Mamshit

Transliterated Name	Source	Name
Mamshit	Hebrew	ממשית‎
Kurnub	Modern Arabic	كورنوب
Kurnub	Nabatean ?
Kurnub	Arabic
Mampsis	Byzantine Greek	Μαμψις
Memphis	Ancient Greek	Μέμφις

Description of Mampsis by Sion and Israeli (2022)

Sion and Israeli (2022:247-250) described Mampsis as follows:

[Mampsis had] three major phases, turning from a guarded roadside caravanserai in the Nabatean period (1st century BC–1st century AD) into a flourishing and rich Middle Roman city (2nd–3rd centuries AD) and eventually becoming a Christian Byzantine city with two impressive churches (4th–mid-6th centuries AD).

The recent identification of a military camp and commander residency point to a military attendance at Mampsis already in the Nabatean Period. Two Latin inscriptions from the military cemetery, dated to Trajan and Hadrian eras, identify two of the burials: a Legio III centurion and an eques of the Cohors I Augusta Tracum.

Based on the sums of money specified in the Nessana Papyrus 39, dated to the mid-6th century AD, it seems that the city’s financial sustenance was based upon payments given by the authorities to the limitanei for their military service. Once this support ceased by Iustinian, probably after AD 532, and no money was available to pay the Saracens off, they invaded Mampsis burning down its main gate.

... Mampsis is situated on the southern margins of a valley in the northeastern Negev Highlands, c. 40km southeast of Be’er Sheva‘ and 5km southeast of Dimona (map ref. 206/548; 460–478 m asl). Three ancient roads led to the city (Figure 2): the road leading from the prosperous Nabatean region around the southern end of the Dead Sea towards Be’er Sheva‘, the road leading from the copper mining district of Faynan (Phaino) by way of Mezad Hazeva, and the road that connected Mampsis with Oboda.

The city (130 × 150–270m), spanning over 10.5 acres, is set on two hills, eastern (6.5 acres) and western (4 acres, Figure 3), with a ravine running between them, and is bordered by two shallow ravines on the east and the west. A deep seasonal streambed, Nahal Mamshit, runs along the southern margins of the city, then turning to the southeast, cutting through the hard dolomite rock of the Ẓafit Formation along the Hatira anticline and exposing rocks that served as building material in the city.

... The residential area covers 15,245sq. m and including public structures sums to 24,540 sq. m (c. 60% of the city territory). The city is divided into four quarters: western, central, southwestern, and eastern. Outside the city wall are a caravanserai (VIII) and buildings that Avraham Negev suggested to be an architecture school (XXIII) and a gymnasium (XXII). A large area (40%) within the city remained unbuilt and the empty grounds between the huge buildings served as streets, alleys and piazzas (Figure 5: 1–10).

Buildings were well-preserved and the quality of workmanship exceeds any other city in the Negev.

Maps and Plans

Incense Road Map with Mamshit from biblewalks.com

Incense Road Map

biblewalks.com
Aerial Map of Mamshit from biblewalks.com

Aerial Map of Mamshit

Google and biblewalks.com
City plan of ancient Mamshit from Korzhenkov and Mazor (2003)

Figure 2a

City plan of ancient Mamshit

Korjenkov and Mazor (2003)
Towns and Forts in the Central Negev and Arava Valley.

Figure 1.6

Towns and Forts in the Central Negev and Arava Valley. Early Roman through Byzantine periods.

Erickson-Gini (2010)

Early Roman through Byzantine periods - from Erickson-Gini (2010)
Fig. 5 - City Plan - from Sion and Israeli (2022)

Figure 5

City plan (O. Sion, S. Israeli and D. Porotsky).

Sion and Israeli (2022)
Fig. 6 - Map of Mampsis according to period - from Sion and Israeli (2022)

Figure 6

Insulae and Structures according to periods (O. Sion, S. Israeli and D. Porotsky).

Sion and Israeli (2022)
Fig. 7 - The eastern hill - from Sion and Israeli (2022)

Figure 7

The eastern hill, looking northeast (photo G. Fitoussi).

Sion and Israeli (2022)
Fig. 8 - Aerial View of Mampsis looking West - from Sion and Israeli (2022)

Figure 8

Mampsis, looking west (photo G. Fitoussi).

Sion and Israeli (2022)
Fig. 9 - Plan of army camp - from Sion and Israeli (2022)

Figure 9

Plan of army camp and commander residency (O. Sion and D. Porotsky).

Sion and Israeli (2022)
Fig. 10 - Commander residency - from Sion and Israeli (2022)

Figure 10

Commander residency (photo G. Fitoussi).

Sion and Israeli (2022)
Fig. 14 - Western Church Compound - from Sion and Israeli (2022)

Figure 14

Western church compound, looking north (photo G. Fitoussi).

Sion and Israeli (2022)

Phasing - Urban Development of Mampsis

Figures

Fig. 6 - Map of Mampsis according to period - from Sion and Israeli (2022)

Figure 6

Insulae and Structures according to periods (O. Sion, S. Israeli and D. Porotsky).

Sion and Israeli (2022)
Fig. 7 - The eastern hill - from Sion and Israeli (2022)

Figure 7

The eastern hill, looking northeast (photo G. Fitoussi).

Sion and Israeli (2022)
Fig. 9 - Plan of army camp - from Sion and Israeli (2022)

Figure 9

Plan of army camp and commander residency (O. Sion and D. Porotsky).

Sion and Israeli (2022)
Fig. 10 - Commander residency - from Sion and Israeli (2022)

Figure 10

Commander residency (photo G. Fitoussi).

Sion and Israeli (2022)
Fig. 14 - Western Church Compound - from Sion and Israeli (2022)

Figure 14

Western church compound, looking north (photo G. Fitoussi).

Sion and Israeli (2022)

from Sion and Israeli (2022)
Sion and Israeli (2022:252 n. 6) note that The urban development is based on Negev’s publications and chronology, updated, and slightly revised by Erickson-Gini (2010: 83–86).

Period	Time Span	Discussion
Nabatean	30 BC— AD 106	Discussion Controlling the Arabian aromatic and spice trade from Arabia to the Mediterranean coast enriched the Nabatean kingdom. In the end of the 1st century BC, Mampsis was founded as a road station on a secondary caravan route leading from the southern coast of the Dead Sea to the Mediterranean. In that period, a fortress was constructed on top of the hill, a caravanserai and watchtowers were erected on the hill slope and a cemetery to the north of the settlement. The fortress (R/5; Building XIV) is the earliest structure dated by Negev from the last quarter of the 1st century BC to mid-first-century AD (Negev 1988a: 30-32). It was built on the southeastern hill — the highest place overlooking the road (Figure 7). The square fortress has a courtyard at the center, in which a small structure of two rooms is situated. The outer walls are extremely wide and other rooms are built align them. The fortress was ruined on several occasions: between the Nabatean and Middle Roman periods, in the Byzantine period when walls were constructed between the Eastern Church and Building XII, and in the 20th century AD when a British Police building was erected. Near the fortress, a small watchtower was found under the southwestern corner of Structure R2 (Building IV). Another watchtower was uncovered in the southwestern corner of the city wall (Building VI) and a third watchtower is in Building II. All three protected the settlement over the steep cliff of Nahal Mamshit (Negev 1988a: 4, 28). Three structures were built to the north of the fortress. The first has a central small courtyard (P/1; Building XIX) with storage rooms at its southern side, resembling those at Masada. Nabatean and Herodian lamps (somewhat rare in Nabatean assemblages) were revealed in the courtyard, in one of the richest assemblages uncovered at the site (Negev 1988a: 34). An ash layer was observed between the Nabatean period remains and the wall, which was constructed in the Late Roman period over the northern and eastern rooms of the building (Negev 1988a: 32-34). Adjacent to the west is Structure P/4 (Building Va) that joins a wall with contemporary Structure P/1. Half of Structure P/4 has rooms surrounding a courtyard and a second floor. Negev proposed that the remains of a stylobate and a staircase-tower attest to domestic architecture that was well developed already in Nabatean period (Negev 1988a: 34-37). To the north, Structure K was revealed, which was later integrated into the city wall (Negev 1988a: 44). A large caravanserai (S; Building VIII) was raised outside the city wall, at the foot of the hill. Negev suggested that the large structure (U; Building XXIII) further north was founded in the 2nd half of the 1st century AD as a school for the study of architectural and applied arts (Negev 1993b: 259). The settlement of the Nabatean period probably extended over a larger area than Middle Roman period (Negev 1988a: 27, 30). Pottery of the earlier Nabatean period was revealed in several buildings of the Middle Roman period, such as Structures D (Building I) and O/1 (Building XI; Negev 1988a: 40-44, 74-75 respectively). The foundations of a watchtower were exposed under the bell tower of the Eastern Church (R/7; Building IX) and early building remains were present under Structure R/2 (Building IV; Negev 1988a: 27, 29, 48-49). A Nabatean cemetery, discovered by Kirk 1 km northeast of the city was excavated by Negev, and it comprised of primary and secondary burials; it was in constant use for the first four centuries of the settlement (Negev 1971a; Negev and Sivan 1977: 111). The finds from the cemetery enhance our knowledge of the funerary customs, the dating of the settlement and the connection between Mampsis and other administrative centers such as Petra, Rabbathmoba and Characmoba (see below; Negev 1969). According to Negev, evidence of a settlement gap between the Nabatean and post-annexation periods were revealed in the `palace' (D, Building I; Negev 1988a: 74-77) and in both stages of the cemetery (Negev and Sivan 1977: 118-119).
Middle Roman	AD 106-300	Discussion Following the death of the Nabatean king Rabel II in AD 106, the Romans annexed the Nabatean kingdom to the Roman Empire and founded the new Provincia Arabia. New security arrangements led to the construction of a new road—the Via Trajana Nova—completed at AD 114-116. The III Cyrenaica Legion paved the road network in the province and Mampsis appears to have been renovated in the late 1st to early 2nd centuries AD (Negev 1971b).⁷ During the Middle Roman period, in the 2nd and early 3rd centuries AD, Mampsis was rebuilt as an unwalled city and most of the preserved buildings are ascribed to that period. They are spacious (average size: 570 sq. m), highly decorated two-story buildings and the rooms surround a central courtyard or several courtyards. They were built like a small fortress surrounded by a wall and the empty spaces between them served as alleys or piazzas. Most buildings on the western hill were residential; shops and workshops were built near the northern gate, while public complexes were on the eastern quarter. A Roman military necropolis dated to the time of Trajan and Hadrian was discovered east of the city. Two Latin inscriptions mention the burial of a centurion of the Legio III Cyrenaica and of an eques of Cohors I Augusta Thracum (Negev 1967a: 52-53; 1971a: 124 and see references therein). The Western Quarter Twenty-one houses were constructed on the western hill (Figures 3, 6). Insulae A and B included 9 and 3 structures, respectively. Three pairs (G, F, E) and three single structures (C, D, I), mostly residential (10 of 15), occupied the quarter. Structure D (Building I; Figure 8), dated to the Middle Roman period (2nd-3rd centuries AD), was identified as a palace with a reception hall and special decorative architectural elements (Negev 1988a: 66; 50-77). Structure F/2 (Building II) was an administrative center containing a tower, a courtyard, halls, and storage rooms (Negev 1988a: 77-78). Structure O/1 (Building XI) had an upper floor, stables, and a private shrine. It was partly destroyed in the second half of the 4th century AD, probably during the earthquake of AD 363, whereafter the western church was built (Negev 1988a: 88-109). Eastwards, Structure O/2 (Building XIa) was probably constructed in the late 2nd or early 3rd century AD (Negev 1988a: 109–110). In Structure S (Building VIII), arches were constructed in a new direction during the 2nd half of the 3rd century AD. Structure A/5 (Building XVI), southeast of the northern city gate, was dated to the current period based on its masonry carving style. Structure F/1 (Building XVII) was identified as stores and workshops of the period along with three rooms in Structure C (Building XVIII), located northwest of Building I (Negev 1988a: 191-197). The so-called gymnasium (T; Building XXII), a one-story structure, has many rooms surrounding a court (Negev 1993: 246-261). Following the survey, other structures were attributed to the current period (A/8, A/5, B/3, E/1, G/1, G/2, H, I, N/1), due to their dimensions, construction style and location (see Figure 6). Three small structures (A/6: 88sq. m; A/7: 84sq. m; F/3: 24sq. m), composed of 2-3 rooms were identified as stores or workshops. Two of them (A/6; A/7) are near the northern gate and the third (F/3) is in the northern fringes of Piazza 6. Footnotes 7 Tali Erickson-Gini suggests that earthquake destruction in the late 1st or early 2nd century AD prompted a wave of construction at Mampsis and Oboda (Erickson-Gini 2014: 100; Erickson-Gini and Tuttle 2017: 141). The Eastern Quarter The eastern quarter is located on the high ground which observes the city and its environment. Its complexes (R, P, Q) are of public nature. Insula P has two Middle Roman period structures (P/1, P/4), over which a bathhouse (P/3) and a pool (P/2) were built later.8 Structure Q, attached to the city wall and excavated at its northern part, is 25m long and has rooms on its eastern and western sides with a courtyard in between. At the southern high ground of the hill are two structures surrounded by a wall (R/1, R/2; Figures 9-10). The first was well preserved, while the other was destroyed in the 4th century AD, when the Eastern Church was erected. Negev assumed that Structure R/1 (Building XII) was the governor's house, while Structure R/2 (Building IV) was a market (Negev 1988a: 75-78). Structure R/1 (Building XII; see Figure 9), the largest and most magnificent structure at the site (c. 1500sq. m), was constructed at the late 1st or 2nd century AD. It was entered from the north with protecting towers on both sides and has three sections (see Figure 10): the northern is a large, domestic two-story residential area with rooms adorned with mosaic floors, ornamented door posts, decorated arches posts and Nabatean capitals with ritual decorations, an inner court with wall paintings and a cistern (Negev 1988a: 146-147); in the west are service rooms; and in the southeast, stables. In its various sections the building has inner courts. Four strata were observed (Negev 1988a: 145-147): phase I, the beginning of the commercial activity, was dated by earliest coins from a hoard of 10,500 silver coins, from the years AD 98-137 to AD 222, hidden in a bronze jar under a staircase. The building was abandoned probably following a local epidemic, and the hidden coins jar, that indicates a wealthy owner, was never recovered. The coins, minted at northern Syria and Asia Mainor, indicate a wide and rich trade network that dealt with Nabatean racehorses (Negev 1988a: 77). At phase II (c. AD 300), based on coins retrieved from the stables (Figure 11) and a reusage of Tomb 108 in the Nabatean necropolis, the complex was re-inhabited in AD 300. In phase III, dated to the 1st half of the 4th century AD, various changes were observed in the complex and entrances between rooms were blocked; it might have been inhabited by the army in the era of Diocletian. In phase IV, the complex was inhabited by Christians, as crosses were carved over lintels and door posts and a surrounding wall was erected around the building, the Eastern Church and Building IV; Negev suggested (1988a: 146-147) that it might have served the bishop. We propose that Structure R/1 (Building XII), adjacent to the army camp (see below), might have served the commander of the Roman garrison and its stables were probably used by the cavalry, one of which — Diogenes the son of Alexander of the Cohors I Augusta Thracum was buried in the Roman cemetery (Negev 1967b: 52, Pl. 9c). Fabian (2005: 219) suggested that the existence of the Roman cemetery indicates a military attendance which participated in roadbuilding. Structure R/2 (Building IV; see Figure 9) has three rows of rooms running north—south along two streets; most of the floors in the rooms were made of beaten earth. Two entrances led to the building from the north and the rooms were entered from the streets with no entrances between them. The building, constructed over the remains of a watchtower of Mid-Nabatean period, was identified as a market (Negev 1988a: 163). Fabian (2005: 217) suggested to identify it as a military camp. We agree with his identification for the following reasons: its location on the highest place observing the road leading to the city; its plan resembling other military camps with long rows of rooms (Fabian 2005: 217, see references therein); and its wide external walls, build of large ashlars on the outer face with two corner watchtowers beside the entrance in the northern wall (1.5m wide; Figures 9: 1; 12; 13). The military camp has a trapezoid plan (53m long, 25/36.5m wide). On its eastern side are six similar rooms (see Figure 9: 2-7). South of the watchtower are three wide, fortified rooms, presumably serving the watchtower and south of it is the southwestern entrance to the camp. The central row has six similar rooms (Figure 14: 15-20) with entrances in the east, apart from room 18, in which the entrance is in the west. Outstanding is room 14 that has benches along its walls (see Figure 13); its northern side is open to the western street, and it might have served as an inner training area. The eastern rooms were well-preserved up to the northern wall of the Eastern Church. In the northern section, that is attached to Structure R/1, are two rooms (see Figure 9: 25, 26), a watchtower (see Figure 9: 23) and a narrow room in the south (2.6 x 2.35m; see Figure 9: 24). It would be reasonable to assume that they were guard rooms. The third room in the eastern row is unique and large (10.0 x 3.8m) and has a connection to the commander's residency at the northeastern corner. It has benches along its walls and might have been a reception room of the command office. A bathhouse (P/3, Building V) was built north of the army camp and the commander residency. It has a Roman imperial plan with two connected units built on two levels. The upper holds an apodyterium, a frigidarium and a tepidarium and the lower holds the partly subterranean furnaces and a caldarium. It was built above two structures of the Middle-Nabatean period (P/1, P/4) and was used during the Roman and Byzantine periods. Water was supplied to the bath by pipes from the public pool (P/2; Building VII), which was built from ashlars, was plastered on the interior, and roofed by arches. The building was integrated with the city wall that was dated to c. AD 300 (Negev 1988a: 167-190).
Byzantine	4th— 6th c. AD	Discussion In the Late Roman period during the reign of Diocletian, the city was surrounded by a wall, dated, based on coins revealed at its foundations and related architecture to around AD 300 (Negev 1988a: 64; 1988b: 9-29; Erickson-Gini 2010: 84-85). In the 2nd half of the 4th century AD, two churches were constructed (see Figure 14; Negev 1988b: 30-63). Other structures were identified in three areas: five structures next to the city wall, west of Structure A/8, eighteen in the central quarter between the eastern and western hills (Insulae L, M, see below) and several others (B/1, B/2, O/4,O/5,O/6,O7, Q, T) next to the city wall. The average dimension of each structure is about 350 sq. m. The Central Quarter Eighteen structures were observed in two insulae (L, M) at the central quarter, and in the ravine between both hills, out of which 5 and 8 structures (respectively) were identified. Most of these structures are domestic and they are relatively small and rather poorly built. Larger structures (L/1 — 925sq. m; L/2 — 750sq. m) have spacious rooms and their plan and location next to the city gate may indicate that they were storage facilities. Outstanding is Structure J, which is poorly built and Structure N/22, which is rather small (320 Sq. m). An estimate of the number of inhabitants at Mampsis during the late 5th and early 6th century AD is based upon the size of the built area within the wall (24540 Sq. m) and the number of structures (56, of which 43 are domestic buildings). Broshi and Finkelstein (1990:4) estimated that 25 persons per dunam sums up with 1135 persons (see also Shor 1988: 252). Safrai (1995: 287-289) surveyed several sites and estimated 30-35 persons per dunam, which sums to 1239-1445. Shershevsky estimated between 1060-1500 inhabitants for Mampsis. Hence, we can assume 1000-1500 inhabitants in Mampsis. According to Avraham Negev, based on numismatic data, the end of Mampsis occurred during Justinian's reign, when the northern city gate was burned.⁹ As limitanei, the existence of Mampsis was based on the annona militaris. When the regular pavement to the limitanei was discontinued by Justinian during the 2nd quarter of the 6th century AD, the compact and isolated city may not have been able to thrive on the agricultural production alone. When the regular payment to the neighboring Saracen tribes ceased, the city was possibly attacked, and the gate was burnt. Footnotes 9 However, later excavations by Erickson-Gini in the courtyard entrance of Building XII revealed ceramic evidence dated to the 2nd half of the 6th and the early 7th centuries AD (Erickson-Gini 1999b: 101; Figs. 17: 6; 18: 5; 21: 3). Subsequent excavations in Building III between 2017–2020 substantiate the presence of ceramic wares beyond the mid-sixth century AD (Erickson-Gini pers. comm. 9.9.21).

Archaeoseismic Study by Korzhenkov and Mazor (2003)

Korzhenkov and Mazor (2003) analyzed damage patterns at Mampsis utilizing 250 cases of 12 different types of deformation patterns which they were able to resolve into two separate earthquake events on the basis of the age of the buildings which showed damage. The fact that the two different events showed distinct directional patterns - the first earthquake with an indicated epicenter to the north and the second with an epicenter to the SW - was taken as confirmation that they had successfully separated out archeoseismic measurements for each individual event. The first earthquake, according to Korzhenkov and Mazor (2003) struck around the end of the 3rd/beginning of the 4th century CE and the second struck in the 7th century CE - at the end of the Byzantine period. They provided the following comments regarding dating of the earthquakes

To determine exact ages of the destructive earthquakes, which destroyed the ancient Mamshit, was not possible by methods used in given study. It has to be a special pure archeological and historical research by specific methods related to that field. Age of the first earthquake was taken from a work of Negev (1974) who has conducted main excavation activity in the site. As concern to the second earthquake – the archeological study reveals that the seismically destroyed Byzantine cities were not restored. So, most probably, one of the strong earthquakes in VII Cent. A.D. caused abandonment.

Mamshit thrived, in spite of its location in a desert, thanks to runoff collecting dams, and storage of the precious rain water in public ponds and private cisterns. These installations were most probably severely damaged during the earthquake, cutting at once the daily water supply, forcing the inhabitants to seek refuge in the more fertile regions. This situation was most probably followed by looting by local nomads, turning a temporal seek of shelter into permanent abandonment.

Deciphering chronology at Mampsis has unfortunately been problematic.

Before the First Earthquake - Early 2nd century CE earthquake

Russell (1985) cited Negev (1971:166) for evidence of early second century earthquake destruction at Mamphis. Negev (1971) reports extensive building activity in Mamphis in the early second century AD obliterating much of the earlier and smaller infrastructure. However, neither a destruction layer nor an earthquake is mentioned. Citing Erickson-Gini (1999) and Erickson-Gini (2001), Korzhenkov and Erickson-Gini (2003) cast doubt on Russell (1985)'s assertion of archeoseismic damage at Mamphis stating that recent research indicates a continuation of occupation throughout the 1st and 2nd cent. A.D.. Continuous occupation could indicate that seismic damage was limited rather than absent.

First Earthquake - Early Byzantine

Negev (1974) dated the first earthquake to late 3rd/early 4th century via coins and church architectural styles however he dates construction of the East Church, where some archaeoseismic evidence for the first earthquake was found, to the 2nd half of the 4th century CE. Erickson-Gini (2010:83) asserted that the Mampsis appears to have experienced extensive damage in one of the Cyril Quakes of 363 CE noting that the damage predates structural changes in buildings at the site and the construction of two churches.

Second Earthquake - 5th - 7th centuries CE ?

The date for the second earthquake also seems tenuous as Negev (1974:412) and Negev (1988) indicate that Mampsis suffered destruction by human agency long before the official Arab conquest of the Negev and the town ceased to exist as a factor of any importance after the middle of the 5th century. However, Magness (2003) pointed out that there is evidence for some type of occupation at Mampsis beyond the middle of the 5th century CE.

The small amount of Byzantine pottery published to date from Mamshit also indicates that occupation continued through the second half of the sixth and seventh centuries. There are examples of dipinti on amphoras of early fifth to mid seventh century date. Early Islamic presence is attested by Arabic graffiti on the stones of the apse of the East Church (Negev, 1988). More recently published evidence for sixth to seventh century occupation, as well as for early Islamic occupation, comes from a preliminary report on the 1990 excavations. The description of Building IV, which is located on the slope leading to the East Church, states that "the building continued to function in the Early Islamic period (7th century c.E.) with no architectural changes ¹²². The large residence, Building XII, contained mostly material dating to the fifth century, but pottery of the "Late Byzantine and Early Islamic periods" was also present ¹²³. In 1993-94, T. Erickson-Gini conducted salvage excavations in several areas at Mamshit, under the auspices of the Israel Antiquities Authority. The pottery she found includes Fine Byzantine Ware Form lA bowls, and examples of Late Roman "C" (Phocean Red Slip Ware) Form 3, African Red Slip Ware Form 105, and Cypriot Red Slip Ware Form 9 (Erickson-Gini, 2004). This evidence indicates that the occupation at Mamshit continued through the late sixth century and into the seventh century. The Arabic graffiti on the apse of the East Church reflect some sort of early Islamic presence at the site, the nature of which is unclear.

Considering this dating difficulty, I am labeling the date for the second earthquake as "5th -7th centuries CE ?".

First Earthquake - Early Byzantine - Lower parts of buildings (built in Nabatean and Roman Periods)

Damage Type	Location	Figure	Comments
Systematic Tilting of Walls	E of West Church Figure 2b The West Church and Administrative Tower Quarters Roman numbers denote residential quarters Arabic numbers the location of Figures Korjenkov and Mazor (2003) Entire Site	3a Figure 3a Tilted walls. South side of the north wall of a room east of the west Church, the wall has a trend of 86° and the average inclination is 79°. Korjenkov and Mazor (2003) 3b Figure 3b Tilted walls. North side of the same wall, the lower stone rows are tilted northward by up to 60° Korjenkov and Mazor (2003) 3c Figure 3c Tilted walls. Summary of measurements: direction of tilting patterns observed at the lower parts of walls at Mamshit (Tab. 1), as a function of the direction of the walls, revealing that out of 30 cases of walls trending ENE (55° to 105°), 26 were found to be tilted northward, 180 and only 4 cases are tilted southward, whereas the perpendicular walls (trending 145° to 185°) show only 9 tilting cases with no preferred direction Korjenkov and Mazor (2003) 3d Figure 3d Tilted walls. Stress directions concluded from the observed tilting patterns Korjenkov and Mazor (2003)	Observed damage pattern: tilted walls or wall segments (Figs. 3 a. b). By convention, the direction of tilting is defined by the direction pointed by the upper part of the tilted segment. Only cases of tilting of most of the wall were included in this study. Statistical observations: The data of surveyed cases of tilting are summed up in Tab. 1. 30 cases of tilting were observed at walls trending 55° to 105°, out of these 26 are tilted northward, and only 4 are tilted southward (Tab. 1 and Fig.3 c). In contrast, only 8 cases of tilting were observed in the perpendicular walls, with a 135° to 185° trend, and out of these the tilting is in 4 cases eastward and in 4 cases westward. Thus, a clear preference of northward tilting is observed at the Roman ruins of Mamshit. Interpretation: Preferentially oriented tilts of the walls is becoming a common technique for recognition of a seismic nature of damage applied in archeoseimology ... An analysis of the seismic motions and resulting stresses in Mamshit is given in Fig. 3 d, leading to the conclusion that a seismic shocks arrived from north.
Lateral Shifting of Building Elements	E of West Church Figure 2b The West Church and Administrative Tower Quarters Roman numbers denote residential quarters Arabic numbers the location of Figures Korjenkov and Mazor (2003)	4 Figure 4 Lower stone of a N-S trending (175°) arch, shifted 8cm northward (the original position is marked by dashed lines). This is the fourth arch of the eastern line of fodder-basins of the Stables. The lowest stone of the arch has been severely cracked during the seismically caused shifting Korjenkov and Mazor (2003)	Observed damage pattern: northward shifting by 8 cm, as well as severe cracking of the lowest stone in a 175° trending arch (Fig.4). Thus, a large building element was shifted, and in addition slightly rotated clockwise. The location is at the eastern line of fodder-basins of a complex of stables, at a residential quarter east of the West Church. Statistical observations: 14 cases of shifting were observed. Interpretation: Displacement of the building elements is a known phenomenon of earthquake deformation in ancient buildings and was used for the determination of the seismic motions’ directions as wall tilt or collapse. The only process that could cause such shifting is an earthquake – no other mechanism is known. In Mamshit the seismic shocks arrived from north and the push movements were transmitted from the ground to the building foundations, causing the arch to move in an opposite direction, i e. towards the epicenter, due to inertia.
Rotation of Wall Fragments around a Vertical Axis	ENE of West Church Figure 2b The West Church and Administrative Tower Quarters Roman numbers denote residential quarters Arabic numbers the location of Figures Korjenkov and Mazor (2003) Near Frescoes House Figure 2c East Church and House of Frescoes Quarters Roman numbers denote residential quarters Arabic numbers the location of Figures Korjenkov and Mazor (2003) Entire Site	5a Figure 5a Rotation patterns at walls of the Roman period. Clockwise rotation of stones of the lowest row of a 172° trending wall of a room ENE of the West Church; stone A was rotated by 5° and stone B by 10°, the horizontal displacement of the two stones being 21.5cm. Korjenkov and Mazor (2003) 5b Figure 5b Rotation patterns at walls of the Roman period. A 2° counterclockwise rotation of stones in the lower part of a 59° trending wall at a yard near the Frescoes House. Korjenkov and Mazor (2003) 5c Figure 5c Rotation patterns at walls of the Roman period. Angle of rotation as a function of wall trends: SES (150°-185°) trending walls reveal mainly clockwise rotation, whereas walls trending NEN (60°-95°) reveal mainly counterclockwise rotations Korjenkov and Mazor (2003) 5d Figure 5d Rotation patterns at walls of the Roman period. Stress directions concluded from the observed rotation patterns Korjenkov and Mazor (2003)	Observed damage pattern: 1. An example of clockwise rotation of stones within a wall trending 172°, in a room located ENE of the West Church (Fig. 5 a). Stone A was rotated 5° clockwise and stone B was rotated 10° clockwise, the horizontal displacement between these rotated stones being 21.5 cm.. An example of a counterclockwise rotation in the northern wall of the Frescoes House (Fig. 5 b); the trend of the wall was 59° and the azimuth of the rotated wall fragment is 57°. Statistical observations: Walls trending 150° to 175° revealed 22 cases of rotation, and out of them 16 are clockwise and only 6 counterclockwise (Fig. 5 c). The perpendicular walls, trending 60° to 95° revealed 27 cases of rotation, out of which 24 cases are counterclockwise and only 3 cases are clockwise. Thus, a clearly systematic picture of rotations is seen: counterclockwise in ENE walls and clockwise in SSE walls (Fig. 5 c). Interpretation: Rotation of individual stones, fragments of the walls, or whole walls around a vertical axis is common phenomenon during strong recent and ancient earthquakes. Pulling out of foundation stones accompanying by their rotation in spite of their solid cement testifies on just dynamic beating out of them in the process of sharp horizontal oscillations of the all wall (and not only of its upper part) relatively the foundation. Seismic ground motion is the only mechanism that can cause rotation of building elements, a conclusion well supported by the large number of observed rotation cases and the obvious directional systematics. The theoretical background of this phenomenon in the buildings was described in details by Korzhenkov and Mazor (1999a) and Korzhenkov and Mazor (1999b). In Mamshit an analysis of the direction of the seismic motion, as derived from the dominant rotation patterns is presented in Fig. 5 d, revealing that the epicenter was approximately at NNE.
Cracking of Door Steps, Staircases and Lintels	Administrative Tower Figure 2b The West Church and Administrative Tower Quarters Roman numbers denote residential quarters Arabic numbers the location of Figures Korjenkov and Mazor (2003) E of West Church Figure 2b The West Church and Administrative Tower Quarters Roman numbers denote residential quarters Arabic numbers the location of Figures Korjenkov and Mazor (2003) Entire Site	6a Figure 6a Cracks in the southern (left) edge of a doorstep and the southern doorpost, at an entrance trending N-S (175°), leading to a room west of the Administrative Tower. Korjenkov and Mazor (2003) 6b Figure 6b A similar set of cracks in the entrance to a room east of the Administrative Tower. Korjenkov and Mazor (2003) 6c Figure 6c Stress directions concluded from the observed cracked patterns, disclosing movement of the adjacent walls Korjenkov and Mazor (2003) 7a Figure 7a Open cracks at the bottom and closed cracks at the top of a E-W (83°) trending staircase inside a Late Nabatean Building, and tilted side walls Korjenkov and Mazor (2003) 7b Figure 7b Stress directions concluded from the observed open and closed crack patterns, disclosing northward movement of southern walls. Thus the seismic push arrived from north Korjenkov and Mazor (2003) 8 Figure 8 Number of cracks observed in doorsteps, staircases and lintels as a function of the direction of the latter. The number of observations in N-S structures is three times greater than the observations in the perpendicular walls, leading to the conclusion that the seismic shock arrived from north Korjenkov and Mazor (2003)	Observed damage pattern: A 175° trending doorstep of the entrance into one of the rooms of the Administrative Tower was cracked at its southern part (Fig. 6 a) and a similar damage pattern is seen in the doorstep of another room, located eastward within the same building (Fig. 6 b). Cracks in a staircase of the Late Nabatean Building, located east of the West Church, is seen in Fig. 7 a. Double arrays there show direction of walls swinging. Because of pressure from tilting wall the doorstep got extra-loading which led to cracking of it. Statistical observations: Fig. 8 reveals that out of 44 observed cases of distinct cracking in Roman buildings, 32 are in northward trending structures (mainly 180°), and only 12 cases are seen in structures included in the perpendicular walls (trending around 90°). Interpretation: Cracks breaking special building elements, like doorsteps, staircases and lintels, are an important indicator for evaluation of the seismic damage. The cracking process of the doorsteps shown in Figs. 6 a. b are suggested in Fig. 6 c, and the damages seen in the staircase shown in Fig. 7 a is discussed in Fig. 7 b. The conclusion in each of these cases is that the southern wall was tilted northward by inertia in reaction to seismic shocks from the north, indicating the epicenter location was northward of Mamshit. The clear preferential occurrence of cracks in N-S trending walls is in agreement with this conclusion.
Slipped Keystones of Arches	W of Eastern Church Figure 2c East Church and House of Frescoes Quarters Roman numbers denote residential quarters Arabic numbers the location of Figures Korjenkov and Mazor (2003) Stables - E of West Church Figure 2b The West Church and Administrative Tower Quarters Roman numbers denote residential quarters Arabic numbers the location of Figures Korjenkov and Mazor (2003)	9a Figure 9a Keystone, slipped down 6cm in a N-S (174°) trending arch in a room west of the Eastern Church Korjenkov and Mazor (2003) 9b Figure 9b Two central stones, slipped 3cm down in an N-S (175°) trending arch above the third fodder-basin in the Stables Korjenkov and Mazor (2003) 9c Figure 9c Diagram of the operating stresses, leading to the conclusion that the seismic wave propagation was approximately parallel to the damaged arches, i.e. along a N-S direction Korjenkov and Mazor (2003)	Observed damage pattern: A 174° trending arch, located in a room west of the Eastern Church, exhibits a keystone that slipped 6cm down of its original position, as can be seen in Fig. 9 a. A pair of keystones slipped 3cm down in a 175° trending arch located above the third fodder-basin in the Stables (Fig. 9 b). An important auxiliary observation is that in these cases the arches themselves were not deformed. Statistical observations: Two cases of slipped keystones were observed, both in N-S trending arches. Interpretation. Hanging keystones themselves are a strong evidence of seismic origin of such type of deformations, but they also can be used as a kinematic indicator telling about seismic motions direction of a historical earthquake. Displacement of an arch keystone reflects an event of brief extension, during which the keystone slipped, followed by rapid return to the regular state of compression that fixed the keystone in its present state. Such a brief state of extension discloses arrival of seismic shocks that was transmitted to the base of the arch, causing its upper part to be momentarily tilted in the direction of the epicenter, the part facing the epicenter being more effected, as depicted in Fig. 9 c. The observed slipping of the keystone could have occurred in a number of steps during a series of oscillations of the upper part of the arch. The observation that otherwise the arch remained in its original position indicates that the seismic push arrived from a direction parallel to the trend of the arch, as otherwise the arch would be tilted and collapse side wards. Thus, the described cases indicate that the seismic motions were parallel to the direction of the respective arches, i. e. along a N-S direction.
Jointing	Administrative Tower Figure 2b The West Church and Administrative Tower Quarters Roman numbers denote residential quarters Arabic numbers the location of Figures Korjenkov and Mazor (2003)	10a Figure 10a Examples of jointing A 88cm joint crossing two adjacent stones in the West wall (175° trending) of the Administrative Tower Korjenkov and Mazor (2003) 10b Figure 10b Examples of jointing A 70cm joint crossing the lower part of a 178° trending arch support in a courtyard west of the Administrative Tower. Such joints are indication of a strong earthquake, around VII–VIII in the EMS-98 scale Korjenkov and Mazor (2003)	Observed damage pattern: At the western wall of the Administrative Tower, trending 178°, an 88cm long joint is seen crossing two stones (Fig.10 a). A 70cm long joint is seen at the lower support stone of a 178° trending arch, located in a room west of the Administrative Tower (Fig.10 b). Statistical observations: 12 through-going joints were observed. Interpretation: Joints crossing a few adjacent stones is one of the strong evidences of seismic origin of the deformations. Formation of such joints has been reported in many macroseismic studies. For example, Korjenkov and I. N. Lemzin described such joints formed in modern buildings during the Kochkor-Ata (Southern Kyrghyzstan) 1992 earthquake of a magnitude M_LH = 6.2. Such through-going joints are formed only as a result of high intensity earthquake – high energy is necessary to overcome the stress shadow of free surfaces at the stone margins (i. e., the free space between adjacent stones). ... At Mamshit the joints occurred together with the other listed seismic damage patterns.
Pushing of Walls by Connected Perpendicular Walls	Entire site	11 Figure 11 Both, clockwise and counterclock-wise rotations of adjacent stones in a wall, caused by a strong seismic push of a connected perpendicular wall Korjenkov and Mazor (2003)	Observed damage pattern: Clockwise and counterclockwise rotations of adjacent stones in a wall, caused by a push of a connected perpendicular wall (Fig. 11). Statistical observations: 6 cases of such pushes were observed in Mamshit ruins. Interpretation: A strong seismic event pushed the perpendicular wall. Hence, the seismic motions came along an axis parallel to the pushed wall. In the case of Mamshit this was along a N-S direction.
Percentage of Heavily Damaged Buildings	Entire site		The destroyed Roman buildings were rebuilt and, thus, many of the destroyed building parts were cleared away. The large number of deformation patterns that seen in the remaining parts of the Roman period buildings makes room to the assessment that practically all houses were damaged. Thus, the intensity of the tremor was IX EMS-98 scale or more.

Second Earthquake - 5th -7th centuries CE ? - Upper parts of buildings (repaired and built in the Byzantine Period)

Damage Type	Location	Figure	Comments
Tilting of Walls	S of West Church Figure 2b The West Church and Administrative Tower Quarters Roman numbers denote residential quarters Arabic numbers the location of Figures Korjenkov and Mazor (2003) Entire Site	12a Figure 12a Upper stones tilted 75° westward at a N-S (176°) trending wall, located at the west in a room south of the West Church Korjenkov and Mazor (2003) 12b Figure 12b Tilting westward of upper stones of the N-S (174°) trending east wall in a room south of the main premises of the West Church – stone A has a dip of 61° and stone B has a dip of 74° Korjenkov and Mazor (2003) 12c Figure 12c Tilting direction of the upper parts of walls of the Byzantine period, as a function of the wall directions – an overwhelming portion of the SES trending walls has been tilted to the WNW, and significantly less cases are seen in the perpendicular walls and with no direction preference Korjenkov and Mazor (2003) 12d Figure 12d Indicating the seismic push arrived from SW Korjenkov and Mazor (2003)	Observed damage pattern: The upper row of stones of a N-S (176°) trending wall, in a room south of the West Church, is tilted westward by an angle of 75° (Fig. 12 a). The upper stones of a wall trending N-S (174°), in a room south of the premises of the West Church, are also tilted westward, in an angle of 75° (Fig. 12 b). Statistical observations: 50 cases of tilting have been found on 145° to 185° trending walls, out of which 47 are tilted WSW and only 3 cases are tilted ENE (Fig. 12 c). In contrast, 50° to 100° trending walls revealed only 14 cases of tilting and with no systematic direction. Interpretation: The seismic pulses arrived from WSW.
Rotation of Wall Fragments around a Vertical Axis	E of West Church Figure 2b The West Church and Administrative Tower Quarters Roman numbers denote residential quarters Arabic numbers the location of Figures Korjenkov and Mazor (2003) House of Frescoes Figure 2c East Church and House of Frescoes Quarters Roman numbers denote residential quarters Arabic numbers the location of Figures Korjenkov and Mazor (2003) Entire Site	13a Figure 13a Examples of rotation Clockwise rotation (on 4°) of stones in the upper part of a N-S (172°) trending wall in a room at the Late Nabatean Building Korjenkov and Mazor (2003) 13b Figure 13b Examples of rotation Counterclockwise rotation (on 5°) of a part of a ENE (62°) trending wall at the House of Frescoes. Korjenkov and Mazor (2003) 13c Figure 13c Examples of rotation Direction and angle of rotation observed at the upper parts of walls, the Byzantine period, as a function of the trend of the walls – a clear directional preference is observable, leading to the conclusion. Korjenkov and Mazor (2003) 13d Figure 13d Examples of rotation Indicating the seismic push arrived from SW Korjenkov and Mazor (2003)	Observed damage pattern: A 4° clockwise rotation is seen in the upper part of a N-S (172°) trending wall, situated in a room of the Late Nabatean Building (Fig. 13 a). In contrast, a counterclockwise rotation of 5° is seen in part of an E-W (62°) trending wall in the House of Frescoes (Fig. 13 b). Statistical observations: Walls trending 60° to 85° reveal 9 cases of counter-clockwise rotation versus just 1 case of clockwise rotation (Fig. 13 c). In contrast, out of 14 cases of rotation in 155° to 180° trending walls, 13 are rotated clockwise, and only 1 counterclockwise. Interpretation: The seismic shocks arrived from SW, i.e. in the direction of the bisector to the trend of the walls (Fig. 13 d).

Additional Imprints of Severe Earthquakes

Damage Type	Location	Figure	Comments
Blocking of Entrances	West City Wall Figure 2a City plan of ancient Mamshit Korjenkov and Mazor (2003) XII quarter Figure 2c East Church and House of Frescoes Quarters Roman numbers denote residential quarters Arabic numbers the location of Figures Korjenkov and Mazor (2003)	14a Figure 14a Examples of rebuilding Blockage by smaller stones to support a damaged entrance at the west city wall, close to the SW corner – tilted stones at the south (right) side of the entrance indicate damage (marked by dashed lines) during a previous earthquake. Korjenkov and Mazor (2003) 14b Figure 14b Examples of rebuilding Blocking of an entrance in the east wall of a room in the XII quarter, in order to support the lintel that was cracked (marked by arrows) at an earlier earthquake. Korjenkov and Mazor (2003)	Observation: Fig. 14 a depicts a gate in the western city wall, close to its SW corner, that was blocked by smaller stones. The wall edge is tilted towards the former entrance, disclosing that the latter was blocked in order to support the wall that was damaged, most probably by an earthquake. The blocking stones are tilted as well, possibly disclosing the impact of another earthquake. Fig. 14 b shows an entrance in the eastern wall of a room of the XII quarter that was blocked to support the lintel that was cracked (marked by arrows), most possibly during a former earthquake. Statistical observations: 4 cases of blocked entrances one can observe in Mamshit ruins. Interpretation: Earthquake(s) is one of possible reasons for such type of building activity. ... So, the entrances in some places at Mamshit were possibly blocked in a number of cases in order to repair observable seismic damage. In other instances damaged structures had to be turn down and occasionally rebuilt.
Mismatch of Lower Stone Rows and Upper Parts of Buildings	E of East Church Figure 2c East Church and House of Frescoes Quarters Roman numbers denote residential quarters Arabic numbers the location of Figures Korjenkov and Mazor (2003)	15 Figure 15 Rebuilding disclosed by the protrusion of the lowest row of stones at the west wall of a room, east of the East Church (the recent restoration line is higher above – marked by arrows) Korjenkov and Mazor (2003)	Observation: The lower row of stones of the western wall of a room, east of the East Church, protrudes from the plane of the rest of the wall (Fig. 15). Statistical observations: 12 cases of mismatching were observed in Mamshit. Interpretation: Two stages of building are disclosed: the original structure was destroyed by an earthquake, dismantled, and a new wall was built, using the old foundation, but following a somewhat different direction. Such phenomenon was also observed in adjacent ruins of ancient Avdat and Shivta, which were damaged by strong historical earthquakes.
Supporting Walls	South City Wall Figure 2a City plan of ancient Mamshit Korjenkov and Mazor (2003)	16 Figure 16 A 66° trending section of the south city wall, tilted by 81° to SES (marked by a dashed line), supported by an added wall (shown by an arrow). Part of the support wall was dissembled during the archeological excavations, to expose the tilting of the original wall Korjenkov and Mazor (2003)	Observation: Fig. 16 discloses a section of the southern city wall (trending 66°) that is tilted by 81° to SES (marked by a dashed line), and connected to it are seen the remains of a special support wall (shown by an arrow). Part of the support wall was dissembled during the archeological excavations, to expose the tilting of the original wall. Statistical observations: One supporting wall was observed in Mamshit ruins. Interpretation: Various segments of city wall were tilted at an earlier earthquake (most probably during the Roman period) and repaired later on (most probably during the Byzantine stage of rebuilding). Such supporting walls were observed in another cities in the Negev desert, like Avdat, Shivta, Rehovot-ba-Negev and Sa’adon. Together with another "pure" features of the seismic deformations, they can be used as additional supportive evidence of earthquake damage.
Secondary Use of Building Stones	East Church Figure 2c East Church and House of Frescoes Quarters Roman numbers denote residential quarters Arabic numbers the location of Figures Korjenkov and Mazor (2003)	17a Figure 17a Examples of secondary use of building stones Section of a column reused as part of a bench along the west wall of the main hall of the East Church. Korjenkov and Mazor (2003) 17b Figure 17b Examples of secondary use of building stones E wall of a room at East Church Quarter: the lower-right part protrudes 7 to 12cm, as compared to the upper-left part, that is built of reused smaller stones (the contact is marked by + signs) Korjenkov and Mazor (2003)	Observation: Fig. 17 a shows a secondary use of a segment of a column, western wall of the main hall of the East Church. Fig. 17 b displays the eastern wall of a room at the East Church quarter, disclosing a lower- right part that protrudes 7 to 12cm, as compared to the upper-left part that is built of reused smaller stones, disclosing a stage of repair and rebuilding. Statistical observations: 9 walls with secondary use of building stones were found in Mamshit. Interpretation: The rather common secondary use of building materials in the Byzantine buildings may well reflect the destruction of the Roman buildings that were severely damaged by the earthquake that is identified by the long list of damage patterns discussed so far.
Incorporation of Wooden Beams in Stone Buildings	Administrative Tower Figure 2b The West Church and Administrative Tower Quarters Roman numbers denote residential quarters Arabic numbers the location of Figures Korjenkov and Mazor (2003)	18a Figure 18a Examples of the incorporation of elastic wooden beams in stone buildings Wooden beam incorporated as second lintel above a door in a room at the Administrative Tower Korjenkov and Mazor (2003) 18b Figure 18b Examples of the incorporation of elastic wooden beams in stone buildings Wooden beam incorporated at the same building between two doorsteps Korjenkov and Mazor (2003)	Observation: A high quality wooden beam is incorporated as a second lintel above a door in a room at the Administrative Tower (Fig. 18 a). Another beam is incorporated in the same building between two door steps (Fig. 18 b). Statistical observations: 2 cases of wooden beams were found in Mamshit ruins. Interpretation: The builders of Mamshit were aware of the seismic danger and incorporated wooden beams to absorb future seismic shocks. Horizontally placed beams lowered mainly the effect of the vertical component of seismic motions. Laying inside the walls of a regular longitudinal-diametrical framework from the wooden beams is a typical antiseismic method of Medieval Turkish construction noticed by A. A. Nikonov (1996) during his archeo-seismological study in Crimean Peninsular.
Bulging of Wall Parts	West City Wall Figure 2a City plan of ancient Mamshit Korjenkov and Mazor (2003)	19a Figure 19a Westward bulging of the central part of the western city wall, trending SES (152°). Korjenkov and Mazor (2003) 19b Figure 19b The angles of the displaced stones Korjenkov and Mazor (2003)	Observation: The central part of the western city wall, trending SES (152°), is bulged westwards, as is seen in Fig. 19 in a photo and a sketch. Statistical observations: 11 cases of bulging of central parts of the walls were observed in Mamshit. Interpretation: The city wall is well built of massive stones and, thus, deformation due to poor building most probably can be ruled out. This seems to be the result of a strong earthquake.
Percentage of Heavily Damaged Buildings	Entire Site		Practically all the buildings of the Byzantine period were damaged, more that 50% are estimated to have been destroyed. Thus, the intensity of the tremor was IX at the EMS-98 scale or more.

First Earthquake - Early Byzantine

Korzhenkov and Mazor (2003) provided the following analysis for the first earthquake:

The Lower Parts of the Buildings, Reflecting Mainly the Earthquake of the End of the 3rd cent. or Beginning of the 4th cent.

The walls of the houses of Mamshit had a general orientation of around ENE (~ 75°) and SES (~165°). Hence, a quadrangle of these directions may serve as the basis for a general discussion of the observed damage patterns, in order to deduce the direction of arrival of the seismic movements.

Arrival of the seismic motions from north has been concluded for the 4th cent. event. Let us discuss in this context three possibilities:

If the strong seismic pulses would have arrived from NWN, the walls perpendicular to this direction (ENE) would experience quantitative and systematic tilting (as well as collapse) toward the epicenter, whereas the perpendicular walls (SES) would have distinctly less cases of tilting and they would be in random to both NEN and NWN (Fig. 20 a

Figure 20a

Analysis of the direction of the epicenter in the devastating earthquake of the 4th cent. Systematic of expected damage patterns

a. If the seismic movement would have arrived from NWN

Korjenkov and Mazor (2003)

). Rotations would be scarce and at random directions. This is not the case of the lower parts of buildings (Roman period) at Mamshit.

If the strong seismic shocks would have arrived along the bisector of the trend of the walls (i.e. from NEN), the walls trending ENE would have undergone both systematic tilting toward NWN and anticlockwise rotation, whereas the perpendicular walls (trending SES) would experience systematic tilting toward NEN and clockwise rotation (Fig. 20 b

Figure 20b

Analysis of the direction of the epicenter in the devastating earthquake of the 4th cent. Systematic of expected damage patterns

b. If the epicenter would have been at NEN

Korjenkov and Mazor (2003)

), but this is not the case of the lower parts of buildings (Roman period) at Mamshit.

If the epicenter was at the north, the ENE trending walls would undergo systematic tilting to the NWN and systematic counterclockwise rotations, whereas the SES trending walls would suffer of a few cases of random tilting but systematic clockwise rotations (Fig. 20 c

Figure 20c

Analysis of the direction of the epicenter in the devastating earthquake of the 4th cent. Systematic of expected damage patterns

c. If the seismic movement came from north – the case met by the observations

Korjenkov and Mazor (2003)

). This combination of damage pattern orientations fits the observations at the lower parts of the buildings at Mamshit, leading to the conclusion that the epicenter of the devastating earthquake at the end of the 3rd cent. or beginning of the 4th cent. was north of Mamshit.

The systematic directional deformation patterns disclose that the hypocenter was not beneath Mamshit, but to the north of it, and the concluded intensity of IX or more, suggests the epicenter was in several-first tens of km away. Future field investigations are recommended to check for evidence of recent tectonic activity in the Judean Desert.

Second Earthquake - 5th -7th centuries CE ?

Korzhenkov and Mazor (2003) provided the following analysis for the second earthquake:

The Upper Parts of the Buildings, Reflecting Mainly the 7th cent. Earthquake

The direction of the epicenter of the 7th cent. strong earthquake has been concluded to have been SW of Mamshit. In this connection let us examine three possibilities, bearing in mind that the walls of the houses of Mamshit had a general orientation of around ENE (~ 75°) and SES (~165°):

If the strong seismic shocks would have arrived from WSW, the walls perpendicular to this direction (SES) would experience quantitative and systematic tilting toward the epicenter, whereas the perpendicular walls (ENE) would have distinctly less cases of tilting and they would be in random directions and not to the epicenter (Fig. 21 a

Figure 21a

Analysis of the direction of the epicenter in the devastating earthquake of the 7th cent. Systematic of expected damage patterns

a. If the epicenter would be at WSW

Korjenkov and Mazor (2003)

). Rotations would be scarce and at random directions. This is not the case of the upper parts of buildings (Byzantine period) at Mamshit.

If the strong seismic pulses would have arrived along the bisector of the trend of the walls (i.e. from SWS), the walls trending ENE would have under¬gone both systematic tilting toward NWN and counterclockwise rotation, whereas the perpendicular walls (trending SES) would experience systematic tilting toward NEN and clockwise rotation (Fig. 21 b

Figure 21b

Analysis of the direction of the epicenter in the devastating earthquake of the 7th cent. Systematic of expected damage patterns

b. If the seismic waves would have come from NEN

Korjenkov and Mazor (2003)

), but this is not the case of the upper parts of buildings (Byzantine period) at Mamshit.

If the epicenter was at SW, the SES trending walls would undergo systematic tilting to the SW and systematic clockwise rotations, whereas the ENE trending walls would suffer of a few cases of random tilting but systematic counterclockwise rotations (Fig. 21 c

Figure 21c

Analysis of the direction of the epicenter in the devastating earthquake of the 7th cent. Systematic of expected damage patterns

c. If the seismic movement came from SW – the case met by the observations

Korjenkov and Mazor (2003)

). This combination of damage pattern orientations fits the observations at the upper parts of the buildings at Mamshit, leading to the conclusion that the epicenter of the devastating seventh century earthquake was SW of Mamshit.

The systematic directional deformation patterns disclose that the hypocenter was not beneath Mamshit, but to the SW of it, and the concluded intensity of IX or more suggests the epicenter was in several-first tens of km away. Future field investigations are recommended to check for evidence of recent tectonic activity along E-W trending faults in the Negev Desert.

First Earthquake - Early Byzantine

Earthquake Archeological Effects chart

Earthquake Archeological Effects (EAE)

Rodríguez-Pascua et al (2013: 221-224)

of Rodríguez-Pascua et al (2013: 221-224)

Effect	Location	Intensity
Tilted Walls	E of West Church	VI+
Displaced Masonry Blocks	E of West Church ENE of West Church Near Frescoes House	VIII+
Folded Steps and Kerbs	Administrative Tower	VI+
Dropped Keystones in Arches	W of Eastern Church Stables - E of West Church	VI+
Penetrative fractues in Masonry Blocks	Administrative Tower	VI+
Displaced Walls	Entire Site	VII+
Collapsed Walls	Entire Site	VIII+

The archeoseismic evidence requires a minimum Intensity of VIII (8) when using the Earthquake Archeological Effects chart of Rodríguez-Pascua et al (2013: 221-224).

Korjenkov and Mazor (2003)'s Seismic Characterization

This was a strong earthquake with an epicenter at the north, and an EMS-98 scale intensity of IX or more. This is a minimum value because the wrecks of the most badly struck buildings had most probably been completely removed, leaving no trace. Thus, our observations are biased toward the lower end of the intensity scale.
...
In the present study the two earthquakes were resolved by the archeological identification that the Roman town was rebuilt at the Byzantine period, and the latter fell into ruins as well. The archeoseismological resolution of the two earthquakes is validated in the present case by the observation that the epicenters were at different directions – north in the first event and SW in the second.
...
The percentage of collapsed buildings of the Roman town is hard to estimate as most of them have been cleared away and rebuilt. Yet, an estimate can be done by the extended rebuilding - most of the second floors or upper parts of high structures were rebuilt at the Byzantine stage, leading to an estimate that at lest 15% of the Roman period buildings were destroyed at the end of the 3rd cent. or beginning of the 4th cent. earthquake. Thus, according to the European Macroseismic Scale of 1998 (EMS-98) an earthquake intensity of IX or more is concluded.
...
Zero distance is ruled out in both studied earthquakes on the basis that most of the observed seismic deformations were caused by lateral movements. Hence, the hypocenter was not beneath Mamshit.
...
The observed dominance of lateral movements in both earthquakes indicates the epicenter was away at some distance from the epicenter. Future studies will have to address this point.
...
The large body of damage patterns surveyed at Mamshit provides a fairly simple picture: devastation was caused mainly by lateral movements that arrived from the fault rupture zone. These observations were made for both earthquakes – the one at the end of the 3rd cent. or beginning of the 4th cent. that had its epicenter at the north, and the second at the 7th cent. that had its epicenter at SW.

Discontinuous Deformation Analysis by Kamai and Hatzor (2005)

Kamai and Hatzor (2005) performed Discontinuous Deformation Analysis (DDA) on a model

Figure 6

The final mesh configuration for the embedded arch in Mamshit. The uniform masonry wall rests on two blocks: the lower is fixed, and the overlying block can be subjected to time-dependant displacements. The height of the wall above the arch is h.

Kamai and Hatzor (2005)

for a dropped keystone in an arch

near the Eastern Church in Mampsis. The optimal model

, using a sinusoidal input with an amplitude of 0.5 g and a frequency of 1 Hz., produced 3.11 cm. of displacement vs. 4 cm. measured in the field. The conclusion was that the keystone dropped due an earthquake with a PGA of ~0.5 g and a center frequency of ~1 Hz.. 0.5 g translates to an Intensity of 8.2 using Equation 2 of Wald et al (1999). In their modeling, Kamai and Hatzor (2005) found that low frequencies (e.g. 0.5 Hz.) resulted in strong fluctuations and high frequencies (e.g. 5 and 10 Hz.) resulted in a "locking" of the structure and very little displacement. Accelerations between 0.32 and 0.8 g produced reasonable values of keystone displacement although 0.5 g produced the most amount of displacement and the closest amount of displacement to what is observed in the field.

Kamai and Hatzor (2007) reiterated the same study at Mampsis noting that keystone displacement only occurred in the frequency range of 1.0 - 1.5 Hz. and that seismic amplification might have been at play at the higher parts of the structure (i.e. the "Sky-scraper effect" mentioned by Korzhenkov), thus amplifying bedrock PGA by as much as 2.5. This led to a bracket of PGA values for the dropped keystone between 0.2 and 0.5 g. These PGA values convert to Intensities of 6.7 - 8.2 using Equation 2 of Wald et al (1999).

Variable	Input	Units	Notes
PGA		g	Peak Horizontal Ground Acceleration
Variable	Output - Site Effect not considered	Units	Notes
I		unitless	Conversion from PGA to Intensity using Wald et al (1999)

Model Values and Lab derived properties

Model Values

Property	Value	Units
Friction angle of arch	35	degrees
Friction angle of wall	40	degrees
Young's Modulus of arch	17	Gpa
Young's Modulus of wall	1	Mpa
Height of Wall above arch	0	m

Model was run in qk.mode. An unusually low model value of Young's Modulus for the wall (1 Mpa) was explained as reasonable when one considers the heterogeneity of the wall where spaces between the wall blocks are filled with soft filling materials.

Lab Measurements of original stones from Mampsis

Property	Value	Units
Density	1890	kg./m³
Porosity	30 - 38	%
Dynamic Young's Modulus	16.9	Gpa
Dynamic Shear Modulus	6.17	Gpa
Dynamic Poisson's Ratio	0.37	unitless
Interface friction angle	35	degrees

Second Earthquake - 5th -7th centuries CE ?

Earthquake Archeological Effects chart

Earthquake Archeological Effects (EAE)

Rodríguez-Pascua et al (2013: 221-224)

of Rodríguez-Pascua et al (2013: 221-224)

Effect	Location	Intensity
Tilted Walls	S of West Church Entire Site	VI+
Displaced Masonry Blocks	E of West Church House of Frescoes	VIII+
Collapsed Walls	Entire Site	VIII+

The archeoseismic evidence requires a minimum Intensity of VIII (8) when using the Earthquake Archeological Effects chart of Rodríguez-Pascua et al (2013: 221-224)

Korjenkov and Mazor (2003)'s Seismic Characterization

At the end of the Byzantine period a second earthquake hit the place, the epicenter being this time to the SW, and the intensity was IX or more.
...
The percentage of collapsed buildings of the Byzantine town can be well estimated as the ruins were left untouched. The survey disclosed that at least 15% of the well built stone buildings of Byzantine Mamshit collapsed – practically no second floor structures survived with no severe damage. Hence, according to the EMS-98 an earthquake intensity of IX or more is deduced as well.
...
Zero distance is ruled out in both studied earthquakes on the basis that most of the observed seismic deformations were caused by lateral movements. Hence, the hypocenter was not beneath Mamshit.
...
The observed dominance of lateral movements in both earthquakes indicates the epicenter was away at some distance from the epicenter. Future studies will have to address this point.
...
The large body of damage patterns surveyed at Mamshit provides a fairly simple picture: devastation was caused mainly by lateral movements that arrived from the fault rupture zone. These observations were made for both earthquakes – the one at the end of the 3rd cent. or beginning of the 4th cent. that had its epicenter at the north, and the second at the 7th cent. that had its epicenter at SW.

References

Articles and Books

Erickson-Gini T. 1999 Mampsis: A Nabataean Roman Settlement in the Central Negev Highlands in Light of the Ceramic and Architectural Evidence Found in Archaeological Excavations During 1993 1994, Unpublished M.A. dissertation, Tel Aviv University.

Erickson-Gini, New Excavations in the Late Roman Quarter in Avdat, Proceedings of the Twenty-Seventh Archaeological Congress in Israel, Bar Ilan University April 2–3, 2001

Erickson-Gini, T. (2004). Crisis and Renewal-settlement in the Central Negev in the Third and Fourth Centuries C. E.: With an Emphasis on the Finds from Recent Excavations in Mampsis, Oboda and Mezad 'En Hazeva, Hebrew University of Jerusalem.

Erickson-Gini, T. (2010:47). Nabataean settlement and self-organized economy in The Central Negev: crisis and renewal, Archaeopress.

Erickson-Gini, T. (2010). Nabataean settlement and self-organized economy in the Central Negev : crisis and renewal, Oxford : Archaeopress.

Kamai, R. and Y. Hatzor (2005). Dynamic back analysis of structural failures in archeological sites to obtain paleo-seismic parameters using DDA. Proceedings of 7th International Conference on the Analysis of Discontinuous Deformation (ICADD-7).

Kamai, R. and Y. H. Hatzor (2008). "Numerical analysis of block stone displacements in ancient masonry structures: A new method to estimate historic ground motions." International Journal for Numerical and Analytical Methods in Geomechanics 32(11): 1321-1340.

Korzhenkov, A. and E. Mazor (2003). "Archeoseismology in Mamshit (Southern Israel): Cracking a Millennia-old Code of Earthquakes Preserved in Ancient Ruins." Archäologischer Anzeiger: 51-82.

A. Negev (1971), The Nabatean Necropolis of Mamshit (Kurnub), IsrExplJ 21, 1971, 110–129

Negev, A. (1974). "THE CHURCHES OF THE CENTRAL NEGEV AN ARCHAEOLOGICAL SURVEY." Revue Biblique (1946-) 81(3): 400-421.

Sion, O., Israeli, Shoshana (2022). Mampsis:Mapping the City Following a New Survey. Cities, Monuments, and Objects in the Roman and Byzantine Levant, Archaeopress: 247-260.

Excavation Reports

Negev, A. (1988). The architecture of Mampsis : final report. 1. The Middle and Late Nabatean periods, Hebrew University of Jerusalem.

Negev, A. (1988) The Architecture of Mampsis, Final Report, Vol. II: The Late Roman and Byzantine Period, Hebrew University of Jerusalem.

Websites

Mamshit at biblewalks.com

Mamshit National Park

Mamshit Park Brochure

Wikipedia page for Mampsis

Photos

Description	Photo
Dropped Keystone of Mamshit Front View Long Shot	Dropped Keystone of Mamshit Front View Long Shot Photo by Jefferson Williams 12 Jan. 2023
Dropped Keystone of Mamshit Rear View Long Shot	Dropped Keystone of Mamshit Rear View Long Shot Photo by Jefferson Williams 12 Jan. 2023
Dropped Keystone of Mamshit Front View Closeup	Dropped Keystone of Mamshit Front View Closeup Photo by Jefferson Williams 12 Jan. 2023
Dropped Keystone of Mamshit Orientation Az = 174°	Dropped Keystone of Mamshit Orientation Az = 174° Photo by Jefferson Williams 12 Jan. 2023
Arch of Dropped Keystone of Mamshit Corner Crack of Voussoir	Arch of Dropped Keystone of Mamshit Corner Crack of Voussoir Photo by Jefferson Williams 12 Jan. 2023
Arch of Dropped Keystone of Mamshit Corner Crack of Voussoir - another side	Arch of Dropped Keystone of Mamshit Corner Crack of Voussoir - another side Photo by Jefferson Williams 12 Jan. 2023
Arch of Dropped Keystone of Mamshit Closeup on Spalled Voussoir	Arch of Dropped Keystone of Mamshit Closeup on Spalled Voussoir Photo by Jefferson Williams 12 Jan. 2023
Cracked Lintel in Mamshit Orientation not recorded Long Shot	Cracked Lintel in Mamshit Orientation not recorded Long Shot Photo by Jefferson Williams 12 Jan. 2023