DOI Page / Citation link:
https://doi.org/10.11588/diglit.3236#0146
CHAP. XII.—THE WEIGHTS.
81
picked up on the denuded surface of the ground
by the Bedawin, who find them both at
Defenneh, and at sites to the north of that on
to Tell Belim. The whole supply there in two
months was 397 stone weights and 1600 metal
weights. Altogether over 4000 weighings were
performed, of which Mr. Spurrell most kindly
did nearly a thousand. Unhappily, owing to
the weights of Defenneh being found exposed
on the surface, they have on the whole suffered
more by weathering than those of Naukratis,
which are found bedded in stiff mud that has
prevented any access of air, and has preserved
all the products of corrosion around the weight.
The metal weights of Defenneh I have therefore
found it needful to set aside for a special
research; though every one has been weighed
in air and in water, and it did not seem a very
serious matter to work out their original values,
yet a fresh difficulty stood in the way, for they
proved to consist of all sorts and conditions of
alloys, from almost pure copper down to
sulphides of copper and tin, with perhaps other
materials. Probably these alloys are the black
and white bronze of the inscriptions. Under
these circumstances even the specific gravity
fails to show us the internal state of a weight;
and fresh modes of examination must be
worked out and formulated before we can say
anything exact as to the original values. The
whole subject of the modes of patination
requires discussion, and some unexpected
results have appeared; as, for instance, that
weights lose by exudation of their more oxidiz-
able alloys from the whole mass, leaving a
sponge of metal; and the last stage of this
course ends in the complete replacement of the
Aveight by a siliceous or calcareous pseudomorph
of the exact form and polish of the original.
Thus are produced those strange casts of coins,
which appear as if made artificially in plaster
(see " Tanis," p. 40). The whole of the stages
and proofs of this alteration I hope to work
out and state in future: suffice to mention it
here, to show the questions which arise in
dealing with 1600 metal weights. The setting
aside of these, however, does not much affect in
any case the results which are here stated; for
the changes of metal weights are in general so
great that they are worthless for showing the
exact standard, though of as much value as any
for ascertaining the distribution of different
standards and the forms. The limit of 2%
error excluding a weight from the curves of .
results, which I saw need to impose in dealing
with the Naukratis weights before, I have still
maintained as essential to a proper treatment
of the results. I have also found the need of
another limit, when a large quantity of minute
weights are in hand ; it is clear that a weight
of 10 grains cannot (with a given imperfection
of balance) show a standard as accurately as a
weight of 100 grains; hence some limit of
smallness is needful, below which weights fall
under much the same uncertainty as when they
have lost any considerable percentage of their
substance. From comparing the curves of
distribution of the |ths, &rds, and other Kat
weights (as the standard most fully repre-
sented), it seemed best to draw the line a,t 40
grains, thus including in the curves all the Jrds
of the shekels and ka/ts. Perhaps, seeing the
number of erratically low and high £rd kats,
this may be too low a limit, and 60 grains
might be safer; but, at least, if the source of
these stray groups is noted, not much harm
will have been done.
82. The general arrangement of the catalogue
of weights is on the same lines as that of last
year, and therefore the prolegomena need not
be repeated. The only difference is that where
a weight is under the limit of size a bar is put
across the change column, " Ch," to call atten-
tion to it; ■ the other entries in that column
call attention as before to weights which have
suffered more than 2% of change, and the
broken ones are marked B. The numbering
M
81
picked up on the denuded surface of the ground
by the Bedawin, who find them both at
Defenneh, and at sites to the north of that on
to Tell Belim. The whole supply there in two
months was 397 stone weights and 1600 metal
weights. Altogether over 4000 weighings were
performed, of which Mr. Spurrell most kindly
did nearly a thousand. Unhappily, owing to
the weights of Defenneh being found exposed
on the surface, they have on the whole suffered
more by weathering than those of Naukratis,
which are found bedded in stiff mud that has
prevented any access of air, and has preserved
all the products of corrosion around the weight.
The metal weights of Defenneh I have therefore
found it needful to set aside for a special
research; though every one has been weighed
in air and in water, and it did not seem a very
serious matter to work out their original values,
yet a fresh difficulty stood in the way, for they
proved to consist of all sorts and conditions of
alloys, from almost pure copper down to
sulphides of copper and tin, with perhaps other
materials. Probably these alloys are the black
and white bronze of the inscriptions. Under
these circumstances even the specific gravity
fails to show us the internal state of a weight;
and fresh modes of examination must be
worked out and formulated before we can say
anything exact as to the original values. The
whole subject of the modes of patination
requires discussion, and some unexpected
results have appeared; as, for instance, that
weights lose by exudation of their more oxidiz-
able alloys from the whole mass, leaving a
sponge of metal; and the last stage of this
course ends in the complete replacement of the
Aveight by a siliceous or calcareous pseudomorph
of the exact form and polish of the original.
Thus are produced those strange casts of coins,
which appear as if made artificially in plaster
(see " Tanis," p. 40). The whole of the stages
and proofs of this alteration I hope to work
out and state in future: suffice to mention it
here, to show the questions which arise in
dealing with 1600 metal weights. The setting
aside of these, however, does not much affect in
any case the results which are here stated; for
the changes of metal weights are in general so
great that they are worthless for showing the
exact standard, though of as much value as any
for ascertaining the distribution of different
standards and the forms. The limit of 2%
error excluding a weight from the curves of .
results, which I saw need to impose in dealing
with the Naukratis weights before, I have still
maintained as essential to a proper treatment
of the results. I have also found the need of
another limit, when a large quantity of minute
weights are in hand ; it is clear that a weight
of 10 grains cannot (with a given imperfection
of balance) show a standard as accurately as a
weight of 100 grains; hence some limit of
smallness is needful, below which weights fall
under much the same uncertainty as when they
have lost any considerable percentage of their
substance. From comparing the curves of
distribution of the |ths, &rds, and other Kat
weights (as the standard most fully repre-
sented), it seemed best to draw the line a,t 40
grains, thus including in the curves all the Jrds
of the shekels and ka/ts. Perhaps, seeing the
number of erratically low and high £rd kats,
this may be too low a limit, and 60 grains
might be safer; but, at least, if the source of
these stray groups is noted, not much harm
will have been done.
82. The general arrangement of the catalogue
of weights is on the same lines as that of last
year, and therefore the prolegomena need not
be repeated. The only difference is that where
a weight is under the limit of size a bar is put
across the change column, " Ch," to call atten-
tion to it; ■ the other entries in that column
call attention as before to weights which have
suffered more than 2% of change, and the
broken ones are marked B. The numbering
M