Acknowledgements. The survey is part of the postgraduate research conducted by Anthi Koutsoukou during her studies at Edinburgh University. Chrysanthos Kanellopoulos, postgraduate student in the Department of Classical Archaeology of Athens University, co-operated in the research. We would like to express our thanks to the Ephorate of the Cyclades and the Ministry of Culture for their permission to carry out the survey. The Baldwin Brown Travelling Scholarship in Classical Archaeology and the Abercromby Fund of Edinburgh University covered the travel expenses of the fieldwork. Finally we are grateful to Dr R.L.N. Barber and Mr P. Valavanis for their valuable comments.

Photographs were taken and printed by Mr N. Raisis.

Chrysanthos Kanellopoulos made all drawings except those in Figures 1, 2, 4, 5 and 9, done by Anthi Koutsoukou.

Abbreviations additional to those in standard use.

Delphoi = Valavanis, P., ‘Archaios Pirgos stous Delphous’, Archaiognosia 1 (1980) 331ff.Google Scholar

Lawrence = Lawrence, A.W., Greek Aims in Fortifications (1979).Google Scholar

Paschalis = Paschalis, D., Andros (1925).Google Scholar

Sounion = Young, J.H., ‘Studies in South Attica: Country Estates at Sounion’, Hesperia 25 (1956) 123ff.CrossRefGoogle Scholar

¹ The assumption that towers had a variety of functions is characteristic in recent research, see particularly Boussac, M.-F. and Rougemont, G. ‘Observations sur le territoire des cités d'Amorgos’, Les Cyclades: Matériaux pour une étude de géographie historique (1983) 117Google Scholar and Osborne, R., ‘Island Towers: the case of Thasos’, BSA 81 (1986) 168–75.Google Scholar

² Pottery and other finds will be fully published in the dissertation of Anthi Koutsoukou.

³ Paschalis, 602.

⁴ Nowicka discusses extensively the cases of houses or complexes with a tower in the countryside, with reference to ancient sources mentioning such cases; Nowicka, M., Les maisons à tour dans le monde grec (1975) 90–139.Google Scholar

⁵ See the anonymous oration against Euergos and Mnesiboulos attributed to Demosthenes, XLVII, 53–57.

⁶ Paschalis, 605.

⁷ Sounion, 134–135. Towers less than 5 m in outer diameter are exceptional.

⁸ At Siphnos tower 10 in Young's catalogue has a buttressed foundation but the height of this external attached wall is not reported; Young, J.H., ‘Ancient Towers on Siphnos’, AJA 60 (1956) 52, fig. 2:10.CrossRefGoogle Scholar Perhaps the external wall in the tower of Ayia Marina is a similar reinforcement.

⁹ The counterweights were probably for an olive press with lever and winding gear, a type dated from the end of the 2nd century AD onwards (Brun, J.-P., “L' Oléiculture Antique en Province-Les huileries du département du Var” Revue Archaeologique de Narbonnaise Suppl. 15 1986 102–4, fig. 59:12Google Scholar). The millstone fragment is identical to those of a large rotary mill pictured in Délos XVII Le mobilier Délien, pl. LII 392, 393 right.

¹⁰ Characteristic of the site in Kalamos is the amount of dressed marbles scattered in the area and widely reused in modern structures. This site cannot be identified as ordinary habitation area and is believed to have served some special purpose.

¹¹ Fiedler, K.G., Reise durch alle Theile des Königreiches Griechenland (1840) 218.Google Scholar Fiedler reports the ruins of massive walls with large ashlar blocks.

¹² Gounaropoulos, K.A., ‘Ekthesis istoriki kai topografiki tou Gauriou’, Pandora 22 (1871) 181.Google Scholar Ruins of ancient buildings are mentioned; apparently Gounaropoulos did not visit the site.

¹³ Sauciuc, T., ‘I Andros’, Sonderschiften des Österreichischen Archäologischen Institutes in Wien VIII (1914) 36.Google Scholar Sauciuc reports ‘ein hellenischer Wartturm’.

¹⁴ Paschalis, 606. Paschalis notes that the tower is ‘probably medieval … but could also be earlier’.

¹⁵ The dismantling of the NE corner and upper courses of the other sides is the result of clandestine activities in recent decades according to local accounts. Some of the stone wedges used for this operation are still in situ between blocks. All blocks lying inside or immediately around the tower were overthrown during this particular incident. The SW corner had been apparently demolished in some earlier destruction.

¹⁶ Clusters of the same marble are on the plateau about 50 m to the SW of the tower and also 30 m to the NE. It appears that marble has been removed from this area, especially from the first location, although there are no signs of systematic quarrying. Only a half-buried rock of blue cipolino in the S part of field bears randomly four traces of wedges. This material has been used for the internal lining of the tower.

¹⁷ Stones of the lining, being of convenient size, have been amply used in the sheepshed in the W side of the field, in terraces and boundary walls of the nearby fields.

¹⁸ Lawrence, pl. 19. See the SW corner of the tower of Ayia Marina.

¹⁹ Wrede, W., Attische Mauern (1933) pls. 69, 70.Google Scholar

²⁰ Bas, M.P. Le, Voyage Archéologique en Grèce et en Asie Mineure (1888) pl. 2.Google Scholar

²¹ The orientation of the entrance to the S or E as a common feature in towers is explained by Young by the need of light for the dark ground floor Sounion, 135–36.

²² Some examples of towers with entrances asymmetrically placed: the Red Tower and Hilltop Tower in Sounion; Sounion 128, fig. 5 and 129, fig. 6 and the Tower of Delphoi; Delphoi, fig. 3.

²³ Young restores towers to 2–2.5 their diameter or side based on observations on towers surviving to or near their full heights, Sounion, 135.

²⁴ The walls of Choreza are among the thicker examples of rectangular towers, supporting together with other examples Young's observation that in towers with diameter or sides up to 9 m, the thickness of walls generally is not directly proportional to the dimensions of the tower; Sounion, 135, n.31. It is more likely that the height of the building is more closely related to the thickness of its walls.

²⁵ This is usually the case in square towers; Sounion, 137.

²⁶ It has been observed that some of the headers (see the header of the upper course in Fig. 7) had a special finish for the inner part of their upper surface. Could this be a treatment intended to receive built-in stairs?

²⁷ Sounion, 137 and Delphoi, 334.

²⁸ Wooden upper floors are implied for most towers by the existence of sockets for horizontal beams; Winter, F.E., Greek Fortifications (1971) 173, pl. 161, 162.Google Scholar

²⁹ Delphoi, 335. No roof-tiles were noticed either in the towers of Ayios Petros and Ayios Georgios (Fig. 1, i and k).

³⁰ Haselberger, L., ‘Dächer Griechischer Wehrtürme’, AM 94 (1979) 110ff.Google Scholar

³¹ Sixty blocks can theoretically be restored to 2–3 courses in this tower.

³² Two blocks, apparently from the internal lining of the tower, have been used in the water source of the village of Chartes, 2 kms away from Choreza.

³³ A small number of quadrangular, rather thin blocks was noticed (sizes ranging between 0.83 × 0.30 × 0.15 m and 0.97 × 0.35 × 0.24 m) more carefully finished with a pick than any of those belonging to the tower: window lintels?

³⁴ In Sounion fig. 139 we see that towers when associated with a courtyard stand in the middle or at one corner of the enclosure with their entrance orientated towards the central area.

³⁵ A similar arrangement of the external space is observed at the Vari house where a veranda is constructed in the S front of the building with a small enclosure. The wall SW of the courtyard of Choreza can be compared with the arrangement of the outer enclosure of the Vari House, where two cross-walls served also as low retaining walls. The morphology of the area at Hellenikon would require similar low terracing. Jones, J.E., Graham, A.J. and Sackett, L.H., ‘An Attic Country House below the cave of Pan at Vari’, BSA 68 (1973) 369–72, fig. 4.Google Scholar

³⁶ One block measuring 1.00 × 0.50 × 0.50 m to be dressed quarry-face requires one man day's labour; Lawrence, 239.

³⁷ Mines of argentiferous lead are mentioned on Cape Selienitis less than 2 kms NW of the tower, in Davies, O.Roman Mines in Europe (1935) 263, n.2.Google Scholar The information is probably untrustworthy since surface investigation of the area has not located anything up to the present.

³⁸ The attribution of masonry styles to specific chrono logical periods is often proved erroneous; Lawrence, 235, 245.

³⁹ Quarry-face as a surface treatment eventually proved to resist attacks of improved siege machines introduced in the end of the 5th and beginning of the 4th century; Lawrence, 239–240, Winter, op.cit. 311–24, Ober, J., ‘Early Artillery Towers: Messenia, Boeotia, Attica, Megarid’, AJA 91 (1987) 569ff.CrossRefGoogle Scholar

⁴⁰ Pouilloux, J.La Forteresse de Rhamnonte (1954) 43–66Google Scholar on the chronology of the circuit and pl. 6, 7.

⁴¹ Chandler, L., ‘The North-West Frontier of Attica’, JHS 46 (1926) 19, fig. 11.CrossRefGoogle Scholar

⁴² W. Wrede, op.cit. 32, 57, pl. 68, 82.

⁴³ Hammond, N.G.L., ‘The main road from Boeotia to the Peloponnese through the Megarid’, BSA 49 (1954) 110Google Scholar on the chronology of tower; pl. XLIXd.

⁴⁴ Haselberger, L., ‘Der Paläopyrgos von Naussa auf Paros’, AA (1978) 348–53, fig. 3–6.Google Scholar

⁴⁵ Haselberger has already noticed the similarity between a group of large defensive towers in the Cyclades and pergamenese defences; ibid. 374.

⁴⁶ Diodorus Siculus, XIX 61 refers to the control of navigation of grain carrying ships, probably sailing through Kaphireus.

⁴⁷ Sounion, 141–143.

⁴⁸ Agricultural activities on Andros would be analogous to those of the Temples Estates associated with Delos where inventories mention four main sources of revenue: livestock, grain, grapes and figs; Kent, J.K., ‘The Temple Estates of Delos, Rheneia and Myconos’, Hesperia 17 (1948) 299–300.CrossRefGoogle Scholar

⁴⁹ Some towers on Siphnos seem to be related to agricultural activities, particularly towers 20, 34; Young, op.cit., 53–55. figs 2:34. 3:20.

⁵⁰ Hammond, op.cit., 110–111, fig. 1.

⁵¹ Lord, L.E., ‘Watchtowers and Fortresses in Argolis’, AJA 48 78ff.Google Scholar

⁵² Delphoi, 336–7.

⁵³ Boussac and Rougemont, op.cit., 117.

⁵⁴ L. Haselberger, op.cit. 365.

⁵⁵ Paschalis reports pottery found in this area (Paschalis, 604) but it is clear that he never visited the site.

⁵⁶ See D.R. Keller Archaeological survey in Southern Euboea 1985 sites No. 6 and No. 119 and discussion 203, 210–211.