NOTATION AND DEFINITIONS
457
that the sign was not duplicated for the plural, although such
duplication was the practice of the Byzantines. That the
sign was merely an abbreviation for the word dpiOpos and no
algebraical symbol is shown by the fact that it occurs in the
manuscripts for dpidpos in the ordinary sense as well as for
dpiOpos in the technical sense of the unknown quantity. Nor
is it confined to Diophantus. It appears in more or less
similar forms in the • manuscripts of other Greek mathe
maticians, e.g. in the Bodleian MS. of Euclid (D’Orville 301)
of the ninth century (in the forms 9 99, or as a curved line
similar to the abbreviation for kul), in the manuscripts of
the Sand-reckoner of Archimedes (in a form' approximat
ing to ?), where again there is confusion caused by the
similarity of the signs for dptdpos and koll, in a manuscript
of the Geodaesia included in the Heronian collections edited
by Hultsch (where it appears in various forms resembling
sometimes £, sometimes p, sometimes o, and once £, with
case-endings superposed) and in a manuscript of Theon of
Smyrna.
What is the origin of the sign 1 ? It is certainly not the
final sigma, as is proved by several of the forms which it
takes, I found that in the Bodleian manuscript of Diophantus
it is written in the form larger than and quite unlike the
final sigma. This form, combined with the fact that in one
place Xylander’s manuscript read ap for the full word, suggested
to me that the sign might be a simple contraction of the first
two letters of dpiOpos. This seemed to be confirmed by
Gardthausen’s mention of a contraction for ap, in the form ap
occurring in a papyrus of a.d. 154, since the transition to the
form found in the manuscripts of Diophantus might easily
have been made through an intermediate form < p. The loss of
the downward stroke, or of the loop, would give a close
approximation to the forms which we know. This hypothesis
as to the origin of the sign has not, so far as I know, been
improved upon. It has the immense advantage that it makes
the sign for dpidpos similar to the signs for the powers of
the unknown, e.g. A' for Svvapus, for kv/Sos, and to the
sign M for the unit, the sole difference being that the two
letters coalesce into one instead of being separate.