Open this page in a new tab

Jerash - Macellum

Orthophoto of the Macellum

Dataset on figshare provided by Stott, Raja, and Lichtenberger (2019)


Introduction
ChatGPT Introduction

The macellum of Gerasa was built in the first half of the 2nd century CE and functioned as a formal food market. Measuring approximately 50×50 m, it was centered on an octagonal courtyard paved in limestone slabs and enclosed by a peristyle of 24 Corinthian columns.

Its eastern entrance opened directly onto the Cardo Maximus and featured a monumental façade with 10 m-high columns, a triple-arched gateway, and a vestibule. Flanking tabernae provided commercial frontage. A fountain in the northeast corner bears a Greek inscription honoring Julia Domna, wife of Septimius Severus, suggesting construction around 193–211 CE.

In the Late Byzantine period, the building was repurposed for industrial functions. Archaeological evidence indicates the presence of a tinctoria (dye workshop), a lime kiln, and pottery production. These transformations reflect urban adaptation to new economic realities by the 6th and early 7th centuries CE.

A major collapse layer was found over the southern exedra and sealed by a stone retaining wall. The authors propose that this may result from a destructive event in the early 7th century CE—possibly linked to the Sasanian invasion in 614 CE, an earthquake, or early Islamic warfare. The exact cause is undetermined.

Despite this destruction, limited reuse followed during the Umayyad and Abbasid periods. Archaeological traces include grain storage pits and ceramic kilns, suggesting that portions of the macellum remained in use until the late 8th century CE, when the complex was ultimately abandoned.

Jerash - Introduction Webpage

Aerial Views and Plans
Aerial Views, Plans, and Sections

Aerial Views

  • Jerash Macellum in Google Earth

Plans and Sections

Site Plans

  • General Plan of Jerash from Wikipedia
  • Fig. 2 - Plan of Umayyad Jerash from Walmsley and Daamgaard (2005)

Area Plans and Sections - Macellum

Normal Size

  • Fig. 1 - Plan of the Macellum from Uscatescu and Marot (2000)
  • Fig. 7.1 - Location of the early Islamic contexts and structures from Uscatescu and Marot (2000)
  • Fig. 2.1 - Plan showing location of mid 7th century CE collapse layer from Uscatescu and Marot (2000)
  • Fig. 6.1 - Islamic wall with mid-7th century collapse layer to the right from Uscatescu and Marot (2000)

Magnified

  • Fig. 1 - Plan of the Macellum from Uscatescu and Marot (2000)
  • Fig. 7.1 - Location of the early Islamic contexts and structures from Uscatescu and Marot (2000)
  • Fig. 2.1 - Plan showing location of mid 7th century CE collapse layer from Uscatescu and Marot (2000)
  • Fig. 6.1 - Islamic wall with mid-7th century collapse layer to the right from Uscatescu and Marot (2000)

Chronology
Phasing

Uscatescu & Marot (2000)

Phase Period Date Description
I Roman early 2nd century CE Construction of the macellum with octagonal courtyard, peristyle, and monumental entrance onto the Cardo Maximus. Greek inscription dedicated to Julia Domna.
II Late Roman–Byzantine late 5th – early 6th century CE Addition of southern row of tabernae, installation of lime kiln, and reuse of interior rooms for industrial purposes. Function shifts toward dye production and ceramic workshops.
III Early Islamic early to mid-7th century CE Collapse of southern exedra and construction of retaining wall. Possible destruction linked to Sasanian invasion, an earthquake, or early Islamic conflicts.
IV Umayyad–Abbasid late 7th – late 8th century CE Reoccupation with restricted functions—grain storage, pottery kilns, limited rebuilding. No architectural renewal. Final abandonment by end of 8th century CE.

mid 7th century CE Destruction

Islamic Wall and mid 7th c. collapse layer at Jerash Macellum Figure 6.1

Islamic wall between south peristyle columns and exedra 3

JW: Mid 7th century CE collapse layer to right

Uscatescu and Marot (2000)


Discussion

Uscatescu and Marot (2000:283) dated the seismic destruction of the macellum “at the latest to the second quarter of the seventh century” based on pottery and coin assemblages.

This destruction layer was sealed, undisturbed, and securely stratified. In a revised discussion, Uscatescu and Marot (2000:281) acknowledged that “this destruction was previously dated to the early seventh century, but now we are compelled to withdraw this date and propose, on the basis of both pottery and coin data, a mid-seventh century chronology for this collapse, that is some 20 years later than the first proposed chronology.” However, they stress that this revision “does not affect the established pre-Islamic phases of the building.”

Pottery and coins used in this redating effort were previously published in Marot (1998) and Uscatescu (1995, 1996).

Further evidence of structural loss includes the removal of exedra 4, which had been cleared prior to the construction of Islamic vaulted galleries. This spatial transformation confirms that the collapse and clearance were complete before early Islamic re-use.

References
Uscatescu and Marot (2000)

Notes by JW

Uscatescu and Marot (2000:283) dated seismic destruction of the Macellum to at the latest to the second quarter of the seventh century based on pottery and coins1. The seismic destruction layer was found in a sealed and undisturbed context and is well-dated. Uscatescu and Marot (2000:281) discussed a dating revision as follows:

This destruction was previously dated to the early seventh century, but now we are compelled to withdraw this date and propose, on the basis of both pottery and coin data, a mid-seventh century chronology for this collapse, that is some 20 years later than the first proposed chronology. However, this corrected date for the Macellum destruction does not affect the established pre-Islamic phases of the building.
In addition to the collapse layers found throughout the Macellum, vaulted Islamic galleries were constructed over where exedra 4 once stood. It can be presumed that exedra 4 was demolished and cleared away before construction of the galleries began.
Footnotes

1 Pottery and coins from the Macellum have been already published separately. Detailed information on the coins and pottery can be found in Marot (1998) and Uscatescu (1995; 1996)

Chat GPT Summary of Archaeoseismic Evidence

A major collapse layer was found over the southern exedra of the macellum. This collapse consisted of architectural debris and rubble that sealed earlier phases of the building. A retaining wall was built above the destruction level to prevent further subsidence.

The authors interpret this collapse as marking a significant disruptive event, though they stop short of attributing it definitively to a single cause. They propose three possible explanations: (1) the Sasanian invasion in 614 CE, (2) the 631–632 CE earthquake recorded in Byzantine sources, or (3) early Islamic conflict-related destruction. The collapse layer provides a terminus post quem for these events.

No explicit faulting or directional collapse patterns were reported. However, the sudden nature of the destruction, stratigraphic sealing, and architectural displacement all suggest a rapid catastrophic event, compatible with earthquake damage. Later reuse of the building in the Umayyad and Abbasid periods confirms that the structure was never fully restored after this event.

2nd half of the 8th / Early 9th CE Destruction Layer

Discussion

After the mid-7th century CE collapse, the Uscatescu and Marot (2000:299) report that the Macellum remained in use, with new Islamic-period structures constructed over earlier ruins. A second destruction layer was identified over the Islamic galleries. It consisted of fallen ashlar blocks and voussoirs, originating from collapsed vaults and walls. However, this layer was disturbed: archaeological analysis indicated that 88.5% of the pottery and 92.68% of the coins were residual finds, not directly associated with the destruction event.

Only four ceramic sherds were securely dated to the late Umayyad period. These included a cooking pot, a handmade grey basin, and a local grey amphora. The rest were earlier survivals, such as Cypriot Late Roman D ware (Hayes form 9B), which terminates in the late 7th century. No Islamic coins were recovered, making it difficult to assign a precise date to the collapse (Uscatescu and Marot 2000:298–299).

Some evidence of limited reuse was found above the destruction layer at exedra 3, where Uscatescu and Marot (2000:299) describe “a very short occupation” in the Abassid period, evidenced by small fireplaces, burnt patches, several intact cooking pots, and dark BGW vessels. This activity, based on stratigraphy alone, is dated to the second half of the 8th century CE. No coins were recovered from this level either.

Dating the final abandonment of the building remains difficult. Although the stratigraphic record suggests a second major collapse, likely in the second half of the 8th or early 9th centuries CE, Uscatescu and Marot (2000:299) caution that “there is not a single undisturbed context that can be surely dated in the early Abbasid period,” aside from the short occupation phase in exedra 3.

References

Uscatescu and Marot (2000)

Notes by JW

Uscatescu and Marot (2000:299) report that the building continued in use after the mid 7th century CE destruction with new structures built over parts of the ruins. A later destruction layer was present which Uscatescu and Marot (2000:298-299) discussed as follows:

The Destruction Layer of Late Umayyad/Early Abbasid Chronology over the Islamic Galleries

The Islamic galleries were covered by a destruction level composed of ashlar blocks and voussoirs from the fallen walls and vaults. Archaeological analysis proves that it is also a disturbed layer, since the residual sherds account for 88.5 per cent of the total. In the case of the coins, the percentage of residuality is much higher, accounting for 92.68 percent (Table 4).

Unfortunately, only four sherds can be dated to the late Umayyad period; the rest are rubbish survival, including some transitional shapes such as imported Cypriot Late Roman D (Hayes form 9B), which has an end date of the late seventh century (Hayes 1972: 382). The Umayyad pottery is limited to a cooking-pot (Figure 9.6), a handmade grey basin (Figure 9.12) and a probably local grey amphora (Figure 9.13). The absence of any Islamic coins within this context does not help when attempting to fix a more accurate date to the collapse.
Some traces of an early Abbasid occupation over the destruction level were found at exedra 3 (Table 5b) which they discussed as follows:
This evidence points to a very short occupation, with some burnt patches identified as small fireplaces and several complete cooking-pots (Figure 9.3, 6) and some dark BGW (Figure 9.9, 11). Therefore, this level should be dated, at least, to the second half of the eighth century on stratigraphical basis. No coins were recorded.
UUscatescu and Marot (2000:299) noted chronological difficulties in dating final (destruction) and abandonment.
It is difficult to ascertain the chronology of the second and final abandonment of the building. But most of the archaeological evidence recorded pointed to the second half of the eighth or early ninth centuries. Unfortunately, there is not a single undisturbed context that can be surely dated in the early Abbasid period, with the exception of sporadic occupation in exedra 3.

Chat GPT Summary of Archaeoseismic Evidence

A second major destruction layer was identified above the Islamic vaulted galleries in the Macellum at Jerash. It consisted of a dense accumulation of fallen ashlar blocks and voussoirs, collapsed from the galleries’ walls and vaults. The layer appears to have been deposited rapidly, but its integrity is compromised: 88.5% of the ceramic material and 92.68% of the coins were identified as residual.

Despite the high residuality, four diagnostic ceramic sherds— including a cooking pot, a handmade grey basin, and a grey amphora—support a terminus post quem in the late Umayyad period, while imported Cypriot Late Roman D (Hayes Form 9B) further constrains the terminus ante quem to the end of the 7th century CE. The destruction event itself is thus tentatively dated to the late 8th or early 9th century CE. The absence of Islamic coins from this level hinders finer chronological resolution.

A short-lived early Abbasid reoccupation was noted in exedra 3, above the destruction layer, and consisted of small hearths and intact cooking vessels. This episode, though ephemeral, supports a terminus post quem in the second half of the 8th century CE.

Uscatescu and Marot (2000:299) acknowledge the difficulty in firmly assigning a cause to this second collapse, noting that no single undisturbed context is available for dating the event with certainty. Nonetheless, the stratigraphic sequence and associated material suggest a sudden and damaging episode, possibly of seismic origin, occurring during the late Umayyad to early Abbasid transition.

Seismic Effects
mid 7th century CE Destruction

Effect Location Image(s) Description
  • Collapsed Vault            
  • Collapsed walls
  • Fallen columns
  • Collapsed arches
  • Collapsed roof
    (displaced walls?)
  • Fallen architraves
    (Collapsed walls)
  • Debris
throughout Macellum

  • The late Byzantine craft-activity was abandoned sometime before the mid-seventh century destruction. It does not seem that this destruction was the reason for the break in human activity within the building, since nothing valuable was found under the collapse layer and the state of the ceramic finds was very fragmentary, with the exception of some broken amphorae found over the earthen floor of the abandoned officina tinctoria (Figure 5.1-2, 7). This extensive destruction is well evidenced by the fallen vaulted and tiled roofs and collapsed walls; a huge collapse that reaches a thickness of more than two and a half metres, and was composed by voussoirs, tiles, ashlars, architraves, column shafts, capitals and other architectonic elements. - Uscatescu and Marot (2000:281)
  • Collapsed Vault            
exedra 4
  • JW: In addition to the collapse layers found throughout the Macellum, vaulted Islamic galleries were constructed over where exedra 4 once stood. It can be presumed that exedra 4 was demolished and cleared away before construction of the galleries began.
  • A construction deposit of Umayyad date was located behind the rear wall of the two vaulted galleries built over the demolished exedra 4 - Uscatescu and Marot (2000:297)

2nd half of the 8th / Early 9th CE Destruction Layer

Effect Location Image(s) Description
  • Collapsed Vault            
  • Collapsed Walls            
  • Destruction layer
Islamic galleries over exedra 4
  • The Islamic galleries were covered by a destruction level composed of ashlar blocks and voussoirs from the fallen walls and vaults. - Uscatescu and Marot (2000:288-289)

Deformation Maps
mid 7th century CE Destruction

Deformation Map

Modified by JW from Fig.s 1 and 6.1.1 of Uscatescu and Marot (2000)

2nd half of the 8th / Early 9th CE Destruction Layer

Deformation Map

Modified by JW from Fig. 1 of Uscatescu and Marot (2000)

Intensity Estimates
mid 7th century CE Destruction

Effect Location Image(s) Description Intensity
  • Collapsed Vault            
  • Collapsed walls
  • Fallen columns
  • Collapsed arches
  • Collapsed roof
    (displaced walls?)
  • Fallen architraves
    (Collapsed walls)
  • Debris
throughout Macellum

  • The late Byzantine craft-activity was abandoned sometime before the mid-seventh century destruction. It does not seem that this destruction was the reason for the break in human activity within the building, since nothing valuable was found under the collapse layer and the state of the ceramic finds was very fragmentary, with the exception of some broken amphorae found over the earthen floor of the abandoned officina tinctoria (Figure 5.1-2, 7). This extensive destruction is well evidenced by the fallen vaulted and tiled roofs and collapsed walls; a huge collapse that reaches a thickness of more than two and a half metres, and was composed by voussoirs, tiles, ashlars, architraves, column shafts, capitals and other architectonic elements. - Uscatescu and Marot (2000:281)
  • VIII+
  • VIII+
  • V+
  • VI+
  • VII+?
  • VIII+
  • ?
  • Collapsed Vault            
exedra 4
  • JW: In addition to the collapse layers found throughout the Macellum, vaulted Islamic galleries were constructed over where exedra 4 once stood. It can be presumed that exedra 4 was demolished and cleared away before construction of the galleries began.
  • A construction deposit of Umayyad date was located behind the rear wall of the two vaulted galleries built over the demolished exedra 4 - Uscatescu and Marot (2000:297)
  • 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).

2nd half of the 8th / Early 9th CE Destruction Layer

Effect Location Image(s) Description Intensity
  • Collapsed Vault            
  • Collapsed Walls            
  • Destruction layer
Islamic galleries over exedra 4
  • The Islamic galleries were covered by a destruction level composed of ashlar blocks and voussoirs from the fallen walls and vaults. - Uscatescu and Marot (2000:288-289)
  • VIII+
  • 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).

Notes and Further Reading
References