me a start. It read:

Sloetten (Cornelius van). The Isle of Pines; or a Late Discovery of a
Fourth Island in Terra Australis Incognita. London, printed by G. S.
for Allen Banks, 1668. With a New and Further Discovery of the Isle of
Pines, 1668; and a duplicate of the Isle of Pines. 1 vol. small 4to,
calf supr., gilt leaves. A most interesting, rare, and valuable work.

Even against the Editor of the Society the Dowse books are kept behind
lock and key, though he is not under more than ordinary suspicion. So
I was obliged to wait till the next day before my curiosity could be
satisfied. I then found a thin volume, less than one-third of an inch
in thickness, containing two copies of this very tract which the auction
expert had identified as an issue of the "Isle of Pines" by Green, and
a London issue of a second part of the "Isle of Pines," with the name of
Cornelius Van Sloetten, as author. For more than fifty years this little
volume had reposed in this well-known yet almost forgotten [9]library,
and no one had suspected or questioned the nature of its contents.

For full fifty years it had been in the care and at the call of Dr.
Samuel A. Green, who claimed to be an expert on New England imprints of
the seventeenth century, and one of the great wishes of whose life had
been to establish his descent from this very printer, Samuel Green. Two
copies within the same covers, of a tract long sought and of which only
a single example had come to light in two centuries and a half--was not
that alone something of a bibliographical coup?

I read two of the pieces--one of the Green issues and the second part as
printed in England--making a few notes for future use. On returning to
the matter some weeks later I found to my annoyance that every reference
to the Green tract but one was wrong as to the page. Cold, haste, or
weariness will account for a single or possibly two errors of reference,
but to have a whole series--except one--go wrong pointed to failing eyes
or mind. Very much put out, I read the tract a second time and corrected
the page references, carefully checking up the result. Some days after I
again took up the matter, and in verifying my first quotation found that
I had again put down the wrong page number, and was surprised to find
that the correct page was the one I had first given. This proved to
be the case in all the references--except one. A book which could thus
change its page numbering from week to week