Sunteți pe pagina 1din 39

A Dark Age Refuge Centre near Pefki, East Crete

Author(s): Krzysztof Nowicki


Source: The Annual of the British School at Athens, Vol. 89 (1994), pp. 235-268
Published by: British School at Athens
Stable URL: http://www.jstor.org/stable/30102572
Accessed: 22/06/2010 07:49

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=bsa.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

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.

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

http://www.jstor.org
A DARK AGE REFUGE CENTRE NEAR PEFKI,
EAST CRETE

39-42)
(PLATES

lQog ovg xaToLxoug ov Ine1xov


TOUXWQLOU
xatL ov aytLO6t[tovQO66bQO
x. Eqtavovulk Kavapdxli

THIS paper presents one component of an on-going study of the topography of dark age sites
on Crete.' This explorationof the end of the Bronze Age and the Early Iron Age on the island
has sought (I) to clarify the broader regional context of well-known excavated dark age
settlementssuch as Karphi, Kavousi Vronda and Kastro, PalaikastroKastri, and Vrokastro;(2)
to establish the location, function, and chronology of previously discovered but poorly
documented sites; (3) to augment this catalogue of known dark age remains with new
evidence; (4) to define patterns of settlement and land use for the period in various areas of the
island; and (5) to establish chronological and regional patterns that may help us begin to
understandthis obscureperiod in Crete.2
The Dark Age on Crete is potentially better representedthan in other areas of the Aegean,
and explorationsin the mountainous hinterland of the island have shown site numbers and a
density of habitation only vaguely suggested by the excavated sample of sites. This study of
dark age topographyhas shown the likelihood that a more complete or representativemap of
Crete might be drawn for the Dark Age than for any other period.
At the beginning of this century three sites representingdark age defensible settlements in
Crete were partly excavated:Kavousi Kastro, Vrokastro,and PalaikastroKastri.3The list of
explored sites was later complemented by a more comprehensive project undertaken by
Pendlebury at Karphi.4 In the I96os new excavations at Palaikastro Kastri revealed
particularly valuable evidence for the problem of the beginning of the Cretan Dark Age.5

I
I am most grateful to the Greek archaeological 2 The problem of defensible sites in Crete, and their
authorities, in particular Dr K. Davaras, for the permits to topography, distribution and evidence as recorded on the
carry out my studies and to draw the plans of the sites surface,will be presented in the forthcoming Defensible
Sitesin
presented here. The work would not be possible without Crete(LMIII c-Geometric).
extremely gracious help and friendship of the inhabitants of 3 H. A. Boyd, 'Excavations at Kavousi, Crete, in I9oo',
Pefki. First and foremost my thanks go to the mayor of this AJA 5 (1901), 125-57; R. C. Bosanquet, 'Excavations at
village, Mr Emmanouil Kanavakis, for his interest in my PalaikastroI', BSA 8 (I901-2), 286; E. Hall, 'Excavations in
work and for his and his wife's Cretan hospitality. I would eastern Crete: Vrokastro', Universityof PennsylvaniaMuseum
also like to thank his brother,Mr Georgios Kanavakis,for all Publications,
Anthropological 3. 3 (Philadelphia,1914).
the information he has given about the archaeology and 4 J. D. S. Pendlebury and M. B. Money-Coutts,
history of the Pefki area, and Mr Nikos Tavladakis for his 'Excavations in the plain of Lasithi, III: Karphi, a city of
hospitality. I would like to express my special thanks to refuge of the early Iron Age in Crete', BSA 38 (1937-8),
Victoria Batten for improving the English of this paper, and 57-145.
to Donald Haggis for many very helpful comments, 5 L. H. Sackett, M. R. Popham, and P. M. Warren,
particularlyconcerning the pottery. 'Excavationsat Palaikastro,VI', BSA6o (1965),269-305.
236 KRZYSZTOF NOWICKI

Besides material from stratifiedcontexts in settlements, there were short reports or notes on
cemeteries (e.g. Erganos, Panagia Prophitis Ilias, Kourtes, Kamares, Siderokephala,Praisos,
Atsipades,Driros, Vronda),6unexcavatedsettlements (e.g. ZakrosEllinika,Kandilioro, Oreino
Ellinika, Stavrochori Kastri, Monastiraki)7and unstratified or very disturbed layers of the
Dark Age unearthed during the investigations at later sites (e.g. Prinias, Ligortino, Panagia
Prophitis Ilias, Elefterna).8Such was the basis for studies of the Dark Age in Crete beyond
Knossos.
Although Pendlebury proposed the reconstruction of events after the collapse of the
Minoan-Mycenaean civilization, he still had too little evidence to support his ideas or to
resolve some contradictionshe found between differentareas. Additionally,the chronology of
the post-Minoan period was very confusing, and the most important aspect, the definition of
LM III c, was not yet established.9Nevertheless the list of refuge or defensible settlements
proposed by Pendlebury remained the most comprehensive one for a long time,'l and was
complemented only by a few sites identified in the 196os and 1970s11Pendlebury'swork was
never fully explored by later scholars, and many of the sites seem to have never been visited
again. Desborough and Snodgrass, when they presented their synthetic books on the Dark
Age in Greece, did not include all the identified sites in their discussions.'2
A new chapter of research on the Cretan Dark Age started with two American projects
covering the areas around Kavousi and Vrokastro.'3The material from the re-excavatedsites

6 F. Halbherr, 'Cretan Warren, 'Ancient sites in the province of Agios Vasilios,


expedition, XI: three Cretan
necropoleis: report on the researches at Erganos, Panagia Crete', BSA 61 (1966), 178.
and Kourtes', AJA 5 (1901), 259-93; A. Taramelli, 'Cretan 12Pendleburygave, for example, a much better context for
expedition, XII: notes on the necropolis of Courtes', AJA 5 Karphi as a defensible settlement. A. M. Snodgrass writes:
(1901), 294-301; L. Mariani, 'Cretan expedition, XIII: the 'this site is shown by its pottery to have been curiously aloof
vases of Erganos and Courtes', AJA 5 (1901), 302-14; A. from many contemporary developments in the island' (The
Taramelli, 'Cretan expedition, XX: a visit to the grotto of DarkAge of Greece(Edinburgh, 1971), 249); and further on:
Camares on Mount Ida', AJA5 (1901),439-42; A. Taramelli, 'Karphi was such an isolated site . . . The evidence of
'Ricerche archeologiche cretesi', Mon. Linc.9 (1899), 403; F Karphi must, as always, be put forward with qualification,
Halbherr, 'Cretan expedition, XVI: report on the researches not only because of its general cultural isolation .. .' (ibid.
at Praesos', AJA 5 (1901);R. C. Bosanquet, 'Archaeologyin 371). He identified the most important links with the outside
Greece, 1900-1901',JHS 21 (1901),399-400; id. 'Excavations world as being with Cyprus, in spite of the fact that they
at Praesos, I', BSA 8 (1901-2), 231-70; E. Petroulakis, were based on a very few items, whereas the connection of
'KplLxij 'ATotatdbagdtpoL', Arch.Eph. 1915,48-50; Boyd Karphi with other Cretan centres is supported by numerous,
(n. 3), 131-6. much better-grounded factors. On the other hand, V. R.
7 L. Mariani, 'Antichithcretesi', Mon. Linc. 6 (1895), 293; Desborough does not want to see Karphi in 'cultural
D. G. Hogarth, 'Excavations in Zakro, Crete', BSA 7 isolation', writing that 'One of the most interestingpoints is
(1900-1), 145;J. D. S. Pendlebury, The Archaeologyof Crete the evidence provided for close contact with other parts of
(London, 1939), 178, 290, 385; H. A. Boyd, 'Gournia', Crete, at least for some periods, and also for overseas
Transactionsof theDepartment
ofArchaeology,FreeMuseumof Science contacts' (The GreekDark Ages (London, 1972), 127-8).
andArt, University i (Philadelphia,1904), 17-18.
ofPennsylvania, Desborough too, however, did not mention other sites,
8 Pendlebury(n. 7), 313-16. similar to Karphi and already identified by Pendlebury,
9 Pendlebury often mentioned the sites as Subminoan or when discussing this site's political role on the basis of its
PG, but their contemporaneity with Karphi allows them to geographical location.
be redated to LM III c-PG. 1' G. C. Gesell, L. P. Day, and W. D. E. Coulson,
0oPendleburynever published a proper list of refuge sites 'Excavations and survey at Kavousi 1978-1981', Hesp. 52
in dark age Crete, but numerous remarksare scattered in his (1983), 389-420; iid., 'Kavousi, 1982-1983: the kastro', Hesp.
works. 54 (1985),327-55; iid., 'Excavationsat Kavousi, Crete, 1987',
" N. Platon, PAE 1956, 239-40; P. Faure, 'Nouvelles Hesp.57 (1988),279-301; iid., 'Excavationsat Kavousi, Crete,
recherches de spe16ologieet de topographie cr6toises', BCH 1988', Hesp.6o (i991), 145-78; B. Hayden, 'New plans of the
84 (1960), 196-219; id., 'Cavernes et sites aux extr6mit6sde early iron age settlement of Vrokastro', Hesp. 52 (1983),
la Crfte', BCH 86 (1962), 39-41; M. S. E Hood and P. M. 367-87.
A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE
237

of Vronda and Kastro has become vitally important for the chronological development of the
pottery,while the surveysaround Kavousi and Vrokastrohave yielded valuable informationon
environmentaland geographical factors as related to the settlement pattern.'4 Both projects,
however, cover restrictedareas on Mirabello Bay, and cannot resolve all questions concerning
the history and characterof the Dark Age in Crete.
More recently, selected Greek Dark Age problems were presented in brief by Coulson.'5
Several questions were posed by this American scholar to illustratehow complex the problem
is and how restricted the material available for further studies. Crete is again discussed
together with the rest of Greece; the island shows many individual characteristics,however,
which should be studied before a new synthesis on this period is presented. New field
investigations lead to the conclusion that many of the key questions must be revised and
analysed on the basis of a wider range of archaeological evidence. For example, the
chronological sequence (particularlythe problem of the beginning of the Dark Age in Crete),
settlement pattern changes throughout the LM III C-Geometric periods, the problem of
immigration and emigration, the development of pottery styles, the partition of Crete into
different political and cultural regions, and the interrelations of the main settlement centres
within the island and their relations with the outer world, are all important issues which must
now be addressed.
The problem of pottery,for example, cannot be restrictedto the fine ware. The coarse ware
appears at some points to be even more helpful, and should not be treated as non-diagnostic
and, as is often the case, excluded from the final publications. The fabric, inclusions,
treatment of surface, and decoration are very important indicators of chronological and
geographical differences.'16It is too early to present here more comprehensive conclusions
about the variation in this kind of pottery, but some preliminary remarks, which will be
illustratedin a more detailed way elsewhere, may be helpful.'7In general the coarse ware, at
least in the mountainous areas, shows characteristics which can be used to show some
distinction between LM III B and c. This phenomenon may have been connected with the
moving of pottery workshops from lower settlements, with a long tradition of ceramic
manufacturing,to the defensible settlementssituated in a different geographical zone. In some
areas, however, particularly where there was a continuity of settlement, as was the case at
Knossos and other central Cretan sites, and perhaps in Chania, the differences between the
LM III B and C coarse ware may be less recognizable. In the mountainous sites the coarse
ware seems to have changed very little during the LM III c-SM/PG period, and certainly it
changed more in shape than in fabric. The next key point in the sequence of the coarse ware
appears somewhere between the late Protogeometric and Geometric periods: the fabric
became much harder, the clay better-fired,the inclusions finer and more homogeneous, and
the range of colour variations more restricted,while new motifs of incised decoration quickly

"4D. C. Haggis, TheKavousi-Thriphti Survey:An Analysisof Survey project, 1986-1989', Hesp.61 (1992),293-354.
Settlement
Patternsin an Areaof EasternCretein theBronzeAgeand 5 W. D. E. Coulson, The GreekDarkAges:A Reviewof the
Early IronAge (Ph.D. diss., Univ. of Minnesota, 1992); id., EvidenceandSuggestionsfor
FutureResearch (Athens, 1990).
'Survey at Kavousi, Crete: the iron age settlements', AJA 95 16D. C. Haggis and M. S. Mook, 'The Kavousi coarse
(Iggi), 291; M. S. Mook and D. Haggis, 'The wares: a bronze age chronology for survey in the Mirabello
Kavousi-Thriphti Survey,1988-1989', AJA 94 (1990), 323; B. area, East Crete', AJA97 (1993), 265-93.
J. Hayden, J. A. Moody, and 0. Rackham, 'The Vrokastro 17Nowicki(n.2).
238 KRZYSZTOF NOWICKI

replace the old ones. The latter changes seem to coincide with the considerablechanges in the
settlementpattern of Crete.
Another point worth a brief mention is the problem of the regional partition of the island.
The picture appears to be more complicated on the basis of new field investigationsthan that
presented by Desborough and Snodgrass.'8Crete was probably divided into more than two or
three areas, and East Crete was not a political and cultural unit. With the collapse of
Minoan-Mycenaean civilization Crete entered the period of partition, and this process
deepened throughout the period in question. On the other hand, at some point and in some
areas the process was reversed, probably in the Protogeometric, but certainly by the
Geometric period. Partition and unification (the latter based on a new political, social, and
perhaps ethnic identity) shaped the new map of Crete during the LM III c-Geometric
periods, and at the end it reveals large towns, probably expanding their outer territoriesby
force, and old refuge settlementseither being abandoned or under pressurefrom outside. Such
a picture brings us, of course, to the question of Dorians, the second major problem pointed
by Coulson in the work mentioned.'19This, however,cannot be resolved before revealing local
variation in material culture, and before the archaeological situation of dark age Knossos, in
particular,and of central and northern Crete in general, is understood. The subject of dark
age Knossos is much more complex, and at present it should be separated from the problem
of the defensible sites in the Cretan mountains.20
To understandthe actual role of defensible settlementsin darkage Crete, one must consider
how many have already been identified and how few are usually discussed in connection with
the subject. The most comprehensive study of the LM III material was published in the
excellent book of Kanta.21Even here, however, the problem of defensible sites is limited to a
few sentences. Other works usually do not mention as many sites as Kanta.22However, how

's Desborough(n. 12) was able to point out only very regions of Crete, e.g. the w Siteia mountains and the
generaldifferencesbetweencentraland E. Crete,andthe w Rethymnon isthmus.
part remainedfor him a blankspot (pp. I13, 115-17, 225, 19Coulson (n. i5).
234-5). Snodgrass(n. 12)proposeda more detailedpicture 20 The picture revealed by two important excavations
when writing about the border between central and E. beyond the palace in Knossos is somewhat different. The
Crete:'the borderbetweenthe two may be placed at the Unexplored Mansion area shows continuing, but restricted,
defileon the roadfromMalliathroughDreros,closeto the occupation during the LM III B, LM III c, and SM phases
moderntownof Neapolis'(p. 164).It is a pitythathe did not (M. R. Popham, 'The Unexplored Mansion at Knossos: a
drawthe restof thisborderor explainhow andwhenit was preliminaryreport on the excavationsfrom 1967 to 1972',AR
formed.MaterialfromLM III c sites on both sidesof S.'s 19 (1972-3), 59-61). The StratigraphicalMuseum excavations
borderis verysimilar,but not that of the PG or G periods. show this area to be abandoned by late LM III B, but clearly
This phenomenonfitsverywell into the generalchangesin reoccupied in LM III c (P. M. Warren, 'Knossos:
the settlementpatternof Creteat the turn of LM III c (or StratigraphicalMuseum excavations, 1978-82, part III', AR
Subminoan)and PG. S. may be rightto pointto the special 29 (1982-3), 69-83). The problem of DA Knossos was
roleof the Neapoliscorridorin the expansionof the central recently summarized by J. N. Coldstream ('Knossos; an
Cretan element towards the E, and to some political urban nucleus in the Dark Age?', in La transizione del miceneo
borderlineat thisplace.The lattercan be drawnaroundthe dalpalazzoalla citth(Roma, 1991),287-99).
all'altoarcaismo:
whole Lasithi range, with the outer settlementsquickly 21 A. Kanta, TheLateMinoanIII Periodin Crete: A Surveyof
developinginto largetowns(PanagiaProphitisIlias,Lyttos, Sites,PotteryandtheirDistribution
(SIMA58; G6teborg, 1980).
Kalo ChorioMaza,Anavlochos,Driros,Lato)and the inner 22 e.g. Whitley mentions only 5-9 sites (including
Lasithiansettlements(KeraKarphi,GoniesTo Phlechtron, non-defensible ones) for SM, 8-12 for PG (A.J. M. Whitley,
AdrianosFortetsa,KritsaKastello,Tapes Kastello,Zenia Style,Burial and Societyin DarkAge Greece:Social,Stylisticand
Kastrokephala, Erganos) experiencing decadence or MortuaryChangein the Two Communities of Athensand Knossos
abandonmentand moving to other locations (e.g. Kera betweennioo and 7oo Bc (Ph.D. diss., Univ. of Cambridge,
Papoura).A similar process can be also noted in other 1986), 261).
A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE 239

these sites relate to the more than one hundred defensible settlementsso far identified for the
LM III c-G periods in Crete needs to be addressed.23At least seventy can be dated to the
early phase of the Dark Age (LM III C-early PG), and this is apparentlyonly a fraction of all
sites of this type once scattered throughout the island. For some areas the picture must be
completely changed, as in the case of West Crete, where typical defensible or even refuge
settlements were identified in both north and south (Rokkaand Agios Georgios Kastri).24We
can say, therefore,that the threat which forced most of the Cretans, at some point in late LM
III B and at the beginning of III C, to abandon their old settlementsand look for safety on the
summits of easily defended mountains was similar in the Zakros bay and Kissamos bay, in
Mirabello and Sphakia. Local differences, like the continuity of settlement in Knossos,
Chania, and some parts of Mesara, may reflect more the strength of particular communities
than the differentcircumstancesin differentparts of Crete.25
The problem of the Dark Age in Crete cannot be revised until more new archaeological
evidence is published. The very limited scope of available information, despite being a
starting-pointfor many ideas arising from the first excavations at early dark age sites, has led
to a situation where each criticism of an earlier theory contributes no more than the theory
itself.6 The only way to give a new impetus to further research on the period in question
seems to be field studies in the Cretan mountains, as a supplementaryproject to the limited
excavationsstarted some years ago.27The newly identified sites near Pefki (Siteia)are the most
recent result of the project which previously concerned the area directly to the west and has
alreadybeen published.28

GEOGRAPHY
Pefki is situated on the southern outskirtsof the Romanati massif, in the SEpart of the west
Siteia mountains (FIGS.I-3).29 Romanati has a similar geological structure to most of the East
Cretan mountains: secondary (Jurassic) limestones overlie an earlier (Permian, Triassic)
schist/phyllite core. The shape of the landscape is primarily due to erosion. The main massif
is cut by many gullies and small gorges; the outer faces of particularrocky islands form high
cliffs. The karsticprocess can be found in crevices and caves, particularlyin the southern part
of Romanati. Seismic activity may have been responsiblefor the most dramaticchanges in the
local landscape, as in the vicinity of Pefki.
The Romanati massif is not as rich in water as the mountains to the west. Springs are
limited to a few areas and appear in a usual geological situation, in the zone between the

23 All these sites will be presentedin detail in Nowicki (n. 2). investigations and the new facts yielded by them indicate
24 K. Nowicki, 'Report on investigations in Greece, VIII: that there are many fewer contradictions than the facts
studies in g99i', (Warsaw),43 (1992),118-19.
Archeologia seemed to imply twenty years ago when D. presented his
25Warren(n. 20), 83. interestingbut controversialpaper.
26 See e.g. V R. Desborough, 'Crete in the first half of the 27 This remark concerns two DA sites: Kavousi Vronda
twelfth century BC:some problems', in Proceedings of the Third and Kavousi Kastro;see reports on the excavationsby Gesell
CretologicalCongress(Rethymnon,18-23 Sept. 1971), vol. A I et al. (n. 13).
(Athens, 1973),62-9, esp. the author's remark 'This is not a 28 K. Nowicki, 'The west Siteia mountains at the turn of
matter of adding to your knowledge of the facts; it is my the Bronze and Iron Ages', Aegaeum,6 (I99o), 161-82.
purpose to bring to your notice certain problems, and even 29 On the geography of the w Siteia mountains see
contradictions, that the facts imply.' Recent field Nowicki (n. 28), 16I-6.
KRZYSZTOF NOWICKI
240

FIG.i. Defensiblesettlementsin LM III c-PG Crete (selection).I. Milatos,2. VrachasiAnavlochos,3. Driros,4. Zenia
Kastrokefala,AdrianosFortetsa,Tapes Kastello,5. Lato, 6. KritsaKastello,7. Vrokastro,8. KalamafkaKastelloand
AnatoliMesokastella,9. Mythi Kastello,io. Arvi Fortetsa,ii. Keratoand LoutrakiKandilioro,12.ErganosKefali, 13.
AfratiProfitisElias, 14.Lyttos,15.Kera Karfiand Gonies To Flechtron,16. Kalo Chorio Maza, 17.ProfitisElias/Lykastos,
18. KrousonasKoufo, 19. PriniasPatela,20. LigortinosKefala,21. KastellianaKastello,22. Rotasi Kefala,23. PobiaVigla,
24. Gortys,25. KourtesKefala,26. Axos, 27. PandanassaVeni and Thronos Kefala,28. Spili Vorizi,FratiKefala,and
AtsipadesFonises,29. MirthiosKirimianou,3o. Ag. GeorgiosKastri,31.VrysesAg. Georgios,32. Rokka.

500 - 700 m

700 - 100 m

LAND OVER 1000m

Kastri,2. ZakrosGorgeKato Kastello,3. ZakrosGorgeEllinika,4. SfakiaKastri,


FIG.2. Defensiblesettlementsin E. Crete.I. Palaikastro
5. Praisos,6. ChandrasVoilaKastri,7. KryaAg. Georgios,8. MyrsiniKastello/Ellinika,9. TourlotiKastri,IO.Pefki:Stavromenos,
Kastellopoulo,and Mega Chalavro,II. Ag. StefanosKastello,12.ChrisopigiKorakia,13.Oreino:Kastri,Kato Ellinika,and PanoEllinika,
14.Ag.IoannisPsychro,15.KoutsounariKarfi,16.VainiaStavromenos,17. Kato ChorioProfitisElias,18. Monastiraki: Chalasmenoand
Katalimata,19.Kavousi:Vronda,Aloni,and Kastro,20. Asari,21.Palaikastro PlakaloniaKalamafka.
A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE 241

.
12
Jr

^
t '
i "
.
".r%

)".- -"c-".

71

, 5 80
.. .. 0.0.

IPF
f

% %1

),
5, Ioo ro te

; / .
/

" / 1000
1
arhelogclst
A ~; - \ [

" a
acheoooca'st
'50

routes

seee

"" KitYA , ."


242 KRZYSZTOF NOWICKI

limestone and schist layers. Many are seasonal; in some areas the water-table elevation has
sunk, and today the springs yield far less water or have even completely dried up. The
existence of springs around Romanati was probably the most importantfactor in the location
not only of early settlements, but also of medieval villages. The Venetian (Pefki)or Turkish
(Agios Stephanos) fountains are still in use, though today more water is drawn from deep
wells. The intensive watering based on the new deep wells is responsible for changes in the
vegetation zones of this area. More olive trees are visible around Romanati than several
decades ago.
The expansion of olive groves has caused not only deforestation, but also the decline of
grain cultivation and even of vineyards. The reduction in grain cultivation is typical of the
whole island, and the process has continued especially since the second world war. A large
area over the top and higher slopes of Romanati is dominated by pasturage. Several mandras
are scattered around the massif, which serves as a main summer herding area for the
shepherdsfrom Daphni, Lithines, and Agios Stephanos. The animals are driven down for the
winter to lower plains like that around Adromyloi or above Analipsi-Makrygialos.The great
number of abandoned and partly destroyedmandras indicates that on Romanati the herding
pattern has undergone considerablechanges during the last few decades. In general, herds are
much bigger but fewer than in the past few generations (this is a characteristicof the whole of
Crete). The number of animals is usually higher than can be supported by the physical
resourcesof the area. The last few dry years have revealed this problem most dramatically.
There are three inhabited villages and one deserted village around the Romanati massif
(FIG.3). Agios Stephanos and Pefki are located on the south side at an altitude of c.400 m,
Daphni on the north slope at c.6oo m. All three villages were mentioned by Pashley as having
the following populations: Ag. Stephanos (Gras) 17 families, Pefki 21 families, Daphni 19
families.30Between Daphni and Agios Stephanos, at c.6oo m above the sea, is the abandoned
and partly destroyed village of Aori. According to the locals, Aori was inhabited, in part at
least, by families who had houses in other villages as well, for example in Daphni. The
travellingtime between the villages range from half an hour (Pefki-Ag.Stephanos)to one hour
(Ag. Stephanos-Daphni, Pefki-Daphni), and all were connected with each other by kalderimia
which followed the higher slopes of Romanati or ran directly across the massif (FIG.3). Many
eroded and destroyed remains of such kalderimia,or simply pathways, can be still seen at
places. Better-constructedkalderimia connected the villages with the 'outer world', for example
the routes from Daphni to Siteia and from Agios Stephanos to Makrygialos.
These paths and kalderimiawere of very local significance,and were apparentlylaid out at
a distance from the main transitroutes. The latter ran around Romanati, followingvalleys and
plains which, with their gentler contours, made travel more comfortable.To the north ran the
eastern extension of the Avgo-Roukakka line, which connected the Mirabello area with the
Siteia plain via Daphni and Krya. Along the southern coast ran an easy route from lerapetra
to the Makrygialos-Goudouracoastal plains; behind Analipsi it branched off, and its northern
line entered the Lithines corridor which gives the easiest access to Siteia on the northern
coast. Of all the aforementioned villages, only Daphni was located close to, and directly
above, the main transit route. It seems, however, that during the Venetian and Turkish

30 R. Pashley,Travelsin Crete(London, 1837),i. 322.


A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE 243

periods, if not earlier, the south coastal line played only an insignificant role in lines of
communication, and that the main route was moved up because of the high location of Agios
Stephanos and Pefki. The coastal area in general was regarded as poor and unprofitablein
comparison to the fields around the villages. The process has been reversed during the last
twenty or thirty years.

EARLY OCCUPATION

Very little is known of the ancient settlement on the massif of Romanati, and so far this
mountain is a blank spot on the archaeologicalmap of Crete. The east Siteia plateau abounds
in evidence dating back to the Neolithic period and the Early and Middle Bronze Ages.
Several sites from these periods are also known from the areas to the north, south, and west.
We can suppose, therefore, that the early Cretan settlers may have penetrated Romanati as
well. If we accept the idea that the geographical setting of human habitations in Crete
depended not only on environmentalfactors, but also (and in some periods first and foremost)
on the political and social situation, we can suggest some topographical criteria for searching
for archaeological sites. All the dark age settlements discussed below may prove that this
method works.Can we also use it in looking for earliersites on the Romanati massif?
Previous archeologicalinvestigationsmentioned Neolithic in one of two caves situated near
Pefki, namely Sto Vreiko (FIG. 3, C I). There is no published material, however,and neither the
dating nor the interpretation of the finds can be proven at the moment.3' The second cave,
Apaloustres (FIG. 3, C 2; FIG.4), was not claimed as an archaeological site, but I recorded a
scatter of probable EM sherds 30 m from the entrance on a visit in 1992 (FIG.4. 14).
We have much better evidence for the Early Minoan period. It comes from the site which is
situated on the very southern edge of the Romanati massif, between Makrygialosand Agios
Stephanos (FIG.3. 2). A small settlement probably of EM I-II date was located on a steep
slope below a rocky knoll which rises beside the approach to the deep valley some 2 km from
the seashore. It can be compared to similarsites on the south coast of Crete. A somewhat later
site is located on the SEedge of Romanati, c.I km NE of Pefki (FIG. 3. 9).32Here, on the summit
of an isolated rocky knoll, Kastello, which is in fact divided into two separate parts, are the
sparse remains of a Middle Minoan I-II site (farmstead or group of houses). The question
arises whether Kastello was a proper habitation place, or a kind of defensible strongholdused
in case of threat by the people living around it. An extensive Minoan settlement is situated
c.5oo m south of Kastello at Akra (FIG.3. 8). The place is defensible on the w and sw sides
(high cliff), but is rather open on the other sides. Sherds visible on the surface seem to
represent mostly the Neo-palatial period, but fragments dating to the Proto-palatial period
can be also seen.
Another Middle Minoan settlement is situated c.I.5 km SE of Kastello (bearing I 17 deg.) at
Kieratia Lagos (FIG.3. Io). Architecturalremains, possibly a number of houses built of large
unworked blocks, cover an area c.6o m in diameter on top of a gentle hillock, which is
encircled on the west by a natural wall of rocky boulders. Some stones, however, may have

31 N. Papadakis, Siteia,Fatherlandof Mysonand Korneros: A shown to me by G. Kanavakis, to whom I also owe all the
andCulturalGuide(Siteia, 1983), 73.
Historical,Archaeological information concerning the history of the destruction of
32 The sites of Kastello, Akra, and Kieratia Lagos were these sites.
244 KRZYSZTOF NOWICKI

FIG. 4. Pefki area (for the sites see text).


A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE 245

reinforced this natural fortification, making it slightly higher. The site of Kieratia Lagos is
partly destroyed by an electricity pylon. Several sherds visible on the surface belonged to the
Protogeometric-Geometric period, but most of the material dates back to Middle Minoan.
Kieratia Lagos seems to have been a border site on the eastern edge of the Romanati massif.
The MM III-LM I period has yet to be identified amongst numerous archaeological sites
scattered around the lower slopes of Romanati. One of the most promising areas is that of
Adromyloi, where a very extensive settlement located on the summit and steep eastern slopes
of To Marino To Kephali (FIG. 3. 13) suggests the local centre of MM-LM occupation.
Another MM-LM I settlement was recorded in the vicinity of Daphni at a place called
Xenotaphia, c.I km E of the village (FIG.3. 12). Its topographical setting is very typical of sites
inhabited during the new phase of the Minoan culture which began in the MM III period.
Diagnostic pottery, including sherds representing a fine-ware class, indicates that the
settlement flourished during LM I. To the same period may belong some of the sites
mentioned by the inhabitants of Pefki in a strip along the slope of Romanati below the village.
This information, however, should be checked by proper field investigations. In the area of
Pefki we were able to identify at least two LM sites. The first is situated directly south of the
village at Kato Geitonia (FIG.3. 4; FIG.4), which is related to the local tradition of a Byzantine
and Venetian predecessor of Pefki.33 LM sherds seem to predominate among the prehistoric
material, though several fragments may be dated to the Dark Age and quite a lot show the
same local variation of fabric as those from the settlement of Mega Chalavro (see below).
Because of the long occupation during the medieval period, neither the character nor the
extent of the site can be easily reconstructed. The second LM settlement was situated on the
aforementioned hill of Akri (FIG.3. 8). LM I occupation is well attested along the southern
coast, where a Minoan villa was excavated and a probable settlement recorded at
Makrygialos.34
The changes in the settlement pattern between LM I and III have yet to be studied for
particular geographical areas, but it seems that this part of Crete was in general similar to the
rest of the island. Coastal sites were occupied throughout LM III A-B. Some settlements or
towns connected with maritime trade may have even developed by LM III A and early III B.
In this group we should mention Diaskari, which is in a similar geographical position on the
southern coast to Palaikastro Roussolakkos on the north coast. In LM III A-B one can observe
a striking concentration of graves within well-definied cemeteries, which are often situated
inland and are not associated with settlements. Such a group of cemeteries is represented by
the sites between Adromyloi and Sykia.35 Its general geographical situation resembles that of
the cemeteries near Episkopi on the lerapetra isthmus, or Armenoi in the Rethymnon district.
All three are situated inland, one to two hours' walk from the seashore, in an area which has
no earlier tradition as a strong settlement centre. To whom did these cemeteries belong?

33The historical tradition of this village goes back to pre- instructivejourneys to the area of Kato and Pano Geitonia,
Saracen times. It was sacked by the Saracens, but reoccupied and for all the information about it.
during the Venetian period and destroyed again by the 34The existence of a settlement near the excavated villa at
Turkishcorsairs at the end of the I5th cent. After this event Makrygialoswas not ruled out by Dr K. Davaras during the
the village is said to have been moved to the present-day 8th International Symposium at the Swedish Institute at
location of Pefki. Surface archaeological evidence seems to Athens, 6-9 June 1992 (paper on 'The function of the
support such a history for this medieval village, whose name Minoan villa').
is unknown. I am much obliged to E. and G. Kanavakis for 35Kanta (n. 21), 185.
246 KRZYSZTOF NOWICKI

There are most probably two possible answers. The first is that an extensive settlement or
settlements are yet unidentified somewhere in the vicinity of these cemeteries. The second
hypothesis is more tentative and is based on the general picture of LM III B settlement, which
seems to have been much more scattered than in the earlier (LM I and LM III AI) periods; on
this hypothesis the cemeteries may have been common burial grounds used by people who
inhabited isolated houses or groups of houses over a wide surrounding area.
At the beginning of LM III C or even at the end of LM III B there was a sudden breakdown
of the settlement pattern in Crete. While the picture of settlement in LM III B-c is still far
from clear, and while the changes in settlement pattern are probably more complex than
usually presented, in general low-lying settlements and towns seem to have been abandoned
and people fled to the summits of neighbouring mountains.36 In the west Siteia mountains the
process is reflected in a series of refuge sites grouped in areas offering safety and the natural
conditions necessary for the existence of any place of habitation (FIG. 2).37 Strong centres were
identified above Kavousi, Monastiraki, Kato Chorio, Vainia, Koutsounari, and in the Oreino
valley. The same phenomenon took place along the north coast of Mirabello Bay and the
Siteia peninsula. The LM III c-Geometric settlements at Tourloti Kastri and Myrsini
Kastello are only elements in a more complex system along the northern slopes of the Orno
range. In the Romanati massif, which is the eastern extension of the Siteia mountains, the
only site dating to the Dark Age was Agios Stephanos Kastello.38 The field investigations
undertaken in 990o-I near Pefki revealed evidence which sheds new light on the problem of
the Dark Age in this part of Crete.

STAVROMENOS

THE SITE
Stavromenos is the mountain which directly towers above the village of Pefki on the NW(PLATE
39 a-b). Its name comes from the chapel which was erected, according to the local tale, to
prevent the Turks from building a fort in this dominating position. Today an easy path leads to
the summit, but the rocky massif must have always been considered an ideal natural refuge
place.
The summit itself is 8o m long and 20 to 30 m wide, and consists of several terraces. The
highest terrace (FIG. 5. 2, 5), 12 to 20 m wide and 30 m long, is occupied in its s part by the
chapel and is nearly completely eroded in the N part, exposing a bare rocky surface. A thicker
accumulation of soil, and perhaps some architectural remains, can be expected on the lower
terraces (FIG. 5. 3-4, 6). All the aforementioned terraces are encircled by a high cliff which
gives the only access on the Nw side. Here the cliff forms several natural steps which were
additionally transformed into terraces supporting a 'stairway' leading to the chapel (FIG.5. 8).
Although no walls were observed on the steep w slope, the spot may have accommodated
several architectural units.
The main settlement, however, was located on the E slope of Stavromenos (FIG. 4. 2, 5)-

36 For the general characteristic of this process see K. 37 For the situation in the w Siteia mountains see Nowicki
Nowicki, 'Topography of refuge settlement in Crete', (n. 28), 16i-8o; id. (n. 24), 116-18.
JahrbuchderRiimisch-germanischen Zentralmuseums Mainz, 34 38Papadakis(n. 31), 73-
(1987), 213-34.
A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE 247

FIG.5. Plan of Stavromenos.

The remains of buildings are visible on the spur about 40 m below the summit. In two or
three cases house plans can easily be drawn. The lower settlement extended onto the slope w
and sw of the spur, where the surface is very eroded. We can guess that this area was arranged
in a series of narrow terraces which once supported proper buildings. There is, however,
nothing but very destroyed terrace walls to support such a reconstruction. Sherds are scattered
over an area about 8o m in diameter, indicating an approximate size for the lower settlement.
248 KRZYSZTOF NOWICKI

CATALOGUE OF SHERDS FROM STAVROMENOS (PLATE 42b)


x. Tripod leg. Clay brown, coarse phyllite inclusions, tempered; herring-bone incised decoration in
straw-tempered;three shallow finger-impressions(FIG. simplifiedfashion (shallowincisions)(FIG.16).
i6). ix. Body fragment. Clay pale brown with grey core,
2. Tripod leg. Clay brown, v. coarse phyllite coarse phyllite inclusions, much straw-tempered;
inclusions, much white quartzite, straw-tempered;five herring-bone incised decoration in simplified fashion
vertical grooves (FIG.16). (shallowincisions)(FIG.16).
3. Body fragment. Clay orange, pinkish surface and 12. Body fragment. Clay orange-pinkish with grey
grey core, coarse phyllite inclusions, white quartzite core, coarse phyllite inclusions, white quarzite grits,
grits, straw-tempered; herring-bone incised straw-tempered; herring-bone incised decoration in
decoration (FIG.16). simplified fashion (shallowincisions) (FIG.16).
4. Body fragment. Clay orange-pinkish, coarse 13. Body fragment. Clay: light buff surface with
phyllite inclusions, white quartzite grits, straw- brown core, coarse phyllite inclusions, white quartzite
tempered; horizontal band with finger-impressions grits, much straw-tempered; horizontal band of
(FIG.16). incised spiral circles (FIG.17, S 13).
5. Clay light, orange-buff, coarse grey phyllite x4. Rim fragment of pithos or pithoid jar. Clay:
inclusions; herring-bone incised decoration in a more brownish pink surface with grey core, coarse grey and
elaborate fashion (FIG.i6). brown phyllite inclusions, white quartzite grits, straw-
6. Rim fragment. Clay pale brown, coarse phyllite tempered (FIG.17, S 14).
inclusions, white quartzite grits, straw-tempered(FIG. 15. Rim fragment of pithos or pithoid jar. Clay dark
16). brown with grey core, coarse phyllite inclusions,
7. Rim fragment of basin (?).Clay reddish brown with straw-tempered(FIG.17, S 15).
grey core, coarse grey phyllite, white quartzite grits, x6. Handle fragment. Clay reddish brown, coarse
straw-tempered(FIG.16). phyllite inclusions, white quartzite grits, two deep
8. Base fragment. Clay orange-red, coarse purple finger-impressionsat junction with body (FIG.13, S 16).
phyllite inclusions (up to 8 mm) (FIG.16). 17. Body fragment. Clay light orange-pinkish, coarse
9. Base fragment. Clay orange-pinkish to brownish, phyllite inclusions, white quartzite grits, deeply sealed
large coarse phyllite inclusions, much straw-tempered circles (FIG.13, S 17).
(FIG.16). x8. Body fragment. Clay pinkish with grey core,
io. Body fragment. Clay red-orange, grey core, medium to coarse phyllite inclusions, white quartzite
coarse phyllite inclusions, white quartzite grits, straw- grits, incised hatching (FIG.13, S 18).

CONCLUSIONS
The pottery on the surface of an archaeological site can usually reflect either intensity of
occupation or progressive erosion. In the case of Stavromenos the reconstruction of the
original appearance of the ancient settlement is particularly difficult. The pattern of the
scatter of sherds on the surface resultsfrom several factors of differentcharacter,and much of
the materialwas apparentlymoved from its originalposition.
Nearly all the fragments recorded on the surface are coarse wares. No clear differences
between particularareas were observed. The fabric is in general homogeneous: coarse phyllite
inclusions, white quartzite grits, and straw-tempering are the typical features; very
characteristicis a grey core which appears mostly in large vessels like jars and pithoi. Sherds
are usually very worn, and no painted decoration has been noted. Relief decoration, however,
is very common (PLATE42 b). The tripod legs belong to the diagnostic LM III c-PG types with
multiple finger impressions and vertical grooves; these are very numerous in the LM III c
material from Karphi, Palaikastro Kastri, Arvi Fortetsa, Erganos, Kritsa Kastello, Oreino
Ellinika, and Kastri.39Stamped circles, visible on x7, were noted together with LM III c

39 M. Seiradaki,'Potteryform Karphi',BSA55 (1960),9 andfig.4; Sackettet al. (n.5), 285 and fig.17.
A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE 249

evidence at Loutraki Kandilioro, Myrsini Kastello, Chandras Voilia Kastri, but were also
observed on a PG-G sherd at Maza and at the PG-G-Archaic site of Stavrochori Skalia.40
Incised spirals (13) are particularlynumerous on the top of Kera Papoura in Lasithi, which
seems to have been founded during very late LM III C or even at the beginning of PG, and
flourished during PG and G.41The herringbonemotif appearsin two versions, more elaborate
or simplified;this is characteristicfor the LM III c-PG periods.42
I could not find any meaningful difference between the material from the upper and lower
settlements. Both seem to represent the same period and the same technology of pottery
production. This conclusion supports topographical observations and the reconstruction of
Stavromenosas a typical darkage site consisting of a lower settlement and a refuge area above
it.

KASTELLOPOULO

THE SITE
This site lies about 8oo m N of Stavromenos,below the isolated rock of Kastellopoulo (FIG.4.
7-8; PLATES 39 d, 41 a). The limit of the ancient settlement is drawn by the scatter of sherds
and terraces ascending from the foot of Kastellopoulo to the s and E (FIG.6). The natural
boundaries are formed by a narrow,dry stream bed to the w and a miniature 'kampos'to the
E. The scatter of sherds is very dense on the terraced area, and the pottery is particularly
abundant at the points where the terrace walls are eroded. Considering the natural relief of
the slope, and the height and width of the terraces, one can assume that the ancient houses
are fairly well preserved. Some of the terrace walls may follow exactly the lines of ancient
supporting walls. The lowest of the latter may have served as a defence.43 The area
reconstructedas a settlement itself can be estimated as covering c.5,500 sq m, and the refuge
area on the top of Kastellopoulo as not larger than 300 sq m.
The lower settlement at Kastellopoulo, described above, is easily approachedbut situated in
an area abundant with isolated rocks and inaccessible summits. We can only guess that it was
no accident that this LM III C site was founded right below the outstanding rock of
Kastellopoulo, which is reminiscent of natural citadels typical of many dark age refuge
settlements.44
At first sight Kastellopoulo seems to be totally inaccessible. In fact, one can attempt to
climb it on its E side, at the place where a natural crack forms an opening in a cliff wall. The
oblong, rocky top leaves no suitable surface for any construction; indeed, no architectural
remains were observed here. The ridge of Kastellopoulo is divided by the aforementioned

40 These parallelswill be published in Nowicki (n. 2). Amsterdam, 1992), 53-76). It is very probable that many
41 Ibid.; for similar examples from Kavousi Kastro see more sites from this period were defended by walls, but
Gesell et al. 1985 (n. 13),pls 95-6. because the wall construction was poor only a few can be
42 Seiradaki(n. 39), 3; Sackett et al. (n. 5), 286. traced on the surface.
43 The problem of fortification walls in DA Crete is 44 We can mention here: Zakros gorge Ellinika,
discussed by B. J. Hayden, 'Fortificationsof postpalatial and Palaikastro Plakalonia Kalamafka, Adrianos Fortetsa,
early iron age Crete', AA 1988, 1-21, and by K. Nowicki, Gonies Porolios, Mythi Kastello, Arvi Fortetsa, Frati
'Fortificationsin dark age Krete', in S. Van de Maele and J. Kephala, and Rokka Kisamou, all with very well-defined
M. Fossey (eds), Fortificationesantiquae(McGill University lower settlements and inaccessible refuge areas on the rocky
Monographs in Classical Archaeology and History, 27; knolls above.
250 KRZYSZTOF NOWICKI

FIG.6. Plan of Kastellopoulo.

crack into two parts with a depression between them, about 2 m deep and 2 m wide, opening
to the precipitous western cliff (FIG. 8). Here can be seen numerous sherds, an ash layer, and
animal bones in substantial concentration. Unfortunately, the place is badly destroyed by
hunters of antiquities.45 A trench dug by them probably explored the whole deposit, and the
only remains visible today are on the surface. The sherds are mostly fine wares, and many

15 According to G. Kanavakis this illegal digging took claimed that it took place in the i96os, but the former
place between 1954 and 1957. Another inhabitant of Pefki informant seems to remember the facts better.
A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE 251

KASTELLOPOULO N
S
Shrine

S e t tl e m e n

0 30m

FIG.7. Section (S-N)of Kastellopoulo.

ROCK
I~

2 1

ROCK

ash
potsherds 0 5m
bones

FIG.8. Plan of Kastellopoulo shrine.


252 KRZYSZTOF NOWICKI

have painted decoration. The animal bones seem to represent goats and sheep, and are
usually burnt. Ash is visible all over, but in some places it appears as a layer in an exposed
section. A cobblestone tool with traces of use on both ends was also recorded (below,59). No
remains of walls were visible during my visit, but two naturalwalls are formed by rock.
The character of the site on the top of Kastellopoulo can be only tentativelyreconstructed
on the basis of finds and topography.In sharp contrast to the pottery scattered over the lower
settlement of Kastellopoulo, which belongs to the coarse-ware group, the material from the
knoll represents almost exclusively a fine-ware class. This kind of sherd, together with the
abundance of animal bones and ash, suggests a special function for Kastellopoulo. It was
mentioned as a beacon or a watchtower.46Its topographical position, dominating the lower
settlement and with a splendid view towards the south, might have justified such an
identification,but the finds and new researchat other similar sites suggest another,and I think
more accurate, reconstruction: the place may have been used as an open shrine where
sacrificeswere offered. Very important information concerning the character of this site was
given to me by Mr Georgios Kanavakis, who visited the place with the police immediately
after it was plundered in the 1950s. According to him, two large pithoi had been unearthed.
One of them was filled with 'burnt animal bones and two elephant tusks' (!). The second was
filled with clay animal and human figurines and with 'round stones the size of apricots [i.e.
cobblestones?] of different colours'. What can we now find to check the plausibility of this
story,which at first sight sounds incredible?
The position of one pithos and its body fragments were still visible at the site during my
visits in i991 and 1992 (FIG. 8, circle no. I; PLATE 40 d). Although no figurines or cobblestones
were recorded nearby, this pithos must have contained, at least, smaller pots, fragments of
which are still preserved in situ. Many broken pieces of the pithos were noted in the crack
directly below to the E. On the basis of the preservedfragmentswe can reconstructthe pithos
as being c.o.8 m in diameter (FIG.13. 34). The clay is brown close to the surface and grey to
black at the core. The outer surface was slightly burnished,the inner surface rough. Although
I could not find any fragment of the second pithos, there is no reason to doubt this part of
Kanavakis'sstory.According to him the second pithos was found about I m furtherw (FIG.8,
circle no. 2; PLATE40 e). It was said to have been filled with bones, and indeed bones and ash
are visible in this part of the trench. More pottery can be seen down the steep rocky slope and
cliff to the w.
The material recorded during recent field investigationspoints without doubt to a special
function for the top of Kastellopoulo. A very similar situation was noted on the summits of
other LM III c defensible settlements, for example Arvi Fortetsa,47Gonies To Phlechtron,48
Oreino Kastri,49and Mirthios Kirimianou. We can conclude, therefore, that the summit of

46 Papadakis(n. 31), 73. shrine, proposed by the Britisharchaeologists.


M. S. E Hood, P. Warren, and G. Cadogan, 'Travelsin 48 K. Nowicki, 'To Phlechtron and other dark age sites
Crete', BSA59 (1964),92. On my recent visit to Arvi Fortetsa near Kera Karphi' (forthcoming).
in May 1992 I saw three holes dug during the winter of 9 The cult place was probably situated on the top of
1991-2, probably by illegal excavators.They destroyedmuch Kastri, where the Hilltop Building was reconstructed
of the area, but revealed more fine ware and many fragments (Nowicki (n. 28), 17Oand pl. 37 b). Many burnt animal bones
of burnt animal bones, which could be seen on the surface. were seen by the author along the s edge of the highest
This supports the original identification of the spot as a terrace of Kastri.
A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE 253

Kastellopoulo was used as an open shrine. It was probably located at the most inaccessible
spot, which may also have been regarded as a good refuge area in case of a sudden raid. It
may have been that the rock of Kastellopoulo was used for refuge, and was only one element
of a refuge system covering the inaccessible rocks of Mega Chalavro and the summit of
Stavromenos.
CATALOGUE OF SHERDS FROM KASTELLOPOULO (PLATE 42 a)
Nos. I-39 and the stone tool 59 are from the rocky knoll of Kastellopoulo, 40-57 from the
lower settlement.

I. Shoulder fragment. Clay dark red-brownish, fine, 14. Rim fragment of one-handled (?) cup. Clay
hard;innersurfaceall paintedwith red, outersurfacered pinkish, fine; surface plain, light brownish; LM III c
paint on pinkishslip,octopusmotif;LM III c (FIG.12). (FIG.
12).
2. Shoulder fragment. Clay light orange-buff, fine, v. 15. Rim and body fragment of deep bowl. Clay light
hard; orange-buff slip, thick layer of black-brownish buff-orange, hard, fine; buff slip on surface; LM III B
paint, black overpainted interior, wide horizontal or C (FIG.12).
band with linear decoration below it on outer surface; x6. Rim fragment of deep bowl. Clay orange-buff,
LM III c (FIG.12). fine; inner and outer surface plain; LM III c (FIG.12).
3. Body fragment (near base). Clay light buff-orange; 17. Rim fragment of conical cup. Clay buff, fine;
outer surface buff slip, black brownish paint, inner inner and outer surface plain; LM III c (FIG.12).
surface all painted with black, two vertical strips x8. Rim fragment of closed jug. Clay coarse
preservedon outer surface;LM III c (FIG.12). brownish, coarse phyllite inclusions, white grits, straw-
4. Body fragment. Clay light buff-orange, fine; inner tempered (FIG.12).
surface buff, outer surface buff slip and red paint, four xg. Rim fragment. Clay v. coarse, brownish orange,
narrowwave (?)strips;LM III c (FIG.12). coarse phyllite inclusions, white quartzite grits, much
5. Body fragment (a lower part). Clay light orange- straw temper; surface sl. burnished (FIG.12).
buff, fine, hard; inner and outer surface black-painted 20o. Base fragment of tube stand (?). Clay red brick, v.
all over (FIG.12). hard, coarse and medium phyllite inclusions, white
6. Kylixstem fragment.Claylight buff-orange,fine; quartzite grits; surface brownish red, sl. burnished;
orange-buff slip, red paint, horizontal strip on LM III (FIG.12).
junctionof stemand bodywithverticalband;LM III 2x. Rim fragment. Clay red brick, hard, coarse and
Bor earlyIII c (FIG.12). medium phyllite inclusions, white quartzite grits;
7. Base fragment.Clay light pinkishbuff,fine, hard; surface brownish red, sl. burnished (FIG.12).
inner and outer surface black-painted (FIG.12). 22. Rim fragment of kylix (?). Clay buff, fine, hard;
8. Rim fragment. Clay brown, grey core, fine to LM III B or C (FIG.12).
coarsephylliteinclusions,sand,whitequartzite,straw- 23. Rim fragment of kylix (?). Clay as 22; LM III B or
tempered(FIG.I2). C (FIG.12).
9g.Rim fragment.Clayreddishbrown,coarsephyllite 24. Clay coarse, brown, hard, coarse phyllite
inclusions,white quartzite,calcite, straw-tempered inclusions, straw-tempered; surface sl. burnished,
(FIG.
12). herringbone incised decoration in a more elaborate
zo. Rim and body fragment of deep bowl. Clay light, fashion; LM III B or early III c (FIG.12).
pale brown to greyish, fine, hard; inner and outer 25. Conical cup fragment. Clay orange-red,fine, little
surface orange brownish slip and black-brownish sand and white quartzite,v. fine straw temper; LM III
paint; LM III B or c (FIG.12). c (FIG.
12).
I1. Body fragment. Clay orange, fine; inner surface 26. Rim fragment of narrow-neckedjar. Clay reddish-
red overpainted, outer surface orange-buff slip and brown, fine to medium phyllite inclusions, quartzite,
brownish red vertical strips (FIG.12). calcite, straw-tempered; LM III (FIG.12).
x2. Base fragment. Clay light buff, fine, hard; orange- 27. Rim fragment. Clay greenish buff, hard, fine to
buff slip and red horizontal band around base; LM III coarse phyllite inclusions, sand, straw-tempered (FIG.
c (FIG.
12). 12).
13. Base fragment. Clay pinkish, fine, hard, v. fine 28. Rim fragment. Clay as 26 (FIG.12).
grits and v. fine straw-tempered;pinkish buff slip and 29. Shoulder fragment. Clay light orange-buff, fine,
wide horizontal band around base, brownish paint; hard; light orange slip, black paint; LM III c (FIG.
LM III c (FIG.12). I3).
254 KRZYSZTOF NOWICKI

30. Shoulder fragment. Clay light orange, fine, hard; 46. Lid knob or false neck of stirrupjar. Clay light,
buff slip, red to dark brown paint, inner surface buff, fine, much worn (FIG.14).
overpaintedwith darkbrown, LM III C (FIG.13). 47. Base fragment. Clay v. coarse, brownish red;
3x. Body or neck fragment. Clay light, fine, hard; buff coarse phyllite inclusions, white grits, much straw
slip, brown to black paint, inner surface overpainted temper; basket-likevertical incisions close to base (FIG.
with darkbrown to black (FIG.13). 14)-
32. Shoulder fragment. Clay red-brown with dark 48. Fragment of human statue (?). Clay pinkish
grey core, fine, v. hard; pinkish orange slip, red orange, fine, hard, v. fine phyllite inclusions, little
lustrous paint, inner surface overpainted with black; straw temper. Probably from the head of a human
LM III B (FIG.13). statue; hair and one eye (?) rendered in applied strips
33. Rim and body fragment of deep bowl. Clay light and knob (FIG.14)-
buff, fine, hard;buffslip, blackpaint; LM III c (FIG.13). 49. Shoulder fragment. Clay orange to brownish
34. Rim fragment of pithoid jar. Clay brown with orange, coarse, many coarse phyllite inclusions, v.
grey-black core; outer surface sl. smoothed, inner large quartzite grits; two bands of hatched incisions
rough (FIG.13). (FIG.14).
35. Tripod leg. Clay coarse brownish, surface sl. 50. Rim fragment. Clay pink with grey core, coarse,
smoothed, orange to brownish in colour; LM III B or medium size (to 2 mm) grey phyllite inclusions (FIG.
early III C (FIG.13). 14).
36. Body fragment. Clay orange-red, coarse phyllite 5x. Rim fragment. Clay orange, white quartzite grits,
inclusions, white quartzite grits; finger-impressed fine to medium phyllite inclusions, straw-tempered
bands and vertical incised hatching (FIG. (FIG.14).
I3).
37. Body fragment. Clay light orange, fine, hard; buff 52. Base fragment. Clay orange, hard, fine, straw-
slip, darkbrown paint (FIG.13). tempered (FIG.14).
38. Body fragment. Clay coarse, orange-brown with 53. Base fragment. Clay orange-brown with pale
brown core, phyllite inclusions, white quartzite grits, brown core, fine, straw-tempered(FIG.14).
straw-tempered,surface sl. smoothed (FIG.13). 54. Tripod leg. Clay reddish brown with grey core,
39. Rim fragment. Clay light brownishwith grey core, coarse phyllite inclusions, quartzite, calcite, straw-
fine to medium, and a few coarse, phyllite inclusions, tempered, a deep (io mm) vertical groove (FIG.14).
white quartzitegrits, straw-tempered(FIG.13). 55. Rim fragment of basin. Clay orange-brown
40. Rim fragment. Clay reddish with brown core, surface, grey core, phyllite inclusions, quartzite,
coarse phyllite inclusions, white quartzite grits, straw- calcite, straw-tempered(FIG.14).
tempered (FIG.13). 56. Body fragment. Clay brownish buff, coarse
41. Tripod leg. Clay coarse brown, grey core, coarse phyllite inclusions,straw-tempered(FIG.15)-
phyllite inclusions, white quartzite grits, straw- 57. Upper part of body fragment. Clay orange-brown,
tempered; three incised vertical grooves;(FIG.14). surface orange, v. coarse (to io mm) phyllite
42. Tripod leg. Clay brownish red, coarse phyllite inclusions, straw-tempered; two horizontal bands of
inclusions, white quartzite grits, straw-tempered; a finger impressions(FIG.I'5)
finger impression(FIG.14). 58. Upper part of body fragment. Clay red-brown
43. Tripod leg. Clay like 4x; a finger impression (FIG. with brown core; coarse (5-6 mm) brown phyllite
14). inclusions; outer surface probably washed with
44. Tripod leg. Clay like 42; a finger impression (FIG. pinkish slip, horizontal band with v. shallow finger
14). impressions(FIG.15)-
45. Tripod leg. Clay brownish red, coarse phyllite 59. Stone tool. Serpentine cobblestone with traces of
inclusions, white grits, straw-tempered; a finger intensive use on both ends. From the rocky knoll of
impression(FIG.14). Kastellopoulo. (FIG.15).

CONCLUSIONS
We have two groups of pottery from Kastellopoulo. The first comes from the very small area
on the top of the rocky knoll, and representsmostly the fine ware. It can be dated, therefore,
more easily than is usually the case with material from unexcavated sites, and gives us a rare
opportunityto find links between the coarse and fine ware on a darkage site.
The majority of the sherds from the first group date back to the LM III C period. Deep
bowls, cups, and conical cups all have parallels in the excavated material from Palaikastro
Kastri, Kavousi Vronda, Vrokastro,and Kera Karphi, to mention only sites of the defensible
A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE 255

type.50The Pefkiassemblageis very similarto evidence from Arvi Fortetsapointing to the same
tradition and indicating very strong links with the LM III B period and good ceramic
workshops.5'Several fragmentsmay be dated LM III B. No. 32 is definitelyone of the earliest
examples, but the kylix stem 6 and two probable rim fragmentsof kylikes(22-3) show fabric,
slip, and paint of LM III B rather than III C character.The tripod leg with a slightlysmoothed
orange to brownishsurfaceis similarto the examplesnoted in Milatos and Mythi togetherwith
very early LM III c and late III B pottery,but also in Kalo Chorio Maza where the range of
occupation is much wider.52It seems to representthe LM III B tradition of tripod legs, which
already have circular section but are still without vertical grooves or finger impressions;they
have a much better texture of clay and are better fired than typical LM III C examples. Such
tripod legs can be found on sites which are claimed to have been abandoned in LM III B (e.g.
ChondrosKephali, Mallia, Sissi, and Palaikastro).53 Also fragment24, with incised herringbone
decoration, seems to be early LM III c or late III B. No definite cult objectswere recorded,but
fragment20 may come from a tube a
stand, very common featureof LM III shrines.54
The pottery from the lower settlement representsthe typical LM III C-PG repertoire.The
tripod legs with vertical grooves and impressedfingers are usual among the most characteristic
examples. The finger impressions appear on the definitely LM III C fragments from Kera
Karphi, Arvi Fortetsa,Kritsa Kastello, and PalaikastroKastri, but can also be found among
the PG material from Kera Papoura.55One fragment, 47, is worthy of special attention: it
belongs to the group of pottery noted at Erganos Kephali and Gonies To Phlechtron. In both
cases only a very few pieces were recorded.56The fabric is very coarse but relativelyhard, the
colour orange to red, and the decoration consists of incised or impressed grooves around the
base of the pot. The incisions, however, are of an entirely different character from those
common for dark age pottery, and were formerly described as 'basketlike'.57Because there is
no published analogy from the excavated sites we can be sure of neither the dating nor the
origin of this class of pottery.The three possibilitiesseem to be (I) that the fragmentsdate back
to a much earlierperiod, probablyEM, (2) that they representa very local production, limited
to their place of origin in mountainous areas of Crete, and (3) that they represent a foreign
(non-Cretan)product, dating, however,to the Dark Age.

MEGA CHALAVRO
THE SITE
The third site is situated about 400 m w of Stavromenos, on the E edge of the curious
labyrinthof a huge heap of rocks (FIG.4. 4; PLATE39 b). The physical appearance of this place
has led to fantastic theories among the local villagers concerning the ancient history of the

50 Sackett et al. (n. 5), 282-3; Seiradaki (n. 39), 21-2; L. P. (SIMA 67; Goteborg, 1985), 50. The nearest contemporary
Day, W. D. E. Coulson, and G. C. Gesell, 'Kavousi, shrine was found at Kavousi Vronda, where fragments of
1983-1984: the settlement at Vronda', Hesp.55 (1986), 363. snake tubes and goddesses with upraised hands were
51 K. Nowicki, 'ArviFortetsaand LoutrakiKandilioro:two unearthed in great number: Gesell et al. 1991(n. 13), 16i f.
refuge settlementsin Crete' (forthcoming). 55 Seiradaki (n. 39), 7-8; Hood et al. (n. 47), 93; Sackett et
52Nowicki(n. 2). al. (n. 5), 285 and fig. 17.
53 I wishto expressmy warm thankstoJ Driessen,A. Farnoux, 56 B. Rutkowskiand K. Nowicki, 'Report on Investigations
S. Muller,andJ. A. MacGillivrayfor showingme potteryfrom in Greece, VI: studies in 1988-89', Archeologia (Warsaw),41
recentsurveysand excavationsin Malliaand Palaikastro. (1990), 120.
54 G. Gesell, Town,Palaceand House Cult in Minoan Creta 57Ibid.
256 KRZYSZTOF NOWICKI

area above Pefki. People believe that the extensive area between three points-Stavromenos,
Mega Chalavro, and Kastellopoulo-was once a level surface and was destroyed in historic
times by an earthquake. According to this story the whole area was occupied by a great
politeia, the rest of which is represented today by the three sites mentioned. As is common, the
inhabitants of Pefki want to believe that huge treasures are still buried under the fallen stones
and rocks; Mega Chalavro is the site pointed out as the most promising in this respect. The
truth, however, apparently differs from this tale. Mega Chalavro is the result of a geological
process which may have had a catastrophic nature, but certainly long before the site was first
settled. A large, deep depression, filled with megalithic rubble and cutting the rocky massif
into several parts, was formed when the ground below sank down into a huge hollow or cleft.
An interesting twist in the local story is that people have accidentally or intentionally explored
the rocky labyrinth and have probably found archaeological evidence which seemed to them
to support their version of the great catastrophe.
The archaeological site was identified as a dark age settlement in 1990, located on the slope
below the characteristic rock rising about 30 m above it (FIGS. 9-II; PLATE39 c).58The area
where the sherds are visible on the surface is bounded by the rocky massif to the w and 5-1o m
high cliffs to the s and E; only the N side is open and not defended by nature. Thus the site
measures c.150 m N-S and c.50 m E-W. Part of this area is overgrown with bushes, but in
general the visibility is good. Architectural remains, however, cannot be easily traced. The
state of preservation is much worse, and the soil deposit thinner, than in the case of the lower
site at Kastellopoulo. Only in some places are scant traces of walls visible among stones and
bushes. One better-preserved section of walling can be seen in the N part of the site. Its
character and position suggests that it may have had a defensive function and may have closed
the only access to the upper refuge area (FIG.9, wall F). This reconstruction, however, is only
tentative.
The part described above is the lower settlement. Far more defensible was the upper, which
can be reconstructed as a refuge area rather than a permanently inhabited settlement. There
was only one very difficult connection between these two parts. It starts from the Nw corner of
the lower settlement, close to the wall described above, and passes beside and above a deep
crevice to a natural gate (FIG. 9, G; PLATE 40 a) formed by two fallen rocks, which one must
crawl through; immediately behind this gate the path ascends a steep, bare rock. At the end of
it one must climb a high, rocky step, reaching the first isolated rocky island (FIG.9, RA; PLATE
40 b), about 20 m higher than the lower settlement. The top of this rock, which is c.io X 12 m,
was occupied by a single building; remains of its walls can still be seen on the surface. Sherds
are very numerous, and some come from pithoi or big jars. The area covered by pottery
extends s as far as the rock RB (FIG.9). Another isolated rock (FIG. 9, RC, RD) has a still more
defensible character, situated directly w of RA and c.3 m higher. It is cut by a deep crevice, only
I m wide, into two parts. All its sides are defended by a precipice and cliffs; the only access
climbs to the area RC via a rocky step up from area RA. This rocky 'refuge island', 50 m by
10-15 m in size, shows no architectural traces, but sherds are very numerous, particularly on
top of RC. It seems this area was not built upon, but used for special purposes and under

58K. Nowicki, 'Report on investigationsin Greece, VII: studies in 1990',Archeologia


(Warsaw),42 (I991),142.
A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE 257

FIG.9. Plan of Mega Chalavro.


258 KRZYSZTOF NOWICKI

E W

Refuge Area
Rc
RA

Set t il m e

0 30 m

FIG. io. Section (E-W) of Mega Chalavro.

i/ !~j e/

0e~
rr,
e

FIG.iI. Perspectiveview of Mega Chalavrofrom NE.


A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE 259

special circumstances. If any buildings existed here, they may have been built of perishable
material.
Sherds are visible on small terraces in the higher part of rock RE, as well as below it along
its base to the w (FIG.9, RE, RF). Both areas were perhaps used as refuge areas as well.

CATALOGUE OF SHERDS FROM MEGA CHALAVRO (PLATE 42 c-d)


i. Tripod leg. Clay pinkish brown, coarse phyllite 13. Rim fragment. Clay as 2 (FIG. i8).
inclusions (to 8 mm), straw-tempered; deep vertical 14. Rim fragment. Clay red-brownish with grey core,
groove (FIG.18). phyllite inclusions, straw-tempered (FIG.18).
2. Tripod leg. Clay pinkish brown, coarse phyllite 15. Rim fragment. Clay red-brown with grey core,
inclusions, white quartzite grits, sand grits, straw- coarse phyllite inclusions (FIG.18).
tempered, v. shallow vertical groove (FIG.18). 16. Rim fragment. Clay brown, coarse phyllite
3. Tripod leg. Clay as 2 (FIG.18). inclusions, straw-tempered (FIG.18).
4. Handle or tripod leg (?). Clay brown, coarse 17. Body fragment.Clay v. fine, hard,v. fine sand grits,
phyllite inclusions, straw-tempered(FIG.18). straw-tempered; buff slip and blackpaint, two horizontal
5. Tripod leg or leg of plate. Clay reddish brown, bandsand unidentifiedmotif abovethem (FIG.18).
coarse brown phyllite inclusions, white quartzite grits, 18. Base fragment. Clay brownish, phyllite inclusions,
much straw temper; sl. burnished surface (FIG.18). white quartzite grits, straw-tempered(FIG.18).
6. Base fragment. Clay reddish brown, phyllite xg. Base fragment. Clay brown fine, hard, phyllite
inclusions, white quartzite,straw-tempered(FIG.i8). inclusions, straw-tempered(FIG.18).
7. Rim fragment. Clay orange, phyllite inclusions, 20. Base fragment. Clay reddish brown, phyllite
white quartzite grits; band of impressed fingers below inclusions,white quartzite,straw-tempered(FIG.i8).
rim (FIG.18). 21. Base fragment. Clay light, buff, somewhat
8. Body fragment. Clay pinkish brown, coarse phyllite greenish, fine, hard (FIG.18).
inclusions, white quartzite grits, straw tempered; two 22. Body fragment. Clay brownish, coarse phyllite
bands of impressed fingers made on v. wet and soft inclusions, white quartzite grits, straw-tempered(FIG.
clay (FIG.i8). 17, m 22).
9. Body fragment. Clay red with grey core, phyllite 23. Body fragment. Clay brownish, coarse phyllite
inclusions, much straw temper; two bands of finger inclusions, white quartzite grits, straw-tempered;
impressions (FIG.18). horizontal band of short hatchings (FIG.17, m 23).
io. Rim fragment of closed vessel. Clay as 2 (FIG.18). 24. Body fragment. Clay as 23; two horizontal bands
ix . Rim fragment. Clay as 2 (FIG.18). of impressed fingers (FIG.17, m 24).
12. Rim fragment. Clay brownish, v. fine and hard; 25. Rim fragment. Clay reddish brown, coarse
black paint preserved below collar on outer surface. phyllite inclusions, white quartzite, straw-tempered
Geometric (FIG.18). (FIG. 15, m 25).

CONCLUSIONS
The pottery from Mega Chalavro is not very diagnostic and differs somewhat from most of
the pottery from Stavromenos and Kastellopoulo. No typical herringbone motif was found on
the sherds-a notable observation at a dark age site. Relief decoration seems to have been
limited to finger-impressed bands (very common) and bands with vertical hatchings (relatively
rare). Tripod legs present a variety of shapes: 2-3 have a circular section, a characteristic
feature of the LM III B-PG periods; 1-2 have vertical grooves. The groove in 2 is very shallow
and is executed below the junction of the leg with the body of the pot. I know only one similar
example, from Milatos.59 No. I has a flat section which is very rare, but not extraordinary, in
dark age types. Although at first sight the drawing suggests the EM III-MM I type, clay and
fabric indicate rather LM III production. Flat tripod legs with vertical grooves were recorded

59Nowicki (n. 2).


260 KRZYSZTOF NOWICKI

2
4 6
8

1
3 5
7 9
7

11

13
12
10
K 25

Si28
17
18 1919
16
14
15 5
27
26
21
2W 23 24
20

FIG.12. Potsherdsfrom Kastellopoulo.

29
30 32

33

34
35 38

36

37

39

40o
S16 S18
S17

FIG.13. Potsherdsfrom Kastellopoulo and Stavromenos(s x6-x8).


A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE 261

42

44
445
5

41 46

43

f49
47
48

52
51

50 53
54

FIG. 14. Potsherds from Kastellopoulo.

Cl
58

56

C3

59
C4
C2
M25

CS
57

FIG.15. Finds from Kastellopoulo (56-9), tholos tombs (c x-5), and Mega Chalavro (M25).
262 KRZYSZTOF NOWICKI

4
9

6 7

1
5-
3

11

2 10

12

FIG.16. Potsherdsfrom Stavromenos.

S14

M22

S13 S15 M23 M24

FIG. 17. Potsherds from Stavromenos (s 13-15) and Mega Chalavro (M 22-4).
A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE 263

1
L1
10 11

7
4
12
13

2
14 15
16

17 18
5

19 20

21
6
3 9

FIG.18. Potsherdsfrom Mega Chalavro.

in Kritsa Kastello, Gonies Porolios, Arvi Fortetsa, Koutsounari Karphi, and Kera Karphi.60
No. 3 is a typical LM III example. The bases are usually not very diagnostic, but most Mega
Chalavro examples belong to the shapes which appear in Erganos, Arvi, and Kera Karphi.
Very few rims were observed on the surface of the site. Nos. io-ix and 13 can be generally
classified as LM III-PG; I2 is geometric, painted with black. Rim 7 is rather unusual in the
dark age repertoire, but can be compared to examples from Arvi Kamini and Chondros
Kephali, both probably LM III B. I can find analogies to the strange rim 25 (from the refuge
area RD) only in the Erganos material (where the shape is similar but not identical) and in
Ridopoulia I (R9).61 The only hatching motif recorded at Mega Chalavro was of a simple
type, represented by 23; the same decoration was observed on fragments from Arvi Fortetsa
and Erganos.62
To sum up the problem of the Mega Chalavro pottery: no certain date can yet be proposed.

60 Ibid. 62 Nowicki
(n. 2).
"' Gesell et al.
1983 (n. 13), 419, fig. 13-. a.
264 KRZYSZTOF NOWICKI

The site was inhabited or visited during the Dark Age, but the date of its foundation and of
the first, and probablymost important,phase of occupation is problematic.63

CEMETERIES

Our present knowledge of the turn of the Bronze and Iron Ages in Crete indicates that stone-
built tholos tombs were the most common, or even exclusive type during that period, at least
in these areas where the Minoans lived in defensible settlements.The graves generally formed
extensive cemeteries, within which the individual constructions were irregularly scattered,
sometimes very close to each other but sometimes far apart. No single dark age cemetery has
been explored in its entirety, and the distribution of tombs, as well as the topographic
relationshipbetween the settlement and its cemetery, must be based on a few examples where
the links between them are well attested.64
The Dark Age settlements identified in the neighbourhood of Pefki must have left a
considerable number of tholos tombs. No find of this type, however,has yet been reported. Of
the three sites described above, Kastellopoulo reveals the best-preservednuclear settlement,
with many topographical elements typical of the period. One can expect, therefore, a tholos-
tomb cemetery or cemeteries somewhere in its
vicinity.
TOMB 2
TOMB 1 One of the most promising spots is a gentle slope
descending from the rocky cliff w of Kastellopoulo
and across a small, dry stream-bed (FIG.4. 15). The
2
area is artificially terraced or levelled in places. A
.

scatter of dark age pottery is visible all over the


slope, sometimes in a dense concentration.Although
the possibility of a suburban complex cannot be
-a ruled out, it seems that the archaeological evidence
here testifies first and foremost to the ancient
cemetery, part of which at least may have been
TOMB 1
qbF
SECTION linked with the settlement of Kastellopoulo.
Two stone tholos tombs were, indeed, identified
on a small lakkosabout 300 m sw of Kastellopoulo
0
40 c, 41 b). Considering the state of
(FIG.4. 9; PLATES
FIG.19. Tholos tombs I and 2. their preservation and the lack of any material

63 It was suggested by Donald Haggis that most of the fortified citadels-not only in Crete, but also in the Cyclades
Mega Chalavro pottery may representMM I-II and LM III and Asia Minor. This probably appears in Egyptian and
B, and I think such a dating reasonable. There are relatively near eastern written sources as the stormy period of the first
few links with LM III c evidence, a fact which may even invasion of the Sea Peoples during Merneptah's reign and
suggest earlier or middle LM III B rather than the end. Such the activities of post-Ahhiyawan troublemakers in the E.
an early dating, if supported by other evidence and further Aegean, but still before the main attack of the Sea Peoples
research, may shed more light on the sequence of events in on Cyprus, Asia Minor, the Near East, and Egypt.
LM III B Crete, particularly the late phase between the 64 This remark concerns Erganos (Halbherr (n. 6),
destruction horizon(s) in III BI/2 and the drastic change in 262-81), Kera Karphi (Pendleburyand Money-Coutts (n. 4),
the settlement pattern at the end of III B and the beginning 100-12), and Kavousi Vronda (Boyd (n. 3), I31-6). The
of III c. This was probably a period of wars and migrations, situation in Vrokastroand Kavousi Kastro is not so clear for
when some coastal areas were abandoned, or at least LM III c, because intensive occupation during PG and G
seriously depopulated, and attempts were made to build destroyedmuch of the earlier layout of the cemeteries.
A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE 265

around them, the tombs must have been robbed a long time ago. They were probably built as
a pair 3 m apart, close to the projecting rock (FIG. 19).65 Tomb I is in a better state of
preservation,and affords a view of its method of construction which can thus be studied. The
chamber was dug in rocky ground, and its walls were of slabs of local schist and limestone.
The external plan is square in shape, with sides 3.2 m long. The interior shape of the chamber
is only partiallyvisible, as it is filled with earth. We can suppose, however,that its ground plan
was also square and measured about 1.8 X 1.8 m; the approximate height of the chamber
must have been 1.5 m (at present c.I m). The robbers reached the tomb in the usual way, by
removing its top slabs. An original, and still blocked, entrance is visible on the SEside. Tomb 2
is far less exposed on the surface (PLATE41 b). Either it was completely demolished and then
filled up with stones and earth, or the robbers ignored it for unknown reasons. It may also
have been looted long before tomb I was found. No sherds are visible in the close vicinity of
these tombs, but some 30 m SEwere a few fragments, one of which is a tripod leg with the
rather uncommon decoration of two grooves incised in a shape of an inverted letter V (FIG. 15,
C 1). The fabric points to the LM III C-Protogeometricperiod.
The third tholos tomb was recorded c.250 m further sw at Glikis Prinos (FIG.4. io). This
identification, however, is based on far less well-preservedevidence. A few stones are all that
remain of the tomb construction (FIG.20 b). One outer corner and the inner circle, or rather
the remains of it, indicate that this tomb was bigger than the other two. Its external diameter
may have been as much as 4 m or even 4.5 m. A hollow in the centre, a heap of earth and
stones around it, a large amount of sherds, and a few very small fragments of bones would
seem to indicate that the tomb was robbed quite recently.66The coarse-warefragmentsshow a
typical grey core and brownish-redsurface;one fine-ware fragment derives from a deep bowl,
while another two (FIG.15, C 3) of deep bowls or cups had their insides overpaintedwith black.
An interesting object is a small piece of
iron which may be a fragment of a dagger
or sword handle with a rivet (FIG.15, C 5)- TOMB 5
An extensive area around tomb 3, TOMB 3
particularly to the E, is covered with
sherds. One fragment, part of a deep bowl,
was noted about 40 m E of tomb 3 (FIG. 15,
C 2). Close by, a piece of molten iron has TOMB 4

been seen. The last find may suggest that


this outskirt of the valley was designated
not only as a cemetery area, but also as a
peripheralindustrialcentre.
Tomb 3 is located between 0o 3m

Kastellopoulo and Mega Chalavro, but


closer to the latter site. At present it is FIG.20. Tholos tombs 3-5.

65 Tholos tombs located in pairs are a very characteristic 66 1 was unable to obtain information as to when this tomb
feature of both the Karphi cemeteries (Ta Mnimata; was robbed, but the inhabitants of Pefki clearly stated that
Astividero); see Ta Mnimata, tombs 1-2, 5-6, 7-8, 17-16; this digging took place much later than the illegal
Astividero, tombs 2-3 (Pendleburyand Money-Coutts (n. 4), excavations on Kastellopoulo and in tombs 1-2.
Ioo-8, pls 12-13).
266 KRZYSZTOF NOWICKI

hard to say which tombs belong to which settlement. It seems that the whole area of Glikis
Prinos and around was a traditional burial ground used by the inhabitants of the whole
district. Other graves or cemeteries may have been located closer to the particularsettlements.
This was the case with two constructionswhich were recorded very close to, and NE of, the
Mega Chalavro site (FIG.4. 6). A long, gentle hillock between the Mega Chalavro slopes and
the stream bed E of it is covered with sherds, which testify that the area lay within the
settlement's activity zone but beyond the limit of its constructions.Two badly destroyed stone
constructionsare located on the N end of the aforementionedhillock (FIG.20 a). The character
of the stone slabs and the shape of these constructions justify their identification as poor
remnants of small tholos tombs (PLATE 41 c). About 15 m further Nw one fragment of a
geometric pot has been seen. Stone tumbles to the N and s may indicate the position of other
tombs. The cemetery of Mega Chalavro may therefore have been located very close to the
settlement and slightly below it, a pattern which is very well representedby the cemeteries of
Karphi and Vronda.67
All the constructionsidentified above as tholos tombs are apparentlyonly a very small part
of the dark age cemeteries linked with the settlements of Kastellopoulo, Stavromenos, and
Mega Chalavro. According to Georgios Kanavakisfrom Pefki, larnax burialswere unearthed
in the small lakkosE of Kastellopoulo (FIG.4. II), and other graves have been found in rock
shelters still furtherE (FIG.4. 12).If this secondhand information is true, these larnax and cave
burials may point to even older (LM III B?)roots for the settlement of Kastellopoulo. More
intensive research, surveys,and excavationsshould reveal new objects, with (let us hope) their
original contents.

SUMMARY

What is the context of the group of Dark Age sites near Pefki?While they do not belong to the
greatest and most important settlementsof that age, the Pefki sites may point to the character
of the changes that governed the general settlement pattern in Crete. Perhapsvery local, but
topographically significant movements in the area, as observed at Pefki, reflect the same
process marked elsewhere by the abandonment of sites such as Kera Karphi, Arvi Fortetsa,
Erganos, and Mirthios Kirimianou and the foundation or very dramatic development of
towns like Lato, Anavlochos, Panagia ProphitisIlias, Kerato, Kato Chorio ProphitisIlias, and
many others.
The topographic studies at Pefki, as presented above, were unable to provide answers to
several essential questions: first and foremost, the chronological relationship between all the
sites. The general field research,however,sheds more light on the Dark Age in East Crete and
allows us to present some preliminary conclusions before intensive surveys and excavations
add new data to the problem.
The material recorded on the surface at Kastellopoulo points to the LM III C or even the
late LM III B period as the beginning of occupation. The best pieces noted on the top of the
rocky knoll may date back to the very early phase of LM III C, and show links with very late
LM III B. Such an early date for the process of population shift to the mountains, following

67 For the topographical relation between the cemeteries (n. 13), fig. I; K. Nowicki, 'The history and setting of the
and settlements in Vronda and Karphi, see Gesell et al. 1988 town at Karphi', SMEA26 (1987),fig.
i.
A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE 267

the collapse of the Minoan-Mycenaean civilization in Crete, is possibly supported by new


evidence from Kavousi Kastro68and fits well with the resultsof the excavationsundertakenat
PalaikastroKastri.69The refuge settlements of Arvi Fortetsa,70Vainia Stavromenos,Krya Ag.
Georgios, Kavousi Kastro,71 and Praisos, to mention only those close to Pefki, were all
founded at the beginning of LM III c if not at the end of LM III B. The settlement at
Kastellopoulo, however,is not very defensible and its geographical setting can be compared to
that of Kavousi Vronda,72 Monastiraki Chalasmeno,73 and Oreino Ellinika.74 All the
aforementioned sites were accompanied by real defensible settlements: Kavousi Kastro,
MonastirakiKatalimata, and Oreino Kastri respectively.The topography of early refuge sites
suggests that the beginning of the process of change in the settlement pattern of Crete was a
particularlyhard time, when people's security was seriously threatened.75Since pottery often
overlapsbetween the more and less defensible sites located close to each other, is hard to say
what their relationshipwas in the functional and chronological sense. The theory of seasonal
settlement must be abandoned, for the sites of Oreino Ellinika and Kastri, Monastiraki
Chalasmeno and Katalimata, and Pefki Kastellopoulo, Stavromenos,and Mega Chalavro are
all located within the same climatic and agriculturalzone.
The answer to the problem of the mutual relationship between more and less defensible
sites in dark age Crete may be clarifiedby the results of the Kavousi project. We may propose
the following reconstruction.First, at the beginning of the Dark Age, at the turn of the LM III
B and III C periods, the very defensible sites like Kavousi Kastro (and subsequently
Monastiraki Katalimata and Oreino Kastri) were founded. Soon afterwards, in early or
middle LM III c, perhaps even within the same generation or the following one, people tried
to go down and built lower-lyingsettlements. Kavousi Vronda, MonastirakiChalasmeno, and
Oreino Ellinika may have been settled at this time. They coexisted, however, with the true
refuge settlementsin the vicinity. This security system probably lasted until some point in the
Subminoan or the beginning of the Protogeometric period, when new troubles forced the
inhabitantsof many LM III C settlementsto abandon their homes and either to move to more
defensiblesites or to look for new locations.
In the case of Pefki we are dealing with three sites located within a circle c.8oo m in
diameter. All three sites consist of two topographically different parts. Kastellopoulo has a
relatively open lower settlement (if we rule out the possibility of a defence wall here) and the
inaccessible, but very small, top of the rocky knoll above; and Mega Chalavro has a lower
settlement which is, however, well protected on three sides, and a particularly inaccessible
refuge area on the rocks above it. The third site was located on the summit and steep slopes of
Stavromenos. The part situated on the SE slope was less defensible, but closer to the arable
land below. The refuge area, and perhaps the upper settlement, occupied the summit and
highest terraces of the mountain. There is still another area which may have been used for

68 W. D. E. Coulson,'Recentexcavationson the Kastro 71See n. 68.


Kavousi,East Crete',paperpresentedto 7th International 72Boyd (n. 3), 131;Day et al. (n. 50), 355-9.
CretologicalCongress,25-31 Aug.i99i, Rethymnon;M. S. 73 D. C. Haggis and K. Nowicki, 'Khalasmeno and
MookandW.D. E. Coulson,'TheLateMinoanIII c pottery Katalimata: two early iron age settlements in Monastiraki,
fromtheKastroKavousi,EastCrete',AJA97(1993),351. East Crete', Hesp.(in press).
69Sackettetal.(n.5), 269 ff. 74Nowicki (n. 28), 172 and pl. 33; id. (n. 36), pl. 39-
70HoodWarrenandCadogan(n.47),93. 75Nowicki (n. 36), 233-4.
268 KRZYSZTOF NOWICKI

protection: the curious rocky labyrinth behind the site of Mega Chalavro (the proper Mega
Chalavro location). The differences in location between the particular sites may reflect the
changing security situation of Minoan refugees during the Dark Age. The surface material
indicates that Kastellopoulo was the most important settlement during LM III c. Its
inhabitants may have used the rock of Kastellopoulo itself as a refuge area or, at a slightly
greater distance, the summit of Stavromenos and the inaccessible rocks of Mega Chalavro.
Stavromenoswas apparentlyinhabited for most of LM III C and until the Protogeometricand
Geometric periods. By that time Kastellopoulo was probably abandoned.
We come to the most difficultpoint when we try to find out the place of Mega Chalavro in
the discussedsystem. A geometric sherd noted close to tombs 4 and 5 suggeststhat the site was
occupied or visited during that period, but in general the material is not very typical either of
LM III c or of the Geometric. The fabric is local and differs from that observed at
Kastellopoulo. Some sherds may even represent a much earlier period (EM-MM), but they
are not very numerous. The Mega Chalavro setting proves that its inhabitants were terrified
or even obsessed about the threat of sudden raids. It would be very interesting to know in
which phase of the Dark Age (or even earlier)people lived under such conditions.
The dark age sites near Pefki fill the gap in the settlement pattern between the very strong
group of LM III c-Protogeometric refugees in the west Siteia mountains and the Praisosarea.
They show the same process of moving away from the shore which is represented by the
Oreino sites, Vainia Stavromenos,and Kato Chorio ProphitisIlias along the south coast, and
by Tourloti Kastri and Asari on the north coast. The abandonment of coastal settlementstook
place under a very serious threat from the sea, but people did not always flee to the interior of
the island. They stayed close to their former locations wherever possible, finding a defensible
mountain towering over the coastal plain. Examples are Vrokastroand Myrsini Kastello on
the north coast, and Arvi Fortetsa,Kerato, and KoutsounariKarphi on the south coast. Some
coastal plains, however, lacked an appropriatelocation for a refuge settlement, and in such
cases people had to move to the nearest defensible mountain in the interior. The Pefki sites
represent the second pattern, and their closest analogy can be found in the Oreino valley.The
settlements situated about two hours' walk from the sea, and controlling all access from this
direction, gave their inhabitants the chance to prepare for fighting by escaping onto the
Romanati massif or seeking safety in the refuge areas above.

PolishAcademyof Sciences,Warsaw KRZYSZTOF NOWICKI


,I

settlement,
=
2

area,
2

refuge
=
(I
iiiii
1iiii~ N

from

Chalavro

Meg4
CRETE
(c)Stavromenos).
N. =
2
EAST
from
(2)
1 2:
PEFKI,
Kastellopoulo,
Chalavro
NEAR=
NOWICKI
(I
(b) (d) N
Mega
and from
CENTRE
(i)
KRZYSZTOF
2

REFUGE
Kastellopoulo
Stavromenos
AGE (d)
(b)

DARK
1 A cemetery?).
Diaskari).
= =
2 3

A
Stavromenos,
=
(I
NW
1
m from

Stavromenos
(a) (c)l (a)
PLATE40

(a)
(b)1

(d)

(c) I (e)
KRZYSZTOF NOWICKI
A DARK AGE REFUGE CENTRE NEAR PEFKI, EAST CRETE
(a)Natural gate to the refuge area of Mega Chalavro.(b)Refuge area of Mega Chalavro (RA). (c)Tholos tomb (i) and
Kastellopoulo (2) from sw. (d)Kastellopoulo shrine (i = position of pithos filled with pots and figurines?).
(e)Kastellopoulo shrine (2 = position of pithos filled with bones?).
It

tholos
a
of

Remains
(d)

Chalavro.
PW
Mega

I
below
(?)

tombs
CRETE

Tholos
EAST
(c)

PEFKI,
Chalavro.
(b) (d)/ NEAR
Kastellopoulo.
NOWICKI
of Mega
sw
2)
CENTREbelow
and(?)
(I
KRZYSZTOF
tomb
REFUGE
tombs

AGE
Tholos
(b)
DARK
A

1 u settlement).
=
2

shrine,
=
(I
SE

from

Kastellopoulo
(a)
(a) (c)
It

Chalavro.
CRETE
Mega
(d)
EAST
(c),

PEFKI,
(d)

(b) NEAR
NOWICKI
Stavromenos.
(b)

CENTRE
(shrine).
KRZYSZTOF

REFUGE

AGE
Kastellopoulo
(a)
DARK
A
Potsherds.

(a) (c)

S-ar putea să vă placă și