² Tournavitou 1995.

³ New data concerning the architecture of this group of buildings have emerged during the recent restoration work conducted by the Ministry of Culture, which the author of this paper has been overseeing for two consecutive years (1999–2000). Further information on the stratigraphy and architectural details of the WH has become available after the even more recent discovery of the missing WH site-notebook for the all-important first season of work during January–March 1958 (Tournavitou 1995, 2 n. 1), accompanied by plans (Fig. 3) and photographs, a number of which were not included in the original archives. The newly discovered documents illustrate the original state of affairs in the west wing of the excavated building, with additional and in some cases more accurate information on individual features that were in subsequent years left out of the official reports and plans. Their contribution to the final reconstruction of the building has been invaluable.

⁴ Mylonas-Shear 1987, 150–4; Tournavitou 1995, 7, 285–7 (see Section III).

⁵ Verdelis 1958; 1959; 1961; 1963.

⁶ Tournavitou 1995.

⁷ Site-notebooks 20/8/59; Verdelis 1958, 161; 1959, 146; 1963, 13, 15.

⁸ The various omissions, discrepancies, even contradictions in the extant season plans (Fig. 4–6), and especially in the final, 1961 season plan (Fig. 6) as regards various architectural features, are one of the major issues and/or tools in the architectural interpretation of the building and have in fact prompted the present reassessment, especially after the recent conservation project (see n. 3). They have also dictated a selective approach as regards their use in the present reconsideration of the remains. Apart from the 1961 season plan (Fig. 6), which, as will become clear in the course of this paper, includes the greatest number of discrepancies/errors (see area of ‘wall’ A, north wall of the house), the two original 1959 season plans, one depicting the Mycenaean and pre-Mycenaean remains (Fig. 4), and one incorporating the post-Mycenaean features (Fig. 5), the former representing the most accurate depiction of the Mycenaean building immediately after excavation (state plan), have been in fact used as a basis for most of the plans illustrated in this paper, with the addition of any features omitted by the original architect, T. Kritharas; accordingly, most of the measurements cited in the text are taken from the 1959 plans, unless otherwise stated. The 1958 preliminary plan (Fig. 3) and the 1958 final season plan (Verdelis 1958, fig. 1), the former recently discovered along with the missing notebook, are selectively used in the course of this paper to substantiate specific points or architectural details. Also included is a recent plan of the house (Fig. 2), the most accurate depiction of the present condition of the remains after the completion of the restoration work conducted by the Ministry of Culture, produced with photogrammetric methods by the architect M. Xypnitou. The omissions, discrepancies, and/or contradictions in the season plans can be attributed to various reasons, such as the preliminary or changing character of the excavator's interpretation of the remains during successive years of excavation, before he had had time to attain an overview of the architectural and stratigraphical issues involved, as well as the degree of accuracy applied to the task by the various architects employed over the years with regard to the dimensions and orientation of the building.

⁹ Verdelis 1958, 160; 1961, 161–4.

¹⁰ The term ‘three-room unit’ was first used by Mylonas-Shear 1987, 150, with reference to the basic Mycenaean domestic unit consisting of a vestibule or anteroom, a main room, and a rear chamber along a single axis, opening into a courtyard or open space. Darcque's objections concerning the historical and political implications and the inconsistencies in the use of the term ‘megaron/megaron apartment/megaron unit’ over the years (Darcque 2005, 318–19) are not entirely without justification and necessitate a more objective and descriptive terminology, hence the adoption of the term ‘three-room unit’, which is in this case used in a much wider sense, for both the domestic (courtyard-vestibule/anteroom-main room-rear chamber) and the palatial/palatial-inspired versions of the unit (court-porch-vestibule/anteroom-main room). Darcque 2005, 332 has chosen a more elaborate term (‘unité principal stereotypée’) to describe the palatial version (court/two-columned porch/vestibule/hall, c.115–150 m² with central hearth surrounded by four columns), which was only attested at the palaces of Mycenae, Tiryns, and Pylos (for further comments on his use of the term, see Section III). For the use of the term ‘megaron/megaron apartment/megaron unit’ by different scholars see Shelmerdine, C., ‘Review of Aegean prehistory, VI. The Palatial Bronze Age of the southern and central Greek mainland’, AJA 101 (1997), 537–85, at p. 558CrossRefGoogle Scholar; Dickinson, O., The Aegean Bronze Age (Cambridge World Archaeology; Cambridge, 1994), 153Google Scholar; Werner, K., The Megaron During the Aegean and Anatolian Bronze Age: A Study of Occurrence, Shape, Architectural Adaptation, and Function (SIMA 108; Jonsered, 1993)Google Scholar; Hiesel 1990, 237–9.

¹¹ Site-notebooks 10/8/59, 13/8/59.

¹² For the post-Mycenaean remains in the area, see Appendix and Fig. 5.

¹³ East wall: depth of foundation trench: 1.55 m, Th. of wall in foundation trench: 1.65 m, Th. of wall above ground: 0.70 m/0.95 m (better preserved sections)–1.25 m (closer to the foundations). North wall: depth of foundation trench: 0.15 m (north side)–0.48 m (south side), Th. of wall in foundation trench: 0.75 m (site-notebooks 29/7/58, 30/7/58, 4/8/58, 5/8/58, 11/10/58).

¹⁴ For further details on the HL graves, see Appendix and Fig. 5.

¹⁵ Site-notebooks 18/10/58. The other reasons include (a) the existence of slab-like stones at the bottom of the opening, which he does not illustrate on any plan or photo, and which do not exist on the ground, and (b) his claim that the Mycenaean wall continues to the south for 1 m with the same width, which is not true (Fig. 4). In both his reconstructions of the house (Fig. 13 a, b), he depicts the gaps left in the east wall of the house by the first two HL tombs, without marking out the second one as the main entrance. Although in the 1959 reconstruction (Fig. 13 a) he offers the option of an entranceway to the house further south, in the 1961 reconstruction (Fig. 13 b) he does not: the east wall continues in an unbroken line to the south limit of the house. In 1963 (p. 15), he eventually acknowledges that, although the main entrance to the house must have been located on the east side, mainly on account of the existence of the passage between the two houses, its precise position is impossible to fix with certainty.

¹⁶ On ‘wall’ A see Section I. F(i).

¹⁷ See Appendix and Fig. 5.

¹⁸ See Appendix and Fig. 5.

¹⁹ According to Verdelis 1963, 18–19 the fourth HL tomb lay ‘between the passage and the east wall of the WH, completely destroying the latter at this point’. This statement, which was made for all four HL tombs that cut through the Mycenaean wall, implies that the east wall of the house did extend up to this point, even if Verdelis never convincingly substantiated his opinion in his reports and, to do him justice, did not actually restore it in this way in his first attempt at a reconstruction (Fig. 13 a). In the site-notebooks for 30/7/59, nothing of the sort is mentioned. French on the other hand (E. B. French, ‘Pottery from LH IIIB1 destruction contexts at Mycenae’, BSA 62 (1967), 149–93; fig. 1), having accepted the excavator's theory about an entrance south of the first HL tomb, seems to have also accepted his original and final views about the course of this wall, and restored the east wall as extending in a straight line, up to the south wall of the house (Fig. 13c).

²⁰ Mylonas-Shear 1987, 16.

²¹ For the post-Mycenaean sections of walls Γ and E, see Appendix and Fig. 5.

²² Site-notebooks 18/8/59, 19/8/59.

²³ 1.45 m below the top of wall Γ.

²⁴ Site-notebooks 19/8/59.

²⁵ Site-notebooks 18/8/59 (a drop of c. 1.70–2.0 m).

²⁶ Panagia group, House I (Mylonas-Shear 1987, 16, 151).

²⁷ Tournavitou 1995, 285–7.

²⁸ Verdelis 1959, 150, fig. 4. For further details on the post-Mycenaean features/layers south of the Mycenaean house, see Appendix and Fig. 5–6.

²⁹ L. 1.65 m, Th. 0.35–0.40 m (site-notebooks 10/8/59).

³⁰ Perhaps this was an afterthoutht, intended to decrease an originally wider south entrance to the west wing of the house, possibly dictated by the presence of the drain; perhaps the two walls (wall Γ and this wall) were joined higher up, leaving an opening only for the drain.

³¹ On the east face of the wall the foundation trench was traced for a length of c. 1 m south of the south porch wall. The pottery from the foundation trench, which had an average width of 0.35 m and an average depth of c. 0.13 m on both sides of the wall (site-notebooks 14/10/58), was mostly Mycenaean, with some late instrusions, especially at the north end, an area with intense overbuilding activity; see Appendix and Fig. 5 (site-notebooks 17/10/58: wall a: post-HL, wall A1: Archaic, wall II: HL).

³² Site-notebooks 14/10/58 (W. 0.35 m, depth: 0.10 m).

³³ See Section II.

³⁴ No further information is provided for this wall (see Appendix and Fig. 3, 5).

³⁵ See Section II.

³⁶ See Section I. F(i), Appendix, and Fig. 5.

³⁷ Verdelis 1963, 26, fig. 55. His argument involves a conical crater in a trench lying on the line of the foundation marking the east boundary of the corridor. He claims that, at this point the actual foundation was missing and that the trench was filled with small stones. It is of course conceivable that when he says ‘in the same position’, he might be referring to the position of the frescoes, which he had been discussing before (for frescoes, see below, Section II. G).

³⁸ In his first attempt at a reconstruction (Fig. 13 a), he does not depict this division beyond the entrance to Room 1. In his final reconstruction of the house (Fig. 13 b), he depicts the entire east limit of the corridor as a solid wall, reaching to the south end of the house, with a jog beyond ‘wall’ A.

³⁹ Verdelis 1963, 19–20, 27; Tournavitou 1995, 4, 15.

⁴⁰ Th. 0.55–0.70 m.

⁴¹ Site-notebooks 11/8/59. Its presence on the final season plan of the house (Fig. 6) is therefore not only superfluous, but also inexplicable.

⁴² Where he omitted the last 0.60 m of ‘wall’ A from the west, its southward return (Aβ), as well as ‘wall’ B and Bα to its south, as post-Mycenaean, while ‘wall’ A itself is shown in dotted lines to indicate a Mycenaean foundation below a later wall.

⁴³ These stones are probably what remained of the southward return of ‘wall’ A (Aβ), after years of exposure to the elements and/or human action, or possibly what was left after the excavator removed most of the post-Mycenaean walls in the area.

⁴⁴ It is only depicted on one of the original 1959 b/w photos, taken after the removal of the post-Mycenaean remains in the area (Plate 4 a).

⁴⁵ The Mycenaean date of the extant remains of ‘wall’ A can be discerned: (a) by the extant masonry style, especially on the north face, where it bonds with the substructure of the paved court to the north, and at the west end, where the ‘wall’ and the substructure of the court have been exposed to a greater depth (Plate 2 b); the fact that the stones on the south face of ‘wall’ A at the west end seem to be sitting on earth, is probably related to the formation of the bedrock; (b) by the preserved level of the remains. According to the recorded levels on the preliminary plan of the first season (Fig. 3) the west part of ‘wall’ A, lay c.0.62 m above the floor level in the corridor, while the top of its southward turn at the south end (Aβ), lay c.0.72–0.74 m above the Mycenaean floor surface and c.0.67 m above the top of the stone paving to the east; the stone pavement itself lay c.0.55 m below the top of ‘wall’ A at the west end. The fact that today, this ‘wall’, stands only an average of 0.16 m above the corridor floor (Fig. 9), implies that the excavator did eventually remove the post-Mycenaean construction off the Mycenaean foundation which lay under it.

⁴⁶ The original trench (Fig. 9) extended over the entire width of the corridor at this point, from the south end of the stone-built stretch, to ‘wall’ A in the north. The area to the west of the two slab-like stones, up to the east wall of Room 3, was only cleaned down to the plesia clay floor surface. Further investigation focused on the areas to the north of the slab-like stones, up to ‘wall’ A, and to their east, up to the southernmost limit of the stone-built stretch. Below the traces of the plesia clay floor surface discovered both to the north and to the east of the slab-like stones, a hard-packed earth layer, c.0.30 m thick, was revealed in the area to the north; this earth layer was identical with that recorded along the east wall of Room 1 (see Section II. G), and probably served as the substructure of the corridor floor. In the area to the east of the two stones was discovered a foundation trench along the stone-built stretch, with a soft earth fill. The test-trench was then extended for c. 1.60 m to the north of ‘wall’ A, within the limits of the corridor (see Section II).

⁴⁷ W. 0.30 m (north end)–0.13 m (south end).

⁴⁸ Verdelis 1958, 161; 1963, 15. Exceptions to this rule, such as Panagia House I, which have been attributed to the need for privacy, or to the avoidance of draughts, are characteristic of a more private/domestic type of establishment (Mylonas-Shear 1987, 16, 145), the palatially inspired model conforming to a more axial, symmetrical arrangement.

⁴⁹ Site-notebooks 30/7/58.

⁵⁰ See Appendix and Fig. 5. The alternative use of the levelled bedrock as a final floor surface has no real parallels in houses of this size and quality, while the pre-Mycenaean features in the area, would render such a solution even more unlikely (see below).

⁵¹ See Appendix and Fig. 5.

⁵² Tournavitou 1995, 8–9, with full discussion of structural details and interpretation.

⁵³ The so-called libation facilities in the throne room at Pylos (Blegen, PN 88) were more modest in dimensions and were not provided with an outlet, while the drain in the main room of the HofCols (Wace 1949, 93), passed under the final floor surface. Finally the drain passing through the south-east corner of the main room (Room 9) in Panagia House II (Mylonas-Shear 1987, 29), was not located in a central area of the room and was apparently associated with a stone paved area to the west. Considering the differences in quality and status between the two houses, domestic activities of the sort envisaged for the main rooms of the Panagia houses, are unlikely to have been the norm in the WH, where a special cooking area and other auxiliary spaces were situated in the west wing.

⁵⁴ Tournavitou 1995, 8 (L. 1.65 m, W. 0.22 m).

⁵⁵ Hiesel's designation of the ground-floor suite of rooms as working areas, solely on the basis of the existence of this minor drain (Hiesel 1990, 111–12, 128–9), is quite unrealistic (Tournavitou 1995, 8–9). The drain in Rooms 3–5 of the HSph (Tournavitou 1995, 60–3) is not really comparable on grounds of size, context, and circumstances.

⁵⁶ Possibly provided with a plesia clay floor, similarly to the rooms in the west wing.

⁵⁷ Wace 1949, 93.

⁵⁸ For buildings inside the citadel walls, see the Tsountas House (Tsountas, Chr., “Ανασκαφαὶ Μυκηνῶν τοῦ 1886”, PAE 1886, 59–79, esp. 74–8Google Scholar; Wace 1949, 66–7). For buildings outside the citadel walls see Panagia House I, Room 5 (Mylonas-Shear 1987, 17–8) and Panagia House II, Room 9 (Mylonas-Shear ibid. 29–30).

⁵⁹ See Appendix and Fig. 5. As regards the pre-Mycenaean features, apart from the egg-shaped stone construction built inside the bedrock in the south-west corner of the room (0.60 m east-west, 0.46 m north-south, depth: 0.15 m, site-notebooks 16/10/58), there were two other oval cuttings in the bedrock, towards the north-east corner of the room (site-notebooks 31/7/58, 6/8/58, 14/10/58, 16/10/58) and three holes (T1, T2, T3: D: 0.20–0.25 m, depth: 0.17–0.28 m), to their south, which the excavator also assigned to the pre-Mycenaean phase of the house (Fig. 4).

⁶⁰ See Appendix and Fig. 5.

⁶¹ The absence of stone column bases from the extant remains of the Mycenaean house is not really surprising, considering the intensive post-Mycenaean reoccupation in the area (Verdelis 1958, 162; 1963, 15; Fig. 5).

⁶² Hiesel 1990, 226–7, table 13; Darcque 2005, 117–18.

⁶³ L. 6.20 m. The stone-built stretch (αβ) was also depicted as a Mycenaean wall running north-south off the west end of ‘wall’ A (see Section I. D).

⁶⁴ L. 5.50 m. Its southward return (αβ) was omitted altogether as a post-Mycenaean feature.

⁶⁵ See Section I. D.

⁶⁶ The ‘wall' itself is preserved for a length almost equal to that depicted on the 1961 final plan (Fig. 6; L. 6–6.20 m).

⁶⁷ Stereo was not reached at this point, either by the excavator, or by the restoration team, since that would necessitate a large scale operation.

⁶⁸ See Section II.

⁶⁹ A similar case, as regards the merging of so-called walls with the court pavement has been made for the section of the west wall of the east wing and of the east wall of the house, south of the porch. The existing differences in the preserved levels between the top of ‘wall’ A and the remnants of the paved court to the south are easily explained by reference to the post-Mycenaean disturbance in the area, which resulted in the extensive displacement of the top course of the pavement at various places, not to mention its complete disappearance in others (see n. 45).

⁷⁰ See Appendix and Fig. 5. Verdelis ignores it as a post-Mycenaean feature in his earlier reconstruction (Fig. 13 a), but for some reason depicts it as a solid wall in the final season plan (Fig. 6) and in his final restoration (Fig. 13 b).

⁷¹ The section of this wall inside the passage between the WH and the HOM does seem to bond with the section inside the house (site-notebooks 6/8/59).

⁷² Site-notebooks 12/8/59 (L. 2.70 m, Th. 0.70 m). The oval cutting in the bedrock to the south of the Mycenaean wall (L. 1 m, W. 0.60 m, depth: 0.25 m; Fig. 4; Plate 3 b), including pottery of Early Mycenaean date and remains of bones, was probably a pre-existing feature (old burial).

⁷³ A view shared by the excavator in only one of his published reports (Verdelis 1959, 148).

⁷⁴ A circular rock-cut hollow, containing a small stone covered by plesia clay near the west end and to the south of this stretch (site-notebooks 1/8/59: 1.70 m from the modern surface; no further details are given; Fig. 4), is very similar in structure to the pivot-holes recorded in three of the doorways of the west wing (see Sections II. A, C, D) and could be part of an entrance to the court area of the house from the east (see Section III and Figs. 14–16).

⁷⁵ See Appendix and Fig. 5.

⁷⁶ It is really surprising that the excavator, despite his one-off statement about the original extent of the court (Verdelis 1959, 148), did not reconstruct the court pavement up to the southern limit of the Mycenaean house (Fig. 13 a, b). Although stone-paved courts are not a common feature of Mycenaean architecture, domestic or otherwise, the evidence for a different material covering the stones in the court area is very uncertain (see Section I F(iv)). The plaster forming the final floor surface of the court of the HofCols was a more or less compulsory choice since it covered the hard packed earth layer, c.0.70 m thick, which sealed the stone fill of the artificial terrace under the court (Mylonas, G. E., “ Ανασκαψὴ Μυκηνὼν ”, PAE 1967, 7–19, at 10)Google Scholar. Similar arrangements were attested inside the court of the Tsountas House (Tsountas (n. 58), 74–5).

⁷⁷ The preserved surface of the stone pavement north of ‘wall’ A, is clearly not the original surface of the court, which in Mycenaean times must have stood at a higher level, equal at least to the level of the stone slabs covering the drain to the south of the porch. The latter now lie only 0.23 m lower than the highest preserved point of ‘wall’ A at the west end, a logical inclination, facilitating the drainage of the rain water towards the north. The stone pavement ranged in thickness between 0.65 m in the area between the porch and ‘wall’ A, between 0.71 and 0.75 m in the area between ‘walls’ A and B, and c.0.45 m at Γ α.

⁷⁸ The south porch wall, the exterior east wall of the house, the east wall of the corridor, ‘wall’ A, and ‘wall’ B.

⁷⁹ On the whole, it seems that the entire corridor, at least between the porch and ‘wall’ A, including the corresponding west edge of the paved court, formed a single, uninterrupted stone substructure, which was differentiated only at floor level by the plesia clay surface in the corridor.

⁸⁰ Site-notebooks 29/7/59.

⁸¹ Stereo recorded here c. 1.35–1.45 m below the surface. On the bedrock were recorded two architectural features antedating the Mycenaean house: (a) c.0.45 m below the top of Γ α and 1.80 m to the south of the HL section of ‘wall’ B, a small east–west drain lined with plaster (pres. L. 0.50 m, W. 0.22 m; Fig. 4) was recorded directly on the bedrock; it apparently continued under the pavement itself, (b) Opposite this drain, to the north, were located two holes of uncertain purpose dug in the bedrock (D: 0.20 m; site-notebooks 31/7/59; Fig. 4).

⁸² D: 0.10 m, depth: 0.04 m, c.0.35 m apart (site-notebooks 29/7/59). The hard earth layer was probably contemporary to the house, since it included a few tiny Mycenaean sherds. The same hard earth layer was encountered over the entire west half of the court area south of ‘wall’ A (see Section I. F(iv) ). Stones, probably fallen from wall Γ, and mostly Mycenaean sherds were recorded over the stones of the pavement, between 0.90–1 m below the modern surface (site-notebooks 28/7/59, 30/7/59).

⁸³ Corroborating the theory which places the original east limit of the house on this line (see Section III).

⁸⁴ This roughly built feature, which did not preserve any proper faces, stopped c.1.30–1.40 m short of the northernmost preserved limit of Γ α to the east, and seems to have also served as the east wall of the drain. On the 1959 season plan including the post-Mycenaean features (Fig. 5), and in the final reconstruction of the house (Fig. 13 b), it is depicted as a post-Mycenaean feature; in the same reconstruction, Γα itself is omitted altogether, while in the earlier reconstruction of the house (Fig. 13 a), it is depicted as a solid wall.

⁸⁵ Site-notebooks 30/7/59. It appeared at the same depth as Γα (1 m below the modern surface), it lay directly on stereo and was preserved to a height of c.0.40 m above the latter.

⁸⁶ Site-notebooks 5/8/59. Two tombs were recorded to the north of this pavement: (a) the fourth HL tomb in the east part of the house (Tomb b; see Appendix and Fig. 5), and (b) a tomb of Early Mycenaean date (L. 0.70 m, W. 0.25 m, depth: 0.10–0.15 m), located just north of the north-west corner of the stone pavement, 0.35 m below its top. It consisted of a hollow in the rock, which contained a highly contracted skeleton with the head to the west (site-notebooks 1/8/59, 5/8/59; Fig. 4).

⁸⁷ There is no architectural, or other feature (drain?) that could explain this arrangement (see Section III).

⁸⁸ Tournavitou 1995, 5–6.

⁸⁹ One of the properties of plesia clay is to facilitate the flow of water, and make surfaces waterproof. The practice of coating the interior of drains with clay plaster is also attested at the Panagia houses (House I, Rooms 2, 3; Mylonas-Shear 1987, 17, 22).

⁹⁰ The discovery of a number of household vessels with strong traces of burning, together with some mudbrick fragments from the superstructure of the Mycenaean house inside the main branch of the drain, near its junction with the west branch (site-notebooks 16/10/58), confirms the unroofed status of the drain south of the porch. The excavator states in two of the published reports (Verdelis 1958, 163; 1963, 18) that a number of household vessels inside the drain, at the junction of the main branch with the west branch, preserved a thick layer of oxidation, not burning, on the surface, presumably, as he concludes, from the dirty water thrown away from the kitchen, i.e. Room 4. The vessels were unfortunately not located among the WH pottery. No traces of cover slabs have been recorded anywhere else in the house.

⁹¹ In the vast majority of the houses with court drains inside the acropolis (HofCols, House M, South House), including the palace, the drains in question were sealed by the floor and were therefore provided with drain traps at floor level, for the collection of the rain water: Wace 1949, 75, 93; Mylonas, G. E., “ Ανασκαφὴ Μυκηνῶν ”, PAE 1963, 99–106, at p. 102Google Scholar; Wace, A. J. B., ‘The South House’, BSA 25 (1921–1923), 86–96, at p. 87Google Scholar. The only exception with an open drain, as well as a drain trap, was the open court area (Room 7) of the House of the Artisans (Mylonas 1965, 91)Google Scholar and possibly House Δ (id., “ Ανασκαφὴ Μυκηνῶν ”, PAE 1966, 103–14 at 106; id., “ Ανασκαφὴ

Μυκηνῶν ”, PAE 1967, 7–19 at 16); neither of these courts was nevertheless stone paved. The big drain in House III of the Panagia group was possibly originally an exterior drain and was only later incorporated into the house (Mylonas-Shear 1987, 48); where it remained open, it either crossed small basement rooms (Rooms 37–8), or the niche in the main room of House II (Room 9), where it was assumed that it was related to activities requiring the use of water (Mylonas-Shear 1987, 29). A similar function was assigned to a much smaller stone-lined, open drain crossing from the court of House I to Room 2 (Mylonas-Shear 1987, 22, 29).

⁹² Tournavitou 1995–6.

⁹³ Site-notebooks 20/7/59.

⁹⁴ Most of these post-Mycenaean features in the area were eventually removed by the excavator (site-notebooks 11/8/59).

⁹⁵ Especially in the east half of the area, between walls A and B.

⁹⁶ In the area between ‘wall’ Β and wall Γ; it extended up to c.0.50 m north of the north line of Γα (FIG. 5).

⁹⁷ Below the depth of 1.10 m from the modern surface in the area between ‘walls’ Α and Β, and below the depth of 1–1.25 m from the modern surface in the area between ‘wall’ Β and wall Γ. The thickness of the Mycenean deposit varied between 0.75 and 0.80 m in the area between ‘walls’ Α and Β, and between 0.25 and 0.60 m in the south part of the court (site-notebooks 31/7/59), increasing in thickness towards the east, where the bedrock dipped (FIG. 7 c).

⁹⁸ Site-notebooks 28/7/59, 29/7/59.

⁹⁹ Chiefly on account of the fragmentary condition of the finds and the ambiguity concerning their date.

¹⁰⁰ Th of layer: 0.60–0.70 m (site-notebooks 31/7/59).

¹⁰¹ See Section III. Immediately below this layer was discovered a north–south MH wall, lying directly on stereo (pres L. 1.10 m, Th. 0.55 m, H. 0.25 m; site-notebooks 31/7/59).

¹⁰² Animal bones, some obsidian, an astragal (site-notebooks 4/8/59) and three fragmentary borders of Linear B tablets, c.0.50 m south of ‘wall’ Β (site-notebooks 30/7/59).

¹⁰³ Pivot hole: D: 0.25–0.28 m, depth: 0.07 m.

¹⁰⁴ The excavator, for some reason, states that the door from the corridor into Room 2 opened inwards from the left (Verdelis 1963, 20), a rather impractical proposition also repeated in the final publication of the house (Tournavitou 1995, 11).

¹⁰⁵ Site-notebooks 17/1/58, 28/2/58, where the position of the pivot hole is indicated on measured drawings. Curiously enough, the particular pivot hole was illustrated as such only on one of the original 1959 season plans (FIG. 5); it has now been restored on the other 1959 season plan too (FIG. 4).

¹⁰⁶ Originally labelled Room N (Νοτιο Δωμὰτιο) by Verdelis (FIG. 11).

¹⁰⁷ W of opening c.1.25 m. The current measurements are somewhat increased (1.35 m), after damage caused by the long exposure of the remains to the elements and possibly by human action (FIG. 2).

¹⁰⁸ Site-notebooks 27/2/58: 0.80 m; Fig. 3: 0.70–0.74 m above floor level.

¹⁰⁹ Site-notebooks 17/1/58.

¹¹⁰ Cleaning operations during restoration work (22/7/99).

¹¹¹ Site-notebooks 13/1/58. Pieces of this lime plaster were allegedly kept by Verdelis, but were not located amongst the WH material in the course of my study of the building. Impresssions of reeds or other organic material on clay/clay plaster/lime plaster can be apparently associated with both flat and pitched roofs as well as with ceilings (Hiesel 1990, 221–5).

¹¹² The fact that the fragments of lime plaster were not found amongst the WH material unfortunately prevents us from making any direct connections/comparisons with the preserved fragments of hardened clay packing allegedly surrounding ceiling beams, discovered in the Panagia group of houses (Mylonas-Shear 1987, 8), where they were used as corroborating evidence in favour of flat roofs.

¹¹³ Also suggested by the excavator (Verdelis 1963, 26).

¹¹⁴ Site-notebooks 15/1/58, 18/1/58.

¹¹⁵ Verdelis 1963, 20; Tournavitou 1995, 9 (L. 3 m, W. 0.75 m, H. 0.10 m). On the 1958 preliminary plan (Fig. 3), the height of this platform is given as only 0.05 m.

¹¹⁶ Site-notebooks 15/1/58.

¹¹⁷ From the top of the walls, presumably the best preserved ones, to the west and south.

¹¹⁸ Site-notebooks 15/1/58.

¹¹⁹ The deep black colour of the deposit suggested to the excavator that wood was one of the main building materials in the construction of the house, while the large quantity of mud-bricks was assigned to the superstructure of the walls, or to the remains of a second floor, built entirely of clay (site-notebooks 15/1/58).

¹²⁰ Similar slabs were mentioned with reference to roofing arrangements in the context of the sacred centre, in the area of the round altar (Mylonas, G. E., Τὸ θρησκευτικὸν κὲντρον τῶν Μυκηνῶν (Πραγματεῖαι τῆς Ακαδημεὶας Αθηνῶν, 33; Athens, 1972), 27Google Scholar: ‘στοὰ κεκαλυμμὲνη ὺπὸ πλακῶν’. For the reconstruction of a roof over west wing, see Section III.

¹²¹ Tournavitou 1995, 9–10. The contents of this room consisted of pottery, mostly small used stirrup-jars in a variety of sizes (69 out of 100 vases), Linear B tablets, and half a seal-stone. It is worth keeping in mind, that both the seal-stone and all the extant tablets inside this room were found in the same context, along the south wall (Fig. 10–11), not in association with the pottery deposit and were possibly indeed originally placed/kept on the raised platform, as the excavator claims (Verdelis 1963, 20). For the association of the contents of the tablets with the stirrup jars in the room, see Tournavitou ibid. 9–10. Shelmerdine, C. W., ‘Workshops and Record Keeping in the Mycenaean World’, in Laffineur, R. and Betancourt, P. P. (eds), TEXNH: Craftsmen, Craftswomen and Craftsmanship in the Aegean Bronze Age, Annales d'archéologie égéenne de l'université de Liège et UT-PASP (Aegaeum, 16/II), 387–95 at 392–4Google Scholar; id. 1999, ‘A Comparative Look at Mycenaean Administration/s’, in S. Deger-Jalkotzy, S. Hiller, and O. Panagl (eds), Floreant Studia Mycenaea, Akten des X Internationalen Mykenologischen Colloquium in Salzburg (Verlag der Ōsterreichischen Akademie der Wissenschaften, 274, Band II, Vienna), pp. 555–600, esp. 572–3, after confirming my basic conclusions (Tournavitou 1995)—especially about the administration of the WH providing monthly support to workers and about the close administrative ties between the four houses (WH, HSh, HOM, HSph), which functioned as part of the palatial administration, serving as repositories and clearing houses—claims that the term ‘secondary archive’, is in this case misapplied, that the majority of the vessels in the room are not connected with the subject matter of the tablets discovered herein (not all were allocations of rations), and that the room was chiefly a pottery storeroom. Her objections as to the use of perfumed oil in small stirrup-jars for monthly rations, is irrelevant in this context because it was never really suggested in the final publication of the house (Tournavitou ibid.); what was suggested was that the small stirrup-jars could have also contained plain oil for rations (Tournavitou ibid. 10). The suggestion made by Darcque 2005, 364–5 concerning the contents of this group of buildings, according to which the fine stirrup-jars in this room could have been ultimately intended as grave offerings, is in this case rather misplaced, both because they are closely associated with the Linear B documents in the same room (see above) and because these particular vases had been used, perhaps more than once.

¹²² Intact, on the floor surface (Tournavitou 1995, 9–10).

¹²³ Originally labelled Room Μ (Μεοαῖο Δωμὰτιο) by Verdelis (Fig. 11).

¹²⁴ At this point, only 0.13 m lower than the west wall of the room (site-notebooks 17/1/58: H. of west wall 1 m; 27/2/58: H. of west wall 0.92 m; Fig. 3: H. of west wall 1.05 m).

¹²⁵ Especially with reference to the lack of post-Mycenaean disturbance and the nature of the burnt debris lying on the floor.

¹²⁶ Preserved top surface of stones at south end, c.0.30 m above floor level (Fig. 3).

¹²⁷ Tournavitou 1995, 11.

¹²⁸ Originally labelled Room B (Βὸρειο Δωμὰτιο) by Verdelis (Fig. 11).

¹²⁹ Although the opening of the doorway was c.1.30–1.35 m, the width of the actual door would be reduced by 0.30 m, the latter representing the diameter of the extant pivot hole (depth: 0.08 m/0.09 m), almost identical with that of the door from the corridor. In this case too, a stone slab was placed at the bottom of the hole (site-notebooks 15/1/58; Fig. 10–11; Plate 16 a).

¹³⁰ Again the excavator maintains that the door into Room 3 opened inwards, to the left (Verdelis 1963, 20).

¹³¹ Site-notebooks 15/1/58. The threshold was also probably wooden, although no traces were recorded in situ. About the popularity and use of wooden thresholds in Mycenaean architecture see the Panagia group of houses (Mylonas-Shear 1987, 16, n. 6).

¹³² North wall: preserved to 0.80–0.90 m above floor level. West wall: preserved to 0.95 m above floor level (Fig. 3)

¹³³ Site-notebooks 16/1/58, 17/1/58, 28/2/58. As depicted on the 1958 preliminary plan (Fig. 3), its top lay c.0.23 m above floor level.

¹³⁴ In the final publication of the house (Tournavitou 1995, 12), which was printed before the discovery of the long-lost data of the first season, no mention is made of other pottery in the room except for the pithos. The presence of the funnel makes perfect sense with reference to the pithos in the same room, and with the stirrup-jars in Room 1.

¹³⁵ Tournavitou 1995, 11–12. The position of the pithos is indicated in the site-notebook drawings (site-notebooks 15/1/58, 17/1/58), but no reference is made to it in the text. A few fragments of this pithos, were apparently found by the pivot hole. In the published reports, on the other hand, the excavator claims that the pithos was found in the centre of the room (Verdelis 1958, 160; 1963, 22, fig. 33); the author of this paper considers the site-notebooks as more reliable.

¹³⁶ Site-notebooks 16/1/58.

¹³⁷ Ibid.

¹³⁸ Verdelis 1959, 152; 1963, 22–3.

¹³⁹ Verdelis 1959, 152; 1963, 23, where he suggests that the opening in the wall was intended to supply air to the hearth and light and heating to a room, evidently used as a storeroom (Tournavitou 1995, 11–12).

¹⁴⁰ This of course does not preclude the existence of a terracotta chimney-pipe over the hearth proper. The absence of chimney-pipe fragments could be explained by the post-destruction disturbance, a phenomenon quite common in Mycenaean domestic architecture.

¹⁴¹ See below, Section II. D.

¹⁴² Site-notebooks 15/1/58.

¹⁴³ Originally labelled Room 1 by Verdelis (FIG. 11).

¹⁴⁴ Tournavitou 1995, 13–14.

¹⁴⁵ The pivot hole (D. 0.22 m, depth: 0.08 m), with the stone slab at its bottom, was located nearer to the inner face of the east wall of the room, with c.0.10 m to spare between it and the south face of this wall (FIG. 3–4, 10–11), just enough space for the wooden section of the doorjamb (site-notebooks 28/2/58).

¹⁴⁶ Verdelis 1963, 20. Yet in the site-notebooks (21/7/59), three slabs are illustrated (FIG. 12). In the site-notebooks for 27/2/58, the threshold is described as λιθινὴ φλοιὰ [sic], without any further elaboration. On the 1958 preliminary plan (FIG. 3), only one slab is illustrated. After the clearing operations in the area, only the remains of two stone slabs were recorded. The top surface of the stone slabs, was at the same level as the floor surface inside the room.

¹⁴⁷ The peculiarity of the east wall, which was built partly on a pre-Mycenaean transverse wall of MH/LH I date running east–west across the corridor (Verdelis 1963, 23) and partly on the rising bedrock to the north (FIG. 3–4, 10–11; PLATE 8), is worth noting and is related to the reuse of earlier architectural features in the construction of the Mycenaean house. The MH/LH I wall appeared c.0.08 m below the floor surface of Room 4 (FIG. 3). The eastward curvature of the east wall of Room 4, which cannot be attributed to a structural priority or severe earthquake damage, was probably an afterthought, related to the accessibility of the north part of the west wing (see n. 188). The east wall was preserved to 0.91–1.28 m above floor level (site-notebooks 27/2/58; FIG. 3). The lack of plaster on the interior face of the east wall of the room was possibly the result of unfinished repairs. The north wall, built on the rising bedrock (PLATE 8), rose by the north-west corner to c.1.74 m above the floor level in Room 4 (FIG. 3).

¹⁴⁸ Site-notebooks 27/2/58 (pres. H. 1 m).

¹⁴⁹ See Appendix.

¹⁵⁰ Verdelis gives conflicting evidence. First he reports traces of a plesia clay floor (site-notebooks 27/2/58), and then in his summary of the room (site-notebooks 28/2/58), he assigns the room a floor of trodden earth. Traces of a plesia clay floor were rediscovered in the south half of the room, during the recent restoration project in the building.

¹⁵¹ For a discussion of the contents of this room, see Tournavitou 1995, 13–14.

¹⁵² 0.45 m north-south × 0.30 m east–west, depth 0.05–0.06 m below floor level (Figs. 3–4, 10–11; Verdelis 1963, figs. 41, 45).

¹⁵³ Verdelis 1958, 158; 1963, 23; Tournavitou 1995, 13.

¹⁵⁴ Site-notebooks 28/2/58; Verdelis 1963, 23, figs. 42–4.

¹⁵⁵ Site-notebooks 21/7/59; Tournavitou 1995, 13. Mylonas was the first to suggest that it was a hearth and not a cupboard or the site of a vertical wall beam (site-notebooks 20/7/59).

¹⁵⁶ In the first 0.04 m of this layer was identified in the west part a quantity of burnt organic material, i.e. seeds and animal bones, mostly of sheep and goats (site-notebooks 21/7/59; Verdelis 1959, 150; 1963, 23; Tournavitou 1995, 13).

¹⁵⁷ c.0.29 m above floor level.

¹⁵⁸ One of the mudbricks was recorded in the west part, by the north wall of Room 3, and two along the south side of the hearth opening.

¹⁵⁹ Site-notebooks 21/7/59.

¹⁶⁰ c.0.24 m above floor level.

¹⁶¹ The actual pot was never registered among the WH pottery finds.

¹⁶² Tournavitou 1995, 13, pl. 3 a. The dimensions of the five logs from west to east are: (a) L. 0.45 m, W. 0.07–0.15 m; (b) L. 0.50 m, W. 0.12 m, including a branch; (c) L. 0.50 m, W. 0.17 m; (d) L. 0.45 m, W. 0. 13 m; (e) L. 0.37 m, W. 0.07 m (site-notebooks 21/7/59). No dimensions were given for the sixth log (Verdelis 1963, fig. 35; Tournavitou 1995, pl. 3 a). Over the east part of the hearth opening, directly over the logs, lay a solidified mass of burnt building debris from the house (site-notebooks 21/7/59), a further argument against the communication of the hearth with the catch-pit in Room 3.

¹⁶³ Originally labelled Room 2 by Verdelis (Fig. 11).

¹⁶⁴ Site-notebooks 28/2/58.

¹⁶⁵ c.2.60 m.

¹⁶⁶ See Appendix and Fig. 5.

¹⁶⁷ As observed by the architect, M. Xypnitou, the east wall of this room does not align with any other wall in the house and its position was probably determined by the formation of the bedrock. The 1958 preliminary plan (Fig. 3) suggests that it was built in a rock cutting, perhaps a foundation trench. It stood c.0.52–0.61 m above the floor level in the room.

¹⁶⁸ See Appendix and Fig. 5. The proposed length of the wall could coincide with the northernmost edge of the HL well (Fig. 15–16).

¹⁶⁹ The top surface of the specially cut bedrock on the north side stood c. 1.05 m above the floor of the room (Fig. 3). For the west side, we have no measurements.

¹⁷⁰ Perhaps the burnt mudbricks reported in the south-west corner of the area (site-notebooks 21/2/58) belong to the superstructure of these walls.

¹⁷¹ The latter lay c.0.81 m higher than that of Room 4 and 0.86 m higher than the floor level in Rooms 1–3, again a consequence of the rising bedrock (Fig. 3). The level of the bedrock in the corridor outside Room 5 was c.0.42 m lower that that inside the room.

¹⁷² Tournavitou 1995, 14.

¹⁷³ Site-notebooks 19/2/58.

¹⁷⁴ Originally labelled Room 2α by Verdelis (Fig. 11).

¹⁷⁵ Site-notebooks 28/2/58; Verdelis 1963, 25; Tournavitou 1995, 14. After the first season of work, the excavator suggested that this area was used as a room, or more likely as a court (Verdelis 1958, 161).

¹⁷⁶ See Appendix and Fig. 5. The assumed Mycenaean foundation was never exposed.

¹⁷⁷ The preserved top of this wall lay c.0.50–0.66 m below the rock surface on the terrace (FIG. 3).

¹⁷⁸ For wall IV, one of the HL walls in the area (FIG. 3, 5; Plate 1), see Appendix. The vertical face of the rock on the east side of the terrace continued for 7 m beyond the north limit of the house.

¹⁷⁹ Verdelis 1958, fig. 1; 1963, 25.

¹⁸⁰ Site-notebooks 22/2/58, 28/2/58. Next to the remains of the wooden beam, Verdelis mentions two sections of thick wall plaster, about whose date he was uncertain. Part of a post-Mycenaean stone pavement he recorded in the south half of the terrace, apparently removed along with the rest of the late features in the area, was only once mentioned as the original flooring of the terrace (site-notebooks 22/2/58: dimensions 3.50 m east-west × 2 m north-south).

¹⁸¹ See Section III.

¹⁸² The position of this staircase is very similar to that of the staircase in House I of the Panagia group (Mylonas-Shear 1987, 23; Tournavitou 1995, 15). The existence of a staircase in this area, leading to an upper floor, was also suggested by the excavator (Verdelis 1959, 154; 1963, 25; FIG. 13 a, b), irrespective of his views about the absence of a roof over the terrace.

¹⁸³ House of the Artisans: W. 0.41 m, H. 0.16 m (Mylonas 1965, 92).

¹⁸⁴ Average W. of step 0.40 m, average H. of ste: 0.15 m.

¹⁸⁵ 1.20 × 1.20 m. Most details concerning the reconstruction of this staircase were supplied by the architect M. Xypnitou.

¹⁸⁶ Verdelis 1959, 154; 1963, 25. This reconstruction implies that the wall lining the south half of the terrace on this side need not have extended any further to the north. Both Verdelis (FIG. 13 a, b) and French (Fig. 13 c) propose the existence of a staircase in this area, closely associated with the cuttings in the bedrock.

¹⁸⁷ See Section I.

¹⁸⁸ The considerable width of the section of the corridor running along the first three rooms of the house, i.e. c.1.80 m, was probably related to the traffic in this area, as well as to the more intensive or public use of these rooms. The drastic reduction of its width in the section between ‘wall’ A and the north limit of Room 4, down to c.0.65–0.70 m, can be only understood in the absence of a full-scale wall marking the west limit of the court, at least up to the level of the porch and was also probably related both to the activities associated with this part of the house (cooking, storage of household utensils and foodstuffs) and to the more private nature of the north part of the wing. The presence of a hearth facing outwards towards the corridor implies the free circulation of air in the vicinity and the availability of some free space around it that would facilitate cooking. In the absence of any concrete evidence of earthquake damage in other parts of the house, which might have explained the eastward curvature of the east wall of Room 4 (see n. 147), the only suggestion that seems remotely possible, might be related to an attempt physically to delimit access to the north part of the west wing, with the pottery storeroom (Room 5) and a perhaps more private sector (north-west terrace). The increase in the width of the corridor before Room 5, to c.2.80–2.90 m, is again probably related to the character of that particular room (storage).

¹⁸⁹ Site-notebooks 27/2/58, 28/2/58.

¹⁹⁰ Site-notebooks 28/2/58. In the 1958 preliminary plan (FIG. 3), he nevertheless illustrated the area of the corridor before Room 5 as if it were covered with a plesia clay foor.

¹⁹¹ According to the findings during the recent restoration work in the area, the plesia clay surface begins c. 0.10 m east of the east walls of Rooms 1–3; in the narrow zone against both walls only the hard trodden earth substructure of the floor is preserved. A similar situation is recorded along the patch of stone pavement to the south of Bα to the east, as well as along the preserved stretch of pavement to its north (Fig. 9; Plate 3 a). In the doorway at the south end of the house, no plesia clay layer was attested, only a hard, thin earth layer, which suggests that, on the analogy of the corridor floor, this was the substructure for the threshold area. A hard-packed earth layer was also reported under the clay floor surface in the Panagia houses (Mylonas-Shear 1987, 55).

¹⁹² Site-notebooks 17/1/58, 28/2/58, 1/3/58, 3/3/58; Verdelis 1958, 16; 1961, 163; 1963, 26; Tournavitou 1995, 15. The majority of the frescoes, c.3500 fragments in all, monochrome in their majority, but including a considerable number of fragments with figured scenes, were found on the plesia floor of the corridor, some apparently facing up and some facing down; more fresco fragments were discovered in the first room of the MH/LH I building to the south of the WH (fig. 4); their distribution so far south, in the absence of any evidence for a serious earthquake destruction, is somewhat peculiar, unless we interpret this as a result of later (HL) disturbance. The frescoes, which were not located during the original period of study of the house (Leonardon storerooms of Nauplion Museum: 1983–7), were rediscovered in the new museum at Mycenae and are currently under study for publication by me. The lower part of a Linear B tablet (AU657), was also recorded in the corridor, by the entrance to Room 2.

¹⁹³ See Section III. Judging by the layout of the south part of the house and the use of the different areas, the frescoes would have decorated the exterior face of the east wall of Room 1 and possibly also of Room 3 further north. The position of the frescoes so close to the east wall of Room 1, does not suggest that they had fallen from an upper storey, above Room 1 for example, nor is the existence of such a storey supported by the stratigraphical evidence (see Section II. A–C).

¹⁹⁴ See Section I. D and n. 46.

¹⁹⁵ 0.06–0.12 m lower than the level of the plesia clay surface attested in the corridor south of ‘wall’ A.

¹⁹⁶ The level of the bedrock, is here c.0.50 m higher than the recorded floor surface of the corridor further south, by the hearth (Fig. 3).

¹⁹⁷ Site-notebooks 21/2/58, 26/2/58; Verdelis 1958, 161.

¹⁹⁸ Site-notebooks 22/2/58

¹⁹⁹ See Appendix.

²⁰⁰ For an alternative reconstruction, see Section I E(i). 3. The only domestic, non-palatial examples with a two-columned porch are attested at Ag. Kosmas, Houses S, T (Mylonas, G. E., Aghios Kosmas: An Early Bronze Age Settlement and Cemetery in Attica (Princeton, 1959), 53, 55 and figs. 14–15)Google Scholar. The HofCols inside the acropolis is a slightly different case because the columns were not placed between antae (Wace 1949, 92).

²⁰¹ Verdelis 1958, 163; 1963, 15. For objections to this theory, see Tournavitou, 1995, 5–16. The alleged incompatibility of the central hearth and chimney opening (terracotta pipe) combination with a pitched roof or with a second floor over a three-room unit, as suggested by Mylonas-Shear 1987, 10, 145, is not architecturally valid and should not be used as an argument in favour of a flat roof or against the existence of an upper floor (Darcque 2005, 125–6).

²⁰² Perfectly adequate thickness of load-bearing walls, manageable distance between lateral walls, structural independence from the west wing.

²⁰³ The resulting difference in levels between the two units, especially in the area of the north-west terrace and Room 5, would according to Xypnitou render the drainage of rain water from the west wing very problematic.

²⁰⁴ According to Xypnitou, the location of the cuttings associated with the framework/support system of the wooden staircase along the east side of the north-west terrace could only serve a staircase aiming no higher than the actual terrace level and not one intended to reach the level of a second storey. One could of course disregard this note of caution if the cuttings were not really associated with the support system of the suggested staircase or if the frame/support system of the upper level of the suggested staircase did in fact utilize a purely conjectural northward continuation of the north–south wall discovered along the east edge of the terrace further south.

²⁰⁵ Built remains and the lack of evidence for the continuation of the east wall south of ‘wall’ B.

²⁰⁶ The east face of the patch of pavement designated Γα by Verdelis is remarkably, and probably deliberately, straight, as if designed to represent some kind of physical limit. Only in the 1959 reconstruction of the house (Fig. 13 a) is such an option left open.

²⁰⁷ For more information on the character of the deposit see Section I. F(iv). The alternative is implied by Verdelis in the notebooks, but nevertheless not depicted on his earlier reconstruction of the building (Fig. 13 a). It appears only in his later version of the house (Fig. 13 b), which was reproduced by French (Fig. 13 c), i.e. an exterior east wall continuing in a straight line up to wall Γ, which would imply that this section of the wall had been completely taken apart in the HL period. Even so, it does not explain the awkwardness or the purpose of the section of pavement to the east of Γα (Γα), or the significance of the black layer to its south.

²⁰⁸ See Sections II and I. D.

²⁰⁹ Site-notebooks 28/2/58; Verdelis 1963, 25; Tournavitou 1995, 14. The same view was uncritically repeated by Darcque 2005, 358 in his brief discussion of the building.

²¹⁰ See Section II. F.

²¹¹ Mylonas-Shear 1987, 19–21, 23, 31–2, 146 n. 30, where the rooms in question (Rooms 7, 10), were located exactly behind the main rooms. In the case of House I (Room 7), the architectural arrangement is very similar in some aspects to that in the WH.

²¹² Unless of course it was used as some kind of treasury for light-weight precious materials, which were removed either before or after the destruction of the house.

²¹³ Although fragments of figured frescoes discovered in the north part of the narrow open-air court in the House of the Artisans were assigned to the upper floor (Mylonas 1995, 91, 94), the existence of the latter in that building has been established beyond doubt.

²¹⁴ See Section II A–C, F, G.

²¹⁵ The dimensions of the slabs, are also not specified. The excavator mentions these slabs only in connection with Room 1 and the section of the corridor opposite Room 5. Although I agree with Darcque 2005, 124, 128 and Küpper, M. (Mykenische Architektur: Material, Bearbeitungstechnik, Konstruktion und Erscheinungsbild (Internationale Archäologie 25), 107)Google Scholar, that the total disappearance/destruction/re-use of the roof-tiles of a building after its destruction is not very likely considering the large number of tiles involved, I do not think that the presence or absence or roof-tiles should be the only determining criterion for the existence of a pitched roof, overriding the architectural parameters involved.

²¹⁶ Also suggested by the excavator (Verdelis 1963, 27). The colonnade would use the west edge of the court pavement as a footing. The absence of stone column bases in this area of the house is not that surprising, considering the degree of post-Mycenaean disturbance.

²¹⁷ Hiesel 1990, 225, drawing on palatial examples, suggests that in Mycenaean architecture stoas/porches lined with columns on one side functioned as corridors.

²¹⁸ Treuil, R., Le neolithique et le bronze ancien égéens: Les problems stratigraphiques et chronologiques, les techniques, les homes (BEFAR 248; Paris, 1983), 257Google Scholar.

²¹⁹ 1987, 9–11, 153.

²²⁰ Küpper (n. 215), 110.

²²¹ Darcque 2005, 124.

²²² Considering the variation in the form of the extant Mycenaean buildings, the existence of more than one type of roof (flat, double-pitched, single-pitched) seems more or less certain. It has been rightly suggested (Darcque 2005, 126) that double pitched roofs are more suited to isolated rectangular/apsidal buildings with a relatively simple plan. Hiesel 1990, 221–5, who discusses both options of flat and pitched roofs for the Mycenaean period, also suggests that flat roofs, a Minoan innovation, were used for more complicated ground plans (palaces, corridor houses), and admits to the existence of single pitched roofs for certain buildings (ibid. 22–3), while Iakovidis, S. on the other hand (‘Mycenaean Roofs: form and construction’, in Darcque, P. and Treuil, R. (eds), L'Habitat égéen préhislorique (BCH supp. 19; Paris, 1990), 147–60 at 147–52)Google Scholar, argues for the exclusive use of pitched roofs.

²²³ Pent roof (Iakovidis ibid. 160), ‘anchored on a pole or beam bedded on the top of the rear wall, and on a similar beam crowning the lower front wall’. The rainwater from the west wing would end up in the space between the two roof-units, and could be easily channelled away to the south. A single pitched roof sloping down towards the west would theoretically represent another option, but it might be problematical with regard to the drainage of the rain water from the west wing.

²²⁴ Mylonas-Shear 1987, 145. In this case, the east wing would again constitute a separate unit, as would the first three rooms of the west wing, Room 4, Room 5 and the north-west terrace. The roof over the latter would of course be at a higher level than the roofs over the rest of the house. It should be stressed at this point that most of the arguments presented by Mylonas-Shear loc. cit. against sloping roofs (close juxtaposition of houses, party walls, problems with the drainage of rainwater, are not really applicable in this case.

²²⁵ Tournavitou 1995, 285.

²²⁶ Hiesel 1990, 111–12.

²²⁷ Tournavitou, I., “Ιδιωτικὸς βὶος και κρατικὸς τομὲας στο τὲλος της υοτὲρης εποχὴς του χαλκοὺ στις Μυκὶνες -Ηὰποψη της αρζιτεκτονικὴς”, Ariadne, 8 (1996), 31–64Google Scholar; ead., “Οι τοιχογραφι1ες σαν αντικεὶμενα κὺρους στην ΥΕ ΙΙΙ σποχη1 Ηπερὶπτωσι της Αργολὶδας”, in Eliten in der Bronzezeit: Ergebnisse zweier Kolloquien in Mainz und Athen (Monographien des Romisch-Germanischen Zentralmuseums, 43; Mainz, 1999), 119–28Google Scholar.

²²⁸ Tournavitou 1995, 286, n. 67.

²²⁹ I agree with Hiesel 1990, 205–9 about the close relationship between town houses and palatial architecture in big palatial centres, as well as in the close conceptual connection between the ‘Corridor-Haus’ and the layout of the palatial units, which can be traced back to progressive changes in LH society.

²³⁰ Hiesel 1990, 128–9, 141–2.

²³¹ Ibid., 129, 141.

²³² Ibid., 1990, 141. According to Hiesel, access to the living quarters would be provided through a staircase (itself conjectural) situated in the courtyard.

²³³ Darcque 2005, 358–9.

²³⁴ He maintains (loc. cit.) that the bad state of preservation of the east wing makes study of the function of the complex impossible, that the south part of the east wing remains very obscure, that the southern limit of the building possibly corresponds to the southern limit of the east wing, that the two wings of the WH appear unequal, that the difference in dimensions amongst the rooms in the two wings probably reflects their function, and that the two reconstructions of the building (presumably by Verdelis and French; FIG. 13), illustrate larger sections of the walls than what is preserved. He further mentions Verdelis's theory about the staircase, without evaluating it. His statement that the dimensions of the building, the placement of the walls, and the overall plan do in fact respond to needs beyond those of everyday life (Darcque 2005, 357) is thankfully true. Finally, his published plan of the house (plan 28) includes a number of mistakes/omissions (in Room 1 he omits the west part of the north wall, he does not depict ‘wall Α’, wall Γ, pavements Γα and Γβ).

²³⁵ Darcque 2005, 325–6. The WH happens to be the smallest building in the group and lies right on the borderline as regards the number and the size of the rooms attributed to this group. It also seems that it is included among the buildings with a second floor (ibid. 330). Of the two criteria used as evidence of the use of advanced construction techniques (monolithic thresholds and fresco decoration), a limited number of criteria in my opinion, the WH posseses only the latter.

²³⁶ Darcque 2005, 339, 357.

²³⁷ His primary classification of Mycenaean buildings, based on the exhaustive application of a series of more or less objective architectural criteria, such as the overall dimensions, the number of extant rooms, the existence of a second floor, the size of the main rooms, the use of complex building and decorative techniques, the presence of a ‘principal stereotyped unit’ and evidence of administrative activities (Darcque 2005, 315), although methodologically sound, has already been tested by me, albeit within the limited area of Mycenae (Toumavitou 1996 (n. 227), 31–64; 1999 (n. 227), 119–28), with very satisfactory results as regards the social and economic stratification of Mycenaean society. Although his final classification of Mycenaean buildings into three categories, palaces, intermediary buildings and houses (Darcque 2005, 336–40), is also quite satisfactory, my own research in the area of Mycenae suggests even further and finer distinctions not made by Darcque. Having said that, I still cannot see the relevance of his passionate objections concerning the difference in status between the houses included in the corpus of buildings used by Mylonas-Shear (1987) in her study of Mycenaean domestic architecture: firstly, his own corpus of the twenty-five intermediary buildings is not fully consistent with his own criteria; secondly, the buildings do not all belong to the same functional category, their only common ground being that they are all located within the acropolis walls; thirdly, this has in my opinion nothing to do with the object of the exercise as understood by Mylonas-Shear (1987), who pursued a study of the underlying architectural concept through architectural form and not through function. Finally, I should like to add a note of caution concerning the use of partly excavated and basement-only buildings in the various classification schemes, a dangerous habit which lends itself to false and occasionally prejudiced interpretations.

²³⁸ Darcque 2005, 339, 357.

²³⁹ Ibid., 332.

²⁴⁰ In the case of the HSh the absence of a domestic unit is directly related to the more specialized function of the building (Tournavitou 1995, 287–9).

²⁴¹ Tournavitou 1995, 262–4, 286.

²⁴² Its position in respect to the other three buildings in the group reflects the close ties, physical and administrative, also attested in the Linear B documents (Tournavitou 1995, 285–92).

²⁴³ The post-Mycenaean features will be discussed in the order followed in the main text.