Molyvoti, Thrace, Archaeological Project: 2014 Preliminary Report

Molyvoti,  Thrace  Archaeological  Project:  “Ancient  Stryme”    2014  Preliminary  Report   Forthcoming  from  Archaiologikon  Deltion     Nathan  T.  Arrington,  Domna  Terzopoulou,  Marina  Tasaklaki,  Thomas  F.  Tartaron     The  second  season  of  the  Molyvoti,  Thrace  Archaeological  Project  (MTAP)  took  place  from  June  9  to  July   18,   2014.   Excavation   continued   in   the   three   areas   investigated   in   2013—a   house,   a   crossroads,   and   a   pithos   installation,   all   dating   primarily   to   the   Classical   period—and   an   intensive   surface   survey   began   that   focused   on   the   urban   area   of   the   city   and   its   hinterland.   The   overall   goals   of   the   project   are   to   explore   the   identity,   form,   and   chronology   of   the   site   inconclusively   designated   ancient   Stryme;   to   understand   its   various   roles   in   regional   communication   and   exchange   networks;   and   to   assess   its   evolving  relationship  with  the  landscape  and  local  populations.     Excavation   In   2014,   20   5X5m   squares   were   excavated   (Figure   1).   826   kg   of   pottery   and   1668   kg   of   tile   were   processed;   611   finds,   including   118   coins,   registered;   8,832   bone   fragments   were   analyzed;   and   757   liters  of  archaeological  sediment  were  floated.     2.5.2   ΑΔ/30β   was   opened   to   contextualize   the   large   inscribed   pithos   discovered   near   a   Bulgarian   war   trench  in  2013.  A  pi-­‐shaped  feature  made  of  field  stones  was  located   about  1  m  from  the  pithos,  and   probably  once  associated  with  a  fragment  of  floor  surface.  The  feature  was  disturbed  by  a  wide,  shallow   cut,   at   the   bottom   of   which   were   fragments   of   an   iron   blade.   This   square   was   heavily   damaged   by   modern  activity.     2.5.4  ΑΓ/5β,  2.5.4  ΑΔ/5α,  2.5.4  ΑΔ/5β,  and  2.5.4  ΑΔ/5δ  were  opened  to  uncover  the  full  extent  of  the   Classical  crossroad  partially  revealed  in  2013  (Figure  2).  Two  large,  unworked  granite  blocks  were  placed   within   the   road   fill   at   the   southwest   and   southeast   corners.   A   road   sounding   0.60   m   deep   produced   material   well   below   the   bottom   of   the   walls   that   securely   dated   the   construction   of   the   road   and   of   the   Hippodamian  grid  plan  to  the  4th  cen.  BC.  In  the  east,  the  road  lay  only  a  few  centimeters  over  an  earlier   structure.   Patches   of   a   pebble   floor   were   preserved   as   well   as   a   substructure   of   large   worked   and   unworked  stones  and  reused  architectural  members.  Although  parallel  to  the  houses  on  the  4th-­‐cen.  BC   grid  plan,  it  clearly  belonged  to  an  earlier  phase  of  the  city,  but  no  loci  have  yet  been  excavated  that  can   provide  a  secure  date.     The  northeast  corner  of  the  crossroads  belonged  to  a  large  Classical  house  on  the  Hippodamian  grid  plan   that  measured  approximately  18.5  X  18.5m  according  to  the  geophysical  survey.  A  shallow  stone  drain   extended  from  the  house  into  the  road.  Excavation  inside  the  structure  revealed  a  large  deposit  of  tile   and   pottery,   particularly   from   transportation,   storage,   and   cooking   vessels   (Figure   3).   At   the   southeastern  edge  of  the  house,  near  the  spine  wall,  a  southeastern  Aegean  (possibly  Knidian)  amphora   rim   dated   the   abandonment   of   the   structure   275–250   B.C.   From   the   upper,   disturbed   strata   of   the   house  was  found  a  Roman  carnelian  gemstone  depicting  Hermes  (Figure  4).    

Work  continued  in  the  adjacent  southeast  house  first  excavated  in  the  1990s,  revealing  8.5  m.  more  of   the   spine   wall,   the   probable   southwest   exterior   wall,   and   several   internal   walls   defining   new   spaces   (2.5.4  ΑΖ/5γ,  2.5.4  ΑΖ/6α,  2.5.4  ΑΖ/6β,  2.5.4  ΑΖ/6γ,  2.5.4  ΑΖ/6δ,  2.6.3  Α/5δ,  2.6.3  A/6β,  2.6.3  A/7α,  2.5.4   ΑΖ/5β,   2.5.4   ΑΖ/5δ,   2.6.3   Α/5α,   2.6.3   Α/5γ,   2.6.3   A/6α,   2.6.3   A/6γ,   2.6.3   A/6δ;   Figure   5).   The   tile   deposit   in  space  α,  which  was  exposed  and  defined  in  2013,  was  removed  this  season.  Excavation  lower  than  the   floor   revealed   a   significant   quantity   of   Archaic   sherds,   though   the   stratum   itself   dates   to   ca.   400   B.C.   Space  γ  is  enclosed  by  walls  which  all  date  to  the  Late  Roman  reuse  of  the  house  (see  further  below),   though  not  necessarily  coterminous.  Space  δ  is  enclosed  in  the  northeast  by  walls  that  are  Late  Roman   but   not   necessarily   coterminous,   and   in   the   southwest   by   walls   that   are   likely   Classical–Hellenistic.   Removal   of   the   tiles   and   associated   features   revealed   red   and   yellow   wall   plaster,   some   of   it   in   situ,   and   a   tile   inscribed   ΑΡΙΣΤΟΚΛΕΟΣ.   Most   of   the   destruction   material   from   space   δ   may   have   been   redeposited  in   space   ζ,   where   debris   was   piled   25   cm   deep.   Among   the   upper   strata   of   the   tiles   were   splendid   antefix   fragments   with   parallels   at   Abdera.     Space   β   is   a   large   area   that   occupies   portions   of   four  grid  squares  and  possibly  once  formed  a  courtyard.  Two  uninscribed  lead  sling  bullets  were  found   here.   To   the   northwest   of   space   β,   in   the   corridor   between   the   spine   wall   and   space   ε,   tiles   were   deposited  in  a  thick  stratum  along  with  remains  from  storage  vessels  and  a  significant  number  of  finds,   such   as   fragments   from   fourth-­‐century   cup-­‐kantharoi,   a   coin,   a   lamp,   an   inscribed   tile,   an   amphora   handle   stamped   [θ]AΣΙΩΝ   dating   ca.   315–310,   and   two   loom   weights.   Underneath   this   debris   was   a   drain  constructed  of  large  pan  tiles.  Some  of  the  debris  probably  was  redeposited  from  space  ε,  because   excavation  here  uncovered  a  loose,  dark  soil  with  very  few  tiles.  In  spaces  στ,  η,  and  θ,  only  the  upper   strata  were  excavated  in  2014,  and  possibly  the  house’s  southwesternmost  wall.         It   appears   that   in   the   Late   Roman   period,   destruction   debris   throughout   the   house   was   moved,   redeposited,   and/or   leveled   in   order   to   modify   and   reuse   it.   Buttress   walls   were   built   to   support   the   spine   wall,   the   upper   courses   of   Classical   foundations   were   re-­‐articulated   and   re-­‐appropriated,   and   new   walls  were  built.  The  longest  wall,  L13-­‐096,  may  span  the  whole  width  of  the  house.  Two  long,  narrow,   deep   cuts   provide   crucial   evidence   for   dating   this   activity.1   They   cut   the   Classical   walls   of   the   house   and   go   under   Wall   L13-­‐096.   It   is   uncertain   what   purpose   they   served;   perhaps   they   removed   drains.   The   latest  pottery  in  them  dates  from  the  later  4th  to  the  early  5th  cen.  AD  and  thus  provides  a  terminus  post   quem  for  L13-­‐096  and  the  associated  first  Late  Roman  architectural  phase.  With  this  phase  belong  two   pithoi   to   the   southeast   of   L13-­‐096,   whose   cut   extends   to   the   southeast,   where   a   third   pithos   once   must   have  been  located.     In  a  second  Late  Roman  phase,  L13-­‐096  and  the  pithoi  went  out  of  use.  The  area  was  leveled,  cutting  the   tops  of  the  pithoi.  The  foundations  of  a  large  circular  feature  were  built  directly  over  L13-­‐096,  and  a  dark   fill  was  laid  in  and  around  the  foundations.  Sherds  from  the  dark  fill  date  the  construction  of  the  feature     to   the   later   5th–   first   half   6th   cen.   AD.   The   fill   consisted   of   rubble,   ceramics,   and   large   quantities   of   redeposited   waste,   with   significant   amounts   of   bones   (some  nearly   complete  skeletons)  and  a   diverse   botanic   assemblage   of   crop   weeds   such   as   rye   grass,   brome   grass,   cereals,   and   legumes.   Among   the   bones   was   a   complete   skeleton   of   a   baby   deposited   between   the   first   and   second   course   of   the   foundations,   placed   chest   down   and   legs   turned   to   the   side   (Figure   6).   There   was   no   sign   of   a   cutting   around  the  baby  or   of  a  surface  on  which  it  was  resting,  but  a  small  blue  bead  20cm  away  is  perhaps   associated  with  it.                                                                                                                             1

 Nicholas  Hudson  is  studying  the  Late  Roman  activity  at  Molyvoti.    

Survey   The   surface   survey   explored   chronological   and   functional   variation   across   the   city   and   its   immediate   hinterland,  seeking  to  determine  the  extent  of  the  city  and  to  investigate  how  different  areas  were  used,   and  beginning  to  integrate  the  settlement  into  a  diachronic  history  of  the  landscape.  Surface  survey  was   conducted   in   Zone   A,   which   includes   the   Molyvoti   Peninsula   and   the   fields   on   the   mainland   in   close   proximity.   This   zone   was   divided   into   “Tracts,”   corresponding   to   a   plowed   field,   which   were   further   divided   into   20X20m   square   “Survey   Units.”   Survey   was   conducted   in   cotton   fields,   which   provided   good,   consistent,   and   comparable   visibility,   and   in   one   area   cleared   within   the   fenced   archaeological   zone.   4   walkers   spaced   at   5m   intervals   counted   pottery   and   tile.   They   collected   (1)   diagnostic   feature   sherds,  including  rims,  bases,  handles,  and  wall  pieces  with  surface  paint  or  decoration;  (2)  an  example   of  unique  pottery  fabrics  if  represented  only  in  body  sherds;  and  (3)  an  example  of  each  unique  type  of   tile.  They  also  gathered  other  artifacts  and  ecofacts  and  recorded  the  findspots  with  handheld  GPS.  20%   of   the   units   were   designated   “Total   Collection   Units,”   from   which   all   material   was   gathered   from   a   central  5  m2  circle.  The  first  collection  technique  measured  the  overall  quantity  of  ceramics  and  tiles  and   provided   presence/absence   data,   while   the   second   collection   technique   can   be   used   to   make   quantitative   comparisons   across   the   site.   1040   20X20m   squares   were   surveyed   in   2014,   with   19%   of   Zone  A’  walked  and  11%  of  Zone  A’  viewed  (243,174.76  m2).    321,752  sherds  and  72,290  tile  fragments   were  counted,  and  258  finds  registered,  including  67  coins.     The  northeastern  wall  of  the  city  was  clear  but  the  northwestern  wall  has  been  destroyed  by  plowing.   Sherds,  tiles,  and  coins  concentrated  within  the  city  walls  and  produced  a  sharp  dropoff  northeast  of  the   city,   while   the   plowing   in   the   northwest   spread   material   across   the   fields,   yielding   a   diffuse   decline   in   material  in  the  region  where  Bakalakis  saw  the  northwest  wall  (Figures  7–8).     The   Classical–Late   Classical   period   was   overwhelmingly   represented   in   the   survey   units,   both   in   and   outside  of  the  city.  Very  few  units  produced  material  from  other  periods.  Both  Archaic  and  Hellenistic   later   than   300   were   extremely   rare.   The   peninsula   was   abandoned   in   the   later   Hellenistic   and   early   Roman  period.  Survey  found  Late  Roman–Byzantine  material,  but  it  concentrated  most  densely  at  a  new   area:   a   site   at   the   southwest   of   the   peninsula.   A   preliminary   assessment   of   the   finds   suggests   a   prosperous  agricultural  estate,  possibly  6  ha  in  extent,  with  access  to  fineware  from  abroad,  glassware,   and  plenty  of  goods  in  local  and  imported  amphoras.  Elsewhere,  Late  Roman  material  clustered  near  the   region   currently   being   excavated.   The   survey   together   with   the   excavation   thus   indicate   that   activity   on   the  peninsula  in  the  Late  Roman  and  Byzantine  periods  had  sharply  nucleated.     From   the   Classical   period,   the   preponderance   of   amphoras   in   the   region   of   the   ancient   city   was   striking.   Nearly   every   survey   unit   in   and   out   of   the   city   produced   amphora   sherds,   often   in   tremendous   quantity   (Figure  9).  Though  the  majority  of  the  diagnostics  are  from  North  Aegean  types,  these  types  were  varied   in  their  fabric.  The  dense  amphora  finds  contrast  with  the  rarity  of  loom  weights:  only  6  loom  weights   were   found   on   survey,   and   only   6   have   been   found   in   the   excavation   of   the   southeastern   house.   Spread   across  the  site  in  much  greater  quantity  than  loom  weights  were  the  remains  of  hopper  mills.  Though   they   may   have   been   used   only   for   subsistence,   the   hopper   mills   could   also   suggest   that   some   of   the   trading  activity  at  the  settlement  in  the  Classical  period  was  oriented  towards  grain  cultivation.      

 

Figure  1.  Ephorate  excava3ons  before  MTAP,  and  MTAP  squares  of  2013  and  2014.     Map:  Eli  Weaverdyck.    

Figure  2.  Classical  crossroads  with  an  earlier  structure  underneath  the  road.  Grid  squares   2.5.4  ΑΓ/5β,  ΑΓ/5δ,  ΑΔ/5α,  ΑΔ/5β,  ΑΔ/5γ,  and  ΑΔ/5δ.  

Figure  3.  Deposit  of  3les  and  poUery  in  the  northeastern  house  bordering  the  crossroads.    

Figure  4.  Carnelian  gemstone  with  a  representa3on  of  Hermes,  ca.  1st–3rd  cen.  AD.     Drawing:  Samuel  Holzman.    

Figure  5.  House  da3ng  to  the   Classical  period  and  with  Late   Roman  reuse.    

Figure  6.  Skeleton  of  a  baby  along  the  founda3ons  of  the  circular  feature.    

Sherd Density

Sherds per square meter 0.01245 - 0.4183 0.4184 - 0.8575 0.8576 - 1.345 1.346 - 2.061

Map by Eli Weaverdyck

0

500

1,000 Meters

2.062 - 4.858

Figure  7.  PoUery  density  on  the  Molyvo3  Peninsula.  Map:  Eli  Weaverdyck.    

Tile Density

Tiles per square meter 0.0000 - 0.16926 0.16927 - 0.53000 0.53001 - 1.0979 1.0980 - 2.1000 Map by Eli Weaverdyck

0

500

1,000 Meters

2.1001 - 3.7691

Figure  8.  Tile  density  on  the  Molyvo3  Peninsula.  Map:  Eli  Weaverdyck.  

Amphora Distribution

Amphoras Present Map by Eli Weaverdyck

0

250

500 Meters

Figure  9.  Survey  units  with  amphora  sherds  present.  Map:  Eli  Weaverdyck.    

Absent