Three
The Dema Tower

We turn now to another feature of the Dema defenses, the tower located on the higher of the two hills crossed by the wall. Because of its proximity to the wall, the resemblance of its construction to that of the wall, and the apparent suitability of the tower as a vantage point for a view along the wall and beyond, the Dema tower has been considered to be a contemporary and integral part of the Dema wall defenses by every commentator to describe it.^[1] In what follows, the evidence of excavation and new observations on the site will be presented in support of the same conclusion regarding the contemporaneity of the wall and tower. With regard to its function, as will be seen, this reappraisal substantially changes our understanding of the purpose of the tower, relating it to a system of mountaintop observation posts along the western frontiers of Attica. Its association with the wall, though reinterpreted, is reaffirmed, and as will be shown, excavation has yielded evidence for the date of construction of the tower which substantiates the circumstantial cases made in chapter 2 for dating the Dema defenses within the first half of the fourth century.

Location

The remains of the Dema tower rest on the summit of the highest hill in the Aigaleos-Parnes gap, a hill known locally, after the tower on it, as

Map 3.
The Dema tower and adjacent wall

Pyrgarthi (elevation 225.17 meters).^[2] The tower sits behind and above the line of the Dema wall, which passes some 100 meters away at the closest point. The position of the tower would be an excellent one for commanding a view along the entire length of the wall but for the existence of a knoll between it and the wall, 130 meters southwest of the tower (benchmark at elevation 224.75 meters in map 3). This knoll, being nearly the same elevation as the summit on which the tower sits, completely blocks the view of more than 600 meters of the wall, where it runs lower down the slope beyond the knoll (see figures 20-22).^[3] Both

[3] The photo panorama from the tower in figures 21 and 22 shows that the wall is out of view from the southern saddle (where the railroad and highway cross the wall) to just north of the highest point ,of the wall as it traverses Pyrgarthi. This corresponds to the distance between sally ports 11 and 20 on the map, DEMA plate 29. The lengths of the wall-sections between these points given in DEMA 167 note 25 totals 632 meters.

north and south of this obscured stretch, the remaining length of the wall can be seen from the tower site. The knoll itself would have been the most suitable position for a view along the whole length of the wall if it were essential for the tower to have such a view. The importance of this point will be stressed when the function of the tower is considered.

The long-range views from the tower are impressive (figures 21-24). To the southwest, most of the peaks of Salamis are visible above the bay of Eleusis. Eleusis itself is in clear view, and beyond it (weather permitting) the mountains of the Megarid. Westward, most of the Eleusinian plain is in view, bounded by the Pateras range and its outrunners, with the peak of Kithairon on the horizon. Closer to the tower, the view of the nearer edge of the Eleusinian plain is blocked by the southwestern knoll of Pyrgarthi and by the western spur of Kalistiri. The slopes of Parnes dominate the view to the north, as do the slopes of Aigaleos to the south. Immediately east of the tower, between Aigaleos and Parnes, the whole plain of Ano Liosia can be seen, except for a small section of the foreground blocked by an eastern knoll of Pyrgarthi, now disappearing in quarry work. Menidi/Acharnai lies at the far end of the plain, and beyond it the upper end of the Athenian plain is in view, with Mount Pentele dominating the view beyond.

The Tower Enclosure

The Dema tower is surrounded by a low enclosure wall that traces an elliptical course, with irregularities on its eastern and western ends where the enclosure wall bends to incorporate bedrock outcrops into its line. The tower sits on the highest ground, near the southwestern end of the enclosure, while the largest open area within the enclosure is the gently sloping ground northeast of the tower. The interior of the enclosure, like most of this limestone hill, is rocky and uneven, with few level areas. The enclosure wall is built of roughly laid limestone rubble, with traces of its inner and outer faces preserved for most of its circumference. Only in one area, on the southern side of the enclosure (bounding areas 2S, 3S, and 4S in map 4), is the rubble so sparse that no trace of either face can be followed. The width of the wall varies from 1.20 to 1.80 meters, and it stands today to a height of 0.60 meters at its highest points. It probably was never very much higher than this, since only a

Map 4.
The Dema tower

small number of stones lie loose around it. No trace of an entrance through the wall is to be found.

Inspection before excavation revealed no evidence for ancient subsidiary structures within the enclosure.^[4] The only artifacts to be found were fragments of Lakonian-type roof tiles, scattered throughout the enclosure, but more abundant close to the tower. Excavation within the enclosure was undertaken in an effort to discover datable occupation debris and to ascertain whether or not there were any subsidiary structures associated with the tower. The area immediately south and east of the tower seemed to be the most promising quarter for investigation, since it was generally covered with a layer of rubble that might preserve features or artifacts beneath it, and the ground here was more nearly level than elsewhere in the enclosure (figures 25, 26). The area immediately north and west of the tower had less rubble cover, and here sterile bedrock could be seen over much of the area. The numerous jagged bedrock outcrops in the open area northeast of the tower made this section seem less likely to yield occupation debris. Work was thus directed toward clearing the rubble and excavating in the areas labeled 2S, 3S, and 4S in map 4.^[5]

Stratigraphy

Limestone rubble covered most of the excavation area to a depth of 0.40 to 0.50 meters, occasionally up to I meter against the face of the tower. This scatter of stones was continuous between the tower and the inner face of the enclosure wall in area 2S, while in 3S and 4S, the scatter generally ended 3 meters from the tower. Roof-file fragments were found

[4] Two crudely built circles of rubble, each about 1.50 m across, standing to 0.50 to 0.60 m, located in areas 4N and 5N of map 4 are probably nineteenth- or twentieth-century constructions, rifle pits (tambouria ) in the opinion of local workmen. Similar constructions have been noted on the Dema wall, see DEMA 171. A hollow in the rubble of the tower may have a similar explanation, according to DEMA 173 note 48, but see below, under Later Activity, for an alternative suggestion.

[5] Rubble was also moved away from the face of the tower in areas 2N, 3N, and 4N, and a strip was cleared through rubble in area 2N between the tower and the enclosure wall in order to see if any structural remains could be identified; none were found. Artifacts previously found at the Dema tower include only tiles and a base fragment of a beehive kalathos; see DEMA 186 and note 121. These sherds are now in the collection of the American School of Classical Studies at Athens, site A-19; see catalog nos. 20, 22, below. Excavation was carried out between October 14 and November 21, 1979, and preliminary reports appear in the Unpublished Papers of the American School of Classical Studies at Athens (Munn 1979a), and in the proceedings of the 82nd General Meeting of the Archaeological Institute of America (Munn 1981).

in this rubble as well as below it, amidst smaller stones on the soil or bedrock surface. Fragments of pottery were also found on the soil or bedrock surface immediately below the scatter of rubble.^[6]

Below the rubble, islands of bedrock protruded from the soil, although not so prominently here as elsewhere in the enclosure. The soil around this bedrock, with the exception of dark-gray surface patches formed by the decomposition of plant remains, was uniformly a loose, crumbly red earth, usually mixed with a moderate amount of small stone chips but sometimes free of them. Alongside the tower, and up to two to three meters away from it, this soil frequently contained large concentrations of roof-tile fragments, filling cavities and depressions in the bedrock (figure 27). Roof tiles and earth together formed a layer usually no more than 0.20 meters thick, depending on the contours of the bedrock below. This red earth is certainly the disintegrated debris of sun-dried mud brick which has eroded and washed over the site.^[7] Part of this mud brick, however, had been deliberately laid down, with the concentrations of files as a packing to level the ground in connection with the construction of secondary structures to be described below. Sherds were found in this mud-brick debris, both with the concentrations of tiles and in earth relatively free of tiles. It is significant that sherds associated with the tile concentrations were always found among the uppermost files of those packings.^[8]

The soil below this, wherever bedrock lay deeper down, was a red earth similar to the mud-brick debris but distinct in that it contained many stone chips and pebbles, all somewhat worn and rounded by water, whereas the chips in the upper layer had rougher edges. This lower soil also contained no sherds or tile fragments except at its uppermost

[6] The following artifacts were found on soil or bedrock surfaces under rubble: one fragment of catalog no. 1, no. 4, no. 11, most fragments of no. 15, no. 17, some fragments of the beehive kalathoi listed under no. 20, and some of the roof-tile fragments described under no. 24.

[7] The nature of this soil is the same as that described as decomposed mud brick on other sites; cf., e.g., "Dema House" 77 and "Vail House" 360. Soil and tile fragments were usually found mixed together at the Dema tower, but in some places a layer of clear soil from 0.10 to 0.18 meters deep overlay concentrations of tiles and could have accumulated there only after these files were in place. Since the site is at the summit of a hill, no topsoil could have been washed there from elsewhere. It must have been deposited there by erosion either from the tower or from structures immediately adjacent to the tower and must therefore represent manmade debris (i.e., mud brick).

[8] The following artifacts were found buried in the mud-brick debris: one fragment of catalog no. 1, no. 2, no. 3, no. 7, no. 8, no. 12, no. 14, some fragments of no. 16, no. 18, no. 19, some fragments of beehive kalathoi listed under no. 20, no. 21, no. 23, the majority of the tile fragments described under no. 24, and no. 25. The following artifacts were found in mud-brick debris atop roof-tile packings: no. 5, no. 6, no. 9, no. 10, no. 13, and some fragments of no. 20.

level. It is evident that this sterile soil formed the original ground level before the deposition of the mud-brick and tile debris.

Secondary Walls

As the clearing of the fallen rubble around the tower proceeded, remains of four rubble walls were discovered (Walls 1-4, map 4). The packings of tile fragments respected these walls, coming up to but not underlying them, indicating that these walls were built before the roof-tile fills were laid down.

Walls 2 and 4 abut the tower face without bonding into it (figure 27). Both walls probably once formed right angles with Walls 1 and 3 respectively. The continuation of Wall 4 was indicated in the roof-tile fills lying on either side but not intruding into the area where the stones of the wall had been removed (see stippling in map 4). A dense roof-tile fill on the western side of Wall 2 did not extend as far as Wall 1, so here evidence for a corner is less dear.^[9] Bedrock nowhere showed signs of having been dressed to form a bedding for these walls or to level the uneven surfaces enclosed by them. Equally notable is the absence of any hard-packed earth or stone-chip floor surface. The only indication that care was taken to provide a level surface in the spaces bounded by these walls and the tower is the presence of the loose packings of roof-tile fragments and mud-brick debris.

Some large roof-tile fragments were found standing upright against the foot of Wall 4, on either side, as well as against the foot of the tower just south of Wall 4. These appear to have been deliberately placed to border and retain the mud-brick and roof-tile fill packed up against these rubble walls.^[10] The fact that the tile and mud-brick fills with their crude revetments were the only evidence of leveling for floors associated with Walls 1-4, and even more significant, the fact that a few isolated tile fragments were found built into Walls 3 and 4, wedged among and under the stones of these walls, indicate that the construction of the walls and the laying of the tile and mud-brick leveling fills all belong to the same construction phase. This construction was undertaken when large

[9] After the fragments were cleaned and sorted, special effort was made to join the tile fragments from this tile packing, with little success. Most of the fragments were small, and few joins could be made, indicating that the fragments had been thoroughly mixed before they were deposited here. This finding tends to substantiate the conclusion, drawn from other evidence, that this and other tile packings were deliberate secondary deposits of tile fragments. If this had been merely the chance resting place of tiles fallen from the roof of the tower, joins should have been much easier to find.

[10] Cf. the use of upright roof tiles to protect the foot of a crude wall at Halieis, Jameson 1969, 323; cf. BCH 89 (1966):787 figure 2, and AD 21 (1966): Chronika plate 143.

quantifies of broken roof tiles and disintegrating mud brick were available on the site.^[11] Since the tower itself is the only structure from which this material could have been taken, it is evident that the walls and associated leveling fills are the remains of a construction phase begun after the tower had fallen into ruin and its roof had collapsed.

The Tower

All that remains now of the tower is a solid circular base built of limestone rubble. The base is founded at ground level, on earth and bedrock outcrops, where it has an average diameter of 7.60 meters. The stones on its face are as carefully fitted as unworked rubble can be, while the solid core is more loosely packed with stones, earth, and limestone chips.^[12] The face presently stands an average of 1.36 meters high, and it has a slight inward batter giving the tower an average diameter of 7.40 meters at the top of the preserved face. There is no evidence for a ground-level doorway into the tower, nor is there any trace of interior walls or chambers. The base appears to have been built as a solid platform.

The top of the base has been disturbed. All stones are loose, and a pit over a meter deep has been carelessly dug into the fill of the base for some purpose. As a result of this disturbance, no trace of the original top surface remains, but it is likely that it stood not much higher than its present overall height of 2 meters.^[13]

Finds

Compared to the ubiquitous roof-tile fragments, the quantity of pottery recovered in excavation around the tower was small. Nevertheless, frag-

[11] Whole tiles were too valuable to have been deliberately broken and used in this manner. Note, for example, how tiles are included in the plunder taken from Attica during the Dekeleian War, HellOxy . 12.4 (London), and note the evidence for prices paid for Lakonian-type roof files collected by Orlandos 1955, 109-12, and by Martin 1965, 82-83.

[12] A single block at the foot of the tower face, on the boundary between areas 3N and 4N in map 4, shows signs of having been roughly dressed into a rectangular shape, 1.20 m long by 0.40 m high. It is the only worked block on the site, and as a block in the lowest course of the tower, it may have been one of the first stones laid, after which it was derided that the base could be more economically built using unworked stones.

[13] Based on a rough estimate of the volume of rubble around the tower, the original height of the base can be calculated at approximately 2.45 meters, or a little less assuming that not all of the rubble was originally part of the base. The calculation is probably not far off, since the rubble base of the Hymettos tower, which closely resembles the Dema tower, has a floor surface partially preserved at a height of 2.15 meters (see Munn 1983, 406-10).

ments of at least eighteen vessels were found, including three black-glazed vessels with complete or nearly complete profiles, one complete beehive kalathos lid, and two intact late Roman lamps. In addition to the identifiable vessels, numerous undiagnostic coarse sherds and a few incised roof-tile fragments preserving portions of gameboards were found. Altogether, these finds make up a significant body of evidence for the nature and date of activity on this site.

Sherds on the surface were usually covered by hard encrustations of lime, making joins nearly impossible. All sherds recovered from the stratum of the mud-brick debris had become discolored and soft, apparently through the actions of soil and water. Black-glazed sherds often had only a few traces of glaze preserved. For these reasons, clay descriptions would be misleading and have been omitted from the catalog. It should be noted, however, that, with the possible exception of no. 6, all of the black-glazed sherds are probably Attic.^[14]

Black-Glazed Pottery

1. Cup with high-swung handles Figure A, Figure 29

DP -2S-5. Two nonjoining handle fragments, each with part of body at handle base. L. of handle 0.045 m, D. of handle 0.008 m. Body is thin walled, with no sign of articulation or offset for rim in handle zone. Black glaze in and out, inside of handles reserved.

Probably a skyphos or stemless cup with plain rim. These handles are not as high swung and attenuated as those on most shallow-bowled stemmed and stemless cups. This was probably therefore a cup with a steeper wall, such as the stemless cups, Agora XII, nos. 467, 468, which date between ca. 430 and 400 B.C. Cf. also the skyphos illustrated by Richter and Milne 1935, figure 173, cited in Agora XII as a parallel to nos. 467 and 468.

2. Cup Figure A

DP -2S-2. Three joining fragments giving a quarter of the circumference of the foot, half of the floor, and part of the lower body. Pres. H. 0.022 m, est. D. of foot 0.07 m. Flaring ring foot, thin-walled floor and body. Light spiral grooves on undersurface. Traces of fugitive black glaze in and out, undersurface probably reserved.

Probably a bolsal or one-handler. Cf. Agora XII, nos. 539 (bolsal, ca. 420 B.C. ) and 755 (one-handler, ca. 400 B.C. ); Agora P 27409 (bolsal, from deposit S 16:1, ca. 425-400 B.C. ; see Holloway 1966, 33-84 and plate 28c). The thinness of the fabric in this specimen, as in examples cited, is

Figure A

appropriate in a cup manufactured in the last quarter of the fifth or first quarter of the fourth century B.C. ; cf. the remarks of Corbett 1949, 301-2, and cf. the one-handler no. 74, p. 330 and plate 93.

3. Bowl with incurved rim Figure A, Figure 29

DP -2S-l. Nine joining fragments giving complete foot and one quarter of wall and rim and five nonjoining fragments. H. 0.041 m, est. max. D. 0.09 m, D. of foot 0.058 m. Torus ring foot, concave on interior with offset, light wheelmade facets on exterior. Deep body, wall rises in convex curve becoming gradually sharper to the incurved rim. Lip rounded. Black glaze in and out, slightly mottled in firing.

Cf. Agora XII, nos. 838, 889: third quarter of the fourth century B.C. For slightly more developed (and presumably slightly later) examples of this deep-bodied shape, with more sharply incurved rim, cf. Agora XII, nos. 840-42; "Vail House" nos. 28, 31; Miller 1974, no. 31; Thompson 1934, no. A 20; in this last example, the torus foot has become beveled.

4. Bowl Figure A

DP -3S-7. Two nonjoining fragments of foot and lower body. Pres. H. 0.024 m, est. D. of foot 0.092 m. Slightly flaring convex ring foot, concave on interior, with grooved resting surface. Undersurface swelling toward nipple at center. Black glaze in and out.

Bowl with incurred rim, type similar to no. 3. For bowls with similar foot, cf. Agora XII, nos. 830, 832, 841, all middle to second half of the fourth century

5. Small bowl Figure A, Figure 29

DP -4S-9. Two joining fragments giving complete foot and small portion of wall to rim. H. 0.024 m, est. max. D. 0.08 m, D. of foot 0.054 m. Broad ring foot, convex on exterior, concave on interior. Undersurface has central nipple. Shallow body, curve of wall turning abruptly inward just below rounded lip. Black glaze in and out, resting surface reserved. Incised graffito on undersurface: D or L , or possibly A.

Cf. Agora XII, no. 887: 350-325 B.C. This shape is commonly found with little variation from ca. 375 into the early third century B.C. ; see Agora XII, nos. 883-84, 886-89; Rotroff 1983, no. 5; Corbett 1949, no. 155; Thompson 1934, no. A 18. Examples dated by context earlier in the series usually have a reserved resting surface, as does this specimen; Rotroff 1983, 265, places examples with reserved resting surfaces before ca. 310 B.C.

6. Squat aryballos Figure A, Figure 29

DP -4S-8. Eleven joining and four nonjoining fragments of body to base of neck. Pres. H. 0.06 m, max. D. 0.086 m. Flat base, squat body beveled 0.006 m above base. Six-toothed comb used to cover body from neck to maximum diameter with haphazard vertical ribbing. Dipped in thin black glaze, drip line on beveled face leaving base reserved; interior glazed.

This type of ribbed aryballos imitates the shape and incised ribbing of Corinthian blisterware aryballoi. Examples with closely spaced "linear" ribbing as in this specimen are known at Corinth in true and imitation blisterware fabric, dated to the second and third quarters of the fourth century B.C. ; see Corinth VII.iii, 147-48 and note 17. Cf. the two examples, apparently true blisterware, of the second half of the fourth century published by Broneer 1962, 24-25, nos. 20 and 21, with plate 12 f, and p. 6 on the date; cf. also Agora XII, no. 1681: second half of the fourth century B.C. by context. An imitation blisterware example from Athens is published by Rotroff 1983, 289, no. 45, dated late fourth to early third century B.C. ; it is somewhat more squat and heavy and presumably, therefore, somewhat later than this specimen.

7. Skyphos Figure A

DP -2S-6. Fragment of rim. Pres. H. 0.025 m. Convex upper wall with slightly outturned rim. Lip rounded. Black glaze in and out.

Compare the rim profiles in Agora XII, nos. 350-54: second through last quarter of the fourth century B.C. The latest of these types remained in use in the first quarter of the third century; see Rotroff 1984, 347.

8. Kantharos Figure A

DP -4S-10. Two joining fragments of rim. Pres. H. 0.033 m, est. D. of rim 0.08 m. Vertical wall flaring to rounded lip. Black glaze in and out.

Type is either a cup-kantharos or spur-handled kantharos with plain rim, which range in date from the second quarter of the fourth to well into the third century B.C. ; see the discussion in Agora XII, 119-20, 122.

Ten nonjoining black-glazed sherds are not identifiable.

Lamps

9. Roman lamp Figure 29

DP -4S-21. Five joining fragments giving less than half of lower body. Max. pres. L. 0.06 m. Moldmade lamp with deep body, thin walls, ovoid shape in horizontal section.

Shape is probably that of Corinth IV.ii, type xxviii, 113-14, dating from the middle of the third to the early fifth century A.D. , but too little is preserved to date this specimen with assurance.

10. Late Roman lamp Figure 29

DP -2S-19. Complete, worn. H. without handle 0.03 m, L. 0.071 m, W. 0.052 m. Moldmade lamp, plain flat base, deep lower body, concave disc with central filling hole, wick hole in narrow end of ovoid body, rounded vertical handle, H. 0.011 m, at back of body. No decoration preserved.

For shape cf. Agora VII, nos. 2440, 2796, 2806, 2807: late fifth to sixth century A.D.

Figure B

11. Late Roman lamp Figure 29

DP -3S-20. Joining upper and lower halves giving complete lamp, worn. H. without handle 0.037 m, L. 0.075 m, W. 0.056 m, handle H, 0.011 m. Moldmade lamp as no. 10 above.

Coarse Wares

12. Water pitcher Figure A

DP -2S-11. Two nonjoining fragments of rim. Pres. H. 0.02 m, est. D. of rim ca. 0.11 m. Outturned, thickened rim, rounded on top, concave below, possibly broken at point where rim springs from a raised ridge around top of neck.

For the shape, cf. "Vail House" nos. 67 and 68, figure 8, p. 382; and Corinth VII.iii, no. 631, plate 24. Evidently of non-Attic and non-Corinthian origin, this type of water pitcher became extremely common from the last third of the fourth century B.C. onward; see the discussions by Thompson 1934, 465, and Corinth VII.iii, 113, with the foreign examples from contexts of ca. 350-250 B.C. cited in note 14. At the Vari house it occurs in an assemblage dated ca. 350-275 B.C. For examples of the later fourth and early third centuries, see Miller 1974, nos. 46, 47, with plate 33; numerous Hellenistic examples are illustrated by Braun 1970, plate 82.2, 3.

13. Water pitcher Figure A

DP -4S-12. Two joining fragments of base. Pres. H. 0.012 m, est. D. of base 0.077 m. Flat base thickened to form offset at bottom of wall.

Probably a smaller pitcher similar to no. 12 above.

14. Transport/storage amphora Figure B, Figure 30

DP -2S-3. Fifteen joining fragments giving toe and lower body, at least seventeen nonjoining fragments from body as high as the shoulder. Pres. H.

0.265 m, est. max. pres. D. 0.368 m. Knob toe with shallow depression in undersurface, rounded flange around knob, D. 0.09 m, narrowing above to junction with swelling convex body.

A similar toe comes from the Athenian Agora deposit A 17: 3, which is dated ca. 320-290 B.C. , Agora XII, 383; cf. P 20472, from deposit D 16:1, dated to the fourth century B.C. , Agora XII, 387. A complete amphora with a roughly similar toe was found at Corinth in fill of the late fourth to early third century B.C. ; see Robinson 1969, 10, no. 4, with plate 2 no. 4.

15. Transport/storage amphora Figure B, Figure 30

DP -2S-13, DP -3S-15. Large fragment of toe and lower body, many fragments from body as high as the shoulder. Pres. H. 0.24 m, est. max. pres. D. 0.28 m. Knob toe with deep depression in undersurface, rounded flange around knob, D. 0.072 m, concave above in junction with narrow body, gradually flaring to convex profile.

Cf. McCredie 1966, 24 no. 12, and no. 25 in plates 4, 20e: late fourth century B.C. Also similar are Athenian Agora nos. P 20509, from the blind passage of Group B, Thompson 1934, 330-32, which is deposit H 16:3, containing much material of the late fourth century B.C. ; see Agora XII, 393; P 20431, fourth century B.C. , from the NW room of the Poros Building; P 25945, from deposit F 17:3—POU (1): second half of the fourth century B.C. , Agora XII, 390.

16. Transport/storage amphora

DP -4S-16. Eight joining fragments of lower body, toe not preserved, other nonjoining body fragments. Pres. H. 0.22 m, est. max. pres. D. 0.28 m. Lower body similar in shape to no. 15 above.

Numerous coarse body fragments recovered throughout the excavated area are likely to be amphora sherds, although some may be uncombed coarse kalathos sherds (see no. 20 below). Two amphora handle fragments were found, but no neck or rim fragments were identified.

17. Basin

DP -3N-27. Body fragment. Single fragment of a large, thick-walled vessel, probably a large basin. Horizontal relief band on exterior, with attachment point for a horizontal handle immediately below relief band.

Possibly a beehive. Cf. the "Orestada" vessel, a basin with horizontal handles from the Rachi site near the Isthmus of Corinth, dated ca. 360-240 B.C. , Broneer 1958, no. 42, with plate 14b; for its identification as a beehive, see Kardara 1961, 264-65, with plate 81 figure 6; cf. also "Vari House" 399, with plate 78c. The identification of the "Orestada" vessel as a beehive has now been questioned; see Crane and Graham 1985, 160-61. Regardless of the original purpose of basins like this specimen or the "Orestada" vessel, it is possible that this specimen was a basin used as a makeshift beehive; cf. note 20 below.

18. Beehive kalathos lid Figure 31

DP -2S-4. Twenty-seven fragments giving complete lid, chips missing. D. varies from 0.40 to 0.405 m. Lid flat on inner side, outer side has two concentric relief bands, D. of outer band 0.285 m, D. of inner band 0.135 m, around a central boss. Crescent-shaped indentation in edge of lid, 0.04 m across, 0.015 m deep. Four holes piercing lid are set in pairs alongside the outer relief band; pairs are opposite each other, in line with the indentation on the edge, each pair of holes 0.08 m apart. Raised lug, H. 0.016 m, extends along the outer relief band between the two holes nearest the identation; outer relief band on side opposite lug is raised slightly to form a second lug.

The association of lids of this type with combed kalathoi, as no. 20 below, and their interpretation as ceramic beehives have been established in the publication of the Vail house in Attica; see "Vail House" 397-414, 443-52. Cf. also Agora XII, 217-18. Beehive lids and combed kalathos fragments are illustrated in "Vari House" figures 13, 18-21, and plates 75-77, 83-86; known parallels are listed in note 21, p. 398; on the form and function of the lids, see 409, 446. The examples from the Vari house are dated by their context to between the third quarter of the fourth century and the first quarter of the third, pp. 414-18. There seems to be little chronologically significant variation in details of shape and size in beehive kalathoi and their lids between the late fifth century B.C. and the Roman period.

19. Beehive kalathos lid Figure 30

DP -2S-22. Single fragment, broken all around. Max. pres. L. 0.143 m. Lid without concentric relief bands. Deeply impressed epsilon in what is probably the center of the upper side, H. of letter 0.053 m, L. of upper and lower crossbars 0.045 m, center crossbar shorter. Trace of hole in lid at broken edge farthest from epsilon.

The identification of this fragment as a beehive kalathos lid with an impressed epsilon in the center is assured by comparison with the larger fragment of such a lid in the collection of the American School of Classical Studies at Athens (provenance unknown), illustrated in figure 30. This fragment of nearly half a lid, est. D. 0.38 m, has a crescent-shaped indentation in its outer edge, 0.027 m deep, adjacent to which is a lug or ledge, H. 0.008 m, L. 0.075 m, tangent to two holes piercing the lid. At the center of the lid is an epsilon, partially broken away, impressed while the day was wet with a blunt tool, not a stamp. L. of lower crossbar 0.048 m, center crossbar shorter. The clay of this lid is buff (5YR 7/6 reddish yellow) with red inclusions.

20. Internally combed beehive kalathoi Figure 31

DP -2S+3S-23, DP -4S-24. Twenty-five fragments, mostly nonjoining, of bases and bodies, no rims. Average est. D. of bases 0.19 m. Base and body fragments completely or partially covered on interior surfaces with vertical combing.

Figure C

For identification of these as ceramic beehive fragments, see no. 18 above. One base fragment is in the collection of the American School of Classical Studies at Athens, a surface find from the Dema tower, site A-19.

Miscellaneous

21. Incised gameboard on roof tile Figure C, Figure 32

DP -2S-25. Three joining fragments, broken all around, of Lakonian-type pan tile with streaky black glaze on concave side, as no. 24 below. Max. pres. L. 0.142 m, max. pres. W. 0.075 m. Square gameboard, almost half preserved, incised after firing on the unglazed convex side, oriented at a 45º angle to original sides of tile, roughly parallel to the top and right-

hand broken edges of the fragment. Gameboard is formed by three concentric squares with sides bisected by perpendicular lines. L. of side of outer square 0.10 m, L. of intermediate square 0.08 m, est. L. of inner square 0.06 m. On glazed side of tile are two lines meeting at right angle, incised after firing; these are oriented parallel to the original sides of the tile, as indicated by the direction of the streaks of glaze and the curve of the tile as it approaches the lateral edge.

The game represented is Nine Men's Morris, a two-player game, also known as Mühle, or the Mill, and as Morelles, or La Merelle. On the play of the game and its wide popularity in antiquity, see Bell 1960, 93-95; cf. also Baran 1974, 21-23. The closest example known to me of this game in time and space to those of the Dema tower occurs at Gordion, incised on the underside of a reused block built into the foundation of the paved court for the Persian gate of the sixth century B.C. , published by Young 1955, 12, and figure 25 in plate 6.

22. Incised gameboards on roof tile Figure C, Figure 33

ASCS A-19. Single fragment, broken all around, of Lakonian-type pan tile, as no. 24 below. Max. pres. L. 0.125 m, max. pres. W. 0.091 m. Glaze not preserved. Lines deeply incised after firing on concave side to form gameboard, about half preserved. Gameboard is similar to no. 21 above, but design is less carefully executed and the configuration is rectangular rather than square; diagonal lines are added to connect the adjacent corners of the concentric rectangles. On the convex side are one or more attempts to outline a similar gameboard, without the diagonal lines, more lightly and more carelessly incised. This sherd is in the collection of the American School of Classical Studies at Athens, a surface find from the Dema tower, site A-19.

See no. 21 above. The variant of Nine Men's Morris with diagonal lines at the corners is also known today; see Gibson 1970, 32-33.

23. Incised roof tile Figure C, Figure 32

DP -4S-26. Single fragment, broken all around. Max. pres. L. 0.072 m. Fragment of a black-glazed Lakonian-type roof tile, as no. 24 below, with hatched lines incised after firing on glazed side.

Possibly a tally; cf. Talcott 1935, 516 figure 28c, and Lang 1956, nos. 2, 3, 63, and 84.

24. Lakonian-type roof tiles Figure D

Thousands of fragments of red- and black-glazed Lakonian-type roof tiles were recovered. There were no complete tiles, and none were restored to their full lengths. The most complete restored example of a pan tile has a width of 0.50 m, a max. pres. length of 0.735 m, and a max. thickness of 0.016 m. The best-preserved example of a cover tile has a width of 0.23 m behind the thickened lower rim, a max. pres. length of 0.28 m, a max. thickness of the lower rim of 0.034 m, and an average thickness of

Figure D

0.015 m. Thickness of fabric and details of rim profiles are variable, but the figures cited here and the pieces illustrated in figure D are representative of the whole lot, with the exception of the unusual, asymmetrical curve of the rim of the illustrated pan tile; typical examples have a symmetrical profile following the curve of the right rim of this tile.

These are typical examples of Lakonian-type roof tiles, the dimensions and profiles of which are fairly uniform through the classical, Hellenistic, and early Roman periods. Attention to details of glaze, fabric, and profile might allow more precise dating of roof tiles of this type. DEMA 185 n. 120 cites the opinion of Mrs. Carl Roebuck that the files of the type found at the Dema tower are datable to the fourth century B.C. See the examples cited by Orlandos 1955, 103; Martin 1965, 68-70; Olynthus VIII, 232, and figure 17, A and B; Stevens 1950, 174-88; "Dema House" 84-85, with notes 9-11. On the basis of the measurements cited in these works, the restored length of both cover and pan files from the Dema tower should be an average of 0.95 m, and it is probable that the cover files had one end, the upper, narrower than the other, tapering from 0.23 m to about 0.19 m. There are in fact cover tile fragments with both lighter rims and sharper curves than that of the illustrated example, and these would have been the upper rims of the cover tiles. Pan tiles may also have had a taper from 0.50 to about 0.45 m in width, but no measurements could be made from the joined fragments to confirm this possibility.

25. Twisted lead strip Figure 33

Two fragments of a lead strip, pointed at one end, twisted into a coil. The strip bears no trace of an inscription. Edge of the strip appears to have been cut by a sharp instrument. Apparently a piece of waste material that had been trimmed off and discarded.

Discussion

Two phases of building and occupation on the site of the Dema tower are distinguishable on the basis of the structural remains: the phase of the tower and the phase of the secondary structures formed by Walls 1-4 adjacent to the tower. The contexts of the roof-tile fragments and of the mud-brick fill, debris from the collapse of the tower reused in the secondary structures, make it certain that these were consecutive phases.

Excavation was undertaken with the hope of discovering closely datable occupation debris from the original use of the tower, and in this respect, the results were initially disappointing. Because of the extensive secondary activity on the site, no undisturbed contexts of first-phase material could be found. The walls and tile packings of the second phase might have covered debris left from the first phase, but no pottery sherds were found buried within or under the tile packings and walls, so there is no pottery that can be associated by context with the tower phase of the site.^[15]

Find contexts do establish a clear connection between many of the pottery sherds and the phase of the secondary structures. Sherds found on top of the dense roof-tile fills adjacent to Walls 1-4 could possibly be first-phase debris that, by chance, was left on top of the files after they were laid down, but more likely they were deposited there only after those fills were in place. This likelihood becomes a virtual certainty in the case of a vessel with many fragments, all found in the same spot on top of a tile packing. The building of the secondary structures, and especially the laying of the tile and mud-brick leveling fills associated with those structures, involved considerable displacement of earlier debris. It is highly improbable that many fragments of any vessel left over from the original occupation of the tower would have remained together in one place on top of the tile packings after the secondary structures were

[15] Excavation in areas 2S, 3S, and 4S involved the removal of all debris—rubble, mud-brick debris, and roof-tile fragments—to bedrock or to sterile soil. Portions of Walls I and 3 were dismantled to see if any sherds or tiles underlay them; a few tile fragments were found in Wall 3, but no pottery sherds were found. Wall 4 was carefully cleaned and examined without dismantling; here too, a few tile fragments were found in the interstices of stones, but no pottery sherds.

built. The same reasons make it unlikely that a substantial number of fragments of an older vessel would have remained undisturbed in one place in the immediate vicinity of the secondary structures during the building and use of those structures. There is no evidence for extensive clearing or construction on the tower site after the abandonment of the secondary structures, so it is reasonable to expect that debris left when the secondary structures were abandoned should be relatively undisturbed. These considerations make it possible to identify the majority of the pottery sherds as debris from the secondary occupation of the tower site.

All fragments of nos. 5 and 6 (black-glazed wares), 9 and 10 (Roman lamps), 13 (water pitcher), and many of the beehive kalathos fragments of no. 20 were found on top of tile packings. Disregarding the Roman lamps as much later material, nos. 5 and 6 both date to the middle or second half of the fourth century, which is possibly also the date of the chronologically less diagnostic coarse sherds, nos. 13 and 20. The aryballos no. 6 is almost three-quarters complete as restored, and all fifteen of its fragments were found together in one place, on top of the tile fill south of Wall 4 close to the tower. As noted above, the find context of the many sherds of no. 6 indicates that it came to rest after the tile fill was laid down. Indeed, this vessel is almost certainly debris left from the period of use of the secondary structures. The two fragments of the black-glazed bowl no. 5, of the same date and from a similar find context as no. 6, support this conclusion, and it is borne out by the analysis of other finds.

The many sherds of three, or possibly four, different artifacts found together as a group provide a second significant context, evidently an abandonment deposit. All identifiable sherds belonging to the black-glazed bowl no. 3, the amphora no. 14, and the beehive lid no. 18 were found in immediate contact with each other, buried within mud-brick debris atop bedrock and under rubble in a small area on the southern side of Wall I (figure 28). The three sherds of the black-glazed cup no. 2 were found very close to, but not directly contiguous with, this deposit. The most remarkable artifact in this group is the complete lid no. 18, all twenty-seven fragments of which were found in place, where the lid had been smashed and left undisturbed until its discovery. The beehive lid is not closely datable, but parallels from datable contexts (such as the Vari house; see discussion under catalog no. 18) demonstrate that this type of lid is very much at home in the period established by the latest closely datable vessel from this deposit. This is the black-glazed bowl no. 3, the shape of which, restored from fourteen fragments, shows that it is a type characteristic of the second half of the fourth century. This date is con-sonant with the parallels for the amphora no. 14, of which at least thirty-

two fragments of the body and toe were found. The black-glazed cup no. 2 dates at least a half-century earlier than no. 3. Because of its date and the smaller number of fragments and proportion of the vessel pre-served, this specimen may well be first-phase debris left by chance in the vicinity of the deposit of later material.

This deposit, dumped here in the second half of the fourth century, confirms the conclusions to be drawn from the material found on top of roof-file fills, namely, that pottery of the second half of the fourth century is associated with beehive fragments and is found in contexts no earlier than the phase of the secondary structures. Indeed, this comparatively abundant material can be identified as debris from the abandonment of the secondary structures.

The rest of the sherds from the tower site were found scattered individually, either in rubble on the surface of the ground or buried in loose mud-brick or tile debris that was not part of a leveling fill associated with the secondary structures. Find contexts therefore have no bearing on the association of these sherds with either structural phase of the site, but comparison of the dates and types of these sherds to those from significant contexts allows most of the remaining pottery to be associated with the period of use of the secondary structures. The rest of the beehive kalathos sherds listed under no. 20, the lid fragment no. 19, and possibly the basin no. 17 can be placed in the group of artifacts associated with the secondary structures. The black-glazed sherds nos. 4, 7, and 8, though not so closely datable as nos. 3, 5, and 6, can easily be placed in the second half of the fourth century and are therefore most likely also part of this group. Likewise, the water pitcher no. 12 and the amphoras nos. 15 and 16, though even less closely datable by themselves, can also be associated with this group through the parallel vessel types of nos. 18 and 14.

There are, however, finds of uncertain association with this large group and other finds that stand apart from it. The lead strip no. 25 is neither datable nor from a context that would dearly associate it with the secondary structures. It may be debris from this phase of the site, but it need not be. The incised roof-file fragments, nos. 21-23, were found in ambiguous contexts, on the surface and in loose tile and mud-brick debris. As roof files, they are certainly material left from the first phase of the site, but they may have been incised at any later time. Arguments will be presented below for identifying these incised files as reused construction debris from the first phase of the site. Only nos. 1, 2, 9, 10, and 11 unambiguously stand apart from the material associated with the second phase of the site by reason of their dates. Nos. 1 and 2 are black-glazed vessels distinctly earlier than the pottery associated with the second-phase structures and are therefore probably debris from the

original period of use of the tower. Nos. 9, 10, and 11, the Roman lamps, are evidence of activity on this site long after the abandonment of the secondary structures.

Altogether, therefore, excavation has yielded evidence for at least three phases of activity on the Dema tower site: the phase of the tower, in which the roof files were originally employed and with which the incised roof tiles and the cups nos. I and 2 are most likely to be associated; the phase of the secondary structures, associated with the reuse of the roof files as leveling fills and with the majority of the pottery; and later activity that accounts for the Roman lamps. Before considering the original form, function, and date of the Dema tower, it will be useful to interpret the evidence for subsequent activity on the site, since that secondary activity provides a terminus ante quem for the date of the original tower.

The Secondary Structures

Throughout the excavated area, fragments of internally combed beehive kalathoi were found, both body and base fragments, described under catalog no. 20, as well as one fragmentary but complete beehive lid, no. 18, a fragment of a second lid, no. 19, and a fragment of a basin, no. 17, a vessel that might also have been used as a beehive. Beehive kalathoi were probably laid horizontally in stacks, either enclosed in a frame of some sort or built into a wall, to provide them with shade and insulation.^[16] Given the prevalence of beehive fragments on the site, it is certain that at least one purpose of the secondary structures was to house stands of beehives.

The hills around the Dema today still abound in wild thyme, eminently suitable grazing for honeybees, and before the arrival of heavy industry and the city dump, this area was noted for beekeeping.^[17] It

[16] See "Vail House" 402-3, 410, 412, 443, 445, and 448 note 246, with plates 79b, c, d, 80a, and 85. Cf. Columella De Re Rustics 9.7.2-6, quoted by Crane and Graham 1985, 38-39.

[17] See Rossiter 1981, 187: "The uninhabited valley between the Aigaleos ridge and the foothills of Parnes is dotted with beehives." This must have been written in the 1960s, for by 1977 beehives were no longer in evidence in the area, which was dominated by the city dump. See also Gell 1819, 25, on the route from the Rheitoi northward toward the Dema, where Gell remarks at what is now known as the Thriasian Lager: "Proceeding up the little valley, see I. on the top of the hill a circular inclosure [sic ] of stone, called Giaverdeli. It is only an enclosure for bees." By the brevity of these remarks, it seems that Cell did not inspect the enclosure himself, and the last comment sounds like the explanation of a local guide, who knew that the spot was deserted except for the beehives that were kept there. The Lager was evidently used as an apiary in antiquity as well, for beehive kalathoi have been found there; see McCredie 1966, 70, on "umbrella stands" (i.e., beehive kalathoi) from the site. At least nine ancient beehive sherds from the Thriasian Lager are in the collection of the American School of Classical Studies at Athens, site A-74.

must have been so in antiquity as well. The suitability of the area and the availability of building material on the site of the abandoned and at least partially collapsed Dema tower evidently prompted some beekeeper to bring his hives to this spot to construct shelters for them.^[18] The solid rubble base of the tower provided a wall to build against and a break against the strong north wind. Stones were on hand for wall building, available either from the base itself or perhaps from the enclosure wall nearby.^[19] Tile fragments in abundance were available to fill the irregularities in the bedrock, as well as mud brick, which might also have been used as packing around the hives as they were stacked. Very likely, enough sizable tile fragments could be salvaged to form a crude roof over the hives.

The remains of the secondary structures are too scant to show if they might have had any function other than sheltering beehives. Given the suitability of the site for this purpose, it is quite likely that beekeeping was the only reason for the reoccupation of the Dema tower site. When we consult the finds, it is noteworthy that there are no fragments of vessels associated with food preparation—for example, cooking vessels, mortars, large lekanai—which suggests that this was not primarily a habitation site. We might speculate that the several drinking cups, small bowls, water pitchers, and juglet were the accumulated discards from the daytime visits of the beekeeper. The amphoras may have been stor-

[18] To judge by the number of beehive fragments that can be found at remote sites in Attica, it would seem that this sort of reuse of abandoned sites was fairly common. I have seen beehive fragments at the Aigaleos tower and on the summit of Beletsi. McCredie 1966, notes fragments of beehive kalathoi ("umbrella stands") at Kynosoura (46 no. 5), at Helioupolis (48), at Yerovouno (62), and at the Thriasian Lager (70). Vanderpool et al. publish an example from Koroni (1962, no. 46). Varoucha-Christodoulopoulou 1953-54, 335 figure 6, provides examples from Koroni and Helioupolis. In the collection of the American School of Classical Studies at Athens, there are also beehive fragments from Kastraki (A-20), the Hymettos tower (A-34), and Patroklos Island (A-59). The occurrence of beehive fragments at the sites of so many temporary military encampments has prompted speculation that these might have been used as water containers; see "Vail House" 402 and notes 44, 45. This is possible, though it is hard to believe that a used beehive would be a very desirable container for drinking water. It seems more likely that most, if not all, of these beehive fragments represent the reuse of abandoned sites as apiaries, as is evident in the case of the Dema tower.

[19] It is noteworthy that, while both faces of the enclosure wall can be traced every-where else, for 14 meters where the wall comes closest to the secondary structures, its remains are scant and neither face can be traced (see above, under Tower Enclosure, and map 4). It is possible that here, where the wall lay close at hand, it was used as a quarry for building material for the secondary structures.

age vessels kept on the site, but it seems likely that they too were used as makeshift beehives.^[20]

As to the date of this secondary activity, it has already been noted that the finds associated with this phase can be dated to the second half of the fourth century. The better-preserved specimens, nos. 3, 5, and 6, in fact find their closest parallels in examples dated by Agora XII and Corinth VII.iii to the third quarter of the fourth century. It is now recognized that some of these examples regarded as characteristic of the third quarter were in use as late as the end of the fourth century, but none of them, and therefore none of the specimens from the Dema tower, needs to be dated later than ca. 300.^[21] The water pitcher no. 12 may be the latest artifact of this group. It is a Hellenistic type, the earliest appearance of which is not clearly established, though it seems to become common within the last third of the fourth century. Considering that this specimen may make a date close to midcentury unlikely, we can, with reasonable probability, place the secondary activity on the tower site somewhere within the period ca. 340-300.

[20] Many fragments could be attributed to each of the three amphoras identified (nos. 14-16), but none of these fragments came from higher than the shoulder—no neck fragments, no rims, and only two small handle fragments were found. This pattern indicates that the tops of these amphoras had probably been removed before they were discarded on this site. The most obvious explanation is that these several amphora bodies were reused as beehives. The fact that all fragments of no. 14 were found resting atop the beehive lid no. 18 reinforces this conclusion. Regular kalathos lids could have been tied to such makeshift beehives, though makeshift lids of wickerwork or wood could also have been used; examples of such lids in modern use are cited in "Vari House" 445, 448 note 246.

[21] Several recent studies by Susan Rotroff have demonstrated the consequences for early Hellenistic Athenian pottery chronology of new evidence, i.e., the finds from Koroni, the revision of the chronology of stamped Rhodian amphora handles, and refinements in Athenian numismatic chronology. Revised dates for the deposits of Athenian Hellenistic pottery published by Thompson (1934) are presented by Rotroff in Agora XXII, 107-10, and in Thompson et al. 1987, 6; reappraisals of several other late-fourth-century and Hellenistic deposits are contained in Rotroff 1983, Rotroff 1984, and Thompson et al. 1987, 183-97. The closest parallels for the most diagnostic pieces from the Dema tower, nos. 3, 5, and 6, come from Agora deposits F11:2, F16:1, and O18:2. Rotroff 1984, 344-46, now places the deposition of F11:2 in 294 B.C. ; Rotroff 1983, 263, places the deposition of both F16: l and O18:2 close to 300 B.C. (cf. Thompson et al. 1987, 184-85). In all three cases, the deposits contain pottery spanning the second half of the fourth century, and nos. 3, 5, and 6 are best paralleled by shapes that are not among the most developed, or latest, forms of vessels from these deposits. The implication that the Dema tower pottery is thus to be dated not later than ca. 300 B.C. is further borne out by the comparison of the Dema specimens with the later forms of those vessels found in deposits closed well within the first quarter of the third century, e.g., Thompson's group A (1934, 313-30; see Agora XXII, 107-8, and Thompson et al. 1987, 6), and Menon's Cistern, published by Miller 1974 and now dated by Rotroff (1983, 262, and 1984, 346-47) to ca. 286 B.C. The appropriate comparisons are cited in the catalog entries.

Later Activity

The Roman lamps, nos. 9-11, provide the only datable evidence of activity on the Dema tower site after the abandonment of the secondary structures. Nos. 10 and 11, both the same type, date to the late fifth or sixth century A.D. No. 9 seems to be earlier, possibly by as much as two centuries. The absence of other identifiable Roman pottery or any traces of building activity suggests that the lamps were left on the site by occasional visitors who did not occupy the site for any prolonged period of time. It seems likely that these were votive lamps, left at the ruins of this hilltop tower which, by the Roman period, must have looked much as it does today, that is, a circular stone heap conspicuously placed on a hill-top. The remains might have been taken for an ancient altar, a tumulus, or some such venerable relic.^[22] Roman lamps have been found at other mountaintop sites in Attica where there are classical remains, so it is not surprising that they should appear in this context as well.^[23]

The appearance of the ruined Dema tower in later times might well have led people to believe that it was a tumulus over a grave, suggesting that treasure of some sort lay buried within. If so, this might explain why the top of the rubble base of the tower is so thoroughly ruined and why a hollow has been dug out of its rubble fill. The deliberate disturbance of other mountaintop tower sites demonstrates that digging of this sort did take place.^[24] This digging probably occurred after the lamps had

[22] Note the remark of Pausanias 5.13.8 on the simple altars of Attica. On mountaintop altars in Attica, see Pausanias 1.32.2 and Langdon 1976, 1-2, on the rubble altar of Zeus, and 98-106 on literary and archaeological evidence for other mountaintop altars in Attica.

[23] See Langdon 1976, 73-74, nos. 337, 339, 340, 342-49, and p. 76: "Fragments of about 120 lamps indicate some sort of activity in the late 4th and early 5th centuries after Christ." Langdon, 100, refers to Roman lamps from the Zeus sanctuary on Parnes, and 106, to late Roman lamps from a possible Christian shrine on Mount Kerata (see also Ober 1987a, 224-25). In the collection of the American School of Classical Studies at Athens, there is a fragment of a Roman lamp from the Hymettos tower (A-34, see also Ober 1987a, 201). I have seen a late Roman lamp fragment among the remains on the peak of Beletsi (see also Ober 1987a, 204). Smith and Lowry 1954 report fragments of Roman lamps from the summits of Pani (24) and Attic Olympos (32).

[24] The remains of the Hymettos tower show dear evidence of exploratory excavation, and the ruinous condition of the Aigaleos tower is most likely due to digging as well (see Munn 1983, 403 and 408). Note that the remains of the Hymettos tower, which closely resemble those of the Dema tower (figure 34), are referred to as a tumulus by Scully 1979, 27-28, 201. Digging for buried treasure in such ruins could have taken place at any time. The Greek workmen, local residents whom I hired for the excavation of the Dema tower, were clearly eager to dig into the tower itself to find treasure there. The lure of buried treasure must have incited people to dig and explore in antiquity as well. Note the story of Timon's gold, which was said to be hidden in his tower, which was also his tomb, Lucian Timon 42. The story of Timon evidently had wide currency in Attica through the Roman period; see Plutarch Ant. 70; Pausanias 1.30.4; Olympiodoros Vita Platonis p. xlvi (Bekker).

been left on the site, since all of the lamps were found buried beneath a thick layer of rubble that must have been thrown from the tower.

First Phase: The Tower

As it was originally constructed, the Dema tower stood by itself on the hilltop, surrounded by the enclosure wall. There is no evidence that any other structure adjoined the tower or stood elsewhere within the enclosure. The rubble enclosure wall was probably not very high and cannot in any event have been designed as a defensive perimeter, for it is too insubstantial and encloses only a small area with no significant natural strength to recommend it as a defensive position. The wall probably served simply to define the precinct of the tower, the area within which men detailed to the tower were bivouacked. The enclosure wall probably never stood to a height of much more than 0.60 meters in stone. It probably had a superstructure, more likely a brushwood charax than a mud-brick wall, which would suffice to keep grazing animals out of the bivouac area.

The mud-brick and roof-file debris found around the base of the tower originally came from the walls and roof of the tower. The erection of mud-brick walls atop a solid rubble base some 2 meters high suggests that the tower was intended to be as lofty as was practicable using the simplest and most economical of building techniques. A mud-brick structure standing two stories above the base seems likely. A third story is possible, especially if timbers were used to reinforce the mud-brick walls, but this is certainly the maximum that could be allowed for mud-brick walls standing on loosely joined rubble. We may imagine, there-fore, that the Dema tower originally stood close to 8 meters in height if it had two stories, or as much as 11 meters if it had three.^[25]

Entrance to the tower and interior communication between its stories were likely provided by wooden ladders, since no trace of stone steps was found built into or against its tall rubble base. If we introduce evidence from the Hymettos tower, the remains of which are similar to those of

[25] DEMA 173 suggests that a structure of one or two smiles stood atop the base. The height of each story of the Dema tower can be approximated at 2.50 to 3.00 meters, to judge by other examples of room heights. This is the suggested height of the ground-floor rooms of the Vail house ("Vari House" 442), and the smiles above ground level in two towers built of cut stone, one on Andros and one on Keos, fall within this range; see Young 1956, 137. DEMA 174 note 52, followed by McCredie 1966, 119 note 3, suggests that the Dema tower was as high as 15-20 meters, which is clearly impossible. The towers of that height cited by DEMA to support this conjecture are all built of well-cut masonry to their full heights and should not be considered suitable evidence for the height of a freestanding mud-brick tower like the Dema tower.

the Dema tower (see figure 34), then it is likely that the Dema tower had a massive central pier taking up as much as half the space within the first story, providing a firm bedding for a central post that probably sup-ported the roof and that might have continued upward as a mast to carry signal flags. The topmost story must have been well provided with windows for observation and to allow flags to be sent up the mast.^[26]

The round shape of this tower gave it more stability than a rubble and mud-brick tower with corners would have had. The roof, however, must have been square in plan, since the roof tiles used in it were canonical rectangular tiles. It may have been single pitched, double pitched (gabled), or pyramidal in form. Each possibility raises questions about how a square roof was erected above a round wall. What little evidence there is suggests that a pyramidal roof was used.

A pyramidal roof would have had the advantage of having all edges of the roof at the same level, that is, with no raking cornices, and would have been the easiest type of roof to erect on a round building. A pyramidal roof is in fact the simplest approximation, in a roof of rectangular files, of a conical roof, which would have required specially shaped tiles.^[27] Even so, a pyramidal roof would have required some of its rectangular tiles to be modified in shape. Pan tiles running down the lateral edge of each of the four triangular facets of the roof would need to be trimmed along a 45-degree line in order for each facet to fit flush against the adjacent facets. Since Lakonian-type tiles have distinct upper and lower ends, the trimmed upper portions would have been discarded. This may well have provided the source of the fide fragments used for the incised gameboards, nos. 21 and 22, and possibly 23, for it is note-worthy that no. 21 is a gameboard incised on a tile fragment with one edge broken along a 45-degree angle to the original sides of the tile. The find contexts of these reused tile fragments do not associate them with any particular phase of activity on the tower site, but it would be most appropriate to assume that they were incised by the idle hands of men on long and uneventful shifts of duty at the tower in its first phase.

Of the pottery finds, only nos. 1 and 2 can be associated with the first phase of the Dema tower. All fragments of both were found in area 2S, where the three fragments of the gameboard no. 21 were also found. The cup no. 1 is represented by two handle fragments, one found on the surface of the ground beneath rubble and the other found a few

[26] The remains of the Hymettos tower with its platform are described in Munn 1983, 406-10; cf. also McCredie 1966, 117-19, and Ober 1985a, 132-33. Textual and archaeological evidence for the use of signal flags at outposts like the Dema tower will be presented in a separate study.

[27] For examples of the special tile systems required for well-built round buildings, see Thompson 1940, 65-73; Roux 1952, 442-83.

meters away, buried in the mud-brick stratum near where no. 2 was found. The handles are from a type of cup produced no later than the end of the fifth century, very probably within the last third of that century. The three fragments of no. 2 were found close by the deposit comprising nos. 3, 14, and 18. The vessel form of no. 2 is not precisely identifiable from the preserved portion, but enough remains to show that this was a type of cup manufactured ca. 420-400. The interval of more than a half-century between the dates of nos. 1 and 2 and the datable pottery associated with the secondary structures clearly separates these pieces from the second phase of the site. Since these cups are unlikely to be merely stray pieces left on this barren hilltop before any structure existed here to make the spot a focus of human activity, these pieces can be identified as debris from the original use of the tower. They provide evidence that the first phase of the Dema tower is to be placed no earlier than ca. 425 and probably no later than a generation after the end of the fifth century, that is, within the range ca. 425-375.^[28]

The evidence of these sherds fits closely with the terminus post quem for the date of the Dema wall established by the Dema wall saltcellar. Taken together, the saltcellar and cups 1 and 2 establish a fairly narrow chronological range within which the construction of the Dema defenses is to be placed. The saltcellar indicates that a date within the last quarter of the fifth century is possible but that a date after the beginning of the fourth century is more likely (appendix I), and the cups nos. 1 and 2, as

[28] As is the case with the sherds assigned to the second phase of use of the tower, especially the best-preserved examples, nos. 3 and 5, an allowance of about a quarter-century beyond the dates assigned to parallel pieces in Agora XII for the period of use and final deposition of individual pieces is appropriate when extrapolating from parallel pieces to the date of an archaeological context. The dates assigned to pieces in Agora XII represent an amalgamation of context, i.e., deposit date, as determined especially by the presence of figured fragments, and an estimation of the place of an individual shape in a stylistic series, i.e., by comparison with parallel pieces from other deposits (Agora XII, 46). As Sparkes and Talcott point out, this system of dating is founded on the "cumulative and interlocking" nature of the evidence of the deposits, and considering the number of de-posits and volume of material consulted by them, it should inspire confidence that the resulting dates are reasonably accurate. It should be noted, however, that since the foundations of their chronology are, first, the chronology of figured pottery and, second, judgments about the nature and rate of stylistic evolution of vessel shapes (a process of dating comparable to the stylistic criteria applied to vase painting), the entire system of dating is strongly influenced by criteria related to the date of manufacture of individual pieces. Al-though, with the manufacture, the period of use of any given piece may well lie within the time range, often of a quarter-century, assigned in Agora XII, it may extend beyond it by as much as a generation. So Sparkes and Talcott caution: "There is often no clear indication of the lag between the time when a vase was made and the time when it was broken and thrown away," and "Ordinary tableware and domestic pottery may  . . . lay claim to a considerable life simply because it remained quite good enough for everyday use" (Agora XII, 46 and note 129).

first-phase occupation debris from the tower, indicate that a date any later than ca. 375 is unlikely.

This conclusion corresponds well with the terminus ante quem provided by the material from the second phase of the site. After its initial use, the tower was abandoned for a sufficient length of time that it fell into ruin, and its debris was reused by a beekeeper to provide shelter for his stands of beehives. The most probable period for this reoccupation of the tower site, ca. 340-300, indicates that original construction of the tower should probably be placed before the middle of the fourth century. A construction date within the first quarter of the fourth century would allow an adequate interval for the tower to have become dilapidated and clearly useless as a watchtower by the time the beekeeper arrived.

Excavation of the Dema tower has thus yielded evidence, corroborated by the Dema wall saltcellar, indicating that the construction of the Dema wall and tower might be placed as early as the last quarter of the fifth century and more probably is to be placed within the first quarter of the fourth century. This agrees well with the analysis of other archaeological criteria and with the evidence for the tactical plan of the wall discussed in chapter 2. This conclusion is based in part on the assumption that the Dema wall and tower were built at the same time as part of the same defensive scheme. The demonstration of this assumption depends upon an understanding of the function of the Dema tower and of its fundamental importance to the defensive scheme embodied by the Dema wall.

Purpose of the Dema Tower

The Dema tower has generally been considered to be a command post for troops manning the Dema wall. Arguments in favor of this view are presented by Jones, Sackett, and Eliot:

The Dema represents one general defensive scheme adopted for the whole pass, based on fluid tactics of counter-attack along a very wide front; it could not operate properly without an effective supervision, exercised from some central vantage point. The tower is the one point behind the wall which commands the greatest length of its undulating course, and was therefore suited for the observation of any hostile advances, and the signalling of local counter-attacks; for a Dema headquarters it was the best (and only) site.^[29]

This explanation is founded on the dual premises that the defense of the Dema required a central command post and that the Dema tower

[29] DEMA 173.

was located in the best (and only) spot to serve such a purpose. Both of these premises are false, the second most obviously so, since it can be disproved by simple observation on the spot.

As was noted in the first section of this chapter, the position of the tower on the highest point of Pyrgarthi leaves more than 600 meters of the wall near the tower concealed behind a secondary summit of Pyrgarthi (figures 21-22). Even if the tower originally stood as much as 11 meters high, not much more of the wall could have been seen from it.^[30] This is a serious objection to the command-post theory. If the purpose of the tower required it to have a view of the entire length of the wall, as it would if it were to observe and signal operations along the wall, then it could have been built on the very secondary summit, only 130 meters away, which blocks the view from its actual site. The fact that the Dema tower could have been but was not so situated is sufficient to refute the command-post explanation.

It is quite doubtful that defensive operations along the Dema wall could have been effectively controlled from a central vantage point, even if the Dema tower had provided such a vantage. Visual or audible signals were effective in conveying only a limited number of previously agreed-upon messages, and on the battlefield signals might prompt the commencement of certain prearranged maneuvers but were usually no more than trumpet calls for the advance or the retreat.^[31] Polybios observes that the sort of simple signal systems employed in his day were useless in the face of unforeseen circumstances, and the point certainly applies to signals on the battlefield, where the opportunities for surprise and confusion would be the greatest.^[32] It is no wonder, then, that no example of an army commanded by signals from a central vantage point can be cited as a parallel for the supposed purpose of the Dema tower. It is significant in this connection that even in the defense of a city-wall perimeter under attack, Aeneas Tacticus assumes that the commanding general will be on the battlements leading his troops wherever they are hardest pressed and not issuing commands by signal from some headquarters, even though he expects that such a commander will have a signal post from which a general alert signal can be seen over the whole city.^[33]

In addition to regarding the Dema tower as the command post of the Dema defenses, previous investigators have believed that signal com-

[30] From the elevations and distances recorded in map 3, it can be calculated that the Dema tower would have to have stood at least 17 meters high in order to command a view of the wall immediately beyond the knoll at elevation 224.75 m. As discussed in note 25 above, the Dema tower could not have been so tall.

[31] See Anderson 1970, 79-81.

[32] Polybios 10.43.5-10; cf. 10.45.1-5.

[33] Aeneas 38.1-5 on the commander leading his troops in battle, and 26.12-14, 27.2, on the central signal post in a city.

munications between the Dema and Athens would have been desirable and that the Dema tower served as the signal point at the wall.^[34] Everywhere within the Aigaleos-Parnes gap, the view of Athens is blocked by the northern end of Mount Aigaleos, so some intermediate relay station would have been necessary for signals to be passed between the Dema tower and Athens. On the northernmost summit of Aigaleos, there are remains of a tower that has long been associated with the Dema defenses, and this has been interpreted as the relay station for communications between the Dema and Athens (maps 2, 5; figure 35).^[35] From the Aigaleos tower there is a clear view of Athens, but it is impossible to see the Dema tower or any part of the main sector of the Dema wall because the summit of the mountain is too broad to allow a view into the valley to the north. Nor could the two towers ever have been tall enough to be in view of each other over the intervening shoulder of Aigaleos.^[36] Other positions on the Aigaleos ridge would have been suitable for such a relay station between the Dema tower and Athens, if such communications were desirable, but the actual Aigaleos tower is not appropriately located to serve this purpose. Furthermore, there is no evidence that any such relay station was built, either on Aigaleos or elsewhere.^[37] There is there-fore no reason to believe that the Dema tower ever served to communicate between the Aigaleos-Parnes gap and Athens.

The Dema tower provided a sheltered vantage point for a few men in a prominent position within the Aigaleos-Parnes gap. Since it must have been an observation post of some sort, it should be possible to see even now the places that were to be observed from the tower. We have just noted why it is doubtful that the tower was intended to communicate by signal with Athens and why it could not have served as a command post for the Dema wall. Considering the view beyond the wall, it is also apparent that the tower could not have been very effective in observing the movements of an enemy force approaching the wall, since like the view

[34] Dow 1942, 206; DEMA 174-75.

[35] DEMA 174-75; see also McCredie 1966, 119. Descriptions of the Aigaleos tower are also given by Munn 1979b, 18-21; Munn 1983, 402-5; Ober 1985a, 148-49; Ober 1987a, 208-9.

[36] DEMA 174 note 52, followed by McCredie 1966, 119 note 3, suggests that the two towers could have been 15-20 meters tall and thus could have been intervisible. The impossibility of this height for the Dema tower is argued in note 25 above, and the same arguments apply to the Aigaleos tower.

[37] The ridge of Aigaleos, the most appropriate location for such a relay station, has been repeatedly investigated by myself and others. See Munn 1979b, 11-23; McCredie 1966, 71-72, 115 note 30; DEMA 174-75, 185-86; Smith and Lowry 1954, 39-40 and appendix III; Curtius and Kaupert, Karten von Attika , sheet IV: Pyrgos (1883), surveyed by Siemens; Milchhoefer 1883, 44, 46. No remains of any sort in a suitable location have been reported on Aigaleos, nor are any known in the plain of the Kephisos east of the Dema.

of the wall itself, the nearer approaches to the wall are not all visible from the tower (figures 21-22). Furthermore, because it is well within the confines of Attica, it is highly improbable that the tower was meant to be a position from which enemy troop movements could be first detected. The tower could have served, however, as a post where men waited to receive and acknowledge visual signals coming from other observation posts closer to the frontiers of Attica.

The far side of the Eleusinian plain and the mountains of the frontier beyond are clearly in view from the tower site (figures 21-22, see also map 5, p. 99). Within this view, the town of Eleusis can be seen to the southwest, and to the northwest, the low summit of Plakoto and, above it, the higher summit of Velatouri stand out, marking the sites of fourth-century towers (see figures 37-40).^[38] From the Velatouri tower, Panakton, Oinoe, and Eleutherai can be seen, and from these positions other outposts closer to the western frontiers are visible, making up a network of observation and signal posts through which the arrival of an enemy force on the frontiers could have been signaled to the interior of Attica. Signals coming to Eleusis from Salamis, or to Velatouri and Plakoto from the Kithairon frontier, could have been relayed across Aigaleos to Athens by the Aigaleos tower or by its companion on the summit of Korydallos to the south of the Sacred Way.^[39] But the Dema tower, which could also receive these signals from the west, is in a very different position from these and other mountaintop towers in Attica. It sits not on a peak with wide long-distance views on all sides but on a hilltop that is comparatively enclosed within the Aigaleos-Parnes gap. Except for the nearer ground within the pass itself, it commands no view not already better surveyed from the Aigaleos tower. This very exception, however, pro-vides the decisive clue to its purpose: the Dema tower served to link the lookout and signal system of the western frontiers directly to Athenian forces at the wall in the Aigaleos-Parnes gap.

The Dema tower must have been built because it was of particular importance that signals from the west should be received and answered from the area of the Dema wall. As discussed in the previous chapter, the wall was only useful when it was manned by a sizable army. Since the defensive scheme of the Dema wall envisioned occasions when an army would be in place at the wall, it would clearly have been desirable for that army to receive the same warning signals or intelligence that might be sent from the frontiers to Athens. And since the commander of that

[38] On the Plakoto tower, see Edmonson 1966, 42-45; McCredie 1966, 72-74; Munn 1983, 420-25; Ober 1985a, 158; Ober 1987a, 212. On the Velatouri tower, see Vanderpool 1978, 238; Munn 1983, 422-25; Ober 1985a, 157-58.

[39] On the Attic signal network, see Munn 1983, 401-63, and Ober 1985a, 130-80. A further study of Attic watchtowers and mountaintop outposts is forthcoming.

army must have been empowered to initiate actions in response to those signals, it would have been important for signals to be sent back from the Dema to Eleusis and to the outposts on the frontiers. The Dema tower was eminently suited to such a purpose. Located on the highest summit within this pass, it would have been readily visible and accessible to forces assembled in the vicinity of the Dema wall, and its position could easily have been discerned from afar. It must therefore have served as the communications center of a major Athenian military camp at the Dema wall.^[40] The Dema tower was thus integral to the defensive scheme of the Dema wall, and furthermore, it was integral to a scheme that called for the placement of lookout and signal towers elsewhere in Attica.

Other towers, noted briefly above, indicated in map 5 below and illustrated in figures 34-40, are appropriately situated for the purposes of the lookout and signal system envisioned here, and these, on the evidence of masonry and surface sherds, were in use within the fourth century. Like the Dema tower, they are round, with solid rubble-filled bases. These resemblances and their functional suitability provide sufficient evidence to consider it probable that these lookout towers were employed along with the Dema wall as part of an integrated scheme for the defense of Attica. The Dema wall and tower are the most specialized works in this defensive system, and the interpretation of the whole depends upon a demonstration of the specific historical function of these key works. Chapter 4 will therefore bring together all of the various forms of evidence and lines of reasoning that, on historical as well as archaeological grounds, converge on the Boiotian War of 378-375 as the time of the creation of the Dema wall.

[40] It is possible that the so-called ghost wall behind the Dema wall (see DEMA 171) can be recognized as the remains of this Athenian encampment; see Munn 1985, 371-77.