Hadra Hydriae and Central Crete: A Fabric Analysis

Author(s): P. J. Callaghan and R. E. Jones

Source: The Annual of the British School at Athens , 1985, Vol. 80 (1985), pp. 1-17
Published by: British School at Athens

Stable URL: https://www.jstor.org/stable/30102817

REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/30102817?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

British School at Athens is collaborating with JSTOR to digitize, preserve and extend access to
The Annual of the British School at Athens

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
 HADRA HYDRIAE AND CENTRAL CRETE:
A FABRIC ANALYSIS

(PLATES 1-4)

INTRODUCTION

HADRA H YDRIAE were named after one of the main Hellenistic cemeteries of Alexandria
where they were first discovered during the last quarter of the nineteenth century. The
since been unearthed in large numbers at other Alexandrian necropoleis and, b
comparatively few have been found elsewhere in the Hellenistic world, most authorities
assumed that they were made in Egypt.1
In Egypt they were used as ash-urns for the burial of visiting dignitaries to the Ptolem
court, foreign mercenaries, and many of the Graeco-Macedonian inhabitants of the c
For religious reasons the natives continued to prefer mummification. Decoration is
comparatively simple. The best painters attain modest standards of care and neatness tho
even the finest hydriae cannot compete with their Classical predecessors in elegance.
these limitations Hadra hydriae are among the most ambitious products of Hellenistic po
and painters, and as such have tended to receive the attention of students of post-Classical
pottery. In a period characterized by a shortage of secure chronological markers,
importance is further augmented by the fact that almost thirty bear inscriptions which i
numbered regnal year of one of the ruling dynasty.3 It is true that no individual king i
named in the formulae, but objective analysis of other information contained in these ep
has resulted in the identification of the Ptolemy involved.4 Scholars have now begun the
arranging the undated Hadra vases, which vastly outnumber those bearing inscript
stylistic groups whose absolute chronology is based on related inscribed vases.5 This
successfully completed, will provide us with the largest body of securely dated pottery i
Hellenistic world.
The term 'Hadra vases' has been applied to two loosely related groups of hydriae, both t
having been found in great numbers at Hadra and the other Alexandrian cemeteries. The f
are called White Ground hydriae.6 They are made of a red, friable clay which closely rese

Acknowledgements. We should like to thank the TrusteesResearch

of Fellow at the University of Newcastle upon
the British Museum and Mr B. F. Cook of the DepartmentTyne.
of Greek and Roman Antiquities for permission to study and
1 For Hadra vases see B. F. Cook, Inscribed Hadra Vases in
sample the Hadra vases in London (PLATE I, samples I-5)-
the Metropolitan Museum of Art. Papers 12, 1966 (henceforth
Dr A. Delivorias and the American School of Classical
IHV); Brooklyn Museum Annual x (1968-9) 115-38; BSA 75
Studies in Athens also allowed study of their material from
(I98O) 33-47; BICS 30 (1983) 123-9; L. Guerrini, Vasi di
Knossos and the Managing Committee of the British School
Hadra, Tentativo di sistemazione cronologica di una classe ceramica
at Athens gave permission for the publication of the
(Seminario di archeologia e storia dell'arte greca e romana
results and the accompanying photographs. Professors
dell'UniversitaSt.
di Roma), Studi Miscellanei 8 (1964), hence-
Alexiou, Dr J. Tzedakis, Dr A. Lembesis, and Mrs
forth A.Hadra.
Vasi di
Karetsou have, at different times, allowed access to their finds
2 Brooklyn Museum Annual x (1968-9) 15-38; IHV.
and have discussed questions of style and provenance.
3 IHV, BICS 30 (1983) 123-9-
Professor D. Levi and Professor Larosa of the Italian Insti-
4 H. Braunert, 'Auswartige Gaste am Ptolemaerhofe', JdI
tute in Athens have also been very helpful. The technical
65/6 (1950-1) 231 if.
assistance of Miss E. Louka is gratefully acknowledged. The
5 BICS 30 (1983) 124; BSA 75 (I980) 33-47; 76 (I981)
photographs were prepared by Matthew Nightingale. This
38-40, 55-7, 59-70. 6 IHV 9-Io.
study was completed while P. J. C. was Sir James Knott

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
 2 P. J. CALLAGHAN AND R. E. JONES

that used for the manufacture of local terracotta

Egyptian origin. The rough and porous surface neede
application of polychrome tempera decoration which
been fired. The motives chosen for these vases are o
and they were all probably made specifically for the
this fabric are never found outside Egypt. There are
the local transport amphora fabric and then whitew
The second group, called Clay Ground hydriae,8 hav
though there is a subgroup which uses a redder,
frequently polished and varies in colour from a b
dark-brown to black paint was applied straight on to
touches of white, red, or pink were also added. Ther
motives. The most popular are rinceaux of vines, ivy,
more ambitious figured scenes representing races
human activities.10 Dolphins figure prominently both
one of these vases was not made specifically for the
made it for the Games.'1 Many others may have been
some iconographic influence from contemporary
look Egyptian and the inscriptions, which are, were
Hydriae of this type have been found in widely sca
Mediterranean.13 Because of the nature of the fabric
examples than anywhere else but Alexandria, R. M. C
might have been an important centre for their manu
receive much attention until recent excavations in C
that the true home of the Hadra workshops once lay
The first clue was provided by a morphological feat
Clay Ground hydriae sharing a technical quirk which
vases: the dropped floor.16 In this group the inter
dropping the floor of the body deep into the high r
thrown in one piece with the rest of the vase. This fe
long history on Crete, where it is often found on cups
century BC.17 Recent excavations in Crete have als
vases and of other shapes decorated in similar techni
surface of the Clay Ground hydriae closely resemble
and other Cretan centres.18s Knossos has also produc
Slope technique where the motives are so similar to
there is good reason to claim that the same painters
it was decided that an objective analysis of the C

7 IHV 7 n. 3: Rhodes. 15 See n. 5 above.

8 IHV9-io. 16 IHV9.
17 BSAfabric:
9 I have seen two examples with this 73 (1978) 6, no. 12; no. di
Vasi 16; i'Hadra
fig. 7. C,
5, now in the National Museum at 18Athens,
For self-polished
andvases see BSA 72 (1977) from
another 99. BSA 73
Rhodes. (1978) nos. 27-41 all have this type of surface treatment as do
10 Vasi di Hadra groups A-C. many of the local domestic hydriae as well as vases of various
11 Ibid. II and n. 3. other shapes.
12 Either inked on the surface of the vase or incised: IHV 19 BSA 76 (I98I) 38-40, 55-7, 59-70; 'Knossian Artists and
passim and Brooklyn Museum Annual x (1968-9) 115-38. Ptolemaic Alexandria', in Alessandria e il mondo ellenistico. Studi
13 IHV n. 3. in onore di A. Adriani (forthcoming).
14 Cook, Greek Painted Pottery (1972) 20o8.

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
 HADRA HYDRIAE AND CENTRAL CRETE: A FABRIC ANALYSIS 3

characteristic of Knossos, was necessary in order to discover the

Hadra vases.
The approach adopted in this investigation of provenance follows a traditiona
identifying the characteristic chemical compositions of reference material from the p
tion centres of interest, in this case Crete and the Nile Delta, and comparing those of
pieces, the Hadra hydriae (and related material), with them. Similarity of comp
term that can be judged by objective statistical criteria, generally implies corres
of origin.
The chemical compositions of the selected reference and test material were determined by
optical emission spectroscopy, a method of analysis that the Fitch Laboratory has used
extensively in recent years for pottery provenance studies. The method is most suited, as is shown
elsewhere,20 to the characterization of manufacturing centres whose products exhibit relatively
large differences in composition; although there are exceptions,21 the spectrographic technique
cannot effectively resolve fine differences in composition that may represent, for example, the
products of individual workshops at a particular site or within a region.22 For the purposes of the
present investigation, however, this problem does not arise since the primary goal was to
differentiate the pottery of central Crete and the Nile delta with respect to chemical
composition. But before presenting the results a summary of what has already been achieved in
these two regions by this and other laboratories is necessary.
The Fitch and Oxford Research Laboratories' survey of Cretan pottery has provisionally
discerned three composition zones on the island;23 the clays from the centre of the island used in
the manufacture of Minoan pottery are more calcareous and have higher Mg, Cr, and Ni
contents than those first in the east (the lerapetra isthmus appears to be the 'border') and
secondly in the Chania and Rethymnon districts to the west. Within central Crete the
compositions are not entirely uniform, but it is not yet possible, at least with the spectrographic
technique, to differentiate the composition of, say, the Knossos region and the Mesara with
confidence.24 A phenomenon that is shared, however, by all the pottery groups in central Crete
and may be no more than a reflection of the natural variations in the compositions of the clays is
the size of the composition ranges in each group. Knossos is no exception; fine, decorated pottery
of LM IB25 and LM IIIB26 has yielded compositions whose dispersions are wider than those
encountered in many parts of the Aegean. Petrological analysis of Knossian fabrics is currently
providing the necessary information which will greatly amplify our understanding of the clays in
the region that were exploited for pottery-making.27 The emphasis of Riley's survey has been
directed towards prehistoric pottery, but in an interim report he presented the mineralogical
compositions of some material that has direct relevance to the present study. Wasters from the
Classical-Hellenistic kiln found on the south-east slope of the Akropolis hill28 contained much
20 A. H. S. Megaw and R. E. Jones, 'Byzantine and Allied op. cit. n. 79 ff
Pottery: a Contribution by Chemical Analysis to Problems of 24 This topic is discussed by Jones, op. cit. n. 23, chapter
Origin and Distribution', BSA 77 (1982) 235-63- 3, C.
21 At Chania in west Crete, for example, some of the classes 25 Analysed as part of the study of Marine Style; P. A.
of Late Minoan pottery have specific compositions; see H. W. Mountjoy, R. E. Jones, and J. F. Cherry BSA 73 (1978)
Catling et al. BSA 75 (i980) 81. 159-69.
22 A recent analytical study of terracotta figurines from 26 Analysed as part of the study of the Linear B inscribed
Myrina by optical emission spectroscopy and X-ray fluores- stirrup jars, Catling et al. 1980, op. cit. n. 2 i, table 4 and p. I I I.
cence spectrometry that attempted to characterize workshops 27 J. A. Riley, 'The Contribution of Ceramic Petrology to
illustrates this problem well; Kalamiotou et al. Journal of Our Understanding of Minoan Society', in Minoan Society, eds.
Archaeological Science 11 (1984) 103-17. O. Krzyskowska and L. Nixon (Bristol 1983) 283-92.
23 The results are set out in full by the present writer (R. E.J. 2s B. Homann-Wedeking, 'A Kiln Site at Knossos', BSA 45
Greek and Cypriot Pottery: a Review of Scientific Studies, forthcoming (1950) 140-92.
chapter 3, C). For a preliminary account see Catling et al.

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
4 P. J. CALLAGHAN AND R. E. JONES

limestone, abundant quartz, chert lumps, and foramini

identified.29
Through the efforts of the Lawrence Berkeley Laboratory and especially the Brookhaven
National Laboratory, a large corpus of neutron activation analyses of Egyptian clays and
pottery is available. Asaro and Perlman30 identified the compositions of burnished red slip,
black top, and coarse wares made of Nile alluvium found at the Had ed Beir, Ballas, and El
Ahaiwah cemeteries in upper Egypt; they also characterized a calcareous buff ware at Ballas.
Tobia and Sayre31 have surveyed the clays of different geological formation along the Nile and,
in connection with the Brookhaven Laboratory's recent study of Tell el Yahudiyeh ware, they
published the composition characteristics of Nile alluvium and the clays of Pleistocene strata
located immediately below the alluvium.32

THE MATERIAL

The three groups are listed in Table I:

I. Hadra hydriae and related vases: i-30o.
2. Nineteen samples acting as a control for Knossos in the Classical and
31-49.
3. Twenty terracotta lamps and figurines ofHellenistic date from a number
Nile Delta and now in the Benaki in Athens. The clay was uniformly fin
was present in the red and brown fabrics; there were six examples of a
fabric.

One modern brick from a brickworks Io km south of Cairo.

TABLE I

Sample number Description Museum number

Group i: Test pieces

I Hadra vase London BM 1926.7-13.1
2 Hadra vase London BM 1927.3-17.5
3 Hadra vase London BM 1922.1-16. I1
4 Hadra vase London BM 68.7-5-136
5 Hadra vase London BM 90go.6-9.28
6 bg jug London BM G 12
7 Hadra vase Athens (Benaki) SLM 190
8 Hadra vase Athens (Benaki) SLM I9i
9 Hadra vase Athens (Benaki) SLM 192
o10 Hadra vase Athens (Benaki) SLM 193
S1 Hadra vase Athens (Benaki) SLM 194
I2 Hadra vase Athens (Benaki) SLM 195
13 Hadra vase Athens (Benaki) 22863
14 Hadra vase Athens (Benaki) 22694

29 J. A. Riley, D. P. S. Peacock,
of Various and A. C. Renfrew,
Egyptian Soils, 'Th
Cl
Petrological CharacterisationPottery of Late by NeutronBronzeActivation', Age Ceram in
Technology
from Knossos and Mycenae', Revue d'Archeomitrie of Materials iii, e
4 (1980); s
also Riley, op. cit. n. 27. 99-218.
30 I. Perlman and F. Asaro, 32 M.Archaeometry
F. Kaplan, G. Harbottle, and E. V. Sayre,
II 'Multi-
(1969) 34-
and H. V. Michel, J. D. Frierman,
Disciplinary Analysis of Tell el and F.Archaeometry
Yahudiyeh Ware', Asaro, ibid.
(1976) table 2. 24 (1982) 127-43.
31 S. K. Tobia and E. V. Sayre, 'An Analytical Comparis

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
 HADRA HYDRIAE AND CENTRAL CRETE: A FABRIC ANALYSIS 5

Sample number Description Museum number

15 Hadra vase Athens (Agora) P7194

16 Medusa bowl Knossos, Stratigraphical
17 Medusa bowl Knossos, Stratigraphical
18 Classical bg vase Knossos, Stratigraphical
19 Tulip cup Knossos, Stratigraphical
20 Tulip cup Knossos, Stratigraphical
21 Hadra base Knossos, Stratigraphical
22 Hadra vase Knossos, Stratigraphical
23 Hadra vase Knossos, Stratigraphical
24 Hadra vase Knossos, Stratigraphical
25 Horse krater Knossos, Stratigraphical
26 Hadra vase Knossos, Stratigraphical
27 Hadra vase Knossos, Stratigraphical
28 Cylindrical jug Knossos, Stratigraphical
29 Cylindrical jug Knossos, Stratigraphical
30 Hadra vase Oxford 1920.250
Group 2: Knossos control
31 Kiln waster Knossos, Stratigraphical
32 Kiln waster Knossos, Stratigraphical
33 Kiln waster Knossos, Stratigraphical
34 Cylinder cup Knossos, Stratigraphical
35 Cylinder cup Knossos, Stratigraphical
36 Cylinder cup Knossos, Stratigraphical
37 Base-ring workshop cylinder cup Knossos, Stratigraphical
38 Base-ring workshop cylinder cup Knossos, Stratigraphical
39 Pseudo-base ring workshop cylinder cup Knossos, Stratigraphical
40 Pseudo-base ring workshop cylinder cup Knossos, Stratigraphical
41 Proto hadra Knossos, Stratigraphical
42 Hydria Knossos, Stratigraphical
43 Hydria or jug Knossos, Stratigraphical
44 Hydria Knossos, Stratigraphical
45 Hydria base Knossos, Stratigraphical
46 Hydria, misfired Knossos, Stratigraphical
47 Hydria Knossos, Stratigraphical
48 Hydria Knossos, Stratigraphical
49 Hydria Knossos, Stratigraphical
Group 3: Egyptian control
50 Modern brick made near Cairo
51 Lamp Athens (Benaki) 11892
52 Lamp Athens (Benaki) 12053
53 Figurine Athens (Benaki) 22126
54 Lamp Athens (Benaki) 12102
55 Figurine Athens (Benaki) 12819
56 Lamp Athens (Benaki) 12062
57 Figurine Athens (Benaki) 128o8
58 Figurine Athens (Benaki) 22638
59 Figurine Athens (Benaki) 22127
60o Lamp Athens (Benaki) 12058
61 Figurine Athens (Benaki) 22128
62 Lamp Athens (Benaki) 1I865
63 Figurine Athens (Benaki) 21943
64 Lamp Athens (Benaki) o6
65 Lamp Athens (Benaki) I1954
66 Figurine Athens (Benaki) 12812
67 Lamp Athens (Benaki) 12079
68 Lamp Athens (Benaki) 11879
69 Lamp Athens (Benaki) 11853
70 Figurine Athens (Benaki) 12834

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29n 1976 12:34:56 UTC
All use subject to https://about.jstor.org/terms
 6 P. J. CALLAGHAN AND R. E. JONES

The Knossian control samples (Table i, group 2) were selected

visually to the normal fabric range on the city site. Care was
shapes for which there is no evidence outside Crete, and which
back at least as far as the Classical period. Three thoroughly m
floor of a Classical kiln on the slopes of Monasteriaki Keph
relatively large number of local domestic hydriae34 were in
provided the closest analogies both in shape and fabric to H
were taken from the ubiquitous, and clearly local, Cylindrical
from the two pottery workshops recently isolated at Knossos
Several aims were incorporated in the selection of test-piece
Classical bg vases (18-2o) were included. These, together w
illustrate local fabrics for a period long before the foundation
way connected with any putative Alexandrian manufacture
event this precaution, given the results of analysis of the Egyp
did add further data for the local control.
Twenty-one complete or fragmentary Hadra vases were also included (1-5, 7-15, 22-4, 2
27, 30). One-third of these were found at Knossos, a single example from the Athenian Agora
and the rest in Alexandria. It was hoped that these would provide a sufficiently large sample
give clear-cut results.
Finally, two sherds from Medusa Bowls were added. This shape is well represented in Crete
and the type has been claimed as a Cretan one.37 The many fragments found at Knossos or alon
the north central coast of the island suggested that at least one of the workshops producing th
bowls may well have been situated in the city. One of the samples (16) conformed to the norm
fabric and glaze type. The other (17) differed slightly in fabric, which was blonder than usua
and had a far less lustrous glaze. It was assumed that these characteristics merely represented
late position in the series.

Method

The spectrographic technique followed the procedures outlined by Schweizer38 and describe
in detail by the present writer.39 The samples were collected (i) by drilling (30 mg) with
tungsten carbide drill head (groups 2 and 3) and (2) by breaking off a chip from the sher
removing paint/slip, and grinding to a fine powder. The samples were heated at I Io OC for tw
hours before mixing with the internal standard mixture. The contents of nine elemen
(expressed as oxides) were determined for each sample. Some of the performance characteristi
of the analytical method are set out by Megaw and Jones,40 and they are treated fully elsewher
by the present writer.41

Results

The Knossian compositions (group 2) conform to the central Cretan type, but there are
nevertheless some interesting differences from those of prehistoric (Late Minoan) pottery. The
clay is less calcareous than in earlier times and indeed all the other measured major and minor

33 Homann-Wedeking, op. cit. n. 28. 37 BCH 95 (1971) 207; BSA 76 (1981) 61-5 and esp. n. 15-
34 For the type see BSA 45 (1950) pl. 12e; 52 (i957) 229 38 A. J. N. W. Prag, F. Schweizer, and J. L1. Williams,
fig. 2. Archaeometry 16 (1974) 153-88.
35 BSA 45 (1950) i8i fig. 20, left; 66 (1971) 260 fig. io no. 39 Jones, op. cit. n. 23, chapter 2, A.
37; 73 (1978) 15-18; 76 (1981) 37f., 55- 40 Megaw and Jones, op. cit. n. 2o, 245 ff.
36 BSA 73 (1978) 15-18; 76 (I98I) 37f. 41 Jones, op. cit. n. 23, chapter 2, A.

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
 HADRA HYDRIAE AND CENTRAL CRETE: A FABRIC ANALYSIS 7

TABLE 2

Percentage oxide

Sample no. Al Ca Mg Fe Ti Na Mn Cr Ni

1 I6.8 o0.6 6.4 7-3 0.74 2.62 0.085 0.042 0.026

2 24.2 25.0 8.3 10.9 0.95 1.I3 0.135 0.066 0.040
3 15.8 13.0 6.o 9.0 0.75 1.64 0.078 0.046 0.033
4 17-9 12.2 5.7 7.6 0.73 1-41 0.083 0.038 0.033
5 20.4 18.9 5-4 7-9 o.85 1.20 0o.o3 0.049 0.029
6 20.0 12.8 6.3 8.2 0.91 0.74 0.087 0.042 0.025
7 I1.5 8.4 6.6 8.3 0.77 2.1 0.094 0.075 0.049
8 17.9 12.0 5-3 8.8 0.87 2.50 0.100 0o.o049 0o.o041
9 17.4 12.0 7-3 10.2 0.91 1.79 0.093 0.079 0.057
Io 16.2 10.4 8.o 9.8 0.95 1.98 0.089 0.064 0.047
II 10.2 4.0 5-4 7.1 0.59 i.6o o.o81 o.o041 0.034
12 14-9 I1.2 8.5 8.I 0.83 1.79 0.096 0.075 0.046
13 14-5 8.2 8.o 8.9 0.75 1.54 0.092 0.045 0.037
14 11.5 9.1 5.3 8.2 0.66 1.6o 0.oo 0.060 0o.o051
15 I8.7 17-5 6.4 7.I 0.77 0.96 o.o8o 0.053 0o.o031
16 19.6 17.2 7.2 9.8 0.87 1.45 o.118 0.089 0.052
17 26.4 18.3 8-4 11.I 1.15 1.54 0.o3 0.o054 0.035
18 11.5 10.2 6.3 7-7 0.79 1.20 0.o03 0.066 0.047
I9 19.6 I7.2 5.7 9.8 0.83 1.34 o.1I8 0.054 o.o041
20 15-4 13.1 6.2 10.2 0.83 1.25 0.113 0.078 0.057
21 12.6 8.0 4-3 7.0 0.64 I.oo00 0.073 0.073 0.047
22 19.7 12.8 9.8 10.5 1.07 1.63 o.118 0.090 0.062
23 16.6 5.2 4-3 9-4 0.83 0.88 0.098 0.052 0.032
24 17.4 4-9 5.0 9.8 0.87 I.I 0.103 o.o61 0o.o041
25 17.9 11.7 8.I io.6 0.99 1.54 0.132 0.097 0.059
26 I6.8 I1.2 8.9 9.8 0.92 1.83 0.098 0.078 0.055
27 11.9 7.0 5-9 8.8 0.73 1.47 o.i18 0.067 0o.o051
28 I1.I 7.8 4-9 7-3 0.66 1.o8 0.084 0.057 0.043
29 15-3 10.2 7.0 9-4 0.83 I.I9 0-079 0.056 0.044
30 I7.9 12.5 6.o 7.5 0.73 1.34 0.o98 0.048 0.025

TABLE 3
rercentage oxiae

Group N Al Ca Mg Fe Ti Na Mn Cr Ni

Knossos (Classical-Hellenistic) 19 ? 13.0 7.6 5.6 9.o0 0.74 1.49 0.102 0.076 0.051
s.d. 2.6 2.4 1.4 1.1 0.12 0.45 0.021 0.01I 0.006
Knossos (LM IB) 16 x 20.3 15.1 5.1 9.5 0.95 1.34 0.077 o.o65 0.043
s.d. 2.4 4.6 2.3 1.3 0.10 0.30 0.015 0.011I o.o16
Knossos (LM IIIB) 24 f 17.6 13.2 6.4 10.0 0.92 1.23 0.092 0.064 0.058
s.d. 3.1 4.3 1.6 1.3 0.12 0.34 0.022 0.010 0.010
Nile Delta 20o x 19.8 7.3 2.o0 9.o0 1.37 i.8o 0.o89 o.o019 o.o010o
s.d. 3-I 7.3 0.6 2.3 0.31 0.73 0.032 0.007 0.005
Cairo brick 20.4 5.7 2.6 10.7 1.o03 237 0.123 0.036 0.012
Thebes (LH IIIB) 20 x I8.7 14-3 5-4 9.0 o.88 1.30 0.101o 0.058 0.053
s.d. 2.7 3-9 1.8 1.6 0.12 0.45 0.024 0.025 0.026
Athens (Hellenistic) 20 x 16.7 4-3 4-4 11.8 o.85 0.92 0.096 o.o65 0.047
s.d. 2.7 1.6 0.7 4.0 0.II 0.I5 o.oI6 0.0II 0.012
Rhodes (LH IIIC) 16 9 12.1 16.I 12.2 11.6 1.03 o.81 0o.149 0.154 0.135
s.d. 1.4 6.7 1.4 1.4 0.09 0.49 0.028 0.032 0.021

(1-mean; s.d.-standard deviation)

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
8 P. J. CALLAGHAN AND R. E. JONES

elements are in lower concentration; the trace-element con

average enriched (Table 3).
The points of interest in the Nile Delta compositions ar
origin-sensitive elements, Mg, Cr, and Ni, and the variability
Ca contents. A correlation between the colour of the fabric and its Ca content was noted.
Comparison with the neutron activation data is limited, but those elements that were me
by both techniques are set out in Table 3. Table 2 sets out the individual composition of g

TABLE 4
Percentage oxide

Groups N Ca Fe Na Mn Cr

LBL 'Nile Mud' 32 X 3.6 9.8 1.80 . 156 0.026

s.d. 1.2 0.5 0.29 0.009 0.002
LBL Ballas buff 20 x 11.6 6.6 0.90 0.050 0.028
s.d. 1.6 0.4 0.23 0.005 0.oo0
BNL 'Nile Alluvium' 161 9 - 9.7 1.70 0.154 0.023
s.d. - 0.9 0.42 0.054 0.005
BNL 'Nile Mixture' ioo 0 - 7.6 1.61 0.Io2 O.ol9
s.d. 2.1 0.60 0.059 0.oo6

(LBL-Lawrence Berkeley Laboratory; BNL-Brookhaven

The full complement of control and test sample

multivariate analysis, following standardization of
trace-element contents. Hierarchical cluster analysis (W
groups I and 2 (cluster I) from group 3 (cluster II),
components analysis.42 In the plot of the first two pr
control and the test-pieces are clearly different from
different scores on the first principal component
contents and accounted for 42 per cent of the total v
Egyptian cluster, II, is attributed to the variable sc
essentially the Fe and Mn contents that have alread
The principal statement that can be made about the
most unlikely to be of Egyptian origin. How well they
with reference to FIG. 2, the results of the principal comp
alone. Although there is inevitably less inherent struct
emerge from the results of the multivariate analysis; a
(Ti, Al, Fe, and Ca) dominate the first principal comp
the second. The local hydriae distribute themselves
control samples; without requiring to define this a
12-14, I8, 20, 21, 23, 24, 26-30 within it and to pr
manufacture.

The analysis were carried out on the ICL 2980 at the Oxford
42 Cluster analysis and principal components analysis are
part of the CLUSTAN package (D. Wishart, CLUSTAN User University Computing service.
Manual, version IC Release 2, Edinburgh University, 1978).

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
 HADRA HYDRIAE AND CENTRAL CRETE: A FABRIC ANALYSIS 9

65 III
cm '(6
0 51'
Nile Delta
I 59
61

IN
17
2
Central Crete 5,
22
25
3i5
145
4_3321 l l is 1
II 1
01
S32' 24 62

41 7113
12 5
70
323
51
52

101 2 ht 43 4 6 54
67
31 3095
28 11 66
-2
21 35
,1I
$3

56

-4

-4 -2 0 2 4 6
PCI

FIG. I. The results of principal components analysis of the combined composition data for Groups I, 2, and 3 (70 samples). The
numbers are as in Table I. The first two principal components, PC I and PC 2, account for 42 and 25 per cent of the total variation in
composition, and they are dominated by the Cr, Ni, and Mg contents (negative values) and by the Fe and Mn contents respectively

The remainder diverge by degree, not absolutely, from the control group, and they appear to
separate into three subgroups:
A. x, 3-6, and 15
B. 2 and 17
C. i6, 22, and 25

Subgroups B and C and i i and xg differ from the control samples in their major and minor
element contents in a way that suggests they may represent a technical variant of the local clays;
more precisely, this is envisaged as a finer, more purified clay than the traditional Knossian
sources. By contrast, the basis of the divergence between the control samples and subgroup A
(together with a 'local' hydria, 49), which lies essentially in the contents of Cr and Ni, may
indicate that the members of subgroup A were made of a Knossian clay that was not otherwise
used there. Alternatively, in view of the inter-site differences in composition within central Crete
being small, a source for subgroup A outside Knossos but within central Crete cannot certainly
be ruled out.
A central Cretan, if not specifically Knossian, origin is argued, then, for the Hadra hydriae.
But we are left to inquire whether there are other centres in the Aegean whose products would b
compatible in composition with these hydriae. Boeotia is the principal candidate since the

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
 IO P. J. CALLAGHAN AND R. E. JONES

c4
C)
0L
25
2
22

23
27 I
31 21
7 16

14
12 II
1is 43
24
O-
21
Q 13
23
13

23
11 17
3
-2

1
4 5
15

-4

-4 -2 0 2 4 6
pci
FIG. 2. The results of principal components analysis of the combined composition data for Groups I and 2 (49 samples). 0 are the
Knossos control samples, and the numbers are as in Table I. PC i and PC 2 account for 37 and 31 per cent of the total variation in
composition, and they are dominated by the Ti, Al, Mg, and Ca contents and by the Ni and Cr contents respectively

pottery compositions from a representative site, Thebes, wholly overlap the Knossian ranges
(Table 3). Attica is another possibility, but comparison of the compositions of Attic Hellenistic
Black Glaze43 and the Knossos control group reveals differences notably in Ca; furthermore,
the composition of the hydria found at the Agora, 15, lies outside the concentration ranges
associated with a 8o per cent level of confidence of the Attic group in Al, Ca, Mg, and Na
(Table 3).

STYLISTIC CONSIDERATIONS AND DISCUSSION

Two clear results have emerged from the analysis programme. The first is that an Egyptian
origin for Clay Ground hydriae can no longer be considered a serious possibility. The second is
that central Crete becomes a prime candidate as the true home of the workshops which
produced the majority of extant Hadra vases with this type of decoration.44
The possible claims of the other two major contenders for this role are weakened on grounds of
distribution. Not a single example of a Clay Ground hydria has ever been found in Boeotia. A
single vase from Eretria is the closest we can come geographically, and its fabric is red and highly
micaceous, little resembling that of the vases under discussion. At Athens, where excavation has

43 Prag et al. op cit. n. 38, 167-72 and table 6. 44 For Clay Ground hydriae see IHV 9 f. and BSA 75 (1980) 33-7.

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
 HADRA HYDRIAE AND CENTRAL CRETE: A FABRIC ANALYSIS II

been both extensive and intense, only two Hadra vases h

tells a different story. To the six examples known t
numerous fragments from recent excavations at Knossos
minimum of thirty to forty individual vases.*" The local
back to the Archaic period,47 and production in Crete las
Hundreds of other sherds decorated in the Clay Ground
They belong to various other shapes not found in Egypt
and dishes and there is every reason to believe that Hadr
Finally, there is evidence, discussed in detail elsewhere,5
local vases in the West Slope technique.
The most significant result, perhaps, is that sample
fabric. These three sherds were selected because they hav
work in both the West Slope and Clay Ground techniq
local history going back to at least the fifth century BC
The first is the Aliki Painter. His work is well represen
techniques may be attributed to him (PLATE 2). The
produced many of his Hadra vases.51 His characteristi
long hook-shaped tails (FIGs. 3 and 4). They are sometim
his equally idiosyncratic version of the bead and reel.52 T
a Hadra vase decorated with this distinctive pattern
The Dromeus Painter was one of the more careful and
figured Hadra vases. He also painted in the West
unfortunate that little pottery recognizably decorated by
status as a local artist is indicated by the shapes upon wh
upper right) is from the rim of a cylindrical jug, a shape
century BC. It was decorated with a dolphin in the W
characteristic of this artist: angled forehead and a bifur
line across the end of the body (FIG. 5). Sampled, too, w
Athens embellished with a line of similar dolphins aroun
Another Knossian painter who worked in both techn
painters to sign a vase.54 Lucia Guerrini gave him the fir
the first two bear his own version of the dolphin motif
From Knossos comes a selection of local shapes whic
Knossos have
45 For a list of Hadra vases found outside producedsee
Egypt hundreds
IHVn.of fragments
3. of kraters and
To these must now be added theother
many fragments
large vessels from
decorated in similar style.
Knossos, Kommos, and other Cretan 50sites.
'Knossian Artists and Ptolemaic Alexandria', in Alessan-
46 Ibid. dria e il mondo ellenistico. Studi in onore di A. Adriani (in press). BSA
47 BSA 68 (1973) 50 fig. 5 nos. L6-II; 73 (1978) 15 no. 42.(1981) 38-9, 55 f
48 Guerrini, Vasi di Hadra (Seminario di archeologia e storia 51 For example, a selection of his figured vases: IHV no. 1 7;
dell'arte greca e romana dell'Universita di Roma), Studi Miscel-
Ann. Mus. Alex. (1940-50) i8 fig. I4, 20 fig. 15; Brooklyn Museum
lanei 8 (Rome 1964) F17. Its associated grave group includes
Annual Io (Io68-9) 118 figs. 5-7.
later first century BC objects such as the piriform unguentarium 52 Clearest in Ann. Mus. Alex. (1940-50) 20 fig. 15.
and several of the jugs and amphorae illustrated on pl. Xa-e. 53 Guerrini op. cit. (n. 48) A5. For others see A3 and A6-8.
At Kommos, late Hadra vases are regularly associated with Detailed photographs on pls. 11-12: C. C. Edgar, Catalogue
pottery of the first century BC. general des antiquit6s 6gyptiennes du Musle du Caire. Greek Vases
49 For a Hadra style krater found in the district of Chalara at(Cairo 1911) pl. 14 no. 26.226.
Phaistos see AD 19 (1964) B3, 448 and pl. 524c; Annuario 45-6 54 Guerrini op. cit. (n. 48) AI-2. A3-8, also attributed to
(1967-8) 65 and fig. 9; BCH ioo (1976) 257 fig. 4. Two-columnPylon by Guerrini, have more recently been identified as
kraters found at Lato are published in BCH Ioo (1976) 254-5.works of the Dromeus Painter on grounds of style (cf. BSA 76
For a miniature oinochoe in dark-on-light style depicting an (I981) 39).
agrimi hunt see BSA 73 (1978) pl. 6 no. 87. Excavations at 55 Guerrini AI-2, pl. I.

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
12 P. J. CALLAGHAN AND R. E. JONES

A
B

E
F

G
FIG. 3- Sherds attributed to the Aliki Painter

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
HADRA HYDRIAE AND CENTRAL CRETE: A FABRIC ANALYSIS 13

B
C

FIG. 4. Hadra vases attributed to the Aliki Painter

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
14 P. J. CALLAGHAN AND R. E. JONES

amphoriskoi, juglets, kantharoi, and cylindrical jugs, a

3a, b). Sample 21, the base of a Hadra vase, is decorated w
manner common to both Pylon and the Dromeus Painter,
This also proved to be of a fabric normal at Knossos.
On stylistic grounds many other Hadra vases found in E
whose Cretan origin is indicated by the shapes they d
Together with the Hadra vases assigned to the three nam
considerable body of material which we might identify a
alone does not allow us to decide between the rival cla
combined evidence of style, fabric, and distribution
overwhelming probability that many Hadra vases we
especially Knossos. It now remains to consider the erratic
Of the three divergent types it is probably best to begi
Knossian despite the minor differences in fabric compos
subgroup C is the clearest since all the sherds in the sam
other material from the city arguably by the same hands.
of this shape and the reasons for locating the centre of m
in detail elsewhere. A very similar bowl by the indisputa
at Lyttos.56 25 (PLATE 4a-c) is from a large and ambitiou
by an artist who also worked on local West Slope shapes (
fabric indistinguishable to the naked eye from that norm
the whole group is merely what it appears: a more purif
Subgroup B is more problematic. I noted that 2 woul
though found there, was also included as an eccentric ver
of their exact provenance within central Crete must, on
Subgroup A should be regarded as foreign to Knossos
Ni in the affected samples seem to rule against a Knossia
group has been found there. On grounds of fabric ty
cluded as Knossian at the time of initial study. On st
however, we can still place them in the northern half of
its borders.
We may exclude the Mesara at the outset. Hadra vases found at Gortys,57 Phaistos,58 and
Kommos59 all have broad strap side handles. Knossian hydriae and those of subgroup A all
possess side handles which are circular in section. The Mesariote vases form the nucleus for a
larger group found in Egypt and probably of south Cretan manufacture. They are characterized
by a sometimes elegant but always unadventurous painted style in contrast to the more
ambitious and innovative products of Knossos and the better quality hydriae of subgroup A.
Lyttos in the centre of the island may also be excluded as a potential source. I have been
fortunate enough to examine its pottery through the good offices of Dr A. Lembesis and the
provinciality of its local ceramic industry is quite clear. The vast majority of its 'black-glazed'
shapes are made of a red gritty clay and are covered with a thin dull wash. Some shapes, which
would have been glazed at Knossos, are left plain. The very few finer buffbg shapes found on the

56 BSA 76 (I98I) 62 and n. 20; 'Knossian Artists and to date from the first century BC. They conform in all respects to
Ptolemaic Alexandria' (see n. 19 above). the shape of Hadra vases from earlier periods but, like similar
57 Guerrini, op. cit. E7. examples from the first century Be at Knossos, are no longer
58 Ibid. E3 and E9. decorated with figured or floral zones.
59 The examples from Kommos, as yet unpublished, appear

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
 HADRA HYDRIAE AND CENTRAL CRETE: A FABRIC ANALYSIS 15

A B

D
C
F

FIG. 5. Sherds and Hadra vase attributed to the Dromeus Painter

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
 16 P. J. CALLAGHAN AND R. E. JONES

site must therefore have been imported, at least one from Knosso
Aliki Painter.60
The pottery of Lato has been compared unfavourably by its excavators with that of Knossos.6
Neither the glaze nor fabric is of sufficient quality to be related to the hydriae of subgroup A, n
can the dull provinciality of the painted decoration be connected with the fine vases making u
the group.
To the west, the site of Ayia Pelagia, ancient Apollonia, has produced a large quantity
early second century Be pottery. Through the kindness of Professor St. Alexiou and Mrs A
Karetsou I have been able to examine many of the finds, and more is now on display in th
Herakleion Museum. The site is characterized by a mixture of fabrics, some probably Knossian
Obviously this relatively unimportant city did not have a flourishing school of ambitious vase
painters and seems to have relied on more than one of its neighbours to supply its fine table
wares.62

The two valleys flanking that of Knossos had once held the two flourishing settleme
Tritonion63 and Lykastrion.64 They appear to have been under Knossian rule for at lea
the time when vases of subgroup A were being produced. The survival of any functionin
centres in these subdued territories seems unlikely65 and it is equally improbable that i
pottery workshops should have existed there. No Hellenistic pottery has yet been note
Lykastrion on the west, but the large site of Astritsi-Tritonion in the Karteros Valley t
of Knossos has produced at least two vases now in the Herakleion Museum.66 Thr
kindness of Dr Y. Tzedakis I was able to examine both: the tondo of a Medusa Bowl and a West
Slope Olpe. Both conform to Knossian type and in no way resemble the fabric of subgroup A
hydriae. Although not decisive, this circumstance may suggest that the settlement relied on
Knossos for its fine decorated pottery.
This leaves but two likely areas in the whole north-central area of Crete. The first is
Chersonesos on the coast east of Herakleion. Although an important Hellenistic city, at one tim
in the third century under Knossian hegemony, no contemporary pottery has been noted, nor
could the local clay be sampled. It does, however, have a continuous history throughout th
period and its importance suggests it as a possible contender as the home of subgroup A.67
The second is the Herakleion Valley itself. The rich agricultural land along the Gazano
contained the territories of the two important cities of Rhaukos and Tylissos, both of which
succumbed to external aggression during the second quarter of the second century BC. Both are
possible sources for the subgroup A vases but it should be noted that none of the fine wares

century BC. The date of desertion cannot be precisely docu-

60 For a small selection of the excavated pottery from Lyttos
mented, but the two vases in the Herakleion Museum were
see AD 26 (I97I) Chron. B2 pl. 513. None of the material
identified here as Lyttian has yet been published. made about 200 BC or a little after.
61 BCH Ioo (1976) 254 ff., esp. 267. 64 Strabo x. 479 implies that the place was sacked when first
62 Praktika (I973) 208ff. and pls. 216-I9; AAA 5 (1972) 238captured by Knossos. This would have preceded the events of
fig. 9. 185 BC. For the fate of the subdued cities of the Herakleion
63 For the identification of Astiritsi with the Tritonion- Valley see TAPA lo6 (1976) 313-30.
Diatonion mentioned in the sources see C. Davaras, Die Statue65 Ibid. BSA (i98I) 63 n. 31. It is possibly significant that
the two sites were handed over to Lyttos and Rhaukos by
aus Astritsi Ant. K Beiheft 8 (1972) 30. Marinatos in Praktika
(1938) 137f. and Faure in BCH 82 (1958) also place Tritonion
Gortys in 185 Be. This might well reflect their lack of size and
in this region. This certainly makes better sense of the status at that time since otherwise an attempt to set them up as
independent
Gortynian dispositions of 185 BC (Polybios xxiii. I5). For a poleis might have been expected.
contrary view see M. Guarducci, Historia 8 (1934) 627-30. The66 For the vases from Astritsi see Annuario 4-5 (1921-2) 173
story of how St Myron was forced to quell the waters of the and BSA 45 (1950) 182.
Triton on his way to Knossos (from his bishopric?) in Acta 67 For Chersonesos see G. Le Rider, Monnaies Cretoises (Paris
1966) 280. For the later second century BC sympoliteia between
Sanctorum xxxvi. Aug. ii, implies a position to the south and west
Lyttos and Chersonesos see RE Suppl. 7 (1940) 88.
of Knossos. The city survived into the first quarter of the second

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
 HADRA HYDRIAE AND CENTRAL CRETE: A FABRIC ANALYSIS 17

collected from this area during early surveys, and now stored in the Stratigra
Knossos, at all resembles the physical appearance of that fabric.
One other intriguing possibility remains. Herakleion itself is documented as
in close alliance with Knossos in the third century BC, and incorporated w
sometime before c. 17o BC.68 As the port of Knossos, Herakleion must hav
connected with the Egyptian trade of its mistress. It would have been only na
enterprising potter to set up shop at the point of departure. Be that as it may,
indicates that subgroup A should be located somewhere in this part of the Knoss
its eastern border. These vases, then, should be regarded as 'Knossian' in the bro
word.69

P. J. CALLAGHAN
R. E. JONES

68 Guarducci, IC i. 58 and 306 for the subjugation of xxx, 23, I and IC i. 291. The new bord
see Polybios
through
Tylissos. The city is absent from the long list of Cretan the middle of the city, which cannot therefor
states
survived
allied to Eumenes II in 183 BC (IC iv. 179) and had as a functioning urban centre. Once again
probably
fallen by that time. If this is so, the later vases of comparatively
group i (see early date of its fall would tend to m
BICS 30 (1983) 'Stylistic Progression in Hellenistic againstCrete'
this city as the production centre for vases of gro
69 For
31-9) cannot have been produced there. For Rhaukos, the few testimonia on ancient Herakleion see TAPA
which
io6in(1976)
fell to a combined Gortynian and Knossian assault 314 and n. 5-
165 Be,

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
B.S.A. 80 PLATE 1

(a) (b) (c)

(d) (e)

(a-e) Samples 1-5 (photographs by courtesy of the Trustees of the British Museum)

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
PLATE 2 B.S.A. 80

(a)

(b)

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
B.S.A. 80 PLATE 3

(a)

(b)

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms
PLATE 4 B.S.A. 80

(a) (b)

(d)

(c) (e)

This content downloaded from

151.250.139.236 on Wed, 26 May 2021 21:29:36 UTC
All use subject to https://about.jstor.org/terms