THE
HESPERIA 75 (2006)
Pages
OF
INTRODUCTION
357~378
THE
MOLDMADE
BOWL
REVISITED
a Hellenistic
Tracking
Innovation
abstract
The date of 224/3 for the introduction
atAthens
of Hellenistic
is traced in detail, using amodel
The implications of the stratigraphie record at the
our
of the introduction of innovations in
Agora for
understanding
introduction
Athenian
inventions.
and
general,
of this
innovation
in
particular,
are discussed.
The
sparse
repre
bowl in later 3rd-century deposits may indicate
of the new type, but more likely reflects the time that it takes
of the moldmade
sentation
slow acceptance
for
relief bowls
reaffirmed?in
subsequently
shift in the chronology of Rhodian
downward
on recent
moldmade
a
light of recent
amphoras. The process of
of the innovation process based
is reexamined?and
to enter
objects
the
archaeological
record.
INTRODUCTION
relief bowl, familiarly known as the "Megar
The hemispherical moldmade
ian" bowl, is one of the most useful ceramic dating tools of the Hellenistic
a radical
it cannot be
innovation,
technological
period.1 The product of
confused with
earlier pottery. Furthermore,
because of the relief decora
tion that covers its exterior surface, it can be identified even in small wall
on any one site, then, is of par
fragments. The date of its introduction
ticular interest; the more closely we can pinpoint
it, the more accurately
we can date strata that were laid down in the years
just before and after
that
event.
20 years ago, I proposed a date in the span 240-220
for the
of the moldmade
bowl at Athens, positing 224/3 b.c. as the
most likely year in that span; recent revisions in the chronology of Rhodian
type of artifact often found in association with moldmade
amphoras?a
Over
introduction
1.1 am grateful toGerald Finkiel
Lawall
for comments
sztejn and Mark
on an earlier draft of this article. All
were
photographs
Excavations.
Agora
?
The
American
provided
by the
a r??valuation of that date. It is also
to
require
appropriate
in the light of more recent scholarship on the
that hypothesis
and ceramics, and of discussions
of the
relationship between metafwork
bowls?now
review
process
School
of Classical
Studies
of innovation
at Athens
itself.
I.
SUSAN
358
ROTROFF
HISTORY OF SCHOLARSHIP
bowls have attracted scholarly interest since early in the last
and
and
century,
chronology has been the subject of ongoing discussion
revision. Fernand Courby, who devoted a large part of his 1922 monograph
on relief ware to the type,
at the end of the 4th century,
placed its invention
Moldmade
on the basis of both
Systematic
stylistic and contextual considerations.2
contexts at the Athenian Agora was to suggest revi
analysis of Hellenistic
sions, however. In 1934, Homer Thompson
published the contents of five
over the years from the
early
(Groups A-E)
spread
deposits
3rd to the late 2nd century (aswe would now date them). From this data
bowls at Athens.
he extracted a new date for the inception of moldmade
He observed that these bowls had not been found in association with red
Hellenistic
figure pottery in any of the many wells and cisterns explored in the first
years of excavation, a fact that argued strongly for an initial date after the
end of the 4th century. He also noted that there were no moldmade
bowls
in his Group B, the deposition of which he placed, largely on the basis of
numismatic
evidence, in "the first half of that [the 3rd] century, perhaps
about halfway along."3 Taking this date, then, as a necessary terminus post
bowl, he located its inception "in the first quarter
quern for the moldmade
of the 3rd century, probably towards
A lower date, around the middle
its end."4
of the 3rd century, was suggested by
in his 1956 publication of the bowls from the Pnyx.5 He
Roger Edwards
did not make his reason for this revision explicit, but general references to
deposits at the Agora suggest that they formed the basis for his chronology.
in 1975, he reiterated the mid-3rd-century
date, though with less
a quarter century
it
certainty, speculating that
might be "late by perhaps
Later,
or
so."6
a
In the meantime, Virginia Grace had been developing
chronology
for the stamps that appear on the handles of Rhodian transport amphoras.
She had published the single Rhodian jar from Group B in 1934,7 but she
in print upon its date until 1963. In an article focused
did not comment
on the east coast of
on the
primarily
amphoras from excavations at Koroni,
as
the
of
B
Rhodian
she
Attica,
amphoras
Group
presented
jar paradigmatic
at about the time when months began to be named on the stamps (the
2. Courby 1922, pp. 360-361. He
cites the following evidence: (1) bowls
from
tombs
Delphian
said to date
Athena
toire with
to
naic
the production
of which
amphoras,
to have ceased at the
then thought
end of the 4th century.
the end of the 4th century {FdD V.l,
was
their contents
include
pp. 174-176;
in
fusiform
indicating
ung?entar?a,
3.Thompson
stead aHellenistic date); (2) the sup
contemporaneity
posed
with West
production
the 25 coins
of moldmade
3rd century
vase found
disintegrated
now
illegible. Five
Slope ware,
together
with
by comparison
in an Olbian
a coin
with
tomb
of Lysamachos
dated 306-281 (Watzinger 1901,
p. 94); and (3) comparison of an
from
either
which Courby dated in the late 4th to
early
metal
reper
figure of the moldmade
on late Panathe
the Athena
IT'-729,
a
tentatively
1934, p. 332.Most of
the cistern
system
or were
in cleaning
remain: IT'-728
identified
and
at the
time of finding as belonging to the
double-bodied owl variety (Svoronos
=
1923-1926, pi. 22:35-46 Agora
XXVI,
dated
pp. 41-42,
330s-322/317),
varieties
41-43,
but now con
sidered illegible; ZT-303, Greek but
not further identifiable; ST 304, owl
in wheat
variety
ST-305,
wreath
p. 45,
{Agora XXVI,
heavily worn;
wreathed
type
piglet
284-270s,
53),
uncertain
{AgoraXXVI, pp. 43-46, variety 48,49,
51, or 55),
the current
ca. 322/317-270s,
and discussion,
worn.
of these
identification
see Kroll
1974,
For
coins,
pp. 202
203; Agora XXVI, p. 309, under H 16:3.
4. Thompson
1934,
p. 457.
5. Edwards 1956, p. 90.
6. Corinth
7. Grace
VTI.3,
1934,
p. 152.
no. 5,
fig.
p. 202,
(jar);p. 235, no. 77 (stamp).
1
THE
INTRODUCTION
OF
THE
MOLDMADE
BOWL
359
of Rhodian period lia), and she applied Thompsons
terminal
beginning
date of ca. 275 for Group B to this jar.8
The excavations at Koroni had shown, however, that something was
The excavations, aimed at
amiss in the chronology devised by Thompson.
the identification
traces that had long been visible on the
in July of 1960, but they bore fruit out
scale and ambition. The numismatic evidence
of architectural
amere
three weeks
surface, occupied
of proportion to their modest
allowed the excavators to date the site within
the reign of Ptolemy II and
the presence of Ptolemaic
troops during
the Chremonidean War, ca. 267-262/1.9
A problem emerged, however,
in the analysis of the pottery: given the terminal date in the 260s, the
to associate
furthermore
itwith
inThompsons
ought to have resembled the material
Group B,
ca. 275; instead,
more
to
in
in
material
they corresponded
deposited
closely
a
at
turn
to
for
4th
which
terminal
date
the
of
the
Thompsons
Group A,
ceramics
s
the 3rd century had been proposed. This was the first hint that Thompson
pottery dates, and consequently Grace's amphora dates (which were in
on
dates in this period), were considerably
large part based
Thompsons
too high.
In her initial response to the publication
of the Koroni excavations,
Grace argued against a downward
shift,10 but she subse
chronological
an
a lower
for
argument
independent
quently developed
chronology.11
Her method was characteristically
clearheaded and logical. She took as a
were annual and that all Rhodian
given that the terms of Rhodian eponyms
of the 3rd and early 2nd century were known. She then counted
back, allocating one year to each name, from 175, the date she assigned
a
to the closing of the Pergamon Deposit,
large and rich assemblage of
eponyms
stamped amphora handles excavated at Pergamon in 1886.12 In doing so, she
arrived at a date of ca. 240 for the point when months began to be named
on
amphoras, and hence for the amphora in Group B. The new chronol
ogy gave a date in the late 270s for the amphoras at Koroni, somewhat
so much earlier that the
occupation, but not
use aswater containers
not have been serving a
secondary
earlier than the documented
amphoras might
at the
encampment.
application of these revisions to the chronology of the moldmade
a lowered date for its introduction. The absence of the
bowl necessitated
type from Group B provided a terminus post quern of ca. 240 for its incep
tion. Small numbers of fragments had been found in deposits inwhich the
The
8. Grace
1963, pp. 324-325,
no. 7,
333-334,
fig. 1.
9. Vanderpool,
and
McCredie,
For a recent
1962, pp. 56-60.
of the evidence
for the dating
summary
of the war, see Badoud
2003, p. 584.
Steinberg
10. Grace
1963,
11. Grace
1974.
pp. 329-332.
12. Schuchhardt 1895; B?rker 1998;
Lawall 2002.
13.Agora XXII, pp. 9-13. The
date of the introduction of the festival
has now
been more
firmly
established
(Habicht 1982, pp. 106-109).
latest Rhodian
aggregation
20:7), which contained
amphora handle stamped
Grace's
new
were
only slightly later, but the first substantial
bowls was in the upper fill of the Altar Well
(B
a
over
of
20
with
Rhodian
bowls,
fragments
along
amphoras
of moldmade
in the term of Xenostratos
This
evidence
(ca. 217 according to
that moldmade
bowls
chronology).
suggested
sometime between ca. 240 and 220. Within
began to be made inAthens
that span, as I argued inAgora XXII, 224/3 seemed a likely date; this was
a festival initiated
the year of the first celebration of the Ptolemaia,
by
in honor of their benefactor, Ptolemy
the Athenians
III Euergetes,
and
an occasion when
Alexandrian
and thus have provided models
to copy in clay.13
silver bowls might have been on display
for an enterprising
and inventive potter
I.
SUSAN
360
ROTROFF
THE MOLDMADE BOWL AND NEW REVISIONS
TO THE AMPHORA CHRONOLOGY
scholars have been inclined to accept the conclusions
summarized
Many
too
is
above, but others have had reservations, believing that 224/3
precise or
too late a date for the earliest bowls.14 Now, however, Gerald
Finkielsztejn
has proposed yet another revision of the Rhodian
amphora chronology,15
if correct, shows that moldmade
not
later. Finkielsztejns
method
late, if
same
the
eponyms
assumptions?that
which,
bowls were
introduced
is not unlike Grace
at least that
s.He makes
the
served a one-year term and that all
known?and
he too counts back from a
eponyms are probably
as a datum
fixed point. He discards the Pergamon Deposit
point, however,
to
his
instead
known
destruction
dates of
sequence
anchoring
historically
various cities (mostly in the Levant). He also points out what is clearly
a
fact: that the period from the end of the Pergamon Deposit
worrying
IV
(175, according to Grace) to the destruction of Corinth
(146)?period
a span of 29 years, but contains
in the Rhodian
chronology?covers
only
14 known eponyms.16 Counting
back from the destruction
of Corinth
Rhodian
in 146, he arrives at a date of ca. 160 for the beginning
15 years later than Grace s estimate.
IV, some
of period
shift of course has an impact on earlier periods aswell.
ca. 205-175, while
dates
assigned period III to the years
Finkielsztejn
to 234-199.17
it between 198 and 161, and period II is shifted from 239-206
This
downward
Grace
so far, have stood up well to
persuasive, and,
expert scrutiny18 Furthermore, Christian Habicht has reviewed the Rhodian
men who
as eponyms and found it
stamped
epigraphical evidence for the
Finkielsztejns
arguments
are
to be largely consistent with the new chronology, although his work reveals
to Finkielsztejns
the necessity for some minor adjustments
dates.19 My
aim here, however, it not to discuss the details of the new chronology, but
rather to investigate the implications of this downward
shift for the date
of the introduction of the moldmade
bowl at Athens.
Finkielsztejns
chronology places the Rhodian
amphora in Group B,
which has always served as the terminus post quern for moldmade
bowls,
in the span 233-220. The handle gives only a fabricant, but Grace long
If this conjecture is
ago suggested restoring the eponym as Philokrates.20
ca. 226)
the
lia
be
the
middle
of
date would
around
correct,
period
(perhaps
in Finkielsztejns
scheme,21 and Group B (thought to be without moldmade
the latest
bowls) must have been discarded after that date. Xenostratos,
(with many fragments of moldmade
bowls), falls
eponym in the Altar Well
either late in period lib or early in period lie; Finkielsztejn
places him in
ca. 211, or
revisions, then, suggest that the
perhaps later. Finkielsztejns
introduction
of the moldmade
14. E.g., ?zyigit
Massa
1992,
bowl
1990, p. 96;
p. 33; Hausmann
1996,
p. 105.
15.Finkielsztejn2001.
16. Finkielsztejn 2001, pp. 127-129.
17. For
a summary
of the changes,
see Finkielsztejn 2001, pp. 196-197,
tables
22.1
and 22.2.
should be bracketed
between
ca. 226
18. Lund 2002; Lawall 2003;
Badoud 2003.
19.Habicht 2003.
20. Grace
21.
1963,
p. 326,
is
the stylistic group, however,
refrains
and Finkielsztejn
conjectural
from assigning
annual dates to most
within
n. 16.
most
of the
assigns
Finkielsztejn
of period
II to
and,
eponyms
subperiods
within
these, to smaller
stylistic groups.
The position
of any one eponym
eponyms
the date,
in this
estimate
period. His
can be
extrapolated
however,
from one of his tables (Finkielsztejn
2001, p. 191, table 18).
of
THE
INTRODUCTION
OF
THE
MOLDMADE
BOWL
361
and 211, making an initial date of 224/3, while not impossible, seem less
likely than it once did.
A new and surprising discovery, however, now comes into play. Despite
assertion to the contrary, Group B does contain amoldmade
Thompsons
bowl: a small fragment (max. p. dim. 3 cm) that can be attributed to the
of Bion (Fig. I).22 It is hard to explain how Thompson
missed
Workshop
an
intrusion and thus not worthy of mention.23
it; possibly he considered it
It comes from lot ET 105, a small amount (one bag) of pottery from the
drawshaft and the passage that connected itwith the northern chamber of
the cistern system.24 Such a small fragment could, of course, have found its
way into the deposit in a number of ways, but let us assume for the moment
Such a supposition
receives support
which
include
(P 10:2),
considerably
bowls (Fig. 2)25 and, as the latest datable
that it is part of the original deposit.
from the contents of another cistern
larger fragments of two moldmade
object, a Rhodian amphora handle naming the eponym Philokrates?the
same eponym
on the
conjecturally restored
Group B amphora.
22. P 34577. For theWorkshop
Bion,
see
Agora
the rim pattern
Agora
103,
XXII,
XXII,
pp. 26-27.
of the fragment,
nos.
pp. 49,55-56,
of
For
cf.
43,
99,
75,98.
pis. 7,17,18,
some later Hellenistic
23. While
it comes
from
elsewhere
in the
fig.
12.
25. P 19705 (AgoraXXII, p. 45,
no. 3,
pis.
1, 92),
about
a
quarter
bowl;
pinecone
XXII,
two fragments of a type 27 D lamp
(L 5215) from the south chamber.
24. See plan inThompson 1934,
p. 331,
material has been identified in the
deposit,
a
of a
system:
fragment
2nd-century
the
northern
from
chamber;
plate
of a
and P 19720
(cf. Agora
no. 252,
pi. 50), five
a small
part of
preserving
pp. 75-76,
fragments
the lowerwall of a figured bowl with a
to be attributed
scene, probably
hunting
to
For Workshop
A.
A, see
Workshop
pp. 28-29.
Agora XXII,
SUSAN
362
I.
ROTROFF
Figure 3.Moldmade bowls from
cistern N 21:9 (P 34499, P 34498,
P 34967, P 34968). Scale 2:3
No moldmade
bowls were found in the substantial upper fill of well
B 13:8, which contains amphora handles stamped with the names of the
and Kallikratidas
I, probably dating slightly after
eponyms Xenaretos
Philokrates.
of
four
bowls, however, occur in cistern N 21:9
Fragments
(Fig. 3),26 which also yielded the handle and neck of an amphora manu
factured by the fabricant Pausanias II, whose activity can be placed largely
within period Ha on the basis of the eponyms with which he is associ
ated (ca. 234-218);
and two small rim fragments came to light in cistern
L 17:7,27 along with an amphora handle naming the eponym Aglokritos
ca. 216). In view of these
(period lib, perhaps
findings, the single frag
ment
in Group
B seems unexceptional,
and we may
26. P 34498, P 34499 (fragments
preserving
the medallion
and
lower
wall of one floral bowl and about a
third of another floral bowl), and
P 34967, P 34968 (three small frag
ments of two figured bowls).
27. Lot
decoration
rim pattern
a
O
cautiously
184. The
that
remains
on one
and
palmette
it.
flanked
the
view
it as
trace of
only
is part of the
the tip of
fragment:
tail of a dolphin
that
THE
INTRODUCTION
OF
THE
MOLDMADE
BOWL
363
that the fragments of amphoras
part of the original deposit. Assuming
and moldmade
bowls in these contexts are approximately
contemporary,28
a date of 224/3 for the introduction of the moldmade
thus
bowl atAthens
remains a possibility, even in light of the downwardly
shifted amphora
chronology.
MOLDMADE BOWLS IN THE ATHENIAN
STRATIGRAPHIC RECORD
A precise date for the introduction of the moldmade
bowl will probably
us.
more
is perhaps
What
important is to track the intro
always elude
duction of the new type with an eye to how it is represented in deposits
of (approximately) known date. Such a process reveals that, although the
bowl may loom large inmodern
moldmade
eyes as an im
archaeological
portant dating tool, it loomed rather small in ancient debris (and perhaps
in daily life) in the first generation or so of its production.
one exception,
discussed below, wells and cisterns dated by
With
Rhodian amphoras of period II and the earlier part of period III (Illa-b,
ca. 182) contain very small numbers of moldmade
bowls, usually
one and six vessels (see Tables 1 and 2);
between
fragments representing
the average of bowls per deposit for the period is about four. There are
or
even some
on in
early in
deposits with amphoras dating well
period II
ending
no bowls at all. Two
nearly complete jars of the
at
B
the
bottom
of
well
13:7
show that itwas filled no
Aristonidas
eponym
as established on the basis of
date of Aristonidas
earlier than 208/7?the
no moldmade
fragments exist among the 57
epigraphical evidence29?but
the nearby drawshaft
identified fine-ware pieces in the deposit. Likewise,
B 13:3, with ca. 25 identified fine-ware vessels, contained no moldmade
period
III that contain
was
no earlier than the first years of the 2nd
although it
deposited
Given the consider
century (the date is given by the eponym Dorkulidas).
able number of deposits with contemporary handles inwhich moldmade
bowls,
a matter of
likely that their absence here is
source
nature
the
of
of
the
the
material, rather
chance, probably reflecting
an
so
than
indication that production began
late.
28. This
seems
safest assump
of any reliable basis
a
to formulate
rule
upon which
general
about the relative ages of transport
tion,
the
in the absence
and the household
amphoras
found with
them. Ethnographic
pottery
studies
of the use life of pottery suggest that
different
ity, and,
longer
different
shapes enjoy
longev
in
general,
large shapes last
than small ones (Foster 1960;
Longacre 1985; Rice 1987, pp. 296
299). Given these findings, one might
therefore
expect
that amphoras
would
be older than the drinking cupswith
were
they
considered
which
vessels
however,
were
vessels.
storage
have a different
shipping
pected
it seems
are present,
bowls
found.
The
in these
mostly
water
Transport
large
studies,
jars and
amphoras
as
serving
and may be ex
containers,
a different
to follow
of
pattern
function,
breakage. They might be discarded
immediately upon being emptied of
their
contents
and
thus have
a short use
life; or theymight be retained for
case
use, in which
they
secondary
would
the
of the water
enjoy
longevity
in the
and storage vessels
ethnographic
data.
In reviewing
pottery
from
the
Palatine inRome, Pe?a (1998) found
of amphoras
similar percentages
to the
tablewares
could be dated
that
and
same
phase
(12.8%
and
re
13.5%,
spectively), and that a slightly larger
of tablewares
percentage
versus
12% for
(15.1%
was
residual
amphoras).
This finding suggests that amphoras
need
not,
ered
older
which
logical
they
as a
than
rule, be consid
general
the tableware with
are associated
in archaeo
deposits.
29. Habicht 2003, p. 546.
1. AGORA DEPOSITS
TABLE
Moldmade
Deposit
and
Wells
H
FOR WHICH
A PERIOD
Bowls
II RHODIAN
AMPHORA
PROVID
Handles/
Rh
Rhodian
Pe
2/1
Ha
2/2
Ha
Kallikratidas
I ep (SS 7764, with
another handle from same jar)
Xenaretos
ep (SS 7269, with another
handle from same jar)
8/8
Ha
Pausanias II fabd (SS 10015,
neck + handle)
1/1
Ha
ep (SS 14279, with another
Aglokritos
handle perhaps from same jar)
7/4
?b
Xenostratos
6/3
Latest
Rhodian
Amphora
Cisterns
16:3 (Group B)
1 small fr (P 34577, Workshop
of Bion)
I fab/[Philokrates?]
ep
(SS 370)
Zenon
nearly complete
P10:2
frr of 2 bowls
ofWorkshop
B 13:8, upper fill
none
N21:9
Philokratesep (SS 7771)
(P 19705, P 19720,1
A)
(at least 65 fine pots represented,
to early 3rd
Classical
century)
frr of 4 bowls
largely
P 34499, P 34967,
(P 34498,
P 34968)
L17:7
2 small rim frr (Workshop A)
(lotO 184)
B 20:7 (Altar
Well),
at least 22 bowls
(P 17511-P
in sizable deposit
17514, P 17622,
ep (SS 9663)
P 34506, lotNN 377)
upper fill
lia
lib
ep (SS 9660)
Xenophantos
lib
H 6:4
POU fill
6:1 nearly whole
(Workshop of Bion)
at least 5 (P 401,4879,
lot A 216).
+ frr of
Xenostratos
ep (SS 148)
8/2
lib
N 21:4
2: 1 ca. half preserved
(lot EE 18,Workshop
cast) + 1 fr
Xenostratos
2 handles)
+
ep (SS 8934, neck
9/8
lib
(Satyr
lower fill
Cistern),
B 13:7, bottom
fill
L 19:2, lower fill
L-M
19:1
none
(P 16221, metal
of Bion)
(at least 57 fine pots
5 frr of 1 bowl, slightly worn;
fine ware (lot AA 16)
frr of at least 7 bowls,
Three
represented)
small amount
small amount
nearly complete
jars of
5/4
Mution ep (SS 7582) and
Aristonidas ep ( SS 7581, SS 7583)
of
of fine ware
Theuphanes II ep (SS 9579)
lie
Ik
3/3
Ik
Pausanias
II ep (SS 9662)
3/3
Ik
Sochares
ep (SS 3784+3811)
29/24
lib
(P 17085, lotX 164)
Fill
N10:2
Total
*
none
(only 3 frr of fine pottery
minimum
of 47 bowls
in deposit)
in 12 deposits
Deposits with period II Rhodian amphoras as the latest datable objects but not included in the table: A 18:1 (contains a substantial 2nd-century
O 20:3, level IV (superimposed over O 20:3, level III, with a period III handle; see Table 2).
a
to Finkielsztejn 2001, p. 191, table 18, with
According
adjustments based on Habicht 2003.
From Grace and Sawatianou-P?tropoulakou
1970; Grace 1974; Agora XXIX, summarized in Finkielsztejn 2001, p. 191, table 18.
c
Dates in brackets give the position of the eponym in
table 18; spans are his dates for the period in question.
Finkielsztejns
For eponyms associated with Pausanias, see J?hrens 1999, p. 17;
Finkielsztejn 2001, p. 76, n. 55.
e
to Habicht 2003, p. 546.
According
V. R. Grace (pers. comm.).
83/63
intrusion); E
THE
H 16:3
P10:2
[226]
B13:8
[226]
the
terminal
N10:2
N21:9
L17:7
OF
THE
MOLDMADE
IB 20:7 I H 6:4
N21:4
[223] [234-218] [218]
[216]
[210]
[211]
[211]
Deposit [approximate date of latest amphora]
Figure 4. Number of moldmade
bowls in deposits for which an
amphora of Rhodian period II
gives
INTRODUCTION
date
BOWL
365
B13:7
L19:2
[208]
[203]
L-M 19:1
[199]
in period II with a substantial representation
of
only deposit
bowls is the upper fill of the Altar Well
terminus
Its
(B 20:7).
I (ca. 210) and
post quern is supplied by handles stamped by Xenophantos
Xenostratos
(late period lib or period lie, ca. 211 or later). As noted above,
the fill contained many moldmade
fragments, representing at least 22 bowls,
a
a
shattered
and
battered;
badly
glance at Figure 4, graphic representation
of the numbers inTable 1, makes the anomaly of these statistics glaringly
The
moldmade
is located in the industrial district southwest of the Agora
fill
and
the
included many fragments of terracotta figurines and one
square,
mold (T 2277), aswell as a possible kiln support (P 17515).30 A workshop
may be the source of at least some of this material, a possibility that would
clear. The well
of moldmade
help to explain the high concentration
ments from two different
workshops?Workshop
bowls
A
(although frag
and theWorkshop
of Bion?are
30. For
son 1959.
the terracottas,
For
the kiln
see
support,
present).
Another, and inmy view more likely, explanation is that the fill entered
the well considerably
later than the date of the latest amphora handle.
Thomp
see
Papadopoulos 1992, pp. 209,211,
p. 360,
no.
1214,
21112 (AgoraXXIX,
fig. 75, pi. 90).
32. Stored in lots NN 377,399,401.
there is no join between
the
Although
two groups
of fragments,
several details
that they come from the same
suggest
bowl. The
shape of the rim, the rim
pattern,
and
the distance
of the rim
below
the lip are the same in
pattern
both groups. Moreover,
both sets of
sherds have a fugitive gloss verging
from black into a dull purplish red on
the rim; display
similar patterns
in the gloss; and share
cracking
identical
fabric color.
of
an
some
this is always a possibility,
tangible support for it in this
from the fact that joining pieces of the same pot31 have been
in the Altar Well
and in the lower fill of cistern B 18:13, some 36
case comes
found
:a.
fig. 4, pi. 51
31. P17517+P
While
m to the north, which contained an
amphora handle dated to the span ca.
198-187 by the Rhodian fabricant Aristion. There are also fragments of
what appears to be the same moldmade
bowl in these two deposits (five
If the source of
fragments from B 18:13 and nine from the Altar Well).32
those two fills was the same, the Altar Well may also have been filled in
the early 2nd century; and in that case the larger number
bowls is easily understood.
Moldmade
taining Rhodian
b, ca. 198-182);
of moldmade
con
to be sparsely
represented in cisterns
handles dating in the earlier part of period III (Ilia and
four bowls per deposit remain the average (Table 2). The
bowls
continue
larger sample provided by fills of this period suggests that the industry was
well established, but the nature of these deposits, extensive and subject to
disturbance, makes their evidence difficult to evaluate. Some contain later
TABLE 2. AGORA
Moldmade
Deposit
and
Wells
A PERIOD
FOR WHICH
Bowls
III RHODIAN
and Molds
Latest
AMPHORA
Rhodian
Amphora
PROVID
Handles/
Rhodian
Rho
Peri
Cisterns
B13:3
O 20:1,
DEPOSITS
none
lower fill
among
ca. 25 identified
3 frr of 3 bowls
B 18:7, upper fill
6 frr of at least 3 bowls
B 13:1, lower fill
1 fr among
B 18:13,
22 frr of at least 8 bowls
lower fill
O 20:3,
fine-ware
(P 20539-P
ca. 44 identified
(P 9399,
K 18:2
4 frr of 3 bowls
O 20:2
13 frr of at least 6 bowls
B 20:2
at least 11 bowls,
1 mold
M 18:10
at least 50 bowls,
1mold
N 20:7
at least 49 bowls
Q_12:l
none,
Sostratos
(lot 110 336)
pots
Aristion
lots Q 21,23)
Kratidas
(P 17027-17030,
fab (SS 9901)
II fab (SS 8133)
ep (SS 7085)
Hieron I ep (SS 8020)
lot Q 29)
(P 14186,
I fab (SS 6523)d
Aristokrates
lot Y 3)
ep (SS 8074)
ep (SS 10145)
Damokrates
(lots NN 399, 401)
(P 14166, P 20581,
ep (SS 6428)
Agloumbrotos
lot NN 787)
20541,
fine-ware
5 frr of 5 bowls
fills III and IV
Dorkulidas
pots
(lot Q 36) in tiny deposit
lot NN 268)
Kleukrates
ep (SS 9383)
Athanodotos
almost no fine ware
ep (SS 14296)
2/2
Ilia
1/1
Ilia
2/2
Ilia
3/2
Ilc-II
5/5
IIIa-b
4/2
IIIa-b
13/10
13/8
Illb
Illb
17/7
Hid
14/5
Hid
Hie
Xenophon
ep (SS 7898)
9/4
Amuntas
fab (SS 11122)
2/1
Illc-IV
Fills
T-U21:l
310 frr of at least 68 bowls,
K7:l
(P 34485, lotsEA 168-170)
5 frrof 5 bowls (P 34500, lotH 62)
including
one
early long-petal
fr
Agloumbrotos ep (SS 15053)
Damokrates I fab (SS 15046)
14/11
37/28
Ilia
IIIa-b
Sodamos ep (SS 15018)
24/16
Ilia
Hieron I ep (SS 10831)
122/57
Illb
Agloumbrotos
ep (SS 4494)
Aristion fab (SS4511)
Damokrates
T21:l
at least 23, with 3 frr of
developed
long-petal bowls (P 34093,
P 34487 from lots EA 149-154,178);
evidence of disturbance
(a globule lamp)
Q8-9
(fill of Square Peristyle)
At
H-K
40 inventoried frr, 6 molds
(no complete tally of uninventoried
class and
fragments has been made),
including M Monogram
early long-petal frr (P 21048, P 21049, P 22858, P 24819,
12-14
(Middle Stoa building fill)
least 190 bowls,
long-petal
including M Monogram
fragments (P 20204, P 31696,
class and early
P 31740)
Ilia
Ilc-III
Athanodotos
I fab (SS 4475)
epe
Ilc-III
1498/885 Hid
P 31698, P 31699, P 31701)
Total
minimum
of 473 bowls
and molds
in 17 deposits
1780/1046
a
to
Finkielsztejn 2001, p. 192, table 19.
bAccording
From Grace 1985; Agora XXII; Agora XXIX, summarized
in Finkielsztejn 2001, p. 192, table 19.
in ca. 198, but Habicht (2003, p. 557) has established that
Finkielsztejn placed Dorkulidas
Theuphanes
"another (very close) slot."
For the eponyms linked with Damokrates,
see J?hrens 1999, p. 40.
e
For detailed analysis of the amphora handles in theMiddle
Stoa building fill, see Grace 1985.
II held office that year.
According
to Finkielsztejn
(pers.
THE
INTRODUCTION
Figure 5. Average number of mold
made bowls per well or cistern in
deposits dated by Rhodian
of periods II and III
OF
THE
MOLDMADE
BOWL
llla-b
amphoras
Rhodian
367
llld-e
period
types (early long-petal bowls inT-U 21:1 and Q8-9;
fragments of theM
are
in
these
class Q8-9),
but whether
intrusions, evidence that
Monogram
those types appear earlier than was thought, or a hint that the deposits are
significantly later than the latest amphora handle is impossible to know.
bowls are the clear earmarks of well
Large numbers of moldmade
and cistern deposits dated by amphoras of Rhodian period Hid (which,
to Finkielsztejn,
begins ca. 175). Here the average is 27.5 per
according
an enormous increase over what had gone before and confirmation
deposit,
bowl was inwide use inAthens
that, by that time, the moldmade
(Table 2,
Fig. 5).
TRACKING AN INNOVATION
clarifying the approximate date of the introduction of the moldmade
account above leaves a number of puzzles unsolved. The origin
the
bowl,
remain
of the models
and the mechanism
that brought them to Athens
While
uncertain.
Nor
do we know how nonelite Athenians
became sufficiently
aware of those costly models
to desire inexpensive copies, or how potters
of the
acquired examples of them for copying. The sparse representation
new invention in
contexts
in
40
the
first
years of
archaeological
deposited
at
A
also requires explanation.
closer look
the details of the
production
process, and some help from innovation
theory, may throw
on these issues.
light
more recent
Using
insights derived from the study of
technological
D.
has
that can be used to track the
A. Spratt
advances,
developed models
innovative
innovative process.33 Although
the analogy between innovation in industrial
and in earlier and simpler societies (and economies)
is not exact,
the models can help to clarify some aspects of the ancient process. Spratt
societies
more momentous
bronze metal
chiefly of
inventions?agriculture,
the
coke-fired
blast
furnace?but
his
models
should
be applicable
lurgy,
to lesser innovations as well, and it is illuminating
to apply them to the
of the moldmade
bowl.
development
writes
33. Spratt 1982,1989.
SUSAN
368
I.
ROTROFF
Spratt breaks the innovation process into six parts: (A) discovery, (B)
and distri
invention, (C) development,
(D) investment,
(E) production
or some
and
obsolescence.
be
These
bution,
(F)
stages may
sequential,
may occur at the same time, and the duration of each phase may differ
from case to case, but all innovations pass through all of these stages, and
we can
of the moldmade
bowl
classify the events in the development
accordingly.
defined by Spratt as "an addition to the body of technical
Discovery,
or scientific
occur simultaneously with invention, or it
knowledge,"34 may
have
taken
before.
The latter is the case with the moldmade
may
place long
technical process of creating a clay vessel in amold was at least
Bronze Age. It was familiar to Athenian
potters of
the Archaic
and Classical periods,35 and had been newly applied to the
manufacture
of lamps in the third quarter of the 3rd century.36 Athenian
bowl. The
as old as the
Aegean
potters had also, since the 5th century, regularly created the round bottoms
of shallow cooking pots by pressing clay into a convex bat (in essence a shal
low mold) and then throwing the upper part of the vessel on the wheel.37
artisans were skilled at taking impressions
As well, Athenian
of
metal
variety
objects and casts of metal objects.38 Although
would
be needed before this technology
of moldmade
bowls, the basic concepts
from awide
adjustments
could be applied to the production
and practices were firmly in place
the second half of the 3rd century.
defines
the second step, invention, as a mental process: "the
Spratt
use of technical
our case, this
perception of the practical
knowledge."39 In
a potter s
constitutes
mold
that
the
could
be
process
insight
adapted to
on the
bowl with relief decoration
the manufacture
of a hemispherical
we
turn
to
must
another of Spratt's points: that
exterior. Here, however,
long before
other developments,
such as social, economic, or military
changes, have
ones
on
as much of an
as
innovation
and on
impact
purely technological
the success and acceptance of the resultant product. In the case under
itwas the presence of new models and an incentive to copy
consideration,
the realm of historical, political, economic,
them?factors
that lie within
and social development?that
led to the invention.
Let us examine this step of the process inmore detail. First, the models.
Their appearance is preserved for us in moldmade
bowls made in molds
taken directly from metal bowls, or from casts of metal bowls (Figs. 6-8).40
The motifs and their arrangement?alternating
leaves, petals, and floral
3rd century a staple
the
later
tendrils springing from amedallion?were
by
of the Hellenistic
artistic koine, and it is not possible to know where the
metal models were produced. Although many, including myself, have argued
37.Agora XII, p. 35;Agora XXXIII,
34. Spratt 1982, p. 80; 1989,
p. 246.
35. E.g.,
for rhyta
in the 5th
and
4th centuries (Hoffmann 1962; Kopeke
1964, p. 56, nos. 334-336, Beilage 8)
and plastic lekythoi in the 4th (Trumpf
Lyrizaki 1969;Williams 1978).
36.Agora IV, pp. 129-131, types 42 A
and 42 B; seeAgora XXIX, p. 505, for
revised
dates.
It has also been argued
forthcoming.
that this process was used for the pro
duction
of many
earlier glass vessels,
in
particular
for Achaemenid
glass
bowls of the 5th and 4th centuries
(Lierke 1993). I thankMarianne Stern
for bringing
attention;
Macedonian
this
information
she adds
glass
that Rhodian
bowls
to my
and
of the second
half
of the 4th
been made
century
in this way
appear
to have
(Triantaphyl
lides 2000). It is possible thatAthenian
potters might
information.
have
had
access
to this
38. Thompson 1939; Reeder 1976.
39. Spratt 1982, p. 80; Spratt 1989,
p. 246.
40.Rotroffl982.
THE
INTRODUCTION
OF
THE
MOLDMADE
BOWL
369
Figure 6. Bowl made in amold taken
directly from metalware (P 5813).
Scale 2:3. Drawing
P. de Jong
Figure 7. Bowl made in a broken
mold (P 16221). Scale 2:3
as a source,41 other centers of
such as Sicily or
production,
or even Athens
to
claim
the
honor.42 This
itself,
Italy,
might
lay
the potters motive
for replicating the metal bowls in clay at this
for Alexandria
southern
makes
particular
41. Zahn 1904, pp. 412-415;
1934,
Thompson
Quarles
Parlasca
van Ufford
1953,
1955; Hausmann
19-22;
Agora
42. Nenna
pp. 135-136)
an Alexandrian
form
p. 455;
13-15;
1959,
pp. 6-9.
and Seif el-Din
model,
of the earliest
30-33).
(2000,
reservations
noting
Athenian
difficult
and T2.2,
pp. 185-194,
(types T2.1
It is also noteworthy
that
pp.
XXII,
express
more
an outturned
with
rim)
(hemispherical
rare in the
is relatively
large corpus of
bowls
Hellenistic
faience
Egyptian
Byvanck
pp.
moment
about
that
bowls
the
pis.
their
no
of faience vessels preserves
catalogue
close parallels
for the decorative
scheme
of the earliest
versely,
that
Athenian
bowls.
scheme?alternating
Con
to fathom.
leaf separated
petal and acanthus
by
on a 3rd-cen
floral tendrils?appears
tury silver bowl of Sicilian
provenance
or South
and probably
of Sicilian
Ital
ian manufacture
(Bothmer
1984-1985,
no. 92) and on the
pp. 54-55,
pyxis
in the Treasure
of Taranto
(Pfrommer
1987, pi. 33).
lotus
SUSAN
I.
ROTROFF
'Vit?*;
The
relationship
considerable attention
between
metal
and ceramic
vessels
has
received
in recent decades. Michael Vickers and David Gill
have argued that a large majority of fine wares are close
replicas of metal
on
their
work
has
the figured pottery
concentrated
originals.43 Although
of the Archaic
and Classical periods, they have also
applied the principle
to the ceramics of other times and
bowls
places, including the moldmade
of the Hellenistic
as
take
it
axiomatic
that
such
imitations
period.44 They
Figure 8. Bowls with three different
rim patterns: dolphin, floral, inverted
ovolo (P11436, P 27436, P 34969).
Scale 1:1
should be the norm, and that there would
versions of precious vessels. While
market
always be a demand for down
this may be so, it fails to explain
in preference to others, and
why they were
why certain forms were imitated
imitated at one time rather than another.
Margaret Miller has explored the forces behind such imitations in her
discussion of the introduction of
imitations, adaptations, and
black-gloss
derivatives of Persian metalwork
in the 5th century.45
the influx of
Although
Persian booty at the end of the Persian wars made eastern metalwork more
inAthens, Miller discounts
its presence as a sufficient
explanation
for the ceramic phenomenon,
for it does not explain why Athenians would
be receptive to those models?or,
more
would
pointedly, why Athenians
want to
of
the
dinnerware
of
their
enemies.
Miller
finds
acquire replicas
the answer in the social and
climate
of
the
She
political
proposes
period.
that perceived
similarities between Athenian
and Persian elites would
have made Persian vessels attractive to
as
wealthy Athenians
"ready-made
of
aristocratic
wealth
and
and
in the 5th
that
later
signifiers
authority";
with
the
of
the
Athenian
Athens'
"middle-class
century,
growth
empire,
imperialists,, constituted amarket for ceramic copies of these same symbols
of imperial power.46
The social and political situation two centuries later was different, but
the question is the same: why these cups, at this time? The
that
hypothesis
the models were Alexandrian
some
an
answer.
toward
goes
way
providing
The departure of theMacedonian
from
the
Piraeus
in 229 opened
garrison
the way for renewed relations with the Ptolemies,
and highly placed Athe
nians surely cultivated those ties
silver was the
assiduously.47 Alexandrian
ware
of
Persian
the
of
opposite
booty:
potent political friends rather
luxury
than of enemies. Public display of Alexandrian
silverware?a
festival such
as the Athenian
Ptolemaia may have provided amechanism
for such dis
have given the ordinary Athenian
a view of these elite
play?would
goods,
which otherwise circulated
within
the
of the wealthy.
households
only
visible
43. Vickers andGill 1994.
44. Vickers and Gill 1994, pp. 178
180;
1986,
see also Vickers,
Impey,
and Allan
pis. 26,27.
45.Miller
46.Miller
1993.
1993, pp. 137-141.
47. For Athens'
Ptolemies
at this
relations
time,
with
the
see Habicht
1982, pp. 105-117; 1992, pp. 74-77;
1997, pp. 179-185.
THE
INTRODUCTION
THE
OF
MOLDMADE
BOWL
371
The copying of those goods, however, requires more than admiration
from afar. The ceramic bowls made in molds
taken directly from metal
ware show that the potter had access to a set of metal cups (or to a cast of
such a set). How did they come into his hands? It may be that someone
an interest
in supporting
close ties with
the Ptolemies
commis
him
the potter to make copies, and furthermore
with
supplied
the models necessary to carry out that commission.48 A similar situation
occurred half a century earlier, when an Athenian
potter produced a series
of hemispherical
with
the
relief
drinking cups
portrait of King Ptolemy
with
sioned
on the floor.49 In my
imagined scenario, the first imitations of metal
ambassadors
cups may have been made at the instance of Ptolemaic
relief
or of
supportive of strong links with the Ptolemaic
kings, perhaps
to serve as gifts or souvenirs on the occasion of the Ptolemaia. Unlike
the
earlier cups with portraits of the Egyptian king, however, the moldmade
bowl was not a short-lived phenomenon.
His commission
finished, the
Athenians
to devise a method
it worth his while
for the large-scale
of
is
of
the
silver
It
here
that public display
cups.
adaptations
production
a
role: the placement of elite tableware before the public
may have played
eye could have created a demand for copies, providing an economic incen
found
potter
this is only a
production of clay relief bowls. Although
a potter
it
has
the
of
both
how
advantage
plausible narrative,
explaining
access
to
would have had
imported silverware, and why there would have
tive for continued
been amarket
for copies.
to
invention?the
According
Spratt's model,
a mold to create
relief bowls?is
hemispherical
idea of throwing within
followed by development
and investment.50 At the development
stage, the idea is put into practice
and through experiment, trial, and error, aworkable process of manufacture
is created. In our case, itwas the open shape of the cups that presented a
vessels, such as rhyta, plastic
special challenge. Earlier Athenian moldmade
were closed shapes.
were made by
lekythoi, and lamps,
They
spreading clay
into a two-piece mold, and the appearance of the interior, which would
not have been visible in the finished vessel, was unimportant. The interior
of an open shape, however, is visible and must be smooth. The existing
was
inadequate here, since itwould have taken too much time to
technique
smooth the inner surface by hand. The solution was to smooth the interior
on the wheel, which
led to the adaptation of the existing
new
a
to
the mold.
technology
practice, that of throwing the bowl within
This process in modern
ceramic manufacture
is known as jollying and is
that makes it possible to throw
accomplished with the use of a machine
mechanically,
a vessel with
a
48.
mission
It is also
was
possible
to an Athenian
that
the com
silver
to
replicas
produce metal
add another
such cups. This would
to the process,
but it has long been
worker,
realized
worked
1966,
that metalworkers
in close
association
p. 55; Barr-Sharrar
of
(Thompson
and it
1990),
too developed
potters
probably would not have been difficult
difficulties
for a potter
casts) from
forced
to borrow
metal
a metalsmiths
vessels
(or
workshop.
49. Rotroff 1988;Agora XXIX,
step
and potters
a mold. It is
thick vessel wall within
likely that
some such device, or at least used a
uniformly
ancient Athenian
pp. 277,281-282,
pis.
nos.
(2002)
Saramago
a potent
presents
s novel
by
wheelmade
a
potter
pottery
and produce moldmade figurines in
considerable
numbers.
While
fictional,
it brings home the challenges involved
333,376,
33,36.
50. Jos?
encountered
to abandon
The Cave
account
of the
new
in
methods
undertaking
working
and in the manufacture
of new products
at a
large
scale.
SUSAN
372
handheld
no more
template, enabling them to make
than 3 mm thick.51
bowls with
I.
ROTROFF
very uniform walls
that could happen, however, the mold had to be made. Unlike
for later moldmade
bowls, the first molds were created as direct
at
from
least
three metal bowls (or casts; see Figs. 6-8).52
clay impressions
was
The impression
taken by completely encasing the decorated surface
Before
the molds
in clay, probably the highly refined clay routinely used for
as the mold dried over its convex model,
figurine molds; to avoid cracking
itmust have had a very small coefficient of shrinkage. The exterior of the
of the model
mold was
on the wheel,
either in the wet or leather-hard
stage,
shaped
since it had to be perfectly smooth and regular if itwas to be centered on
it.
the wheel and bowls were to be thrown within
of such amold must have been a delicate process, for
itwould have been difficult to remove the model from the enclosing clay
mold without damaging the impressions left on the mold s inner surface.
The
construction
The potter could have solved this problem by slicing the mold in half, creat
a
aswas done for
ing two-piece mold,
lamps and figurines; but the absence
of a seam in the one nearly complete bowl that was made in such amold
was not
(Fig. 6)53 shows that this solution
applied. These prototype molds
were
use after
as
to
in
continue
sufficiently prized
breakage and mending,
a
on the surface of P 16221 attests
diagonal ridge
(Fig. 7).
It required a costly
The prototypical process had some disadvantages.
or
a
an
a
cast
of
and
metal
bowl
such
bowl)
exacting procedure
patrix (a
of mold
a one-time
commis
adequate for completing
have been difficult to adapt to a larger scale of production.
it afforded no possibility of innovation in the decoration of
creation. While
sion, itwould
Furthermore,
the resultant bowls. In any event, a different technique soon replaced it. In
this process, a plain wheelmade
bowl served as amold blank, and its inner
surfaces were decorated with a combination of wheel-run,
hand-drawn, and
some time, for
decoration.
Its
would
have
stamped
development
required
stamps used to decorate the wall would have had to be designed,
and
tested and a sufficient number of molds made to support
created,
as
production. This would constitute the investment stage of the process,
time and work were diverted to this project from other tasks in the pottery.
the many
It seems likely that this new technique was a response to demand for the
the potter s
newly invented bowls, for it would hardly have been worth
a
to
of production unless he thought
while
streamlined method
develop
that there was amarket for the product. So the first prototypes had to be
and circulated, they had to gain sufficient attention to create
and
and
the
demand,
potter then had to respond by further development
we
con
innovation. Taking all this into consideration,
may conjecture that
of inspiration (stage B of the
siderable time elapsed between the moment
manufactured
bowls
process) and the production of any significant number of moldmade
was
to
become standard (stage E). If that inspiration
in the technique that
was
in 224/3, we would
that arrived inAthens
metal
bowls
prompted by
not expect any volume of production
later.
220 or even considerably
of clay copies before perhaps
about
a cash-flow curve?a
simple plotting against time of
Spratt presents
a process and then income as the innovation
at
of
the
outlay
beginning
a useful model for
is established?as
innovations.54
tracking commercial
51.SeeRotroff2000,p.499.
52. Their
floral decoration
nearly
identical,
but
three
was
different
bowls can be distinguished through
differing
rim patterns
on bowls
made
in these first molds: leaping dolphin
(Figs. 6, 7, and 8, left: P 5813, P 16221,
P 11436; Rotroff 1982, pp. 335-336,
nos. 1-3, pi. 83); floral scroll (Fig. 8,
middle: P 26664, P 27436; Rotroff
1982,
p. 336,
nos.
4,5,
pi. 83);
inverted
ovolo (Fig. 8, right: P 34969).
53. P 5813 (AgoraXXII, p. 50,
no. 49, pis. 8, 73). See Rotroff 2000 for
further
comments
on
this process.
54. Spratt 1982, pp. 83-84, fig. 3;
1989, pp. 248-249, fig. 12.1.
THE
INTRODUCTION
Stage
a?
JO
a
u
*?
ri "
t/?"~u
E
c
?
E
a
o
h- a
.2 01
SB
-a E a
m
i
vO
9 to
= 2
1
?
(N
(N
Figure 9. Stages of innovation in the
history of theAthenian moldmade
bowl
THE
BOWL
MOLDMADE
iQJ E
m
^
c
fU
Oi-? +-?
?^
o?- ?4
Q. O
o fc
C W
?0> TO
u
C O? c
S S C o
u
? I 13 S
"g
3
Si?
00 ?
7 5
O(N
(N
-S
2o- 13
*-'
?
2
01
= 'S
?- o
5
2
1
(N
(N
I
-*
rsi
(N
o
QJ
?
c S g
o c E
vj
=
+* O
Se
373
of innovation
tu
S??
-S
O
c
o
03
?<D<?
OF
?o
Q..Q
VO
00
I
o
00
o
?
si
JS?
o <v
U
curve for a successful innovation, annotated with
Figure 9 gives his generic
of the moldmade
the stages in the development
bowl. The archaeological
contexts allow us to control the time factor, and suggest that stages B and
C were of very short duration, but that there was a surprising lag in time
before stage E?profitable
achieved. That at least seems to
production?was
be the message of the small number of fragments found inAgora deposits
dating within the first 40 years of production.
CONCLUSION
A
review of the chronology of the Athenian moldmade
bowl in light of
lowered
for
Rhodian
handles
indicates
Finkielsztejns
chronology
amphora
in 224/3 is still a viable hypothesis.
that introduction
Innovation
theory
that help in following
the process; its application to an
case inwhich
objects and deposits can be dated within a very
archaeological
narrow range, however, poses
and explana
challenges of interpretation,
tions of the observed phenomena
remain elusive. In particular, it is difficult
provides models
to know how best
to interpret the small number of moldmade
bowls in
of their production.
deposits laid down during the first two generations
it mean that over 40 years passed before people were making
Does
and
are
in
moldmade
bowls
numbers?
Or
the
data
any significant
using
merely
the artifact of the depositional
process, a delayed visibility reflecting both
the small percentage of material that has survived from antiquity and the
amount
of time
record and hence
it takes for objects to accumulate
in the archaeological
to become visible to modern
investigators?
ROTROFF
374
Figure 10. Plot of a successful (x)
and a relatively unsuccessful (y)
innovation,
laid
the
against
time
scale for the Athenian moldmade
bowl
at face value, we would have to conclude that
little attention and achieved only modest
seems
standards a very long time. This can be
sales for what
by modern
a
envisioned with the help of another of Spratt's models,
graph in which
are
innovations
successful (x) and relatively unsuccessful
compared in
(y)
terms of the growth of sales over time (Fig. 10).55 The two curves overlap
at first, but at some point the curve of the successful innovation diverges
Ifwe
take the numbers
the moldmade
bowl
attracted
as customer satisfaction creates aword-of
significantly and rises steeply,
factors influence the point atwhich these two
mouth campaign. Complex
our case, it appears that the moldmade bowl followed
curves will
diverge. In
a
curve
a
for
the y
long time, before
fairly sudden increase in production
in the second half of the first quarter of the 2nd century.
is plausible, what would explain the delay? The novelty
bowl may have encouraged some consumers to acquire
it, but the shape had to compete with a large variety of drinking shapes:
West
Slope angular and baggy kantharoi, kantharoi with molded rims, and
to mention
wheelmade
cups with interior decoration,
only
hemispherical
the most numerous classes.56 This vigorous ceramic tradition may have
If this scenario
of the moldmade
bowl inAthens. After the
impeded the rapid acceptance of the moldmade
decorated cups
end of the 3rd century, however, production of wheelmade
kantharoi
ceased to
decreased dramatically. West
and
angular
Slope baggy
rims
Slope kantharoi with molded
altogether, andWest
were
made in only small
and hemispherical
cups with interior decoration
in the 2nd
while
made
and the net-pattern
numbers. The mastos
cup,
be manufactured
in significant
century, are rarities.57 The only drinking shapes manufactured
were
two
of
different
two-handled
numbers
undecorated:
cups
designs
55. Spratt 1982, pp. 82-83, fig. 2;
1989, pp. 250-251, fig. 12.3.
56. Agora XXIX, pp. 100-107,
110-117,
summary
see
cups,
pis. 17-27,31,34-37.
of the chronology
Agora
XXIX,
For
57.Agora XXIX, pp. 108-110,
pis. 32,33.
a
of drinking
p. 481, graph 4.
THE
OF
INTRODUCTION
THE
MOLDMADE
BOWL
375
rim and a strap handle.58 For
and an ungainly kantharos with a molded
the buyer in search of something elegant and decorative,
there was now
little choice other than the moldmade
bowl. The new shape had, in the
end, beaten out its competitors, but it took more than a generation for this
to happen.
be interesting to know if this proposed pattern is a common or
Spratt charts the delay in the acceptance of various innova
over
tions
the past 10,000 years.59 His figures indicate that the rapidity with
Itwould
an unusual one.
innovations are accepted has increased dramatically
through time,
he sees as a reflection of "the acceleration of social processes generally
terms a
times.,,6? Perhaps what is in modern
from prehistory into modern
40 years?was
long delay?some
typical in the Hellenistic
period.
It may be, however, that the delay suggested by the stratigraphie data
which
which
reflects the time it takes objects to go from the systemic to the ar
record and to accumulate in that record in sufficient numbers
chaeological
merely
to be recovered
today. If so, the steep increase in numbers documented
by
in
the
first
which, on the basis
quarter of the 2nd century?and
deposits
of the cash-flow model, we would associate with established production
and distribution?actually
took place considerably earlier.
It is difficult to know which of these alternatives to prefer, but the ex
an earlier
some support for the
ample of
innovation?red-figure?provides
recent
scenario.
A
and
authoritative
handbook dates the
delayed visibility
or
soon after 525."61We have evidence,
introduction of red-figure "around
in the long lists of their works inARV, that the early red-figure painters
were
successful, yet this ware is only sparsely
productive and commercially
in
laid
down between 525 and 480. Leslie
Agora deposits
represented
Shear Jr.'s review of Persian destruction debris tallies 100 red-figure sherds
from 21 wells
and pits filled with this material.62 Kathleen Lynch, including
context pottery from
in her tally the uninventoried
larger deposits such
as the
Shaft and the Stoa Gutter well, adds about
Rectangular Rockcut
30 sherds to the total.63 It is, in any case, amodest
in 480.
that the craft was well established
collection,
but it shows
is almost completely absent, however, from contexts with
a
one
dates. Mary Moore
fragment from
pit
published
ca. 490, and
filled
in
notes
two
small
from
well
perhaps
Lynch
fragments
V 24:2, deposited ca. 500.64 Red-figure
is not represented among the ca.
255 pieces of Archaic fine ware in the building fill of the Old Bouleu
no later than
and black-figure
terion, dated on the basis of black-gloss
Red-figure
earlier deposit
500,65 and only two uncertain fragments were recovered from below the
original clay floor of the nearby Building J,which Shear places in the early
5th century.66 In short, if red-figure did originate ca. 525, it took 25 years
for it to make even a very small appearance in Agora deposits. Indeed, if
63. Kathleen
5S.AgoraXXD?, pp. 106,117-119,
59. Spratt 1982, p. 92, fig. 6; 1989,
pp. 255-256,
60.
Spratt
Lynch
(pers.
comm.).
64. P 17531 {AgoraXXX, p. 320,
pis. 29,30,38,39.
fig.
12.5.
1989,
61.Boardman2001,p.
no.
1418,
attributed
a
of a cup
fragment
pi. 132),
to the Colmar
Painter,
from pit J 18:6.The deposit is other
p. 255.
79.
62. Shear 1993, p. 392, table 4.
wise
not
unpublished
the basis
give
and Moore
for
its
dating.
does
For
well V 24:2, Lynch (pers. comm.);
Agora XII, p. 399;Agora XXIII, p. 336.
65. Shear 1993, pp. 419-422,
472-473, deposit H 10:7.
66. Shear 1993, pp. 426,477-479,
deposit H 12:18, lower fill,with ca. 256
pieces
of Archaic
fine ware.
SUSAN
376
I.
ROTROFF
its inception were to be dated only on the basis of this evidence, itwould
that is a discussion
that reaches
probably be placed somewhat later?but
an effect on
limits
of
have
the
Other
factors
had
this
may
beyond
paper.67
the depositional history of red-figure. Vigorous export would have removed
in the late 6th century
much from Athens,
and the nature of the Agora
the overall relief
differed from what it was in the late 3rd. Furthermore,
of moldmade
decoration
bowls makes
even small
fragments recognizable,
But if the moldmade
bowls followed
is not the case for red-figure.
the same general pattern, we need not take the small numbers of fragments
in deposits of the last quarter of the 3rd century as evidence for a slow
growth of the industry.
which
One final conclusion remains to be drawn. Whatever
may be the case
rise or slow accep
bowl?meteoric
of the moldmade
of the marketing
for the way in which
archaeolo
tance?the
statistics have implications
or
use moldmade
bowls in applying dates to strata
gists
deposits. A single
moldmade
bowl may indeed indicate that a deposit was laid down after
224/3; but it is important to keep inmind an obvious but often overlooked
the chances that the date is shortly thereafter are slight. Strati
fact?that
bowl is mainly a marker
graphically, the moldmade
it iswell represented
in
which
and
century,
deposits
after ca. 175.
be
dated
should
large ones)
probably
of the 2nd
(unless they
(and 1st)
are very
Results
of
the Ameri
by
Classical
Studies at
of
Princeton
Athens,
Greek
IV = R. H. Howland,
can School
and Their
= B. A.
Lamps
XII
and L. Tal
Sparkes
co tt, Black and Plain
Pottery
of the
Centuries
b.c., 1970.
6th, 5th, and4th
= S. I.
XXII
Rotroff, Hellenis
tic
and Imported
Pottery: Athenian
Moldmade
1982.
Bowls,
XXVI = J.H. Kroll, The Greek
Coins,
XXIX
1993.
= S. I.
Rotroff,
tic
Pottery:
Wheelmade
Athenian
Hellenis
Imported
and Related
1997.
=M. B.
Moore,
Material,
XXX
Attic
and White-Ground
Figured
Red
Pottery,
nistic
= S. I.
Rotroff,
The Plain
Pottery:
forthcoming.
N. 2003.
Badoud,
chronologie
riques
587.
des
rhodiens,"
Helle
Greek
J. 2001. The History
of
and Pictures,
Vases: Potters, Painters,
London.
B?rker,
C.
plex,"
1998.
"Der Pergamon-Kom
hellenistischen
Ampho
aus
{PF 11),
Pergamon
in Die
renstempel
pp. 3-69.
D. von.
Bothmer,
1984-1985.
Treasury,"
1-72.
"A Greek
BMMA
n.s.
42:1, pp.
van Ufford,
L. 1953.
Byvanck-Quarles
"Les bols m?gariens,"
BABesch
28,
pp. 1-21.
= G. R.
Corin
Corinth VII.3
Edwards,
VII.3),
Princeton
F. 1922.
Courby,
Paris.
Wares,
"Remarques
36.
Boardman,
thianHellenistic Pottery {Corinth
1997.
XXXIII
B. 1990.
"Coroplast,
in The
and Metalsmith,"
and Roman
and
Table Ware
Potter,
sArt: Greek Terracottas
Coroplast
Hellenistic
ed. J. P.
World,
of the
New
Rochelle,
Uhlenbrock,
pp. 31
1958.
Survivals,
Barr-Sharrar,
sur la
?ponyms
ampho
REA
105, pp. 579
Edwards,
G.
In a recent
R.
1975.
Les vases grecs ?
reliefs,
1956.
"Hellenistic
in L. Talcott,
B.
tery,"
Philippaki,
and V. R. Grace,
G. R. Edwards,
Small Objects
from thePnyx: II
Pot
study Richard
Neer
(2002, pp. 195-205) points out that
almost all of the 56 attributed pots
published from Persian debris in the
Agora
date,
on
the
ogy, well before 480;
with
other evidence,
standard
on
chronol
this basis,
he
suggests
of the Pioneers
ing of the work
he does
years 510-480,
although
the date when
advocate
lowering
red-figure
REFERENCES
= The Athenian
Agora:
Agora
Excavations
Conducted
67.
technique
was
along
a shift
to the
not
the
introduced.
THE
figur?s,
petits
diverses
vierten
Monuments
AM
terre-cuites,
(FdD V: Monu
antiquit?s
Kroll,
mentsfigur?s 1), Paris 1908.
G.
Finkielsztejn,
d?taill?e
2001.
(BAR-IS 990), Oxford.
1960.
G. M.
"Life-Expectancy
of Utilitarian Pottery inTzintzunt
zan, Michoac?n,
AmerAnt
Mexico,"
25, pp. 606-609.
Grace, V. R. 1934.
Amphora
"Stamped
Handles Found in 1931-1932,"
-.
ria 32, pp. 319-334.
-. 1974. "Revisions
in
Early
Hellenistic Chronology," AM 89,
1985. "TheMiddle
Dated
by Amphora
54, pp. 1-54.
peria
V. R.,
Grace,
and M.
grecs,"
amphoriques
et al., L'?lot de laMaison
"Les
timbres
in P. Bruneau
des Com?di
ens (D?losXXVII), Paris, pp. 277
1982.
C.
zur Geschichte
Studien
Athens in hellenistischerZeit (Hypo
mnemata
-.
1992.
-.
and Rhodian
U.
Bau
Eponyms,"
Rhyta,
1999'.
J?hrens,
Nationalmuseum
von H
G.
Metal
M.
AMIran
Seif
en
faience
ed. M.
d'?poque
duMus?e
Lolling
aufgenommenen"
aunedierten Henkelinschriften,
Mainz.
-.
in B'Em<jT7]juoviicr?
yia
Fr?nkel,
pp. 423
Berlin,
K.
megarischen
drinischen
129-154.
1989.
"Innovation
Plain,"
A Closer Look
npo?Xrj
Innovation,
in What
der
alexan
Kunsthandwerk,"
Jdl
9,
Theory
sNew:
at the Process
of
?d. S. E. van der Leeuw
and R. Torrence,
zum
62,
Hesperia
pp. 245
London,
257.
"A?tGocva,
61, pp. 203-221.
1955. "Das Verh?ltnis
Becher
Deposits,"
Made
rrjv eXXr\viaxiKT\
pp. 94-97.
J. K. 1992.
Hesperia
pp.
"Amphoren
von
Spratt, D. A. 1982. "The Analysis
of Innovation
Processes/'y/fS
pp. 79-94.
de Per
Tuy?res, and Kiln Firing Supports,"
Parlasca,
1895.
Inschriften
Zeit. In
R?mische
from Agora
pp. 383-482.
hell?
"C?ramiques
les fouilles
d'apr?s
juara xr\? eXXr?viOTiicr?c K?pa?jLencr?c,
Amphorenstempel
von Athen: Zu den
York.
Shear, T. L., Jr. 1993. "The Persian
Destruction
of Athens:
Evidence
el-Din.
(Etudes
Kepafieacr}: XpovoXoyac?
im
trans.
The Cave,
schriftenauf Thon (AvPV??I.2),
Craft
b.c.e.,
Catalogue
dAlexandrie
game/Kestel,"
Red-Figured
Athens,
Kepa?iiaj,
in Die
Pergamon:
The
4), Cairo.
Ivv?vrr]<7?]
Athens,
C.
Schuchhardt,
in
yia
Ivv?vxr?cr?
New
J. Costa,
and
Production
and Spread,"
xr?v eXXr?viaxiicr?
stempel,"
and M.
nistiques
Papadopoulos,
"Molds:
pp. 496-501.
J. 2002.
Saramago,
Black-Gloss
Century,"
Copenhagen,
Preparation?Use
E' EniGxr\iioviKr\
and
0.1990.
Ozyigit,
Mainz.
G.
-.
of the
Greek and
ed. J. Christiansen
and T Melander,
pp. 516-523.
2000.
Portrait
Proceedings
on Ancient
Symposium
Related Pottery,
5), Rome.
"Adoption
of Achaemenid
La vaisselle
gr?co-romain
alexandrines
1959. Hellenistische
1962. Attic
Bottega
Tableware,"
499.
gr?co-romaine:
denerFundpl?tze inOlympia
(OlForsch27), Berlin.
H.
in Oriente
in
3rd
109-146.
M.-D.,
2000.
and
"Silver, Glass,
Luxury
in Athens,"
Cambridge.
n?rdlich von
Brunnenf?llung
C und
verschie
Reliefkeramik
Hoffmann,
26, pp.
Chicago.
S. 1.1982.
Hellenistic
e delle Mis
di Atene
in Attic
Pottery Analysis:
51, pp. 329-337.
Hesperia
-. 1988. "A Ptolemaic
ellenistica
a relievo della
of the Fifth
Ware
Nenna,
aus attischen und b?oti
Reliefbecher
schen Werkst?tten,
Stuttgart.
-. 1996. Hellenistische
Keramik:
Eine
cer?mica
1993.
C.
Clay
Metal
45, pp. 41-66.
Hesperia
1987.
Sourcebook,
Rotroff,
pp. 334-346.
Vase-Painting:
ca. 530-460
ofDemocracy,
REA 105, pp. 541-578.
Hausmann,
M.
"Ancient
from Greek
Clay: Evidence for theDating of
Athenian
11, pp. 68-90.
Alexander
\997. Athens
from
to
Mass.
Cambridge,
Antony,
2003.
"Rhodian Amphora
Stamps
Miller,
P. M.
A
vessel
Berichte
Neer, R. T. 2002. Style and Politics in
the Ptole
ClAnt
mies,"
-.
73), G?ttingen.
"Athens
and
archeologica
sioni italiani
Adaptation
ware Forms
382.
Habicht,
work,"
Rice,
di Efestia (Monografie della Scuola
Sawatianou
1976.
Impressions
"Pottery Use
the Kalinga,
Northern
among
in
the
Luzon,
Philippines,"
Decoding
ed. B. A. Nel
Prehistoric
Ceramics,
La
E. D.
Reeder,
Life
1992.
M.
Massa,
con decorazione
Hes
Stamps,"
1970.
P?tropoulakou.
Stoa
zu alex
Studien
ischer und grossgriechischer
Toreutikfr?hhellenistischer Zeit
(?F 16), Berlin.
66, pp. 321-329.
W. A. 1985.
Longacre,
son, Carbondale,
1987.
M.
andrin
Lund, J. 2002. Review of Finkielsztejn
2001, BMCR 2002.11.23.
pp. 193-203.
-.
Glass
Glastechnische
Production,"
3, pp. 197-310.
Hesperia
1963. "Notes on the Amphoras
from the Koroni
Peninsula,"
Hespe
290-291.
Pfrommer,
2001, m AJA 107, pp. 131-133.
Was theTurning
Lierke, R. 1993. "It
Wheel, andNot the Lathe: Mold
Pressing andMold Turning of
in Ancient
residui
(CEFR
archeologico
C. Pavo
?d. F. Guidobaldi,
lini, and P. Pergola,
pp. 5-19,
Amphora
Stamps,"
71, pp. 295-324.
Hesperia
2003. Review
of Finkielsztejn
Glass
in / materiali
249),
Appen
of Rhodian
Hot
of Residual
"Aspects
Assemblage,"
nello scavo
89, pp. 201-203.
dix," AM
M. L. 2002.
"Early Excavations
and
the Chronology
Pergamon
-.
1998.
J. T.
377
ity in the Palatine East Pottery
v. Chr.,"
Jahrhunderts
79, pp. 22-84.
J. H. 1974. "Numismatic
BOWL
Pe?a,
Lawall,
at
Chronologie
et r?vis?e des
?ponymes
rhodiens, de 270 ? 108
amphoriques
bilan
environ: Premier
av.J.-C.
Foster,
MOLDMADE
des
Schwarzfirniskeramik
attische
bronzes,
THE
Kopeke, G. 1964. "Golddekorierte
(Hesperia Suppl. 10), Princeton,
pp. 79-112.
= P.
FdD V.l
Perdrizet,
OF
INTRODUCTION
70,
J. N. 1923-1926.
naies dAth?nes, Munich.
Svoronos,
Thompson,
turae:
D.
1939.
Impressions
Metalwork,"
316.
B.
Hesperia
Les mon
"Mater
Caela
from Ancient
8, pp. 285
-.
1959.
Hellenistic
"Three
Centuries
of
II B: The
Terracottas,
Altar Well,"
28, pp.
Hesperia
127
152.
-.
1966.
"The
of Tana
Origin
gras,"4/^70,pp.51-63.
H. A. 1934. "Two Centu
Thompson,
ries of Hellenistic
Pottery," Hesperia
3, pp. 311-480.
P. 2000.
Triantaphyllides,
Po?iaicrj
I: Ta ev Oepju
vaXovpyia
?ia/uop
(pco?eva oia(pavr? ayye?a noXvre
Xeia?. O? kXocgiko? Kai o? np?ifioi
eXXr?viariKo?xp?voi,
Athens.
Susan I. Rotroff
Washington
University
department
campus
st.
I.
ROTROFF
1969.
Griechische
SUSAN
378
louis,
of
box
classics
io50
missouri
[email protected]
6313o
Trumpf-Lyrizaki,
vasen
Figuren
sp?ten Klassik,
E.,
Vanderpool,
A.
Steinberg.
A Ptolemaic
M.
des reichen Stils
und der
Bonn.
Watzinger,
J. R. McCredie,
1962. "Koroni:
on
and
the East
Camp
of Attica,"
31,
Hesperia
pp. 26-61.
and D. Gill.
1994.
Vickers, M.,
Artful
Ancient
Greek Silverware
and
Crafts:
Oxford.
Pottery,
Coast
Vickers,
M.,
O.
1986. From
Potters
Debt
Graeco-Roman,
Worlds,
Impey, and J. Allan.
Silver to Ceramic: The
toMetalwork
in the
Oriental,
and Islamic
Oxford.
C.
1901.
"Vasenfunde
aus
AM
Athen,"
26, pp. 50-102.
1978.
E.
R.
Williams,
"Figurine
from the Athenian
Agora,"
Vases
Hesperia
47, pp. 356-401.
in
R. 1904.
Zahn,
"Thongeschirr,"
Priene:
der
Ergebnisse
Ausgrabun
in den
gen und Untersuchungen
ed. T. Wiegand
1895-1898,
Jahren
and H.
468.
Schrader,
Berlin,
pp. 394