the archaeology of scattered wreck-sites: formation processes and

The International Journal of Nautical Archaeology (2002) 31.2: 211–227doi:10.1006/ijna.2002.1044

The archaeology of scattered wreck-sites: formation processesand shallow water archaeology in western Lake Huron

John M. O’SheaMuseum of Anthropology, University of Michigan, Ann Arbor, Michigan 48109-1079, USA

This paper explores the application of archaeological site formation theory to stranded wooden vessels and the scatteredwreck-sites they produce. Shallow water wrecks and wreckage sites in the Au Sable Shores region of western Lake Huron areused to develop a preliminary classification of the processes operating on the breakup and deposition of wooden vessels.

� 2002 The Nautical Archaeology Society

Key words: shipwreck, Lake Huron, Great Lakes, site formation, scattered wreck

Introduction

E very wreck is unique. The site typicallyrepresents a single vessel, on a single voy-age from a particular port, that succumbed

to some specific condition or set of conditions atone particular time, which caused it to sink and bedeposited in a one particular place. The sense ofthe unique is enhanced when there are historicalaccounts of the vessel, its fate, and its crew. Thisview of shipwrecks supports a general image ofshipwreck-sites as unique ‘time capsules’ thatprovide an instantaneous, but inherently unique,snapshot of a single time in the past (Bass, 1983).This perspective is equivalent to what has beentermed the ‘Pompeii premise’ (Binford, 1981); aview of the archaeological record as a frozenmoment in time.

It is difficult, however, to apply such a perspec-tive to most shallow water wrecks, particularlywhen the vessel has been broken up and scattered.Indeed, some underwater archaeologists havesuggested that shallow water wreck-sites are notworth investigation, since they have so littlecoherent information content (Dumas, 1972: 32–33; but see Tomalin et al., 2000). Ironically, suchshallow water wrecks have much in common withconventional terrestrial archaeological sites.

Terrestrial sites are increasingly viewed notas a frozen moment in time, but rather as apalimpsest of distinct, short-term events andbehaviour, created under the joint force of inten-

1057–2414/02/020211+17 $35.00/0

tioned cultural activities and natural formationprocesses. In essence, each deposit or feature on atypical terrestrial site could be described in thesame terms as a shipwreck. It is a unique event, ona specific historical trajectory of individual actionand local circumstances, which results in a par-ticular deposition of material remains within aspecific spatial matrix. Yet all this uniquenessdoes not mean that the deposits are unrelated nordoes it mean that they cannot be combinedsystematically with other observations to build upa more dynamic picture of human activity in thepast. It is the role of archaeological theory toguide the archaeologist in understanding thedeposit itself, and to determine how the depositcan appropriately be used to understand eventsand activities in the past.

Underwater research, perhaps more than eventerrestrial archaeology, is characterized by anextensive but implicit and anecdotal understand-ing of the factors that affect sites, such as storms,waves, woodworm, looters, or reef formation.The implicit character of this understandingcauses it to be applied inconsistently and furthercontributes to the illusion of site uniqueness. Theneed to develop a more systematic understand-ing of the archaeological record is not an exercisein theory for theory’s sake. For the underwaterarchaeologist, it is rather the necessary founda-tion on which any historical or anthropologicalunderstanding of the vessel and its wreck must bebased (Fontenoy, 1998).

� 2002 The Nautical Archaeology Society

NAUTICAL ARCHAEOLOGY, 31.2

A ship, as a large, complex and expensiveartefact (even a simple dugout canoe represents amajor investment of time, labour, and materials),shares a number of characteristics with othercomplex, long use-life artefacts, such as housestructures. It can be expected to retain informa-tion relating to its initial construction and place oforigin, and to its life history of use and modifica-tion, as well as the evidence of its final loss orabandonment. Therefore, the processes of vesselconstruction, use, loss, deposition, and recoveryshould be understandable in terms of the body ofarchaeological theory that is subsumed under theheading ‘formation theory’ (Schiffer, 1987). Theissue of formation processes looms particularlylarge when the focus of research shifts from theinvestigation of a single, intact wreck to a regionalstudy, as in the Au Sable Shores project describedbelow.

In essence, formation theory is designed to dealwith two related archaeological problems: (1) howdo materials pass from a systemic context, wherethey are part of an ongoing behavioural system,into a static archaeological context; and (2) whathappens to these material remains and theirspatial interrelationships between the time theyare deposited and the time they are recovered bythe archaeologist.

The first question relates to what can be termeddepositional theory. How an item comes to beincorporated into the archaeological record has amajor impact on the potential meaning that canbe attached to it by the archaeologist. Was theitem intentionally placed? Or discarded? Or lost?Did the deposit occur as a result of naturalprocesses or human agency? Clearly, to answerquestions of this kind we need to know some-thing about the systemic context and the waymaterial items were used and consumed prior todeposition.

The second question relates to post-depositional and recovery processes. Post-depositional theory is concerned with whathappens once an object has left the systemiccontext. What portion is preserved (or not pre-served)? How is the deposit later modified orredeposited? To what extent has scavenging orrecycling of materials taken place? The naturalelements of these questions are effectively tapho-nomic, while the human aspects may reflect theactivity of the same or different cultural groups, atnear or more distant points in time.

Archaeologists have increasingly realized theimportance of recovery theory (Clarke, 1973).

212

Recovery theory is concerned with how the actualprocess of archaeological discovery and recoverycan distort or bias the perception of the archaeo-logical record. In terrestrial archaeology, anappreciation of recovery processes is reflected inthe practice of random sampling in field survey, toensure the sample of discovered sites is represen-tative, and the use of screening during excavationto standardize recovery. For underwater research,an entire host of additional issues relating tohuman physiology and perception in an under-water environment must also be considered(Muckelroy, 1978: 182; van Tilburg, 1994).

Shipwrecks as archaeological phenomena:the archaeological systematics ofshipwreck-sitesAn overview of formation theory for shipwreck-sites must necessarily extend from pre-depositionprocesses through to the archaeological recoveryand analysis of remains from an underwater site.The goal of this overview is to provide a generalsense of the framework or outline of the theoreti-cal process, rather than to provide an exhaustivetreatment of any of the particular aspects oftheory. In the next section, based on recent workin the Great Lakes, some specific aspects ofdepositional and post-depositional theory forstranded wooden vessels will be offered. It shouldalso be noted that many of the concepts drawntogether here have already been suggested andemployed in one guise or another in underwaterresearch[1] so the effort here is directed more atshowing how these different elements can beintegrated into a single unified approach to under-standing underwater archaeological deposits.

Prior to the wreck event itself, the vessel, itscontents, and crew all operate within an ongoingsystemic context. Souza refers to activities in thisstage as ‘pre-depositional processes’ (1998: 47–48). These activities occur in the period immedi-ately prior to the wreck event and reflect efforts bythe crew to compensate for risky local conditionswhile the ship is in passage. These activities aresometimes referenced as part of the ‘wreckingprocess’ (Souza, 1993: 308) or simply as an aspectof the ‘formation processes’ (Murphy, 1993: 268–269). In a strict sense, such operations are occur-ring within the systemic context and do notthemselves result in archaeological deposition.They do, however, produce an imprint on thematerials and spatial relationships that come to be

JOHN M. O’SHEA: THE ARCHAEOLOGY OF SCATTERED WRECKS

manifested in the archaeological record. Souza’sexample is a case in point for understanding thecircumstances of a given wreck.

Since all activities in the systemic context maypattern the material remains in the archaeologicaldeposit, these pre-depositional processes can beviewed as part of a much broader range ofactivities and practices. The technique and styleemployed in construction, alterations and repairsexperienced over the life of the vessel, the currentuse of the vessel, all as well as the immediatepre-wreck attitude of the vessel, all fall withinthis realm (Conlin, 1998). A consideration ofthese pre-depositional features can serve to high-light diverse classes of information that may berecoverable from the wreck-site.

It is in the consideration of such ‘pre-depositional’ aspects of the vessel that the com-plementary role of historical research becomesmost prominent. And as with the archaeologicalinvestigation, such historical investigations needto be pursued on a wide range of scales, fromthe global social history (Gould, 2000) to theminutiae of idiosyncratic nailing techniques ofparticular shipyards (McCarthy, 1996).

Depositional processesDepositional theory should account for thecharacter and circumstances of the wrecking andfor any other factors that are relevant to the initialcreation of the archaeological deposit. There are,obviously, many ways that these depositionalfactors might be categorized. One approach is tobegin with a classification of wreck types, sincethis should have important implications for theform and character of the resulting deposit(Thompson, 2000: 34). For example, in theAnnual Reports of the United States Life SavingService (1876–1914) wrecks are divided into thecategories of Foundering, Stranding, and Other.As a starting point, the contrast between vesselsthat have foundered and stranded seems quiteuseful since it implies very different trajectories forthe final disposition of the wreck. A vessel thatfounders is usually in deeper water and will tendto arrive at the bottom in a more or less intactcondition. By contrast, a vessel that is stranded ismuch more likely to be deposited in a broken upcondition. One can easily make the same charac-terization for the different categories of circum-stances subsumed under the rubric of ‘othercauses’, such as wrecking due to fire, collision,hostile action, or scuttling. The systematics

of scuttled vessels presents its own interestingcomplexities, since it would be important todistinguish intentional abandonment of a vesselfrom the use of scuttling as an emergency pro-cedure to save the vessel for subsequent salvage(which of course failed, since the vessel remainedon the seabed for archaeological investigation).

Within each category of depositional process,the task of the nautical archaeologist is twofold:(1) to determine the unique characteristics of agiven depositional pathway that would enable oneto identify the process from the resulting wreck-age; and (2) to ascertain what other aspects ofsite formation can be inferred once the specificdepositional pathway has been established. This isthe basic dialectic or interplay between obser-vation and theory that constitutes the applicationof formation theory in archaeology (Muckelroy,1978: 189–190). How does one know that thisvessel was intentionally scuttled? And knowingthat it was intentionally scuttled, what else can beinferred about the character and condition of thevessel at the time of deposition?

Environmental conditions at the time and placeof the wreck represent another relevant set ofdepositional processes. Clearly, if the localweather and water conditions are known, or canbe determined, a great deal can be predictedregarding the expected aspect and condition ofthe wreck, and the distribution of wreckage.Conversely, if these factors are not known apriori, it may be possible to infer them from thecondition of the wreck. There may be otherdepositional factors that are worth consider-ing, such as the suddenness of the sinking, orwhether the loss of the vessel was controlled orcatastrophic.

Post-depositional processesNautical archaeologists have devoted consider-able thought to the changes that wrecks undergoonce they are deposited (Gould, 2000: 8–9). Muchof this effort has been anecdotal and site specific,but there have been efforts to systematize theseprocesses. Muckelroy’s study of the Kennemerlandwreck is a pioneering example (1975). In this case,Muckelroy employed cluster analysis to identifysets of artefacts that had similar distributionalproperties, either as a result of their originalposition on site or as a result of their being actedupon by similar forces in the underwater environ-ment. Muckelroy identifies two classes of post-depositional processes, which he terms ‘extracting

213


filters’ and ‘scrambling devices’ (1978: 165–182).Following this usage, extracting filters are pro-cesses that remove material from the site so thatthey are not present for discovery. Scramblingdevices, by contrast, are processes that movematerials from their primary context. One prob-lem with the concept of scrambling devices is thatit implies a randomizing or pattern diminishingeffect. Yet, many natural and cultural processesthat act on shipwrecks in situ have the effect ofintroducing new organization or pattern, such asthe simple process of size sorting (Murphy, 1990:53). Similarly, there are a number of factors thatcan produce an artificial concentration of related,or unrelated, cultural materials at the site ofthe wreck, in effect, providing the inverse toMuckelroy’s extracting filters. An example isthe case of the Adelaar (Martin & Long, 1975;Muckelroy, 1978: 180). Any systematic treatmentof post-depositional factors must account not justfor the weakening of patterns and associations,but also for the creation of new and potentiallyspurious patterning.

Given the great number of processes and fac-tors that can exert an influence on wrecks oncethey have been deposited, it is difficult to developa classification that is both simple and yet retainsanalytical value. One useful starting point may beto distinguish those processes that are the result ofdirect human intervention as opposed to naturalgeophysical and biological processes (McCarthy,2000: 53).

Direct human intervention can take a variety offorms, ranging from salvage and scavenging, tointentional breakup of the wreck, to new con-struction and stabilization. In many cases, thisintervention is intentional and is oriented towardsachieving an identifiable end, whether to clear achannel for navigation or to remove artefactsfrom a wreck for personal profit. Yet there willbe other instances where the disturbance is inad-vertent, as when a fisherman’s net is snagged orwhen a developer drives the pilings for a newcondo project through a buried wreck. In all ofthese cases, however, the likely biasing effect onthe archaeological record can be predicted andcontrolled, at least analytically. Direct humanintervention can also have interesting temporaland cultural dimensions. Is the interventionthe result of actions by the vessel’s own crew,by contemporary salvagers, or other, possiblyindigenous or unrelated, cultures? Equally, is theactivity occurring near in time to the wreck or ata long time removed? Here, again, the interplay

214

between observation and theory leads the archae-ologist both to seek diagnostic attributes associ-ated with differing kinds of activities and to makeinferences about other aspects of the wreck thatmay not be visible at the site.

There are so many potential geophysical andbiological processes that can affect or alter ashipwreck-site that any effort to summarize themhere would be futile. These processes are perhapsbest thought of as the ‘ecology’ of the wrecklocality (Muckleroy, 1978: 181). This definitionwould include all of those factors (climatological,geological, biological) that may affect the site.While these processes are general in nature, theireffects and significance will be extremely local incharacter. They may also be variable, as in thecontrast between normal bottom conditions andconditions that occur during storms (Murphy &Jonsson, 1993). It should also be anticipated thatthe wreck itself will have an impact on the siteecology, as a sediment trap, a habitat for marinelife, or even as a source of toxic materials (LaValleet al., 1999). These processes will have a majoreffect on what remains of the wreck, how it isdistributed, and indeed, whether the wreck is infact discovered. Yet as natural processes they arelargely predictable and their likely effects can beanticipated.

Recovery processes

The final step in the chain of reasoning is recoverytheory. Included under this heading would be howa given site comes to be discovered, and whatfrom the wreck’s archaeological deposit comesto be recorded and recovered for analysis. Each ofthese questions also implies a converse—whatdeposits are not discovered, and what remains arenot recovered or recorded? Questions of this kindare particularly significant when the analysis stageof research begins, or when the analyst seeks tomake comparisons or to draw conclusions basedon a body of data. Is a particular search techniquemore likely to find one kind of wreck-site thanother types? Are wrecks in shallow water morelikely to be recorded than wrecks in deep water?Was the recovery of artefacts controlled (was thedeposit screened, for example)? To what extenthas the site already been scavenged or collected?These are precisely the same questions that anarchaeologist working on land must ask beforeany meaningful analysis can be done.


Theory and systematicsIn this brief overview of archaeological formationprocesses and underwater sites, one might wellask, so where is the theory? What has beenoutlined is more a structure for organizing inquiryand data, than an explanatory or predictivetheory. In fact, the presentation has been more adiscussion of meta-theory (or a theory of theory)than specific theory itself. The meta-theory high-lights the blanks that the investigator must fillin with substantive information relevant to theparticular site’s circumstances. This substantivetheory will come from a number of related disci-plines: oceanography, geology, zoology, physicswhich must be adapted to the specific conditionsof the particular wreck under investigation. Theusefulness of formation theory, for both the ter-restrial and nautical archaeologist, comes fromthe systematic structure it imposes. In effect, itprovides a checklist of conditions that must besatisfied to ensure that the information recoveredcan be used as evidence to understand past activi-ties. Formation theory also provides a structurefor applying information that is gained from awreck-site. Nowhere is this systematic structuremore useful than in cases where formation theoryenables the archaeologist to work back andforth between the written records and thearchaeological deposit, amplifying both.

Identifying formation processes forscattered wrecks: the Au Sable Shoreregion of western Lake HuronThe study of scattered wrecks is a far cry from thepopular image of shipwreck archaeology, with itsfocus on the discovery or identification of a singleintact vessel. Instead of the single vessel, researchis focused on a region in which wrecks occurred.Wrecked vessels and wreckage within the area arenot viewed in isolation, but are viewed instead asan aspect of the regional environment. An import-ant goal of the Au Sable Shores survey was thediscovery and identification of vessels that hadbeen lost in the area. But the survey was alsointended to investigate how the archaeologicaldeposits came to be in their present locations, andto identify systematic elements in these distribu-tional processes that might be applied to securethe future protection and preservation of thewreck-sites. The following sections present aninitial classification of the systematics of stranded

wooden vessels in the Great Lakes. While thediscussion is cast in terms of the research inwestern Lake Huron, information is drawn from amuch wider region and the results will also, it ishoped, have more general applicability.

0 100km

Au Sable ShoreSurvey Region

LakeH

uron

N

Figure 1. The Au Sable Shores region, western Lake Huron.(Drawing: J. M. O’Shea)

BackgroundThe Great Lakes region in North America wasan area of intensive water-based commerce andtraffic during the 19th and early 20th centuries(Mansfield, 1899). Lake vessels were responsiblefor moving major bulk goods, such as coal,timber, grain, and ores throughout the Midwest.Prior to the coming of the railroads, they werethe primary source of food, manufactured goods,and transport for the peoples living around thelakes. Vessels during this time were overwhelm-ingly of wood. They comprised wind-drivenschooners, steam-driven propellers and side-wheelers, and towed barges (often modifiedschooners) (Thompson, 2000). The Au SableShores project[2] is concerned with the wrecks ofsuch vessels in western Lake Huron within a

215


Table 1. Near shore wrecks in the Au Sable shores region of Lake Huron

Name Commissioned Sunk Type Capacity Cargo Loss type Comments

May Queen 1855 1859 Schooner 246 tons Wheat and barley StrandingWater Witch 1861 1863 Bulk Freight Propeller 458 tons Copper, passengers FounderedEureka 1869 Sloop 10 tonsSummit 1856 1872 Schooner 226 tons Iron Ore Stranding SalvagedTable Rock 1853 1872 Schooner 179 tons Lumber StrandingWhite Squall 1852 1872 Schooner 241 tons CollisionOcean 1873 Barge FounderedWayne 1848 1875 Schooner 80 tons StrandingAthenian 1856 1880 Schooner 372 tons Lumber StrandingStranger 1872 1880 Schooner 16 tons StrandingJohn Prindiville 1867 1882 Freight-Propeller/Tug 270 tons Ballast StrandingJames C. Harrison 1870 1885 Schooner 518 tons Iron Ore Stranding Salvaged?Susan Ward 1863 1885 Barge 365 tons Lumber Stranding Refloated?Ferguson 1886 Schooner-Barge FounderedGeorge Hand 1868 1888 Tug 25 tons None FireJane Mason 1880 1889 Schooner 33 tons Bricks StrandingMears 1869 1889 Schooner-Barge 429 tons Lumber Stranding Located?Midnight 1856 1889 Schooner-Barge 288 tons Lumber StrandingSea Gull 1864 1890 Steam-barge 289 tons Ice FireJohn Shaw 1885 1894 Schooner 928 tons Coal Stranding Salvaged?Volunteer 1889 1896 Schooner 31 tons Lumber StrandingGeorge Steele 1855 1898 Schooner 271 tons Lumber StrandingBaltimore 1881 1901 Bulk Freight Propeller 1160 tons Coal Stranding LocatedThomas P Sheldon 1871 1901 Schooner 669 tons Collision Refloated?Mary E. Pierce 1871 1906 Tug 22 tons None StrandingWawanosh 1876 1906 Schooner 370 tons Lumber StrandingAndrew McLean 1890 1916 Tug 24 tons FounderedGoshawk 1866 1920 Schooner-Barge 550 tons Lumber Foundered LocatedOwen 1881 1921 Tug 44 tons None FireLinden 1895 1923 Bulk Freight Propeller 894 tons Ballast Fire RemovedLangell Boys 1890 1931 Bulk Freight Propeller 387 tons Ballast FireHercules 1904 1932 Dredge 559 tons NoneDudley 1926 1934 Dredge 95 tons None Foundered

38 km zone between Tawas Bay and the Au SableRiver (Fig. 1). This is an area of relatively shallowcoastal water and sandy seabed that lies betweenthe notorious wreck zones at Point aux Barquesand Thunder Bay. While Tawas Bay was theprincipal harbour of refuge in this portion of thelake, vessels, particularly those that were wind-powered, often ran aground here during easterlystorms. Given cooperative weather, most vesselsin this predicament were refloated and saved, butin the case of violent weather, stranded shipswere rapidly broken up by the waves and lost.Table 1 summarizes the vessels that are currentlythought to have been wrecked and actually lost tostranding in this region. Vessels that foundered inthe deeper waters of Lake Huron are not includedin this list.

The unusually low lake levels of 1999–2001provided a unique opportunity to investigate the

216

distribution of the near-shore wrecks in the re-gion. The aim of the research has been to identifythe wreck-sites and wreckage along the near-shoreregion and to link these finds to specific vessels.Shallow water survey has been conducted both toidentify areas of major wreckage concentration,and to document the surviving vessel wreckage.The research has also included the search ofcontemporary accounts of vessel loss and ananalysis of historic maps to chart the rapidlychanging shape of the near-shore region.

As can readily be imagined, the archaeologicalinvestigation of vessels that have been strandedand broken up for a century or more presentsformidable problems of interpretation and analy-sis. Indeed, it is because of this complexity, thatthe understanding of formation processes is socritical. The discussion that follows summarizesthe principal depositional and post-depositional


processes that have been identified during thecourse of the Au Sable Shores research.

Figure 2. The wrecked schooner Advent, Coos Bay, Oregon, 1913. Letters A, B, and C mark the aft, fore, andside planking of the vessel. The arrow marks the broken ends of the lower frames. (Photo: courtesy James A.Gibson.)

Wreck creation: depositional processesThe great majority of the wrecks in the Au SableShores region represent strandings (Table 1),although for several vessels the proximate causefor loss was fire. Yet, even when fire was thecause, the burning vessels either drifted, or weretowed, into shallows where they subsequentlygrounded and broke up. However, the very firstissue that must be assessed is whether a givenvessel that was reported as stranded was in factlost. Since the lake bottom is sand, vessels wereoften refloated or salvaged, with minimal damage.Indeed, captains occasionally scuttled their vesselsin shallow water as a way of protecting them fromsevere weather. In the case of true losses due tostrandings, the vessels typically were broken up bywind and wave action, as they were held fast tothe bottom. This may occur as a single event, or itmay occur in stages, with the vessel first beingstranded, and then being broken up by a subse-quent storm. Clearly, if there is an interval, it ismuch more likely that portions of the vessel andits cargo will be salvaged.

A central concern for the archaeology of scat-tered wrecks is the manner in which a wooden

vessel breaks up once stranded. While there hasyet to be a systematic study of this process forstranded wooden vessels, some regular elementscan be offered based on anecdotal accounts andpractical considerations.

A striking visual example of vessel breakup isprovided in a contemporary photo of the Advent,a 431-ton schooner wrecked on the Coos Bay bar,Oregon, in February 1913 (Gibbs, 1986: 82) (Fig.2). While this is not a Great Lakes vessel, its sizeand the circumstances of loss make it comparableto strandings in the Great Lakes. The vessel ranaground while crossing the bar, and was subse-quently broken up by waves. Inspection of thisphotograph reveals the major components of abroken vessel. To the left, the aft portion of thevessel is intact [A]. To the right, the bow issimilarly visible [B]. In the centre foreground is alarge section of planking [C], which extends fromthe gunwales to the turn of the bilge. This seg-ment of planking presumably contains the upperfuttocks of the vessel’s frames, but does notappear to retain much if any of the decking.Immediately to the right of the planking is asegment of the fore interior beams and decking. Inthe centre of the photo, just above the mass ofloose timbers, can be seen the broken ends of thelower frame elements of the vessel [arrow]. Theseare particularly significant insofar as they suggest

217


that the keel assembly and bilge remain intact andburied beneath the sand. Finally, surroundingthese largely intact portions of the vessel is a cloudof loose timber and other debris.

In viewing a photo such as this, the archaeo-logical questions are obvious; what comes to bedeposited and where? Certainly some systematicelements can be suggested. For example, the loosetimbers and associated wreckage can be expectedto be carried by the water from the wreck-site andscattered widely. This distribution ought to reflectboth the normal water currents in the locality andthe specific weather conditions associated with thewreck, as when wreckage is tossed high up on abeach during storms or high water. At the otherend of the spectrum, it appears that the moststable portion of the wreck might be the keelassembly and bilge, which is already buried in thesand and which would not be expected to moveunless the sea bottom itself were to be exposed. Assuch, the loose timbers may be expected to pro-vide the most extensive evidence of the wreck, butto have the least spatial fidelity to the wreck-site,while the keel assembly should have the closestspatial association to the actual wreck-site, buttypically with the lowest visibility. These expecta-tions are not qualitatively different from those fordeep-water wrecks except perhaps for the quantityof loose material and the distance it travels.

This characterization of the loose wreckage iscertainly in accord with 19th-century accounts ofGreat Lakes wrecks, where often the first evidenceof a marine disaster was the appearance of float-ing debris along the lakeshore. A good localexample of this scattering process is provided bythe wreck and wreckage from the steam barge D.M. Wilson.[3] The Wilson foundered in a northerlygale about 3 km NNE of the light at Thunder Bayon 27 October, 1894 (Swayze, 2001). The vesselapparently was broken up during a second, north-erly storm on 10 November. Quantities of wreck-age washed ashore between Au Sable Point andTawas Point, nearly 160 km to the south. Thecharacter of the wreckage that came ashore isrecorded in the weekly Iosco County Gazette,published in East Tawas on 17 November, 1894:

What was evidently a large steam barge has found-ered off Sable, her hull and deck parting. A terriblenorthern storm and snow raged during the night shewas wrecked. The boat’s deck is broken in, andseveral pieces of the pilothouse is gone. The hurri-cane deck and the rail forward of the pilot housewere washed up in one place, the after hurricanedeck with the whistle pipe, wire and whistle in

218

another place and the port bulwark and parts of themain deck in still another place . . . Most of thewreckage was washed ashore 3 miles south of FishPoint and 5 miles from East Tawas. About 4,000cedar ties and a lot of timber also came ashore.

In this instance, not only light debris but alsolarge segments of the vessel were rapidly trans-ported and deposited a significant distance fromthe wreck-site.

A more detailed, although tragic, example isprovided by the wreck of the wooden steamerBaltimore, which ran aground in a storm andbroke up just south of Oscoda on LakeHuron’s western shore in 1901 (Ferris, 1999). TheBaltimore was a 201 ft (61 m) wooden steamerwith a rated capacity of 1160 tons. The vessel ranaground while attempting to reach Tawas Bayduring a severe northeasterly storm on 25 May,and wreckage from it was scattered widely. Apublished statement by the two survivors describ-ing the vessel’s final break-up fits well with theprevious image of the Advent:

. . . She labored and pounded on the bottom forabout two hours after this when she began to go topieces, and finally broke in two. Her rails broke onboth sides just aft of the forward house and as shebegan to split in two, we could see that many of herbeams had started and pulled loose so that she wasalso breaking in two athwartships. . . . When we leftthe vessel she had practically broken all to pieces.From the morning of the wreck until noon May 27,Deponent Murphy and others patrolled the beachand found it strewn with great quantities of wreck-age from the vessel for a distance of eighteen milessouth of Tawas. (Marine Review, No. 22, 1901: 16)

Here, as with the Advent, a stranded vesselcomes apart at the seams, and breaks into free-floating fore, aft, and side portions, along with animmobile bilge and keel. There is also a vividsense of the floating ‘cloud’ of loose wreckage thatis subsequently distributed over great distancesdown lake from the wreck-site.

In the case of the Baltimore, along with thegreat quantities of wreckage went the bodies of 12crew, and it is through their identification that it ispossible to trace positively the distantly scatteredwreckage to the Baltimore. The distribution ofrecovered crew bodies and wreckage is shown inFig. 3.

What is most striking about this map is how farand how rapidly the debris travelled from thewreck-site. The first identified body was foundamid wreckage at Point Lookout on 26 May, one


0 10km

Saginaw Bay

Sebewaing

Crew member body recovered

Wreckage

Bay City

Wreckage

CharityIsland

PointLookout

TawasBay

AuSablePoint

Oscoda

Wreckage

Lake Huron

Baltimorewreck site

Figure 3. Distribution of wreckage and recovered bodies from the wrecked steamerBaltimore, 1901. (Drawing: J. M. O’Shea)

day after the storm and some 50 km distant fromthe wreck-site. On 27 May, a second body wasrecovered along with large pieces of wreckagenear Bay City, 100 km from the wreck-site, andon 2 June, a week after the wreck, a body andsubstantial portions of wreckage, including a seg-ment of the stern and furnishings, was found onMaisou Island on the opposite side of SaginawBay (65 km direct but 200 km via shore currentcirculation). As regards the breakup of the vesseland its dispersal, it is also interesting to note that

the two survivors of the wreck were found offTawas Point lashed to a 3�4 m segment of thepromenade deck. The fisherman who rescued thesurvivors reported that the decking was drifting ata rate in excess of eight miles per hour (13 km/h)(Ottawa Point Life Saving Station Log, 24 May,1901).

Based on these examples, loose timbers andwreckage should be distributed widely from awreck-site, moving in the direction of prevail-ing currents and winds. A locality that has

219


experienced a number of wrecks, therefore, maypresent a broad scattering of planking andtimbers, with concentrations of wreckage morelikely reflecting debris ‘traps’ rather than necessar-ily indicating the location of true wreck-sites.

The distribution of wreckage from theBaltimore and the Wilson touches on a secondsystematic element in the deposition of brokenwrecks. As was seen in the photo of the Advent,large portions of a vessel may remain intactduring its initial breakup and, as the Baltimoreand Wilson accounts indicate, these large seg-ments might be scattered widely, and in muchthe same manner as smaller pieces of wreckage.This observation is borne out in numerous otheranecdotal accounts.

On reflection, this occurrence is not difficult tounderstand. As the stranded vessel breaks up,these large buoyant segments with a high degreeof structural integrity break free from the con-straint of the bottom and drift with the prevailingcurrent and winds. It would be expected that someportions would move into shallow water near thewreck-site and be beached, while other portionsdrift away from the site. It might also be expected,given their large surface area, that they may beaffected more by winds in this process than wouldindividual timbers. These vessel fragments woulddrift either until they are beached, or until theylose their buoyancy and sink to the lake bottom.The significant point, from the perspective ofarchaeological site formation, is that even largeand intact segments of broken vessels may movegreat distances from the wreck-site.

Arguably, the least mobile portion of astranded and broken vessel is the bilge and keelassembly. This is the heaviest and least buoyantportion of the boat, and is the part that in mostcases will be firmly embedded in the lake bottom.Heavy deposits of ballast and cargo may alsooverlie it. Given these factors, and its size, the keeland bilge assembly is the least likely portion of thevessel to move from the initial wreck-site and maytherefore be the best indicator of the actual wrecklocation. Unfortunately, these same factors makethe keel assembly the most difficult element tolocate visually (although its linear shape and theextensive fastenings used in the construction ofthe keel assembly make it a good target forlocation via remote sensing). This is not to say,however, that keels or bilges never move, but thatmovement will more likely be the result of post-depositional changes in the lake bottom, ratherthan to the initial conditions of deposition.

220

The rudder assembly represents another cat-egory of heavy and relatively immobile vesselelement, although it shares much in common withother portions of the midline structure. Rudderswere typically constructed of massive timbers thatwere held together by heavy metal fasteners andmetal strapping. In most instances the rudder willbe detached and deposited as a structural unitand, given its weight, it would be expected to sinkrapidly to the bottom and be deposited in situ.Lying flat on the lake bottom, the rudder mayrapidly be covered by sediment and completelyobscured from view. As such, the rudder, like thekeel and bilge assemblies, may be expected toretain a close fidelity to the original wreck-site. Itwill also have a relatively low visibility profilealthough the quantity of metal fastenings mayfacilitate the use of remote sensing techniques forits discovery.

While intact rudders are readily identifiableas ship wreckage, their great weight and sizemitigates against their removal from wreck-sites.Unlike other central portions of the vessel,though, rudders often were lost from vesselsthat were not wrecked. So while rudders can beexpected to remain at the site of loss, they may befound at locations that are not vessel wreck-sites.

Heavy machinery and furnishings from astranded vessel may behave in a similar manner.Heavy machinery can be expected, in most cases,to break free from moorings and to sink quicklyto the lake bottom at the site of the wreck. Thesemay include anchors (deployed or undeployed),capstans, windlasses, cannon (if armed), engines,boilers and other metal appliances. Smaller fit-tings, however, may float off with their attachedtimbers. In any event, large metal pieces shouldbe deposited at the immediate wreck-site. Hereagain, though, is an instance where the site ofinitial deposition may be quite different from thesite of eventual recovery, depending on a varietyof post-depositional processes.

Cargo in the hold or on the deck of a vesselpresents another interesting category of deposit.Heavy materials will sink rapidly to the bottom,and may also pin large portions of the vessel tothe bottom, while lighter materials will float awayonce the holds are broken open. In the GreatLakes, the most common cargos were ore (heavy),coal (intermediate), and lumber and grain (light)(Table 2). On stranding, vessels carrying ore werepinned to the bottom and the cargo tended eitherto stay in the intact hold, or to spill out on thelake bottom once the holds were broken. Such


cargo will resist movement from natural forces.On the other hand, ores are the cargos that weremost likely to be salvaged from the vessel, bothimmediately after stranding (when a lighter mightbe used to unload the vessel in an effort to floatthe ship, and to save the valuable cargo) or laterwhen salvagers sought lost wrecks. Lumber,which often was piled high on vessel decks ortowed in great rafts, was typically scattered duringthe wrecking process, and was recovered withdifficulty from the lakeshore and the open lake.Little would remain in the vicinity of the wreck-site itself. Coal presents a more interesting prob-lem. Coal, like ores, would often be removed fromthe stranded vessel during recovery efforts, butthere was less economic interest in salvaging itonce the vessel broke up. In such a situation, thecoal would sink to the bottom, but then betransported by underwater currents, presumablyundergoing continuous mechanical breakdowninto progressively smaller sized chunks. As such,coal might be expected to have an initial distribu-tion near the site of the wreck and then betransported progressively farther way. The size ofcoal chunks might be expected to decrease overexposure time, and the exposed outer surface ofthe material should show progressive evidence ofweathering.

To summarize briefly, as the name implies, atscattered wreck-sites, vessels are broken up intolarger and smaller pieces, which may then driftsubstantial distances from their point of origin.The large size of a vessel fragment does notpreclude it from being deposited at a great dis-tance from the initial wreck-site (Reedy, 1991: 53).The most likely portions of a vessel to bedeposited in the immediate vicinity of the wreck-site are (1) the keel and bilge assembly of thevessel, and (2) heavy machinery and other non-buoyant items and cargo. As such, only the lattertwo classes of items can provide reliable evidencefor the original location of a wreck.

Table 2. Great Lakes vessel cargo deposition and post-deposition potentials

Cargo typeCommon Great

Lakes cargosDepositional

potentialPost-depositional

movement potentialSalvage

potential

Heavy Ore, Stone, Brick High Low HighIntermediate Coal High High High/Low*Light Wood, Grain Low High High/Low*

*Potential for salvage is high immediately after stranding, but becomes low once vessel has gone topieces.

Wreck-site alteration: post-depositional processes

Post-depositional processes, both natural andhuman, can further alter the location and condi-tion of scattered wreck remains. These variedprocesses work on a number of different timescales, but most have relatively predictable effects.As previously noted, post-depositional pro-cesses associated with human agency can be sum-marized under the terms salvaging, scavenging,and intentional or unintentional scattering.

Salvage efforts were common in the GreatLakes and elsewhere, particularly when the vesselretained some structural integrity. Even in caseswhen the vessel could not be recovered, it waseconomical to salvage not only valuable cargoand machinery, such as boilers and engines, butalso anchors, rigging, and even sail cloth (Souza,1998: 43–45). The relatively shallow depth of moststrandings made salvage a particularly viablepossibility. In some cases, strandings also pro-vided the opportunity for salvage efforts to occurbefore the final breakup of the vessel. As such, it islikely that items, particularly items of high econ-omic and reuse value, will be removed from thewreck-site whenever conditions permitted.

Scavenging, by contrast, is the removal ofmaterials not intended for direct reuse.Scavenging may be contemporary or much laterin time, and may take myriad forms. The removalof copper plates from the bottoms of Britishvessels by Native Americans on the NorthwestCoast for the manufacture of tools and ritualobjects is one example, the theft of vessel bells,nameplates and furnishings for decorations or theremoval of timbers for furniture or curiosity areothers. Scavenging, like salvage, results in theremoval of materials from the site of deposition,which may or may not be the original wreck-site.Perhaps more than salvage, scavenging removesmaterials from the site that directly affect vesselidentification and dating, and it is for this reason

221


that the looting of wreck-sites is particularlyregrettable. Finally, casual scavenging or collect-ing may also result in the creation of new concen-trations of debris, as scavenged materials becomeincorporated into ‘yard’ or shore collections.

Not surprisingly, a number of post-depositionalprocesses result in a further scattering of wreckremains. This can occur as an intentional outcomeof human action, as when wrecks are blown up toensure the safety of navigation, or unintentionallywhen they are unexpectedly encountered duringshore modification activities, such as dredging orbuilding. On a smaller scale, it can occur as anoutcome of activities such as fishing, cable laying,or later vessel strandings. In the Great Lakes, thescattered wrecks resulting from coastal strandingsdid not commonly require active removal fornavigational purposes; indeed, some strandingsrepresent the towing of a hulk out of a navigationarea with the intent of stranding it harmlessly inthe shallows, as was the case with the schooner-barge Thomas P. Sheldon (Annual Report of theUnited States Life Saving Service, 1902: 67). Themain result of intentioned scattering would be toreduce large, intact portions of the vessel to asmaller size, which in turn would free them fromtheir attachment to the lake bottom and permitthem to drift as previously described.

The unintentional scattering of the wreck-siteas a result of shoreline development can have adevastating impact on archaeological deposits,since it uncovers and scatters wrecks that havealready been buried and stabilized. Obviously, thenature of the disturbance will determine whetherthe wreck-site is simply scattered or whethermajor portions of the wreck are removedpermanently from the record, as in dredging.

As mentioned previously, there are a largenumber of natural processes that may affect thewreck-site. Of particular importance to thepresent discussion of Lake Huron is the ongoingprocess of shoreline and lake bottom remodellingthat is associated with the recurved spit features ofTawas and Au Sable Points (Wadsworth, 1975;Meadows, 2000). These spits exist in a very activeshore and near-shore environment, and have wit-nessed a repeating cycle of aggradation and scour-ing of the near-shore lake bottom. This processresults in the periodic exposure and reburial ofwreckage and debris. During the scouring phase,previously buried debris is exposed on the lakebottom, where it becomes vulnerable to deteriora-tion, movement, and scavenging. The settling ofheavier debris and artefacts during these scouring

222

episodes also results in the accumulation andmixing of materials of differing ages on theimmobile lake bottom substrate. During the sub-sequent aggradation phase, the materials thathave accumulated on the bottom are reburiedas a lag deposit beneath a thick layer of sand.Depending on the stage and the speed of theprocess, the location may be incorporated into theaccumulating spit and eventually be found on orbeneath dry land.

Survey in the Au Sable Shore area of LakeHuron provided several examples of the predict-able character of these geological and hydrologi-cal processes. One is a keel and keelson assemblythat was recorded to the east of Au Sable Point inthe spring of 1999. The keel assembly becameexposed in about 1m of water as the pointmigrated in a southwesterly direction. As LakeHuron experienced its normal seasonal drop inwater level into the autumn, progressively more ofthe keel assembly was exposed and, during thewinter, the upper surface was above the ice. Yetby the following March, the wreck was completelyburied beneath the sand, and was now locatedsome 3 to 6 m inland from the lakeshore. A severeeasterly storm in May reclaimed much of thenewly created shore, and re-exposed a small por-tion of the keel assembly on the immediate shore-line. By August, 2000, the keelson was backbeneath a half metre of water, and roughly 6 mbeyond the shore. So in the space of little morethan a year, the wreck was exposed, buried,re-exposed, re-submerged, and currently is in pro-cess of being buried once more. While it didundergo major damage and deterioration due toexposure during this period, the wreckage did notmove, although its location relative to the modernshoreline changed.

The debris-concentrating effect of specific shorefeatures can be documented on a number ofspatial scales. At a macro-scale, the effect isclearly illustrated by plotting the location ofdebris encountered during the near shore survey.Figure 4 represents the Au Sable Shore regionfrom the mouth of the Au Sable River to TawasPoint, and indicates the location of all modifiedtimber observed during survey in May of 2001.(The western half of Tawas Bay was not surveyedduring the 2001 season.) The bulk of moderntimber is treated lumber from sea walls, while thevessel timbers derive from wrecks with an averageage of more than 100 years. For clarity, the figuredoes not show the distribution of natural debris,although these materials exhibited the same


Figure 4. Distribution of vessel debris and modern timber along the Au Sable shorearea of western Lake Huron, recorded during survey, May 2001. Background satellitecoverage is photo-mosaic produced from USGS digital ortho-quads (DOQ), 1993.

pattern of occurrence. Regardless of age orsource, debris is not distributed evenly over theregion, but is strongly concentrated in majordebris fields in the vicinity of the two recurvedspits. Since free-floating wreckage from a brokenvessel, like any other debris, tends to accumulatein such areas, the spatial association of wreckage

and artefacts within these locations will have nonecessary relationship to the original source of thematerials.

A closer look at one such debris trap atAu Sable Point clearly illustrates this effect. Onthe inner side of the Au Sable Point spit, andimmediately opposite the site of the previously

223


Table 3. Portable artefacts from the Au Sable Point debris trap

Object Dating range

Nu-Grape Soda Bottle 1922–1946Cork Stopper Olive Bottle 1900–1917Coke Bottle, 6 oz. Patent D Type 1931–1951Coke Bottle, 6 oz. Patent D Type 1931–1951Coke Bottle, 26 oz. 1951–1965 (?)Wood Caulking Mallet Common late 19th centuryCoal (abundant, large unweathered chunks) 1901?Fishing lures, Monofilament line, Lumber, Aluminum beer can,Plastic soda bottles

Modern

described keel and keelson assembly, both scour-ing and redeposition was occurring. Water actionon the inside of the recurved spit produced a smallarea of relatively deeper water, from which mostof the lake sand had been transported, leaving animmobile substrate of lake clay and gravel. In thiszone, several pieces of vessel wreckage, derivingfrom more than one vessel, were observed. Theyincluded a 10�2 m segment of planking, a seriesof frames still in positional relationship suggestingthey were deposited as a unit, several spar frag-ments and a number of portable artefacts. Thesematerials were first recorded in 1998 and by thespring of 2000 were entirely buried beneath AuSable Point.

The concentrating effect of this environment iswell illustrated by the portable artefacts recovered(Table 3). It should be noted, however, that thislist is not a complete representation of the port-able artefacts present. Almost immediately onexposure, the site was overrun by looters who,using diving gear, dredge hooks, and jet skiscarted away a large, but unrecorded, portion ofthe site collection. Table 3 lists all the recordeditems and their associated date. What is immedi-ately obvious from this table is that there arematerials of widely differing ages that have cometo rest on the same depositional surface alongwith the vessel wreckage. Several of these itemsmight be associated with one or more of the AuSable Point wrecks, the vessel fragments andcaulking mallet being the least ambiguous. Threeof the soda bottles date from within a span that isconsistent with the Langell Boys wreck, while theolive bottle predates the Langell Boys wreck bymore than 15 years but overlaps with the wreckof the Baltimore above. Much of the coal foundin the area is locally attributed to the Baltimore.

224

To these historic artefacts can be addedmodern fishing lures, lumber, and aluminium beercans.

Archaeologically, this deposit, which currentlyis sealed again beneath Au Sable Point, providestwo points of interest. The obvious one is thatitems found together in a single deposit need notderive from the same event, nor be of similar age.This is the deflating and concentrating effect of adynamic near-shore environment (Murphy, 1990:15–16). The second point relates back to thedeposition of differing portions of broken vessels.All the vessel fragments encountered in this debristrap are portions that are likely to have driftedaway from an initial wreck-site. As such, theirpresence in the debris trap may be every bit asserendipitous as the occurrence of any of theportable artefacts. In fact, it appears that at leastsome of these vessel fragments may relate to thenearby keel assembly and thus have moved only ashort distance from the apparent wreck-site. Onthe other hand, since the vessel fragments donot derive from a single vessel, some portion ofthe wreck assemblage must derive from a moredistant location.

There is a possible parallel here between theconcept of a ‘ship trap’, which has evolved fromthe idea of a particularly hazardous locality forships (Throckmorton, 1964: 51–61) to localitieswhere shipwrecks are concentrated (Gould, 2000:82–91), and the ‘debris traps’ which are observedin Lake Huron. Probably in most instances whereclusters of intact vessels are encountered, it isreasonable to search for conditions that wouldhave resulted in repeated wrecks in the locality,but for deep-water wrecks in more open waters,the possibility of natural processes producing aconcentration of vessels might also be considered.


A third variety of post-depositional alterationalso relates to the movement of large vessel seg-ments, although at a time far removed from theinitial wreck. In December 1994, a large panel ofintact vessel planking (12�3 m) was observed,moving just below the surface of Lake Huron.Over the next 6 months, the fragment was drawnand photographed by local residents as it workedits way along the shore, damaging several smallboat docks in the process. Eventually the frag-ment was dragged ashore, cut up, and burned toprevent further property damage. In May 2000, afragment of this original panel was relocated on aprivate beach. The specimen was identified fromthe earlier photographs and the saw marks atboth ends of the fragment were clearly visible. Theproperty owner later confirmed the identificationof the fragment and its source. In this case, a largesegment of an unidentified schooner, presumablywrecked more than a century earlier, was freedfrom the bottom and retained sufficient buoyancyto migrate with the current an unknown distance,then to be intentionally scattered, scavenged, andrelocated during archaeological survey.

Despite the many confounding influencesaffecting the distribution and deposition of scat-tered wreckage, there are systematic elementsthat can assist archaeological interpretation. Forexample, once debris breaks free of the strandedvessel and begins to drift, it will drift in thedirection of the prevailing winds and currentsuntil it ultimately becomes stuck in one of thesedebris traps. Even through successive cycles ofdeposition and release, the debris will continue tomove following this directional pattern. In the AuSable Shores region, the prevailing direction ofdrift generally coincides with the long shore cur-rent, which transports debris in a north to southdirection. As such, it is unlikely that debristrapped on Au Sable Point would derive fromwrecks located south of this point. Such generalpatterns, of course, are subject to specific localconditions. For example, in Saginaw Bay, thecombination of current and easterly winds canrotate debris and wreckage north along the shoreof the ‘Thumb’ and then back west across the bay.Given this local circulation system, it was notuncommon for the debris from wrecks at Pointaux Barques on the east side of Lake Huron toappear in Tawas Bay. This was the case with theschooner Table Rock in 1872, which was lost fromits tow near Point aux Barques and went ashoreon Tawas Point (Iosco County Gazette, 2 Oct.,1872).

Discussion

Beyond their obvious cautionary value, thesecases highlight systematic elements in the deposi-tional and post-depositional processes at work onscattered wreck-sites. In many of these instances,from the Advent photograph to the case just cited,large panels of planking are broken off and moveas a unit. In these instances, the interior orexterior planking, or both, remains attached tothe associated frames, and survives as large flatpieces of timber. These panels seem most often toderive from the central portion of the vessel (asopposed to the fore or aft sections) and to repre-sent the upper portion of the ship’s side and rail,above the turn of the bilge. This pattern ofbreakup makes sense both in terms of the struc-ture of a wooden vessel, and in terms of where thebreaking force would be exerted to the side of astranded vessel (Labadie, 1989: 126).

The structure and shape of these panels mayalso partially explain their mobility. The panelshave a very large surface area on which windand current may exert force. This, coupled withresidual buoyancy, would seem to provide themeans for the panels to be moved along in thewater column. Clearly, there is a necessary rela-tionship between the size and weight of the paneland the amount of energy in the lake environmentthat is needed to produce movement. By contrast,the more complex, three-dimensional geometry ofthe fore and aft portions of a vessel would presentless surface and more drag to counteract theforces for movement. As such, fore and aft seg-ments should be less likely to move in the waterafter their initial deposition and loss of buoyancy.If these portions break up further, they will pre-dictably separate into the heavier axial elements,which will tend to remain in place, and moremobile and buoyant planking, which may resumedrifting.

While a realistic assessment of the depositionand post-depositional processes acting on shallowwater wrecks might lead one to conclude thatnothing useful can be done with these sites, thatassessment is far from accurate. Instead, thesefactors simply force the nautical archaeologist toabandon Pompeii and to adopt the normal con-ceptual tools that terrestrial archaeologists applyto the understanding of complex archaeologicaldeposits. Even with badly scattered wrecks,nautical archaeologists still enjoy an importantadvantage over their land-based counterparts,and that derives from the planned structure of the

225


vessel itself. Indeed, the very complexity andredundancy of vessel construction and functioncan be the critical key for untangling the com-plexities of deposition and post-deposition. Withknowledge of deposition and post-depositionalprocesses, one can use the canons of vessel con-struction to determine whether, for example, vari-ous pieces of wreckage can be attributable to thesame, similar, or different vessels. This attributionmay be based on the style or period of construc-tion, the type of vessel and the overall size orcapacity. All of this information is encoded in theconstruction and dimensions of the vessel itself,and requires only accurate observation and thecorrect identification of the vessel portion beingexamined. In essence, the nautical archaeologistadopts a kind of forensic analysis of vesseldebris.

The present research also stresses the impor-tance of understanding the depositional environ-ment as another line of ‘forensic’ evidence. In theAu Sable Shores study area, for example, theprevailing long shore current moves wreckageand other materials in a southerly direction.Therefore, if the approximate location of thestranding is known or can be inferred from his-torical records, it will normally be possible toeliminate fragments of vessels that are discoverednorth of the wreck-site, while remembering ofcourse, that historical records have their ownlimitations on accuracy. Such information, whichcan often be found formalized in general environ-mental models of lake circulation and climate, canbe coupled with historic meteorological observa-tions at times relevant to known wrecks to formconcrete predictions of wreck location and debrisscattering.[4]

Since scattered wrecks and wreck localitiescan yield important historical and archaeologicalinformation, it is necessary to reassess the kinds ofmanagement and protection such sites receive. Inprinciple, wrecks of a scattered character are

226

afforded the same legal protection as intactwrecks. In practice, however, and given limitedresources, only intact wrecks receive any signifi-cant research and preservation effort. Such apolicy, while perhaps understandable, creates aserious obstacle to the historical understanding ofthe nautical past, since it effectively eliminates anentire class of vessel loss from research considera-tion. Current efforts within the Great Lakesregion to create underwater preserves, which aredefined regionally and contain a variety of wrecktypes, represent one important means by whichthe full variety of wrecks can be examined andprotected.

ConclusionsWhile many factors work to disperse and intermixthe wreckage from stranded wooden ships, thedocumentation of the depositional and post-depositional processes provides the bedrock onwhich an archaeological study of scattered wreck-sites can be based. An understanding of howthese sites are formed and modified will not onlyenhance their research potential, but will alsoenable culture resource managers to betterevaluate scattered wreck-sites and justify theirprotection.

AcknowledgementsThe Au Sable Shores research has benefited fromthe advice and steadfast support of John R.Halsey (State Archaeologist) and Kenneth Vrana(Center for Maritime and Underwater ResourceManagement). The historical advice of DaveSwayze and Neal Thornton is gratefully recog-nized. It is perhaps also fitting to acknowledge theunnamed looters, whose plundering of a newlyexposed wreck provided the impetus for this studyof western Lake Huron’s scattered wrecks.

Notes[1] For example, see Gould, 2000; Hulse, 1981; McCarthy, 2000; Muckleroy, 1978; Murphy, 1990; Quinn et al., 2002; Ward

et al., 1999.[2] The Au Sable Shores project is based in the Great Lakes Division of the Museum of Anthropology, University of Michigan.

The research described was funded in part by the Coastal Management Program of the Michigan Department ofEnvironmental Quality (DEQ), Project #01–309–13.

[3] There is some uncertainty whether the wreckage described derives from the D. M. Wilson or from the schooner John Shawwhich ran aground and broke up on 13 November, 1894 near Greenbush, Michigan, roughly 45 km north of Tawas Bay.In either case, the wreckage observed at Tawas Bay had scattered quickly and over a considerable distance.

[4] See Forsythe et al., 2000 for a comparable case in Ireland.[5] This is maintained as an online database at http://greatlakeshistory.homestead.com/home.html


BibliographyAnnual Reports of the United States Life Saving Service, 1876–1914. Washington.Bass, G., 1983, A plea for historical particularism in nautical archaeology. In R. Gould (Ed.), Shipwreck Anthropology, 91–104.

Albuquerque.Binford, L., 1981, Behavioral archaeology and the ‘Pompeii premise’. Journal of Anthropological Research, 37: 195–208.Clarke, D., 1973, Archaeology: the loss of innocence. Antiquity, XLVII: 6–18.Conlin, D., 1998, Ship evolution, ship ‘ecology’ and the ‘masked value hypothesis’. IJNA, 27: 3–15.Dumas, F., 1972, Ancient wrecks. In UNESCO. Underwater Archaeology, a Nascent Discipline, 27–34. Paris.Ferris, P., 1999, The Shipwrecked Baltimore, Forest Park, Illinois.Fontenoy, P., 1998, A discussion of maritime archaeology. In L. Babits & H. van Tilburg (Eds), Maritime Archaeology: A

Reader of Substantive and Theoretical Contributions, 47–52. New York.Forsythe, W., Breen, C., Callaghan, C. & McConkey, R., 2000, Historic storms and shipwrecks in Ireland: a preliminary survey

of severe synoptic conditions as a causal factor in underwater archaeology. IJNA, 29: 247–259.Gibbs, J., 1986, Peril at Sea: A Photographic Study of Shipwrecks in the Pacific. West Chester, Pa.Gould, R., 2000, Archaeology and the Social History of Ships. Cambridge.Hulse, C., 1981, A Spatial Analysis of Lake Superior Shipwrecks: A Study in the Formation Processes of the Archaeological

Record. PhD. Dissertation, Michigan State University, East Lansing.Labadie, C. P., 1989, Submerged Cultural Resources Study: Pictured Rocks National Lakeshore. Southwestern Cultural

Resources Center Professional Papers, 22. Santa Fe.LaValle, P., Brocks, A. & Lakhan, V. C., 1999, Zebra mussel wastes and concentrations of heavy metals on shipwrecks in

western Lake Erie. Journal of Great Lakes Research, 25: 330–338.Mansfield, J., 1899, History of the Great Lakes (reprinted 1972). Cleveland.Martin, C. & Long, A., 1975, The use of explosives on the Adelaar site. IJNA, 4: 345–352.McCarthy, M., 1996, Ship fastening: a preliminary study revisited, IJNA 25: 177–206.McCarthy, M., 2000, Iron and Steamship Archaeology: Success and Failure on the SS Xantho. New York.Meadows, G., 2000, Tawas Area Sediment Budget Study, Report OEL-0002. University of Michigan Ocean Engineering

Laboratory, Ann Arbor.Muckelroy, K., 1975, A systematic approach to the investigation of scattered wreck-sites. IJNA, 4: 173–190.Muckelroy, K., 1978, Maritime Archaeology. Cambridge.Murphy, L., 1990, 8SL17: Natural Site-Formation Processes of a Multicomponent Underwater Site in Florida. Southwest

Cultural Resources Center Professional Papers, 39. Santa Fe.Murphy, L., 1993, Windjammer site (FOJE 003). In L. Murphy (Ed.), Dry Tortugas National Park: Submerged Cultural

Resources Assessment. Southwest Cultural Resources Center Professional Papers, 45: 245–272.Murphy, L. & Jonsson, R., 1993, Environmental factors affecting vessel casualties and site preservation. In L. Murphy (Ed.),

Dry Tortugas National Park: Submerged Cultural Resources Assessment. Southwest Cultural Resources Center ProfessionalPapers, 45: 97–110. Santa Fe.

Ottawa Point Life Saving Station Log. United States Life Saving Service, Records of the United States Coast Guard [USCG],NARG 26.

Quinn, R., Breen, C., Forsythe, W., Barton, K., Rooney, S. & O’Hara, D., 2002, Integrated geophysical surveys of the Frenchfrigate La Surveillante (1797), Bantry Bay, Co. Cork, Ireland. Journal of Archaeological Science, 29: 413–422.

Reedy, J., 1991, Field procedures developed for a scattered, deeply buried site in a high energy location: the Whydah(WLF-HA-1), Cape Cod, Massachusetts, USA. IJNA, 20: 53–58.

Schiffer, M., 1987, Formation Processes of the Archaeological Record. Albuquerque.Souza, D., 1993. East Key Construction Wreck (FOJE 011) 1990 investigations. In L. Murphy (Ed.), Dry Tortugas National

Park: Submerged Cultural Resources Assessment Southwest Cultural Resources Centre Professional Papers, 45: 303–324.Souza, D., 1998, The Persistence of Sail in the Age of Steam. New York.Swayze, D., 2001, The Great Lakes Shipwreck File: Total Losses of Great Lakes Ships, 1679–2001. Lake Isabella, Michigan[3].Thompson, M., 2000, Graveyard of the Lakes. Detroit.Throckmorton, P., 1964, The Lost Ships. Boston.Tomalin, D., Simpson, P. & Bingeman, J., 2000, Excavation versus sustainability in situ: A conclusion on 25 years of

archaeological investigations at Goose Rock, a designated historic wreck-site at Needles, Isle of Wight, England. IJNA, 29:3–42.

van Tilburg, H., 1994, Zero-visibility diving on the Maple Leaf: tricks of the trade. IJNA, 23: 315–318.Wadsworth, J., 1975, Transport Mechanisms Operating on a Recurved Spit, Tawas Point, Michigan. MS Thesis, Dept of

Geology, University of Michigan, Ann Arbor.Ward, I., Larcombe, P. & Veth, P., 1999, A new process-based model for wreck-site formation. Journal of Archaeological

Science, 26: 561–570.

227

the archaeology of scattered wreck-sites: formation processes and

Documents