² The settlement history of the Vrokastro area has been published in B. J. Hayden et al., Vrokastro Reports 2. Primary publications include Hayden, B. J., Moody, J., and Rackham, O., ‘The Vrokastro survey project, 1986–1989: research design and preliminary results’, Hesp. 61 (1992), 293–353CrossRefGoogle Scholar; Hayden, B. J., ‘Rural settlement of the Orientalizing through early Classical period: The Meseleroi valley, eastern Crete’, Aegean Archaeology, 2 (1995), 93–144Google Scholar; ead., ‘Final Neolithic-Early Minoan I/IIA settlement in the Vrokastro area, Mirabello, eastern Crete’, AJA 107 (2003), 363–412CrossRefGoogle Scholar.

³ See Vrokastro Reports 2, ch. 15, ‘Future Exploration of the Area’.

⁴ Exploration of key sites could provide important data concerning the chronology and nature of the earliest occupation of Mirabello; the depth and nature of Roman occupation; and the relationships, territorial and otherwise, between the small regional powers and the larger city-states of Lato and Hierapytna. Other analyses might combine fieldwork with a new analysis of the Vrokastro survey pottery and exploration through survey of the territories of neighbouring city-states. Some of these issues are best approached through excavation, others through survey.

⁵ The geophysical prospection programme is being conducted by A. Sarris. E. Kokkinou was involved in the processing of the 2002 geophysical measurements. E. Aedona carried out the magnetic susceptibility measurements. K. Kouriati and S. Soetens were involved in GPS measurements. Exploration of the geological environment and resources is being directed by I. Bassiakos, assisted by K. Pavlopoulos, K. Athanassas, and N. Zacharias (soil scientist, Demokritos Laboratory). Involved in the analyses of metal, ceramics, and stone are I. Bassiakos; H. Dierckx; B. Hayden; M. Kostoglou; and E. Nodarou.

⁶ IM15 belongs to the ist to 7th cc. AD. Walls, pottery, and plaster are eroding out of the beach and large walls and other features are partially submerged. This site contains large and small cisterns floored with thick hydraulic cement above lime plaster mixed with beach pebbles and crushed granodiorite. Bronze fragments and bits of faience have also been recovered. Fresh water sources for these cisterns may have been the Istron River or a seasonal torrent bed that encircles the base of the Ioannimiti promontory. Istron is a toponym mentioned in the late Roman-Byzantine guide to harbours and water, the Stadiasmos; Pendlebury, J. D. S., The Archaeology of Crete (London, 1939), 30Google Scholar, and this may have been the anchorage referred to, with large stores of fresh water for coastal ship traffic.

⁷ The earliest harbour sites are discussed in Hayden 2003 (n. 2), 383–4. Elias to Nisi encloses a small cove that was utilized as a protected harbour during the Early Iron Age, as the large contemporary defensive wall encircling the site indicates. Further east, an early sheltered harbour in deep water may be located between the tip of Vrionisi and a small islet, especially if a narrow land-bridge once existed linking the two.

⁸ Vrokastro Reports 2, 168–9, 347–9; fig. 39.

⁹ Vertical subsidence of 2 m has been estimated for the nearby Malia region; see Raban, A., ‘Minoan and Canaanite harbours’, in Laffineur-Basch 1991, 139Google Scholar. Schafer provides a similar estimate for Amnisos, J. Schäfer, ‘Amnisos—harbour town of Minos?’, ibid., 114. Shaw, , ‘Bronze Age Aegean harboursides’, in Hardy, D. A. (ed.), Thera and the Aegean World III (London, 1990), i. 425–7Google Scholar, suggests a sea level change of 2.5 m at Kommos, and 3 m at Malia.

¹⁰ A similar reconstruction has been posited for Amnisos, where a deeper estuary may have formed a protected harbour behind a headland; see Schäfer (n. 9), 111–15; pl. 28 b. Shaw (n. 9), 427, does not believe that river estuaries in Crete were utilized as harbours. Promontories can create and protect natural harbours, however, as at Malia. See M. Hue and O. Pelon, ‘Malia et la mer’, in Laffineur-Basch 1991, 117–27; pls. xxxi, xxxii. For another view concerning harbours at Malia, see Raban (n. 9), 129–45, who states (p. 133) that by the Roman period, most harbours associated with river estuaries in the Mediterranean had been silted in.

¹¹ A preliminary study of the Istron River between the coastal highway and the estuary indicated that the channel may have been in place since the Venetian period; see G. Postma, ‘The Holocene evolution of the Istron area, Mirabello’, in Vrokastro Reports 2, Appendix 6. Measurements taken by Bassiakos and Athanassas suggest that two tectonic blocks oriented north-south along the river may be moving apart. It was therefore thought probable that the river and its estuary had shifted location over time, and that the course of the river might have been closer to Priniatikos Pyrgos, or even west of this small headland (see also ‘Establishing Limits of the Site’, below). This hypothesis has been supported by the results of recent subsurface probes undertaken in 2004, directly south of the estuary of the river, in an area called Kambos (FIG. 1). Fieldwork will continue in 2005 in an attempt to locate the earlier river channel.

¹² Hall 1914, 84–5; ead. ‘Excavations at Vrokastro, Crete, in 1912’, Art and Archaeology, 1 (1915), 36Google Scholar.

¹³ Pendlebury 1932–3; Schachermeyer, F., ‘Vorbericht über eine Expedition nach Ost-Kreta’, AA 53 (1938), 469Google Scholar; and Sanders 1982, 142. More recent publications of Hall's pottery from the site include Betancourt, P. P., ‘LM IA pottery from Priniatikos Pyrgos’, in Doumas, C. (ed.), Thera and the Aegean World I (London, 1978), ii. 381–7Google Scholar; id., Minoan Objects Excavated from Vasilike, Pseira, Sphoungaras, Priniatikos Pyrgos, and Other Sites: The Cretan Collection in the University Museum, University of Pennsylvania (University Museum Monograph, 51; Philadelphia, 1983), i. 15–21; id., The History of Minoan Pottery (Princeton, 1985), 130–1; Niemeier, W.-D., ‘The master of the Gournia octopus stirrup jar and a Late Minoan LA pottery workshop at Gournia exporting to Thera’, TUAS 4 (1979), 21–2Google Scholar; fig. 7; Driessen, J. and MacDonald, C., The Troubled Island. Minoan Crete Before and After the Santorini Eruption (Aegaeum, 17; Liège, 1997), 211Google Scholar; Popham, M. R. and Coldstream, J. N., ‘Individual sites: session 9B: discussion’, in Doumas, C. (ed.), Thera and the Aegean World II (London, 1980), ii. 389–90Google Scholar. Recent publications include Hayden, B. J., ‘The coastal settlement of Priniatikos Pyrgos: archaeological evidence, topography, and environment’, in Betancourt, P. P., Karageorghis, V., Laffineur, R., and Niemeier, W.-D. (eds), Meletemata II (Aegaeum, 20; Liège and Austin, 1999), 351–6Google Scholar; Vrokastro Reports 2, 62–3, 74–6, 374–5.

¹⁴ Results of the Vrokastro survey indicate that activity or occupation at the site probably began in the Final Neolithic, and may have continued through the Late Roman period. A few LM III C–7t h c. pithos-fragments found along the west side of the promontory testify to some activity between the prehistoric and early Greek periods. Recent use of the promontory includes the erection of a Venetian- or Turkish-period chapel (now destroyed) dedicated to the Panagia, according to local tradition. Recent to modern activity or construction on the promontory includes the one-week excavation by Hall in 1912 (one trench remains open on the central western side); the construction of a WW II bunker in the central part of the promontory (FIG. 4); agriculture and related construction (field walls, a threshing floor, two wine presses, a shed, a well); and the recent and illegal scraped road and field (4) along the eastern side of the promontory.

¹⁵ The subsurface probes were undertaken in 2004 from Katevati, south-west of Nisi Pandeleimon, west, to Ioannimiti. The subsurface investigation is intended to document the sequence and depth of colluvial debris flows down the river, possibly establish a chronology for their sequence, provide data concerning the long-term position of the river and the environment and configuration of the palaeo-coastline.

¹⁶ The mapping program was directed by T. Brennan, formerly of the University of Pennsylvania Museum.

¹⁷ These techniques have been used at Itanos, eastern Crete: Sams, A. et al. , ‘Ancient Itanos: creating an archaeological site in a remote sensing laboratory’, 157–64, in Jerem, E. and Biró, K. T. (eds), Archaeometry 98: Proceedings of the 31st International Symposium of Archaeometry (BAR S1043: Archaeolingua Central European Series, 1; Oxford, 2002), i. 157–64Google Scholar; Vafidis, A. et al. , ‘High resolution geophysical imaging of buried relics in the Itanos archaeological site’, in Doerr, M. and Sarris, A. (eds), Proceedings of the 30th CAA 2002 International Conference: Computer Applications and Quantitative Methods in Archaeology. (Herakleion, 2003), 25–50Google Scholar. In Boeotia: Gaffney, C. F. and Gaffney, V. L., ‘From Boeotia to Berkshire: an integrated approach to geophysics and rural field survey’, Prospezioni archeologiche, 10 (1991), 65–70Google Scholar. In the Laconia survey: Cavanagh, W., Jones, R., and Sarris, A., ‘The phosphate and geophysical surveys’, in Cavanagh, W. G. and Crouwel, J. (eds), The Laconia Survey 2 (London, 1996), 235–62Google Scholar. In the Pylos survey: Zangger, E. et al. , ‘The Pylos regional archaeological project, part II: landscape evolution and site preservation’, Hesp. 66 (1997), 549–642CrossRefGoogle Scholar.

¹⁸ For a brief description, Archibald 2002, 341–2.

¹⁹ Measurements of the magnetic field are made with a proton or caesium magnetometer, which measures the total magnetic field strength, or by a gradiometer, which measures the vertical or horizontal gradient of the total magnetic field or one of its corresponding components. The measurements (of the first derivative of the vertical component, or vertical magnetic gradient) are taken with a fluxgate gradiometer. The magnetic field is measured at a constant distance and close to the ground surface, and measurements are taken every half-metre within rectangular grids. At Priniatikos Pyrgos a Geoscan FM36 Fluxgate Gradiometer was used (this described in the unpublished 2002 technical report, Sarris I), and it was also used at Vetren in Bulgaria; see Archibald 2002, 343. Mapped results can tentatively be separated into archaeological, natural, and modern features. One potential problem is that igneous rock, such as the granodiorites known in the geology of the Vrokastro region, would be magnetic and thus detected using this technique.

²⁰ This technique is implemented by inserting metal probes a short distance into the ground equal distances along a traverse. Mares, S., Introduction to Applied Geophysics (Prague, 1984), 263CrossRefGoogle Scholar.

²¹ The electrode array that was chosen for survey at Priniatikos Pyrgos was the Twin Probe array, with a 0.75 m separation between the mobile probes. A Geoscan RM15 soil resistance meter was also employed for measurements of soil resistance.

²² Using this device, vertically polarized sinusoidal oscillations are transmitted and received. An alternating primary field is transmitted, which induces in rocks with dissimilar electric resistivities eddy currents or secondary magnetic fields. These secondary fields can be directed against or parallel to the primary field; by interference of the primary or secondary field, another field is produced, and finally received on the surface. Changes in the in-phase and out-of-phase components allow conclusions concerning the position of good or poor conducting subsurface bodies. This technique was used on the western side of the promontory (grid G), and in the field to the west (grid D).

²³ Each data set was coded according to a grid number, and data sets were given coordinates according to the position of the adjacent grids and an area code was given for each cluster of grids. A specific map coordinate system was chosen for the geophysical grids, registered to a geodetic system of coordinates based on the GPS/EDM mapping data.

²⁴ Data was preprocessed to create a common base level for all grids. This was necessary as data experienced a shift of average value within each grid surveyed, owing to differences in balancing the instrument and shifting of base/reference stations. Statistical analyses of both the common rows and the calculation of the average level of adjacent grids were made in order to provide a correction factor for each grid. The change of coordinates and the correction factors were used to create a synthetic mosaic of the grids, and in this way processing of the adjacent grids could be conducted simultaneously. Kriging interpolation was used for gridding data and to produce the geophysical maps. Selective compression of the dynamic range of values allowed for isolation of anomalies close to background level.

²⁵ Rectification was done by K. Kouriati. Satellite images were obtained by Ikonos platform, 2000.

²⁶ The earliest sherds from this area are Final Neolithic; a large well-burnished Final Neolithic bowl sherd was found in area B, directly south and east of the German bunker.

²⁷ Hayden 1999 (n. 13), 351–6.

²⁸ Hayden, B. J., ‘Elias to Nisi: a fortified coastal settlement of possible Late Minoan IIIC date in the Vrokastro area, eastern Crete’, in Karageorghis, V. and Morris, C. E. (eds), Defensive Settlements of the Aegean and the Eastern Mediterranean after 1200 B.C. Proceedings of International Workshop held at Trinity College Dublin, 7th–9th May, 1999 (Nicosia, 2001), 61–83Google Scholar; Vrokastro Reports 2, 138, 360.

²⁹ This area described in an unpublished report by Sards (n. 19), 11–13, figs. 5. 2–5. l. 9; walls and the features illustrated in figs. 5. 1. 7–5. 1. 9. These anomalies were confirmed by electromagnetic measurements; the vertical magnetic gradient map accorded with the magnetic susceptibility values obtained through measurement of the in-phase mode of the registered electromagnetic waves. High magnetic susceptibility values are correlated to the suggested kiln features.

³⁰ Hall (n. 12). A landowner (N. Kavousanos) who owns a central strip of the promontory stated that, according to local memory, field 3 (Area B) was excavated by Hall.

³¹ The top of this mass is 0.75 m above the level of field 3; this may not, however, indicate the pre-excavation level of the field. The area is discussed in the unpublished report by Sarris I, 14–15, figs. 5. 2. 1–5. 2. 4.

³² These blocks, which may have belonged to a large Bronze Age structure located on the headland, measure 0.40 × 0.40 × 0.90 m; one is still located on the surface directly south of Hall's trench. This kernos stone was found in the field wall framing the west side of Field 2.

³³ Sarris I, 14–15, figs. 5. 2. 1–5. 2. 2.

³⁴ Sarris I, 15–16, figs. 5. 2. 3; 5. 2. 4. Small sherds recovered during survey of the vineyard directly north of area C were predominantly Neopalatial, although tiny prismatic obsidian blades were also recovered. These tiny sherds were brought up from considerable depth in the course of deep cultivation for these vines. All small artefacts from the site (slag, pottery, furnace fragments) collected during the Vrokastro Archaeological Survey remain in storage at the INSTAP Study Centre, Pacheia Ammos.

³⁵ Vasilike vases have also been recovered from the east side of the promontory, however (pers. comm. J. Soles, K. Davaras).

³⁶ Other pieces of worked stone on the promontory include marble fragments, granodiorite, the harder, darker melanodiorite, andesite, and greenstone. The geologist C. G. Fassoules identified melanodiorite from the slope Kopranes near Vrokastro (pers. comm. H. Dierckx); it is one of the hardest stones in the island, and might be ideal for crushing ores, or other industrial activities.

³⁷ Measurements and Munsells of the tripod feet: E: L. 18.4; Th. top 6.9 cm; exterior surface was at least smoothed and there is one flat side. Temper includes quartz, limestone, granodiorite, abundant chaff. Surface: Munsell 2.5 yr 3/2; core: 2.5 yr N5-N4/. F: L. 20.3; Th. top 8 cm; one flat side, but more round in section than A. Plant temper, quartz/feldspar, some granodiorite?, gold mica, grey pebbles; also siltstone or phyllite (one large piece). Surface: Munsell 7.5 yr 6/6–5/6; core: 2.5 yr 4/6. M. Kostoglou (pers. comm.) has suggested these might have been used to plug holes placed in metal smelting furnaces, based on similar examples she has observed. Two similarly shaped objects with spalled surfaces from Quartier Nu, Malia, were on display in the Agios Nikolaos Museum in 2003, but are not part of a new display (2004). Moveable ovens from Akrotiri have feet half as large as these, though they are a different shape, and thin in section. See Sarpaki, A., ‘Processed cereals and pulses from the Late Bronze Age site of Akrotiri, Thera; preparations prior to consumption: a preliminary approach to their study’, BSA 96 (2001), 39Google Scholar figs. 6, 7. The scale indicates these feet measure c. 17 cm in length, 14 cm wide, and 5 cm thick. Large, probable metal cauldrons with tripod feet this large are seen in a miniature fresco depicting a coastal town, from Kea; see Morgan, L., ‘Island iconography: Thera, Kea, Milos’, in Hardy, D. A. with Doutrias, C. G., Sakellarakis, V. A., Warren, P. M. (eds), Thera and the Aegean World III (London, 1990), i. 255 fig. 2Google Scholar.

³⁸ These sherds also include possible beehive fragments or vases that were used for food processing (with incised, widely cross-hatched interiors), and these may be Middle or Late Minoan (pers. comm. K. Nowicki). The existence of this new feature was reported to the 24th Ephorate. Its stone walls are built of conglomerate and limestone, with pieces varying from 0.10 to 0.40 m. Very burnt stones to 0.20 m protrude through a shallow earth layer in the interior of the kiln, and bits of burnt twigs/roots were seen emerging from wet earth at its base.

³⁹ The kiln is type 1b, hemispherical with stoking tunnel, in Evely's catalogue of kiln types; R. D. G. Evely, Minoan Crafts: Tools and Techniques (SIMA 92. 2; Jonsered 2000), 301. For the Mochlos kiln, T. Brogan (pers. comm.) and Soles 1997, 425–31; Soles, J. C. et al. , Mochlos IA. Period III. Neopalatial Settlement in the Coast: The Artisan' Quarter and the Farmhouse at Chalinomouri. The Sites (INSTAP Prehistory Monographs 7; Philadelphia, 2003), 80–9Google Scholar, figs. 51–53.

⁴⁰ Pleiner 2000, 69, 141; Tylecote, R. F., ‘Iron smelting in pre-industrial communities’, Journal of Iron and Steel Institute, 203 (1965) 340–8Google Scholar.

⁴¹ Pleiner 2000, 109. Ore roasting occurs at 400–800° in an oxydizing environment. Roasted ore is more amenable to reduction and smelting, but is not always required, though with the harder to smelt magnetite ores, it might have been (ibid. 87, 89).

⁴² Ibid. 118–21.

⁴³ Ibid. 216.

⁴⁴ Concerning this area, Sarris had no comment about Grid 7, though magnetic susceptibility values were high along the southern edge. Sarris I, 20–21, figs. 5. 5. 2–5. 5. 7.

⁴⁵ These walls in shallow water indicate the existence of at least two large structures; rooms of the western building are still preserved. Walls are limestone and rectangular to square slabs of beach rock and marl were laid flat and appear to form a levelling course.

⁴⁶ One or two clay-lined channels appear in the scarp contiguous to or within this structure, and the kiln as a consequence might be interpreted as a cross-draft or channel kiln of Neopalatial date. V. Kilikoglou (Senior Research Scientist specializing in ceramic technology, Demokritos Laboratory) examined the kiln in 2004. He agreed that the feature was a ceramic kiln, but observed that the apparent floor of the kiln was too close to the top of the channels to interpret the kiln as a cross-draft type. The channels therefore may be related to another intrusive buried feature. On the other hand, the kiln was built against a slope, and the possible ‘floor’ may no longer be in situ.

⁴⁷ Sarris I, 19, figs. 5. 4. 2; 5. 4. 3.

⁴⁸ Part of a Late Minoan tripod cooking pot with thin tapered feet, rectangular in section, was observed in situ in a corner of one of these partially submerged rooms in 1999. All pottery within the first metre above the waterline in the eroded western scarp of the promontory is also Minoan (EM-LM). Thus associated pottery and elevation suggest a Bronze Age date for these submerged limestone walls. At 1.25–1.50 m above the beach, a thin stratum of Roman sherds mixed with slag was observed, During the Roman period, the western slope above the modern scarp must have extended further west, and may have overlain these massive walls now visible in the water, This does bring into question, however, the topographical situation of the headland in the Bronze Age if the walls in the water are of Bronze Age date; perhaps the surface of the western slope was much lower, and its present high elevation the result of millennia of occupation and deposition. It should also be noted that at the western end of the same embayment, conglomerate walls on the beach define a Roman site eroding into the sea (IM15). The siting of these Roman walls leaves open to question the date of the submerged walls just west of Priniatikos Pyrgos. See also n. 49.

⁴⁹ Difference in elevation between walls in the water and the surface of field D ranged to 1.96 m. A trench excavated at the south-east edge of field D in 2004, as part of the exploration of the palaeo-coastline, revealed Greek pottery and a clay stratum (possible surface) at a depth of 2 m. The Greek pottery associated with this stratum at this depth suggests that the nearby walls in the sea could belong to historical antiquity. This area is discussed in the unpublished report by Sarris I, 16–18, figs. 5. 3. 2–5. 3. 7; walls in D are illustrated in fig. 5. 3. 7.

⁵⁰ Cisterns and other holding tanks or containers as at Gouves; Vallianou 1997, 336, 337, 339. These features for pottery production are also mentioned by Michaelidis, P., ‘Potters' workshops in Crete’, SMEA 32 (1993), 15, 17Google Scholar.

⁵¹ The best-preserved architectural elements appear to be in area G, along the eroding western side of the promontory. Here prospection detected two phases, the earlier one likely belonging to the Bronze Age.

⁵² Vallianou 1997, 333–4.

⁵³ Ibid. 339.

⁵⁴ Ibid. 337–9. The channel kilns at Gouves are described as the older type B; for this see Davaras, K., ‘A Minoan pottery kiln at Palaikastro’, BSA 75 (1980), 115–26Google Scholar, with other Neopalatial examples found at Vathypetro, Agia Triada, Phaistos, Knossos, Zakros, and Kommos. Vathypetro: S. Marinatos, “Αυασκαφημεγὰρου Βαθυπὲτρου Κρὴτης ”, PAE 1951, 270; id., “Ανασκαφὴ ὲν Λυκὰστω καὶ Βαθυπὲ τρου Κρὴτης ”, PAE 1955, 310; id., Crete and Mycenae (New York, i960), 41; Agia Triada: Levi, D. and Laviosa, C., ‘Il forno minoico da vasaio di Haghia Triada’, ASA, n.s. 41–42 (1979–1980), 7–42Google Scholar; Phaistos: Levi, D., ‘La conclusione degli scavi a Festòs’, ASA, n.s. 27–28 (1965–1966), 351–4Google Scholar; figs. 43, 44; id.Festòs e la civiltà minoica I (Rome, 1976), 318, 327, 509; Knossos: Hood, M. S. F., ‘Archaeology in Greece’, AR 24 (1957), 24Google Scholar; fig. 6 b; Warren, P. M., ‘Knossos: Stratigraphical Museum excavations, 1978–1980. part I’, AR 47 (1980–1981), 75–9Google Scholar, figs. 6, 8 a-c; Tarling, D. H. and Downey, W. S., ‘Archaeomagnetic study of the Late Minoan kiln 2, Stratigraphical Museum extension, Knossos’, BSA 84 (1989), 345–52Google Scholar; Zakros: N. Platon,, “ΖΑΚΡΟΣ” Ergon, 1973, 106; id., “ΖΑΚΡΟΣ”, Ergon, 1975, 180–1; id., “Μεταλλουργικὸ καμὶνι στὴν Ζακρὸ τῆς Κρὴτης ”, in Πεπραγμὲνα τοῦ Δ᾿ Διεθνοῦς Κρητολογοκοῦ Συνεδρὶου, i (Athens, 1981), 436–46. Shaw 2001, 20 dates the Gouves channel kiln to the Neopalatial period.

⁵⁵ Vallianou 1997, 342. Vallianou believes the common sewage system and shared open courts suggest an administrative or social system (guild?) responsible for organizing the industrial area.

⁵⁶ Ibid.

⁵⁷ Geophysical prospection undertaken in 2003 in parking lots immediately south and east of the headland yielded little evidence for subsurface relics because of disturbance and filling. Coring undertaken in 2004 in this area, however, revealed significant evidence for continual flooding from the nearby river, which might have precluded settlement. The tops of buried walls of unknown date protrude through the surface at the south edge of one car-park, however, thus subsurface construction (recent?) still exists here. Roman sherds were visible on the surface in this area 10–20 years ago, and in Pendlebury's day (Pendlebury 1932–3).

⁵⁸ These resource include clay beds on the Vrionisi (I. Bassiakos, pers. comm.) and Nisi Pandeleimon promontories, to the east (Day 1997, 160).

⁵⁹ Although Hall's excavation recovered high-quality Minoan pottery, no reference has been made to the presence of potters' wheels from the site; Hall 1914, 84–5.

⁶⁰ Vallianou 1997, 341. She lists other Minoan sites with evidence for ceramic production in the form of wheels, kilns, or workshops.

⁶¹ For the relationship between habitations and workshops at Mochlos, see Soles 1997, 425.

⁶² At least the kiln in the scarp, which may not be a channel kiln, could be earlier than the Neopalatial period; see n. 46. Driessen and MacDonald suggest (n. 13), 211, based on published pottery, that LM IA is an important destruction phase at Priniatikos Pyrgos, since much of the published pottery (lacking a context) is of MM III–LM I A (early) date. Pottery from the site is discussed by Betancourt 1978 (n. 13), 381–7; id., 1985 (n. 13), 130–1; fig. 99. Also Niemeier (n. 13), 21–2; fig. 7; Popham and Coldstream (n. 13), 389–90; Warren, P. and Hankey, V., Aegean Bronze Age Chronology (Bristol, 1989), 64Google Scholar n. 28.

⁶³ Soles 1997; 2003 (n. 39).

⁶⁴ Ibid. 427.

⁶⁵ Pers. comm. H. Dierckx, and Appendix 2, Vrokastm Reports 2.

⁶⁶ See n. 45 for a brief description of this area.

⁶⁷ J. W. Shaw, A. Van de Moortel, P. M. Day, and V. Kilikoglou, ‘A LM IA pottery kiln at Kommos, Crete’, in Laffineur-Betancourt 1997, 323.

⁶⁸ Ibid. 323–4; Shaw 2001, 2. In some cases these crossdraft or channel kilns may have been used for lime production or metal; ibid. 20 n. 29.

⁶⁹ Shaw 2001, 15, 18, 20. See also Evely (n. 39), 298–312.

⁷⁰ Soles 1997, 426. Carbon 14 analysis of olive pits used as fuel date the Mochlos workshops to the 15th c.

⁷¹ Ibid. 430.

⁷² Ibid. 429–30. Soles believes that the palace at Gournia belongs to an advanced stage of the LM I period, indicating the importance of this phase at the site; Soles, J. C., ‘The Gournia palace’, AJA 95 (1991), 22, 24CrossRefGoogle Scholar. For the structure, see Hawes, H. B. et al. , Gournia, Vasilike and Other Prehistoric Sites on the Isthmus of Hierapetra, Crete (Philadelphia, 1908), 31–6Google Scholar.

⁷³ Driessen and MacDonald 37 fig. 4. 3 argue that LM I A, not I B, is the strong period at Priniatikos Pyrgos; which is one of their new foundations or a ‘New Era’ settlement of LM IA. The Protopalatial pottery from the site recovered through survey suggests occupation continued here from the Protopalatial period, however. See also above, n. 62.

⁷⁴ The earliest possible evidence for wasters at the site could be the FN/EM I quartz-tempered sherds found on the eroded northern tip of the headland. These appear burnt to blackened, but this could also be caused by millennia of exposure to salt water.

⁷⁵ The Roman kiln excavated by Davaras just west of Pyrgos village, for example, was 4 m in diameter and was preserved to a height of 3.15 m. Davaras, C., “Κεραμεικὴ καμῖνος εὶς Ιστρῶνα ὰνατολικῆς Κρὴτης”, ArchDelt 28A (1973), 110–15Google Scholar; Sanders 1982, 142.

⁷⁶ Davaras (n. 54), 122, discusses these round to oval kilns; many have a central pier, as, for example, one Orientalizing kiln at Knossos; also Coldstream, N. C., ‘Knossos: area of south–west houses, early Hellenic occupation’, BSA 92 (1977), 191–245Google Scholar. This kiln, with a diameter of 1.50 m, may be of comparable size to the C4 anomaly and the Minoan kiln on the north-west slope.

⁷⁷ For techniques to be used, see below, n. 130; these are discussed in Maniatis, Y. and Tite, M., ‘Technological examination of Neolithic to Bronze Age pottery from central to south-eastern Europe’, JAS 8 (1981), 59–76Google Scholar.

⁷⁸ Pleiner 2000, 149; fig. 36. 5, describes a chimney or shaft smelting furnace built of ceramic bricks, though this type is not known in the southern Aegean. The vitrified fragment was found on the western beach.

⁷⁹ P. Day (pers. comm.) has indicated that the granodiorite temper in sherds analysed from the KARP excavation at Istron does appear to differ somewhat from granodiorites used in Gournia pottery. Thus it may be possible to draw some distinctions. Conversely, in the modern era clay was carried several kilometres on donkey back, so Gournia and Priniatikos Pyrgos could have drawn on the same clay beds, or sources so similar as to not be distinguishable. See the comment on the proximity of two centres near Petras, and the difficulty of differentiating the products of these two sites: Day 1997, 223 n. 30. In terms of the resource base, Bassiakos has identified three clay beds within the Vrokastro area that could have been used for pottery manufacture. The one closest to Priniatikos Pyrgos is located on the coastal promontory of Vrionisi; one is in the Xeropotamos River, toward the eastern edge of the Vrokastro survey area (just south of a modern quarry in the river); and one is near the western edge of the survey area at the chapel of Agios Silas, on the national highway. It is also possible that pottery clay could be obtained from the marl hills of the Istron River valley, directly south. Day apparently test-fired some Neogene clays from Vrionisi (Day 1995, 160), and from the Nisi Pendeleimon promontory (Day 1997, 226). The Vrionisi sample is discussed in Hein, A., Day, P. M., Quinn, P. S., and Kilikoglou, V., ‘Geochemical diversity of Neogene clay deposits in Crete and its implications for provenance studies of Minoan pottery’, Archaeometry, 46 (2004), 357–84CrossRefGoogle Scholar. The other factor is the variety of granitics and diorites that can be contained in one source—a huge granodiorite outcrop revealed through new road construction just west of Gournia contains a wide range of granitics and diorites, from fine to coarse, dark to light, highly micaceous to less micaceous, with milky feldspars and pink quartz—thus one well-exposed source may contain many variants for the potter to chose from. Thus specific granodiorite sources used for clay tempers may be difficult to identify for three reasons: their ubiquity, the potential variety within one source; outcrops used in antiquity could now be deeply buried.

⁸⁰ Ceramic production at Gournia has been evidenced by potters' wheels, Evely, D., ‘The potters' wheel in Minoan Crete’, BSA 83 (1988), 84Google Scholar n. 8, 89, 90, 99, 103, figs. 4, 6. The industrial area at Gournia may remain unexcavated, perhaps at the fringes of the settlement (near the sea, to the west?). Hawes also suggested the metalworking area was at the north-western limits of the excavated area (n. 71), 33.

⁸¹ For a discussion of the distribution of granodiorite tempered ceramics in EM III–MM I, see Haggis 2000. For the absence of granodiorite temper in Neopalatial Gournia pottery (pers. comm. L. V. Watrous and D. C. Haggis). See also Watrous, L. V. and Blitzer, H., ‘The region of Gournia in the Neopalatial period’, in Betancourt, P. P., Karageorghis, V., Laffineur, R., and Niemeier, W.-D. (eds), Meletemata III (Aegaeum, 20; Liège, 1999), 908Google Scholar. Watrous and Blitzer mention the possibility that these granodiorite vessels may have been imported to Gournia (from the Meseleroi area), and that by LM I A they were no longer being imported; phyllite and quartz-tempered vessels at Gournia then take their place (the source for these vessels the east side of the Isthmus?). When granodiorite tempered ceramics were found at Mochlos and Pseira, they were generally described as likely imports from Gournia or the Isthmus area; Brogan, T. (pers. comm.) and Barnard, K. A., ‘Transition, Production, and Standardization in Minoan Ceramics’ (PhD. diss., University of Pennsylvania, 2001), 269–70Google Scholar n. 20. At Mochlos clay fabrics 6 and 7 are tempered with granodiorite; Barnard 2001, 269–70 n. 20, and chart II.3. Type 6 is Mirabello cooking fabric (EM II–III), and type 7 is from a closed vessel (this dates more generally to MM). Type 6/7 Barnard associates with Kavousi fabrics II, III, FV. All Mochlos granodiorite-tempered imports to the site total only 5% of the assemblage, and most are of EM-MM date, which means that Gournia could be the source. Barnard's late deposit contained more MM closed granodiorite vessels, but these are no more than 3%. They are described as medium-sized jars. See also Barnard, K. A. and Brogan, T. M., Mochlos 1B: Period III (INSTAP; Prehistory Monograph 8; Philadelphia, 2003), 7–8Google Scholar. Mirabello imports to Pseira are also mentioned in Floyd, C. R., Pseira V: The Plateia Building (University Museum Monograph, 102; 1998), 59, 74, 249–53Google Scholar. Percentages of granodiorite-tempered sherds in parts of Pseira town are sometimes higher than Mochlos totals for this import (a LM I context in Room BS9, for example, produced a high 20% of undiagnostic Mirabello cooking fabric, p. 249); if these vessels were manufactured in LM I, Gournia was not the likely source.

⁸² Granodiorite fabrics have been characterized as containing quartz, biotite, amphibole, plagioclase feldspar, and hornblende, and occur in a buff jar, jug, and cup fabric (Mirabello ware) and in a cooking fabric with a red clay matrix (Mirabello cooking ware). This fabric has been analysed by Myer, G. H., ‘Ceramic petrography’, in Betancourt, P. P., East Cretan White-on-Dark Ware (University Museum Monograph, 51; Philadelphia, 1984), 60–6Google Scholar (he analysed a MM IA cup from Priniatikos Pyrgos, p. 65). The fabric is discussed by Day, 1995, 159–62; id. 1997, 224–6. For another analysis, see P. P. Betancourt, C. Davaras et al., ‘Appendix A’, Pseira I. The Minoan Buildings on the West Side of Area A (University Museum Monograph, 90; Philadelphia, 1995), 144–5. Granodiorite fabrics are also described in Haggis, D. C. and Mook, M. S., ‘The Kavousi coarse wares: a Bronze Age chronology for survey in the Mirabello area, east Crete’, AJA 97 (1993), 273–4CrossRefGoogle Scholar.

⁸³ During the LM I–II I B periods, ceramics tempered with phyllite in the Vrokastro area are recognized as ‘imports’ to the survey area, although these ceramics may only be from nearby Gournia or the Isthmus (the lack of phyllite in Gournia's environment means that these vessels must have been imports there as well). During the Protogeometric and Geometric periods, phyllite- tempered jars and cooking fabrics are much more frequent at Vrokastro, the primary centre for the survey region from the 12th to the 8th c. BC. Granodiorite and quartz occur again as pottery temper in pithoid jars and amphorae during the Greek and Roman periods in the Vrokastro region, and this fabric predominates in jar fabrics from the 7th to 4th cc. BC. Quartz appears to dominate as temper in the cooking fabrics of the historical period, however.

⁸⁴ Identifying individual pottery production centres by motor habits and idiosyncratic manufacturing techniques has been suggested by Van de Moortel. This is not feasible for an unexcavated site, however. See Van de Moortel, A., ‘Pottery as a barometer of economic change’, in Hamilakis, Y. (ed.), Labyrinth Revisited, Rethinking Minoan Archaeology (Oxford, 2002), 206Google Scholar. One slight indication that Priniatikos Pyrgos fabrics may differ from other regional workshops can be taken from analysis of one MM I dark-on-light cup, which contained more hornblende than other east Cretan pottery groups; see Myer (n. 82),65.

⁸⁵ This study concentrated on eastern Crete and sherds from 19 sites underwent analysis, which included 600 samples and 150 examples of raw material (clay). Day 1997, 222, 227. In the Achladia publication, Day 1995, 150, mentions analysis of samples from 30 sites in the island.

⁸⁶ Day 1997, 223, 225. Haggis 2000.

⁸⁷ Driessen, J. and MacGillivray, J. A., ‘The Neopalatial period in east Crete’ in Laffineur, R. (ed.), Transition: le monde égéen du Bronze moyen au Bronze récent (Aegaeum, 3; Liège, 1989), 99–111, esp. 101Google Scholar. Knappett, C., ‘Assessing a polity in Protopalatial Crete: the Malia-Lasithi state’, AJA 103 (1999), 638–9CrossRefGoogle Scholar.

⁸⁸ Haggis 2000. For granodiorite-tempered imports to Malia in MM-LM, Knappett (n. 87), 632. Many other references to this pottery fabric found in central Crete (Knossos), Palaikastro, Myrtos (Fournou Koriphi and Pyrgos) belong to early periods (EM II/III–MM I); Day 1997, 225. Day 1995, 159–62 discusses the coastal distribution of Bronze Age granodiorite-tempered pottery exported from Mirabello; these exports were found at Knossos in EM II B, EM III, and now perhaps in EN, as well as in Lasithi and other areas of eastern Crete (Vrokastro Reports 2, 99). For granodiorite temper identified at Knossos in a Neolithic sherd, Tomkins, P. and Day, P. M., ‘Production and exchange of the earliest ceramic vessels in the Aegean: a view from early Neolithic Knossos, Crete’, Antiquity, 75 (2001), 259–60CrossRefGoogle Scholar.

⁸⁹ Chrysokamino: Muhly, J. D., ‘Chrysokamino and the beginnings of metal technology on Crete and in the Aegean’, in Day, L. P., Mook, M. S., and Muhly, J. D. (eds), Crete Beyond the Palaces: Proceedings of the Crete 2000 Conference (INSTAP Prehistory Monographs, 10; 2004), 283–9Google Scholar. Betancourt, P. P. et al. , ‘Research and excavation at Chrysokamino, Crete, 1995–1998’, Hesp. 68 (1999), 343–70CrossRefGoogle Scholar. Poros, Herakleion: N. Dimopoulou, ‘Workshops and craftsmen in the harbour-town of Knossos at Poros-Katsambas’, in Laffineur-Betancourt 1997, 434–5. Malia: Driessen, J. and Farnoux, A., ‘Mycenaeans at Malia?’, Aegean Archaeology, 1 (1994), 61Google Scholar.

⁹⁰ Muhly (n. 89), 283–9.

⁹¹ Soles 1997, 426. Evidence included hoards and implements for casting, but no crucible fragments, bellows, or tongs.

⁹² Copper prills appear to be present in small pieces of slag recovered through survey from Nisi Pandeleimon. More iron slag was observed in a road-cut on the east facing slope above; this slope does not face updraft winds from the north-west. Copper-smelting slags contain isolated unreduced ore particles, and sporadic copper prills. In general, analysed copper slags yield a small percentage of copper, and 30% iron, which is added in copper-smelting as flux (that is, haematite or limonite). The presence of a small amount of iron in copper slag probably led to the first identification of iron in antiquity; Pleiner 2000, 12, 254, and also perhaps to the development of iron as a by-product of working copper in Cyprus; see Sherratt 1994. Also F. Tholander, ‘Appendix II: Hypothesis on Iron-Extraction from Ancient Copper Slag’, in Muhly et al. 1982, 180–1; Gale, N. H. et al. , ‘The adventitious production of iron in the smelting of copper’, in Rothenber, B. (ed.), The Ancient Metallurgy Copper (London, 1990), 182–91Google Scholar. On the other hand Madden says the iron produced as a by-product of Cypriot copper smelting would have been of little use, because of its ‘unacceptable inferior properties’: R. Madden, ‘Early iron technology in Cyprus’, in Muhly et al. 1982, 303. Cyprus is described by Madden as not having iron ores, while Waldbaum reports that iron-rich ores have been located in Cyprus; see Waldbaum 1982, 326 and Bear, L. M., The Mineral Resources and Mining Industry of Cyprus (Nicosia, 1963), 53–4, 56–7, 64, 119Google Scholar; figs. 2, 6. According to Snodgrass, Cyprus is rich in iron found in copper ore (copper ores were found containing up to 73% iron ore); A. M. Snodgrass, ‘Cyprus and the beginnings of iron technology in the eastern Mediterranean’, in Muhly et al. 1982, 293. The last word on Cypriot iron production may have been offered by Kassianidou, who describes abundant iron in the form of the yellow ochre, umbers, and gossans of Cyprus. She also states that iron can be derived as a by-product of copper smelting, but in insufficient quantities to make this a probable long-term source for the manufacture of iron artefacts; see Kassianidou, V., ‘Could iron have been produced in Cyprus?’, RDAC (1994) 76–78Google Scholar.

⁹³ The major evidence for iron-working at Kommos dates to the 8th c. BC; there was a resurgence in the 4th c. BC, and then an increase towards the end of the first millennium BC. See Rehder 2000, 80.

⁹⁴ Haematite was the source for the Kommos ore; Markoe 1998, 236 n. 7.

⁹⁵ This evidence is summarized by Markoe 1998, 235; Faure 1966, 52–4 reported that the site near Lasaia was rich in magnetite and described the bowl furnace. Magnetite would be the iron ore associated with igneous (granodiorite) rock, which is present in two outcrops in the south-west Asterousia mountains according to the IGME map (1972) by G. Katsikatsos, with E. Davi and M. Bonneau. Varoufakis reports that analysis of Lasaia slag indicates a high phosphorus content, which suggests that the ore was Cretan; G. J. Varoufakis, ‘The origin of Mycenaean and Geometric iron on the Greek mainland and in the Aegean islands’, in Muhly et al. 1982, 318.

⁹⁶ Faure 1966, 52–4, though he believed this to be copper slag; Bassiakos has recently collected iron slag from this site (pers. comm.); Sanders 1982, 160 and bibliography. This site is 8 km west of Lebena.

⁹⁷ Savignoni, L., ‘Explorazione archeologica delle provincie occidentali di Creta’, MA 9 (1901), 395–6Google Scholar; fig. 85; Davies 235, 267–8. Davies suggests that copper or gold was smelted at this site, and this account is repeated by Sanders and Raab; see Sanders 1982, 171, with bibliography concerning the site. Raab, H. A., Rural Settlement in Hellenistic and Roman Crete: the Akrotiri Peninsula (BAR S984; Oxford, England, 2001), 13Google Scholar. Davies found Roman and later pottery and suggests (268) a medieval date for the smelting operation.

⁹⁸ Chalcopyrite and galena produce copper; Shepherd, R., Ancient Mining (London, 1993), 93Google Scholar.

⁹⁹ Davies 1935, 267. At Priniatikos Pyrgos, the Roman concrete platform south of the bunker, described above, is probably part of a building, not a support for smelting furnaces.

¹⁰⁰ J. Moody (pers. comm.) believes that iron was smelted at Kantanos, and possibly Sklavopoula, where Roman tombs and small adits were found. Copper production was initially proposed by Davies (see n. 97). For analysis of the slag and description of Sklavopoula and Kantanos, Davies 1935, 267. The high iron content of the Kantanos slag might mean that iron was not the objective of the smelt, but iron slag with a high iron content was reported from mainland sites by Davies (1935, 241 n. 1; 246 n. 3; 252 n. 8) and also by Pleiner 2000, 252, who reports a typical 40–70% iron content for iron bloomery slag. Conversely, Pleiner 2000, 12 mentions that copper slag often has a high iron content, and that it can be similar in appearance to slag produced by bloomery iron smelting; see also Snodgrass (n. 92), 293. This ambiguity as to the type of ore smelted at Kantanos (and other sites) indicates the need to investigate these smelting operations in the island more thoroughly.

¹⁰¹ This slope is north of KM 1; see Vrokastro Reports 2, fig. 14. Much of this slope has been destroyed by recent highway construction. Sites containing Greek and Roman components were identified by survey in the Kendromouri hills, and a large Roman site is located directly south of Gournia (pers. comm. V. Watrous). Possibly related to this, a large piece of slag found at Gournia was analysed, and had an average iron content of 46.5%; G. Sperl, ‘Ein Fund aus Gurnia (Kreta) und das Problem der Schlackenbeurteilung’, in id. (ed.), The First Iron in the Mediterranean (Proceedings of the Populonia/ Piombino 1983 Symposium; Strasbourg, 1988), 163–7, and 164 n. 5 for comment on abundant slag found within the excavated site. This could suggest that the Minoan site on the Gournia ridge was used for iron smelting in the historical period. That would parallel historical metallurgical activity over the Minoan site on Priniatikos Pyrgos.

¹⁰² Pleiner 2000, 90.

¹⁰³ These mines are modern, but ores from these sources can be analysed for comparative purposes. For their location, Markoe 1998, 234, map illustrated in fig. 1. This map was prepared by C. Zervantonakis, who shows other ore sources in “Συμβολὴ του ορυκτοὺ πλοὺτου τνς Κρὴτης στην ανὰπτυξη του πολιτισμοὺ της αρζαιὸτητας ”, Πεπραγμὲνα του Ε' Διεθνοὺς Κρητολογικοὺ Συνεδρὶου, i (Herakleion, 1985), 124Google Scholar. Another 1965 map published by the Institute for Geology and Subsurface Research in Athens shows the mineral resources of the island. Two iron-ore sources are indicated on this map just north of the location of the mines; Markoe 1998, 235 fig. 2. This map is also shown in A. M. Snodgrass, ‘Iron and early metallurgy in the Mediterranean’, in Wertime-Muhly 1982, 353 fig. 10. 2.

¹⁰⁴ Markoe 1998, 236 reports that the maps of iron sources in the island and results of fieldwork indicate there is not much iron ore in eastern Crete except for the Gulf of Mirabello area. Faure 1966, 62, mentions possible iron sources near Katharo, Kritsa, and at Agios Antonios, Mirabello.

¹⁰⁵ Magnetite is disseminated through most igneous rocks and if the ore smelted is from near the survey area, this is a probable source, but Hall reports iron present in the form of limonite in a charge found in the Vrokastro setdement (see below, n. 120). Magnetite ore can be dense and difficult to reduce, but this ore could be roasted to prepare for smelting; Pleiner 2000, 87, 89, 265. This ore source is also discussed by Healy, J. F., Mining and Metallurgy in the Greek and Roman World (London, 1978), 39–41Google Scholar. Rapp defines granodiorite and explains that, with the solidification of igneous rock from the crystallization of the magma, associated metal-rich solutions form important metal deposits. Rapp, G. R., Archaeomineralogy (Natural Sciences in Archaeology; Berlin, 2002), 43CrossRefGoogle Scholar table 3.1.

¹⁰⁶ Oxide ores can be found in river flood plains, bogs, river banks, lakes, and marshes; Pleiner 2000, 88 and I. Bassiakos (pers. comm.).

¹⁰⁷ Varoufakis, G. J., ‘Greece: an important metallurgical centre of iron in antiquity’, in Pleiner, R. (ed.), Archaeometallurgy of Iron (International Symposium of the Comité pour la sidérurgie ancienne de l'UISPP; Prague, 1989), 279–80Google Scholar; 281 fig. 1 for a map of mines in west Crete. Some of these iron sources have been exploited in the modern era by Halyvourgiki Inc., Eleusis.

¹⁰⁸ This phosphorus-rich iron was used in a tripod found at Olympia dating to the Geometric period; ibid. 282, and Varoufakis (n. 95), 318, 320 map 1.

¹⁰⁹ Markoe 1998, 234 and Drews, R., ‘Phoenicians, Carthage and the Spartan eunomia’, AJP 100 (1979), 46Google Scholar. For mythological references concerning iron ore in Crete, the importance of Cretan early ironworking, and areas in Crete that maintained contact with the Orient, including Mirabello, see Morris, S. P., Daidalos and the Origins of Greek Art (Princeton, 1992), 151, 162Google Scholar.

¹¹⁰ The possibility of a northern route is mentioned by Betancourt, who noted the presence of possible imports (‘scarabs’) from Vrokastro; Markoe 1998, 241. These are faience seals with convex backs decorated with two shells and pseudo-hieroglyphs on sealing surfaces; Hall 1914, 135–6, fig. 81, pl. 35. Hoffman, G. L., Imports and Immigrants. Near Eastern Contacts with Iron Age Crete (Ann Arbor, 1997), 86–7Google Scholar.

¹¹¹ Snodgrass (n. 103), 349.

¹¹² Waldbaum 1982, 325–6 and bibliography. Cyprus seems to have priority because of the number of early iron artefacts demonstrating sophistication in manufacture, though the type range is limited (belonging to the 11th c., and a few to the 12th); see nn. 115, 124, below.

¹¹³ For a discussion of the origin of these artefacts, see Hoffman (n. 110), 150–89; also Morris on the history of scholarship concerning production of metal artefacts from the Idaean Cave—whether by Cretan or Oriental craftsmen (n. 109) 152 nn. 10, 11.

¹¹⁴ A few earlier pithoi or jar sherds found on these promontories may belong to the LM IIIC-Geometric periods. For evidence of activity in this zone during these periods see Vrokastro Reports 2, 138–9.

¹¹⁵ These early artefacts include iron knives or daggers with bronze rivets; there are examples from Vrokastro corbel-vaulted tombs 1, 5, and 6, which belong at the earliest to LM III C (late); Hall 1914, 139 (tomb 1), 151 (tomb 5), 153 (tomb 6); pl. 21 a, f. Also Hoffman (n. 110), 102 n. 83, 103–4; Waldbaum 1982, 343 nos. 93, 94, pl. 34: 15. Bibliography for these is presented by Hoffman 1997 (n. 110), 102–3 n. 83, 104. (Note: the iron knife from corbelled tomb 5 is published by Hoffman and Waldbaum as having two bronze rivets; mentioned by Hall 1914, 151, only as iron.) The tombs are also discussed in Vrokastro Reports 1, 6–7, 8–9. These knives with bronze rivets found in Crete are now generally regarded as having been made in Crete; Waldbaum 1982, 325–49. Sherratt 1994, 59–106, argues that the knife with bronze rivets was first a product of Cyprus and was widely disseminated in the Aegean. The Cypriot production belongs to the 12th and 11th cc, and the early use of iron may derive from iron found as a by-product in Cypriot copper (see discussion, n. 92). Sherratt (p. 81) argues that Crete lacked the copper sulphide ores that produced iron, and would therefore not have had an iron technology earlier than the Cypriot; also see Snodgrass (n. 92) on the iron content of Cypriot copper ore.

¹¹⁶ Hall 1914, 123–39; Corbelled tomb 1 is discussed in Vrokastro Reports 1, 8–10; most of the published pottery from the tomb indicates that it is not one of the earliest in the series, yet one stirrup jar from this tomb is described by Hall 132 as comparable to an example from a LM IIIC (late)-SM tomb in Chavga.

¹¹⁷ Hall 1914, 138–9; length 0.22 m, weight 2.475 kg. Pleiner 2000, fig. 2: 3, illustrates another large example from Lindos, length 28.5 cm. Hall describes the shaft hole for the Vrokastro example as possibly too small for mounting on a shaft, suggesting it was possibly a symbolic or ceremonial piece; the large double axe illustrated by Pleiner does not appear to have a shaft hole (which could indicate a later date). Other iron implements from corbelled tomb 1 included an adze, spear-end, curved iron knife, slender iron knife, a chisel, and an iron mass containing many spear-ends, knives, and swords; all told, at least 25 weapons were recovered from this tomb.

¹¹⁸ Snodgrass (n. 103), 335, 368, points out that iron is used for jewellery in an experimental stage of its development (stage 1); possible iron fibulae from Vrokastro are in the Herakleion Museum (in poor condition). All are from the bone enclosures, which might indicate a 9th-c. or later date for them (though some could be heirlooms); these vary in type from simple bow to examples that could be later with knobs or mouldings; Herakleion Museum cat. nos. 3309, 3310, 3312, 3313, 3314. 3318. Hall 1914. 140, 156, 165, pl. 19 d, g.

¹¹⁹ Hall 1914, 109. This was compared by Hall with another iron bar from an 8th-c.-BC tumulus (III) at Gordion; Körte, A. and Körte, G., Gordion: Ergebnisse der Ausgrabung im Jahre 1900 (JdI, Ergänzungsheft 5; Berlin, 1904), 79Google Scholar; fig. 69 6. The cult function of room 17 was suggested by Mazakaris-Ainian, A. J., From Rulers' Dwellings to Temples: Architecture, Religion and Society in Early Iron Age Greece (1100–700 b.c.) (Studies in Mediterranean Archaeology 121; Jonsered, 1997), 213–214Google Scholar, n. 1683.

¹²⁰ The bowl contained iron in the form of limonite, lime in the form of calcite, silica quartz-sand (the flux), and a small amount of aluminium; Hall 1914, 110. The contents of the bowl were obviously tested, but the bowl has not been found in the collection of the University of Pennsylvania Museum. This charge must have been intended for a furnace.

¹²¹ Hall 1914, 115–16. A cache of bronze weapons (and two iron spears) was also found near a shrine (room 11) at the south-eastern edge of the site; shrine discussed in Hayden, B. J., ‘Vrokastro terracotta figurines, figures, and vase attachments’, Hesp. 60 (1991), 105–9CrossRefGoogle Scholar, figs. 2, 3). The metal can be viewed either as a possible dedication, or as a cache for recycling. Hall 1914, 102–5; illustrated in fig. 59. P. 105.

¹²² This seen by Hayden just above the lower settlement area; Bassiakos has also seen slag on the north slope.

¹²³ Partially worked marble bowls (‘blossom bowls’) and a stone cup were found within the south-central and south-western summit, rooms 30, 36 (Hall, 1914, 114). It is probable that these bowls were manufactured during the Protopalatial occupation of the summit. Similar bowls (described as serpentine heirlooms) were also found in a LM III C context at Kephala Vasilikis; Eliopoulos, T., ‘A temple complex of the Dark Ages at Kephala Vasilikis’, in Karageorghis, V. (ed.), Proceedings of the International Symposium Cyprus in the 11th Century BC (Nicosia, 1994), 306Google Scholar. Sockets cut into rocks placed in some rooms within the Vrokastro settlement could have been used for textile manufacture; these are less probably for pottery wheels, as no fragments of wheels were identified from the site.

¹²⁴ With one exception, a fragmentary iron sword from Vrokastro, MS 4766 (max. L. 20.2 cm, max. W. 2.2 cm, max. Th. 4 mm.); this is identified as a ‘slender iron knife’ from corbel-vaulted tomb 1 on the catalogue card (Hall 1914, 138 n. 14), but Hall's dimensions, 13 cm long, 1.2 cm wide, do not accord with the artefact, which on the basis of testing by Z. Carrol and Dr. S. K. Nash for MASCA at the University of Pennsylvania Museum (Dr. S. Fleming, Director) dates at the earliest to the 11th c, but more probably belongs to the 10th c. BC. Although the blade contains some low-carbon steel, it is not evenly distributed. Carrol and Nash report that one cutting edge is carbon-free while the other contains less than 0.1% carbon. This lack of homogeneity is probably indicative of the composition of the smelted iron and does not appear to be the result of decarburization of the actual artefact. Such mixtures occur when pieces of raw iron and steel are ‘bundled’ together at forging temperatures to create one piece of iron. Twins (Neumann bands) appear in the decarburized areas indicating coldworking, that is, the object received a sharp blow at cold temperatures. Carburization or steeling results when the ore absorbs carbon from a charcoal fire; carbon is primarily found in the outer layers of the artefact and produces a stronger iron. It can be the accidental result of carbon trapped in the iron bloom and forged into a finished object, but when deliberate it indicates a knowledge of how to make iron stronger. Testing of Cypriot artefacts indicates that carburizing was practiced by smiths in the 11th c; Madden (n. 92) 304, 307, 311 and tables i–ii, pp. 308–9. Also Sherratt 1994, 68 n. 11; J. Waldbaum, ‘The first archaeological appearance of iron and the transition to the Iron Age’, in Wertime-Muhly 1982, 88; Tholander, E., ‘Evidence for the use of carburized steel and quench hardening in Late Bronze Age Cyprus’, OpAth 10 (1971), 15–22Google Scholar. Carburization might have been used intermittently depending on artefact function; in the case of this Vrokastro knife/sword, to strengthen the cutting edge. Because the degree of carburization varies from side to side, however, deliberate carburization is questionable, though the knife is in poor condition and the edges only partially preserved.

¹²⁵ Slags are heavy to light, dark grey, brown-black heterogeneous silicate complexes formed from gangue minerals of iron ores. They have significant quantities of iron oxides.

¹²⁶ See also above, n. 105 (magnetite ore related to granodiorite in the regional geology). Composition and melting temperature of slag may indicate quality of the metal and yield of the smelting process; Pleiner 2000, 265. Kresten's division of slags into categories with descending quantities of iron (wustite, olivines, pyroxenes) includes quartz-types at the bottom; the last may be the primary source for the Vrokastro area; P. Kresten, The Mineralogy and Chemistry of Selected Ancient Iron Slags from Dalarna, Sweden (Jernkontorets Berghistoriska utskott, H 29; Stockholm, 1983).

¹²⁷ Pleiner 2000, 254–5, 263.

¹²⁸ Techniques include wet chemical analysis that gives average values of contents, that is, overall composition, and relative proportion of iron, which allows for comparisons with other slags; electron microprobe analysis (EPMA) reveals composition of microscopic areas. Optical microscopy can determine the crystalline phases of slag, and results can be confirmed through X-ray diffraction analysis; Pleiner 2000, 251–2. Also Bachmann, H.-G., The Identify of Slag from Ancient Archaeological Sites (Institute of Archaeology Occasional Publication, 6; London, 1982), 31–3Google Scholar.

¹²⁹ LIBS testing under the supervision of S. Chouveraki, chief conservator at the INSTAP Study Centre, and I. Bassiakos.

¹³⁰ This technique is discussed in Ferrence, S., Melessanaki, K., Mateo, M., Betancourt, P. P., and Anglos, D., ‘Adaptation of laser-induced breakdown spectroscopy (LIBS) for the analysis of archaeological artefacts’, in Fostem, K. P. and Laffineur, R. (eds), Metron: Measuring the Aegean Bronze Age. Proceedings of the 9th International Aegean Conference (Aegaeum, 24; Liège, 2003), 111–15Google Scholar.

¹³¹ On arsenic in ore, Pleiner 2000, 265; Piaskowski, J., ‘Das Vorkommen von Arsen in antiken und frühmittelal terlichen Gegenständen aus Renneisen’, Zeilschrift für Archäologie, 18 (1984), 313–26Google Scholar. Varoufakis comments on the impurities of iron ores in the Aegean, especially in ores from the Cyclades, which include arsenic, phosphorus, sulphur, copper, though the haematite ores of Laconia are generally free of these. Note: no data were forthcoming on the presence of phosphorus, which is found in iron ores from Crete (see n. 95); see Varoufakis 1989 (n. 108), 315–22. Haematite iron ore is phosphorus-rich, and phosphorus can be derived in part from charcoal ash (Pleiner 2000, 265).

¹³² In this regard it is interesting that Waldbaum cites early slag containing arsenic from Anatolia and the mainland. Arsenic in slag has been reported from Tiryns in a LH III B context and speiss from Boğazköy (speiss results from smelting iron containing arsenic and sulphur): Waldbaum, J. C., ‘The coming of iron in the eastern Mediterranean, thirty years of archaeology and technological research’, in Pigott, V. C. (ed.), The Archaeometallurgy of the Asian Old World (University Museum Monograph, 89; Philadelphia, 1999), 31Google Scholar. Davies 1935, 240 n. 6, 241 n. 2, 253 n. 3 also reports testing iron ore or slag from mainland sites that contained arsenic.

¹³³ Pleiner 2000, 251.

¹³⁴ The size of the west slope metalworking area does not include the slag found in the north-east part of Area D to the west, or the beach between the headland and D, where slag was found, or the eroded north-west tip of the promontory.

¹³⁵ Slag can be differentiated from iron ore by weight, appearance, and lack of magnetic attraction; Rehder 2000, 81, 88.

¹³⁶ Tylecote, R. F., The Early History of Metallurgy in Europe (London, 1987), 311Google Scholar; Pleiner 2000, 172, 257, 259, 263 fig. 70.

¹³⁷ Tylecote, R. F., ‘Furnaces, Crucibles, and Slags’, in Wertime-Muhly (eds), 216Google Scholar.

¹³⁸ Its condition precludes certainty, and in the hundreds of pieces observed at the site, this one possible example is scarcely sufficient to establish the technique.

¹³⁹ Tylecote (n. 137), 311 has pointed out that the dimension of slag ‘block’ can be used to establish the approximate size of the furnace; also Pleiner 2000, 142. This slag at the base of the furnace is located below the bloom, which contains the smelted iron. This one example from Priniatikos Pyrgos suggests a furnace no larger than the partially preserved example from Kommos.

¹⁴⁰ These were found in a rock crevice close to the northern tip. The large spike head was split into four quadrants; this technique suggests a medieval to modern date.

¹⁴¹ This material was recovered with excavation at Kommos, related to the ironworking area. Rehder 2000, 80.

¹⁴² These fragments vary from flat to convex/concave; one appears to be thrown (A) and one is made of two coil-built walls, joined together (C). Dimensions and Munsells are (A) D. 15.2; Th. 4 cm; lighter swirls of grey-green on surface, 5 y 7/2–6/2; core 7.5 yr N3/–N2/; burnt out limestone can be seen in the clay fabric. (B) D. 10.4; Th. 4.8 cm; vitrified surface 5 y 4/1; clay core 10 r 4/3–2.5 yr 5/4; other less burnt surface 10 yr 6/4–6/6. Temper includes granodiorite, plant, quartz/feldspar, gold mica (C) D. 8.7; Th. 5.4 cm Swirls on outer surface 5 y 5/2. 6/2. Other darker convex side, surface 5 y 3/1; core 7.5 yr N4/–N3/. Burnt-out limestone holes visible. (D) D 2.8; Th. 4 cm; surface 5 yr 5/6; core 10 yr 4/1–4/2.

¹⁴³ Rehder 2000, 85 pls. 1. 194–1. 196; the furnace was found 25 m east of the temples, near building V. Early shaft furnaces are usually 0.60–0.80 m high; Pleiner 2000, 172.

¹⁴⁴ Vitrification occurs because of the low heat resistance of clay walls. The linings react with molten fayalitic slag to create a glassy surface (Pleiner 2000, 257). In some cases this might have been intentional, to influence slag formation and have lower viscosity in the slag; ibid., 258.

¹⁴⁵ Analysis revealed this was a typical bloomery slag with a ferrous oxide content, and indicated that a moderately low carbon-content iron was being produced at Kommos. The temperature required was c. 1150°, and this type of furnace can be used for copper as well; Rehder 2000, 85–8; the furnace is illustrated in pl. 1. 199 and the fragment of furnace with tuyère slot is illustrated in pl. 1. 196: Mi 178. Early Iron Age blooms from Europe usually weigh 1–3 kg; Pleiner 2000, 231. It is not known if the excavated area containing the slag at Kommos was as large as the distribution of slag visible on the surface at Priniatikos Pyrgos. Related tools or tuyères were not found in the Kommos excavation. Rehder 2000, 85, 86, 88.

¹⁴⁶ Rehder 2000, 86.

¹⁴⁷ Healy 1978 (n. 105), 184; Pleiner 2000, 141.

¹⁴⁸ Rehder 2000, 86. Calculations involving the size of the slag heap in relationship to the number of furnaces and iron produced are discussed in Pleiner 2000, 267. Iron to slag ratios have been calculated at 1: 5, but the three-dimensional depth of the slag heap must be known as a preliminary to these calculations.

¹⁴⁹ pleiner 2000, 131–2; carbon dioxide and monoxide in a reduction atmosphere release the metal; temperatures required are c.1200°.

¹⁵⁰ Ibid. 142.

¹⁵¹ One round cistern or basin, probably Roman, was noted along the south side of the open trench Hall left on the site. Its small size suggests it could have been used for washing or quenching.

¹⁵² Pleiner 2000, 13, 145–6. Bowl furnaces are cobble built oval enclosures no more than 1 m across. They produced a soft iron, and could have been roofed with clay for smelting or simply covered with charcoal.

¹⁵³ In fields south of the headland, Sarris noted no distinct variations of magnetic susceptibility, and most values were above 40 emu/gr. Measurements of the frequency-dependent susceptibility were above 10%, reaching a maximum of 29.51%, compared to that of 1765% measured in the probable kiln complex of Area A (Grids 3, 4). Similarly, there were only a few isolated areas of low susceptibility, with a slight drop in susceptibility measurements along the farthest traverse, to the south-west.

¹⁵⁴ Unpublished technical report, Sards II, 20–3, figs. 5. 4. 1–5. 4. 10. Architectural relics were found in grids C–I, and high soil resistance values in 4-m wide anomalies in area C (south-east) and H–I (south-west) suggest the earlier location of the river (or overflows from the river), which may have flowed to the sea on both sides of the headland.

¹⁵⁵ See n. 49 for comments on finds observed in a subsurface probe placed in the south-eastern corner of Area D.

¹⁵⁶ Similar difficulties in assessing site size were described for the recent geophysical prospection around a emporion (Vetren) in central Bulgaria; this site also located near a river. Archibald 2002, 349.

¹⁵⁷ Pendlebury 1932–3.

¹⁵⁸ Possibly related to this area is a LM III A2 tomb found in 1941 at Katevati and published by Mavroeidis, E., “Εὶδὴσεις ” Επετηρὶς Εταιρεὶας Κρητικῶν Σπουδῶν, 4 (1941), 273–5Google Scholar, cf. Kanta, A., The Late Minoan III Period in Crete. A Survey of Sites, Pottery and their Distribution (SIMA 58; Göteborg, 1980), 140–1Google Scholar fig. 55. 7–8. Its exact location is unknown, but the toponym is applied to an area directly east of Palekmo and the Kambos, south-west of Nisi Pandeleimon. For geophysical prospection of this area, Sarris II, 14–16, figs. 5. 2. 1–5. 2.7 and overall colour map of grids, p. 74. His results are illustrated in fig. 16.

¹⁵⁹ Sarris II, 14–16. In 2005 the feature SOC-6, in a 20_20 m grid, was re-examined using Electrical Resistivity Tomography (ERT), which produces a two-dimensional resistivity image (Syscal of Iris Instruments and a Sting/Advanced-Smart Electrode Module were used). Also used was a soil resistance multiplexer technique (Geoscan Research RM15) employing a Twin-probe configuration of electrodes. This uses three different electrode spacings to increase the depth of the subsurface probe and produces a three-dimensional resistivity distribution of the subsoil. The goal of this geophysical survey was to investigate the application of different electrode configurations in mapping specific buried structures to produce an image in three dimensions (i.e. to maximize and integrate horizontal and vertical data). The parallel tomographies (with 20 electrodes) and multiplexed resistance techniques were run west-east (tomography) and south–north (multiplexer). This combination of techniques allows reconstruction of the feature by taking horizontal images or ‘slices’ at different depths. The results of the techniques were in accord, with the top of the very large structure wall SOC-6 beginning to appear at c.0.35 m and continuing to a subsurface depth of 1.5 m. This depth (in an area where up to 2 m of surface has been removed) indicates a Bronze Age date for this massive building corner. These data were presented in the third unpublished technical report by A. Sarris, ‘Geophysical Prospection Survey at Priniatikos Pyrgos—Istron (E. Crete) (2005)—Phase III’ (Laboratory of Geophysical-Satellite Remote Sensing and Archaeo-Environment; IMS-FORTH; Rethymnon, 2006), 5, 8, 9, 19, 22, 46, 50.

¹⁶⁰ The prospection report was available shortly after the coring and trenching around the football pitch was initiated. The project would like to thank Kostas Zervantonakis of KZ Geotechnical Engineering and Consulting, who provided the expertise, equipment, and personnel for the coring and trenching programme.

¹⁶¹ J. Moody did observe, however, sherds of MM–LM date in the earth cleared by bulldozing for the field in the late 1980s.

¹⁶² Vallianou 1997, 342; Soles 1997. At Gournia houses within the settlement contain evidence of workshop activities (potters' wheels and moulds), although ceramic kilns and furnaces were not identified. Also Shaw 2001, 20 n. 25. Day 1997, 225, mentions the dispersed locations of pottery workshops serving such centres as Palaikastro and Zakros.

¹⁶³ Other Bronze Age sites in or near the coastal zone include Kato Arniko hill (KA1), about 1 km directly south of Priniatikos Pyrgos; the site involves several phases of the Bronze Age (fig. 1). Unfortunately, its size has been obliterated by later settlement and new construction. Another contemporary LM I–III settlement (IS1) was located south of the national highway and south-east of Nisi Pandeleimon, this again overbuilt in the Greek period and slightly over 1 km distant from Priniatikos Pyrgos. These are too distant from the industrial site and perhaps too small to have been the main settlement associated with Priniatikos Pyrgos.

¹⁶⁴ Day, P. M. and Wilson, D. E., ‘Landscapes of memory, craft and power in Prepalatial and Protopalatial Knossos’, in Hamilakis, Y. (ed.), Labyrinth Revisited: Rethinking ‘Minoan’ Archaeology (Oxford, 2002), 157Google Scholar, and Vallianou 1997, 342, who notes that the lack of kilns at palaces (except for Phaistos) means that manufacturing was elsewhere; also n. 160.

¹⁶⁵ Mcrousis, N., ‘Changes in the economic and administrative organization of Crete in the Late Minoan II–III period’, BSA 97 (2002), 167Google Scholar.

¹⁶⁶ See n. 110 and Markoe 1998.

¹⁶⁷ See n. 113.