spelling, and, to some extent, in
algebraical notation; it also seems that conjectural methods of introducing
interpolations into the text have been necessary. For all this we are
sorry: the scientific value of the collection is little altered, but its
literary value is somewhat lowered. But it could not be helped: the
printers could not work from the originals, and Professor Rigaud had to
copy everything himself. A fac-simile must have been the work of more time
than he had to give: had he attempted it, his death would have cut short
the whole undertaking, instead of allowing him to prepare everything but a
preface, and to superintend the printing of one of the volumes. We may also
add, that we believe we have notices of _all_ the letters in the
Macclesfield collection. We judge this because several which are too
trivial to print are numbered and described; and those would certainly not
have been noticed if _any_ omissions had {305} been made. And we know that
every letter was removed from Shirburn Castle to Oxford.

Two persons emerge from oblivion in this series of letters. The first is
Michael Dary,[560] an obscure mathematician, who was in correspondence with
Newton and other stars. He was a gauger at Bristol, by the interest of
Collins; afterwards a candidate for the mathematical school at Christ's
Hospital, with a certificate from Newton: he was then a gunner in the
Tower, and is lastly described by Wallis as "Mr. Dary, the tobacco-cutter,
a knowing man in algebra." In 1674, Dary writes to Newton at Cambridge, as
follows:--"Although I sent you three papers yesterday, I cannot refrain
from sending you this. I have had fresh thoughts this morning." Two months
afterwards poor Newton writes to Collins, "Mr. Dary is very solicitous
about mathematics": but in spite of the persecution, he subscribes himself
to Dary "your loving friend." Dary's _problem_ is that of finding the rate
of interest of an annuity of which the value and term are given. Dary's
_theorem_, which he seems to have invented specially for the solution of
his problem, though it is of wide range, can be exhibited to mathematical
readers even in our columns. In modern language, it is that the limit of
[phi]^{_n_}_x_, when _n_ increases without limit, is a solution of [phi]_x_
= _x_. We have mentioned the I. Newton to whom Dary looked up; we add a
word about the one on whom he looked down. Dr. John Newton,[561] a sedulous
publisher of logarithms, tables of