Edinburgh Research Explorer
Sasanian walls, hinterland fortresses and abandoned ancient irrigated landscapes Citation for published version: Sauer, E, Usher-Wilson, LS, Ratcliffe, J, Shabani, B, Omrani Rekavadi, H, Wilkinson, T, Abbasi, GA, S, P, Safari Tamak, E, Ainslie, R, Mahmoudi, M, Galiatsatos, N, Roustai, K, Jansen Van Rensburg, J, Ershadi, M, MacDonald, E, Fattahi, M, Oatley, C & Shabani, B 2008, 'Sasanian walls, hinterland fortresses and abandoned ancient irrigated landscapes: the 2007 season on the Great Wall of Gorgan and the Wall of Tammishe', Journal of the British Institute of Persian Studies, vol. 46, pp. 151-178.
Link: Link to publication record in Edinburgh Research Explorer
Document Version: Publisher's PDF, also known as Version of record
Published In: Journal of the British Institute of Persian Studies
Publisher Rights Statement: © Sauer, E., Usher-Wilson, L. S., Ratcliffe, J., Shabani, B., Omrani Rekavadi, H., Wilkinson, T., Abbasi, G. A., S, P., Safari Tamak, E., Ainslie, R., Mahmoudi, M., Galiatsatos, N., Roustai, K., Jansen Van Rensburg, J., Ershadi, M., MacDonald, E., Fattahi, M., Oatley, C., & Shabani, B. (2008). Sasanian walls, hinterland fortresses and abandoned ancient irrigated landscapes: the 2007 season on the Great Wall of Gorgan and the Wall of Tammishe. Iran : Journal of the British Institute of Persian Studies., 46, 151-178
General rights Copyright for the publications made accessible via the Edinburgh Research Explorer is retained by the author(s) and / or other copyright owners and it is a condition of accessing these publications that users recognise and abide by the legal requirements associated with these rights.
Take down policy The University of Edinburgh has made every reasonable effort to ensure that Edinburgh Research Explorer content complies with UK legislation. If you believe that the public display of this file breaches copyright please contact [email protected] providing details, and we will remove access to the work immediately and investigate your claim.
Download date: 04. Oct. 2021 British Institute of Persian Studies
SASANIAN WALLS, HINTERLAND FORTRESSES AND ABANDONED ANCIENT IRRIGATED LANDSCAPES: THE 2007 SEASON ON THE GREAT WALL OF GORGAN AND THE WALL OF TAMMISHE Author(s): Hamid Omrani Rekavandi, Eberhard W. Sauer, Tony Wilkinson, Ghorban Ali Abbasi, Seth Priestman, Esmail Safari Tamak, Roger Ainslie, Majid Mahmoudi, Nikolaos Galiatsatos, Kourosh Roustai, Julian Jansen Van Rensburg, Mohammad Ershadi, Eve MacDonald, Morteza Fattahi, Chris Oatley, Bardia Shabani, James Ratcliffe and Lucian Steven Usher-Wilson Source: Iran, Vol. 46 (2008), pp. 151-178 Published by: British Institute of Persian Studies Stable URL: http://www.jstor.org/stable/25651440 . Accessed: 16/12/2013 07:24
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp
. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
.
British Institute of Persian Studies is collaborating with JSTOR to digitize, preserve and extend access to Iran.
http://www.jstor.org
This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions SASANIANWALLS, HINTERLAND FORTRESSES AND ABANDONED ANCIENT IRRIGATED LANDSCAPES: THE 2007 SEASON ON THE GREATWALL OF GORGAN AND THE WALL OF TAMMISHE
By Hamid Omrani Rekavandi (HO), EberhardW. Sauer (EWS), TonyWilkinson (TW), GhorbanAli Abbasi (GA), Seth Priestman (SP), Esmail SafariTamak (EST), Roger Ainslie (RA),Majid Mahmoudi (MM), Nikolaos Galiatsatos (NG), Kourosh Roustai (KR), JulianJansen Van Rensburg (JJ),Mohammad Ershadi (ME), Eve MacDonald (EM), Morteza Fattahi (MF), Chris Oatley (CO), Bardia Shabani (BS), JamesRatcliffe (JR) and Lucian StevenUsher-Wilson (SU) Iranian Cultural Heritage and Tourism Organisation (HO, GA, EST, MM, KR, ME and BS), University ofEdinburgh (EWS and EM), University ofDurham (TW and NG), BritishMuseum (SP), Abingdon Archaeological Geophysics (RA and CO), Cambridge Archaeological Unit (JJ),Universities ofOxford and Tehran (MF) et al
Abstract The 2007 season yielded significantnew insights into settlementexpansion into the land north of the line of theGorgan Wall and the laterabandonment of these sites in the steppe, prior to the constructionof theWall. It also provided us with a betterunderstanding of Sasanian hydraulic engineering and the date and strategic role of large square fortificationssouth of this linearbarrier. Via underwater archaeology,we explored instal lations associated with theTammishe Wall and now submerged in theCaspian Sea. A detailed study of the pottery from a variety of sites associated with theWalls, as well as of settlements in the hinterland, is beginning to provide us with a clearer picture of pottery typology and the sequence of building projects and settlementpatterns in theGorgan Plain.
Keywords Forts; landscape archaeology; linearbarriers; pottery; Sasanians; underwater archaeology; water supply.
I. INTRODUCTION theWall aimed at shedding new light on their relative chronology. It was also with the question of dating in This results of on paper analyses ongoing research the mind thatwe explored a large three-aisled hall in a fort linear barriers of northern Iran.While our great previous near the Tammishe Wall. Diving expeditions in the fieldwork in 20051 and 20062 focused on the date of con Caspian Sea promised to provide furtherinsights into the structionof the linear barriers and the interioroccupation submerged remains of installations associated with this of the associated in 2007 the focus of our activities forts, shorterWall, its extent and sea-level changes in antiquity. shifted to exploring the relationship between thesewalls and sites in the hinterland and to the north of theGorgan Wall. We examined the extent, economic significance II. SASANIANAND EARLIER LANDSCAPES IN and chronology of settlements and other landscape THE HINTERLAND OF THE GORGANWALL features north of theGorgan Wall, as well as the date and (TW,HO, SP,NG AND KR) possible strategic role of one of the hinterland fortresses south of it.A comparative study of the pottery from one of theWall forts and the hinterland sites on either side of ILL Introduction
1 Nokandeh etal 2006. season The 2007 field of the Gorgan Wall landscape 2 Omrani^al 2007. survey benefited from the availability of a wide range of
This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions 152 HAMID OMRANI REKAVAND ET AL
1 18 17 15 Fig-2 70 1Q 14^3J11 Gunbad 21_a~\ 16^X_y\^ i-Qabus \ *2/^7??^>^^^
26 \ Qaleh" ?J -\ GorganWall ] VGumishan Kharabeh A7 _^ _ . kl \ j__ |_ CASPIAN\ ~y/_I ^* j _ D l^^s v^-^^-^ -\ Gumishan
N _/ Area of underwater / in2007 __yy\ /Survey / ^-^ / G?rgan 0 20km ' ^ --^^ ^^^^^^^^^^
F/g. 7.Map of Gorgan Wall showing the location of the2007 investigations.
remote sensing data, the most useful being the CORONA and GAMBIT3 and orthorectifiedLandsat 74 CORONA images taken in 1968 and 1969. These from theGlobal Land Cover Facility (GLCF) of the we images provide photographs of the earth's surface at a University of Maryland. Furthermore, downloaded resolution which is almost as good as air photographs, elevation data from the Shuttle Radar Topography and importantlyprovide a view of the landscape before Mission (SRTM) version 2.5 much of ithad been destroyed or disturbed by modern The SRTM 3-arcsec Digital Elevation Model (DEM) a source agriculture, urbanisation and industrialisation. is particularly valuable of height data for parts are The availability of satellite imagery and the incorpo of the world where topographic maps not readily ration of ceramic collection strategies into the survey available. The most up-to-date information suggests that on increased the data sources significantly so that it is now itsvertical accuracy is approximately 5m., depending possible to sketch trends of settlement associated with the relief of the ground.6 However, various applications theWall as well as in the dry steppe to the north. In from Hungary,7 Portugal8 and Turkey9 have all demon m. a addition, satellite images have made it possible to strated an accuracy of less than 5 in variety of unravel phases of development of theWall, as well as the terrain.The SRTM 1-arcsec is not yet available outside layout of the water distribution systems that were associated with it (Fig. 1). 3 Galiatsatos 2004. 4 Tuckers al 2004. 5 ftp://e0srp01u.ecs.nasa.gov/ II. 2. The use of satellite imagery 6 Farr 2004. 7 Kay et al 2005. 8 The satellite imagery used for this project included the Gon^alves and Fernandes 2005. 9 declassified imagery from the reconnaissance programs Jacobsen2005.
This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions THE GREAT WALL OF GORGAN AND THE WALL OF TAMMISHE 153
^ O GWS4 ^==^^^^^^^^^^ NorthCanalJ*f " ? ^_^?*-? <9GWS16 C ??T^^-^^^^^^v ^/^^^ River 0 Re.ictGorgan J -^^^^ SouthCanal Qaleh Kharabeh ? A O o ^ Q a J\ ^^^^C;
Legend VQ ^ ^ ? -Modern_river QJ*^Palaeoriver selection
= >-* Canal indeterminate 55=5 o Modern Gorgan River n SouthCanal ^s^mf/^r^ \ S^^lfy^^ \ J\ |Sites ofinterest \ r'-0 1.252.5 5 7.5 10 r?n Wv -- ^y^--^ =^^ ^^ Kilometers I -IForts
F/g. 2. 77zewestern part of theGorgan Wall in relation to the north and south canals.
theUnited States. A detailed description of the SRTM as former channels as broken lines. In addition to a minor well as an evaluation of theDEM product quality can be channel shiftnear Fort 23, therewas a major movement found inRabus et al.10 that resulted in the abandonment of a branch of the Unlike the 2005 and 2006 field seasons, the 2007 Gorgan River between Burak Tepe (GWS-2), Altin Tepe season benefited from the use of CORONA satellite (GWS-28) and Gumishan to a more southern course. images thathad been taken inMay 1968 (mission 1103, The location of the first two sites within protective KH 4B) and inOctober 1969 (mission1052, KH 4A). meander loops suggests that theywere in use when the These images made fieldwork more efficientby enabling river was actually flowing, and from this it can be us to targetkey features thathad been recognised; these inferred that this course was in use during the Sasanian featureswere then recorded in the field and sampled or and Ilkhanid periods. By the Safavid period, when the collected for artefacts. Visits were made to sites and bridge of Agh Ghale was built, the bulk of flow had landscape features to the north and south of theWall in shifted to a course considerably further south. order to describe them and fix theirposition using a hand Nevertheless, the observation by Yate11 that some flow held GPS. The ground control and image analysis enabled still continued through toGumishan when he visited the preliminary maps to be made of canals, archaeological town at the end of the nineteenth century, suggests that sites and theWall itself,as well as ancillary features. the shift in channels was slow and progressive, rather than abrupt and total.
1.3. The changing courses of theGorgan River II. 4. Sites north of theWall The Gorgan River has shifted its course a number of times during recent millennia, and its history is Limited archaeological surveywas undertaken in 2007, to important for an understanding of the Gorgan Wall. record key sites recognised on theCORONA images, as Relict courses of the river place itmuch closer to the well as to provide preliminary dating of settlement to the Wall than is the case today, and Fig. 2 indicates the north and south of theWall. A total of 28 sites were present river course as a dark sinuous line and the recorded and allocated GWS (Gorgan Wall Survey)
10 Rabusefa/.2003. 11 Yate 1900: 274.
This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions 154 HAMID OMRANI REKAVAND ET AL
Fig. ^^^^^^^^H^^^^^|H|^^^^^^^^^^^^^^^^^H Abadan ^^^^^^^^^^^^^^^H^^^^^l^^^^^^^^^^^^^^l^^^^^^^^^^^fl GWS CORONA ^^^^^^^^^^^H^^^fl^^^^^^^^^^^^^^^^l the HHHHHI^^^^^^^^^^HIIHH^^^^^^HHBilHI^^^^^^^^^^^I0 0.10.4 0.2 0.6 0.8 , . , 0 /mage6ycourtesy = ^? ?^?Kilometers Geological Survey.
numbers. All descriptions and dates are preliminary, and before, perhaps as early as themid-first millennium B.C., must await a more intensive and thorough survey. settlement extended from a core area of the plain to the south into the semi-arid steppe lands to the north of the Surveyed sites fell into thefollowing morphological classes: Wall. In the case of sites located to the north of theWall, therewas a clear geographical distinction as follows: 1) Qaleh (fort), sites dominated by conspicuous, upstanding mounds, usually with a concave-up Those to the east appear to have developed within a dished interior. rain-fed agricultural regime. These siteswere 2) Upper and lower qaleh: sites containing both a frequently associated with "hollow ways" which standard qaleh mound as well as a second outer qaleh appear to represent routes radiatingfrom the area, equipped with high and well-developed settlements,perhaps to theirfields and pastures ramparts. beyond.12 3) Sites with a qaleh, and an outer town, the latterbeing On the other hand, sites in thewest were associated contained within ramparts (Fig. 3). with irrigationcanals as well as a latticeof what 4) Similar sites, but without ramparts. appear to have been fields and gardens (Fig. 4). Such 5) Geometric sites such as square fortifications,or field-like features are unusual, and appear to have fortifiedtowns as well as fortsalong theWall. resulted from thepatterning of residual saltwithin 6) Simple, prominent tepes. saline soils. 7) Low, rounded mounds. In addition, many sites, in both the east and thewest, Of these, categories 1-5 appear to have been mainly were surrounded by extensive flat areas or slight Parthian and/or Sasanian indate, and a minority included depressions which would have been ideal places for the some Islamic occupation of probably middle Islamic assembly of livestock. Overall, it therefore appears that date. Categories 6 and 7 appear mainly to have been of the sites to the north of theWall were largely those of earlier date, being IronAge or Bronze Age. agro-pastoral communities, with increased irrigation to Although our ceramic typology is still in theprocess of thewest and rain-fed farming to the east. compilation by Seth Priestman and our Iranian colleagues, 12 it is already evident that during the Parthian period or Omrani et al 2007: 104, fig. 12.
This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions THE GREAT WALL OF GORGAN AND THE WALL OF TAMMISHE 155
0 0.150.3 0.6 0.9 1.2 KiK3l^^K=Z3^MM Kilometers
Fig. 4.Landscape palimpsest in thewestern part of theGorgan Wall to the north ofQaleh Kharabeh. Note that theWall cuts through thepath of the south canal at Fort 29. CORONA satellite image by courtesy of theUS Geological Survey.
II. 5. collection ceramics Surface of main focus during the initial season of the pottery study was towards establishing a secure and reliable basis upon An additional taskwas to date sites theWall key along which to characterise the assemblage for themain period ceramics collected from the surface. Of using specific of theGorgan Wall's use. In futureseasons it ishoped that interestwas the collection recognition, description and the chronological range can be extended beyond a narrow from sites north of the in order to an Wall, gain impression late-Sasanian horizon to encompass groups of pottery of what in activity had occurred the steppe north of the dated by absolute determinations to the late Wall and how it relate to theWall itself.At this might Sasanian/early Islamic transition and the early Sasanian we were not to a stage, aiming undertake comprehensive and Parthian periods. Without this information, and but to a of the sites and then survey, get rapid appraisal indeed some work on ceramic groups extending beyond to the earlier assessments of dates, supplement Kiani.13 this range, it is clearly premature to provide any detailed Site visits thereforesimply entailedmaking field sketches assessment of the pottery recovered from outside the of the GPS on sites, taking readings key points, and taking tightlyfocused range thathas been considered so far (see of ceramics from areas or "grab samples" specific section VII below). What it is possible to provide at this localities within each site. stage, is a reasonably accurate indication of whether the The state of the knowledge regarding classification assemblages recovered from the landscape survey are the and dating of pottery for late historic periods innorth-east same or differentto those recovered from the excavations Iran is at well understood. As a the not, present, result, at Fort 4 or what appears to be its immediate chronologi cal precursor at Qaleh Kharabeh (see section III below). 13 Kiani 1982. In addition, it is possible to identifywith some confidence
This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions 156 HAMID OMRANI REKAVAND ET AL
those assemblages that contain elements that are either appears that the sites and the twomajor canals were pre earlier or later than theWall's use, drawing on a more dominantly Parthian or earlier (late Iron Age to generalised scheme relating to thedevelopment of pottery Seleucid?) in date (see section VII below). across Iran. 7/.7. The Sasanian landscape
II. 6. Canals in relation to theWall Fieldwork in earlier seasons had demonstrated that the Gorgan Wall ditch had received itswater supply from a Several major sites to thenorth of theWall were supplied series of so-called "cross-canals" thatderived theirwater by water from two major canal systems located in the from the Gorgan River to the south.15 One additional western part of theProject area. One of these (the northern cross canal (termed locally the "Band-i Vali") was canal), had already been recognised in the 2005 season, at recognised on the CORONA satellite images and which time itwas unclear whether the canal had derived confirmed in the field by its large and distinctive soil up itswater from theGorgan Wall ditch or directfrom a relict cast banks. Located to the south of Fort 17 and north of course of theGorgan River.14 These northern and southern the village ofYesergichen, the canal had been dug along canals (Fig. 2) were recognised as such on both the the course of a long-abandoned channel of the Gorgan satellite imagery and in thefield by thedistinctive up-cast River to lead water from theGorgan River directly into banks alongside. Both canal systemswere evidently cut by the ditch of theGorgan Wall (Fig. 6). Unfortunately, the theGorgan Wall and its associated ditch. relationship between the Gorgan Wall ditch and the The southern canal had derived itswater from the Band-i Vali is no longer evident because of the presence Gorgan River to the south of Fort 23. It flowed from east ofmodern irrigation lakes near the junction of the canal towest, and near itswestern end it had supplied water and theWall ditch. probablyto the major siteof Tokh Mogh (GWS-4)and its In previous seasons16 itwas inferred that the cross surrounding fields. This canal was cut by theWall and canals had received theirwater from theGorgan River via ditch in the vicinity of Fort 29, which appears to have a series of major earthen dams, best exemplified by the been built over the top of the canal (Fig. 4). Pottery Sadd-i Garkaz, near Fort 6. However, the recognition on a collected from Tokh Mogh suggests that the site and CORONA images of major canal system to the south associated south canal had been abandoned for several east of Sadd-i Garkaz, and running towards it,prompted a was evidence in the generations, or even centuries, when the Wall re-evaluation of the evidence. The field constructed. form ofmassive soil banks and a central channel (Fig. 7: The northern canal collected water from a relict A), located several kilometres to the south-east of the was a channel of theGorgan River, east of Fort 23 to thewest of Sadd-i Garkaz, demonstrated that it indeed canal, a which theWall and ditch cut through the canal (Fig. 5). which had collected water from vigorous left(i.e. south) Almost the entire lengthof the canal lies to thenorth of the bank tributaryof theGorgan River, theRudkhane Dough. to cross aminor it Gorgan Wall and itcould be tracedboth on theground and Where the canal needed tributaryvalley so via an using CORONA images westwards to the large site of may have done embankment, although any signs a dam Mangali (GWS-15), the surrounding fields of which of thishave been obscured by (theMonajim Dam; built a local landlord some 50 probably received theirwater from the canal (Fig. 2). The Fig. 7: B), by years ago. trace of the north canal becomes indeterminatewhere the Downstream and to thewest of themodern dam the canal a distinctive which became canal negotiated a broad shallow valley. From its relation followed trace, progressively the north and as the ship to the IronAge to early historic (Parthian?) siteGWS fainter towards west, presumably a shallower in 7: Nevertheless 5, which apparently had received itswater via distribu canal became depth (Fig. C). followed itsweak trace on theCORONA tary canal from the northern canal, the canal was in use it could be by earthen dam of Sadd-i well before theGorgan Wall was built through the area. satellite images towards the large the of the "dam" Because the pottery collected from both Tokh Mogh Garkaz (Fig. 7: D). Although narrow, top m. in the 3^4 m. in the revealed andMangali pre-dates the ceramics excavated at the forts (5-6 wide south, north), a fine small and numerous along theWall and Qaleh Kharabeh, and included evidence of gravel, potsherds, it significant number of pre-Sasanian diagnostic types, 15 Nokandeh et al 2006: 138-41; Omrani et al 2007: 95-98. 16 i4 Nokandeh et al 2006: 146. Nokandeh et al 2006: 140-41; Omrani et al 2007: 98.
This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions THE GREAT WALL OF GORGAN AND THE WALL OF TAMMISHE 157
PSm! showing theGorgan 111^^ d/fc/j w/zere// cwto course Bgli|h throughthe pyf^RSflBp^ ^? canal CORONAsatellite hBBPII*^^ ||3B8m11^^ /mageZ?y courtesy of BMSresilPllliTO //ze Lft 0 0.125 0.25 0.5 0.75 1 Geological 0 i im ? ?? ??ii owrvey.
0 0.15 0.3 0.6 0.91.2
F/g. 5. 77zerowte 0/ *7*eBand-i Vali canal where itjoins theGorgan Wall ditch in thevicinity of Fort 17. CORONA satellite image by courtesyof theUS Geological Survey.
This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions 158 HAMID OMRANI REKAVAND ET AL
0 0.3 0.6 1.2 2.41.8 BEZZMHZZ3?B?BBI^?==Z==^=MBBBBB?BiKilometers
Fig. 7. CORONA image of theSadd-i Garkaz andfeeder canal from theSE (north to the top).A: canal where it ismost visible on theground; B: theMonajim Dam and associated lake; C: thecourse of canal where it is shallow and theassociated banks are less visible;D: Sadd-i Garkaz; E: theChai Ghushan Kuchek canal leading towards theGorgan Wall. CORONA image by courtesyof US Geological Survey.
small freshwater gastropod shells. These demonstrate that a dam to storewater from theGorgan River and raise it to the Sadd carriedwater via a canal which ran along itscrest the level of the loess plateau, it functioned as an earthen from south to north and thence onwards towards the aqueduct to convey water from theRudkhane Dough to Gorgan Wall ditch via theChai Ghushan Kuchek canal.17 the south, across theGorgan River, to supply water for North of Sadd-i Garkaz, one branch of the Chai brick-making and defence, for theGorgan Wall ditch, as Ghushan Kuchek (Fig. 7: E) was traced in the field to well as perhaps for the supply of Fort 7 and local irrigation. lead into theGorgan Wall ditch. In addition, CORONA West of Fort 5 the Gorgan Wall ditch must have imagery demonstrated that a second branch of the canal conveyed water for some ormost of its length,and must followed an alternative course, turning through 90 thereforehave functioned as a canal. It therefore suffered degrees to take water parallel to theWall, and a short from some of the disadvantages of canal design, namely distance to its south, toward Fort 7. Survey along this that it required a constant "design gradient" to function branch canal parallel to theWall demonstrates that a efficiently. If valleys crossed itspath itwas necessary to number of brick kilns occurred along theWall towards negotiate these by the use of engineering works to Fort 7. In addition, a gap through theWall at its junction maintain an even gradient. For example, between Forts with the Chai Ghushan Kuchek suggests that it also 14 and 15 a large "embankment" of silt18 appears to supplied water to theGorgan Wall ditch. represent the soil cast up during either the construction of The 2007 season thereforewitnessed a significant rein theGorgan Wall ditch, or of a subsidiary canal. In other terpretationof the Sadd-i Garkaz. Rather than operating as words, where the ditch or itswater supply canals needed
17 18 Nokandeh et al 2006: 138-41. Omrani etal 2007: 96 fig. 2.
This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions THE GREAT WALL OF GORGAN AND THE WALL OF TAMMISHE 159
F/g. 5. Satellite image ofQaleh Kharabeh south of theGorgan Wall CORONA satellite image by courtesy of theUS Geological Survey.
HHHHHHHHHHHHHHHHH
This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions 160 HAMID OMRANI REKAVAND ET AL
i A N37?06.134' e 54?25.684' -1-1-1?.?!-f-1-1-h? ' 1' _'JmB _ \_B N37?06.2ir ~\ 1 1 1 JBBkl 1 I | E 54?25.796' c N37?05.966' -1-.-.-1-hs-^HBHkr^-1-1-1- 4 I E 54?25.629' J^HHBkD N37?06.063' -f-H-1-^iN^al^-!-!- <* _ E54?25.769' < Fig. 10. Qaleh Kharabeh: Results of themagnetometer survey (byRA, CO, ME, MM and SU, Abingdon Archaeological Geophysics and theICHTO) and location of the trenches(L, M and N) excavated in 2007. to negotiate a side valley, first the canal or ditch was dug together, these features suggest that during the Sasanian deeper so thatwhen thevalley was encountered, the level period, water was gathered from a brick off-take in the of the water was of sufficiently low elevation for the bed of the Sari Seyyid (WP 438), and conducted water to have been conducted across the valley without downstream via an earthen channel, reinforced at the need for an embankment. intervals by masonry structures specifically where the The brick-built installation exposed in the bed of the canal channel was vulnerable to riverine erosion. The Rudkhane Sari Seyyid (Rudkhane Qarnaveh; WP 438),19 location of this canal, just within the alignment of the was revisited in 2007. A section of a former canal or river Gorgan Wall, places itwithin the protective confines of channel, partly infilledwith fragments of Gorgan Wall theWall, rather than being located outside itsboundaries, brick, was exposed in the bank of the Sari Seyyid River where itcould have been vulnerable to sabotage. to the south-west of thebrick structure.This channel lines up with thenorth-south element of thebrick structureWP 438 and appears tobe the course of a former canal that led II. 8. The Eastern Extension of theWall water downstream towards a line of brickmasonry which formed either part of a canal, or a reinforced brick The landscape of theWall changes significantly towards structurebuilt to protect the canal from erosion. Viewed its east end. First it follows the crest of the Pishkamar Rocks, to then drop down from a fortleton thehigh scarp !9 Omrani et al 2007: 105-6, figs 13-14. east of Pishkamar, down to a valley-floor compound This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions THE GREAT WALL OF GORGAN AND THE WALL OF TAMMISHE 161 P^yaBMh GreatWall of N37?i5.686jVj@|H Gora^n Kilns(?)^j^^^HH 5.645'^^BUHBH^^^^HHr N! ^E54?58.546' 54?58.493J^^^JHH Edgeoffort|^M^K ^* platform ^^J^J^ Fig. 11.Magnetometer survey ofFort 16: unclear traces of two barrack blocks and clear traces offour rectangular anomalies under thefort platform, almost certainly brick kilns (by RA,ME andMM, Abingdon Archaeological Geophysics and theICHTO). (excavated by Nokandeh and Omrani). At this point the when the region fell close to the core of the Parthian Wall makes a dramatic change from a position north of the Empire, settlements extended some way into the arid Gorgan River to one thatoccupies the south (or left)bank northern steppe. These settlementswere then abandoned, of the riverwhere it is overlooked by high loess cliffs to so thatwhen theGorgan Wall was built, and the area lay thenorth. This would have placed theWall in a position of closer to thenorthern fringes of the Sasanian Empire, the military vulnerability.This impression is reinforcedby the Wall cut throughwhat a few centuries earlier had been a small size of theGorgan River in this area (2-4 m. wide), well occupied landscape. It is the intentionof futurefield and the lack of a ditch. The point where theWall switched seasons to test thismodel further. from thenorth to the south bank of theGorgan River (and where itdrops from the command of the limestone ridge) is thereforevery significant. III.HINTERLAND FORTRESSES (EWS,EST, RA,MM, ME, CO, SU andBS) II. 9. Discussion In contrast to the Gorgan Wall forts, littlewas known about the date and function of large square fortifications During the Parthian period, or perhaps somewhat earlier, south of the Gorgan Wall.20 We thus selected Qaleh the lands north of theWall were well populated by agro Kharabeh, a c. 650 x 650 m.-large square fortification pastoral communities living in numerous settlements up 20 to some 25 ha. in area. In otherwords, perhaps at the time Omrani a/. 2007: 130-31. This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions 162 HAMID OMRANI REKAVAND ET AL south of Fort 29 (Fig. 8), because of its good preserva compound are known to us. It seems possible that the tion, for survey and excavation in 2007. As with most of enclosure ditches served as drainage or boundary ditches the other hinterland fortifications, itwas provided with a around tents of a temporarymilitary garrison. The use of raised corner citadel (Figs 8-10). Trench L (Figs 9-10) tents in late antique Persian military camps is attested by revealed that this citadel was heavily defended with a Ammianus Marcellinus21 and Procopius.22 Why these 4.40 m.-wide mud-brick wall. A second wall parallel to it enclosures are confined to the eastern geophysics transect and measuring 4.90 m. inside (inner to inner edge) and a is not clear, especially as there are no obvious differences third perpendicular to the two, partially unearthed in in elevation and potential drainage problems between the trenchL, may conceivably form part of a corner tower. two transects. Possible explanations include that the Seven graves, four of them covered with reused Gorgan seemingly empty sections of the interior space were Wall bricks, were left in situ, thus preventing us from occupied by tents or yurtswithout drainage ditches?or excavating down to ancient levels in significant parts of that theywere indeed leftempty?for example as a space the trench. for horses or livestock. A magnetometer survey (Fig. 11), covering about one The strikingdifference between the absence of finds third of the interiorof Qaleh Kharabeh, revealed in the from the (tent?) enclosure in trenchN and the richness of eastern transect of the compound a large number of findsfrom trenchM calls for an explanation. The central rectangular anomalies, arranged inneat parallel west-east position near the crossroads of the trenchM house might running rows?with c. 17 m.-wide access corridors suggest that itwas a high-status building, but the small more between pairs of two rows each. A section throughone of size would argue against this. Perhaps it is likely or the rectangular anomalies (trenchN) brought traces of a that it represents one in a row of shops, workshops possible ditched enclosure to light,but no associated finds. magazines, which tend to linemain urban trafficarteries Permanent buildings seem to have lined the main in old bazaars and colonnaded streets in the ancient Near towns in roads. Notably the road leading into the fortification East,23 and roads in ancientMediterranean and from the north (Fig. 10: "ancient N-S road") is lined by early and high imperial legionary fortresses24 The wide a row of small high magnetic anomalies. We excavated assemblage of finds, notably the evidence for extensive or one (trench M), which proved to be a mud-brick meat consumption, suggests that,whether not the as one structurewith a hearth. It yielded a rich assemblages of proposed interpretation central supply facility of or room or units finds, notably pottery, animal bones (including fish from sort another is correct, these small house rivers, canals or the Caspian Sea, now c. 45 km. further also served as dwellings, probably for those operating or centres. west), charcoal, some glass shards and a significant these postulated storage, repair redistribution M not quantity of iron fragments (unfortunately too corroded The finds from trench shed significant light for identification, even after x-raying). just on themanner of occupation of the compound, but on The that Before venturing any hypotheses on the purpose and also its chronology. pottery suggests Qaleh as historical context of Qaleh Kharabeh as a whole, it is Kharabeh dates to around the same time the earliest necessary to evaluate the possible functions of its occupation at Fort 4 (see section VII below). Qaleh the individual architectural components. Without knowing Kharabeh, which may well be representative also for the what, if any, structures occupied the citadel's interior other square hinterland compounds south of Gorgan to have been constructed con space in antiquity, any functional explanation has to Wall, thus appears roughly as a some or all of the Wall remain hypothetical. Perhaps it served defensible temporaneously with Gorgan This and other architectural the retreatfor high status occupants or a garrison during times forts. parallels (notably canals section IV when the vast main compound was unoccupied. The association of roads with [see below] observation that it is surrounded by a ditch, cutting and the projecting mud-brick towers, regularly spaced Kharabeh's and the Wall forts' through the ramparts of the main compound, adds along Qaleh Gorgan that the Wall and the strength to the latter theory, though does not prove that defences) suggests strongly Gorgan as a hinterland fortificationswere of the same thiswas its original purpose. A function watchtower square part not and/or for signal transmission is also conceivable, and theories. 21 necessarily mutually exclusive with the above AmmianusMarcellinus 19, 7,11. more are 22 Even puzzling theneatly aligned rectangular Procopius, Wars 1,5, 21-25. 23 anomalies in the east of the compound. No parallels for Ball 2000: 261-76. 24 Petrikovits1975: 96-97. such an internal layout of a Sasanian town or defensive 51-54, 58, This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions THE GREAT WALL OF GORGAN AND THE WALL OF TAMMISHE 163 architectural scheme. Three possible interpretations are If Qaleh Kharabeh was occupied by the army, then worth considering: the question as towhether we are dealing with field army (hypothesis 2 above) or frontier troops (hypothesis 3) Qaleh Kharabeh was an urban foundation created at deserves furtherinvestigation. The magnetometer survey about the same time as theWall?perhaps part of a of Fort 16 (Fig. 11) suggests that this particular fort programme to deliberately bring new settlers into overlies four brick kilns and thuswas most probably not these fertile, and now more secure, borderlands, part of the original design. Whether this suggests that all Itwas a heavily defended compound with fortswere added to theWall only after its completion and temporary accommodation for the Persian field troops originally stationed in hinterland sites, likeQaleh army,whilst fightingmajor wars in the north or Kharabeh (hypothesis 3), or whether many forts It served as a defensible retreat for the frontierarmy belonged to the original design and only some added during the construction of theGorgan Wall and later, is open to debate. Fort 16 is so far the only such before troopswere moved to the forts on theWall. military compound on theWall known to overlie kilns. This observation, the fact that the pottery evidence for Of these options, the first seems the least likely.Not Qaleh Kharabeh and the earliest phase of Fort 4 are only are no parallels known for an urban foundation, roughly contemporary and the large size of thehinterland with a significant proportion of the interior space filled fortresses, ideal for strong troop contingents, render with small and presumably temporary enclosures, but, explanation 2 themost likely. It is perfectly possible, as most of the settlements north of theWall appear to even if far from proven, that compounds, like Qaleh have been abandoned long before its construction (see Kharabeh, could have played a key role in one or more section II. 6 above), there is no obvious local population war(s), such as those fought by king Peroz against the which needed resettling. This does not disprove the Hephthalites. hypothesis that theQaleh was a failed urban foundation, Qaleh Kharabeh is, of course, only one of several as the intended urban population could have come from square hinterland fortresses.27While furtherfieldwork is south of theWall or from furtherafield. Indeed, there needed to test towhat extent our conclusions about the appear to have been urban foundations, even ifnot yet foundation date and interior occupation of Qaleh precisely dated, notably the c. 338 ha. large Dasht Kharabeh hold true for its counterparts furthereast, the Qaleh25 with its recently detected vast structures, striking similarities in overall layout suggest that they possibly consisting of parallel rows of brick pillars should probably all be attributed to the same building enclosing possible courtyards.26 In contrast to the programme. Whether Qaleh Kharabeh's early rectangular plans of the Gorgan Wall forts and the abandonment is mirrored in its "siblings" is as yet square shape of the hinterland fortresses,Dasht Qaleh is unknown. Even if the hinterland fortresses and theWall a surrounded by polygonal town wall. Rectangular fortswere not occupied simultaneously for long, ifat all, compounds are ideal formilitary purposes, as they allow the sheer scale and number of the certain and probable accommodating the maximum number of troops in military compounds in the Gorgan Plain demonstrate or rectangular tents permanent dwellings, whilst powerfully what resources the Persian Empire could enabling them to quickly reach the defences via a simple muster in the fifth/sixthcentury A.D. to defend as littleas network of straight roads. In addition to Qaleh a c. 200 km. stretch of vulnerable frontier. It suggests Kharabeh's ideal square plan, there are other indications strongly that the late Sasanian armymatched itsRoman to a an suggest military function: its abandonment after counterpart in numerical size and technical capabilities. apparently shortperiod of occupation and the scarcity of The recent hypothesis, that the Sasanian army was inmost finds of the interiorof the compound, are much probably substantially smaller than that of the Eastern easier to reconcile with a military installation? Roman Empire,28 now seems doubtful in the extreme especially bearing in mind that army members on (especially as the result of our excavations in Fort 4 had reasons to non a campaign good avoid carrying suggest high occupation density over a long period of essential and bulky fragile goods. time and not just during the construction phase). 25 27 Kiani 1982: 48-52, figs 30-31, 33-35, pis 17,2-22,1. Omrani et al 2007: 130-31. 26 Amin Pour andOmrani 2007. 28 Borm 2007: 161with no. 8. This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions 164 HAMID OMRANI REKAVAND ET AL A: I I 7? II I I I I I N|ST2T2WlE 56*26.196* Ik B: N3T27.122- // ftfrnn*MA 30m E66*26,16Br ft ? ^ ^ <5s-HZTTTJO&Ct-/' Modem\ Modem Modem / g "\ bt/rfa/ j ditchedS. enclosures brick-built Mck-bullt & k. fo/ffe/ bu/te/ / enclosure /"v^^ ^^fl^V Modem^^ ~ ^^^^ ^^^^^ ^^^^^^^^^ Modem ^ monument a ^V,^ ^S^g^ ^^O'/SC/^V^ ^^^^ ^^V/ // / F/g. 72. P/#? ofFort 4 with the location of trenchJ. IV.THE INTERIOROCCUPATION OF clarifying the function of a linear anomaly, first thought A GORGANWALL FORT to be part of an enclosure. Excavation, however, revealed (GA,EWS, EM, EST, SP,BS andMF) that this anomaly was caused by a subsidiary road with paving consisting of stone cobbles and brick fragments The pottery analysis (see section VII below) suggests and a c. 0.30 m. wide and c. 1.80 m. deep stone-lined that the earliest features in Fort 4 and inQaleh Kharabeh gully on itswest side (Fig. 13). The gully was within an are of broadly similar date. The 2007 excavation in earlier filled-up ditch, originally c. 2.10 m. wide and flat trench J (Fig. 12) revealed also a possible parallel in their bottomed. The greenish iron staining of thebottom fill of infrastructure (i.e. their roads and associated canals), theditch (J.020)and thatof thegully (J.018)suggests suggesting that some or all of theGorgan Wall forts,and thatboth successive features had carried water. While a some or all of thehinterland fortresses,may belong to the drainage purpose cannot be excluded, the elaborate same grand building programme. The trench aimed at stone-lining of the gully, its (for such a purpose) This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions THE GREAT WALL OF GORGAN AND THE WALL OF TAMMISHE 165 Fig. 13. TrenchJ with the subsidiary road and stone-lined gully, looking in a south-easterlydirection. The road overlies one certain and a second possible pit/well (half-sectioned in the sondage; note the cluster of brickfragments in itsfill, between the sondage and the road). One of the ovens is visible near the rightmargin, a second is (scarcely visible) just behind the road, a third(half-sectioned), at a much higher level, on the baulk on the left,where thereare also parts of a late brick paving. The red and white segments of the scales measure 0.50 m. each. unnecessary depth and impractical narrow profile with occupation, and themuch lower quantity of occupation vertical sides, and the observation that the ditch and later debris building up, has preserved the twomain roads as gully line the road on one side only, render itmore likely distinct causeways to thepresent day (Fig. 8). No section thatwe are dealing with water supply canals rather than across the north-south road has been surveyed, but the a drainage ditch and gully. west-east road in the Qaleh displays a distinct linear Natural soil was encountered at c. 95 m. OD in a anomaly close to its crest (Fig. 10), almost certainly sondage in trench J (matching strikingly the base of a caused by a canal. Such raised canals would have barrack block's mud-brick wall in nearby trenchH at allowed the use of water (e.g. for brick-making) without 95.04 m. OD?probably the approximate level of the employing water-lifting devices. The gully fell out of use ancient surface in the area of trenchesH and J). That the well before the abandonment of Fort 4, and subsequent ancient surface level in the area of trenches H and J lymuch occupation debris accumulated in itsmiddle appears to have been similar is furthercircumstantial (J.021) and upper fill (J.009) (see section VII below). evidence to suggest that the barrack blocks were built at Complex water supply systems are one of the hallmarks ground level, had no cellars and were at least two storeys of Sasanian engineering. The famous ancient university high.29The remarkable elevation of the road paving, 1.80 town of Jundi Shapur in Khuzestan, for example, was m. above thepostulated ancient ground level,may indicate supplied with water from the perennial River Dez via a that itwas on top of a causeway, built specifically to c. 14 km. long canal and an elaborate tunnel system30? channel water into the fort. mirroring the scale and technical sophistication of the The long occupation of Fort 4 led to sediments installations supplying theGorgan Wall fortswith water building up on either side of the causeway and its from theGorgan River. eventual disappearance under rising levels of debris. By contrast, Qaleh Kharabeh's presumably much shorter 30 Adams andHansen 1968: 59-63; The authorsare gratefulto Mr Housein Yousefifar for showing us the impressive 29 Omrani et al 2007: 121-23. remainsof the systemin December 2007. This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions 166 HAMID OMRANI REKAVAND ETAL If we are right in thinking that Fort 4 and Qaleh Caspian Sea.31 The 2007 season provided the Kharabeh were, initially at least, supplied with flowing opportunity to explore the flooded remains of theWall in water, then this need not have been their only source of more detail. Notably a satellite image (Fig. 14), first water. Sealed underneath the road paving was the studied by Tony Wilkinson and Nikolaus Galiatsatos, probable corner of the curb of a pit orwell at c. 96.36 m. shows a well-defined rectangular to rhomboid platform OD, consisting of large fired brick fragments, the of c. 60 x 110m. size on the seabed, itswest (enemy) side unbroken straight sides facing inside. Our sondage in almost perfect alignment with theTammishe Wall, and revealed that this curb overlay a deep pit, which we also an oblong one, extending further into the Sea. We explored down to a depth of c. 92.74 m. OD. Its first carried out a systematic depth survey of transects dimensions (c. 1.25m. north-south and c. 1.50 m. west through the Gulf of Gorgan, on the premise that any east, as far as they can be reconstructed on the basis of a preserved or collapsed section of theWall should form partial excavation) would be perfect for a well. If so, it is detectable ridges above the sea floor.Wliile furtherwork likely to have been much deeper, and thehazardous shaft is needed to firmly link the results of the depth survey would probably have been filled up deliberately soon with the features visible on the image, several promising after itsabandonment. The early potteryfrom the deepest featureswere explored by diving transects. In one area, excavated deposit in its fill (J.027) (see section VII centred on 36 x 48.616' north and 54 x 01.251' east, below) provides us with a terminuspost quern for the bricks from a 12 x 2 m. large area were systematically date of its abandonment and, possibly, a terminus ad collected and weighed, yielding an average of 6.3 kg. of quern for any redeposited early occupation debris accu bricks per square m. (or 8.3 kg. per square m. in four mulating within the fort.Too littlewas excavated of a adjacent 2 x 2 m. squares). The sheer quantity and wide second feature in the sondage to the west of this spatial distribution of Sasanian brick fragments, occa postulated well to offer a certain interpretation.Yet, its sionally associated with Sasanian pottery, from the depth (reaching down to the bottom of the sondage and seabed in theGulf of Gorgan, eradicates any doubt that possibly beyond) and the early potteryfrom its lowest fill there were extensive Sasanian fired brick structures. a (J.030) suggest a possible second well. WTiilst flooded now by theCaspian Sea, due to rise in The trench also contained two ovens on either side of itswater level, theymust have been on dry land (or, the road and at a similar level, as well as a thirdoven next possibly partially in shallow water) when constructed. zones more to a paving of fired bricks (of 0.40 x 0.40 m., the typical That these brick scatters and of shallow size of bricks from this section of the Gorgan Wall) at water cluster along the alignment of theWall is hardly nor 0.65-0.90 m. above the road paving (Fig. 13), suggesting fortuitous.As neither thewide distribution of brick that the area of trench Jmay have played a role in food theplatform can be explained with a single linear feature, assume to a preparation over some time. The complexity of the we have to that, in addition continuation of were multi-phase stratigraphy and the large quantities of the Tammishe Wall, there other Sasanian features. a to pottery and animal bone add strength to our assumption These may well have included fort (the rectangular on that Fort 4 was occupied for a considerable period of rhomboid platform visible the CORONA satellite would have time, perhaps from the construction of theGorgan Wall image?). The seaside terminal of theWall an a (in the fifthcentury?) to its abandonment (possibly in the formed attractive location for potential harbour.Any firsthalf of the seventh century). Further samples for associated brick-built installations might have to a scatter even if scientific dating, taken from trenches J,G, L, M and O contributed wide of brick fragments, and Sadd-i Garkaz in 2007, should hopefully allow us to much more extensive mapping of the ancient debris will to test It of also refine the chronology of theWalls and associated instal be required this hypothesis. is, course, north?and lations, but are not yet available. possible that theWall extended much further while the alignment of the Tammishe Wall might argue against it joining up with the Gorgan Wall,32 further V UNDERWATER SURVEY research is needed to settle the question. (JJ,HO,EWS, BSandJR) 31 From a previous underwater survey we know that the Nokandeh et al 2006: 152;Omrani et al 2007: 112-13. 32 Wall of Tammishe runs into terrainnow flooded by the Omrani et al 2007: 112-13. This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions THE GREAT WALL OF GORGAN AND THE WALL OF TAMMISHE 167 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^H Fig. the ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^M Tammishe, probably ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^H floodedfeatures the ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^M CORONA bycourtesy HiHHHHHHHHIi^lHHHHiHHHHHIHIHHHHHHi the Interesting in this context is that theArab geographer If the identification is,nevertheless, correct (and itwould Dimishqui (A.D. 1256-1327) attributes the foundation fit the descriptions of latermedieval authors),35 then any at to I of the Caspian Sea port Abskun Kavad (A.D. associated harbour has to be sought furtherwest. If 488-531).33 Traditionally, Abskun has been identified Dimishqui36 is right in postulating the foundation of a western with Gomish Tepe, presumably just south of the harbour at the south-east corner of theCaspian Sea under In section of theGorgan Wall.34 the lightof the Caspian Kavad I, despite writing some eight centuries later, then Sea's lowerwater table at the time of theWall's construc theWalls may also have served to protectmaritime trade tion, this prominent settlementmound would have been across theCaspian Sea. Frye37 argued, albeit on thebasis even some now. furtherinland 1,500 years ago than it is of circumstantial evidence, that trade across theCaspian 33 35 Mehren 1874: 314; cf. Frye 1972: 267; Schippmann 1990: Bosworth 1985. 90-91. 36 Mehren 1874:314. 34 37 Bosworth 1985;Omrani et al 2007: 112. Frye 1972; cf.Harper 2000: 48^19. This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions 168 HAMID OMRANI REKAVAND ET AL TftmchO. xfflfkt ^sjhre+aialed hall iflllfflffiii ^ F/g. 75.Magnetometer survey of the northernpart of theBansaran Fort with dense interior ~ occupation, including a prominent three-aisled hall (by CO, RA,MM, ME and JR, 30m 4 Abingdon Archaeological Geophysics _1 1 1 1 1 1^ 1_I andthe ICHTO). Sea and along theVolga and Kama Rivers expanded in VI. CONTINUITY OF OCCUPATION INTHE the late Sasanian period, and that it involved the export BANSARAN FORT of Persian silver vessels into territories far beyond the (GA,EWS, RA,ME, MM, CO, HO and SP) Caspian Sea's northern shore. If so (and this theory has gained furtherstrength by the discovery of predominant Parts of the interiorof the Bansaran Fort, 500-700 m. lyfifth to seventh-century Sasanian coins from graves in west of the Tammishe Wall at the foot of the Elburz the same area),38 theGorgan and Tammishe Walls may Mountains, had already been explored before, most have created the secure and stable conditions needed for recently in 2005, when a c. 33 x 53 m. large three-aisled creating, in a literal sense, a safe haven for long-distance brick hall was discovered.39 The 2007 season offered the trade, even if theirmain functionwas undoubtedly to opportunity to extend themagnetometer survey over all protect other economic and agricultural assets of the rich parts of the fort (Fig. 15), except for the forest-covered and fertileGorgan Plain. south. Whilst the three-aisled hall remains the most prominent feature, it appears that it is part of a larger complex filling much of the interior.Whether some of 38 Goldina andNikitin 1997. 39 Nokandeh et al 2006: 156-58. This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions THE GREAT WALL OF GORGAN AND THE WALL OF TAMMISHE 169 H^^^^^^^^^^^^^^^^^^^^H^^^H Fig. parts ^^^^^^^^^^^^^^^H|^^^^HHH^|HH|BH|^^HH^^^^H| theentrance facade of the ^^^^^^^^^JMBB|BHB|BpB hall inthe Bansaran Fort (trenchO), ^^^^^^^^BMMBWmWIMlooking south, with the forest-covered ^^^^^^^W^BH^H^BmIM partof thefort and theElburz ^^^^^^^H^B^^^^^^^^^^^^^^^I^H^^HH^^^^H^flHHBH Mountainsthe ^^^^^^^^^^^^^^^^^^^^^^^^^^HH|^^h segmentsthe WKK^^^KKK^^BKIKKt^^^^Kt^^^^^tKKKKKBBBSSMWmKKKUmeasure or not is the remains are those Ibn Isfandiyar saw in the thirteenth will establish whether this occupation layer Bansaran Fort on a century, and thought to be the ruins of a legendary Sasanian. The location of the to those of the Wall forts and palace,40 is open to debate. The sparing use of fired brick platform similar Gorgan 196441 in (to judge by themagnetometer survey) suggests that this finds from excavations in indicate, any case, The reasons the Bansaran expensive material was mainly employed for load Sasanian origins. why Fort, on continued bearing elements of the tallest buildings, e.g. the pillars unlike those theGorgan Wall, presumably have of the hall and two possible rectangular features south to be densely occupied in the early Islamic period to in the of the east of it, possible elements of a monumental fagade, be sought exceptional fertility minarets or other towers. surrounding coastal plain and its humid climate, not to We excavated one of the pillars, measuring 3.00 x mention the vicinity of the important early medieval the forts on the 3.30 m., excluding the adjoining parts of the entrance town of Tammishe. By contrast, Gorgan not fa$ade of the three-aisled hall (Fig. 16). The pillar was Wall, in land of far lower agricultural potential, and on were far less stabilised by a deposit of softmid-yellowish brown silt any major trafficroute, attractive places once with c. 60% of fine to coarse pebbles and fragments of to inhabit, theWall had been abandoned. ceramic building material in the foundation trench?a furtherindication that itwas designed to carry substantial weight: probably the high roof of an imposing VII. POTTERY FROM THE GORGANWALL: A monument? The size of the pillar's bricks (28 x 28 x 5 to LATE SASANIAN "MILITARY"ASSEMBLAGE 6 cm.) differedfrom those on theGorgan and Tammishe (SP) Walls, suggesting that the hall was not part of the same as a result of excavations building programme. A later date is also indicated by a The study of pottery recovered with the of the preliminary analysis of the small pottery assemblage. and survey associated investigation an tomake in Scientific dating should hopefully confirm and refine, or Gorgan Wall has important contribution as a correct, our current assumption that the hall is later than relation to the project whole. Of primary importance a to theWalls and that itdates to the early Islamic period (or, is the potential to establish relative chronology the As the possibly, just before). Should itprove to be early Islamic, compliment that of absolute dating program. two then itmay well be an earlymosque. We encountered an relationship between the is developed and strength rela earlier charcoal-rich occupation layer underneath the ened, the ability to recognise significant temporal level of the three-aisled hall's foundations, but did not tionships "on the spot" becomes an increasingly reach the natural soil.We hope that radiocarbon dating powerful tool in investigating and recording. 40 Browne 1905: 16. 41 Bivar and Fehervari 1966. This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions 170 HAMID OMRANI REKAVAND ETAL More immediately,pottery associated with theGorgan significant feature.Most of these sherdswere concentrat Wall provides a significant opportunity to characterise the ed between 96.38-96.45 m. OD, which falls within the assemblage from contexts that are securely dated. The upper part of the occupation layers and at a similar studyof Sasanian ceramics ingeneral has been plagued by elevation to a "red storage vessel" (H.025).44 This is insecure dating, insufficientattention to detail, and the actually a medium sized jar with relatively thinwalls, not prevalence of chronologically contaminated contexts. by regular definition a storage-jar, and the "organic The main aim for the first season of the pottery study residue" noted inside is a patch of bitumen thathas been was to define the ceramic assemblage typically associated used to seal a crack in the base of the vessel thatopened with theuse of theGorgan Wall. A number of fundamental during firing. Part of a recycled vessel and burnt sherds questions needed to be answered at the outset: may be the sort ofmaterial thatwould accumulate within a building as itwas being abandoned. Was there any detectable change in the ceramic assemblage between the earliest and latest deposits excavated? VII2. Trench J Did the ceramic assemblage compliment the scientific dating thatplaces all activity associated This second trenchwas opened in 2006 and completed in with theGorgan Wall within a relatively narrow 150 2007 over one of the subsidiary routeways inFort 4 (see year time period? section IV above). The upper layers of the stone-lined Does the ceramic assemblage provide an indication gully west of the road produced a particularly rich accu a of the type of activity associated with theWall, for mulation of pottery, and the addition of larger body of a example domestic occupation? material sealed within securely dated sequence proved to be of great value indefining the assemblage relating to During the first two seasons of fieldwork, the main the occupation of Fort 4. emphasis was on the absolute dating of theWall, and most excavations were targeted at low finds-yielding contexts, such as brick kilns and theWall itself. This VII. 3. Processing pattern changed at the end of the second and continuing was into the third seasons, as greater emphasis was given to The ceramic assemblage from the two trenches order to the use and occupation of theWall. studied in detail during the 2007 season, in lay a of the So far, the major body of material that has been down precise framework for the classification studied in detail comes from two trencheswithin one of pottery associated with the period of theGorgan Wall's a theWall forts,Fort 4. use. In the analysis of the material there is clear disparity between the two areas, with trenchH producing 143 sherds,while trench J yielded 2,125. Trench Jwas VILL TrenchH thereforeused to provide the primary basis for the clas sification. Close to the end of the season, large amounts to from Kharabeh Trench H encompassed the adjacent corners of two of pottery started be recovered Qalah a Bansaran Fort. Time rooms in one of the c. 228 m. long buildings detected by and to lesser extent the only of this material and the magnetometer survey (Fig. 12).42 The excavations did allowed basic processing of these awaits not produce an abundance of finds, but the 143 sherds, recording and analysis assemblages season. concentrated mostly towards the floor level, do suggest another field a for the clas domestic use. This is consistentwith the interpretationof In order to provide reliable foundation the itwas to see all of these buildings as barrack blocks.43 Seven sherds from sification of assemblage, necessary out This seemed trench H showed signs of having been burned after the pottery spread together (Fig. 17). core. for an on first breakage with black staining across the Evidence of particularly important assemblage, which, limited obvious variation. On the burning was not noted as a prominent featurewithin the impressions, displayed a one is with a landlocked excavation deposits, and this may or may not be Gorgan Plain essentially dealing 42 Omrani et al 2007: 113-26. 44 43 Omrani et al 2007: 114^16. Omrani al. 2007: 125. This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions THE GREAT WALL OF GORGAN AND THE WALL OF TAMMISHE 171 Fig. 17. All of thepottery from trenchJ spread out togetherfor sorting by class. area, where a similar range of fabrics appear to have been and Table 1). The most common classes are plain vessels used over long periods of time. with a fine, hard fabric and burnished surfaces (REDBUR, PINBUR, MOTBUR). The most common form for these classes is a small tomedium sized jar with VII. 4. The assemblage an off-set rim (Type Jl). There are also two well represented classes with an unburnished cream coloured The assemblage from Fort 4 contains no glazed pottery45 fabric (SELSCEP, CREWE). The most distinctive class and very littlematerial with surface decoration (Fig. 18 is a coarse white grit-tempered cooking-pot (WIGTEM). These vessels all have the same form with a round 45 bottom, a low squat profile, a flaring rim and rounded The absence of glazed potteryup until the early Islamic handles attached at the shoulder period has also been noted atMerv (Herrmannet al 1993: (Type CP1). 52) and inprevious excavations of theGorgan Wall (Kiani 1982: 36). In the case of theGorgan Wall this requires furthercritical examination. At least threesherds seemed to have very small tracesof a green alkaline-glaze adhering. It may be that a small portion of the classes with cream coloured fabric (CREWE, SELSCEP) were originally difficultto differentiate between post-depositional accretions glazed and thatthe glaze has exfoliated.In some cases it is and degraded glaze remnants. This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions 172 HAMID OMRANI REKAVAND ET AL TABLE 1. Summary of the classes recognised within the Fort 4 assemblages from trenches J and H. Class Code Class Name Types Description CREWE Cream Ware Hard, cream-coloured fabric with plain unburnished surfaces. MIGTEM Mixed GritTempered Ware CP2 Brittle,heavily reduceddark-brown or orange cookingpots temperedwith a mixtureof greyand red schistflakes and crystalline grits. Surfaces are decorated with impressed horizontal and vertical groves. MOTBUR Mottled Brown/Purple Hard, brown,purple, dull orangeor grey slightlysoft fabric. Burnished are Ware Surfaces burnished with heavy striations. Mostly undecorated. Oven OVEN Fragment Small roughlyformed blocks derivefrom thebroken up sections of crudely formed coil-built oven walls. REDBUR Fine Burnished Red Ware Jl, J3 Hard, red, orange or pink earthenware with occasional patches of reducedgrey. Walls tendto be thinand are always burnished with strong striations. Undecorated. REDPLI Plain or Incised Red Ware J1,J2, J4 Hard, orange earthenware sometimes with a reduced grey core. Surfaces are mostly unburnished and have incised decoration. REDREG Reduced Red andGrey Ware Hard, slightlybrittle fabric fired through the core to a dull oxidised orange-brown. Exterior surfaces are consistently reduced to a smoky grey. Undecorated. PINBUR Fine BurnishedCream/Pink Jl, J3 Hard, pink,cream or lightorange fabricoften reddening Ware towards the core. Surfaces are smooth and burnished with strong striations. Undecorated. SELSCEP Self-SlippedCream and Pink B1, B2 Hard earthenware,light pink to cream throughthe core turning Ware to cream on the surface. Surfaces are unburnished. Mostly undecorated. Smooth SMOG Grey Ware Hard, consistently reduced light grey earthenware with a distinctive smooth chalky feel. Undecorated. SOBRO Soft Brown Ware Hard, orange/brown earthenware with a smooth leathery feel. Surfaces are heavily worn. SOREB Soft Red Burnished Ware Hard, brown earthenware with a smooth reddish brown slipped and burnished exterior and thick cream slipped interior. WIGTEM CoarseWhite GritTempered CP1 Brittle,coarse, white grit-tempered,hand-built cooking-pots Ware fired either to a buff-brown or black. VII. 5. Analysis icantly less diagnostic, and within this assemblage almost all bases are flat and otherwise featureless. This leaves There are two factors thatmake attention to vessel form just 6% of rims as diagnostic sherds. and diagnostic type particularly importantfor the ceramic Breaking the assemblage down firstby class and then assemblage of Golestan. First, the lack of significant into broader class families, it is possible to gain some an diversity caused by the absence of imports in impression of its functional composition (Fig. 20). essentially landlocked setting, and second, the apparent Cooking-pots form an important component, though continuity of fabric composition and firingmodes across theirconcentration tends to vary depending on thenature chronological periods, which ismost clearly observed in of the contexts. This is likely to partially reflect context multi-period surface survey assemblages. Unfortunately use, butmay also be related to conditions of preservation. only a very small proportion of pieces from the Conditions in the gully,where there is a high concentra excavations are diagnostic. From the collection of over tion,may have been better suited to the survival of this 2,000 sherds from trench J, just 10% come from rim or particularly fragile class. Alternatively, if the gully base portions (Fig. 19). In general bases tend to be signif functioned as a drain rather than as a water channel, then This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions THE GREAT WALL OF GORGAN AND THE WALL OF TAMMISHE 173 m 1 I j.256,9 v| I s r-1?Jk 2 \ j.256,10 Wl I j.371,6 Hfl| \ 6 \ I j.148,13 V 3 ) j.202,2 W ^ 0 'I I A y/ \~~ j.215,10 " \^ 'W^SmS^P-^kf^ ^ 12 / u?/ j.215,48 8 /,313,1 > -1 j.450,1W 10 f j.283,2 13 9 % j.423,45 \\ ' 11 ^"^^^^^^ j.364,2 cm = = = = 0 = 3 Fig. 18. Illustration of thediagnostic typeswithin theFort 4 assemblage. 1-2 Jl; 3-4 J2; 5 J3; 6-7 J4; 8-9 = = CP1; 10= CP2; 11 Bl; 12-13 B2, drawn byMohaddeseh Mansouri Razi. Where thereare repeated examples a a of a particular form, two at minimum, type number has been assigned and recorded. Type groups are subdivided by broad vessel category into cooking pots (CP 1>), jars (Jl>) and bowls (Bl>). broken cooking-pots may have been deliberately added Another feature thatmay be of importance is the in order to break up the accumulation of fine sediment composition of the assemblage in terms of vessel shape and prevent drain blocking.46 and size. Just 4% of the sherds in the assemblage come from bowls. Almost all vessels are jars and most fall 46 within a small tomedium size Notable is the I am grateful to Dr JonathanTubb for providing this range. rapid suggestionbased on his experience at Tel es-Sa'idiyeh in fall off in theproportion of vessels within the assemblage Jordan and other excavations. over the 21-25 cm. rim diameter range (Fig. 21). There This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions 174 HAMID OMRANI REKAVAND ET AL Proportion of Rim, Body and Base Sherds from Trench J OH Base Rim Side Fig. 19. Relative proportion of the rims, bases and body _I sherds in the trenchJ assemblage. Proportion of Main Ceramic Groups in the Trench J Assemblage ^^"^^H^^^^^^^^^ Other ^^^^^^^^^^^^H CreamColour 1 ^mB^^^^^^^^^M Cooking-Pots ? Wafe \ 20. class Fig. Proportionof families )i 'r$S^^^^^ represented within the trench J assemblage. are very few large vessels and no sherds thatone would VII. 6. Early Material characterise as storage-jars. This seems to be a peculiar feature, and there remains a distinct possibility that the By far themost interestingchronological dimension that functional composition represents a particular hallmark of was established, came from two deposits of differing soil the late Sasanian military assemblage. colour and with a greater concentration of finds thatwere Apart from the functional composition of the picked up at either end of the deep sondage in trench J assemblage, one of themost importantfeatures to note is (contexts J.027 and J.030). Unfortunately due to the that the classes from both of the trencheswithin Fort 4 constraints of time and safety, itwas not possible to are the same. Also all of the classes are represented from continue the exposure of these features or to provide a the uppermost layers down to the lowest deposits, with conclusive interpretationof theiruse. What is clear is that one important exception (see below). This seems to fit they occur at an elevation below the foundations of the well with the idea that themajor occupation and use of barrack blocks in trenchH and are probably, therefore, the Gorgan Wall and its associated forts, falls within a pits or wells reaching underneath the original ground relatively narrow time-span. surface. This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions THE GREAT WALL OF GORGAN AND THE WALL OF TAMMISHE 175 Rim Diameter Range from Trench J 50-j 0-5 6-10 11-15 16-20 21-25 26-30 31-35 36-40 41-45 46-50 Fig. 21. Number of rimswithin the trenchJ assemblage broken down Rim Diameter (cm) according to estimated diameter _I range. Amongst the small assemblage recovered from thanMIGTEM). Also, what is the broader geographic contexts J.027 and J.030, therewere just three sherds of distribution of this assemblage? Some of the material to a very distinctive cooking-pot (MIGTEM). These have a excavated appears have good parallels with the different form and fabric to themain cooking-pot class assemblage described by Kiani, particularly the group of found through the rest of the sequence (WIGTEM) and "Sasanian red wares".47 The exception here is any have impressed markings covering the exterior. The mention of cooking-pots, however the pottery report is material did not immediately stand out until pottery selective and these may have been encountered, but startedbeing recovered from a second site thatwas being omitted from the description. most excavated at Qaleh Kharabeh, notably from trenchM The substantial and well documented sequence (see section III above). for theGorgan region to date comes from Tureng Tepe It is significant that all aspects of the ceramic situated south of theGorgan Wall. The uppermost 5m. of assemblage from Qaleh Kharabeh are different to the deposits in the "Grand Tepe" cover the period from the 48 main deposits associated with theuse of theGorgan Wall third-fourteenthcenturies The major structurewithin a at Fort 4. Also, the ubiquitous cooking-pot from Qaleh thisperiod is Sasanian mud-brick fortification.After the Kharabeh is the same as the distinctive sherds recovered fortwas abandoned, the sitewas reused as the setting for a from the deepest deposits at Fort 4. Although the small fire temple, dated to the seventh/eighthcenturies. evidence at present rests on just a few pieces, the Minor activity, mostly involving the excavation and association is very strong and the implications are clear. filling of pits, continued into the Islamic period. Qaleh Kharabeh was occupied in the same period as the It is surprising, given the chronological overlap firstactivity at Fort 4, most likely its foundation. Could between the Tureng Tepe sequence and the use of the itbe thatQaleh Kharabeh acted as a temporary installa Gorgan Wall, that there is not a close association in terms tion used by soldiers involved in theWall's construction of the pottery. There are some indications of a link or in shoring up the border regions tomake construction between the latest occupation of the fort (Periods VIA/B) possible? and the assemblage from Qaleh Kharabeh, particularly in the two pieces of coarse incised cooking-pot49 and numerous jars with flaring necks and a simple bevelled VII. 7.Discussion lip.For the assemblage from themain part of the sequence at Fort 4 thebest parallels arewith Period VTIC, with, for Clearly this is just thebeginning and much more remains example, "lid-bowls", a very specific lamp form and a to be learnt about the sequence of events associated with the construction and use of theGorgan Wall. One of the first thatneeds to be iswhether the 47 questions, answered, Kiani 1982: 21, 36, figs 27.3,28. main Fort 4 is 48 assemblage replicated along the length of Boucharlat and Lecomte 1987;Whitehouse 1992: 380-81. theWall those 49 (i.e. classes represented in Table 1 other Boucharlat and Lecomte 1987: fig. 54.14-15. This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions 176 HAMID OMRANI REKAVAND ET AL glass flask.50Period VIIC also contains numerous coarse also to a significant extent in terms of vessel forms. round bottomed cooking-pots, "ceramique grossiere". Presumably thismust be a basic reflection of differing Some have a distinctive notch on the inside of the rim for patterns of consumption. Although this is a small and a lid,but the simple formwith an everted lip51could be the rathermodest contribution, it is particularly encouraging same as the main cooking-pot class from Fort 4 to see how the study of pottery can feed rapidly into (WIGTEM,Type CP1). The problemwith Period VIIC is much larger questions relating to the project as a whole. that the pits were excavated after the abandonment of the fire temple and they also contain quantities of material clearly dating to the ninth-tenth centuries. This question VIII. CONCLUSIONS needs be looked into inmore detail. Elsewhere on theGorgan Plain, thematerial from the While more fieldwork is required to obtain an absolute upper strata at Shah Tepe52 appears to be dated later than chronology for settlement expansion and abandonment in the occupation of Fort 4 and finds better association with the strip of land north of theGorgan Wall, itnow seems the early Islamic levels represented at theBansaran Fort thatmost sites in this area had long been derelict by the on theTammishe Wall.53 Moving furthernorth, there are time theWall was built. By this time theWall had cut no obvious parallels with a small selection of material across two redundant canals, which had once supplied a presented from survey in theAtrek Valley, though there thriving agricultural landscape with water. By contrast, are superficial points of similarity in the description.54 Qaleh Kharabeh south of theWall may date to around the One site with good parallels for both themain jar and time of theWall's construction. This observation, and the cooking-pot forms represented at Fort 4 (Types Jl and discovery of remarkably regular rows of enclosures in the CP1) is the unpublished site of Khalni Depe, excavated interior,indicate that this fortress (and probably the archi by Soviet archaeologists in the Misrian Oasis in tecturally similar other large square hinterland fortresses) Dehistan.55 Dehistan clearly falls within the same geo played an importantmilitary role during the construction graphical catchment as theGorgan Plain. and/or early occupation of the Wall. New evidence some Looking more widely tomaterial of comparative date emerged to suggest that of thewater, needed for from Merv56 to the east, or Shahr-i Qumis57 and brick production to build theWall, was channelled there Damghan to the south,58 there are no obvious points of over some distance via an aqueduct (rather than coming a as Fort 4 comparison. This points again to the fact that the distri from reservoir, thoughtbefore). Fieldwork in into the bution of pottery across the Sasanian Empire was firmly and Qaleh Kharabeh yielded yet furtherinsights the to bound by themajor geographic lines of demarcation.59 If sophisticated technology used by Sasanians supply one goes south of the Elburz or east of theKopet Dagh, these sites with water. Diving operations confirmed the it appears that one enters a new cultural world. Not just presence of extensive collapsed brick-built structures runs in terms of clay sourcing and product distribution, but where the roughly contemporary Tammishe Wall into terrain now flooded by the Caspian Sea. The exploration of a three-aisled hall in a fort near the 50 Boucharlat and Lecomte 1987: 103.14. figs8La-f, j-k, to which 51 Tammishe Wall added strength literarysources, Boucharlat and Lecomte 1987: fig. 84.a-c, h-j. 52 suggested that installations associated with thisWall Arne 1945: pl. 88.723a-c. 53 outlasted the Wall. Pottery analysis and a Bivar and Fehervari 1966: 46, pl. IV.a-f. Gorgan 54 scientific are to build a Venco Ricciardi 1980: 67, note 24, figsK.l-10, L.l-3. dating programme beginning up 55 the relative and absolute Sarianidi 1952: figs 7.7-8,12-13,17; I am gratefulto Dr St detailed picture of sequence John Simpson for bringing this site tomy attentionand dating of these key developments: the early settlement providing thereference. expansion into the steppe, followed by the construction, 56 In addition to the Seth Priestmanhas been publishedpottery occupation and decline of twomajor linear barriers and of the from able to look throughthe largecollection pottery theirassociated fortsand fortresses. Merv held in theMiddle East Department at theBritish Museum. I am extremelygrateful to Dr St JohnSimpson for guidingme throughthe material. 57 Hansman 1968: 127;Hansman and Stronach 1970: 55-56, Acknowledgements 61. 58 Trinkaus 1986. We are indebted to Dr Seyed Taha Hashemi, the vice 59 and Tourism Simpson 1997: 74; Huff 1987: 307. director of the Iranian Cultural Heritage This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions THE GREAT WALL OF GORGAN AND THE WALL OF TAMMISHE 177 R. and O. 1987.Fouilles de Organisation (ICHTO) and the head of the Research Boucharlat, Lecomte, TurengTepe Department of the ICHTO, and toDr Hassan Fazeli, the sous la direction de Jean Deshayes 1. Les periodes director of the Iranian Center ofArchaeological Research Sassanides et Islamiques, Paris. (ICAR), for theirkind permission to carry out our project Browne, E.G. 1905.An Abridged Translation of theHistory of about 613 and for facilitating and supporting our work in many Tabaristan compiled AH. (A.D. 1216) by B. B. Leiden and ways. We are grateful toDr Seyed Mehdi Mousavi, the Muhammad Al-Hassan Isfandiydr, vice-director of theResearch Department of the ICHTO, London. in users for granting permission to process samples in theUK, Farr,T. 2004. "Reply thediscussion forumfor of SRTM and toMr Fereidoon Faali, the director of theGolestan data." URL: http://pub7.bravenet.conVforum/537683448/ 21 ICHTO, for his help, advice and personal interest in our fetch/394357/Oast day accessed: June2005). work. We would like to thankLeyla Safa'ie, of the inter Frye,R.N. 1972. "Byzantineand Sasanian Trade Relations with national section of the ICAR, Mr Fereydoun Unagh, the Northeastern Russia", Dumbarton Oaks Papers 26:263-69. director of ICHTO at Gonbad-e Kavus, and members of Galiatsatos, N. 2004. Assessment of theCORONA series of the local ICHTO office for all their support of our satellite imageryfor landscape archaeology: a case study research project. from theOrontes valley,Syria, PhD dissertation,Durham We are most grateful for the generous and essential University. support by the British Institute of Persian Studies, the Goldina, RD. andNikitin, A.B. 1997. "New findsof Sasanian, AHRC, the ICHTO and Edinburgh University's School CentralAsian andByzantine coinsfrom theregion of Perm, ofHistory, Classics and Archaeology. the Kama-Urals area", inK. Tanabe, J. Cribb and H. Wang We would like to thank Maryam Hussein-Zadeh, (eds.), Studies inSilk Road Coins and Culture.Papers in Mohaddeseh Mansouri Razi and Atieh Esmaili for their honour ofProfessor Ikuo Hirayama on his 65th birthday, excellent drawings, to our drivers, notably Mr Kamakura: 111-30. Abdolhussein Badpa, and tomany academic supporters Goncalves, J. and Fernandes, J.C. 2005. "Assessment of SRTM and our workmen, which space does not allow us to list. 3 DEM in Portugal with topographic map data", Proceedings of theEARSeL Workshop 3D-RemoteSensing, 10-11 June, Porto, Portugal (European Association of Bibliography Remote SensingLaboratories), unpaginated CD-ROM. Hansman, J. 1968. "The problems of Qumis", Journal of the Adams, R.McC. and Hansen, D.R 1968. "Archaeological Royal Asiatic Society of Great Britain & Ireland 1968: Reconnaissance and Soundings in JundT Shapur", Ars 110-39. ? Orientalis 7: 53-70. and Stronach, D. 1970. "Excavations at Shahr-e Qumis, Amin Pour, B. and Omrani Rekavandi, H. 2007. 1967",Journal of theRoyal Asiatic Society ofGreat Britain & Ireland 1970: 29-62. lAjjO^ JO u^^? t5^?j jj." inArchaeological Reports Harper, P.O. 2000. "Sasanian SilverVessels: The Formationand 7,On theoccasion of the9th Annual Symposiumon Iranian Study of Early Museum Collections", in J. Curtis (ed.), Archaeology (ResearchCentre for ICHHTO, IranianCenter Mesopotamia and Iran in theParthian and Sasanian Periods: forArchaeological Research), vol. 1,Tehran: A t-VV Rejection and Revival c. 238 BC-A.D. 642, London: 46-56. Arne, T.J. 1945. Excavations at Shah Tepe, Iran, Reports from Herrmann, G., Masson, V.M., Kurbansakhatov, K. et al. 1993. theScientific Expedition to theNorthwestern Provinces of "The InternationalMerv Project: preliminaryreport on the China under the Leadership of Dr Sven Hedin, firstseason (1992)", Iran 31: 39-62. Stockholm. Huff, D. 1987. "Archaeology iv. Sasanian", EIr 2, London and Ball, W. 2000. Rome in theEast. The transformationof an New York: 302-8. Empire, London andNew York. Jacobsen,K. 2005. "Analyses of SRTM elevationmodels", Bivar, A.D.H. and Fehervari, G. 1966. "The Walls of Proceedings of theEARSeL Workshop 3D-RemoteSensing, Tammlsha", Iran 4: 35-50. 10-11 June, Porto, Portugal (European Association of H. Borm, 2007. Prokop und die Perser: Untersuchungen zu den Remote Sensing Laboratories),unpaginated CD-ROM. romisch-sasanidischen Kontakten in der ausgehenden Kay, S., Spruyt,P., Zielinski, R, Winkler, P.,Mihaly, S. and Ivan, Spdtantike,Oriens etOccidens 16,Stuttgart. G. 2005. "Quality checkingof DEM derivedfrom satellite Bosworth, C.E. 1985. "Abaskun", EIr 1, London, Boston and data (SPOT and SRTM)", Proceedings of theEARSeL 69-70. 3D-Remote Henley: Workshop Sensing, 10-11 June, Porto, Portugal This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions 178 HAMID OMRANI REKAVAND ETAL (EuropeanAssociation of Remote Sensing Laboratories), Rabus, B., Eineder,M., Roth, A. and Bamler, R 2003. "The unpaginatedCD-ROM. shuttleradar topographymission?a new class of digital Kiani, M.Y. 1982. Parthian Sites inHyrcania. The Gurgan elevationmodels acquired by spacebome radar", ISPRS Plain, Archaologische Mitteilungen aus Iran Journal of photogrammetry and remote sensing 57: 241-62. Erganzungsband 9, Berlin. Sarianidi,V.J. 1952.Ancient City ofKhalni-Depe, Unpublished Mehren,A.F. (ed.) 1874.Manuel de la Cosmographie duMoyen Diploma, Historical Faculty,University of Tashkent. Age, traduitde TArabe "Nokhbeted-dahr fi cadjaib-il-birr Schippmann, K. 1990. Grundziige der Geschichte des wal-bah^r" de Shems Ed-din Abou-Abdallah Mohcammed Sasanidischen Reiches, Darmstadt. de Damas. Copenhagen. Simpson, S.J. 1997. "Partho-Sasanian ceramic industries in Nokandeh, J., Sauer, E., Omrani Rekavandi, H., Wilkinson, T., Mesopotamia", in I. Freestone and D. Gamester (eds.), Abbasi, G.A., Schwenninger, J.-L., Mahmoudi, M., Parker, Pottery in theMaking. World Ceramic Traditions,London: D. ,Fattahi, M., Usher-Wilson, L.S., Ershadi, M., Ratcliffe, J. 7^-79. and Gale, R. 2006. "Linear Barriers ofNorthern Iran:The Trinkaus,K.M. 1986. "Potteryfrom theDamghan Plain, Iran: GreatWall ofGorgan and theWall of Tammishe", Iran 44: chronologyand variabilityfrom theParthian to theEarly 121-73. Islamic periods", Stadia Iranica 15.1: 23-88. Omrani Rekavandi, H., Sauer, E., Wilkinson, T., Safari Tamak, Tucker, C.J, Grant, D.M. and Dykstra, J.D. 2004. "NASA's E. ,Ainslie, R, Mahmoudi, M., Griffiths,S., Ershadi,M., global orthorectifiedLandsat data set",Photogrammetric Jansen Van Rensburg, J., Fattahi, M., Ratcliffe, J., Engineering & Remote Sensing 70.3: 313-22. Nokandeh, J., Nazifi, A., Thomas, R., Gale, R. and Venco Ricciardi,R. 1980. "Archaeological survey in theupper Hoffinann,B. 2007. "An ImperialFrontier of the Sasanian Atrek valley (Khorassan, Iran): preliminary report", Empire: furtherfieldwork at theGreat Wall ofGorgan", Iran Mesopotamia 15: 51-72. 45: 95-136. Whitehouse, D. 1992. "Review of Boucharlat and Lecomte Petrikovits, H. von 1975. Die Innenbauten romischer 1987",American Journal ofArchaeology 98: 380-81. Legionaslager wdhrend der Prinzipatszeit, Opladen. Yate, C.E. 1900. Khurasan and Sistan, Edinburgh and London. This content downloaded from 129.215.19.197 on Mon, 16 Dec 2013 07:24:24 AM All use subject to JSTOR Terms and Conditions