1.1. Available information and degree of preservation

4Appendix 11-1 summarises the degree of burying, the shape, measurements and the principal features and arrangements observed for the 38 excavated buildings of the Central, B and F Sectors, including the four successive phases of the Communal building. The buildings in the road profile (chap. 2, appendix 2-2) and the 2009 test excavations on the upper terrace (chap. 6, appendix 6-1) were not taken into consideration because they were not fully excavated. The 38 buildings were unequally distributed between the Central Sector (N = 8, including the southeast area), the B (N = 27) and the F (N = 3) Sectors. They were also very disparately documented as a result of different degrees of preservation (and excavation for four of them: B20, B22, B906, B907). Nearly half (N = 18, 47%) were so eroded that only a few rounded foundation trench arcs subsisted, a metre long, associated with scant or no preserved internal features.

1.2. Shape of the basal outline

5The complete or almost complete outline of the building was only preserved for eight buildings (21%; “Buildings 1 and 2”, B05, B07, B08, B09, B18, B28, B800; fig. 11-1). Five of them describe an almost perfect circle but three are clearly ellipsoid to oval-shaped with either a N-S (B08, B28) or E-W (B18) major axis. If we consider the six other buildings with more than 2/3^rd of their (half) outline preserved, we find approximately the same proportion of circular (“Building 3”, B24, B906, B907) and ellipsoid (B10, B29, both with N-S major axis) base outlines. In total, four out of nine well-characterised outlines are circular and five are ellipsoidal/oval-shaped (4 N-S, 1 E-W). If we accept the hypothesis of successive versions of the Communal building, this makes seven circular versus five ellipsoidal/oval-shaped buildings, which means that we cannot estimate the shape of any of the 22 buildings with only 1/3^rd or less of their upslope outline preserved, that is to say 67% of the recognised buildings. We can only suggest that all the buildings describe a rounded upslope profile, except for B29, which displays a more “angular” shape.

Fig. 11-1 – Synoptic and schematic representation of the shape, size, orientation and main internal features of 35 excavated PPNA buildings at Klimonas, according to data in chap. 5 to 8 and appendix 11-1.

1.3. Sizes, surfaces, volumes; functions

6Unsurprisingly, based on the measurements or estimated sizes of the buildings in appendix 11-1, we observed two groups of buildings, the communal ones, at 78.5 m², and the rest, with a median size of 14.8 m² (i.e., a useable surface of 13 m² and a mean diameter of 4.35 m; fig. 11-2, A). The size distribution of the 30 residential buildings did not significantly differ from normal distribution (Shapiro-Wilk W = 0.95, p > 0,1), but the kernel density distribution suggests two groups (fig. 11-2, A) that are weakly supported by the mixture analyses (Akaike criterium = 151 for one group, 149 for two groups). Two-thirds of the buildings (63%) cluster in the small-sized group (mean = 11.9 ± 3,6 m²), and the other 10 buildings (24.6 ± 3.2 m²) are disseminated across the different excavated areas: Sector F (B800, B906, B907), Sector B (B03, B10, B13, B14, B18) and the southeast area of the Central Sector (B20, B23, B39).

Fig. 11-2 – A, Histogram and kernel density of the 33 building surfaces which could be measured or estimated (on the assumption of circular buildings for those where the downslope part was missing). The models of the two groups resulting from the mixture analysis were calculated on the basis of the 30 residential buildings only. B, Histogram and kernel density of upslope digging depth estimates, including successive versions of the Communal building.

7In Sector B, there was no correlation between building size and distance from the Communal building (Welch’s t-test unequal var. = 0.7343, df = 7.562, p = 0.5). It is clear, however, that the three constructions found in Sector F, located much further away from the Communal building, are significantly larger than the average size of the buildings in Sector B (and present a more systematically circular shape).

Fig. 11-3 – A, Distribution of buildings according to their diameter or to their main and small elliptical axes, and the correlation between the size and presence of a hearth pit, a cupmark, a “mosaic” pit and a bench. B, Notched box-plot distribution of the (estimated) total area of the buildings from the three groups visible in A.

8Fig. 11-3 shows that oval-shaped buildings are represented in all size ranges. There was no significant correlation between building size or shape and the presence of a hearth pit, except that the two buildings without hearth pits were also the smallest ones and they were both located on the same terrace (B05, B24). However, no “mosaic” pits have been found in the middle or large residential buildings: they were only observed in the very small (B02, B17, B28) or very large (Communal) buildings. The two buildings containing cupmarks (B28, B29) were also among the smallest ones (10 and 12 m²). These were also the only two buildings with an upper foundation trench forming an ogive shape, without upslope digging. The absence of upslope digging prior to construction was also noted for the very small building B24 (4 m²). It seems that the nine smallest buildings, where all the “mosaic” pits and cupmarks were found, constitute a rather coherent group of buildings (group 1) built for craft or food activities rather than habitation. However, we do not yet know exactly what these artisanal functions were.

9At the other end of the size gradient for residential buildings, we observe another group of 12 large buildings (group 3: 22-26 m²), five of which contained earthen bench(es). Half of them comprised a hearth pit, but the other six were too eroded to determine whether or not they contained a hearth pit. No specific technical features were observed, except in B18, where we found a unique deep mortar similar to the ones interpreted as fermented beverage mortars in the South Levantine PPNA (Liu et al. 2018). This group only contained dwelling buildings (“houses”).

10In between, we identified another putative group of nine medium-sized round or elliptic buildings (group 2: 14-16 m², fig. 11-3, B). Six of them included a hearth pit, the rest were too eroded. This group was included with the group 1 “small-sized” buildings by our mixture analysis (fig. 11-2). If we admit that these medium-sized buildings constitute a coherent group (and not a mixture of groups 1 and 3), they would appear to be more similar to the small-sized building group than to the large residential buildings, and they were also probably used for various technical activities.

11These observations indicate a variety of buildings with different functions. Consequently, each social entity appeared to carry out activities in several buildings, or may even have owned several buildings.

1.4. Depth and excavated volumes

12Upslope digging depth estimates, including successive versions of the Communal building, also display a bimodal distribution (fig. 11-2, B). The group with very little upslope excavation corresponds to second-generation buildings set on the rubble floor of an already existing building, either within the Communal building (“Buildings 2 and 3” benefited from the excavation of “Building 1”), or in Sectors B and F (e.g., B807 on B806, B14 and B29 on B13, B28 on B18). Consequently, it was not necessary to excavate more than 1-2 m³ for 11 of the 19 buildings for which we could estimate the depth and corresponding volume (appendix 11-1). Conversely, for the initial development of the other eight buildings, 70 m³ had to excavated from the Communal building and between 4 and 7.5 m³ (B800) in the residential areas (fig. 11-4).

Fig. 11-4 – Comparison of the excavated volumes (pink areas) for building the semi-buried Communal building (A) and a residential building (B800, B) notched and terraced in the slope.

1.5. Partitions and bench seats

13It was often difficult to detect partition walls or benches as remains were discrete and poorly preserved: shallow foundation excavation (B10, B18), narrow foundation trench (B800), line of small stake holes (B03, B800) or post holes (“Buildings 1 and 2”). It was sometimes even difficult to distinguish partitions from benches.

14We detected assumed radial partitions in each of the first two phases of the Communal building, and one tangent in B800 (fig. 11-1). We found well characterised earthen benches in the first three phases of the Communal building and in B10 and B800, and what we assumed were benches in B04, B14 and B18. But we can be certain that there were no benches in B5 and B24. Micromorphological investigations evidenced that the material used for making the Communal building benches was more carefully sorted than for the B10 benches (chap. 9).

15Even though the number of detected partition walls and benches was probably under evaluated, it seems that they were either exceptional or only present in a small number of buildings.

1.6. Entrances

16Fourteen of the 20 well-preserved buildings (more than a third) yielded evidence of a lateral entrance. They were oriented to the northeast, i.e., opposite the dominant western winds, on the higher parts of the site (Communal building, B800, St 23, B18, B28), and to the southeast in the west and central areas of Sector B (B04, B05, B07, B08, B09, B10, B24; appendix 11-1, fig. 11-1). The only exception to this is B800, with an opening to the south, in addition to a northeast entrance.

17As the building floor levels were always more or less beneath the external circulation level, these entrances were all above the floor level, especially if they were located to the northeast, where the drop was higher because of the slope of the natural terraces.

18These entrances were characterised by the interruption of the foundation trench (except B07, B10, B18). This interruption was the only (preserved?) feature for the two superposed small buildings B05 and B24. For the other 12, the entrance structures were more or less sophisticated (and well preserved). Some consisted of only two post holes indicating wooden doorframes (B07). A more sophisticated doorframe was found in St 23, where the post was associated with a return of the foundation trench, setting the door slightly back under the roof in order to protect it from the weather. A similar system was better preserved in B08, where the doorframe was composed of paired posts on each side of the door, which, in addition, opened onto a stone step. We assumed that several stone steps descended into B08 because of the probable high drop at that time between the external circulation level and the building floor. Two other buildings contained a stone threshold (B10) or the imprint of a stone step (B09). In the successive versions of the Communal building, B18 and B800, an earthen ramp (slope or steps) replaced the stone step. The most complex structure was the narrow corridor found in phase 2 (and probably 3) of the Communal building, consisting of a frame made up of four posts and covered by a roof. However, it seems that such a complex system was not exclusively reserved for the Communal building, since a similar corridor might have been associated with the small B28 and maybe the stairs of B08.

19With the exception of the 3-m-wide entrance of “Building 1“, the entrance width for all the measurable 10 buildings was comprised between 70 and 100 cm.

2.1. Building materials

20The building materials were primarily building earth and wood, and secondarily stones.

21The walls of the Communal and domestic buildings were made of earth (chap. 9). At the beginning of site occupation, the sediment was taken locally from the Pleistocene levels, either composed of carbonated colluviums or of pedological calcrete (havara). It was at least partly extracted from the digging performed for burying or notching the building in the slope. Later on, earthen materials from abandoned or destroyed building were reused several times, becoming richer in detrital and organic components, making them darker and darker through time (especially in the Communal building).

22Indeed, regardless of the origin of the sediments, the materials were prepared in a wet state by incorporating plant components, mostly cereal hulls represented in the form of phytoliths. This mixture was applied by direct shaping, sometimes by assembling sub-rounded elements, sometimes by beds of raw earth, using the massive earth construction technique called cob (de Chazelles 2008, Mylona et al. 2017a; chap. 9 and 22). Building earth was also used after mixing with limestone or sometimes flint stones of several centimetres and gravels for making substructures, either stony floors or embankments. Otherwise, well-sorted and fine earth with no stones was used for covering the floor or bench surfaces. The earth material used for the Communal building was on average richer in temper than that of the two residential buildings subjected to micromorphological analyses (B800, B10; chap. 9).

23Water and cereal hull temper were also important building material components. The presence of fresh water diatoms in the earthen material (chap. 9) suggests that it came from the Athiaki River, as well as from water reservoirs in the village itself, owing to systems of water drainage and tanks (see chap. 12). As in many other PPN sites, cereal hulls were probably carefully collected and conserved after harvesting (Willcox and Stordeur 2012, Parès 2017).

24The use of wood is attested by hundreds of stake, pole or post holes. Stakes and poles were on average implanted 15-30 cm deep, frequently in pairs (then probably bound with plant fibre ties), and wedged with small stones, macro-tool fragments or depleted flint cores. The thickest posts were located in the peristyle of the Communal building, at the southern periphery of the residential buildings, where they were used as doorframes or supports for partition walls, bench front edges (B800, B04) or other smaller features located inside the buildings. Poles and rods may also have been used for roof frames. Larger posts holes were found inside the buildings, especially in a central position, and particularly in B29, B800 and the three first phases of the Communal building, where they were implanted 50 to 70 cm deep and wedged by large and heavy stones. In these buildings, some of them were probably included in earthen pillars.

25Basing on the reconstructions of B800 (chap. 6) and the Communal building (chap. 5), we can estimate that some of the largest poles (8-15 cm thick) and the central posts (18-25 cm) were 1.8-2 m long. Charcoal and seed analysis shows that tree species were massively dominated by pistachio trees, secondarily olive trees and oak (chap 23). The latter was the best adapted species for making long and thick supporting posts, but did not appear to be abundant in the site’s surroundings. This is probably why several large post holes displayed clear marks of removal for reuse after demolition, especially in B800 and the Communal building.

26Stones were rare in the constructions of Klimonas in comparison to other PPN sites. They were only used for wedging poles and posts (or positioning them in one case: “Building 2”), for making the threshold or entrance steps (B08, B10), for cupmarks (B28, B29), for the unique buttress of B10 and possibly for stabilising the edge of the roofs (B07). However, stones were easily accessible in the immediate vicinity of the village, either in the Miocene limestone cliff of the upper terrace, or in the numerous outcrops of ultrabasic or compact Lefkhara limestone rocks along Throumbovounos hill, two minutes’ walk from the centre of the village. This parsimonious use of stones can therefore only be interpreted as a deliberate choice (see discussion in chap. 33).

2.2. Substratum, levelling embankments, substructures, floors

27We found no evidence of any PPNA construction built on the green/orange clays of the Moni formation regolith, an outcrop on the west of the Klimonas terraces (Sectors A and C, chap. 2). This may result either from the poor resistance of this substratum to erosion, making it impossible for any PPNA digging to subsist until now, or from the reluctance of PPNA populations to build on this clay because of its plasticity and sensitivity to the degree of wetness (desiccation cracks).

28All the buildings that we found were constructed on (from the softest to the most resistant substratum): (i) the grey-greenish Lefkhara marl regolith (western half of Sector B, chap. 3, fig. 3-3), (ii) carbonated Pleistocene colluviums (east of the Communal building, southeast area of the Central Sector, part of Sector F), (iii) the havara (centre and east of Sector B, part of Sector F, west half of the Communal building), or (iv) the dense vertical limestone banks of the Lefkhara formation (south part of Sector F). All these formations were resistant enough to support earthen wooden-framed buildings and to allow for the preservation of remains until now.

29All of these formations, except for the last one, were excavated for buildings buried or notched into the slope. The products of these excavations were used for making building earth, as already mentioned, but also sometimes for creating an embankment (fig. 11-4, B; B800, but also B04, B08, B10, B18; appendix 11-1) to raise the lower part of the notch so that the downslope part of the floor was at the same level as the upslope one (and to reduce the depth of the upslope excavation in order to save time and effort).

30The floors of the buildings lying on the havara were cautiously levelled then covered with a thin layer of plaster. The latter was sometimes coloured red/orange (B18, fig. 7-65, B). It was always impregnated by more recent pedological carbonated films that formed during the Holocene at the boundary of the havara and the overlying soft sediment. It was almost impossible to tell without micromorphological analyses if the floor was actually plastered or only covered with this natural deposit. Much thicker plasters or fine formal earthen floors were found, sometimes in the form of relict flaps, on most of the buildings on the grey-greenish marls (B5-24, B09, B17, B13-14, etc) or on the carbonated Pleistocene colluvium (“Building 1”, B800). It was also sometimes entirely absent (B07, B08), probably due to erosion.

31Buildings constructed on the same artificial terrace as the previous one generally reused the same floor as it was (e.g., B26, B28, B29), or sometimes after additional shallow excavation (B08, B24, B26, B907). However, surprisingly, several buildings were set on the soft earthen rubble of more or less completely destroyed earlier buildings, either across the entire surface (phases 2-4 of the Communal building, B12, B25) or only part of the surface (B10, B14). In these cases, we systematically observed the same technique consisting in laying a more or less thick (5 to 25 cm) and horizontal substructure (stony floor or embankment), covered by a more or less refined floor (informal or formal floor; chap. 5, 9).

32In all these cases, whatever the type of floor (plaster, informal, formal), we found frequent marks indicating refilling (Communal building; B10, fig. 7-45).

2.3. Roof and load-bearing systems

33Apart from some limestone platelets in the B07 demolition and primary fills, we did not find any unquestionable roof remains, possibly because only a small number of the buildings underwent fire damage (B09 and B28). However, pole and post hole analyses strongly suggest that most, if not all, the buildings described in chap 5 to 8 were covered with a roof, probably made of vegetal rods or poles incorporated into the building earth, as was usual in the PPN. We argue that these roofs were most likely flat and slightly sloping to the south (chap. 5 and 6). What architectonic solutions and systems did the Klimonas villagers use for sustainably supporting architecture made from cob and wood?

34For the small buildings in group 1 (fig. 11-3), the frequent absence of central poles (B24, B05, B07; tab. 11-1) and the lack or absence of partially burying the buildings (B24, B28, B29) may seem to indicate that they were not strong enough to support a roof but might have been suitable as simple enclosures used for technical activities, in the same way as some of the Asprokremnos buildings (McCartney 2017). However, this was not the case because all the buildings contained at least one roof-supporting feature: a central pole (B07, B29, the two largest ones), a more or less deep support due to the upslope digging ledge (B05, B07 and maybe B17), or a series of poles implanted in the peripheral foundation trench indicating wooden-framed earthen walls (B24, B02, B28 and maybe B29). The presence of a plastered soil in B24, B05, B28, B29 and probably B17 also points to covered buildings rather than simple enclosures. For the majority of small buildings with no central pole, the roof frame may have been composed of rods or sticks of a maximum length of 4 m (largest diameter of the building), but layers of crossed 1.5-2 m-long rods would also have provided good coverage (Stordeur 2015).

35We estimated the roof weight of the small buildings in the same way as for the Communal building (chap. 5), by multiplying the building area by roof thickness (estimated to be 7-10 cm) and by an earth density of 1.6 t/m³. The results fluctuated between 0.8 and 2 t (tab. 11-1), except for the miniscule B24 (450-640 kg). By dividing these estimates by the number of observed poles and the preserved area, we found that, on average, each of the columns supported a load comprised between 45 and 130-150 kg. These fall below or in the lower range of estimates for the Communal building (100-200 kg, chap. 5), in accordance with the smaller average size of the poles used in these buildings.

36It was not possible to apply the same calculation to the larger residential buildings as the structures used to support their roofs were more diversified. They consisted of five different types (see appendix 11-1).

• Peripheral poles in the upslope half of the foundation trenches were absent (B01, B03) or rare - less than three (B08, B09, B14, B26, B27). B04 (4-5 holes) and B15 (6-7 holes) were the exception, maybe because these two buildings were slightly buried. B10 and B13 also displayed more upslope peripheral pole holes (5-7), but they were probably used for supporting peripheral earthen benches rather than for framing the perimeter wall. The three large buildings with a preserved downslope portion clearly showed that peripheral poles were mostly used for supporting the southern half of the buildings. Each of them displayed different systems: an internal peristyle of paired poles for the Communal building, a series of regularly-spaced poles included in the peripheral earthen wall for B18 and a mixed system for B800, with a combination of posts included in the south wall and external wooden braces forming a small awning above the south entrance.
• Central posts existed in the first three phases of the Com-munal building and in five of the largest residential buildings (B09, St 23, B18, B800; and in the two largest technical buildings B07 and B29). We could not determine the presence or absence of a central pole for the vast majority of the largest buildings (N = 22 in groups 2 and 3) because their south and central parts were not preserved. However, supporting the roof of a more than 4-4.5 m building (12-15 m²) requires crossing tangential and radial rods, necessitating the presence of a central pole. It is therefore probable that most, if not all of the 22 incomplete buildings had a central pole.
•  Other internal load-bearing columns were identified in the northern half of the first three phases of the Communal building, in the form of earthen pillars set on stony bases and probably strengthened by wood. Nothing similar was found in the residential buildings.
• Earthen partition walls framed by lines of poles were found in the first two phases of the Communal building (radial) but also sporadically in some residential buildings: B07 (two probable light radial partitions), B800 and maybe B18 (tangent partitions). These might have contributed to supporting the roofs.
• Burying or notching the buildings in the slope might have provided a strong upslope roof support. It seems, however, that the depth of these excavations, estimated to be 100-120 cm at most with a median between 50 and 60 cm (fig. 11-2), was definitely not enough to lay the roof frame directly on the edge top. Short poles could have been implanted on this edge to raise the level up to 170-200 cm above the floor, but no such pole holes were found in this position. Therefore, we have to conclude that the upslope earthen wall, even though it was only slightly or not at all framed by wood, actually contributed to supporting at least the upslope limit of the roof. This was possible because most of the wall’s height leaned against the vertical limit of upslope digging, and because its load-bearing part was limited to less than one metre high.

37The strongest evidence for this interpretation was provided by structures built for large buildings that could not, or not sufficiently, lean against the upslope digging wall (chap. 7). This was the case for buildings B14 and B26 built on the terraces of previous buildings (respectively B13 and B27), the remains of which might have weakened the support of the upslope peripheral earthen wall. The builders adopted two different strategies to strengthen this support. For B26, they cautiously removed the remains of the previous building wall directly against the limit of the original digging in the west, and packed the demolition products from the previous building against the excavation in the east (fig. 7-73 and 7-74). For B13, they shifted the building 50 cm to the west compared to the previous one, by re-excavating the slope (fig. 7-48), in order to lean the upslope wall against a new cleaner and stronger limit. Even more convincing is the case of B10, a large building for which the builders constructed a several-metre-cubed buttress made of stones and building earth (St 6780, fig. 7-39) to prolong the excavation to the west, where the deep digging of building B08 prevented them from supporting the wall on a digging ledge.

38In addition, we must consider that the massive internal earthen benches backed against the northern walls in B10, B18 and B800 contributed to strengthening them.

2.4. South earthen walls and foundation trenches

39The north peripheral earthen wall could therefore have played an architectonic role in supporting the roof, but it seems that the southern halves were too fragile to play the same role, except when they included poles. This was especially clear for the Communal building, where the southern half of the wall, founded in a shallow trench, was internally lined by peristyle paired poles. In this case, and also partly in B800, the role of the wall was mostly devoted to separating the internal from the external space, without any significant architectonic function.

40The discovery of several fragments of plaster coloured with red ochre suggests that some walls may have been coated and painted.

41The architectonic function of the foundation trenches should also be questioned. These were often less than 20 cm deep and 15-30 cm wide and therefore could not have played a major role in stabilising the earthen wall. This mirrors the situation 2000 years later in Shillourokambos (de Chazelle 2011, 2021). These foundation trenches seem to be the manifestation of a foundation ritual rather than an architectonic component. This interpretation is further strengthened by the important number of remarkable deposits found in them (chap. 13).

3.1. Artificial terraces for managing the dwelling space

42We have seen that each of the buildings was implanted on a terrace specifically dug for that purpose, on a surface corresponding exactly to that of the edifice. No overall development plan was implemented before the installation of the buildings, except probably the location and function of the surrounding buildings.

43However, in Sector B, at least nine buildings were built on the terrace of a previous edifice (B09, B12, B14, B19, B24, B25, B26, B28, B29), with little or no modification of the initial digging. In addition, together with the chrono-stratigraphic succession of the buildings (chap. 7), the analysis of the average altitudes of the foundation trenches brings to light the relative organisation of the altitudinal levels of construction (fig. 11-5). The lowest preserved level was represented by the very eroded low terrace at around +97 m (B01, B02). Then comes another very dismantled complex situated about 90 cm higher (altitude +97.9 m: B27, B26, B13, B14, B29, B17, B15, B23), then a better-preserved complex, situated 80 cm higher, on a gentler slope (+98.7 m: B04, B06, B05, B24, B09, B07, B08, B25). The high complex was located at the top of the preserved buildings (altitude +99.2 m: B03, B10, B28), to which we can undoubtedly add B20, which has not been excavated. The foundation trenches of the buildings of Sector St 10 are at +100.33 m and the base of St 10 is at +100.50/100.70 m. We can thus discern whether or not a -pre-established plan existed, at least with regard to a certain organisation around the terraced levels of the slope every 50-80 cm.

Fig. 11-5 – Altitudinal distribution of foundation trench bases from the Sector B buildings.

44Building on a pre-existing terrace followed varied patterns: the surface was reduced in relation to the previous construction, and the new edifice was built in the centre of the terrace (B24, B28, B29), moving slightly higher up the slope in order to create a new support for the upper wall (B19), widening the digging towards the west (B14) or towards the east (B08), or moving downslope (B09), sometimes leaning against the still solid rubble of the previous upper wall (B26, but also B14 and B29). Two buildings were installed on a thick fill left by previous edifices, which were undoubtedly destroyed then levelled (B12, B25). As we explained above, most of these architectural solutions denote the resolve to lean the upper part of the peripheral wall against a steep feature of a certain height.

3.2. Size of the village, number of buildings, demography

45The PPNA residential building remains found in the Central, B and F Sectors, in the 2009 test trenches and in the road profiles describe a total area of preservation of the PPNA village estimated at 5583 m², i.e., slightly more than half a hectare (fig. 11-6).

Fig. 11-6 – Sectors of Klimonas with remains of PPNA buildings (red) and delimitation of the preserved area of the village (light blue).

46This is probably an under-evaluation due to erosion of the initial village’s margins. Indeed, we already argued (chap. 6) that the village could probably have stretched to the north of the small upslope Miocene cliff (i.e., 80*16 = 1280 m² more). The road profiles also suggest that it extended to the other side of the road, at least to still relatively flat areas today overhanging the deep erosion gully (see chap. 2, fig. 2-2); that is an additional c. 2000 m². Conversely, we argued that the village might not have extended on the green-orange clays that constitute the substratum to the south and west (chap. 3, fig. 3-3). We can thus estimate that the maximum original area would have been less than one hectare (8863 m²; tab. 11-2).

47In Sector B, we found 27 often overlapping buildings over a surface of 450 m² (chap. 7). In the small Sector F trench, there were three buildings on 77 m², two of which overlapped (chap. 6). In the very eroded southeast area of the Central Sector (chap. 8), at least four residential building foundation trenches were identified over 85 m². We also found three hearth pits and one floor portion with no links to any of the above-mentioned buildings, which should be considered as evidence of four additional buildings (not counted in our total of 38 for the whole site). Even in the upper road profile (chap. 2, appendix 2-2), two overlapping buildings were observed. Wherever we excavated deep enough elsewhere within the village, delimitations were evidenced fig. 11-6, especially in the 2009 test trenches (168-3, 168-6, 168-7, 169-3) where we found PPNA buildings remains. Consequently, we hypothesise (tab. 11-2, B) that the maximal density of buildings across the site was the same as the cumulated density of the buildings found in Sectors B, F and in the southeast area of the Central Sector. That makes 38 buildings for a total of 600 m², i.e., a total number of buildings estimated between 64 and a maximum of 153, depending on the total village area, respectively 5583 or 8863 m².

48The proposed slope occupation scenario for the succession of buildings in Sector B, using the Harris matrix (chap. 7, fig. 7-78), is summarised in tab. 11-3. Taking into account the three best documented phases (1, 4 and 5), the proportion of dwelling buildings can be evaluated between 25 and 50%. We can therefore estimate the number of dwelling buildings and households between 16 and 77. Depending on the number of people per household (5 to 10, including children, this makes a population of a few hundred people (80 to 770; tab. 11-2).

Tab. 11-2 – Detail of the hypotheses and calculations for estimating the size of the village, the number of residential buildings and the population of Klimonas.

49Of course, these evaluations are based on a series of hypotheses and the result is thus very tenuous and questionable. However, we can bear in mind the order of magnitude of this estimation, that is to say a few dozen households and a few hundred people, probably less than 500.

50The radiocarbon measurements and their Bayesian processing indicated that the successive residential buildings of Sector B, as well as the four successive versions of the Communal building range in the same short time interval of 60 years maximum, c. 8800 cal BC (chap. 10), i.e., about three generations. Based on the above estimation of 80-770 people, we can estimate that between 240 and 2310 people lived on the site during the total duration of the village. However, in our scenarios for Sector B, the number of buildings seemed to fluctuate during the different phases. It should be assumed, therefore, that the village and its population also fluctuated through time.

4.1. Size and duration of the village

51The size of the village can thus be estimated at between 5500 and 9000 m², i.e., several dozen households and a few hundred people, probably less than 500, making it the largest known PPNA village in Cyprus. The only other site, Asprokremnos, comprised a much smaller number of buildings and working spaces, and was probably not a permanent village (McCartney et al. 2008, Manning et al. 2010, McCartney 2017). On another hand, compared to the preserved surface of the PPNB site of Shillourokambos (c. 4100 m²), Klimonas also appears to be a quite large village, even though the former was highly eroded on its west and south edges (Guilaine et al. 2011, 2021). Conversely, the 7^th millennium Khirokitia site was much larger than Klimonas, extending over a surface of 3 ha (Le Brun and Daune-Le Brun 2009).

52Compared to continental PPNA sites, Klimonas appears to be much smaller than the largest ones, which were between 1 and 2.5 ha (Çayönü Tepesi, Jericho, Gilgal I Netiv Hagdud; Noy 1989, Bar Yosef et al. 1991, Aurenche and Kozłowski 1999, Goring-Morris and Belfer-Cohen 2011), or even more (Göbeckli Tepe, Schmidt 2000). The Klimonas area, however, falls within the same range of values as Hatula (3000 m², Lechevallier et al. 1989), Jerf El-Ahmar (1 ha for the two villages, Stordeur 2014), Körtik Tepe (5500 m², Özkaya and Coşkun 2009), Nemrik (4500 m²; Kozłowski 1989), M’lefaat (0.6 ha; Kozłowski et al. 1998), Hallan Çemi (0.7 ha; Rosenberg et al. 1998) or Qermez Dere (7600 m², Watkins 1990). Klimonas was therefore a rather large PPNA village, comparable to most of the Mureybetian or Soltanian villages. It is even slightly above the mean settlement size for the beginning of the ninth millennium, as modelled by Kuijt (2000, p. 83, fig. 2).

53Another important characteristic of Klimonas is the short duration of the site (maximum 60 years). This aspect of the site cannot be usefully compared with the continent due to the lack of data. Site duration provides invaluable information on the degree of mobility or sedentariness, a question that will be discussed in the concluding chapter of this volume (chap. 33).

4.2. Communal building

54The 10-m-large Communal building St 10 was described in detail and compared with its continental counterparts in chap. 5. It should be recalled here that it shared many characteristics with continental PPNA communal buildings, especially those of the northern Levant. However, it showed a larger than average diameter, a lower burying depth and thinner walls, was only made of cob and was built without any stones. These specific features suggest greater architectural fragility than on the mainland. They could be explained by the exacerbated importance of the destruction/reconstruction-renovation ritual, as illustrated by the superimposition of four different buildings in the initial 10-m²-ground-surface area. This resulted in lighter buildings in order to save efforts in construction and destruction.

4.3. Residential buildings

55The residential buildings were characterised by:

1. an almost systematic but moderate (no burying) digging of the upslope slope for terracing the building, with the downslope part heightened by an embankment in order to lay a well-levelled horizontal floor across its surface without excavating too deep;
2. a circular or ellipsoid/oval-shaped outline, with no clear angulation;
3. a ground area comprised between 4 and 30 m², with possibly two or even three group sizes (8-10, 14-16 and 22-26 m²);
4. the presence of peripheral earthen benches in the largest buildings;
5. very little or no internal partitions and no multicellular buildings, but virtual groups of different kinds of (technical or dwelling) buildings in the same unit;
6. rather sophisticated lateral entrances opposite the dominant winds, sometimes with steps, ramps or roofed corridors;
7. the use of building earth made from the local sediments, tempered with cereal hulls, and applied by direct shaping (cob), with wooden stakes, poles and sometimes posts;
8. stone was only used occasionally for wedging poles, for thresholds or steps, cupmarks and a unique massive buttress (B10);
9. informal or formal floors directly laid on the most compact substratum or, when the substratum was soft, on thick substructures (stony floors or embankments);
10. frequent reuse with little or no modification of terraces created for previous buildings, though the upslope wall always leant as securely as possible on the upslope vertical limit of the notch;
11. no pre-established terraces or plan for the spatial organisation of buildings, but, due to the frequent reuse of terraces, an a posteriori organisation of buildings according to informal stepped levels;
12. the systematic presence of light (and probably flat) roofs, even for the technical buildings, which were supported by peripheral poles included in the earthen walls for the smaller buildings. For middle and large buildings, the roof load was distributed over several structures, including a central pole. The upslope earthen wall was strengthened by leaning against the upslope vertical digging, and some walls contained downslope poles;

56The shape, size, simplicity of the general organisation and rather low variability were very similar to the only other PPNA site found in Cyprus (Asprokremnos, Manning et al. 2010, McCartney 2017) and to observations for the Early PPNA everywhere in the Levant and Southeast Anatolia, in line with the Natufian tradition. They appear, however, to be very different to constructions in these areas from the end of the tenth millennium onwards, with more angular or even quadrangular residential buildings, and pluricellular buildings, either by the agglomeration of several cells or by partitioning, at least in the north (Kozłowski 1989, Aurenche and Kozlowski 1999, Ibañez 2008, Özkaya and Coşkun 2009, Abbes 2014, Stordeur 2014, 2015). Even though the same tendency was observed to a lesser degree in contemporaneous late PPNA sites in the South Levant (e.g., Nahal Oren, Noy et al. 1973; Netiv Hagdud, Bar Yosef et al. 1991), more similarities could be found with the Jordanian site of Wadi Faynan (Mithen et al. 2011). We reached a similar conclusion based on analyses of the characteristics of the Communal building (chap. 5).

57Lateral entrances with ramps or stone steps have been found here and there in different Levantine PPNA sites, but rarely as systematically as in Klimonas. This structural feature may thus be another local characteristic. The absence of cob and the very limited use of stone, in spite of the presence of abundant stone materials near the site, also seem to differentiate Klimonas from most of the continental PPNA sites, even later ones. The situation is the same at Aprokremnos (McCartney 2017), whereas both stones and building earth were used at Shillourokambos from the early phase onwards, i.e., a couple of centuries after Klimonas. The absence of adobe and stones may explain the short life of buildings, estimated at less than 10 years according to the Bayesian processing of radiocarbon dates.