The EB-MB Transition at Tell Afis: a Reappraisal

The passage from the 3rd to the 2nd millennium BC in Inner Syria is still a matter for debate both in historical and archaeological studies. This is only partially due to lack of evidence: it is much the way one interprets the extant evidence that affects the discourse, emphasizing a break or a transition, and viewing this time as a period of crisis or not. In this article relevant archaeological evidence from the site of Tell Afis is revised and some new data, including 14C dates, are offered in order to resume discussion on the question. Focus will be specifically on pottery: by recovering various strands of development in the local pottery repertoires from the Early Bronze to the Middle Bronze, the article aims to contribute to the reconstruction of this archaeological phase at a regional level as well.


INTRODUCTION
On November 29 th , 2018, a small group of scholars gathered at the University of Florence to confront and discuss the pottery repertoire of inner Syria between the second half of the 3 rd to the first half of the 2 nd millennium BC. The title of the meeting -From the kingdom of Ebla to the kingdom of Yamkhad: data from the pottery repertoire -made the deliberate choice to refer to the kingdoms of Ebla and Yamkhad, two historical and geographical entities at the same time (taken, admittedly with some liberty), to provide both the spatial and temporal scope for the enquiry and stimulate discussion about possible political and historical implications. 1 The sites involved, Afis, Ebla, Shiyukh Tahtani, Tuqan, Tilmen, Zincirli, represented just a small but significant selection of settlements known for the period of the Early and Middle Bronze Ages in the region which could offer data to start a real engagement with the evidence 2 and eventually set the stage for a larger debate involving a greater number of sites. 3 The inclusion of sites such as Shiyukh Tahtani, Tilmen, Zincirli was meant as an attempt to extend our inquiry even beyond the possible extents of the above mentioned kingdoms, in order to explore connections to the north and the east and counterbalance the more heavily exploited, west-(i.e. 'Amuq) and southoriented (i.e. Hama), perspectives.
From the point of view of pottery traditions and assemblages, it has to be said that some phases within that lapse of time, such as the EB IVB and full MB, are relatively well established, both at a local and at a regional level. The definition of the passage from one period to the other remains more uncertain. 4 The question of the transition between the Early Bronze and Middle Bronze has been the subject of discussion, especially during the late '90s and early 2000s; in relation to the site of Tell Afis, in particular, the question has been addressed in a preliminary way on a number of occasions, especially focusing on materials associated or next to a large pottery kiln uncovered at the site, kiln A (Felli, Merluzzi 2002;Mazzoni, Felli 2007). The reason for resuming the issue here lies in the need for an update in light of the most recently published excavation reports with data from the last excavations in Syria 2007Syria -2010. It is also in response to the stimulus provided by a number of focused articles on the topic on behalf of scholars directly engaged in the field in the area (inter alii Iamoni 2014;Schwartz 2017;D'Andrea 2018a, b;2019a, b).
In the following pages I shall present the evidence from Tell Afis, integrating the most recent excavations and the results of relevant 14C analyses, in an effort to provide a better definition of the assemblage characterizing the passage between the 3 rd and the 2 nd millennium at the site. On the occasion of the workshop I circulated a list of individual ceramic types which in my view are significant because their period of attestation bridges -at Afis at least -the crucial period under examination. Emphasis will be thus on morphology: it is hoped that these data will stimulate further investigations of ceramic repertoires from other sites.

FOCUS ON THE TRANSITION: THE ARCHAEOLOGICAL EVIDENCE OF TELL AFIS
Tell Afis is located in the agriculturally rich Jazr plain of northwestern Syria (Fig. 1). It is just north of Ebla, certainly the main regional centre during both the EB and the MB periods. Perhaps on account of its lesser role, Afis did not seem to experience a disruption at the end of the EB period like that postulated for Ebla (see Matthiae 2020;2021: 18, 145). Nevertheless, Afis allows for a diachronic perspective needed to appreciate possible settlement discontinuity (Schwartz 2007: 48) since it provides a long stratigraphic sequence spanning the period under enquiry and the periods before and after. The situation is not ideal from the point of view of sampling, however: most information from both the EB and MB periods derive from peripheral areas of the settlement, and thus no data are presently available on the urban layout in the central part in these phases. We know quite well that the MB settlement was as large as the actual tell, composed of an acropolis and a lower town (Mazzoni 2005: 9). The size of the EB settlement is less clear, since 3 rd millennium levels have only been substantially exposed in area E3, on the western side of the acropolis mound. The most recent excavations in area N on the eastern slope of the acropolis mound, however, have provided evidence suggesting that the 3 rd millennium acropolis was at least as wide as it is nowadays (Fig. 2). 5 Moreover, a tannur with associated EB IV materials retrieved in excavations in area B located in the northern part of the fortifications of the Lower Town may be evidence for the existence of a lower town already in this period (Mazzoni 2002a: 15, fig. 6.11-14). Relying on a minimum scale, i.e. the excavated area 2 See the most welcome articles by Marta D' Andrea and Agnese Vacca in this volume. 3 A first outcome in this direction has been the organization of a workshop at the 12th ICAANE of Bologna in 2021 by the title "Entangled Narratives: Regional Developments and Inter-Regional Connections during the 3rd Millennium BC in the Northern Levant" along with A. Vacca,M. D' Andrea and G. Mouammar. 4 For a quite recent overview on excavations of EB sites see Ascalone, D' Andrea 2013; for a general synthesis on the MB period see last Morandi 2013. 5 See Affanni, di Michele 2007: 11, 12. E3, we realize that the existing data could be changed by future investigations in other parts of the tell. However, assuming that the history of a single area is significant per se, the resulting picture of continuity allows at least to argue reasonably that no break in the occupation at the site took place in the period in question.
Knowledge about the history of occupation of this part of the acropolis in the EB age (Afis IV) has been greatly expanded thanks to excavations in the area in the years 2006-2010 (Fig. 3). Now an articulated sequence spanning at least EB IVB-EB III has been brought to light in the northern portion of the area (Mazzoni 2013a: 95;Mazzoni in press). The industrial character of the area, already revealed by the excavations of a large EB-MB working area connected to pottery manufacturing (Fig. 4) and food production activities in the southern portion of area E3, is now verified for the earlier phases as well, thus confirming earlier suspicions (Felli, Mazzoni 2007: 212;Grifoni, Spinazzi Lucchesi in press). Investigations in both sectors of the area in the last campaigns at the site have provided a better understanding of its wide extension and complex articulation. The sequence of events in the area can be reconstructed as follows: the erection of the fortification wall M. 1115 in the MB period (Afis V) erased part of the stratification underneath, initially preventing the recovery of traces of occupation dated between the latest EB level reached (level 17a: Giannessi 1998: 103, figs 6, the section above, 11 ;Mazzoni 1998: 32, 33) and the MB rooms brought to light to the east of the fortification wall, most likely part of the contemporary MB settlement (L. 1110(L. , 1120(L. and 1122(L. : Giannessi 1998. On the basis of the material recovered, this latter settlement should be dated to the end of MB I or beginning of MB II, at the latest (Mazzoni 1998: 35). It is impossible to say when the earlier city wall M. 1117 was built, but apparently its construction cut one of the last floors of the EB IVB rooms (Giannessi 1998: figs 6, section above, 11), thus indicating at least a slightly later date for its construction. Presumably the wall was still functioning until a change in political circumstances possibly urged a new construction. In this outer part of the acropolis, a domestic quarter with open-areas for industrial installations linked to pottery production and food preparation continued through the end of the third millennium and into the beginning of the following second millennium (Afis IV-V). 6 As already pointed out, at some point in time, activities ceased in the working area and it was abandoned. With the erection of wall 1115 and a further stretch to the south (Afis V), the space immediately inside was given a different arrangement (Felli 2000: 12, figs 7, 8.1;Felli, Merluzzi 2002: 25, fig. 16;2008: 103, fig. 10). These works were part of a large initiative to strenghten the whole fortification system of the town, also attested on the opposite part of the tell, in area N (Affanni, Di Michele 2007). This construction probably took place around the same time that the lower town wall was built, this indicated by the dating of the graves found both beneath and above the rampart in area B, mostly ascribed to MB I, apart from one dated to the initial MB II (Repiccioli, Giannessi, Aletta 2002: figs 4, 5.2;Mazzoni 2002a: 14, fig. 6.1-4;Aletta 2005: 28, 31).
The results of the most recent excavations underneath the "transitional" structures (Felli in press) have reached lower levels which are characterized by materials which correspond well with the local EV IVB repertoire. Such findings confirm the existence of a separate phase on its own.
In particular, an earlier firing installation dug underneath kiln A itself ( Fig. 5.1) has provided a quantity of late EB IVB materials, including Smeared Wash sherds, a triangular ledge rim pot and a rare example of a Simple Ware collared bowl . The latter, with a smooth carination and probable plain flaring rim (Fig. 5.4) is most likely an early attestation of the later MB(?) type (cfr Ansari: Suleiman, Gritsenko 1987, tav. I.26). 7 Interestingly the latter is still found at Afis at the beginning of MB (Fig. 5.6), evolving into more sharply carinated versions. 8 In CqIV18-19, beneath the floors of two rooms in the southernmost part of the fortification dated to the MB I (US 1890and 1877: Felli, Merluzzi 2002), at least two superimposed floors have been brought to light, these directly above a level clearly associated to late EB IVB pottery (Felli in press). A similar situation has been encountered in every place where it has been possible to penetrate below the MB phase in this area.

CHRONOLOGY: SOME NEW DATA
Although the dating and character of the passage between the two (EB and MB) periods will exhibit some variation among sites (Schwartz 2017: 120), chronological synchronization at a regional level is nonetheless deemed necessary and much impulse has been given in recent years to this end.
Traditionally, EB IVB, now equated with phases ENL 5 and 6 in Arcanian terms, is dated 2300-2000 BC, although a possible further subphase, EB IVC, covering the last century of the 3 rd millennium, has been suggested (Schwartz 2017: 91, 93). Considering that MB is conventionally made to start at around 2000 BC, not much space is left to accomodate a transitional period between the EB and MB, as desired. A shared idea is that the transition should be a short period, maybe dated from 2000 BC to 1900 BC, thus extending into the second millennium BC (Porter 2007: 94). 9 If radiocarbon determinations are now increasing in Inner Syria for the third millennium, they are generally coming from earlier EB contexts (Manning et al. 2020: 6). One exception are the 14C dates for Qatna Phase 18, Area J, which indicate the interval 2040-1930 BC for the transition between the EB IVB and MB I (Morandi 2013: 414). 14C available dates for Qarqur equally place the end of EB IVB occupation around 2000-1950 BC (Karoll 2011: 97).
In order to provide radiocarbon dates for the EB-MB transition at Tell Afis, analyses were carried out on a larger number of samples than the three already obtained from the context of kiln A (Felli, Merluzzi 2005: 52): given here are uncalibrated dates so that the raw data can be considered independent of adopted calibrations. 10 The results offer a chronological anchor separate from pottery: on the basis of the available results, the life span of the kiln falls between the 3 rd and 2 nd millennium, fitting quite nicely with Qatna radiometric determinations. Dates available from the immediately underlying EBIV B levels in the northern sector of the area, seen in the table below, confirm the proximity of the two phases: Unfortunately, we lack dates from the overlying MB levels. If we consider other sites, it is worth mentioning the radiocarbon dating of one of MB IB grave from area V at Ebla, D.6384 (Baffi 2000: fig. 8.1), which is 2026-1896 cal BCE (3605 ± 25 BP) (Skourtanioti et al. 2020: e18). Additional datings from Ebla come from the EE midden, ascribed to the beginning of MB in archaeological terms, 2140-1910 cal. BCE (3652 ± 35 BP) and 1980-1740 cal. BCE (3545 ± 45 BP) (Peyronel 2019: 750, note 36), although the peculiar nature of the context here makes the chronological indications less tight.
Although more dates are needed, it is nonetheless clear that the issue concerning the EB-MB transition cannot be solved through radiometric determinations alone; the intervals are still rather large compared to the prescribed archaeological phase, which should be quite short. In terms of relative chronology, and therefore archaeological phases, terminology certainly plays an important role: the aspects each scholar sees predominant in the material culture characterizing these phases in terms of tradition or innovation are determinant in deciding how to call it, if still very late EB, thus stretching this phase at its maximum, or early MB. There is space for reconciling these different positions, but for now, we leave the matter open and maintain the term EB-MB transition.

THE EB-MB POTTERY ASSEMBLAGE
The previously held view concerning the passage between EB IVB and MB in inner Syria was described as follows: "the material culture registers a crucial transformation, with a complete substitution of the Caliciform tradition, that undoubtedly represented a cultural high point, not comparable with the initial Middle Bronze ceramic culture, still rough and uneven, especially as concerns the clays and the fabrics" (Nigro 2002a: 102). This perspective can now be mitigated in the light of the increase of pottery evidence from a number of sites filling in the gap. In her recent synthesis on the question from the point of view of Ebla, Marta D'Andrea has argued for strong cultural continuity between the EB and MB and for a more refined development of the pottery repertoire of the EB IVB period at the site than previously suggested. Through an integration of the existing evidence with new data from the most recent excavations in areas HH and D, she now counts four subphases (2004)(2005)(2006)(2007): see last Matthiae 2020). 11 Nonetheless, though acknowledging the coexistence of "standard EB IVB types, late evolutions of these types, and new innovative shapes" in local EB IVB3-4 contexts, as we shall see below, the absence of a phase comparable to the transitional one identified at the site of Afis, Qatna, Tell Nebi Mend (Kennedy 2015: 302-310) and possibly Qarqur (Karoll 2011), is confirmed as well as the clear-cut distinction between EB IVB and MB I pottery repertoires, ascribed to the break caused by the second destruction of Ebla at the end of the EB IVB (D'Andrea 2019b: esp. 279). 12 Without questioning the validity of this reconstruction, one aspect which should probably be taken into greater consideration is the impact made by different types of contexts, for example official buildings at Ebla versus houses and working areas at Afis and Qatna. 13 Such differences may perhaps explain the presence or absence of certain types, thus constituting an alternative scenario to chronological discrepancies/variances.
In the absence of substantial breaks, a pottery repertoire is not expected to change all at once through time. Rather, an evolution from earlier to later attestations is a more usual pattern than sudden "pop-ups", although the latter cannot be totally discounted; paces of development may vary greatly if we are dealing with short-lived productions. They are more subject to fashion trends than luxury wares, for example, or with long-lived types such as storage vessels. Although some changes did occur over the long run, at Afis no clear sharp line can be drawn to indicate two totally different pottery repertoires.
Just summing up the general characteristics of the corpus found in association with the installations of the working area and related rooms in area E3, we can observe a sharp decrease in painted wares, especially of the specific type which is the hallmark of the EB IVB in the region, the painted and incised beaker. 14 It is true that some sherds are still found in this intermediate phase, but there are no complete vessels, and some of the smaller fragments could represent residuals; the same applies to Smeared Wash Ware and Metallic Ware sherds. 15 13 See for example the case of richly decorated trefoil mouth jugs found in abundance in the HH temple at Ebla and only sporadically encountered at Afis, namely in a probably ruined grave: Giannessi 1995: 256, fig. 10.1-4. 14 See last Cooper 2020. 15 Percentages of distinctive classes such as Reserved Slip, Pattern Combed and Smeared Wash are already rather low (especially the first two) in the late EV IVB phase also at Ebla: see for example, the assemblages of the first phase of the Archaic Palace; Area T next point of view of less specialized productions, Simple Ware is by far the most common, followed by Cooking Ware, these being common traits of the late EB and early MB levels at the site as well (Felli, Merluzzi 2005: 51, 52). These two macroscopically defined wares are clearly identifiable at a petrographic level. First analyses conducted on samples from kiln A and the room to the east showed that Cooking Ware fabrics are characterized by a reddish matrix with abundant calcite crystals, calcareous and marble fragments. Simple Ware ones gather around a brown to greenish brown or reddish brown groundmass rich in volcanic and calcareous rocks fragments with further mineral components (quartz, plagioclase, olivine, pyroxene, biotite, calcite, iron and titanium oxides); some sherds show many clay nodules (chamotte), some also vegetal temper or fossils (Morbidelli 2005). The definition of sub-fabrics within these groups is now in progress and will be the subject of a dedicated study. 16 Certainly there is a trend towards a reduction of fabrics in the passage from 3 rd to 2 nd millennium, although this already started in the EB IVB or possibly even earlier (Mazzoni 2002b: 77-79). 17 It is worth mentioning that a "semplificazione dei processi produttivi" is also acknowledged at Ebla by Cristiana Pezzetta in the passage between EB IVA and EB IVB, when limestone grits are found in the paste of fine wares, including the painted ones (Pezzetta 2003: 28, 29; see also the observations in Pinnock 2009: 75). A similar situation is also encountered at Qatna, where some EB morphological types at the end of the 3 rd millennium sequence show pastes more similar to those in use in the MB phase (Iamoni 2014: 12, fig. 3.7). 18 As to the disappearance of specialized wares, it is also worth recalling that phase J of the Amuq sequence is characterized by the disappearance of 'white-on-black' decoration in Painted Simple Ware and of Smeared Wash Ware, in addition to the disappearance of distinctive Simple Ware features (cylindrical spouts and goblets) (Welton 2014: 343).
Turning to ceramic morphology, when first publishing the Afis material, we defined it as "transitional", following upon Porter's observations of the pottery from Building 2 at Tell Kabir on the Euphrates River (Porter 1995: 149). 19 This was not so much based on the presence of "transitional" types as it was the coexistence of EB IV types alongside those anticipating MB developments (Felli, Merluzzi 2005: 52;2008: 103). Nonetheless, when commenting upon individual types, we were able to identify some types or rather, variants, which could be taken as quite specific components of the assemblage of this facies. Of course, any reconstruction should take into consideration the casualties behind presence/absence of single pieces. Therefore, too rigid slots into which material should be fitted have been avoided, allowing instead for more permeable limits in order to enable comparison with other sites where different developments may have occurred.
In this perspective, it is important to make reference to current knowledge about the two assemblages with which our material is most associated namely, the late EB IVB on one hand, and MB I on the other. Similarities among types attested in both phases have already been underlined both at Afis and Ebla: Stefania Mazzoni, for example, has suggested that expanded rim bowls found quite frequently in early MB contexts at Tell Afis may descend from a similar type encountered in EB IV levels (Mazzoni 1998: 34, fig. 20.2;24.28). As for Ebla, Cristiana Pezzetta noticed that the EB IVB pottery of the Archaic Palace 20 showed traits of future developments, in the less accurate usage of the wheel, for example. The double rim bowls are a bit different from the EB IVB ones and more similar in the development of the lower part of the rim to the ones with high carination at the beginning of the MB (Pezzetta 2003: 33). 21 "The similarities between typologies overlapping from one horizon to the other" to the Northern Gate and the Northern Fort which are illustrated in the distribution graphs in Pezzetta 2003, pp. 28, 37-39. A distinction can instead be made for painted wares which are still relevant in Area T and in the Northern Fort. On Smeared Wash Ware see Rova 2014: 205-208. 16 A reexamination of the samples from Syria kept in the storerooms at the University of Florence is carried out by Simona Raneri, ICCOM-CNR, Pisa. 17 For analyses on EB IVB pottery from Tell Afis see Falcone, Lazzarini, Galetti 1995;Falcone, Lazzarini 1998: 485-487. 18 A first evaluation of the EB-MB transition at the site can be found in Morandi 2008. 19 On Kabir see also Cooper 1998;Porter 2007. 20 Now ascribed to EB IVB3 in the new phasing of Ebla: see last Pinnock 2020: 74, table 4.1. According to the most recent reconstruction the building was abandoned at the time of the destruction of Ebla and then rebuilt in MB I: Matthiae 2020: 91, 98. 21 D' Andrea 2018a: 228 acknowledges bowls with modelled slightly ridged rims ( fig. 13, on the right) and bowls with vertical rims have also been acknowledged by Frances Pinnock (2009: 75). More recently, it is Marta D'Andrea who has focused on the subject in a number of contributions (D'Andrea 2018a, b;2019a, b). In particular, she identifies among the types ascribed to the very late phase of EB IVB at Ebla, especially from area HH 4 and area T, those which appear a bit later than others. In addition to the typical bowls with vertical grooved rims, she notes goblets with modelled bases, small jars with triple-grooved rims and combed surfaces (D'Andrea 2018a: figs 10. 23-24, 11. 4-5, 12), and jars with double everted rims (D'Andrea 2018a: fig. 10.25; see also Pinnock 2009: 75, Fig. 5.2), these anticipating a very common type of MB jars (here Fig. 7.2). Other forerunners of the MB types include carinated bowls, bowls with slightly incurving rims, bowls with incurving slightly carinated walls and everted rims (D'Andrea 2018a: 228, 229, fig. 10.18, 20, 22; see also D'Andrea 2019b: figs 1.4; 2.13, 14, 17; 8.7; 9.3) and jars with vertical triple rims, as already suggested by Nigro (2009: 308, fig. 6.7). Nonetheless, the hallmarks of the Simple Ware repertoire of these late EB IVB phases at Ebla (EB IVB3-4) are still the unpainted goblets, either plain (D'Andrea 2019b: fig. 10: 7-9, 13) or with combed decoration (D'Andrea 2019b: figs 10: 10-12, 11: 1-3), sharing shapes with the same goblet types of the Painted Simple Ware, which are still very frequent. 22 This aspect is also encountered in late EB IVB at Afis, where similar types have still some quantitative significance, but not in the transition (apart from residuality), and that is one main difference which makes the two assemblages separate from one another.
If we look instead at later types, we admit that while the ceramic assemblage of the full MB period is well known in the region (Nigro 2002a, b;Pinnock 2005), that of the beginning of the MB is still somewhat obscure due to the limited amount of materials which can be securely ascribed to this phase. At Ebla, the pottery from the intermediate floors of the Archaic Palace (phases II-III: Matthiae 2006: figs 9, 10), which has been for long "the only stratified assemblage available from Inner Syria for the time span 2000-1900 B.C." (Nigro 2002a: 101), is too limited in kind to provide a large sample for comparisons with other sites (Nigro 2002a: 102). The materials from the houses in area T, where a continuous sequence from EB to MB has also been brought to light, are not yet fully published as far as the Middle Bronze period is concerned, 23 and even so, it consists mostly of "small jars and proto-collared bowls" (Nigro 2002a: 101 note 50). Further, more specific assemblages of the early MB come from funerary contexts: Tomb 16 in area G, redated to MB IA by Lorenzo Nigro (Nigro 2002a : 99  table 7 25 The most recently added material to the Ebla MB I corpus, from the midden from area EE (Peyronel 2019: see also below), is much more abundant and varied, but lacks the requirements of a closed context which would allow secure chronological attribution.
On the basis of the Ebla evidence, the MB IA pottery horizon is defined by the following diagnostic types: Simple Ware slightly carinated bowls, "Gublite bowls," "proto-collared bowls," craters with expanded rim and combed decoration, jars with collared rim, and, slightly later, collared bowls and bowls with sharp carination and ovoid jars with double rims; cooking pots, mostly with rounded bodies and unburnished surfaces. Finally, no specialized wares are attested, since Black Burnished ware is considered an outcome of MB IB (Peyronel 2019: 748-750). 26 As we have seen, many of these types are in fact anticipated in the local late EB IVB levels: the Afis transitional evidence may contribute to providing further ties bridging the apparent gap.

SIMPLE WARE
At Afis, the repertoire of small-sized open vessels of the late EB IVB includes, apart from goblets, different types of bowls. Among them is a cup-like form, with rather vertical sides and flat bases (Fig. 6.1). Another is carinated, with a slightly inturned, beaded or plain vertical rim and shallow ring base (Fig. 6.4). Both forms are still attested in the EB-MB assemblage which we are trying to define here (Figs 6.2, 3 the cup-like form and 5, the inturned rim bowl; for the first, see also Felli, Merluzzi 2005: fig. 42.4). They apparently survive the disappearance of painted and incised beakers. An interesting specimen of the second type has bitumen dots on the outer surface (Felli, Merluzzi 2008: fig. 6.7) which recalls a decorative technique common in the late third and early second millennium BC levels (EJ5) at some Jazira sites, such as Tell Brak and Tell Mozan (Rova 2014: 19). The former type seems to disappear in the early MB levels at Afis. It is also not attested in MB IA at Ebla, unlike the other which is considered, as we have seen before, a good indicator of MB IA, disappearing thereafter.
The relationship between the carinated bowl with inturned rim and the so-called "gublite bowl", a biconical bowl with everted rim and ring base which, in the Ebla sequence, marks a more advanced phase of the MB IA that continues into MB IB (Nigro 2002a: 102;2009: 373;Peyronel 2019: 749; but see contra Pinnock 2005: 120) requires further investigation. The dating of the gublite bowl is based on the fact that it is thought to be a derivation from metal propotypes, of which the earliest example known is attested in the Montet jar from Byblos, dated, on the basis of the glyptic enclosed there, to MB I (Nigro 2002a: 102, note 60). However, at both Afis and Ebla, the earliest examples of gublite bowls are made of the same paste as the other Simple Ware bowls and no "metallic", i.e. metal-inspired, specimens are attested, the latter being a later phenomenon so far encountered only at Ebla. 27 In addition, at Afis, both the narrow (Fig. 6.6, 7) and the wide variety ( Fig. 6.8; for a comparison from Ebla see Nigro 2009: fig. 6:4 on the right; Pinnock 2014: fig. 1) make their appearance in the transitional level where they are found together with the types above mentioned and seem to represent a local evolution within that tradition (Fig. 6.9), antedating the beginning of the MB, as it is now clear also at Ebla. 28 According to the reconstruction offered by Nigro (Nigro 2009: 343, 344: see also Pinnock 2005: 37;2014: 231), the gublite bowl is, in its turn, replaced by the collared bowl, this becoming the typical drinking vessel in Middle Bronze II, anticipated by protocollared bowls already in MB IA (Nigro 2002b: 301, fig. 4). 29 We have seen above the possibility that a plain rim variety of collared bowl, or biconical beaker as it can also be termed, has an antecedent in late EB IVB levels ( Fig. 5.4). In transitional and early MB levels, the most common collared rim type has a straight or slightly flaring rim, usually with flat base (Fig. 6.10); the height of the rim is not relevant until the end of the sequence (see for example Gabarrini 1995: fig. 15.5). A more rounded and usually larger variety (Fig. 6.11) is also found in the transition and recalls Middle Euphrates examples attested already during the EME 5 and continuing in the EME 6 (Sconzo 2015: 133, type 121).
Speaking of carinated forms, there is another type, the everted rim bowl with high carination, which is found in late EB IVB levels not only at Afis (Felli, Merluzzi 2008: fig. 5.6), but, as seen above, also in other sites such as Ebla (D'Andrea 2019b: fig. 1.4, 12, 13), where it is thought to anticipate the development of everted rim carinated bowls characteristic of MB IA (Nigro 2009: 302, fig. 6:3 top left; 30 see also Pinnock 2005: 34). At Afis an intermediate type is represented by the carinated bowl, with flaring rim, slightly convex sides and flat base (Fig. 6.12). This bowl, also encountered in the Hama J1 assemblage, was manufactured in the pottery kiln A and has a still dense, well fired fabric recalling the tradition of fine wares of the Early Bronze Age (Felli, Merluzzi 2002: 100 and27 For Ebla, Pinnock 2005: 121. It is perhaps not out of place to recall here that if the vessel in the hands of the stone seated statues found at Ebla is a gublite bowl as it has been suggested (dating to MB IA: Pinnock 2005: 120 note 14, and 38, note 65, for the possibility of metal specimens), it has to be considered that Rita Dolce dates at least two of them well ahead the end of the third millennium: Dolce 2008: 179, 180, figs 4.a, b, 9. 28 For late EB IVB attestations at Ebla see for example D' Andrea 2018b: fig. 10.22; 2019b: fig. 8.7. 29 See already Marchetti, Nigro 1997: 12, notes 37-39, fig. 6.13, where however the vessel is dated to MB IB. 30 Here called "coppa con carenatura arrotondata" and assimilated with the inturned rim carinated bowl, in my opinion a different type, for which see above.
note 12). This suggests a continuing favour of the shape throughout the period and not a MB inception. We also note the presence of some thicker and more rough, in terms of manufacture, examples which possibly reflect an experimental stage of the more canonical type of carinated bowl (Fig. 6.13, 14).
Moving to larger vessels, deep bowls with horizontal grooved rims have been found at Afis in a phase just after the dis-use of kiln A (Mazzoni, Felli 2007: fig. 4.5, 6). They are similar to MB specimens, although the latter are generally taller and have more pronounced rims (see for example Nigro 2002: pl. XVII.20, 21). The biconical shape of the Afis examples is probably due to separate manufacture of the two parts of the vessel body, built first by coils and then finished on the wheel. Those retrieved in the earlier phase ( Fig. 6.15, 16) have generally a less distinct rim from the wall and can be ascribed to the transition or even to the very late EB IVB. Examples with less pronounced rims are in fact encountered in late EB IVB contexts at Afis (Fig. 6.17), but also at Tuqan, in area N, where grooved rim deep bowls are attested already in EB IVB levels along with triangular rim jars (Baffi, Peyronel 2013: fig. 9.15, 16 and 19), the latter also encountered in transitional contexts at Afis (see below). Large, carinated bowls with grooved rims on the top are ascribed to the transitional period at sites further afield, such as Šavı Höyük in the Birecik area. 31 Regarding closed forms, jars at Afis offer investigations of a long term perspective, since they are frequently attested in both the 3 rd and 2 nd millennia, perhaps due to the functional character of the structures brought to light in area E3 in which they were found. This situation is different than EB IVB Ebla, for example, where jars are on a whole less abundantly represented, with few complete forms. 32 First of all, it is to remind that, as shown also by the new data from Ebla, jars with double everted rims, once acknowledged as diagnostic of the MB (Fig. 7.1), can also be ascribed to the late EB IVB period (Fig. 7.2). At Afis, however, they do not seem to be very frequent before the transitional level in comparison to simple or thickened, everted rimmed jars and then grow in frequency during the MB occupation.
We already noticed that globular jars attested in transitional contexts are much in the EB IV tradition (Felli, Merluzzi 2008: 100), but we can now try to be more specific. Round-bodied jars had already been found in an EB IVB context during earlier excavations at Afis (L. 1112: Mazzoni 1998: 32, fig. 17.2, 3). Another similar, almost complete, jar, with everted rim, large rounded body and convex base, and potter's mark (Fig. 7.3), has been found in situ in a room excavated in 2009, to the east of L.1112. This context can now be now securely dated by 14C to a time range between 2435 and 2205 BC. Smaller specimens are also found in pieces in the context of kiln A, similarly provided with pottery marks, another trait of the late EB tradition (Felli, Merluzzi 2008: fig. 6.5, 10, 12). This type of medium-sized serving jar is still attested in the earliest phase of the fortification, to the north of the complex (Felli 2000: fig. 10.5), originally dated to MB I, then ascribed tentatively to the EB-MB transition (Felli, Merluzzi 2008: 102). It is not attested in the phase just above, apparently losing popularity in comparison with more ovoid shapes. A possible late evolution of the type could be the large jar with sloping shoulder (Fig. 7.4) found in a narrow room (L. 1444), which presumably gave secondary access to the inner rooms to the west in the MB fortification. The jar, which has a simple incised and impressed decoration at the point of maximum expansion of the body, finds a close comparison, though on a smaller scale, to the specimen found in the tomb of the Princess at Ebla, which has been dated to MB IB (Matthiae 1979: fig. L.6; see also Nigro 2009: 139, pl. VIII.3). The latter piece appears rather isolated within a pottery equipment composed mostly by medium-sized jars with more ovoid bodies and flattened bases (Nigro 2002b: fig. 7.19, 20).
Interestingly, round bodied jars are explicitly mentioned along with neckless ovoid jars with "flower pot" bases as diagnostic types of the transitional period in the assemblages of Ebla area T and Afis level 17 by Anne Porter (2007: 85). Jars with "flower pot" or perforated bases are in fact attested at Afis both in the transitional and early Middle Bronze phases (two are attested in the same contexts as the ones mentioned above for the round bodied jars : Mazzoni 1998: fig. 17:4;Felli 2000: fig. 10.6), presumably responding to the specific need linked to the preservation of cereals and protection from humidity (Sollee 2020: 628 note 11). Leaving aside this aspect, the latter have generally a more barrel shaped body, little or no neck and a thickened, externally double profiled rim type, which seems to have been particularly fashionable during this intermediate stage, as attested by the retrieval of several specimens in the firing installation/kiln B in the southern part of the area (Felli, Mazzoni 2007: 214-216, figs 4.7-10). Generally the rim profile is less marked than in the full MB specimens, some showing a shape which is still close to their EB IVB antecedents (Fig. 7.5, 6). 33 The last type of rather common jar both late EB and transitional levels is distinguished by a thickened, roughly triangular shaped rim (Fig. 7.7; see also Felli 2000: fig. 10.6). Some of these jars have a double or triple series of horizontal grooves on the walls, a type of decoration which has usually been considered an MB feature (Pruss 2007). This assumption can now be questioned and not only on the basis of the Afis evidence. In fact this type of decoration appears already on big jars in the last levels of Amuq J (Braidwood, Braidwood 1960: fig. 341.7, 9); it is attested in the lowest depth at Tayinat and included in the Second Mixed Range, ascribed either to the very end of phase J or, less convincingly, to the Middle Bronze period, phase K or L (Braidwood, Braidwood 1960: 463, fig. 363). Since at Tayinat no occupation later than EB IVB was found until the Iron Age, Robert Braidwood, persuaded that this decoration had to be dated to the MB period (phase K or L), put forward the hypothesis that someone had brought along these vessels from Tell Atchana, though admitting, as a less probable option, they could represent the very last part or Amuq J. 34 That comb incised decoration dates back to the late EB IVB, as confirmed by the Ebla evidence mentioned above, is in much better accord with the inception of this decoration in the Jazira, where it appears before the end of the 3 rd millennium (Schwartz 2007: 260 for the chronological implications).

COOKING WARE
Cooking Ware is another realm in which patterns of steady development through time can be followed. The EB IVB Afis Cooking Ware repertoire includes large, open bowls with thickened rims, often provided with ledge handles (Mazzoni 1998: 33, figs 16.18;18:11, 12 and 19), and round bodied pots, some with a triangular lug at the rim (Mazzoni 1998: figs 16.17;17.1) 35 . Fragments of bowls and platters of the former type are still present in transitional contexts (Fig. 8.1-3), including an interesting, somewhat hybrid, open vessel with an EB form, but with burnishing on its surface, akin to MB cooking vessels (Fig. 8.4). It is worth mentioning that at Ebla, some specimens of wide bowls with more hammerlike, thickened rims and outer grooved surface, unburnished, appear in late EB IVB contexts (D'Andrea 2018a: fig. 8.5) and seem to anticipate the classic MB hammer rim platters and bowls, with ring bases and heavy burnishing (Nigro 2002a: 103). A similar bowl, without grooving, is also attested at Hama J1 (Fugmann 1958: fig. 103, 3D697) and possibly represents an early attestation of the MB type as well.
Cooking pots with rounded bodies and simple or thickened everted rims are clearly innovations of the EB IVB (Mazzoni 1998: 33), gradually replacing earlier hole-mouth pots and starting a tradition which lasts throughout the Middle Bronze period. EB IVB specimens are generally unburnished, but this again seems not to be a rule. The same applies to transitional (Fig. 8.5) or very early MB cooking vessels, where burnishing, when present, is less pervasive and accurate than on later examples. Triangular ledge rim pots, typical of the EB IVB period (Fig. 5.5), are instead customarily burnished and continue to be so when still attested in the transitional contexts, though apparently disappearing thereafter (Felli, Merluzzi 2002: fig. 18.6).
Cooking ware also includes peculiar vessels that take the form of long-stemmed censers, whose paste can resist thermal shock: 36 at Afis they are found stratified in the earliest MB contexts, such as in the southernmost part 33 A comparison can possibly be found in the burial container of grave D 6922 at Ebla, now dated to MB IA, of which only an artistic drawing is published but is described as a "giara con orlo doppio derivato dall'orlo triplice del Bronzo Antico" (Nigro 2009: 299). 34 The exact length of the Amuq J phase is an issue still to be investigated, although the idea is that ended earlier than at Qarqur and Afis : Welton 2014: 359. 35 Further examples have been encountered in the most recent campaigns: Mazzoni in press. 36 This is the common trait of all cooking ware vessels: see however the use of the more neutral definition of coarse ware at Taynat to go round the difficulty that not all similarly tempered vessels serve cooking purposes (Welton 2014: 352). of the trench, on a floor on which MB I materials were found (Fig. 8.6). Similar vessels are found in the favissae F.5237 and F.5238 (lower levels) at Ebla, where they were originally ascribed to MB IB (Marchetti, Nigro 1997: figs 6.16;7.33-36;2000: fig. 6.1-3, 5 37 ; see also 2002b: 305), but now to MB IA (Nigro 2009: 318, 319). 38 The type, however, is not an entire novelty of this period, but evolves out of a tradition of stemmed censers in Cooking Ware attested already in EB IVA at Ebla (see for example Marchetti 2013: fig. 27.5.6, 7) and continuing also in EB IVB (see for example Qarqur: Dornemann 1999: 93, fig. 167, on the right; Tayinat: Welton 2014: fig. 9.3, 4). To this date should be also ascribed the large fragment of bowl, most likely the upper part of a censer, found at Afis embedded in the foundations of one of the MB walls (Fig. 8.7).

CONCLUDING REMARKS
In sum, we observe that the changes in the use of the space attested in area E3 at Tell Afis (houses/workshop/ fortifications) do not require an explanation which falls outside the normal reshaping of a continuously settled site; the scale and specialization of the pottery (and possibly other craft) manufacturing area corresponds with the standards expected within a thriving community of an urban society. The reinforcement of fortifications may be a sign of a critical time, but not necessarily of a crisis of the site itself. As far as pottery production is concerned, the "transitional" assemblage speaks for some degree of continuity between the EB and MB traditions and shows no signs of impoverishment: the decrease of specialized, luxury wares has to be seen within the context of an incipient process of standardization which finds its full achievement in the Middle Bronze Age. Several morphological developments can be followed within both the simple and cooking ware from the EB IVB to MB. Beyond Afis, the possibility of a similar, gradual pattern of development can be envisaged at other sites where both EB and MB occupation are attested, once the ceramic types described above are not all seen as exclusively MB in date. 39 To provide but one example, one can see the presence of profiled rim jars, grooved rim deep bowls and carinated bowls and cups similar to those found at Afis at the nearby site of Mastuma, in stratum V, now currently dated to the MB period. These types are also encountered in the preceding stratum VI, which is ascribed to the EB (strata VI-X EB IV: Wakita 2009; see also Ascalone, D'Andrea 2013: 219). This evidence could well be an indication, not of an interruption in settlement between EB and MB, but of some gradual passage between the two phases, and maybe suggesting also an earlier dating of level V.
The results of the survey in the Jazr plain around Afis have shown that the EB IVB was a period of intensification of settlement which most probably continued up to the beginning of the MB period (Mazzoni 2006: pls 6, 7). Some sites, in particular, such as Tell Suffane, Tell Serji ( Fig. 1), 40 have yielded materials mostly ascribable to that time span, thus indicating that occupation in favoured agricultural zone continued to thrive even at a critical time elsewhere (Mazzoni 2013: 50-52). A similar pattern has also been suggested by the archaeological research undertaken in the Matkh region (Peyronel 2014: esp.116, 123-125). Here excavations in area N in the lower town of Tell Tuqan, located just to east of Afis, have yielded a sequence from EB IVB to MB IIA which attests to continuing occupation of the settlement during the time period of interest (Ascalone 2011). No transitional phase, however, has been detected at the site (Peyronel 2014: 124-126). The succession of phases assigns phase 7 to EB IVB, phase 6 to MB IA and phase 5 to MB IB, along with a number of graves similar to those at Ebla and Afis (Ascalone 2011;. At the regional level thus the question of the passage between the 3 rd and the 2 nd millennium still remains open, although continuity would seem to be a widespread trend. The evidence from Tell Afis challenges the