers in their spelling of foreign names. The Abbe
Sabatier de Castres,[423] in 1772, gives an account of an imaginary
dialogue between Swif, Adisson, Otwai, and Bolingbrocke. I had hoped that
this was a thing of former days, like the literal roasting of heretics; but
the charity which hopeth all things must hope for disappointments. Looking
at a recent work on the history of the popes, I found referred to, in the
matter of Urban VIII[424] and Galileo, references to the works of two
Englishmen, the Rev. Win Worewel and the Rev. Raden Powen. [Wm. Whewell and
Baden Powell].[425]

I must not forget the "moderate computation" paradox. This is the way by
which large figures are usually obtained. Anything surprisingly great is
got by the "lowest computation," anything as surprisingly small by the
"utmost computation"; and these are the two great subdivisions of "moderate
computation." In this way we learn that 70,000 persons were executed in one
reign, and 150,000 persons {268} burned for witchcraft in one century.
Sometimes this computation is very close. By a card before me it