HESPERIA 78 (2OO9) THE 2005 CHIOS 2 Pages 69-3 05 ANCIENT SHIPWRECK SURVEY

New Methods for Underwater

Archaeology

abstract

In 2005 a Greek and American interdisciplinary team investigated two ship B.C. wrecks off the coast of Chios dating to the 4th-century and the 2nd/lst century.The project pioneered archaeological methods of precision acoustic, an digital image, and chemical survey using autonomous underwater vehicle (AUV) and in-situ sensors, increasing the speed of data acquisition while decreasing costs.The AUV recorded data revealing the physical dimensions, age, cargo, and preservation of thewrecks. The earlierwreck contained more than 350 amphoras, predominantly ofChian type,while theHellenistic wreck contained about 40 Dressel 1C amphoras. Molecular biological analysis of two amphoras from the 4th-century wreck revealed ancient DNA of olive, oregano, and possibly mastic, part of a cargo outbound fromChios.

introduction

summer an In the of2004 during archaeological and geological-geophysical in survey theChios Strait, archaeologists of theHellenic Ephorate of Under water Antiquities (EUA) and oceanographers of theHellenic Centre for a Marine Research (HCMR) discovered the remains of 4th-century B.C. shipwreck (Chios wreck A) between Chios and Oinousses.1 Recognizing the sonar significance of thewrecks side-scan and sub-bottom signatures, the team deployed theHCMR SuperAchilles remotely operated vehicle (ROV) to collect video images of the site. EUA maritime archaeologists reviewed the video and confirmed the archaeological importance of thewreck.2 After the scientists Greek made the initial discovery and characterization of the

1. There are In the EUA mat multiple shipwrecks points. naming convention, (chromophoric dissolved organic around so for the sake of Chios, clarity Chios wreck A is the Chios-Oinousses ter),DVL (Doppler velocity log), and we have to an chosen apply alphabetic wreck and Chios wreck B, described LBL (longbaseline). For these and convention in this article. The is the Lithi wreck. term out naming below, others, the full is spelled upon EUA uses a are officially differentnaming A number of acronyms used firstmention. convention for on in the text. Some that be unfamil shipwrecks, relying may 2. Sakellariou et al. 2007, pp. 371, the names of iar to nearby geographical archaeologists include CDOM 373.

? The American School of Classical Studies at Athens 270 BRENDAN P. FOLEY ET AL.

at site, they contacted personnel theWoods Hole Oceanographic Institu team tion (WHOI). Together, in 2005, the international jointly planned a an and conducted survey of the site using autonomous underwater vehicle a (AUV). Preliminary investigations of second, laterwreck (Chios wreck B), located off thewestern coast of Chios, were also carried out. We present here the results from this innovative archaeological ship an wreck survey, conducted with autonomous underwater robot carrying sensors. at a in-situ Chios wreck A lies depth of 70 m, deeper than is feasible to ever explore with scuba equipment. We do not expect that divers will we use visit this site, and therefore sought to this project to determine the present limits of underwater remote sensing for archaeological purposes. was One of the project goals to extract useful archaeological information as as from the site rapidly possible, while leaving the site undisturbed. We a new believe that this approach presents paradigm formaritime archaeo logical investigations, for itwill enable study of large numbers of directly over a comparable sites spread wide geographic range.With these powerful new technologies and surveymethods, underwater landscape archaeology becomes possible. In the pages that follow,we attempt to reach across academic and sci as entific disciplines in order to engage broad audiences in the humanities to well as scientific and engineering fields.Our intent is alert archaeologists new new to methods and to catalyze approaches to shipwreck investiga tions. At the same time, we wish to pique the interest of engineers and physical scientists in the hope of stimulating future collaborations with a social scientists.We begin with detailed explanation of the technologies a and methods involved in this type of survey.To establish professional standard of quality for other investigators interested in deepwater archaeol we ogy, compare the precision of themeasurements obtained in the 2005 to In the project accepted professional shallow-water practice. following section, we describe technologies enabling environmental sensing and the state and their artifacts.We next prediction of the of preservation ofwrecks present the archaeological information obtained from Chios wreck A and our to the interpretation of the findings,with special attention given large we with a vision cargo of amphoras and their contents. Finally, conclude for underwater archaeology's future. to waters Scuba-based underwater archaeology is limited shallow (less out reach. than 50 m), leaving approximately 98% of the seafloor of Deeper coastal waters hold vast numbers of shipwrecks, and historical data indicate that the seafloor far offshore contains 20%-23% of all wrecks.3 The large one number ofwrecks is only attractive aspect of deepwater archaeological wrecks are survey.Our experiences have shown that deepwater typically in water because are immune to better preserved than those shallow they disturbance from surfacewaves and from intrusion by divers. Unlike shal zone 100 low-water wrecks, artifacts lying beyond the phototropic (ca. m) or no marine on for easier visual typically have little growth them, allowing identification. reason to sites is the A compelling investigate deepwater possibil novel information. It re ity of encountering completely archaeological an whether carried bulk fromWalker mains open question specialized long-haul ships 3. Data compiled across routes. We know from ancient texts 1848. cargoes ancient open-water THE 2005 CHIOS ANCIENT SHIPWRECK SURVEY 2J1

enormous across that large grain carriers moved quantities of foodstuffs we to theMediterranean during the Roman period, though have yet dis cover a wrecked grain ship.4 These vessels are somewhat analogous to are to modern oil supertankers and container ships, which designed ply ocean routeswith minimal time spent in port or near shore.The wrecks of ancient long-distance vessels may best be sought in their natural deepwater environment. Future investigations may demonstrate that the near-shore and offshore seafloors contain wrecks of divergent vessel types carrying different cargoes. access to To gain shipwrecks in deep water, archaeologists occasion ocean use or ally have partnered with scientists and engineers to an ROV human occupied vehicle (HOV).5 These projects have demonstrated the accu strengthsof submersible vehicle operations for archaeology: the speed, racy, and repeatability with which mapping and photographic surveys can a be accomplished, and capability for robotic manipulation and recovery of artifacts. on reasons. Reliance these systems, however, is problematic for several costs of are to The ROV and HOV operations high due the capital and rate operational expenses of the technology and the day for the dedicated oceanographic research vessels necessary to support the systems. ROVs a cable connection to an require the surface ship and expensive and heavy to on level-winding winch carry the cable the ship.When using ROVs in must a deep water, the support vessel be equipped with dynamic po a sitioning system, system that uses computers to control and coordinate the a ship's thrusters and propellers. Position fixes from global position are fed to ing system (GPS) the computer, which then directs the ship to a on hover within few meters of its intended position the sea surface, movement preventing unintended vessel from pulling the ROV off-site.6 are Dynamic positioning ships very expensive to charter. American re search cost vessels with dynamic positioning approximately $40,000 per day, while commercial dynamic positioning vessels can cost as much as $100,000 per day.The number of ROVs and HOVs suitable for archaeo use logical is low, limiting the number of projects that can be undertaken. Because of these factors and others, archaeologists rarely employ these technologies. robotic to some Free-swimming AUVs offer solutions of these prob lems and new open opportunities for archaeological exploration in both and shallow waters. no deep Because AUVs have tether, they do not require dedicated can dynamic positioning vessels. Nearly any vessel support these vessel charter costs down to operations, bringing $10,000-$20,000 per day, or even less.We next foresee, within the few years, deployment ofAUVs from the need for the shore, entirely eliminating support ship.While long AUVs waters for range navigate deep large-area and site-specific survey, smallAUVs can over perform daily mapping and photography sites at diver 4. For a discussion of the size of from depth, freeing archaeologists these time-consuming functions. AUV Roman vessels, with references to technology ismaturing, and the number of vehicles available worldwide is ancient texts, see Houston 1988. steadily As the use of we will 5. Ballard et al. 2000,2002; increasing. archaeologists adopt AUVs, gain access to more McCann andOleson 2004. substantially submerged sites.The project described here is 6. one of the firstto AUVs for Mindell and Bingham 2001, employ archaeological purposes, and itpoints 555. to a direction p. promising for archaeology under water. BRENDAN P. FOLEY ET AL. 272

~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^HHt~| Figure1. Overview detail ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^M the * ^^^^^^^^^^^^^K^^K- ^ ^5^^^ j mentionedinthe text. Original photo m (STS078-732-53) courtesythe Image ^^^^HHJH^V ^^^K^^^Sj^^^^L V fl^^^^^^^^^HScience andAnalysis Laboratory, Johnson Aegean Center,NASA; B.P. X^^^^^^^Bk J^HHH Space adaptedby Foley

k?vHISiHUHi Chios 1ki^^^^H

METHODS AND TECHNOLOGY

Overview of the Survey

As noted above, Chios wreck A is located in the strait between Chios and a near Oinousses, roughly kilometer from the shore the village of Langada a (Fig. 1)7 The wreck is embedded in flat, silty seafloor in approximately 70 m of water. The team selected thiswreck forAUV survey because it is archaeologically significant, its depth exceeds conventional scuba div and the flat seafloor is a environment forAUV robotic ing limits, benign 7. Near is Langada the Classical operations. farm siteof Delphinion: Garnett and a new Survey byAUV is approach to archaeological site investigation. Boardman 1961, p. 106;Yalouris 1986, To test the technology and develop methods for its use, the team budgeted pp. 157,163,166. the 2005 chios ancient shipwreck survey 273

b.c. eight days of ship time to complete the survey of the 4th-century wreck, to overcome deeming that period sufficient any equipment failures. Fortu were to nately,weather conditions ideal and theAUV performed expectations. on We completed operations the Chios wreck A sitewithin three calendar days, and allAUV operations within 24 hours.With the remaining ship time, sonar we conducted additional investigations in the Chios Strait: side-scan sonar and sub-bottom profiling surveys, and video examination (with the a HCMR ROV) of geological features and modern-era shipwreck site. After survey of theClassical-period wreck and three days of operations in the Chios Strait, the team moved to thewestern side of Chios, near the a village of Lithi, to investigate reports of Roman wreck (Chios wreck B) was to to a.d. that thought date the 4th century We deployed the ROV, located thewreck site at depths of36-42 m, and conducted AUV and diver at on on operations the site (Fig. 1). Based the style of amphoras observed Chios wreck B, the site dates not to the 4th century a.d., but to the late 2nd/early 1st century b.c.

The Autonomous Underwater Vehicle and Onboard Sensors

measures m The AUV deployed for the Chios project less than 2 long and 2 m high,and is lightweight(approximately 200 kg).This allows it tobe a or deployed from wide variety of vessels, including small coastal craft boats a fishing (Fig. 2). The robot's flotation and pressure housing contain ing computers are mounted in an upper hull, while its batteries and other heavy components aremounted in a lower hull. The two hulls are connected are by struts,upon which mounted two fore-and-aft thrusters.The lower hull contains a vertical thruster.This double-body arrangement separates the center of from center so buoyancy the of gravity, the robot is passively stable in and roll. pitch Combined with precise control ofmultiple thrust ers, passive stability grants theAUV capability for extremely slow-motion operation, thereby allowing dense data collection during surveys.8 sensors were on Three types of board the AUV during the 2005 sensors survey: navigation for real-time positioning and guidance, optical sonar sensors and formapping the seafloor and its features, and in-situ chemical sensors for quantifying the oceanographic environment.9 A downward-facing digital camera was mounted forward in the lower hull of the and its was robot, single synchronized incandescent strobe light posi tioned aft in the lower hull. This arrangement optimized camera-to-light separation, reducing optical backscatter in the digital images. A multibeam sonar was mapping mounted just aft of the camera, and the DVL dead sonarwas reckoning navigation and altimetry fixed in the rear of the lower

8. et al. and sonar sensors were a Singh 2004, pp. 289,294 payload tivity-depth-temperature(CDT) 295. Cooke 12-bit Chelsea Corporation Pixelfly sensor; Technologies Group - 9. sensors included Tele 1.2 Navigation megapixel single-chipcolor digital Aquatracka aromatic hydrocarbon Benthos 6000 Series baseline camera with and dyne long synchronized strobe, fluorometer; Seapoint Sensors chloro a RD an Delta T multi and a (LBL) transponders,Teledyne Imagenex (260 kHz) phyll fluorometer; Seapoint Sen Instruments 1200 kHz veloc beam sonar. The in-situ chem sors Doppler imaging chromophoric dissolved organic an ical sensors a itylog (DVL), and IXSEA Octans payload consisted of matter (CDOM) fluorometer. fiber-optic gyrocompass. The optical Sea-Bird Electronics SBE49 conduc BRENDAN P. FOLEY ET AL. 274

2. The SeaBED autonomous hull.10 Chemical sensors mounted within the lower hull measured salin Figure underwater vehicle, the dissolved matter operated by ity, temperature, chlorophyll, chromophoric organic of sensors were Deep Submergence Laboratory (CDOM), and aromatic hydrocarbons. Because all of these theWoods Hole Oceanographic into a incorporated single, passively stable, precisely navigated platform, Institution. In four missions con the data sets could be overlaid, thus in resulting coregistered enhancing ducted within a 24-hour period, the and of thewreck and its environment. terpretation understanding AUV mapped multiple parameters of Chios wreck A. Photo M. Grund

Precision Autonomous Navigation for Archaeology

Precision navigation makes possible the coordination of observations from sensors multiple into accurately overlaid maps, transforming purely obser vational exploration into systematic scientific investigation.11The requisite on sensors positioning precision of the underwater vehicle varies depending and data products. The navigation systems functional requirements for are as archaeological site investigations offChios follows:

1.The AUV positioning must be sufficiently accurate to locate thewreck site on the seafloor at the beginning of each survey, i.e., the absolute real-time horizontal positioning must be accurate towithin 3 m.12 This was achieved through LBL

10.The DVL takes advantage of the the radiusof a circledefining the 50% provides a 95% confidencethat the to estimate over confidence in the AUVs will be centered within 6.3 m Doppler effect velocity position. survey a see our accu This method the seafloor. For full explanation, We define desired position of the target location. on our assumes there is zero Gordon 1996. racy based the footprint of implicitly that et initial AUV the in our This 11. Singh, Whitcomb, al. 2000, missions, expected bias positioning uncertainty. extent and pp. 144-145. spatial of the wreck site, assumption is justifiedby the fact that actual value of the desire to ensure that each mission LBL acoustic is a relative 12.We quantify the positioning on covers the location. absolute GPS reference. thisdesign goal based the notion completely target of circularerror probability (CEP), The chosen parameterof 3 m CEP THE 2005 CHIOS ANCIENT SHIPWRECK SURVEY 275

acoustic positioning, capable of absolute repeatability of approx imately 1 m.13 to 2. The AUV must have sufficientlyprecise guidance survey the sonar sensor wreck sitewith guaranteed optical and overlap, i.e., must the real-time guidance be capable of positioning adjacent tracklines with a relative uncertainty of less than 0.35 m.14 The Chios survey used precise dead-reckoning, combining DVL measurements and Octans fiber-optic gyroscope heading ref x m erence to estimate position. For the 30 45 fine-resolution we wreck survey, consider the relative position uncertainly as a between the start and end points of adjacent tracklines worst-case scenario. The estimated standard deviation for the

real-time positioning of the start and end points of adjacent tracklines is 0.26 m.15 This translates to an 80% confidence m interval for satisfying the 0.35 requirement. 3. For AUVs to be useful for archaeology, the precision and accuracy of the final data product must be comparable to the state of the art in underwater archaeology, i.e., the post-processed accuracy of the localization solution should be within 0.10 m in three

dimensions.

to Compared the direct surveymethod, the professional standard for mapping underwater archaeological sites, theAUV navigation results are uses acceptable. The direct surveymethod fiberglass tape measurements to of multiple distances various datum points. These measurements are to combined via three-dimensional trilateration estimate position.16 A thorough consideration of the uncertainty in this standard technique shows that positions can be determined to within 0.043 m (95% confidence), but human mistakes can lead to gross errors in 20% of themeasurements (outliers).17 The Chios survey utilized an off-line combination of LBL absolute positioning with DVL dead-reckoning. The final navigational a data product from the Chios surveys has positioning uncertainly of m approximately 0.145 with 95% confidence,18 higher than the best-case direct surveymethod results.19 Because both the real-time and off-line (post-processing) navigation measurements on rely the combination of redundant and complementary sensing modalities, the AUV survey execution is robust with respect to measurement. outliers in The vehicle autonomously niters the data in real time to remove spurious observations. The post-processing algorithms also remove measurement errors. over any We believe this increase in uncertainty

13.Whitcomb et al. 442. merit standard deviation is 3 and metric of 0.3% 2000, p. mm/s, distance traveled is only 14. This is based on the Octans standard requirement heading accuracy applicable because the heading refer a 1.85 m across-track deviation is 0.1? secant latitude. Dead ence to image footprint uncertainty tends dominate the and a 1.5 m trackline See the is time on spacing. reckoning uncertainty depen position uncertainty these scales. fine-resolution described below. dent and nonlinear. survey The uncertainty is 16.Rule 1989. 15.We estimate the to dead-reckoning constrained less than 0.3% of dis 17.Holt 2003, p. 251. a measure tance traveled a on m position uncertainty using (using simple random 18. Based 0.074 standard ment model for the combination of walk model integrating survey velocity deviation. DVL andOctans of 0.2 m/s at 38? velocity heading north latitude for 19. Bingham 2003. reference. The measure DVL velocity distances greater than 75 m). This BRENDAN P. FOLEY ET 276 AL.

3 LBL trans the direct surveymethod is acceptable for four reasons: the site remained Figure (opposite), (a) in relation towreck undisturbed by the survey, allowing for future repeated investigations; the ponder layout site, and (b) ofwreck within resulting positioning is absolute and referenced to GPS coordinates; the position over a area a LBL network showing precision positioning data range larger spatial and at greater depth than navigated tracklines ofAUV based is feasible the direct surveymethod; and the realistically using navigational on comparative position data from precision and accuracy were sufficient to generate useful sitemaps. LBL, DVL, and LBL/DVL combi nation. B. S. Bingham Shipwreck Surveys off Chios

b.c. The survey of the 4th-century Chios wreck A consisted of fourAUV a missions. Repeatable absolute positioning within stable reference frame a common un provided coordinate space among missions. As the teams on derstanding of the site improved, effortsfocused increasingly finer-scale to new awareness surveys generate and knowledge of the site. For example, sur bathymetry measurements from the first survey informed subsequent a veys, allowing for gradual increase in the resolution of the investigation. on Chemical and optical data collected in later surveys could be overlaid was early bathymetric maps because the positioning consistent in each of themissions. team During the first investigation phase, the deployed the HCMR on in and SuperAchilles ROV theGPS coordinates recorded 2004 relocated the wreck. With the site s location reestablished, the team moored LBL on near a transponders the seafloor thewreck in geometry that optimized were the accuracy of the long baseline solution. Once these transponders team deployed, the surveyed their actual locations by repeatedly inter rogating the range and bearing of the transponders from the surface ship while recording the ships GPS position. The resulting GPS locations of on a each transponder the seafloor established stable, global coordinate frame for all subsequent observations. Figure 3 shows the layout of the meters an LBL transponders and the AUV positions in measured from established latitude and longitude origin. This local coordinate frame is aMercator projection relative to this fixed point, and it allows for easier an GPS coordinate navigational data processing while providing absolute reference.20

After locating the sitewith theROV and placing seafloor transponders a from the ship, the team initiated the second survey phase: large-area reconnaissance intended to document the wrecks environmental con coarse text.During this investigation, the AUV collected photographic, over an area of 50 x 100 m. The bathymetric, and chemical observations was mis duration of thisAUV dive nearly three hours. This initialAUV sion established the location of thewreck within the local LBL coordinate an an check of theAUV's system and provided opportunity for empirical the team reviewed camera settings and lighting.After recovering theAUV, the collected the dive and refined the camera settings and images during measure 20. We survey positions survey in for the second mission. plan preparation based on theWorld Geodetic System The third and final consisted of three fine-resolution survey phase 1984 (WGS-84), using the standard in order to multi coordinates of AUV missions produce comprehensive digital imaging, GPS geodetic latitude, beam sonar, and chemical maps of thewreck and the seafloor immediately longitude, and height. THE CHIOS ANCIENT SHIPWRECK SURVEY 2005 277

LBL /DVLNavigation 25501-1-1-1-1-1-1-1

2500 " Transponder - 2450 Locations

- 2400

- - ^-^ 2350

2300- ^ - 2250

- 2200-

- 2150

- 2100

a 2050'-'-'-'-'-'-1-1 21002050 2150 2200 2250 2300 2350 24502400 X(m)

LBL /DVL iNavigation i (Closeup) 22601-1-1-1- ?-1 jDrjr* I-lbl/dvl! 1f*l/fiDno * lbl Only

b 22201-1-1-1-1-1 2180 2170 2190 2200 22102220 2230 X(m) BRENDAN P. FOLEY ET AL. 278

surrounding it.The relatively small scale of the site and its artifacts dic tated positioning precision sufficient to resolve fine features (i.e., features < a 0.10 m).To produce photomosaics, this resolution is primarily function camera of the optical parameters, the resolution of individual images, and cameras most the distance from the seafloor. For forms of bathymetric and chemical observations, the resolution of the finalmap is dependent upon sensor parameters, survey parameters, and positioning accuracy. team To obtain fine-resolution data, the again programmed theAUV over an to swim in grid patterns thewreck at altitude of 2.5 m, collecting as over at overlapping data the vehicle transited the site four different orien was x tations.The total area of coverage during the survey 30 45 m, centered on over was m or thewreck site.The AUVs speed ground 0.20 per second, 0.39 knots. The AUV s camera collected images every three seconds, syn camera chronized with its strobe light.At an altitude of 2.5 m, the footprint was m m on the seafloor approximately 1.50 along-track by 1.85 across-track. over This altitude, image collection rate, and speed ground resulted in ap in were proximately 60% overlap along-track successive images. Tracklines m a in spaced 1.5 apart, theoretically providing 20% digital image overlap sonar adjacent tracks.The multibeam collected data continuously throughout m more themission, with an average swathwidth of 5 providing than 100% sensors measured overlap between adjacent tracks.Onboard environmental water temperature, salinity, aromatic hydrocarbons, concentrations of dis solved organic matter, and chlorophyll levels.

TABLE 1. DURATION OF AUV MISSIONS

AUV Mission Duration (hr:min)

1 ca. 03:00 2 02:42 3 02:19 4 01:23

re These successive surveymissions, of varying duration (Table 1), thewreck and sulted inmore than 7,000 high-resolution digital images of the surrounding seafloor.After color correcting and histogram equalizing raw team of the wreck digital images, the assembled photomosaic strips site.21Partial mosaics of thewreck were in the hands of the archaeologists team within hours of data collection. At the same time, the engineering used the information derived from theAUVs multibeam sonar and navigation of thewreck site.With sensors to generate preliminary bathymetric maps these data, the surveywas complete, and the archaeological interpretation of the Chios wreck A site and its features commenced. a sonar and ROV in After conducting few days of side-scan surveys sonar in the Chios the team shifted spection of targets Strait, operations to thewestern side of Chios to the but undocumented et al. investigate reported 21. Singh, Adams, 2000; noted we the Roman-era shipwreck, Chios wreck B. As above, deployed Pizarro and Singh 2003; Singh,How at 36 to 42 m. This site and Pizarro 2004. ROV and located the site depths ranging from land, THE CHIOS ANCIENT SHIPWRECK SURVEY 2005 279

10cm ^^^^^^^H^^T I "^V'J

an / amphora, with detail of incuse ^^^^^F J> on its recovered divers stamp rim, by ^^^m^K$--J from theHellenistic wreck B site off V 1 Lithi, Chios. PhotoP. Vezirtis;drawing hi E. P. Oberlander

consisted of a scatter of approximately 40 amphoras, apparently all of the same a type, lying very close to shore at the foot of steep, rocky slope?a difficult environment for robotic survey and ship operations. Landslides had disturbed the site, breaking amphoras and partially obscuring them were under rocks and sediment. The visible artifacts heavily encrusted with marine growth and posidonia grass. was Because this site shallow enough to achieve DVL bottom lock from was a was the surfacewhile theAUV also receiving GPS position fix, it recon not necessary to deploy the LBL transponders. The teams divers a noitered the site and recovered single Dressel 1C (Will Type 5) amphora an on a with incuse stamp the rim (Fig. 4). This amphora type indicates much earlier,Hellenistic, date for the wreck than expected, placing it in the late 2nd/early 1st century B.C.22We then deployed the AUV. Dur a two-hour the vehicle followed the contours 22. Peacock andWilliams 1986, ing mission, depth along to the underwater The AUV collected the same p. 91; McCann et al. 1987, pp. 201 headings parallel slope. as more 203; Sciallano and Sibella 1991, p. 34. types of data during previous missions, including than 2,000 280 brendan p. foley et al.

were images of the debris field and surrounding seafloor.These images later assembledinto While therobot the photomosaicstrips (Fig. 5). surveyed ^^^fl^^^^^Bi diverscollected video of theAUV and scatter.After site, footage amphora ^^^B^^b^^^^L the we discoveredsea water intrusionin someof the recovering vehicle, ^^^Ei^E^^^H AUV As thiswas the last of our allotted we components. day ship time, ^HI^^HI^^BIb concludedAUV operationsoffChios. ^^BB^BflflH^Bf

Digital ImageMapping E^BBshPIBp^ While GPS allows surfaceand air vehicles to tracktheir global posi- HBiBBramP^^ tion towithin a few GPS radio do not under meters, signals penetrate BKSB^Bn^flH water.23 methods forunderwater Therefore,typical navigationusually rely ^BS^HHSBBraB beacon-based networkssuch as baseline to offer upon navigation long ^B^BBBB&Bb bounded error measurements.Use of thismethod comes at an position I^^BB^^BBIB because it the and calibration expense,however, requires predeployment j^BBfi^^^Bf^B of thebeacon networkas describedabove. dead- Conversely,although ^^^^^BBB^fii such as DVL do not reckoning-basednavigation technologies require ^^^^SBpBSmS** a their decreaseswith the predeployedinfrastructure, position accuracy ^^SBP^PHBHfl^ distancetraveled. ^H^flH^^^^pPi^ To combatthese limitations infrastructure-basedand navigational (i.e., flBBBBBBBE unboundederror teammembers at WHOI havebeen growth), developing ^^IBBBBBBh a camera-based This uses thevehicle-collected navigationsystem. system ^^IBBBB^B^ of the seafloor to extractmeasurements of vehicle motion for fu- imagery ^^HB^BhBBBi sionwith theonboard data inorder to a bounded- dead-reckoning produce I^^^BHBraSBB error measurement.24In theAUV buildsa of navigation essence, digitalmap ^InBBPllBSB theseafloor by registeringoverlapping digital-still images (both along-track Bfe^^PSjB^^mm and across-track thatare imagery).Images successfullyregistered produce ^BBBBiShI^B a relativemeasurement of both thevehicles attitude and (heading,pitch, ^^^BB^B^^B and translational When fusedwith theonboard roll) (x,y, z) displacement. ^^^2SlS3^^Sf data fromthe theresult is a whose navigational DVL, navigationsystem Bt!BBHBP^!3B error is commensurate or much better than but which is long baseline, jBj^^SPPWlWM1' ! freeof externalinfrastructure such as LBL The transponders. significant i^EBSH^^liB? of this is thatthe AUV can bemore eas- advantage navigationparadigm ^IBBt^^^BBw for to sites ilydeployed exploratorysurveys investigatetarget shipwreck ^j^BMSP^BEjiii^^ withouthaving to investsignificant ship time to deploy an acousticbeacon networkfor In these allow m^BBHBNF^iF precisionnavigation. layman'sterms, algorithms ^^B^B&BBlIk theAUV to much likea humandoes, with navigate bynavigating visually J^^HMW|jH^^P& to the seafloorenvironment. respect ^KKtB3^l^Bm An and useful of this important by-product navigation methodology Mm^^^HHh is thatthe canbe used to constructan overlappingregistered imagery opti- ^BBH^H^^fcll^K This can thenbe used toconstruct a callygenerated bathymetric map. map ^^^BHB^^^^W accuratethree-dimensional quantitatively photomosaicby back-projecting ^^^BflBI^^^^^V the over the 6 the re- imagery opticalbathymetric map. Figure displays ^^^Bj^^Bj^^^^^M suitof this to theChios wreckA site.In applying technology particular, ^^HE^^^^^^v 6:a and 6:b showthe derived fora 15 x Figures optically bathymetricmap S^^BBI^^^^^r 45 m swathcentered thewreck site. 6:a shows the raw overtop Figure KB^EKB^^^^^^M

23. and 24. and Kinsey, Eustice, Whitcomb, Eustice, Pizarro, Singh ^^^^^MPF forthcoming. 2004; Eustice et al. 2006. THE 2005 CHIOS ANCIENT SHIPWRECK SURVEY 281

-69.2

2185 East[m]

North [m]

a I

1-69

-69.2

-69.4

-69.6 _

-69.8 ? & ? -70

-70.2

-70.4

-70.6 2185 East[m] North[m] b

C

Figure 5 (opposite).Photomosaic Figure 6 (above). Chios wreck A site: (c) quantitatively accurate three Raw stripof Hellenistic Chios wreck B (a) three-dimensional triangu dimensional photomosaic obtained near Lithi. B. P. Foley and Deep Submer lated point cloud of optically derived by back-projecting digital imagery Woods Hole Oceano gence Laboratory, bathymetry; (b) three-dimensional onto thegridded surface shown in (b). graphic Institution R. M. Eustice bathymetricmap gridded at 5 cm; 282 BRENDAN P. FOLEY ET AL.

a three-dimensional triangulated point cloud, while Figure 6:b displays cm. a bathymetric map gridded at 5 Figure 6:c displays quantitatively accurate three-dimensional photomosaic obtained by back-projecting the imagery onto the gridded surface. It should be emphasized that this result measure is fully automatic and metrically quantitative. In other words, can ments of object size and geometric relationships be derived. While current this technology is still very much in the active research stage, its and future value for in-situ, rapid, quantitative documentation of marine archaeological sites cannot be overstated.

Acoustic Mapping

a Multibeam sonar systems collect bathymetric data in fan-shaped swath that iswide in the across-track direction and narrow in the along-track sonar are sets direction. These systems capable of providing dense data of can to bathymetric soundings that be used quantify the fine-scale charac as as teristics of objects on the seafloor, well the seafloor itself Bathymet are sonar ricmaps generated by merging these data with high-precision data.25 navigational a There are number of instrument-specific variables that affect the a sonar resolution of multibeam system, including sound frequency, beam as pattern of the sonar dictated by the transducer design, seafloor roughness, a and the range (distance) to the bottom. The swath width of multibeam system is a function of the angular sector of the transducer and the distance on can from the seafloor. The size of the acoustic footprint the seafloor as a acoustic greatly affect the resolution of the finalmap, large footprint not the details of the over fine-scale complex seafloor terrainwill resolve seafloor,but will reveal broader bathymetric patterns.The acoustic footprint a the becomes largerwith increasing distance from the seafloor.As result, near center and increases acoustic footprint is smaller the of the swath acoustic increases toward the edges of the swath. Similarly, the footprint with increased vehicle altitude. Additional variables that affect the resolution of the finalmap are on For the dependent data-acquisition protocols. example, spatial density on vehicle and of bathymetric soundings is dependent ping rate, speed, is vehicle altitude.While along-track data density primarily dependent data is on on survey speed and ping rate, across-track density dependent characteristics of the multibeam system (e.g., beam width) and distance from the seafloor. were Multibeam sonar data collected during the Chios survey gridded is sufficient to reveal the detailed at 5 cm resolution (Fig. 7). This resolution There is no characteristics of thewreckage and the surrounding seafloor. a wreck A. The wreck itself is sign of debris trail around bathymetrically even in the initial sonar individual complex; but plots, amphoras spatially or from the could be identified. isolated (horizontally vertically) wreckage of the sonar individual artifacts With substantial post-processing data, can within the amphora mound be discerned (Fig. 7, inset).

25. See Roman and Singh 2007. ,21852200 2205221022152220

2245M^^^^^^^^B^^^^|^^^^^^HEj*'U696 2250-^^^^^^^^^^^^^^^^^^^HHuSBr^;^I East 2255r_j^h^^^^^^hl/^^ I'^^^^^^^^^^^I^^^^^^^^^^^H^f ..^2235

FigureChiosMultibeambathymetric7.TheAthesonarbottomofdepictsfirstimagethewrecksite.phasemapsofand wrecksurveyits .^ritftiflfl^^^^^^^^^^^HHI2230'-f**^n'' 1-1-1

Theinsetimageissurroundings.thedatapost-processedfinaltheincollectedwreckphases.oversurvey FerriniV.L.andC.Roman 284 BRENDAN P. FOLEY ET AL.

wreck I?? B^^^^^^^H I-? wreckB^H^^^^^^H ^^^recl^^^^^^^^H |?? wreckA^^^^^^^^J

30 30 ^^B^^^^^H ^^^^^^^^^H

40 40^^^^^^^^^^^^^^^R

50~^^^^^^^^^^^^^^^H 50

60^^^^^^^^^^^^^^^H 60^^^^^^^^^^^^^^^^^H Figure8. Salinity and temperature pim^^l^HlH ~H^^^^^^^HHI ^atapl?ts tne 39 39.05 39.1 39.15 39.2 16 17 18 19 20 21 22 23 24 25 environments of the two wreck sites Salinity (parts per thousand) Temperature (centigrade) offChios. R. Camilli

ENVIRONMENTAL CONTEXT AND IMPLICATIONS FOR PRESERVATION

During theChios AUV surveys, in-situ chemical parameters were measured at a both shipwreck sites using suite of onboard sensors. The chemical pa rameters included salinity, temperature, chlorophyll, CDOM, and aromatic are hydrocarbons. These parameters useful in determining the thermal as and kinetic energy inputs and mixing of thewater, well as the levels of biological and anthropogenic activity at the sites.Although the two Chios are wreck sites of roughly comparable age, lie in close proximity, and both in sit relatively shallow water, the data show that they have significantly different patterns of environmental degradation. Salinity and temperature measurements at the older Chios wreck a A site indicate continuous gradient of decreasing temperature and de as creasing salinity functions of depth. These parameters exhibit absolute m waters homogeneity in the overlying 2.5 altitude of the site, showing no = discernible temperature variability (resolvable precision 0.01?C) and no = variation in salinity (resolvable precision 0.01%o) (Fig. 8). Along with the absence of any detectable benthic currents, this suggests that thewaters at to to this depth have had sufficient time mix equilibrium. Although the Bora and Meltemi seasonal winds routinely reach extremely high veloci in ties this region, modeling of the bathymetric shape and fetch limit of this section of the Chios Strait predict that storm energy imparted into water the column would be dissipated before reaching this wreck site. THE CHIOS ANCIENT 2005 SHIPWRECK SURVEY 285

aromatic CDOM I hydrocarbons^^^k

2220Vv^^^^^^^^^^^^Bl 222o\ ^^^^^^^^^^^^^^^BL l^^^^^^^Hpi ' \ ^^^^^^^^^^^^HP^ 220o\ l^^HPP^ ^<^ 2255

2240 ^ \ 2225^ 2180\V"^ 2225^

^^^^^^^jjjj^^^k 22*^S^\^^^^ \ ^^^^^^^^^^^^^H^ 2200' ^B|^^ ^^ 2255 2200 2255 V \ ^^^^^^^ \ \ ^ 2240 A* ^C.n^ V 2240\^ % 2225^ 2180\^<^ a0<^ 2180\^^ 2225?

Figure 9. Chios wreck A survey, AUV dive 3: chemical area a parameters Optical and acoustic multibeam imagery of the reveals large contigu measured over the B.C. ous an area 4th-century amphora pile generally resembling ellipse. The surrounding R. wreck site. Camilli Chios wreck A lacks the obvious furrowmarks in the sediment charac otter on no teristicallymade by the doors trawl fishing nets, and there is a scour crater indication of around the perimeter of the amphora mound. storm Overall, the data suggest that thewreck site has been undisturbed by or events, benthic currents, trawl fishing since its sinking. to The Chios wreck A site also appears support limited photosyntheti cally derived biological activity.Optical imagery shows very littlebiological encrustation ofwreck artifacts. In contrast, imagery from theChios wreck B site clearly shows widespread encrustation, with sizable beds of posidonia grass around the artifacts.However, chlorophyll concentrations measured throughout the overlying water column at both siteswere approximately 2.5 mg/1, with little variability. occurs near A step change in aromatic hydrocarbon concentrations a the seafloor around Chios wreck A, increasing by factor of three in the m water com lowest 10 of the column (Fig. 9). The seafloor is generally m posed of silty sediments exceeding 15 in thickness.26At thewreck site et 26. Sakellariou al. 2007, p. 371. itself, aromatic hydrocarbon concentrations decrease. A plume of low 286 BRENDAN P. FOLEY ET AL.

a source over 10 Photomosaic of concentration advects from point the amphora mound, Figure (opposite). a Chios wreck with features fanning out westward along straight line from the mound. This lower A, major identified. B. P. and Submer concentration plume is also found in CDOM values at the site. Objects Foley Deep gence Laboratory, Woods Hole Oceano on this ancient wreck structural habitat marine life; the provide attracting graphic Institution hard ceramic material acts as a holdfast substrate for sponges and other invertebrates.The lower aromatic hydrocarbon and CDOM plume maybe waters caused by advecting past benthic filter feeders (i.e., sponges). In contrast, the Chios wreck B site averaged a sixfold decrease inwater a column aromatic hydrocarbons and 16% decrease inCDOM concentra tions relative to the Chios wreck A site.An obvious explanation for the characteristic differences in aromatic hydrocarbon and CDOM concentra com tions is that the seafloor underlying the Chios wreck B site is largely no prised of rockymaterial instead of siltyorganic sediments. Furthermore, anomalous depth-dependent aromatic and CDOM concentration changes were observed at the Chios wreck B site, nor was there any indication of localized variability in concentrations. Further investigation of the environmental states at these sites is needed before definite conclusions can be reached as to the transformativemecha

nisms affecting them. Preliminary in-situ data suggest thatClassical Chios wreck A may be better preserved than nearby Hellenistic Chios wreck B because of sediment chemistry, and because the basins physical shape to restricts kinetic energy inputs.The basin bathymetry appears decouple Chios wreck A fromwind and wave energy at the surface.The photographic a and chemical data suggest that Chios wreck A may existwithin hypoxic are at two basin. Although the chlorophyll levels similar the sites, the concentra evident lack of photo synthetic organisms and enriched organic tions within the sediments could enable a sustained microbial drawdown in dissolved oxygen. This sustained decrease in oxygen would greatly limit bio biological activity/encrustation of thewreck site, thereby attenuating at the site. logical degradation of organic materials and artifacts

ARCHAEOLOGICAL EVIDENCE

B.C. Multibeam sonar data show that the 4th-century Chios wreck A site an on a of 21 awidth of forms ellipse the seafloor,with maximum length m, 8 m, and a vertical relief of 1.4 m. The cargo remains are tightly contained within the dimensions of the wreck site; there is no debris trail around thewreck. The visible cargo remains consist of more than 350 amphoras to a of two types.The firsttype is attributed Chios, with generally oblong with a neck, vertical handles with attachment shape high straight long high 27.Many of thebroken Chian conical carinated and base with a conical have fractured at the points, body, shoulder, spiked cup amphoras cleanly is unknown. This has shoulder or at the base of the neck. toe.The origin of the second amphora type amphora and a computermodeling of that a relatively short neck, short vertical handles, globular body, spiked Dynamic conducted atWHOI confirms the base in a knob toe. style terminating observation that these are the weakest of in 10 represents a composite image assembled from hundreds to Figure points of the shape.When subjected dividual an otherwise view of thewreck or digital pictures, providing impossible external pressure shock, the Chian Chian 94 are and 73 are fractures the seams at and its cargo. Of the observable amphoras, intact, amphora along the or the base of the neck. broken.27 In the area containing theChian amphoras, another 101 amphoras shoulder Anchor? Modern trash Modern trash Possible Thasian amphora toe Possible Thasian amphora toe Jug BE2005/4, within area of decreased amphora density Findspots of amj: iensity Findspots of amphoras BE2004/4.1 and BE2004/4.4 004/4.1 and BE2004/4.4 the chios ancient shipwreck survey 2005 287

Figure1.View of full-sized ^^^^^^^^^^^^^^^^^^^^^^^BPss^^^^^^H smallerChian amphoras onChios ^^^^^^^^^^^^^^^^^^^^^^HfeiSl^^^^^^l wreckA,shown insitu, viewed from ^^^^^^^^^^^^^^^^^^^^^^^Hb^II^^^^^H southeast.Sponges are growing on ^^^^^^^^^^^^^^^^^^^^^^^^^^Hb&S^S^^M bothamphoras, ^^^^^^^^^^^^^^^^^^^^^^^^^^^HHjBjj^^^H canbe seen center ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^|Hn^H Photo teamHiil^^lflliHili^li^liHHiiliHHIHHHH

are visible. They appear to be Chian, aswell, but are too covered in sediment to are or determine if they intact broken. The digital images show that the one Chian amphoras have similar dimensions, with exception (Fig. 11, lower right) that is somewhat smaller than the predominant type. In 2004, HCMR and EUA scientists recovered from the site a well preserved Chian amphora, designated BE2004/4.1 (Fig. 12). The vessel is intact except for one missing handle, but traces of that handle s attachments on remain the neck and shoulder. The clay color is red (Munsell 2.5YR are 5/6), and the amphoras dimensions overall height 0.915, maximum recess diameter 0.340, handle height 0.320, toe height 0.070, depth of in m. toe 0.040, and neck inner diameter 0.087 EUA archaeologists measured ac the volume of this amphora by filling the jar with water, according to cepted practice.28 The volume to the base of the neck is 19.0 1, and the total volume to the rim is 22.01. an va The second amphora type found atwreck A is of unattributed are riety.There 30 intact examples of this type discernible in the imagery, at are an least 12 that broken, and many fragments representing unknown one number of additional amphoras. In 2004, HCMR and EUA recovered amphora of this type, designated BE2004/4.4 (Fig. 13). The clay color is are light red (Munsell 2.5YR 6/6), and the amphora dimensions overall height 0.665, maximum diameter 0.400, handle height 0.145, neck height m. was 0.140, and neck inner diameter 0.085 The capacity of this amphora also measured by filling itwith water.29 The capacity at the base of the neck is 32.41, and the total internal volume measured at the rim is 33 1. more am The digital images of Chios wreck A show mouths of 37 in it to 28. Docter 1990, p. 149. phoras visible the sediments, but is impossible determine their type 29. Docter 149. or 1990, p. condition. They appear to be standing upright, and theymay still be in 288 BRENDAN P. FOLEY ET AL.

i BE2004/4.1recovered from Chios ^Jjr

Oberlander

?W 13.Unattributed ^C^^Jr Figure amphora \ 1 BE2004/4.4recovered from Chios W^Wfl I wreckA HCMR andEUA in iocm I by 1 1 2004. PhotoP. Vezirtis; drawing E. P. WJ0r mW^k I Oberlander THE 2005 CHIOS ANCIENT SHIPWRECK SURVEY 289

Figure anchorwestern H^^^^^^^^^^^I^^^^^^UHh^^^H ofChios wreck A, alongside ^^^^H^^HH^^^^^^^H^^BfiH^^^H amphoras.Photo Chios team HHHHiHHfliH^HHHHHHHHjH

their original loading positions, possibly indicating that additional layers of amphoras maybe buried in the sediments below the observable artifacts. a B.C. at By comparison, 5th-century wreck excavated Alonnesos contained m four tiers of amphoras occupying 1.5 of vertical relief.30 The extreme western end of Chios wreck A contains a feature that is an arm suggestive of anchor shank and (Fig. 14). This feature is less than m 2 long, however, and may be too small for ground tackle.31 a area The photomosaic (Fig. 10) reveals small (about 1.5 m2) of near western decreased amphora density along the centerline the end of the wreck. This might represent the location of the ships mast. Nearby, one or the toe of amphora of perhaps Thasian northern Aegean type is area an evident (Fig. 15).32This also contains assemblage of several small or a pots possibly used by the crew, perhaps representing secondary cargo consignment in addition to the amphora cargo. 30.Hadjidaki 1996, p. 574. A of a BE2005/4) was recovered from 31. Casson 1995, pp. 254-256. single example jug (designated Chios wreckA theHCMR ROV in2005 BE2005/4 is 32.Monakhov and Rogov 1990, using (Fig. 16). a or p. 144. plain-ware jug missing the rim, part of the neck, the handle handles, 29O BRENDAN P. FOLEY ET AL.

^^^^^^^^^^^^^^Kmmj^^^^^^^^^^^^^^^^^^^^^^^^m FigureToe an ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^H amphora on Photo ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^H^^H^^^^^HIHHHHI^^^^^^HHHHHI^I^^^^H^^^^^^^^^HHIHHHI team

Figure16. Small jug BE2005/4 ^^T^T recovered fromChios wreck A in Itt^S^^ i- _ _ _ _ _w) 2005byHCMRSuperAchillesROV. EEliiESSIS PhotoP. Vezirtis; drawing E. P.Oberlander

a a and part of the body. It has compressed spherical/ovoid body and ring base. The clay is reddish brown to dark reddish gray (Munsell 5YR 5/3), are and the dimensions of the jug overall height 0.175, diameter 0.170, m. was a and base diameter 0.095 The inner surface coated with thin layer not a of earthlike material, which has yet been analyzed. This jug is plain a one wide-neck typewith slightly angular shoulder; the base of handle is preserved. A close parallel of the vessel had been found earlier in theChios can harbor.33The Chios wreck A jug be dated within the 4th century B.C. a and placed stylistically between typeA jug of the Ephesos Tetragonos a Agora (firsthalf of 4th century B.C.)34 and series of jugs from Rhodes (last quarter of 4th century B.C.).35

was re are max. max. 33.This unpublished jug icalMuseum (inv.4063) and later in sions p.H. 0.20, Diam. trievedin 1985 fromthe southernpart 1985 was moved to the Conservation 0.18, base Diam. 0.09 m. was re of theChios harbor along the insideof Laboratory of EUA, where it 34. EphesosXIII.2, pp. 89-90, the southernmole during an illegal numbered as 85/34.The neck and rim fig. 101, the firstvessel to the left;see was are no. dredgingoperation that halted by broken and thehandle ismissing, also p. 328, and pi. 14, 95. EUA. The jug initiallywas placed in but the jug has the same shape as the 35. Giannikouri, Patsiada, and Phili the collectionof theChios Archaeolog jug fromChios wreckA. Its dimen monos 2000, p. 75, pi. 23:0, no. 118181. THE 2005 CHIOS ANCIENT SHIPWRECK SURVEY 2QI

TABLE 2. SHIPWRECKS ROUGHLY CONTEMPORARY WITH CHIOS WRECK A

B.C. Location ofShipwreck Est. Date SiteLength (m) SiteWidth (m) Amphora Cargo

western 4200 12 Tektas Burnu, Turkey 440-425 Alonessos, Sporades, Greece 410 25 14 1,000 (top layer); 4,000 total? Porticello,Messina, Sicily 390 looted16? El Sec,Mallorca, Spain 375 similarto Kyrenia similarto Kyrenia total475, but looted Chioswreck 350-330A 218 350 (top layer) Kyrenia,northern Cyprus 295-285* 4.5total 14 400+ 9 600+11 Serce Limani, southwestern Turkey 280 total

was as to ca. see *The Kyrenia wreck originally published dating 325 B.C.; Katzev 1972. et Sources: Carlson 2003, 2004; Hadjidaki 1996; Lawall 1998a; Arribas al. 1987; Susan Katzev, pers. comm.; Pulak andTownsend 1987. are not in was not or a Note: Two other Classical wrecks included this table, either because the cargo primarily composed of amphoras was not an complete survey performed. The Ma'agan Mikhael wreck (ca. 410-390 B.C.) off Israel carried building materials rather than see amphora cargo; Linder, Kahanov, and Black 2003. The Phagrou wreck inGreece (ca. 450 B.C.) carried amphoras, but awaits survey.

In addition to the objects carried on the vessel when it sank, thewreck has trapped modern trash, including plastic bags, cups, and water bottles (see Figs. 10 and 11, above). This material has been carried onto the site currents?a on by benthic phenomenon long observed other deepwater across shipwreck sites.Ancient debris also travels the seafloor; anachronistic a artifacts encountered in the supposedly closed context of wreck sitemay be explained by thismechanism.36 In to comparison other roughly contemporary shipwrecks, Chios wreck A is larger than all but the Alonnesos vessel (Table 2). The number of on was a amphoras visible wreck As top layer is evidence that it relatively large ship for its time, but further investigation is necessary to determine the total number of amphoras contained within the 1.4 m tumulus. As on are described above, theChian amphoras thewreck site straight a toe necked with conical cup (see Fig. 12). Intact examples and sherds B.C. of similar jars have been excavated from 4th-century contexts in the northern Black Sea. Chian toes a stratum "dunce cap" appear in dating to the middle of the 4th century B.C. from the Belozerskoe settlement near the mouth of the Dnieper River, Ukraine.37 Complete examples were excavated from child burials in the necropolis tumulus at Panskoe I westernmost in Crimea, dating to the third quarter of the 4th century B.C.38 to Closer Chios, sherds of similar cup-toed amphoras have been studied at Ilion. were no They excavated from strata dating later than 330 B.C.39 The Chian amphoras contained within Chios wreck A conform closely to the mid-century form. An exact parallel has also been found in the Athenian Agora: Chian amphora P 25947.40This amphora was excavated

36. Parker 309. 1981, p. Investiga 38.Monakhov andRogov 1990, tions of deepwater sites reveal that the p. 138.Hind (1992-1993, p. 92) re seafloor ismore deep dynamic than portsChian amphoras importedinto Parker and ancient and mod B.C. suspected, 5th- and 4th-century Nikonian ern artifacts move distances and in may greater Ordessos, modern-day Ukraine. thanpreviously imagined; see Foley and 39. Lawall 2002, pp. 202,204 Ballard 2004. 205. 37. Bylkova 1996,2005, pp. 221, 40. Grace 1979, fig.46. 236; Lawall 2005, pp. 43, 45. 292 brendan p. foley et al.

cen from fill in deposit F 17:3 that is dated to the second half of the 4th seems most tury b.c.41 Taking all of this into account, it likely that the no ship sank sometime between 350 and 330 b.c., and certainly later than 315 b.c.

Earlier Chian amphoras differed significantly in style and capacity from to themid-4th-century form.42From the end of the 6th century themiddle a of the 5th century b.c, Chian amphoras featured distinctive bulbous seven neck, and their internal volume averaged Chian choes. In the later 5th century, the capacity of thisChian amphora stylemight have increased to seven new to eight Chian choes, equivalent Attic choes. Then the form of the of straight-necked Chian amphoras appeared. Production bulging the new ca. 425 neck stylemight have overlapped with style until b.c., was when the bulging-neck amphora fully superseded by the straight a necked style.43The straight-necked style had capacity of eight Chian choes, equivalent to seven Attic choes.44 The initial change in capacity and subsequent change in style reflected in the bulbous-to-straight neck transition might have resulted from theAthenian Standards Decree, under to the which the Chian producers conformed theAttic standard. Among as reason in scholars who point to this decree the for the change measure, a there is disagreement about the date of its issuance. Some scholars posit more date of448-446 b.c.,45 but recently,others have convincingly argued for a date around 424 b.c.46 can a Political changes be reflected in the objects produced by society, and the history of Chian coinage provides good examples of this phe nomenon. Chios minted coins on its own standard from 525 b.c. until the Athenian Standards Decree, when a break occurred in coin production. In 412 b.c. the Chians revolted against the Athenian League, and they on own between resumed minting coins based their standard.47 Relations a a new Athens and Chios warmed for period with Second Athenian tensions soon built with League alliance in 379/8,48 but again, erupting the Social War of 358/7-355/4 b.c. At this point, Athens attacked Chios an insurrection of the sizable slave on and possibly fomented population in the island.Ultimately, Persian intervention resulted Chian independence the fromAthens.49 Throughout the politically tumultuous period following 412 b.c. revolt,Chian coins conformed to the island s own standard.50 measures the trends evident Changes inChian amphora might amplify from to in the islands coinage standards. The transition bulging-neck

these to thoseof Chios 47. Hardwick 1993, 216-218, 41.Agora XII, p. 390. See also ing amphoras pp. stated wreck A. 220. Agora XXIX, p. 451, where the from 42. Lawall 78. 48. Hornblower 2002, 233. provenance of the amphoras 1998b, p. p. a 43. Grace Lawall 81. 49. Peake Ruzicka deposit F 17:3 (thewall of well rilled 1979; 1998b,p. 1997; 1998, a terminus ante 44. Wallace Hornblower 2002, 175,256. after 315 B.C.) provides Mattingly 1981, p. 80; p. 60; pp. this The 88-89. See also Koehler 50.Ashton et al. 2002, 216. The quern for amphora. shapes 1986, pp. p. 284-285. of several other mints and dimensions of the Agora amphora 1982,pp. output regional Chios 45. Barron followed theChian standard in the and amphoraBE2004/4.1 from Mattingly 1981, p. 80; are am 98. 4th B.C., Kni wreck A similar. The Agora 1986, pp. 96, early century including are overall H. 46. Hardwick Mat dos, Kos, Samos, Thasos, phora's dimensions 0.947, 1993, p. 216; Kyzikos, and Rhodes. See max. Diam. 0.332, handle H. 0.342 m. tingly1993, p. 102; Schonhammer Ephesos, Miletos, Koehler and Susan Lawall Ashton et al. 2002, pp. 242-243. We thank Carolyn 1993, pp. 189-191; 1998b, 94-95. Rotroff for their assistance in compar pp. THE 2005 CHIOS ANCIENT SHIPWRECK SURVEY 293

straight-neck amphoras around the time of theAthenian Standards Decree an has been linked to the shift to Attic standard measure, presumably in the 420s. To cut tieswith the defeated Athenians after the 412 B.C. revolt, own theChians might have changed amphora measures back to their stan seven new dard of Chian choes51 while maintaining the straight-necked two amphora style.From survivingChian wine measures, Forrest measured a as the volume of Chian chous 2.881.52 Seven Chian choes therefore equal sowe see 20.16 1, would expect to amphora capacities in this range. Some evidence for this capacity change exists inChian amphoras excavated in a context mid-4th-century in the lowerDon River region. Brashinskii notes that the capacities of the Don River jars range from 18.6 to 20.2 l.53As Lawall points out, definite conclusions about this possible capacity change must a wait until larger sample size of Chian amphoras dating before and can after theAthenian Standards Decree be investigated.54 Amphora BE2004/4.1 (Fig. 12) recovered from Chios wreck A adds one more to data point this debate. As noted above, the volume to the base of the neck is 19.01, and the total volume to the rim is 22.01. Docter were tentatively suggests amphoras filled halfway up the neck of the jar.55 were If thisChian jar filled halfway up itsneck, the resulting volume would be or 2% of seven 20.5 1, within Chian choes. However, it is statistically to unsound base conclusions about the entire class of Chian amphoras on a a from thiswreck single sample.56To verify capacity standard for this of Chian measurements style amphora, additional will be required from on the amphoras this shipwreck and other contemporary sites. were Chian products traded extensively in theClassical and Hellenistic strata eras.57Excavations of 5th-century inAthens and Corinth show large of quantities Chian amphoras, and the islands amphoras have been found in the southernMediterranean atNaukratis in theNile Delta,58 and farther on eastward, Cyprus.59 Evidence from excavations in the northern Black Sea area markets for indicates Chian products in that region during the 4th and was a century B.C., papyri record that Chios port of call for ships between trading Alexandria and the Black Sea during the 3rd century B.C.60 Chian have Additionally, amphoras been recovered from several shipwrecks; as Carlson this indicate were notes, may thatChian products widely traded, or it may simply reflect the fact that the distinctive Chian amphoras are

51. Boardman in sea 1958-1959, pp. 306, covered the off Ashkelon, Israel, 2000; Garlan 2002. The best synthesis 309. had a of 20.5 1.No further of Chian capacity trade in the 5th century, 52. Forrest 64. One of information is available based on a 1956, p. regarding study of amphoras, is Lawall these Chian wine measures was on these finds. See Zemer 1977, pp. 37 1998b. at theChios 39. 58. display Archaeological Archontidou-Argyri and Kokki Museum at the time we were 54. Lawall 7. preparing 2000, p. noforou 2004, p. 93. this see Archontidou 55.Docter 155. manuscript; 1990, p. 59. A wreck with a cargo of similar and 56. An discussion of Chian was Argyri Kyriakopoulou 2000, interesting amphoras located and brief 160-161. in pp. problems determining amphora ly studied in2007 off the coast of 53. Brashinskii and volume standards southeast 1980, pp. 16-17; capacities ap Cyprus (Stella Demesticha, see alsoWhitbread 36-37. inMatheson and Wallace 1995, pp. pears 1982, pers. comm.) See also Archontidou a Chian am n. 21. Additionally, cap-toed esp. p. 300, Argyri and Kokkinoforou2004, pp. 92 excavated in the harbor at 57. phora Athlit, See Empereur and Garlan 1987, 95. had a of 17.2 a Israel, capacity 1, and 1992,1997; Drougou, Zervoudaki, and 60. Sarikakis 1986, pp. 121-123. later Chian re pointed-toe amphora Touratsoglou 1998; Marangou-Lerat 294 BRENDAN P. FOLEY ET AL.

more easily recognized and therefore frequently reported than other

types.61 as Chian amphoras excavated frommost sites have been interpreted a wine containers, although few (possibly reused) examples recently studied were ca. B.C. by archaeologists found to carry other products.62 From 500 an a until the Roman period, amphora accompanied by bunch of grapes was a on as an persistent feature Chian coinage. This has been interpreted advertisement for the islands wine.63 Ancient writers across the centuries

noted the quality of Chian vintages. Strabo declared that the best Greek was wine produced inAriusia, the mountainous northwestern region of the island (Strabo 14.1.35); and Pliny noted that Caesar provided Chian wine at his triumphal banquets (HN 14.16.97). Theopompos credited the Chians with the invention of "dark" or "black" wine, as differentiated from or white yellow wine.64 In the Deipnosophistae, Athenaeus quoted other was ancient authors' claims that Chian wine best of all, again singling out texts theAriusian variety (Ath. 1.26b, 1.29e, and esp. 1.32f).The ancient indicate that the distinctive Chian wine found a wide and eager market throughout theGreek world for centuries. was not an However, amphora BE2004/4.1 lined with resin, indica tion that it did not hold wine. To determine its we probably contents, 61. Carlson 2004, pp. 142-143. a from the wall of the 98-99. A extracted small (< 1 g) sample of ceramic interior jar 62. Barron 1986, pp. 5th to an tech B.C. Chian recovered and subjected it molecular biological analysis, unprecedented century amphora ancient DNA from the sea off Knidos contained hun nique fully described elsewhere.65 These analyses revealed dreds of olive However, it is not of olive and This result is it demonstrates that this pits. oregano.66 intriguing; known if this the use represents original of did not contain wine in its first for or reuse type amphora exclusively export. contents the of the jar (Cemal of this was of olive carried a Some portion ships cargo composed products Pulak, pers. comm.). Similarly, single must and Chian olive in Chian amphoras. Additional samples be collected analyzed amphora containing pits more on broader conclusions about was recovered from the wreck near from amphoras this shipwreck before Asian Burnu; see Kerameikos IX, the cargo can be drawn. pp. 23-24; and Carlson 2004, p. 139. The of the second recovered from Chios wreck A, capacity amphora 63. Hardwickl993,p.211. is 32.41 at the base of the neck unattributed amphora BE2004/4.4, (11.2 64. Boardman 1967, p. 252;Whit at Chian choes of 2.88 1each).The total internal volume measured the rim bread 1995, p. 144. seems to a 65. Hansson and 2008. The is 33 1 (11.4 Chian choes). It possible that this jar conformed Foley of this lies exactmeasure of great potential technique differentvolume standard, perhaps theAttic standard.The in itsability to detect the residueof a standard Attic chous is unknown.67 Lang and Crosby's estimate of3.2761 ancient substances in empty amphoras. is often rounded down to 3.20 or 3.26 l.68 Ifwe generally accepted, accept 66. Hansson and Foley 2008. as from the standard Attic chous 3.26 1, then the unattributed amphora 67.The Attic kotyle isbelieved to Chios wreck A measures within ? 1% of 10 Attic choes if filled halfway be 0.2711 and 12 times that volume about the 3.262 1, equaling an up the neck. As with the Chian amphora, broader conclusions equals perhaps Attic chous. on Chios wreck A cannot be made amphoras of this unattributed type 68.Agora X, pp. 47,58; Forrest without a size. larger sample 1956, p. 64; Koehler andWallace 1987, The bottom of BE2004/4.4 contained a small of amphora lump p. 55. a of the resin in 69.We theNational hardened shiny brownish resin.We collected sample May acknowledge Sci Science Foundations of 2006 and submitted it for carbon-14 analysis to theNational Ocean support NOSAMS underNSF ences Accelerator Mass (NOSAMS) atWoods Hole Cooperative Spectrometry Facility no. OCE-9807266. Institution.69 The radiocarbon date of the resin is 2400 Agreement Oceanographic 70. The calibration curve for the ? 30 calibrated to a calendar date of 410 B.C. ?30 The fiat and years b.p., years. Classical period is relatively is earlier than for radiocarbon calibrated calendar date somewhat expected, considering therefore problematic on The calibrated of artifacts from this era. the estimated date of the Chian amphoras thewreck.70 analysis THE 2005 CHIOS ANCIENT SHIPWRECK SURVEY 295

calendar date range should be tempered with the knowledge that wine was sometimes aged for years in amphoras. In addition, the resin could some have been collected and stored for time before being used to preserve a thewine, assuming that the jar did contain wine.71We acknowledge that a can not radiocarbon date from single sample be suggestive, but conclusive, evidence for the date of the vessels sinking. un The origin of the amphora type represented by BE2004/4.4 is a near known, but itmight have been location Chios. There has been some on speculation that the type originates Chios itself72The unattributed am phora fromChios wreck A shares general similarities with amphoras from the Tektas Burnu wreck, located only 20 nautical miles away. The cor respondence between those amphoras, however, is not exact.Whereas the overall body shape is similar, substantial differences exist in the rim. A toe can seen range of shapes be among theTektas Burnu amphoras. None precisely match the toe of the unattributed amphora from Chios wreck A, although the bottom half of the rounded teacup, double-beveled toe of some Tektas Burnu amphora lot 0911 does bear stylistic similarity to the toe ofBE2004/4.4.73

The variability between the features of the two amphora typesmay be to due the passage of three-quarters of a century between the two wrecks, or are perhaps the amphoras simply from different production centers.The were Tektas Burnu amphoras initially termed "pseudo-Samian" and then as reconsidered and reclassified likely being Klazomenian or west Asian. as a Recently, scholars studying thatwreck have proposed Erythrai possible 71. Barron 1986, p. 98. ofmanufacture We to 72.Monakhov 1999, p. 368. place (Fig. I).74 encourage amphora specialists perform additional research into the unattributed 73. Carlson 2004, p. 175, fig.47. amphoras from Chios wreck A 74. Bass Carlson to 2002, p. 96; 2003, in order establish developmental links to other types and to determine p. 586, 5, bottom. See fig. Ozyigit the origin of the type.Recovery of additional examples of the unattributed 1990 for from amphoras Erythrai, esp. typemay be necessary before firm conclusions can be drawn. p. 145. Chios and Erythrai maintained Molecular of a ceramic from the interiorwall economic, political, and occasional biological analysis sample of BE2004/4.4 an of Chios or military relationsin the5th and 4th amphora may support origin nearby. Our centuries b.c. followed Chios's that Erythrai results show the jar contains ancient DNA from the Pistacia genus, revolt Athens in 412 stone against b.c; which includes pistachio nut and mastic.75 Given the presence of resin in from Erythrai appears in Classical we the jar, suspect that theDNA represent mastic. wrote that Chios and the Persian might Pliny constructions; was Chios the best-known source of gum mastic in the an satrap Mausolos is credited both with high-quality cientworld. He differentiated Chian "whitemastic" from othermastic-like refoundingErythrai in themid-4th on substances for its century and minting its coins the color, consistency, and method of collection (HN 12.36). Chian standard. Mausolos also entered This the of whether was a opens question Chian mastic distinct species into a defensive alliance with Chios and a within general family ofmastic-like substances described in ancient texts. other east Greek states in the 350s. The Fresh ofmodern samples Chian mastic have been collected by the research mixed cargo of ceramic material from and are Chios on team, additional genes from that and Erythraian amphoras the chloroplast being sequenced spe cies and others within thePistacia Successive DNA are now Tektas. Burnu wreck is another sugges genus. analyses tion of links between the two centers. in the being performed hope of further refining the identification of the 75. Hansson and 2008. material in Foley amphora BE2004/4.4. If this jar did contain Chian mastic, it 76. An unresolved is the problem a local for the contents of suggests origin the unattributed amphora, and apparent difference in stan capacity therefore the itself.76 dards between the perhaps jar Chian amphora and the unattributed The and of thewreck show the dis amphora from bathymetric maps photomosaic Chios wreckA. We tribution of the other hope that future amphoras and artifacts (Figs. 7 and 10). The Chian research will to this fact. form the of the visible help explain amphoras majority cargo remains, and they are P. ET 296 BRENDAN FOLEY AL.

distributed along three-quarters of the length of the wreck beginning at a thewestern end, with few scattered along the remaining quarter of the at eastern are site the end. The unattributed amphoras located exclusively in the eastern end of thewreck. Since Chian amphoras are located among and perhaps underneath amphoras of the unattributed type, itmay be that were the Chian amphoras loaded first, or at the same time as the other a amphora variety.This may be clue to the origin of this other amphora a or type,providing circumstantial evidence that it could be Chian style, else produced somewhere nearby. Another possibility is that the unattributed were to on to amphoras shipped Chios another vessel, only be re-exported at by the shipwrecked offLangada. If so, theymight have been loaded the same as on time the jars originating the island. Alternatively, it is possible now that the Chian amphoras in the eastern end of thewreck arrived in eastern their current positions during the wrecking event. If the end of the vessel hit the seafloor first, these Chian amphoras might have tumbled area through the ship into the containing the unattributed amphoras. This account area might also for themany crushed jars in that of thewreck; the a effectwould have been similar to bowling ball knocking down tenpins. Intrusive investigation of the wreck may be necessary before these ques tions can be answered. can on reasons We speculate the for the vessels loss from the informa sonar area tion collected during the survey.The survey of the surrounding thewreck site and the high-resolution photographic survey of thewreck a itself both indicate a distinct absence of debris trail.Apparently there was no attempt on the part of the crew to throw cargo overboard during no fire the ships lastminutes. The digital images show evidence of among any of the artifacts observed on the seafloor, nor is any blackening evident on the three artifacts raised in 2004 and 2005. Perhaps thewrecking event m are occurred suddenly.The islands 1,100 elevations located within 12 km of the wreck site. The terrain funnels violent downdrafts and gusts of a wind across thewaters surface.The wreck is oriented at heading of300?, winds at have been described as pointing toward the shore. The Langada helmsmen to "fluky and gusty," and theywould have required ships' pay careful attention while transiting along the coast.77A square-rigged Clas on a the helmsman sical vessel sailing beam reach would heel, requiring a to increase the rudder angle to leeward. If strong gust caught the sails, the helmsman could lose control of the vessel as it swung into and maybe in sail taken through thewind. This would result the square being aback, to In worst this could cause the causing the vessel heel sharply. the case, vessel to take on water over the rail and sink.78 77. Hunt 1946, p. 49. The exact of and route of the vessel to are place lading prior sinking 78. Roberts 1995, p. 311. B.C. unknown, as is the site of 4th-century amphora production. Analysis 79. Jones 1986, pp. 283,287-288. it in the beds 80. See Yalouris 1986. of the clay of Chian amphoras suggests might originate near in the southeastern of the island 81. One in the area of Letsaina in Emporio and Armolia, portion (see the modern town of Chios and a sec have conducted extensive surveys and excavations Fig. I).79Archaeologists ond one in the area of Limnia on the on the Kofina to the north and west ofmodern Chios Town without ridge northwest side of the island, near the of kilns. Chios holds scores of sites finding evidence archaeological dating harbor of Vblissos; see Archontidou two kilns or from theNeolithic through the Byzantine periods,80 but only Argyri and Kyriakopoulou 2000, sites have been located.81 It has been pp. 158-159. amphora-production long suspected SURVEY THE 2005 CHIOS ANCIENT SHIPWRECK 297

thatClassical kilns may underlie and be obscured by modern Chios Town, and in 1986 archaeologists discovered evidence of amphora production were within the town.82 It ispossible that the amphoras fromChios wreck A loaded aboard the ship at Chios Town, located 15 km south of thewreck. on Locating additional kiln sites may shed light the conveyance of and the island. agricultural products and transport containers around from were to It is not known ifempty amphoras transported thewine- and olive areas or or were across land to producing by boat cart, if the liquids carried on near points the shore for loading into amphoras the ships. to are The closest villages theClassical wreck site Delphinion and Lan a gada, about 2 km distant. During 1954 underwater survey,archaeologists noted the presence of "heaps ofChian amphoras dating from about 400 B.C." m in underwater gullies 20 deep off the islet ofTaurus, just offshore from team Delphinion.83 More recently, the EUA located Chian amphora sherds dated earlier than Chios wreck A in the area offTaurus. Remains of harbor can on near installations and scattered potsherds be found themain island Taurus, along with the remains of theTemple ofApollo Delphinios. The as B.C. ac Athenians used Delphinion their naval base in 411-406 in their was as an or a tions against the Chians.84 It routinely used anchorage place B.C. of lading in the late 5th century and throughout the 4th century

CONCLUSIONS AND FUTURE DIRECTIONS

We show here that precise, accurate, high-quality nondestructive surveys of can a shipwrecks at any depth be delivered within single day. Our methods and technologies permit the detailed quantification of the physical dimen sions, environmental setting, contents, and condition of wreck sites.The on a data collected the 70-m-deep Chios wreck A show site dating to the second half of the 4th century B.C. thatmeasures 21 x 8 x 1.4 m.The wreck's cargo is composed of approximately 350 amphoras in the surface layer,and probably contains additional layers of amphoras under the sediment. The are two an amphoras of types,Chian and unattributed type.One example was of each recovered by EUA and HCMR in 2004. Molecular biological analysis of ceramic samples from the jars indicates the presence of ancient DNA of olive and oregano in the Chian amphora, and Pistacia genus, pos in siblymastic, the unattributed amphora.85 In addition to the amphoras, five small are on a jugs evident the site, perhaps representing secondary or a crews ware. cargo consignment part of the galley One example of was season. these small jugs recovered during the 2005 The wreck site and its artifacts show no evidence of fire, and there is no debris trail near the wreck. The vessel sank to in an area close shore known for unpredictable wind All evidence a gusts. suggests sudden sinking event. a amount Although fair is known about the political and economic 82.Anderson 1954, p. 123; Jones of Chios in the history 5th century B.C., archaeological and textual evi 1986, p. 284;Tsaravopoulos 1986. dence is scant for Chios in the second half of the 4th B.C. The 83. Garnett and Boardman 1961, century information derived from Chios wreck A to fill that Based on p. 106. begins gap. 84. Roisman the visual a 1987, p. 29. evidence collected in 2005, thewreck carries cargo dominated 85. See Hansson and 2008. two Foley by Chian amphoras. Only amphora types are evident on thewreck, 298 BRENDAN P. FOLEY ET AL.

we unseen though acknowledge that cargo may differ in composition a from the observable layer. In any case, the ship carried cargo outbound from Chios. Given the wrecks proximity to the island of origin and the on assume large number of amphoras found it, it is logical to that this was were the first time these Chian amphoras used. This amphora type typically is associated with thewine trade; however, evidence of original our new amphora contents delivered by technique to detect ancient some DNA proves that portion of the cargo carried in these amphoras was not composed of olive products, wine. We therefore question the assumptions made in the past regarding the contents of unlined Chian are amphoras. Classical-era Chian amphoras quite distinctive and there fore commonly identified in terrestrial and underwater sites.We plan to conduct additional DNA studies ofChian amphoras fromChios wreck A a can se and other sites. If access to significant number of these jars be to trace cured, itmay be possible Chian agricultural export production a through long time span. a was During the 2005 expedition, second wreck also partially surveyed. was to not The site reported to the EUA prior the 2005 project, but had a to been investigated before our survey.Chios wreck B is late-2nd- early near lst-century B.C.wreck located off thewest coast of Chios Lithi. It is a to marked by scatter of approximately 40 amphoras that all appear be of theDressel 1C type (WillType 5). The amphoras lie along the bottom of a a m. steep, rocky slope at depth of 36-42 The site has been disturbed by wave inmarine rock slides and possibly action, and it is heavily encrusted a was and it bears growth. One example of Dressel 1C amphora collected, an on wreck was cut incuse stamp its rim (Fig. 4). The 2005 survey of this to continue to short by damage theAUV, but EUA archaeologists study this site. to autono The 2005 Chios project points the futurepotential of rapid, a mous archaeological survey regardless of depth, but therewill always be water With autono place for archaeological divers when depth permits. mous robotic methods complementing standard shallow-water practices, divers may be freed from mundane and time-consuming mapping tasks, to on fine allowing them concentrate instead excavation, manipulation, focus on and archaeological interpretation. Divers should archaeological cannot means of robotic and questions that be answered by technologies methods. For instance, our procedures for retrieving ancient DNA from the to the surface for In shallow amphoras entail bringing jars sampling. We envision DNA of waters, this is best performed by humans. analysis contents a common tool for original amphora becoming archaeologists. will be able Through the study of ancient DNA remains, archaeologists use of to draw more precise conclusions about the amphoras, agricultural ancient trade. production, and patterns of that in-situ methods offer The 2005 project demonstrates survey great cannot answer all of the ask. A promise, but yet questions archaeologists remote and exca combined approach of sensing, artifact recovery, perhaps vation offers the best balance. The in-situ data recorded in 2005 relate the From wrecks environmental context, size, shape, and general composition. THE 2005 CHIOS ANCIENT SHIPWRECK SURVEY 299

can these data archaeologists estimate thewrecks state of preservation, its age, and the origin of its amphora cargo. In contrast, analyses performed on the artifacts recovered in 2004 generate specific conclusions. Detailed narrow examination of artifacts helps to the date of sinking and provides amphora volumetric information. Pioneering molecular biological inves new tigations of the amphoras' contents provide important information about ancient trade goods. Today it is still necessary to combine in-situ and intrusive methods for comprehensive understanding of underwater archaeological sites; in the future thatmay change. are Many of the technologies and methods described here in their more infancyThe next generations of sensitive chemical and environmen sensors are can tal in the final stages of development atWHOI, and they a detect wide range of compounds at very low concentrations. Combined un with precision navigation, chemical maps will provide unprecedented derstanding of shipwrecks' physical environments. Eventually, theymay be able to discern the materials contained within the wrecks. The novel

digital-image mapping techniques outlined here deliver three-dimensional photomosaics and pixel-resolution maps of sites. Soon this technique will be to near new able generate these data products in real time. Likewise, high sensors can frequency acoustic imaging be utilized for archaeology, making to measure on it possible accurately the volume of objects the seafloor. In we to a future projects, hope determine the volumes of statistically signifi on a cant number of amphoras wreck through acoustic and optical digital imaging, with the intent of assessing capacity standards without removing the artifacts from their archaeological contexts. For archaeological investigations in deep water, AUV surveys offer over sonar benefits towed surveys and ROV operations. A fast-running can two as a AUV survey to three times much seafloor in more systematic manner a can than towed system, and it produce better data by main an over taining optimal altitude the terrain during the survey.Multiple can a AUVs be operated from single ship, again increasing the amount of coverage and the number of targets identified. Eventually AUVs will be costs deployed from shore, reducing project by eliminating the support can ship. Because each seafloor target be characterized and surveyed in a an single day with AUV, the number of shipwrecks that archaeologists can examine will increase significantly. Presented with data sets collected in-situ by sensors, archaeologists will be able to prioritize sites deserving intrusive investigations and better allocate the resources necessary for those operations. to us new AUVs applied underwater archaeology may permit to ask a questions by assessing statistically significant number of comparable, sites. Instead of resources on a over contemporary concentrating single site as has been the in soon we many years, practice underwater archaeology, be able to a will survey dozens of sites in single field season. This will allow broad of inter- and study intraregional trade, exchange, and contact through time. Combined with the information gleaned from complete excavations of land and underwater new sites, these technologies and methods will alter our fundamentally understanding of the ancient past. 300 BRENDAN P. FOLEY ET AL.

ACKNOWLEDGMENTS

The Chios 2005 team thanks colleagues Jonathan Adams, Sir John Board man, Elizabeth Greene, Maria Hansson, Ross Holloway, Carolyn Koehler, Mark Lawall, Anna Marguerite McCann, and Aristotle Tympas for shar two ing their knowledge, and the anonymous Hesperia reviewers whose suggestions greatly improved thismanuscript. We thank the two students who participated in the 2005 cruise: Ballard Blair, graduate student in the a WHOI-MIT Joint Program; and Christopher Murphy, then student at now a Franklin W. Olin College of Engineering and graduate student in theWHOI-MIT Joint Program. We gratefully acknowledge the support of theHellenic Ministry of Culture, theHellenic Centre forMarine Re search, theU.S. National Science Foundation, George Chronis and Vangelis crew Papathanassiou ofHCMR, and the captain and of R/V AEGAEO. were The SeaBED AUV and the efforts of its team funded by the Censsis no. ex Engineering Research Center under NSF grant EEC9986821. We our press sincere appreciation to sponsors Susan and Robert Bishop, Jane and James Orr, and others, whose generosity made this project possible. The 2005 Chios AUV Shipwreck Survey included participants from the Hellenic Ministry of Culture, Ephorate of Underwater Antiquities (EUA); theHellenic Centre forMarine Research (HCMR); and theWoods Hole Oceanographic Institution andMassachusetts Institute ofTechnol ogy (WHOI/MIT). EUA team: Katerina Dellaporta, Director; and archaeologists Diony sis Evagelistis, Dimitris Kourkoumelis, Paraskevi Micha, and Theotokis Theodoulou.

HCMR team: Dimitris Sakellariou, geologist, Chief Scientist; The odoros Fotopoulos, engineer, ROV team; Panos Georgiou, geologist; Kostas Katsaros, Thetis pilot, ROV team;Aggelos Mallios, chief diver,ROV pilot, engineer; Prokopis Mantopoulos, engineer, side-scan sonar/sub-bottom profiler; Ioannis Pampidis, engineer, side-scan sonar/sub-bottom profiler; and Spyros Volonakis, head ofHCMR underwater activities. WHOI/MIT team:Brian Bingham, engineer, precision navigation; Bal lard Blair, graduate student; Richard Camilli, engineer, chemical sensing; Ryan Eustice, engineer, AUV operations and precision mapping; Vicki ar Ferrini, engineer, software,mapping, and mosaicking; Brendan Foley, Catherine chaeologist, project leader; David Gallo, program development; conservator and Matthew Giangrande, archaeological project development; Neil vehicle Grund, engineer, acoustic modems; McPhee, engineer, sys tems;David Mindell, engineer, precision navigation; Christopher Murphy, student; Catherine Offinger, logistics and administrative support; James Rakowski, program development; Christopher Roman, engineer, AUV AUV operations and precision mapping; Hanumant Singh, engineer, op administrative erations and underwater imaging; Ann Stone, logistics and support; and David Switzer, archaeologist. THE 2005 CHIOS ANCIENT SHIPWRECK SURVEY 301

REFERENCES

= The Results P. 1986. "Chios in the Athe Agora Athenian Agora: of Barron, J. Excavations Conducted theAmer nian in A by Empire," Chios: Confer at ence at Homereion in ican School of Classical Studies the Chios, 1984, ed. Boardman C. E. Athens, Princeton J. and Vapho X =M. 89 Lang and M. Crosby, poulou-Richardson, Oxford, pp. and 1964. 103. Weights, Measures, Tokens, = L. XII B. A. Sparkes and Tal Bass, G. F. 2002. "The Excavation of Black and Plain a B.C. cott, Pottery of the Fifth-Century Shipwreck: Interna 6th, 5th, and4th Centuries B.C., 1970. 1999 Campaign," in 7th XXIX = I. Hellenis tional on Construc S. Rotroff, Symposium Ship tic Athenian and tion in Pottery: Imported Antiquity:Pylos, 26, 27, 28, Wheelmade Table Ware and Related 29 August 1999. Proceedings(Tropis Material, 1997. 7), ed. H. Tzalas, Athens, pp. 93 Anderson, J. K. 1954. "Excavations on 102. theKofina Ridge," BSA 49, pp. 123 Bingham, B. 2003. "Precision Autono 182. mous Underwater Navigation" (diss. M. Institute Archontidou-Argyri, A., and Kok Massachusetts of Technol Col kinoforou. 2004. Archaeological ogy) lectionat Volissosof Chios (Ministry Boardman, J. 1958-1959. "Excavations at of Culture, 20th Ephorate of Pre Pindakas in Chios," BSA 53-54, and historic Classical Antiquities), pp. 295-310. -. Mytilene. 1967. Excavations in Chios, Archontidou-Argyri, A., andT. Kyria 1952-1955: GreekEmporio (BSA kopoulou, eds. 2000. Xiog r'evaXog Suppl. 6), London. noXiq Oivomcovog (Hellenic Min Brashinskii, I. B. 1980. Grecheskii kera of micheskii na Nizhnem Donu istry Culture, 20th Ephorate of import v V-IIIvv. do n.e. Prehistoric and Classical Antiqui [Greek Ceramic ties), Mytilene. Imports to theLower Don in the Arribas, A., M. G. Trias, D. Cerda, and 5th-3rd Centuries B.C.], Lenin J.De La Hoz. 1987. ElBarco de El grad. Sec Estudio de los V. P. on (Calvia, Mallorca): Bylkova, 1996. "Excavations materiales, Mallorca. the Eastern Boundary of the Chora Ashton, R. H. J.,N. Hardwick, P. Kinns, of Olbia Pontica," EchCl 40, n.s. 15, K. Konuk, and A. R. Meadows. pp. 99-118. -. "The of 2002. "The PixodarusHoard (CH 2005. Chronology inCoin Hoards IX: Settlements in the Lower 9.421)/' Greek Dneiper Hoards 400-100 (Special Publications of Region, B.C.," in Chro theRoyal Numismatic Society 35), nologiesof the Black SeaArea in the ed. A. Meadows and U. B.C. Wartenberg, Periodc. 400-100 (Black Sea London, pp. 159-243. Studies 3), ed. V. F. Stolba and R. Ballard, D., A. M. McCann, L. Hannestad, Aarhus, pp. 217 D. Yoerger, L. Whitcomb, D. Min 247. H. B. dell, J. Oleson, Singh, Foley, Carlson, D. N. 2003. "The Classical D. and C. Gian Greek at J.Adams, Piechota, Shipwreck Tektas^ Burnu, grande. 2000. "The Discovery of Turkey,"AJA 107, pp. 581-600. Ancient in the -. History Deep Sea 2004. "Cargo in Context: Advanced Using Deep Submergence The Morphology, Stamping,and Research a Technology," Deep-Sea 1:47, Origins of theAmphoras from 1591-1620. B.C. Ionian pp. Fifth-Century Ship R. L. E. Ballard, D., Stager, D. Master, wreck" (diss. Univ. of Texas, D. Yoerger, D. Mindell, L. L. Whit Austin). H. and D. Piechota. L. 1995. and comb, Singh, Casson, Ships Seamanship 2002. "IronAge Shipwrecks in in theAncient World, Baltimore. Water off R. F. 1990. Deep Ashkelon, Israel," Docter, "Amphora Capaci AJA 106, pp. 151-168. ties and Archaic Levantine Trade," 302 BRENDAN P. FOLEY ET AL.

in A' Kerameikos IX = Der Sud HBA 15-17 (1988-1990), pp. 143 ocpxouac, P65od," Zwavrrjcrri U. Knigge, 188. yia xr\veXXr\vioxiKr\ Kepap.eiKrj, hiigel{Kerameikos IX), Berlin 1976. Drougou, S., I. Zervoudaki, and I. Tou Icodvviva, AeKejipprfg 1986, 2nd ed., 1998. ed. L. 63 R. and L. Whit ratsoglou. BipXwypayia Marangou, Rhodes, pp. Kinsey, J., Eustice, eXXr\viKr\gKepajuiKijg 1980-1995 I: 85. comb. Forthcoming. "Underwater R. L. 1996. Acoustic Vehicle Recent Ad EXXdg-Kvnpog, Thessaloniki. Gordon, Doppler Navigation: vances and New in Empereur, J.-Y., and Y. Garlan. 1987. CurrentProfiler: Principles of Oper Challenges," ation. A 2nd San "Bulletin archeologique: Amphores Primer, ed., Diego. Proceedingsof the 7th IFAC Confer et V. R. 1979. and the ence on and Control timbresamphoriques (1980 Grace, Amphoras Manoeuvering Marine 1986),"REG 100, pp. 58-109. AncientWine Trade (AgoraPicBk6), of Craft, September 20-22, -. rev. Princeton. Lisbon. 1992. "Bulletin archeologique: ed., 2006, et E. 1996. "Underwater Exca C. G. 1982. on Amphores timbres amphoriques Hadjidaki, Koehler, "Amphoras a (1987-1991)," REG 105, pp. 176 vations of Late Fifth Century Amphoras," Hesperia 51, pp. 284 Merchant at 292. 220. Ship Alonnesos, -. C. and M. B. Wallace. 1997. "Bulletin archeologique: Greece: The 1991-1993 Seasons," Koehler, G., et 1987. The Amphores timbres amphoriques BCH120, pp. 561-593. "Appendix. Transport (1992-1996)," REG 110, pp. 161 Hansson, M. C, and B. P. Foley. 2008. Amphoras: Description and Capac R. F. 209. "Ancient DNA Fragments inside ities," in C. Pulak and Town XIII.2 = P. Sherrer and E. Classical Greek Reveal "The Hellenistic Ephesos Trinkl, Amphoras send, Shipwreck Die in at Serci eds., Tetragonos Agora Ephesos: Cargo of 2400-Year-Old Ship Limani, Turkey: Preliminary von archaischer 1169-1176. 49-57. Grabungsergebnisse wreck,"/^ 35, pp. Report,"AJA 91, pp. bis in Zeit. Ein Uber N. 1993. "The of M. L. 1998a. Men byzantinische Hardwick, Coinage Lawall, "Bolsals, und Funde klassischer to dean and the Date of blick: Befunde Chios fromthe Vim the IVth Amphoras, inActes du XT Con thePorticello Zeit (EphesosXIII.2), Vienna Century b.c.," Shipwreck,"IJNA 27, international de 16-23. 2006. gres numismatique: pp. H. a Voccasion du 150e anni -. 1998b. "Ceramics and Posi Eustice, R. M., O. Pizarro, and Singh. Organise Navi versaire de la Societe ale de Nu tivism Revisited: Greek 2004. "Visually Augmented Roy Transport in an Environ de and in gation Unstructured mismatique Belgique, Bruxelles, Amphoras History," Trade, ment a His 8-13 M. and theAncient Using Delayed State septembre 1991,ed. Hoc, Traders, City, tory,"in Proceedings of the 2004 Louvain-la-Neuve, pp. 211-222. ed. H. Parkins and C.J. Smith, on G. F. 1992-1993. New 75-101. IEEE International Conference Hind, J. "Archaeology York, pp. 26 of the Greeks and Barbarian -. 2000. Wine Robotics and Automation, April Peoples "Graffiti, Selling, New USA the and the Reuse of in May 1, 2004, Orleans, LA, around Black Sea (1982-1992)," Amphoras AR 82-112. the Athenian ca. 430 to 1, Piscataway, N.J., pp. 25-32. 39, pp. Agora, P. 2003. "An Assessment of 400 3-90. Eustice, R. M., H. Singh,J.J. Leonard, Holt, Qual B.C.," Hesperia 69, pp. in -. 2002. "Ilion before Alexander: and M. Walter. 2006. "Visually ity Underwater Archaeological and Economic Archaeol Navigating theRMS Titanic: Con Surveys Using Tape Measures," Amphoras Stadia Troica 197 servative Covariance Estimates for IJNA 32, pp. 246-251. ogy," 12, pp. SLAM Information Filters," Inter Hornblower, S. 2002. The Greek World 244. Robotics Research 479-323 b.c, 3rd London. -. 2005. Chronol nationalJournal of ed., "Negotiating Black 25, pp. 1223-1242. Houston, G. W. 1988. "Ports in Per ogies," in Chronologies of the D. 2004. Some Mate Sea Area in the Period c. 400 Foley, B. P., and R. Ballard. spective: Comparative on and 100 b.c. Sea Studies "AmphoraAlley II," inMcCann rials Roman Merchant Ships (Black 3), ed. V. F. Stolba and L. and Oleson 2004, pp. 183-194. Ports," AJA 92, pp. 553-564. Hannestad, 31-67. Forrest, W. G. 1956. "A Chian Wine Hunt, D. W. S. 1946. "An Archaeo Aarhus, pp. Classical Y. and E. Black, Measure," BSA 51, pp. 63-67. logical Surveyof the Linder, E., Kahanov, of the Island ofChios eds. 2003. The Mikhael Garlan, Y. 2002. "Bulletin archeologi Antiquities Maagan et timbres Carried Out between theMonths The a 2400-Year que: Amphores ampho Ship: Recoveryof BSA Old Merchantman: Final riques (1997-2001)," REG 115, ofMarch and July1938," 41, Report 1, 30-52. Haifa. pp. 149-215. pp. R. E. 1986. Greek and A. 2000. "Le Garnett, R., and J. Boardman. 1961. Jones, Cypriot Marangou-Lerat, plus des vins inLavven "Underwater Reconnaissance off the Pottery:A Review ofScientific Stud agreable grecs," Island ofChios, 1954,"BSA 56, ies,Athens. tura del vino net bacino delMediterra M. L. 1972. "The neo, ed. D. Tomasi and C. Cremo pp. 102-113. Katzev, Kyrenia 71-85. Giannikouri, A., V. Patsiada, and Ship," inA History ofSeafaring Based nesi, Venice, pp. on ed. G. F. P. and M. B. Wallace. M. Philimonos. 2000. "Taquicd Underwater Archaeology, Matheson, M., 1982. "Some Rhodian GvvoXa ano tic, veicpoTio^eic, thc. Bass, London, pp. 50-52. Amphora THE 2005 CHIOS ANCIENT SHIPWRECK SURVEY 303

293 and R. F. Townsend. 1987. Roman. Capacities," Hesperia 51, pp. Pulak, C, C. 2004. "Seabed AUV 320. "TheHellenistic Shipwreck at OffersNew Platform forHigh H. B. 1981. Serce Trans Mattingly, "Coins and Limani, Turkey: Preliminary Resolution Imaging," EOS: and 31-57. actions American Amphoras: Chios, Samos, Report,"4^ 91, pp. of the Geophysical in O. T. P. 1995. "An Thasos the Fifth Century B.C.," Roberts, Explana Union 85, pp. 289-295. JHS 101, pp. 78-86. tionof Ancient Windward Sailing: Singh, H, J.Howland, and O. Pizarro. -. 1993. "New on the Some Other 2004. in Light Considerations," IJNA "Advances Large-Area Athenian Standards Decree 307-315. 24, pp. Photomosaicking Underwater," (ATL 11,D 14),"Klio 75, pp. 99 Roisman, J. 1987. "Kallikratidas: A IEEE Journal ofOceanic Engineering 102. Greek Patriot?" CJ S3, pp. 21-33. 29, pp. 872-886. A. E. K. and H. 2007. "A L. D. McCann, M.J. Bourgeois, Roman, C, Singh. Singh, H, Whitcomb, Yoerger, P. and E. L. Will. Self-Consistent Gazda, J. Oleson, Bathymetric Map and O. Pizarro. 2000. "Microbathy 1987. The Roman Port and Fishery pingAlgorithm," Journal of'Field metric Mapping from Underwater Cosa:A Center Ancient Robotics 26-51. in the of of Trade, 4, pp. Vehicles Deep Ocean," Com Princeton. N. 1989. "The Direct Vision and Understand Rule, Survey puter Image A. and P. eds. 143-161. McCann, M., J. Oleson, Method (DSM) ofUnderwater ing^, pp. 2004. Water and Its A. 1986. "H Deep- Shipwrecksoff Survey Application Tsaravopoulos, ocpxoda SkerkiBank: The 1997 Survey Underwater," IJNA 18, pp. 157 n6Xr\ tnt; Xkro," Horos 4, pp. 124 (JRA Suppl. 58), Portsmouth, 162. 144. R.I. S. 1998. and R. 1848. the Secre Ruzicka, "Epaminondas Walker, J. Report of D. and B. the Genesis of the Social the inAnswer to a Mindell, A., Bingham. War," tary of Treasury 2001. "New Uses of CP 60-69. Resolution the Senate Archaeological 93, pp. of Calling for Autonomous Undersea in P. A. Mal in toWrecks Vehicles," Sakellariou, D., Georgiou, Information Regard of the2001 MTS/IEEE V. D. Kourkoume Vessels to Proceedingsof lios, Kapsimalis, Belonging the United States Oceans P. T. and Conference 1, Piscataway, lis, Micha, Theodoulou, (30th Congress, 2nd Session, Senate 555-558. K. 2007. for Executive Document N.J, pp. Dellaporta. "Searching 3),Washing S. 1.1999. Grecheskie Monakhov, amfory Ancient Shipwrecks in theAegean ton, D.C. v Prichernomore: kera Sea: The of Chios and M. 1986. in Kompleksy Discovery Wallace, B. "Progress Mea micheskoi VII-II vekov do n.e. tary KythnosHellenistic Wrecks with suring Amphora Capacities," in in theBlack Sea the Use ofMarine Recherches sur les [GreekAmphoras Geological amphores grecques Area: Complexes of Ceramic Con GeophysicalMethods," IJNA 36, (BCHSuppl. 13), ed. J.-Y Empe 7th-2nd Centuries reur tainers, B.C.], pp. 365-381. and Y Garlan, Athens, pp. 85 Saratov. Sarikakis,T. C. 1986. "Commercial 94. Monakhov, S. I., and E. I. Rogov. 1990. Relations between Chios and Whitbread, I. K. 1995. Greek Trans Panskoe I Other Greek Cities in A and "Amfory nekropolia Antiquity," port Amphorae: Petrological of thePanskoe I in Chios: A at theHome [Amphoras Conference ArchaeologicalStudy (BritishSchool Mir i reion in ed. Boardman and at Fitch Necropolis],"Antichnye Chios, J. Athens, Laboratory Occa 128-153. C. E. Arkheologiia7, pp. Vaphopoulou-Richardson, sional Paper 4), Athens. 6.1990. "1988Yih 121-131. D. H. Ozyigit, Erythai Oxford, pp. Whitcomb, L., Yoerger, Singh, Kazi M. 1993. Sondaj ?ali?malan," Sonuf/ari Schonhammer, "Some and J.Howland. 2000. "Advances 125-150. on Toplanttsi 11, pp. Thoughts theAthenian Coinage inUnderwater Robot Vehicles for Parker, 1981. "Stratification and inActes du XI6 Ocean A.J. Decree," Congres Deep Exploration: Naviga Contamination inAncient Medi international de numismatique: Orga tion, Control, and Survey Opera terranean Shipwrecks," IJNA 10, nise a Voccasion du 150e anniversaire tions," in Robotics Research: The 309-335. de la Societe de Ninth International pp. Royale Numismatique Symposium, D. P. and D. F.Williams. de 8-13 Peacock, S, Belgique, Bruxelles, septembre ed. J.M. Hollerbach and D. E. 1986. and the Roman ed. M. Amphorae 1991, Hoc, Louvain-la Koditschek, London, pp. 439 An Economy: Introductory Guide, Neuve, pp. 187-191. 448. London. Sciallano, M., and P. Sibella. 1991. Yalouris, E. 1986. "Notes on the Peake, S. 1997. "A Note on the comment les Dating Amphores identifier? Topography ofChios," inChios: of the Social War," GaR, 2nd ser, Aix-en-Provence. A at Conference theHomereion in 161-164. 44, pp. Singh,H.J. Adams, D. Mindell, and Chios, 1984, ed. J. Boardman and Pizarro, O, and H. 2003. B. 2000. Under C. E. Singh. Foley. "Imaging Vaphopoulou-Richardson, "Toward water for Large-Area Mosaicing Archaeology,"//^ 27, Oxford, pp. 141-168. for Underwater Scientific 319-328. A. 1977. in Applica pp. Zemer, Storage Jars Ancient IEEE Oceanic tions," Journal of Singh, H., A. Can, R. Eustice, S. Ler Sea Trade, Haifa. Engineering28, pp. 651-672. ner, N. McPhee, O. Pizarro, and 304 BRENDAN P. FOLEY ET AL.

Brendan P. M. Eustice Foley Ryan

Woods Hole Oceanographic Institution University of Michigan

department of applied ocean physics department of naval architecture

and engineering and marine engineering

deep submergence laboratory department of (ms j) electrical engineering 266 woods hole road and computer science

woods hole, massachusetts o2543-io5o 204 na&me building 2600 draper drive

[email protected] ann michigan arbor, 48109-2145

[email protected] Katerina Dellaporta Hellenic Ministry of Culture klepsydras 2 Dionysis Evagelistis 105 55 athens Hellenic Ministry of Culture greece ephorate of underwater antiquities

kallisperi 30 [email protected] 117 42 athens greece Dimitris Sakellariou

for Hellenic Centre Marine Research VickiLynn Ferrini institute of oceanography Lamont-Doherty Earth Observatory p.o. box 712 of Columbia University 190 13 anavyssos marine geology and geophysics division greece 6l route 9W new york [email protected] palisades, IO964-80OO

[email protected] Brian S. Bingham University of Hawaii at Manoa Kostas Katsaros department of mechanical engineering Hellenic Centre for Marine Research 254o dole street box p.o. 712 holmes hall 304 190 13 anavyssos honolulu, hawaii 96822-2303 greece [email protected] [email protected]

Richard Camilli Dimitris Kourkoumelis Woods Hole Oceanographic Institution Hellenic Ministry of Culture department of applied ocean physics ephorate of underwater antiquities and engineering kallisperi 30 deep submergence laboratory (ms j) athens 117 42 266 woods hole road greece woods 'hole, massachusetts o2543-io5o

[email protected] [email protected] THE 2005 CHIOS ANCIENT SHIPWRECK SURVEY 305

Hanumant AggelosMallios Singh

Hellenic Centre for Marine Research Woods Hole Oceanographic Institution

p.o. box 712 department of applied ocean physics

19O 13 anavyssos and engineering greece deep submergence laboratory (ms j) 266 woods hole road [email protected] woods hole, massachusetts o2543-io5o

Paraskevi Micha [email protected]

Hellenic Ministry of Culture David S. Switzer ephorate of underwater antiquities

kallisperi 30 Plymouth State University

117 42 athens department of social science greece rounds hall

17 high street [email protected] new hampshire plymouth, o3264-1595

David A. Mindell [email protected]

Massachusetts Institute of Technology TheotokisTheodoulou program in science, technology,

and society Hellenic Ministry of Culture

building e-51 ephorate of underwater antiquities

77 massachusetts avenue kallisperi 30 massachusetts cambridge, 02139-43oi 117 42 athens greece [email protected]

[email protected]

ChristopherRoman

University of Rhode Island, Narragansett Bay Campus department of oceanography

south ferry road

rhode island narragansett, o2882-ii58

[email protected]