a systematic method for the identification of historic era shipwrecks · · 2011-06-23a...

A Systematic Method for the Identification of Historic Era

Shipwrecks

The Bark Cortland

(Photo: Center for Archival Collections, Bowling Green State University)

David M. VanZandt

Graduate Diploma Maritime Archaeology

Graduate Certificate Maritime Archaeology

Bachelor of Science Nuclear Engineering

Department of Archaeology

Flinders University

Thesis Advisor: Associate Professor Mark Staniforth

November 2009

ii

Table of Contents

Table of Contents............................................................................................................. ii

List of Figures ................................................................................................................. vi

List of Tables.................................................................................................................. vii

Declaration of Candidate............................................................................................. viii

Abstract ........................................................................................................................... ix

Acknowledgements.......................................................................................................... x

Chapter 1 Introduction ................................................................................................... 1

Chapter 2 Shipwreck Identification and Currently Employed Methods................... 6

Why Shipwreck Identification is Important .................................................................. 6

Shipwreck Misidentification ......................................................................................... 7

Two Methods used in Archaeology to Identify Shipwrecks ......................................... 8

O’Shea’s Method....................................................................................................... 8

Bayesian Probability.............................................................................................. 9

Ahlström’s Method.................................................................................................. 12

Conclusions ................................................................................................................. 14

Chapter 3 Historic Shipwreck Identification Method (HSIM) ................................. 15

Method Development Based on a Forensic Science Approach................................... 15

Evidence Types ........................................................................................................... 18

Direct Evidence ....................................................................................................... 18

Indirect Evidence..................................................................................................... 19

AM Ship Checklist Development................................................................................ 20

AM to PM Shipwreck Data Comparison Categories .................................................. 21

Dating ...................................................................................................................... 21

Vessel Construction and Rigging ............................................................................ 22

Cargo and Cargo Handling Equipment ................................................................... 22

Crew Personal Effects ............................................................................................. 23

iii

Location................................................................................................................... 23

Condition of Shipwreck and Site............................................................................. 23

AM to PM Shipwreck Data Comparison Scoring ................................................... 24

Shipwreck Identification Matrix ................................................................................. 24

Final Shipwreck Identification Scoring................................................................... 25

Conclusions ................................................................................................................. 26

Chapter 4 Case Studies ................................................................................................. 27

Case Study 1: The Griffon Cove Wreck ..................................................................... 27

Case Study Introduction .......................................................................................... 27

Shipwreck Site......................................................................................................... 27

PM Archaeological Data ......................................................................................... 29

AM Historical Research Data.................................................................................. 31

AM to PM Data Evaluation and Scoring................................................................. 36

Dating .................................................................................................................. 36

Vessel Construction and Rigging ........................................................................ 37

Cargo and Cargo Handling Equipment ............................................................... 37

Crew Personal Effects ......................................................................................... 37

Location............................................................................................................... 38

Condition of Shipwreck and Site......................................................................... 38


Conclusions ............................................................................................................. 39

Case Study 2: The Beaufort Inlet shipwreck............................................................... 40


Shipwreck Site......................................................................................................... 40


Vessel Construction and Rigging PM Data......................................................... 43

Cargo and Cargo Handling Equipment PM Data ................................................ 45

Crew Personal Effects PM Data .......................................................................... 49



Dating .................................................................................................................. 52

iv




Location............................................................................................................... 54



Conclusions ............................................................................................................. 56

Case Study 3: The Bark Cortland ............................................................................... 57


Shipwreck Site......................................................................................................... 58


AM Historical Research Data on Cortland ............................................................. 61

AM to PM Data Evaluation..................................................................................... 66

Dating .................................................................................................................. 66




Location............................................................................................................... 68



Conclusions ............................................................................................................. 70

Case Study 4: The Sidewheel Steamer Anthony Wayne.............................................. 71


Shipwreck Site......................................................................................................... 72




Dating .................................................................................................................. 82




v

Location............................................................................................................... 83



Conclusions ............................................................................................................. 85

Chapter 5 Discussion of Results ................................................................................... 86

Case Study 1................................................................................................................ 86

Case Study 2................................................................................................................ 88

Case Study 3................................................................................................................ 91

Case Study 4................................................................................................................ 92

Conclusions ................................................................................................................. 94

Chapter 6 Conclusions .................................................................................................. 95

Appendix 1: AM Historical Data Worksheet for Wooden Ships.............................. 97

Appendix 2: Shipwreck Identification Matrix with Rating Key............................. 111

References .................................................................................................................... 113

vi

List of Figures

Figure 1: Location of Griffon Cove Wreck..................................................................... 28

Figure 2: Map of the Niagara Area Drawn in 1688 ........................................................ 33

Figure 3: Proposed Rig of Griffon by Joy ....................................................................... 34

Figure 4: Location of the Beaufort Inlet Shipwreck........................................................ 41

Figure 5: Beaufort Inlet Shipwreck Site Plan circa 1998................................................ 42

Figure 6: Location of the Bark Cortland......................................................................... 57

Figure 7: Sidescan Image of Cortland Wreck Site.......................................................... 58

Figure 8: Scroll Head of the Bark Cortland .................................................................... 59

Figure 9: Cortland Official Enrollment Document ......................................................... 64

Figure 10: Only Known Picture of the Bark Cortland .................................................... 66

Figure 11: Location of Anthony Wayne........................................................................... 71

Figure 12: Sidescan of Anthony Wayne Wreck Site........................................................ 72

Figure 13: Pittman Arm Linkage of Anthony Wayne ...................................................... 74

Figure 14: Bow of Anthony Wayne ................................................................................. 75

Figure 15: Anthony Wayne Site Plan............................................................................... 75

Figure 16: Anthony Wayne Crosshead Linkage .............................................................. 76

Figure 17: Anthony Wayne Poppet Steam Valve and Linkage........................................ 77

Figure 18: Anthony Wayne High Pressure Steam Engine Cylinder Head....................... 77

Figure 19: Wood Cut of Anthony Wayne ........................................................................ 79

Figure 20: Scaled Drawing of Anthony Wayne ............................................................... 80

vii

List of Tables

Table 1: Shipwreck Identification Matrix ....................................................................... 25

Table 2: Griffon Shipwreck Identification Matrix .......................................................... 39

Table 3: Beaufort Inlet Shipwreck Cannon Data ............................................................ 46

Table 4: La Concorde Shipwreck Identification Matrix ................................................. 55

Table 5: Cortland PM Archaeological Data.................................................................... 60

Table 6: Cortland AM/PM Data Comparison................................................................. 67

Table 7: Cortland Shipwreck Identification Matrix........................................................ 69

Table 8: Anthony Wayne PM Archaeological Data......................................................... 78

Table 9: Anthony Wayne AM/PM Data Comparison ...................................................... 82

Table 10: Anthony Wayne Shipwreck Identification Matrix ........................................... 84

viii

Declaration of Candidate

I certify that this thesis does not incorporate without acknowledgment any material

previously submitted for a degree or diploma in any university; and that to the best of

my knowledge and belief it does not contain any material previously published or

written by another person except where due reference is made in the text.

_________________________________

David M. VanZandt

November 2009

Flinders University

Adelaide, South Australia

ix

Abstract

This thesis will develop a systematic method for the identification of historic era

shipwrecks based on traditional forensic science techniques. These traditional techniques

are used in the identification of human remains when the application of advanced

forensic techniques, such as DNA analysis, is not possible or practical.

The thesis will also demonstrate that developed method adheres to the principles

of the scientific method and can be easily applied to the shipwreck identification

process. The application of the method is illustrated in four case studies of shipwreck

identification.

The identification of a shipwreck allows it to be placed in the overall marine

landscape, or mariscape, and to synergistically make a contribution to the historical

perspective instead of being just an isolated island of archaeological and historical data

unto itself.

x

Acknowledgements

I would like to thank the following individuals for their support, encouragement,

and/or advice. Without their help this thesis would never have been completed.

My thanks go out to Associate Professor Mark Staniforth and Ms Jennifer

McKinnon for providing an ‘old country boy engineer’ with the education and skills

necessary to become a successful maritime archaeologist. I would also like to thank

them for their ideas, guidance, and support in regard to this thesis.

Many people went out of their way and gave their time unselfishly to help me

develop this thesis and accomplish this goal. Claire Dappert, PhD candidate, Flinders

University, Carrie Sowden, Nautical Archaeologist, Great Lakes Historical Society,

Christopher Horrell, Marine Archaeologist, Mineral Management Services, Larry

Murphy, Chief, Submerged Resources Center, National Park Service, Nathan Richards,

East Carolina University, Brad Rogers, East Carolina University, Pat Labadie, Historian,

Thunder Bay National Marine Sanctuary, and Chuck Meide, Director, Lighthouse

Archaeological Maritime Program (LAMP) all provided invaluable support in the

development of the premise and, in addition, provided examples and references to

misidentified shipwrecks and their personal views on this topic.

Special thanks go out to John O’Shea, Professor Anthropology, University of

Michigan for making his time available to discuss his statistical identification method

for multiple shipwrecks, the importance of shipwreck identification, and the shipwreck

identification process. I would also like to thank Hristina S. Lekova, Forensic Scientist,

Parentage and Identity Supervisor for the Cuyahoga County Coroner’s Office, for her

time and recommendation regarding the development of the Historic Shipwreck

Identification Method (HSIM) based upon traditional forensic identification methods.

xi

To Jason Raupp and Ian Moffat, many thanks for proof reading, editing, and

providing encouragement and ideas to help make my thoughts appear coherent on paper.

Thank you very much for your time, energy, expertise, and friendship.

To Jim Paskert, Chief Researcher, Cleveland Underwater Explorers (CLUE),

many thanks for putting up with my faulty use of the English language. Your patience,

editing, and advice in the construction of this thesis are sincerely appreciated. It would

not have come out as well as it did without your help.

My heartfelt thanks go out to my team, the Cleveland Underwater Explorers,

who make discovering, documenting, and diving shipwrecks a dream come true.

Finally, I would like to thank my family, Stephanie, Shelby, and Toni, for their

unselfish support of this endeavor. Without their support this thesis would never have

happened. Thank you again for putting up with me for all of these years.

1

Chapter 1 Introduction

What is the importance of identifying and placing a name on a shipwreck? The

unidentified and nameless shipwreck is a singular entity, a self-contained site, a time

capsule, if you will, of what happened at the time of wrecking. Identification allows the

collected data to be evaluated, not only as it relates to the shipwreck as a singular entity,

but also as it relates to the broader context of the surrounding maritime landscape or

mariscape. The name of a ship provides access to its associated historical documentation

and establishes the historical era in which the ship operated.

At times, the process of shipwreck identification lacks a methodological

framework for data collection, data analysis, and presentation of results. The process

may or may not be transparent in its application when reported in the literature. Thus a

methodological process is needed in order to achieve this transparency, to reach a

sustainable conclusion as to a shipwreck’s identity, and to remove as much of the

subjectivity involved in data interpretation and analysis as possible. Such a

methodological approach can also assist in preventing the incorrect identification of a

shipwreck.

The archeological identification of shipwrecks is often conducted without

appropriate use of the scientific method and therefore lacks rigor. Some shipwrecks have

been identified by utilizing only one piece of archaeological evidence, as in the case of

the Napoleonic brick Mercure (Beltrame & Gaddi 2002, p. 70). The identification was

based solely on the discovery of a single carronade marked with a French foundry stamp

and the date of1806. While the identification of the shipwreck on this occasion was

correct its finders made no use of supporting historical and archaeological evidence.

That supporting evidence being the wreck’s location, an analysis of the artifacts

recovered relative to a Napoleonic brick, and a comparison of the ship’s remains with

known construction techniques of the period. A methodological analysis of the available

archaeological data in comparison with the available historical data would have

2

presented a much stronger case rather than just relying on a single artifact to determine

the wreck’s identity.

In other situations, where an appropriate process is used to identify a shipwreck,

the process often lacks transparency in its application. The identification of William

Salthouse, as reported by Staniforth and Vickery (Staniforth & Vickery 1984, p. 31),

concluded that ‘The archaeological investigation categorically established the identity of

the wreck site as the William Salthouse’. While this conclusion is undoubtedly correct

based on the wealth of supporting evidence collected and analyzed against the historical

record, the report fails to show how this process was applied to support the conclusions.

There are other examples where the identification process used to identify a

shipwreck was rigorous and scientific. The identification of the Danish Frigate Mynden

(Auer 2004, p. 265) demonstrates a methodological, scientific approach to the

identification using archaeological evidence in conjunction with the historical record.

Based on dendro-chronological dating of the ship’s wooden remains, Auer conducted

extensive time frame focused, archival research in the Danish National Archives and

determined a possible candidate for the wreck. Careful comparison of the archaeological

and historical data led to the conclusion that the wreck was that of Mynden.

A second example is that of HMS Ready. As reported by Milstead Post (Milstead

Post 2007, p. 89), though this shipwreck had been misidentified as Dorthea careful

archival research along with interviews of local residents, when compared to the

archaeological data led to the correct identification.

There are many reasons for the misidentification of a shipwreck site including:

rushing to judgment during the identification process, failing to collect or consider all

available evidence, focusing on a research and/or field search predisposition relative to

the identification, or manipulating the accumulated data, whether consciously or

subconsciously, to fit a preconceived identity despite the lack of sufficient

archaeological and/or historical evidence. Additionally there nay not be enough

3

archaeological and/or historical evidence available to make identification possible.

While the identification of a shipwreck is important to place the wreck in the

surrounding mariscape, misidentification of a shipwreck misplaces the wreck in the

surrounding mariscape and damages the historical record by interpreting the

accumulated data using a false assumption.

This was the case in the misidentification of the British merchant ship Western

Empire. The shipwreck was discovered during a pipeline survey in the Gulf of Mexico

in the early 1980’s and reported to the Mineral Management Service (MMS) (Levin

2006, p. 1). The wreck site was surveyed in 1999 in a cooperative project between

MMS, Deep Marine Technologies, and Texas A&M University. The results obtained

from this survey were compared against MMS’s, GIS-based shipwreck database and the

wreck was determined, with a high degree of confidence on the basis of its location, to

be that of the British merchant ship Western Empire (Horrell 2009).

The shipwreck identified as Western Empire then became the thesis topic for

Joshua Aaron Levin of Texas A&M University (Levin 2006, p. i). Levin’s thesis,

entitled ‘Western Empire: The Deep Water Wreck of a Mid-Nineteenth Century Wooden

Sailing Ship’, focused on the history of the ship and the deep water survey of the vessel

(Levin 2006, p. 1).

Further research into the historical record determined that the shipwreck was not

Western Empire (Horrell 2009). This research revealed that Western Empire sank and

was salvaged off the Eastern Coast of Florida (Horrell 2009). The identity of the

shipwreck misidentified as the Western Empire remains unknown. This demonstrates

why a methodological approach and critical analysis of both the available archaeological

data and the historical record is necessary in the shipwreck identification process.

This thesis will develop a methodological framework and a systematic method

for the identification of historic era shipwrecks. The framework is based on evaluating

the available archeological and historic data using the principles developed by forensic

4

science for the identification of deceased individuals (Baraybar 2008, p. 533). The

forensic science identification process has been used to identify individual and multiple

victims from mass graves discovered in Kosovo, Peru and other countries (Baraybar

2008, p. 533). It is normally employed during the investigations of human rights

violations where enough time has elapsed to allow heavy or complete decomposition of

the cadaver or cadavers (Baraybar 2008, p. 533). The decomposition process usually

renders the cadaver unrecognizable and requires the use of forensic anthropology or

archaeology techniques to acquire Postmortem (PM) data (Baraybar 2008, p. 533). This

decomposition process is analogous to shipwreck sites. The forensic identification

process involves hypothesizing an identity or identities for the individual cadaver,

collecting Antemortem (AM) data for the proposed individual, then categorically

comparing the AM data to the PM data (Baraybar 2008, p. 534). The data comparison is

assessed relative to its consistency or inconsistency to each hypothesized identity

rendering a positive or negative identification (Baraybar 2008, p. 537).

Borrowing from this method of identification, an initial hypothesis or hypotheses

about the identity of a shipwreck is used to develop the AM data from the historical

record. The AM data is then compared with the PM data recovered from archaeological

investigation. The data is then categorically evaluated for its consistency or

inconsistency to each hypothesized ship posited for identification and a positive or

negative identification is determined.

It is apparent that shipwreck misidentification happens and that it muddies the

historical record with suppositions that masquerade as facts. That is why there is a clear

need to take a methodological and scientific approach to the identification process and to

postpone suggesting a probability relative to the speculated or hypothesized identity

until the process has been completed. The proposed identification method is based on

the scientific method of data evaluation and is easily applied to identifying individual

shipwrecks. This is not to say that the proposed method is immune to subjectivity or

erroneous data which can still creep into the process intentionally or unintentionally, but

it makes these errors more visible when they have been systematically evaluated in a

5

process that is transparent and provides a framework in which the data can be easily

reanalyzed by others.

Chapter 1 provides a brief introduction to the identification topic. Chapter 2

discusses the importance of shipwreck identification and several additional examples of

shipwreck identification and misidentification along with two methods that have been

used in archaeology. Chapter 3 develops the proposed method based on the forensic

science method and discusses evidence types together with the development of

checklists for both historical and archeological data with evaluation categories and data

evaluation techniques. Chapter 4 contains four case studies to which the developed

method for identification is applied. The first case study is that of LaSalle’s’ ship Griffon

and its numerous reported discoveries, including an evaluation of the discovery of a

wreck found in Griffon’s Cove. The second case study explores the identity of the

Beaufort Inlet shipwreck purported to be Queen Anne’s Revenge. The third case study

evaluates the identity of the recently discovered bark Cortland. And the final case study

examines the identity of the recently discovered wreck of the sidewheel steamer Anthony

Wayne. Chapter 5 contains a discussion of the methods application and the results from

the case studies and Chapter 6 assesses the strengths and weakness of the method and

whether or not it has been successfully adapted from the forensic science approach.

6

Chapter 2 Shipwreck Identification and Currently Employed Methods

Why Shipwreck Identification is Important

Why place a name on a shipwreck? What makes the identification important and

why care about it? The moment a ship sinks it effectively becomes a time capsule of the

material culture associated with that time period. A wealth of archaeological information

including ship construction data, cargo type, sailing conditions for the crew, and, if a

warship, types and kinds of armament, can be found within the boundaries of a

shipwreck site. All of this important data is interpreted in a very limited cultural and

historical context. Placing a name on a shipwreck opens up entirely new avenues for

data interpretation. Identification places the shipwreck in the broader context of the

mariscape thus permitting and encouraging data interpretation beyond the limits of the

shipwreck as a singular entity. This allows the shipwreck to be examined not just locally

but globally with respect to maritime culture and history.

The importance of this global aspect of data interpretation suggests that a

shipwreck should be examined, in so far as possible, within the context of the entire

mariscape. Archaeology seeks to understand past human activity based on the

investigation of material remains. Placing the ship within the maritime community

provides a much better understanding of how the ship operated, the people who operated

it, and the ship’s role within the community. Additionally, placement within the

community permits the study of the crew, their families and related socio-economics and

social behaviors. This new venue for data interpretation stimulates numerous additional

questions such as: how did the loss of this ship impact the community?; what cargo was

the ship carrying and why?; was the ship smuggling or engaged in some other

clandestine behavior?; what was the cultural make up of the crew and officers? These

few examples demonstrate the importance of accurate shipwreck identification in the

interpretation of the archaeological record.

7

Shipwreck Misidentification

Shipwreck misidentification takes on many shapes and forms spanning the

gambit from naivety to personal greed. Robert Marx, a treasure hunter and author,

reportedly found the wreck of USS Monitor while snorkeling off Cape Hatteras in the

1950’s and purportedly stuck a soda bottle into one of its guns (JP Delgado, 2009, pers.

comm., 16 January). Monitor was later discovered in deeper water in 1979. The wreck

Marx found was probably that of another Civil War era wreck, USS Oriental (JP

Delgado, 2009, pers. Comm., 16 January). Marx, knowing that the wreck of Monitor

was in that area and eager to promote his treasure hunting activities, probably did not

hesitate to name his discovery Monitor.

The Beaufort Inlet shipwreck, purported to be the remains of Queen Anne’s

Revenge, flagship of the infamous pirate Blackbeard, is an example of shipwreck

identification (intentional misidentification) for the purpose of stimulating interest and

securing grant funding for a project that otherwise would have received little attention

and likely little funding. David D. Moore (Moore 2005, p. 338), Curator of Nautical

Archaeology at the North Carolina Maritime Museum states:

They additionally claim that our research efforts towards identifying the wreck

as Queen Anne’s Revenge are perhaps ‘politically-motivated’ (p.26). We assume

that they are accusing the project of using the possibility that the site is

associated with Blackbeard to somehow trick the general public into

maintaining a high level of interest, and various funding sources into providing

monies to continue the project. To this we will readily plead guilty and counter

with the question: why not?

The identification (misidentification) of this wreck is discussed further in Chapter 4.

These are just a few examples of wreck misidentification. No matter what the

motive or cause, misidentification leads to erroneous data becoming part of the historical

8

record. Subsequent investigations and research based on this data then leads to a further

corruption of the historical record.

Two Methods used in Archaeology to Identify Shipwrecks

Two documented methods of shipwreck identification have been used in the past.

The first method was developed by John O’Shea and is based on the identification of

individual ships in an assemblage of shipwrecks (O'Shea 2004). The second method was

developed by Christian Ahlström and is based on the comparison of historical to

archaeological data to identify individual shipwrecks (Ahlström 1997).

O’Shea’s Method

O’Shea’s method is described in ‘The identification of shipwreck sites: a

Bayesian Approach’ (O'Shea 2004, p. 1533) and calculates a probability of identification

or degree of belief employing Bayesian theory in an attempt to quantify the identities of

an assemblage of shipwrecks using historic and archaeological data. The method uses a

narrow range of input data (O'Shea 2004, p. 1536) consisting of:

• Maximum observed length

• Estimated gross tonnage – calculated from various scantling tables

• Propulsion system – sail or steam

• Cargo

• Location of the specific wreck site

This data, along with subjectively determined prior probabilities, is used to

calculate a posterior probability of identification for each individual vessel.

9

Bayesian Probability

Bayesian probability is based upon Bayes Theorem developed by the Reverend

Thomas Bayes in the mid-1700s. The theorem was published three years after Bayes’

death in an work entitled ‘An essay towards solving a problem in the doctrine of

chances’ in the Philosophical Transactions of the Royal Society of London in 1763

(Withers 2002, pp. 553-4).

Bayes’ theorem mathematically stated is:

P(H|E) = P(E|H) * P(H) / P(E)

Where:

H represents a specific hypothesis

P(H) prior probability of H being true

P(E|H) conditional probability of seeing evidence E if H is true

P(E) marginal probability of E of seeing new evidence for all hypotheses

P(H|E) posterior probability of the hypothesis H based on the evidence E

P probability ranging from 0 being false to 1 being true

This equation states that for a given hypothesis H and specific evidence E a

posterior probability P(H|E) can be calculated based on the conditional probability of

seeing evidence P(E|H) if the hypothesis is true times the prior probability of the

hypothesis being true P(H) divided by the marginal probability of seeing evidence based

on all the hypotheses.

The posterior probability P(H|E) represents a degree of belief or confidence in

the hypothesis based on the evidence available. As further evidence is acquired the

posterior probability P(H|E) becomes the new prior probability P(H) and a new posterior

probability P(H|E) is calculated. If enough evidence is available, this iterative process

develops a value for P(H|E) independent of the original P(H). Conversely, when only a

10

very limited amount of evidence is available the process remains very subjective as the

calculated posterior probability P(H|E) is heavily biased by the initial choice of the prior

probability P(H).

To calculate the probability of a shipwreck being identified based on observed

archaeological data compared with historical data obtained for a possible candidate ship

the equation can be rewritten as follows:

P(H|E) = P(E|H) * P(H) / P(E|H) * P(H) + P(E|not H) * P(not H)

Where:

H hypothesized ship identity

E evidence from physical or calculated data

P(H) prior probability of the identity being the hypothesized identity H

inferred before E became available

P(not H) prior probability of the identity not being the hypothesized identity H

inferred before E became available

P(E|H) conditional probability of seeing evidence E if the hypothesized identity

H is true

P(E|not H) conditional probability of seeing evidence E if the hypothesized identity

ship H is not true

P(H|E) posterior probability of the hypothesized ship identity H being true based

on the evidence E

P probability ranging from 0 being false to 1 being true

This is the equation that O’Shea used to calculate the probability of identification

for individual shipwrecks in an assemblage of shipwrecks. The problem with using this

approach is that it is only a once through method based on the data available. Since all of

the archaeological data is first gathered and then subsequently passed once through this

equation, the resulting posterior probability P(H|E) is heavily influenced by the prior

probability P(H) which was subjectively determined beforehand.

11

Examining the equation we can see that, when P(H)=0: P(H|E)=0 and, when

P(H)=1: P(H|E)=1. Thus, if we absolutely believe the ship has been positively identified,

the equation strongly supports this belief. Likewise if we are absolutely sure it is not the

ship and assign a value of 0 based on that belief, the equation provides the answer we

expect – the wreck does not represent the remains of the historic ship posited. To avoid

this problem we must assign an initial prior probability value greater than 0 and less than

1. To further demonstrate the impact of this subjective process on the posterior

probability we will assign the following values to a shipwreck assuming, for the sake of

the example, that we have a strong positive belief in the prior identification probability

and that the gathered evidence found supports this:

P(H) = 0.75, a strong belief that this is the ship

P(not H) = 1-0.75 = 0.25

P(E|H) = 1, 100% chance of seeing evidence E if this is the ship

P(E|not H) = 0.5, a 50% chance of seeing evidence E if this is not the ship

Thus:

P(H|E) = 1 * 0.75/[(1 * 0.75) + (0.5 * 0.25)] = 0.86

The posterior probability P(H|E) of this being our hypothetical ship is 86 percent

based on our initial prior belief of 75 percent.

Using the same data we can take a more skeptical approach by assigning the

following probabilities:

P(H) = 0.25, a skeptical belief that this is the ship

P(not H) = 1-0.25 = 0.75

P(E|H) = 1, 100% chance of seeing evidence E if this is the ship

P(E|not H) = 0.5, a 50% chance of seeing evidence E if this is not the ship

Thus:

P(H|E) = 1 * 0.25/[(1 * 0.25) + (0.5 * 0.75)] = 0.40

12

The posterior probability P(H|E) of this being our hypothetical ship is 40 percent

based on our initial prior belief of 25 percent.

Despite using the same evidence to calculate our posterior probabilities, two

widely varying answers (86 percent and 40 percent) result demonstrating the inherent

problem with using this type of analysis on a ‘once through basis’. The outcome is

greatly influenced by the assigned prior probability P(H) which could be manipulated,

consciously or unconsciously, to skew the results one way or the other.

This method does not lend itself readily to the identification of an individual

shipwreck because the data set for each individual wreck is variable and requires the

generation of a new set of equations. Also, the method fails to manage the subjective

influence embedded in the prior probability selection and evidence evaluation processes.

The method works well when evaluating an assemblage of shipwrecks to determine an

individual identification out of the group but performs poorly when trying to identify an

individual shipwreck.

Ahlström’s Method

The method described by Ahlström’s in the book, ‘Looking for Leads:

Shipwrecks of the past revealed by contemporary documents and the archaeological

record’ (Ahlström 1997), is based on the comparison of archaeological and historical

data to determine the identity of a shipwreck. This process consists of five stages:

1. An analysis of the material and artifactual finds from the wreck in question.

2. An assessment of the historical situation at the central, provincial and local levels during

the time period to which the wreck can be assigned with reference to dateable finds and

analyses.

3. The establishment of the range of published and hand-written sources.

13

4. The construction of hypotheses concerning the identity of the wreck on the basis of what

can be logically deduced from the material and written evidence.

5. A test of the hypotheses. (Ahlström 1997, p. 33)

In the first stage, material and artifactual finds are divided into four groups and

then analyzed. The groups include: objects from the vessel itself, objects originating

from the cargo, vessel equipment that perform a function (navigation, defense, etc.), and

personal effects and property (Ahlström 1997, pp. 38-9). Using the analysis from stage

one, the second stage establishes time period boundaries for the historical documentation

search. The third stage identifies all the available historical resources for the time period

and where they can be found. The fourth stage uses these historical resources to

determine all of the possible candidate shipwrecks and then evaluates them against the

archaeological data to narrow the field of hypothetical candidates. The final stage

rigorously tests these candidates’ historical background against the archaeological data

to attempt to determine the identity of the wreck.

Ahlström’s method is a systematic way of comparing archaeological data to

historical data to determine the identity of a shipwreck. It relies very heavily on dating

the artifactual finds to establish a time period in which to do archival research. Ahlström

states, ‘The first and most important keys to establishing the identity of a vessel are finds

indicating its age. Without a date, it is obvious that no reliable identification is possible’

(Ahlström 1997, p. 37).

Sometimes it is not possible to reliably date finds, or the work performed on the

site is insufficient to establish a Terminus ante quem, or to even narrow the time period

enough to even begin archival research. This does not, however, prevent a reliable

shipwreck identification from being made. In lieu of artifactual dating, a wreck’s

location, dimensions, construction, and cargo can all be used to determine the initial

identification candidates and establish the time period parameters for historical research.

The data obtained can be compared to historical wreck lists, wreck data bases, and other

14

archival sources. Local and oral histories can also be used as a starting point for

selecting candidates.

Conclusions

This chapter has described the two currently employed shipwreck identification

methods used in archaeology. It has also evaluated the applicability of these methods to

the process of identification of a single shipwreck based on the availability of

archaeological and/or historical data. The O’Shea method is not directly applicable to

the identification process of a single wreck because the method is probability based and

requires the use of prior probabilities. The prior probabilities in this method are

subjectively determined and greatly influence the identification outcome when they are

applied in a single wreck identification case as opposed to determining a single wrecks

identity in a collection of wrecks. The Ahlström method is useful in the identification of

single shipwrecks. It does, however, rely too heavily on establishing narrow time period

parameters in which archival research must attempt to collect the necessary historical

data. The restrictions and limitations of the two described methods demonstrate the need

for a systematic method not dramatically affected by such problems that can be used to

evaluate the available archeological and historical data to determine the identity of

historic era shipwrecks.

15

Chapter 3 Historic Shipwreck Identification Method (HSIM)

Method Development Based on a Forensic Science Approach

The proposed method for shipwreck identification, the Historic Shipwreck

Identification Method (HSIM), is based upon the framework used in forensic science for

body identification using ‘traditional’ forensic science methodology. This method is

normally applied when mass graves are encountered (i.e. during the investigation of

human rights violations) and the cadavers are heavily decomposed and rendered

unrecognizable thus requiring the participation of forensic anthropologists and

archaeologists (Baraybar 2008, p. 533). The method is generally used when local,

forensic laboratories, capable of performing advanced procedures such as DNA analysis,

are unavailable or the cost of such procedures is deemed prohibitive (Baraybar 2008, p.

533). ‘Traditional techniques generally consist of combining witness testimony, personal

effects and clothing, anthropological and dental data to corroborate or to exclude the

identity of the individual’ (Baraybar 2008, p. 533). This traditional forensic science

technique will be adapted to the identification of shipwrecks.

The forensic science method for body identification consists of five distinct

phases (Baraybar 2008, pp. 533-37).

1. Antemortem (AM) data collection

2. Postmortem (PM) data collection

3. Selection of possible identities

4. Antemortem/postmortem data comparison and evaluation

5. Identification, positive or negative

The collection of antemortem (AM) data, normally conducted during the

investigative phase of an incident, involves the gathering of data from a variety of

sources including official documents, personal documents and interviews with relatives

and friends. The data is collected systematically using checklists developed by various

16

organizations such as the International Committee of the Red Cross. In order to obtain as

much AM data as possible these checklists are tailored to the region in which the

investigations are taking place and are emic to the population being interviewed. The

checklists (forms) utilized for AM data collection are tested and refined over time to

assure that the data collected is the information needed by the scientists for proper

evaluation of the PM data (Baraybar 2008, p. 535).

The PM data collection phase consists of documenting the grave site through

forensic examination of the human remains, clothing, and any artifacts directly

associated with the body. Any artifact that was not directly associated with the body is

noted but is not artificially associated to it (Baraybar 2008, p. 536).

A clothing exhibit is used to select possible identifications. The victim’s clothing

and directly associated artifacts are displayed. This can be done in a public or private

setting. Relatives and friends are allowed to view the display(s) in the hope that they will

recognize the belongings. If a family or families recognize the belongings an attempt at

identification can be made (Baraybar 2008, p. 535).

Once the possible identification of a person has been proposed the

antemortem/postmortem data comparison is conducted. Both the PM and AM data files

are selected and the family is interviewed. During the interview process a number of

specialists are present. These normally include ‘one police officer specialized in

identifications, one anthropologist, one dentist, and one pathologist’ (Baraybar 2008, p.

537). They are present to ask specific questions of the family in relation to their field to

help with their assessment of the AM to PM data comparison.

After the interviews are completed each specialist scores the case on a scale from

1 to 3 as a good, medium or bad match to the AM to PM data. ‘Identification is carried

out by consensus and the pathologist is the only person who may accept or reject an

identification’ (Baraybar 2008, p. 537). Even though the scoring system is based on 1 to

3 there can be only ‘good’ or ‘bad’ data matches (Baraybar 2008, p. 537).

17

This systematic process will be modified and adapted to develop the HSIM. The

first modification will be the order in which the phases of the process occur. Since the

first data available from the site of an unidentified shipwreck is the PM data, the

collection of PM data is logically the first phase. Once the PM data has been collected

and analyzed the second phase, selection of possible identities, can begin. One or more

working hypotheses relative to the vessel’s identity are developed. The selection of

hypothetical identities leads to the third phase, AM data collection. The proposed

identities provide starting points for the historical record search and are used to develop

checklists for data collection. Once the AM data is collected it can then be compared and

analyzed against the PM data and a determination of identification of each hypothesis

made. A matrix with data consistency criteria is developed to aid in the data evaluation

and identification process.

The second modification to the forensic science approach changes the evaluation

criteria of the scoring range. The current scoring range of 1 to 3 (good, medium or bad

match of AM to PM data) is changed to a range of 0 to 3 (no data available, consistent

data, insufficient data, and clearly inconsistent data).

The third modification to the forensic science method of identification is the

development of a final identification matrix. This matrix is used to record the scoring of

the different data comparison categories and to determine a final identification rating

score for each hypothesized identification candidate.

The final modification to the forensic science approach allows the person(s)

making the identification attempt to be the final judge in the matter.

Thus for the purposes of the HSIM the hierarchy of the shipwreck identification

process has been changed to:

1. Postmortem (PM) data collection

2. Selection of possible identities

18

3. Antemortem (AM) data collection

4. AM/PM data comparison and evaluation

5. Identification, positive or negative

This process is consistent with the ‘scientific method’ and follows the basic

premise of:

• Posing the question of a shipwrecks identity

• Conducting the background research

• Constructing a hypothesis

• Testing the hypothesis

• Analyzing the data and drawing a conclusion

• Communicating the results

Evidence Types

There are two evidence types to consider when collecting PM and AM data:

direct and indirect evidence. These evidence types influence the comparison and

evaluation process phase of the identification process. It is important to note that the

skill, expertise, and thoroughness of the AM and PM data collection processes

ultimately control the value of any accepted or rejected identification.

Direct Evidence

Direct evidence is evidence based on fact or the observation of a fact. Relative to

ships or shipwrecks, an example of direct evidence would be the statement on the

enrollment papers of a particular vessel indicating the vessel had a scroll head or the

observation of a scroll head on a particular vessel by a person in port and reported in the

press or other historical document. This type of evidence can be used in the

19

identification process for a direct comparison of fact (AM data) with the observed data

found at a shipwreck site (PM data). In this particular case it is evidence that would be

hard to refute in the identification process. If the AM factual data matches the PM data

obtained from the wreck site, it is a good indicator of a factual match for that particular

piece of data. This is not to say that this piece of data is indeed true. In fact, the vessel

might have been involved in a previous accident or rebuilt after some time in service and

the scroll head subsequently removed or changed. Although it appears that the data

obtained is indeed true, there is the possibility that further research may be needed to

assess the ‘goodness’ and accuracy of AM data obtained in the initial data collection

activity.

This assessment of data and data collection methodology is readily apparent in

Case Study 4. Anthony Wayne was originally built with a vertical, high pressure, steam

engine (US Treasury Department 1838, p. 339). The archaeological data collected from

the site indicates that the wreck has a horizontally mounted steam engine. This is in

direct conflict with the initial AM data. Subsequent research determined that Anthony

Wayne was extensively rebuilt and the original, ‘vertical’, high pressure, steam engine

was replaced with a ‘horizontal’, high pressure, steam engine from the steamer

Columbus (Heyl 1956, p. 93; Sandusky Clarion 1849, p. 2; Thunder Bay National

Marine Sanctuary 2009b).

Indirect Evidence

Indirect evidence (or circumstantial evidence) is evidence that may reasonably

imply the existence or non-existence of a fact from another fact. An example of this type

of evidence is seen in Case Study 3. One of the pieces of evidence used to determine the

identification of the shipwreck Cortland was location. A theoretical location was

inferred from the direct evidence obtained from the AM data collection process

(Cleveland Leader 1868b, p. 3). The AM data showed that the steamer Morning Star,

which was involved in the accident that sank the bark Cortland, left the dock at a certain

20

time maintaining a certain speed and course (Cleveland Leader 1868b, p. 3). The time

the accident took place was also known along with the distance and direction Cortland

was from Morning Star when Cortland sank (McDonald 1958, pp. 311-3). This data was

used to calculate a dead reckoning position and establish an approximate location for the

final resting place of Cortland.

AM Ship Checklist Development

The development of a checklist for AM ship data can be general or very specific.

The checklist developed here is specific in nature. It takes into account the possible AM

data that would be available in official records from the Great Lakes region based on

vessel enrollments and surveys. Though this checklist is specific it can be easily adapted

to different ship types, time periods, and geographical regions as needed since the

majority of the information recorded is applicable to any ship.

These main categories were used to develop the AM checklist in Appendix 1

which is used to collect AM data for a specific ship:

• Vessel Name • Original Construction Data

• Cargo • Incidents

• Final sinking data • Crew

• Construction drawings • Imagery

• Inventory list • Manifest

• Provisions • Personal Belongings

• Models • Survivor and eyewitness accounts

21

AM to PM Shipwreck Data Comparison Categories

The following are the main categories used to compare the AM to PM data for

identification:

• Dating

• Vessel Construction and Rigging

• Cargo and Cargo Handling Equipment

• Crew Personal Effects

• Location

• Condition of Shipwreck and Site

These six categories are evaluated as follows:

Dating

The AM and PM data are compared for consistency in dating and scored

accordingly. This category is used if sufficient dating data is available from the

shipwreck site to establish a Terminus ante quem or to sufficiently date the remains to

establish a likely time period in which the ship operated and was lost. The AM data

gathered from historical research establishes the time period and date of loss for the ship

proposed for identification. This data is compared to the gathered archaeological data.

Dating the archaeological finds can be accomplished by a variety of means such as

stratigraphy, dedrochronology, coinage, artifact seriation, archeomagnetic (from the

galley hearth or in case of the accident being caused by a fire), date or date stamps on

artifacts or casting dates, etc. The accuracy of the data comparison is then scored

accordingly.

22

Vessel Construction and Rigging

The AM and PM data are compared for consistency relative to the vessel’s

construction and rigging and scored accordingly. This category is used if sufficient data

is available from both historical sources and shipwreck remains. If sufficient AM data is

available it is then compared and evaluated against the collected archeological data. The

process may include comparing the deck hardware found, number of masts, scantling

data, or fastening techniques to assure they are consistent with the acquired AM data.

The accuracy of the data comparison is then scored accordingly.

Cargo and Cargo Handling Equipment

The AM and PM data are compared for consistency relative to cargo and cargo

handling equipment and scored accordingly. This category is used if sufficient cargo and

cargo handling equipment data has been acquired for the ship proposed for

identification. The archaeological data collected from the wreck site is compared and

evaluated against the AM data for consistency and the accuracy of the data comparison

is then scored accordingly.

In certain circumstances this category may be eliminated from the identification

matrix. There is at least one documented instance, and probably many more, where the

cargo reported in the AM documentation did not match the cargo discovered at the

shipwreck site. The Canadian steamship Homer Warren, which sank in Lake Ontario in

1919, was reported to be carrying a cargo of coal (Kennard 2009, p. 132). After its

discovery and positive identification in 2003 its cargo was determined to be grain

(Kennard 2009, p. 134). This is a case where a ship, having plied the same route

carrying the same cargo for many years, happened to be in port and its standard cargo

was unavailable. To avoid the losses associated with laying idle or running light, the

ship was loaded with a different cargo.

23

Crew Personal Effects

The AM and PM data are compared for consistency concerning crew personal

effects and scored accordingly. This category is used if sufficient AM data about the

crew is discovered and crew personal effects are found at the wreck site. If demographic

information about the crew is established through AM research, the personal effects

found at the wreck site (if any) can be compared for consistency as to the crew

nationality and cultural backgrounds. The accuracy of the data comparison is then scored

accordingly.

Location

The AM and PM data are compared for consistency based on the location of the

shipwreck and scored accordingly. This category is used if sufficient location

information is discovered during the AM data collection process. The accuracy of the

data comparison is then scored accordingly.

In certain circumstances this category may be eliminated from the identification

matrix. Some anecdotal reports suggest the purposeful falsification of reported locations

or areas of operation in certain cases. These cases normally involved insurance claims

for the vessel and its cargo. Ships nearing the end of their useful operating lifetime were

often scuttled to collect the insurance money. In other cases the ships were operated

outside of the bounds of their insurance coverage. If they were reported lost, the owners

would claim they were operating on a certain route when truthfully they were not.

Condition of Shipwreck and Site

The AM and PM data are compared for consistency relative to the conditions of

the shipwreck and the shipwreck site and scored accordingly. This category is used if

sufficient AM data has been collected that reflect any damage that was sustained from

24

an accident, collision, salvage operation on the site, or any observations about the

condition of the wreck or the surrounding bottom recorded by salvage operatives. This

AM data is compared with the current condition of the wreck and wreck site. The

accuracy of the data comparison is then scored accordingly.

AM to PM Shipwreck Data Comparison Scoring

As discussed above, the data comparison scoring from the original forensic

science identification approach has been modified to encompass the data that would be

available for shipwreck identification needs. The detailed data comparisons are scored as

follows:

0 = Historical and/or archaeological data is not available or is insufficient to perform a

meaningful comparison.

1 = The historical and archaeological data are consistent in sufficient detail to establish

that they belong to the proposed ship and there are no irreconcilable discrepancies.

2 = The historical and archaeological data are not consistent in sufficient detail to

establish that they belong to the proposed ship.

3 = The historical and archaeological data are clearly inconsistent.

Shipwreck Identification Matrix

The Shipwreck Identification Matrix (Table 1) is the culmination of the

identification process. The AM/PM shipwreck data comparison scores are recorded and

a final score is determined for the accuracy of identification of the hypothesized ship

25

candidate. A complete copy of the Shipwreck Identification with rating key can be found

in Appendix 2.

Table 1: Shipwreck Identification Matrix

Final Shipwreck Identification Scoring

Final scoring is based on the AM/PM data evaluation scores for each category. It

reflects the highest score obtained in any of the evaluation categories. This score then

determines if a positive identification has been achieved.

The final identification rating scores are defined as follows.

1 - Identification: The historical and archaeological data in all categories where

comparisons are possible are consistent in sufficient detail to establish that they belong

to the proposed ship and there are no irreconcilable discrepancies. The shipwreck is the

ship proposed.

Shipwreck Identification Matrix Candidate Ship for Identification Ships Name Here AM to PM Shipwreck Identification Categories Data

Consistency Scoring

Dating




Location


Final Scoring (Highest number in right hand column)

26

2 - No conclusion: The historical and archaeological data in all or part of the categories

where comparisons are possible are not consistent in sufficient detail to establish that

they belong to the proposed ship.

3 - Exclusion: The historical and archaeological data in all or part of the categories

where comparisons are possible are clearly inconsistent. The shipwreck is not the ship

proposed.

Conclusions

This chapter discussed the development and the usage of the HSIM. The method

was developed from a ‘traditional’ forensic science method used in determining the

identification of human remains when advanced forensic scientific techniques are not

available. The forensic method was adapted for the use in the identification of

shipwrecks when sufficient AM and PM data are available. Scoring systems for data

comparison and final shipwreck identification were developed and guidelines for data

evaluation established.

27

Chapter 4 Case Studies

Case Study 1: The Griffon Cove Wreck

Case Study Introduction

Case Study 1 involves a shipwreck that was discovered in about 1835 by the

grandfather of Orrie Vail (Hundley 1984, p. 9) in what is today, Griffon Cove, Russel

Island, Ontario, Canada. The wreckage was examined in 1955 by Rowley Murphy, John

MacLean, and C. H. J. Snider and determined to be that of Griffon, the first ship to have

sailed the Upper Great Lakes (Murphy 1955, p. 236). Murphy (Murphy 1955, p. 232)

stated, ‘… we now do think that the wreckage recovered by Mr. Vail in August of this

year is from that ship.’ The wreckage was later examined in 1978 by Paul Hundley

(Hundley 1984, p. 15), a graduate student at Texas A & M University, to determine if

the remains were indeed that of Griffon. Hundley’s research determined that the

wreckage was ‘not’ that of Griffon (Hundley 1984, p. 40) but probably ‘a variation of a

Mackinaw boat, consistent with the type of vessels used on the Great Lakes during the

mid-1800’s’ (Hundley 1984, p. 48). Even though the wreckage has been proven not to

be that of Griffon, this case study will posit the candidate identification to be that of

Griffon and the HSIM, for the purpose of demonstration and exercise, will be applied

based on the PM archeological data available from the previous Griffon Cove wreckage

studies.

Shipwreck Site

The Griffon Cove shipwreck was first discovered by the grandfather of Orrie

Vail (Hundley 1984, p. 9) in MacGregor Cove (Hundley 1984, pp. 1-9), Find Out Island

(Murphy 1955, p. 236), Ontario, Canada about 1835 (Hundley 1984, p. 9). After the

examination of the remains by Murphy and at his suggestion, MacGregor Cove was

renamed Griffon Cove in 1956. Find Out Island has also since been renamed Russel

Island (no reference to date of name change). Russel Island is located about one mile

28

(1.6 km) northwest of Tobermory, Ontario, Canada, the northern-most town on the

Bruce Peninsula (Figure 1).

Figure 1: Location of Griffon Cove Wreck

(Hundley 1984, p. 9)

The greater part of the wreckage found at Griffon Cove was brought to the

surface by Vail in 1955, numbered, and taken by motor boat to Vail’s big boathouse on

Vail’s Point for storage (Murphy 1955, p. 237). After Vail’s death in 1977, the material

was passed to the Minister of Natural Resources (MNR), Canada and ultimately to Five

Fathom Provincial Park (FFPP) (Hundley 1984, p. 14). A subsequent investigation in

1978 by Hundley and the FFPP found more wreckage and artifacts (Hundley 1984, pp.

15-6). These were transferred to the FFPP facilities at Tobermory for cataloging,

conservation, and storage. A short follow-up investigation in 1979 recovered a number

of artifacts from the outer part of the cove (Hundley 1984, p. 16).

29

PM Archaeological Data

The wreckage recovered by Vail and described by Murphy (Murphy 1955, pp.

237-8) consisted of:

… the keel; part of the stem, including scarph; part of the stern-post; and about

2/3rds of the apron tying it into the keel. There were parts of at least 13 frames,

of several planks, including about eight feet of the garboard strake right aft at

the sternpost on the port side. This, when the sternpost and apron were fitted

easily into the keel, slipped into place perfectly. There was also a knee which

could possibly have tied together a side and lower timber of the transom; also a

piece about 16 feet in length and about six inches wide which certainly appeared

to have been part of a bilge stringer. All of the timbers or pieces of plank were

of sound white oak, considerably worn down by attrition, grinding on the rocks,

etc. If there was any piece used in her bulwarks and ceiling, it had quite

disappeared.

Vail also recovered several additional artifacts including an iron padlock and a small

wooden keg (Hundley 1984, p. 16).

The wreckage and artifacts recovered in 1978 by Hundley and the FFPP staff

during the excavation phase of the project included 131 artifacts found at the cove

entrance. Ninety three of the one hundred thirty one artifacts were European ceramic

sherds (Hundley 1984, p. 16). Deeper water outside of the cove yielded five timber

frames, two pieces of planking, and the gripe (Hundley 1984, p. 16). The dredged area at

the wreck site produced 244 artifacts of which 234 were iron nails and nail fragments

(Hundley 1984, p. 16). The 1979 investigation of the outer portion of the cove produced

a number of clay pipe stems, ceramic sherds (one with a maker’s mark of ‘Baker and

Son’), and a padlock similar to the one Vail had found earlier (Hundley 1984, p. 16).

30

Analysis of the hull construction conducted by Hundley (Hundley 1984, p. 52)

concluded that: ‘The Griffon Cove vessel had an extreme length overall of 44 ft 9 in

(13.6 m), a beam of 14 ft 7 in (4.4 m) and a reconstructed depth at the sheer of 3 ft. (0.9

m.) at the midships, 3 ft. 5 in. (0.91 m.) in the bow, and 4 ft. 9 in. (1.4 m.) in the stern’.

The calculated tonnage, using old carpenter’s measure, is 19 tons (17.2 tonnes).

The keel is constructed from a single log of white oak. There is no mast step

present, but it is notched for floor timbers (Hundley 1984, pp. 32-3). A mortice cut is

found 13 feet (3.96 m) ahead of the stem knee which is placed too far aft for a mast and

most likely used for a stanchion or boom crutch (Hundley 1984, p. 33). Also noticed, the

garboard is not rabbeted into the keel (Hundley 1984, p. 36). On the stempost there is

another mortise which suggests the placement of a foremast (Hundley 1984, p. 36).

‘There were no remains recovered from the Griffon Cove vessel indicating there was a

deck on this ship’ (Hundley 1984, p. 40). No ballast stones were recovered from the site.

The discovered artifacts described in the Hundley report (Hundley 1984):

…indicate a date, based on the ceramic types, of 1840-1860. One piece bore a

maker's mark, Barker & Son, a British kiln producing only from 1850 to 1860.

The only other datable artifact was a lock identical in shape and dimensions to a

lock found on the site by Vail. The Vail lock bore a mark of a crowned "GR"

and the word "Patent" on the keyhole cover. The escutcheon refers to George III

or George IV in whose reign the patent was granted. This gives a date range of

1790-1830 for the date of first manufacture. The lock found by FFPP was

identical to this except it was missing the keyhole cover.

31

AM Historical Research Data

Le Griffon or Griffon was the first vessel to sail the upper Great Lakes. Its keel

was laid 22 January 1679 (Hennepin 1903, p. 90) on the orders of René Robert Cavalier,

Sieur de la Salle who intended to use the vessel to explore the upper Great Lakes for

trade and to search for a western route to China (Mansfield 1899, p. 79). The

construction site chosen for Griffon was located about two leagues above Niagara Falls

just off the Niagara River on Iroquois Indian land (Anderson 1901, p. 21). Upon its

launch ‘the Iroquois Indian were amazed, being unable to comprehend how the

Frenchmen could so easily build so large a canoe of wood, although the craft was of

only about forty-five tons [forty-one tonnes]’ (Anderson 1901, p. 27). There is a

reference in Hennepin (Hennepin 1903, p. 94) to Griffon being ‘60 tuns’ along with a

foot note that states ‘In his Louisiane (ed. 1683, p. 46), Hennepin says that it was forty-

five tons [forty-one tonnes], - Ed.’ (Hennepin 1903, p. 94).

There are questions as to the ‘originality’ of Hennepin’s writings as much of the

content appears to be a literal word for word restatement of la Salle’s official report to

the Minister of Marine. The assertion of plagiarism on the part of Hennepin is brought

up by Pierre Margry (Anderson 1901, p. 5) who officially published la Salle’s writings

during the period 1876-1886, and included these allegations in his first footnote:

It was, moreover, a matter of real interest to ascertain the plagiarisms of the man

who was afterwards to endeavor to deprive the discoverer of the honor due to

his labors; and if it was on this account curious to note the similarities between

Father Hennepin's book and our document, it was not less curious to remark the

differences between this same document and the text of La Salle's letters, from

which it must have been drawn, but without depriving them of their intimate and

confidential character.

Very little is known of the actual construction of Griffon. As stated above, it was

a sailing vessel of 45 tons (41 tonnes). Hennepin (Hennepin 1903, p. 93) described the

32

vessel as decked: ‘…our Men, who immediately quitted their Cabins of Rinds of Trees,

and hang’d [sic] their Hammocks under the Deck of the Ship.’ The ship was armed

(Anderson 1901, p. 29) and la Salle stated that it had ‘seven small pieces of small

cannon.’ A more detailed description by Hennepin (Hennepin 1903, p. 102) stated:

We were very kindly receiv'd [sic], and likewife [sic] very glad to find our Ship

well rigg'd [sic], and ready fitted out with all the Neceffaries [sic] for failing

[sic]. She carry'd [sic] five fmall [sic] Guns, two whereof were Brafs, and three

Harquebuze a-crock1. The Beak-head was adorn'd [sic] with a flying Griffin,

and an Eagle above it; and the reft [sic] of the Ship had the fame [sic]

Ornaments as Men of War ufe [sic] to have.

The rigging of Griffon is unknown. There are several hints to the rigging in both

la Salle’s and Hennepin’s descriptions of a storm they encountered while sailing up Lake

Huron. La Salle (Anderson 1901, p. 37) stated:

…he was caught in a furious westerly gale, which obliged him to tack under

foresail and trysail, and finally to bring her to until daylight.

On the 26th the violence of the wind compelled him to lower his topmasts,

to make fast the yards to the deck, and to heave to…

During that same storm Hennepin (Hennepin 1903, pp. 113-4) said, ‘we brought down

our Main Yards and Top-Maft [sic] and let the Ship drive at the Mercy of the Wind…’

These statements suggest a square-rigged vessel with stepped masts. The

iconography of a map drawn in 1688 by the order of the Governor of New France

(Figure 2) shows Griffon entering Lake Erie (Remington 1891, p. 43). It is depicted as a

two-masted vessel, square-rigged on the foremast and lateen rigged on the main mast.

One of the legends on the map describes the cabin la Salle stayed in during the

construction of Griffon. The legend reads, ‘Cabane on le Sr de la Salle a fait faire une

1 A harquebuze is a type of matchlock rifle and a-crock is French for a prop or support

33

barque. (Cabin where the Sieur de La Salle caused a bark to be built.)’ (Remington

1891, p. 43).

Figure 2: Map of the Niagara Area Drawn in 1688

(Remington 1891, p. 43)

An article in the Marine Review and Marine Record, 1903, by Richard P. Joy

(Joy 1903, p. 24) described the possible rigging of Griffon based upon a 1697 woodcut

depicting its construction as well as the description of Hennepin. Joy (Joy 1903, p. 24)

based his rigging design (Figure 3) on the following:

34

I have seen many pictures purporting to be the Griffin, all fanciful pictures,

some showing the vessel rigged as a schooner, but as the schooner rig was first

introduced in New England in 1714, some years after the Griffin was launched,

such pictures could not have been accurate.

It is probable that La Salle obtained the design of the vessel when in

France, and that her rig was the prevailing rig of vessels of that period, as shown

by the drawing. On the mainmast the triangular lateen sail, used then

universally, and on the foremast the two square sails, also common on vessels of

the time. As the triangular jig, or staysail, did not come into use until the early

part of the eighteenth century, it is probable that the Griffin carried a spritsail on

her high bowsprit. Even as late as the year 1750 the spritsail was common to all

sea-going vessels, as many old prints will show.

Figure 3: Proposed Rig of Griffon by Joy

(Joy 1903, p. 24)

35

Griffon, with la Salle in command, departed on its maiden voyage 7 August 1679

carrying 34 men (Hennepin 1903, p. 107), arms, supplies, and goods (Anderson 1901, p.

29). It sailed the length of Lake Erie, up the Saint Claire River, and into the lower

portion of Lake Huron. La Salle continued his voyage up Lake Huron, through the

Mackinaw Straits, and arrived 27 August 1679 at Missilimakinak, what is now present

day Pointe la Barbe and the town of Saint Ignace, Michigan (Anderson 1901, p. 37).

‘On the 12th of September he left Missilimakinak; entering Lake Illinois (present

day Lake Michigan), he arrived at an island situated at the opening of Green Bay (or

lake)’ (Anderson 1901, p. 43). The island la Salle landed on is present day Washington

Island (Hennepin 1903, p. 119). Here la Salle rendezvoused with some of his men that

he had previously sent to trade with the Illinois Indians (Anderson 1901, p. 43). They

had accumulate a ‘great quantity of Furrs [sic] and Skins’ (Hennepin 1903, p. 119) and

these were loaded on Griffon.

‘La Salle had planned originally to sail to the end of Lake Michigan, build a

sister ship to the Griffon, and continue down the Mississippi …’ (Hundley 1984, p. 5)

and for this reason Griffon was ‘… carrying in its hold the tackle, rigging and anchors of

the unbuilt sister-ship …’ (Hundley 1984, p. 5). Anticipating the approach of winter la

Salle decided to send Griffon and a small crew back with the furs to be unloaded at his

storehouse at the eastern end of Lake Erie while he continued his journey down Lake

Illinois in four canoes (Anderson 1901, p. 43). Griffon was to make an intermediate stop

at Missilimakinak to unload the provisions for the second ship which la Salle would

construct and use for his return (Anderson 1901, p. 44).

Griffon sailed on 18 September 1679 (Anderson 1901, p. 45) for Missilimakinak

and was never seen again. Scant information is available as to Griffon’s fate. The

following year la Salle (Anderson 1901, pp. 45-7) learned that:

The vessel having cast anchor in the northern part of Lake Illinois, the pilot,

against the advice of some Savages who warned him there was a great storm

36

outside, persisted in setting sail, without considering that the sheltered position

of the ship prevented him from perceiving the violence of the wind. Scarcely

was the ship a quarter league from land, when the Savages saw it tossing

frightfully, unable to make head against the storm, although all the sails had

been struck; a short time after, they lost it from sight, and think it to have been

driven upon shallows near the Huron Islands, where it has been buried in the

sand.

Referencing the collection of la Salle’s writing compiled by Margry and

translating same from the original French, Marshall (Marshall 1879, p. 287) provides

one account of wreckage being reported in this location:

A hatchway, a cabin door, the truck of a flag-staff, a piece of rope, a pack of

spoiled beaver skins, two pair of linen breeches torn and spoiled with tar, were

subsequently found and recognized as relics of the ill-fated ship.

AM to PM Data Evaluation and Scoring

Dating

Based on ceramic type the European ceramic sherds recovered from the Griffon

Cove shipwreck were dated to the 1840 to 1860 time period (Hundley 1984, p. 16). One

sherd bore a diagnostic maker’s mark and was dated to the 1850-1860 time period

(Hundley 1984). Another dateable artifact was the lock found by Vail. Diagnostic

markers on the lock gave a date range of 1790-1830. These date ranges are

chronologically much later than the loss of Griffon in 1679. The AM to PM data

comparison is clearly inconsistent and exclusionary. This category is scored a 3.

37


The AM/PM data comparison relative to vessel construction and rigging is very

inconsistent and exclusionary. This conclusion is based on several key points. First, the

notching of the keel for floor timbers was a technique not used by French shipwrights

until after the second half of the 18th century (Myers 1956, p. 144). Second, Griffon was

a decked vessel and no decking or indication of decking was found during the

archeological investigation. Third, no evidence of ballast stones was found at the wreck

site and, even at 45 tons (41 tonnes), Griffon would have required ballasting to sail

efficiently. Fourth and finally, no evidence of mast steps was found in the wreckage.

Griffon would have required substantial mast steps even with its light tonnage. Based on

these clear inconsistencies the category is scored a 3.


Although several artifacts have been recovered (see dating category) there is not

enough data to classify them as either cargo or crew effects. This category is scored a 0

due to lack of data.


Although several artifacts have been recovered (see dating category) there is not

enough data to classify them as either cargo or crew effects. This category is scored a 0

due to lack of data.

38

Location

The location of the Griffon Cove wreck site is inconsistent with the AM data

which indicates that the probable wreck site of Griffon is located in northern Lake

Michigan. Thus the category is scored a 3.


There in no AM data for the condition of the shipwreck. Thus the category is

scored a 0.

39


Completing the Shipwreck Identification Matrix (Table 2) for the wreckage

found yields a final score of 3 for the candidate ship for identification, Griffon. This

rating is exclusion: The historical and archaeological data in all or part of the categories

where comparisons are possible are clearly inconsistent. The shipwreck is not the ship

proposed.

Table 2: Griffon Shipwreck Identification Matrix

Conclusions

Case Study 1 used the HSIM to determine if the Griffon Cove shipwreck was

that of Griffon, the first ship to have sailed the Upper Great Lakes. The AM and PM data

were gathered and compared for consistency. The dating, vessel construction and

rigging, and location data comparison categories all indicated inconsistencies between

the AM and PM data. Based on these inconsistencies the conclusion is that the Griffon

Cove shipwreck is not Griffon.

Shipwreck Identification Matrix Candidate Ship for Identification Griffon AM to PM Shipwreck Identification Categories Data

Consistency Scoring

Dating 3

Vessel Construction and Rigging 3

Cargo and Cargo Handling Equipment 0

Crew Personal Effects 0

Location 3

Condition of Shipwreck and Site 0

Final Scoring (Highest number in right hand column) 3

40

Case Study 2: The Beaufort Inlet shipwreck


This case study involves applying the HSIM to a shipwreck that was discovered

in November of 1996 at Beaufort Inlet, North Carolina, USA (Wilde-Ramsing 2006, p.

160). It was discovered while searching an area that had a high probability of containing

ten eighteenth-century shipwrecks (Lawrence 2008, p. 15) of which four were armed.

These include an unknown brig, the Spanish snow El Salvador, and two ships belonging

to the pirate, Blackbeard, the sloop Adventure and his flagship Queen Anne’s Revenge

(QAR) (Wilde-Ramsing 2006, p. 164). The initial findings at the wreck site included a

ballast pile, two large anchors, and nine cannon (Wilde-Ramsing 2006, p. 164). Several

artifacts were recovered one of which was a ships bell dated 1705 (Wilde-Ramsing

2006, p. 164).

Since the site dated to the same period as the two armed ships Blackbeard lost, it

was assumed that the wreckage had to be Adventure or QAR (Wilde-Ramsing 2006, p.

164). Subsequent investigations of the wreck site in 1997 indicated that the wreckage

was that of QAR (Wilde-Ramsing 2006, p. 172). QAR was formerly the 200-300 ton

(180-270 tonnes) French slaver La Concorde which was captured by Blackbeard in

1717 (Lusardi 2006, pp. 196-7). Since the wreck was discovered in an area where QAR

(aka La Concorde) supposedly sank, La Concorde is the candidate ship posited for the

identity of the shipwreck found.

Shipwreck Site

The Beaufort Inlet shipwreck is located approximately 1.5 miles (2.4 km) south

of Beaufort Inlet, North Carolina, USA (Figure 4) and is identified as state

archaeological site 31CR314 (Rogers, Richards & Lusardi 2005, p. 24). It was

discovered in the fall of 1996 by Intersal, Inc. in a search area developed from historical

data compiled by Mike Daniels (Wilde-Ramsing 2006, p. 164). The wreck site

41

encompasses an area of about 110 feet by 75 feet and rests in 23 feet of water (Wilde-

Ramsing 2006, p. 177).

Figure 4: Location of the Beaufort Inlet Shipwreck

(Rogers, Richards & Lusardi 2005, p. 24)

Prior to excavation the initial site consisted of exposed wreckage above the sea

bed that measured

25 feet by 15 feet, and consisted of eleven cannon, two anchors, a grappling

hook, numerous iron cask hoops, several rigging elements, a cluster of cannon

balls, and a large amount of ballast stone and concretions (Wilde-Ramsing 2006,

p. 168).

Fieldwork at the site began in 1997 and continues to this date. At the completion of

the 1998 field season it was determine that the bow was pointing north toward the inlet

of the harbor and a site plan (Figure 5) was developed (Wilde-Ramsing 2006, p. 173).

42

Figure 5: Beaufort Inlet Shipwreck Site Plan circa 1998

(Wilde-Ramsing 2006, p. 174)


Since the inception of the Beaufort Inlet Shipwreck Project ‘tens of thousands of

artifacts have been recovered and many of those have received at least a preliminary

level of analysis’ (The QAR Project 2009b). The analysis of the artifact assemblage will

be discussed with regard to the identification categories: Vessel Construction and

Rigging, Cargo and Cargo Handling Equipment, and Crew Personal Effects. The arms

43

and armament, ships tools and instruments, and food preparation and storage equipment

will be covered under the Cargo and Cargo Handling Equipment category.

Vessel Construction and Rigging PM Data

Only a small amount of the hull structure remains due to the wreck’s location in

an active shallow saltwater marine environment. It consists of an articulated structure

approximately 31 feet (9.4 m) by 9 feet (2.7 m) which was raised in the early part of

2000.

This comparatively small section of ship's structure was made up of numerous

fragments of frame components and bottom planking with associated sacrificial

planking or sheathing. Unfortunately no sign of either the keel or keelson has

been revealed to date. Additionally, it is difficult to determine exactly where on

the original ship the surviving hull structure fit. The only diagnostic feature in

this regard is the presence of sacrificial sheathing that at least indicates a

position below the waterline (The QAR Project 2009m).

Other wooden remains include the lower portion of the stern post, twenty four

partial hull frames, several pieces of hull planking, sacrificial planking, and treenails.

‘Construction techniques including treenail and nail fastenings and composite frames are

consistent with 18th-century construction techniques as is the red pine sheathing

covering the outer hull planking’ (Rogers, Richards & Lusardi 2005, p. 28).

A wood species analysis was performed on the wooden remains and two species

were identified: oak (Quercus sp.), white-oak type, and Scots pine (Pinus sylvestris sp.)

(Newsom & Miller 2009, pp. 4-5). The geographic origin of the white oak is

indeterminate. ‘The anatomical structure is consistent with North American “live oaks,”

e.g. live oak (Q. virginana), and that of other species from similar habitats, including

certain southern European and Mediterranean species, e.g. Q. cerris’ (Newsom & Miller

2009, p. 5). The Scots pine geographic distribution includes ‘Spain, France, Scotland,

44

Northern Europe and Scandinavia, Yugoslavia, Romania, Turkey, and Russia’ (Newsom

& Miller 2009, p. 6).

Both a large and small bell have been recovered from the wreck site. The large

bell was recovered during the initial dive on the wreck site. Its exact position was not

recorded and it is speculated that it came from the central portion of the wreck site

(Wilde-Ramsing 2007, p. 6). The bell weights 21 lbs (9.5 kg) and is 8.66 inches (21.99

cm) high with a base diameter of 8.27 inches (27.01 cm) (Wilde-Ramsing 2007, p. 2).

The bell is made of a bronze alloy with a composition of 81% copper and 19% tin alloy

(Wilde-Ramsing 2007, p. 2). This type of bronze alloy has the same characteristics as

‘bell metal’ commonly used in Spain during the eighteenth century (Wilde-Ramsing

2007, p. 5). The bell has an upper inscription that reads ‘HIS Maria’ and a lower

inscription which reads ‘ANO DE 1705’ (Wilde-Ramsing 2007, p. 2), which is

indicative of Spanish origin. The small bell weighs 14.64 lbs (6.64 kg) and was raised

from the stern of the wreck and has no diagnostic markings (The QAR Project 2009d).

There are four stock-type anchors and one grapnel-type anchor associated with

the wreck site:

Perhaps the most easily identifiable features on the shipwreck are the two large

anchors that lie on the main portion of the site. A third anchor lies

approximately 50 feet [15 m] north of the primary concentration of material

while a fourth that appears to be associated with the wreck site is located about

400 feet south of this area. A smaller grapnel-type anchor is discernible lying

atop the main pile. Although accurate recording must wait for recovery and

eventual cleaning and conservation, all of the anchors have been measured to

facilitate appropriate site plan placement and initial interpretation efforts. Both

Anchors 3 and 4 have fairly well preserved wooden stocks remaining in situ

(The QAR Project 2009a, p. 172).

From the available physical data it was determined that the three large anchors at

the site were rated ‘for a vessel of 250 to 350 tons (225 to 320 tonnes) and would have

45

been too large for use on the much smaller Adventure, El Salvador, or other vessels

reportedly sunk in the inlet’ (Wilde-Ramsing 2006). The smaller fourth anchor is 8 feet

(2.4 m) in length and weighs about 600 lbs (272 kg). The grapnel-type anchor is 4 feet

10 ½ inches (1.49 m) in length and weighs about 125 lbs (57 kg) (The QAR Project

2009a).

‘A perforated semicircular lead pump sieve with three flanges was found in

1996, and a second sieve fragment was recovered in 1999. Both fragments resemble

sieves found on El Nuevo Constante, wrecked off Louisiana 1766’ (Lusardi 2006, p.

202). This is an indication that the wreck was of Spanish origin.

Other construction and rigging material recovered from the wreck site includes

lead patches and tar caulking with imbedded bovine or cow hair, iron hardware

including a rigging hook, sail cloth, cordage, and lead numerals used for draft or

compartment marking. The analyses of these artifacts have not been formally reported

but in general they are consistent with materials commonly used in eighteenth century

ship construction.

Cargo and Cargo Handling Equipment PM Data

Armament:

‘As of 2008 twenty-four cast iron artillery pieces have been located. Ten are

recovered and five of these have been cleaned and analyzed’ (The QAR Project 2009c).

The diagnostic data from the five cannon that have been analyzed is presented in Table

3. Three of the five cannon are of English manufacture, one of Swedish manufacture and

one of Swedish or French manufacture. With the exception of the diagnostic markings

on cannon 3, the manufacturing dates are consistent with a wrecking date of 1718.

Cannon 3

46

exhibits crudely chiseled numbers “17 3 0” running lengthwise along the first

reinforcement. According to Angus Konstam, C-3 may have originated in

France; if so, “1730” may represent the year of manufacture. Eighteenth century

French cannon were often chiseled with a date near the breech (Lusardi 2006, p.

202).

There is debate on whether the last numeral is a zero or an ‘aftermarket’ weight. The 173

may also represent the weight of the cannon in old English measure. If indeed the

numbers are a date, the wreck could not be La Concorde.

Table 3: Beaufort Inlet Shipwreck Cannon Data

Cannon Number

2 3 4 19 21

Material Cast iron Cast iron Cast iron Cast iron Cast iron

Size (lbs) 6 6 4 1 ½ Size (kg) 2.7 2.7 1.8 0.5 0.2 Length (ft) 7.5 7.0 5.5 4.0 3.5 Length (m) 2.3 2.1 1.7 1.2 1.1 Manufacturer English Swedish or

French English Swedish English

Date of Manufacture

Before 1716 1675-1700 Possibly

1730

1690-1715 1713 Not reported

Diagnostic Markings

No Yes Yes Yes Yes

Several pieces of small arms have been recovered along with one gun barrel. The

barrel is similar to an English musketoon with diagnostic markings indicating English

manufacture (The QAR Project 2009g). These types of shipboard weapons were used

from the mid-17th century through the 18th century (The QAR Project 2009g).

47

Ships Tools and Instruments:

These artifacts correspond to activities that took place on any armed, wooden

sailing vessel. The categories for these artifacts are: carpentry, gunnery, medical,

navigation, restraining devices, sail making/rigging, sharpening, survey, and those of an

unknown function such as tube and wire (The QAR Project 2009n).

Three artifacts of this assemblage have yielded diagnostic information: a pair of

dividers, a syringe, and an ointment container. The dividers are similar (The QAR

Project 2009f) to a pair found on the pirate ship Whydah which sank 26 April 1717

(Lusardi 2006, p. 132). The syringe and ointment container are French in origin (The

QAR Project 2009h).

Food Preparation and Storage Equipment:

This assemblage of artifacts consists of cask hoops, staves and heads, sherds

from ceramic jars, pewter flatware, pewter utensils, glass bottles, and stemware. Several

of these artifacts provide diagnostic information.

The pewter flatware is of English origin as reported on the QAR website (The

QAR Project 2009i):

So far a total of fifteen pewter flatware artifacts, whole or fragmentary, have

been analyzed. They represent six plates, three dishes, and three chargers. From

their legible makers marks at least four of the flatware vessels were made in

England by London pewterers George Hammond, John Stiles and Henry

Sewdley during the late 17th and early 18th century. Pewter flatware could have

been carried on a ship for a number of reasons: as cargo, as trade goods (for

example on slavers), as scrap metal or as tableware. The locations of most of the

flatware (recovered and still in situ) on this wreck are in or towards the stern of

the ship. The Captain and higher-ranking crew normally ate in the stern area of

ships and might be expected to use pewter tableware.

48

Several glass bottle and bottle fragments have been analyzed; these date to the

17th or 18th century and were manufactured in England and France (Carnes-McNaughton

& Wilde-Ramsing 2008, pp. 8-9). Ceramic sherds recovered from the site have been

analyzed and found to be consistent with French and Italian ceramics from the 17th

through 19th centuries (Carnes-McNaughton 2008, p. 19). The last piece of diagnostic

data comes from a wine glass stem. ‘The style of this molded stemware wine glass has a

baluster stem and a straight-sided funnel bowl. It exhibits embossed crowns and

diamonds and is associated with the coronation of George I’ (The QAR Project 2009l).

Slavery Equipment and Trade Goods:

It is unsure whether any artifacts associated with the slave trade have been

recovered although several leg irons have been recovered and are hypothesized to have

been from the slave trade.

Restraining devices in the form of leg irons are common on ships of the eighteenth

century as a means to deal with unruly sailors. Two examples of leg irons, also called

shackles or bilboes, have been recovered from the QAR site so far. One shackle was

identified via X-radiography and still remains in concretion. The second has been

cleaned and exhibits cord wrapping in a manner as those recovered from the wrecked

slave ship Henrietta Marie. The wrapping may have helped protect the legs of the ship's

enslaved human cargo (The QAR Project 2009k).

Considering that ‘almost ninety separate sets’ (Moore & Malcom 2008, p. 28) were

recovered from the slaver Henrietta Marie and twenty four sets from the former slaver

Whydah (Hamilton 2007, p. 144), it is likely that the two sets recovered from QAR were

just used for crew discipline.

Five glass beads, three complete and two damaged, have been recovered and

conserved from the wreck site to date (Carnes-McNaughton & Myers 2007, p. 2). The

analysis of these beads by the QAR team concluded that: ‘Clearly the archaeological

presence of both tobacco accoutrement and now the presence of glass beads, including

49

one of African manufacture, at least suggest that the vessel was once associated with the

transportation of slaves during its lifetime’ (Carnes-McNaughton & Myers 2007, p. 8).

Beads of this type were common in the slave trade as trade goods and adornments. The

slaver Henrietta Marie had been carrying 2,074 lbs of beads, much of which were sold

on its last voyage before it wrecked and over 11,000 beads have been recovered from the

wreck site (Moore & Malcom 2008, p. 29). The recovery of just five beads from the

Beaufort Inlet site leads to the inference that these beads were probably part of the

crew’s personal property in the form of adornments.

Crew Personal Effects PM Data

This assemblage of artifacts consists of crew apparel, recreational items, tobacco,

currency, and jewelry adornment. Several of these artifacts provide diagnostic

information. A pair of cuff links were found in concretion and date to the 18th century

time period (The QAR Project 2009e). Pipe fragments recovered from the wreck site

appear to be of English origin based on observable manufacturing traits. No

fragments exhibit a maker's mark of any type. Therefore dating of these pipes

fragments was based on two observations (bowl shape and bore diameter) and

calculated in three ways to assess a chronological dimension to the assemblage.

Based on bowl shape using Oswald's observations (as seen in Noel Hume's 1969

text), three pipes date to the general period of 1690 to 1750 in manufacture,

while one appears to date earlier to the 1680 to 1710 period. Using Harrington's

observation of bore diameter changes from 17th to 18th century (generally bore

diameters became proportionately smaller - narrower - as the stem grew longer)

which are subdivided into five distinct categories of diameter (ranging from

9/64th of an inch to 4/64 of and inch as the latest), the nine measurable stem

fragments fall into the 1680 to the 1710 period of manufacture (The QAR

Project 2009j).

50


La Concorde first appears in the historical record in 1710 when it was acquired

by René Montaudouin (Lawrence 2009a), considered to be ‘one of the most successful

slave traders in Nantes (Lusardi 2006, p. 196)’, France. The ship was described as being

a ‘300-ton frigate armed with 26 cannon’ (Lawrence 2009a). Montaudouin acquired La

Concorde during the War of Spanish Succession (Queen Anne’s War) and operated it

initially as a privateer (Lawrence 2009a).

French historical records recount three slaving voyages: 1713, 1715, and the final

voyage as a slaver in 1717 (Lusardi 2006, p. 196). It was during this final voyage in

1717 that La Concorde was captured by Edward Thatch or Teach, also known as

Blackbeard the pirate (Lusardi 2006, p. 196).

The final voyage began on 24 March 1717 in Nantes (Lusardi 2006, p. 196). The

vessel was described at the time by two of its officers as a 200 ton (181 tonnes) vessel

armed with sixteen cannon (Lawrence 2009b). La Concorde sailed to Judah, Africa and

loaded slaves on 8 July 1717 (Lusardi 2006, p. 196). It then embarked on the final leg of

the outbound voyage destined for the island of Martinique (Lusardi 2006, p. 196).

Nearing the completion of its transatlantic journey La Concorde was captured by

Blackbeard just off the Island of St. Vincent on 28 November 1717 (Lusardi 2006, p.

196). The pirates released most of the French crew on the island of Bequia (Lusardi

2006, p. 196). ‘The cabin boy and three of his fellow French crewmen voluntarily joined

the pirates, and ten others were taken by force including the pilot, three surgeons, two

carpenters, two sailors, and a cook’ (Lawrence & Wilde-Ramsing 2001, p. 2).

Blackbeard then renamed the ship Queen Anne’s Revenge (Lawrence & Wilde-Ramsing

2001, p. 2).

Late in November of 1717 the pirates captured and burned Great Allen near St.

Lucia or St. Vincent Island (Lawrence & Wilde-Ramsing 2001, p. 3). Their plunder

51

from the Great Allen consisted of ‘a great deal of plate and a very fine cup’ (Lawrence

2009b). QAR was then descried as a ‘French ship of 32 guns’ (Lawrence 2009b).

On 5 December 1717, off Crab Island (Present day Isla de Vieques, Puerto Rico),

the sloop Margaret of St. Christophers was seized (Lawrence 2009b). Henry Bostock,

master of Margaret of St. Christophers, described seeing the plunder from the Great

Allen and described QAR as a Dutch-built French guineaman with 36 guns (Lawrence

2009b).

‘From there, a former captive reported that the pirates were headed to Samana

Bay in Hispaniola (Dominican Republic). No historic records have been located to

chronicle Blackbeard’s movements during the first three months of 1718…’ (Lawrence

2009b).

During this period, Blackbeard may have switched his French flagship La

Concorde for a forty-gun English ship, based on a deposition reporting a third

hand account (Morange 1718). While there is no further documentation to

corroborate this vague mention, the possibility cannot be dismissed, thus adding

to the uncertainty surrounding the origins of the vessel known as Queen Anne’s

Revenge (Wilde-Ramsing 2006, p. 192).

Using his flagship QAR and three smaller sloops, Blackbeard blockaded the Port

of Charleston, South Carolina, in May of 1718 (Lusardi 2006, p. 197). Having plundered

numerous ships and after demanding and receiving a ransom of medical supplies, he

headed northward to North Carolina (Lusardi 2006, p. 197). During the Charleston

blockade Blackbeard’s ship was described as ‘a large French ship mounted with 40

guns’ (Lawrence 2009b).

Attempting to enter Topsail Inlet (present day Beaufort Inlet, North Carolina) on

10 June 1718, ‘the Queen Anne’s Revenge struck on the bar and became a total wreck.

Of three sloops in company, one was also wrecked on the bar’ (Laughton 1898, p. 1).

52

This was the sloop Adventure which was lost when Blackbeard ordered it back to assist

QAR in getting off the bar (Lusardi 2006, p. 197).


Dating

The artifact assemblage collected to date, with one exception, is consistent with a

wrecking date of 1718 or before. Cannon 3, with a possible date of 1730, is the singular

exception. There is debate over whether the chiseled numbers on the cannon represent, a

weight or date, and until this is resolved this category is scored a 2.


The analysis of the vessel’s remains and rigging hardware indicate that it was

probably constructed in the late 1600’s or early 1700’s, an estimate consistent with a

wrecking date of 1718. The AM data collected for La Concorde places the vessels size

between 200 and 300 tons (181 and 272 tonnes). The analyses of the anchor and

scantling data are consistent with a vessel of between 200 and 300 tons (181 and 272

tonnes) as well. Wood analysis conducted indicates that the ship was constructed of

European timber suggesting a European build. Unfortunately the analysis of the

construction techniques does not help narrow down the origin or nationality of the

wreck since ship building techniques from this era are not documented sufficiently to

make such determinations possible (Rogers, Richards & Lusardi 2005, p. 34).

It is assumed from the AM data that La Concorde was a French-built ship yet the

bell and pump sieve analyzed are of Spanish origin. If the wreck proves to be Spanish-

built then the Beaufort Inlet wreck is not La Concorde. Since La Concorde’s history

before 1710 is currently unavailable and the data comparison is based on the available

53

AM data, this category is scored a 2 - the historical and archaeological data are not

consistent in sufficient detail to establish that they belong to the proposed ship.


The comparison of the AM to PM data for this category has yielded several

inconsistencies, the first being the cannon. Of the twenty four pieces that have been

located to date, ten have been recovered and five of these have been analyzed. Of the

five, three are English, one is Swedish, and one is either Swedish or French. This is

inconsistent with La Concorde being a French vessel. Why would a French naval vessel,

a frigate, be armed with English cannon? Montaudouin acquired La Concorde in 1710

during the War of Spanish Succession (Queen Anne’s War) (Lawrence 2009a). One

would assume that the French navy would have procured French manufactured

armament for their warships not English, since the French were at war with the English.

This inconsistency may be explained by Blackbeard transferring additional armament

from his other ships or they may have been captured and added to the vessel during the

course of the war. Be that as it may, until this is resolved the inconsistency in armament

nationality remains.

The pewter flatware analyzed so far is of English origin and no personal

markings or inscriptions that would indicate that these pieces belonged to the French

ship La Concorde have been reported (The QAR Project 2009i). Pewter flatware from

this time period was occasionally stamped or inscribed with the initials of the ship it was

associated with (Lusardi 2006, p. 217). A plate recovered from the pirate vessel Whydah

(1717) bore the initials, ‘WG’, and a plate from the British slaver Henrietta Maria was

stamped with an ‘HM’ (Lusardi 2006, p. 217). ‘None of the Beufort Inlet pewterware

has the initials “LC” (for La Concorde) or “QAR” (Queen Anne’s Revenge)’ (Lusardi

2006, p. 217).

54

There is also an absence of slave trade goods. La Concorde was captured

delivering slaves to Martinique and purportedly wrecked seven months later. There

should be some evidence of slave trade activity in the artifact assemblage. To date only

two sets of leg irons have been recovered from the site. When compared to the twenty

four sets that have been recovered from Whydah, a former slaver that was captured and

turned into a pirate vessel, it is unlikely that the two sets of leg irons found were

remnants of an earlier function as a slaver. Beads were also common in the slave trade as

trade goods but only five individual beads have been recovered from the wreck site so

far (Carnes-McNaughton & Myers 2007, p. 2).

There are inconsistencies in the armament, pewterware, and slave trade goods

AM to PM data comparisons. Since the historical and archaeological data are not

consistent in sufficient detail to establish that these items necessarily belong to the

proposed ship, the category is scored a 2. Again this is based on La Concorde’s history

before 1710 being currently unavailable and the data evaluation being based on the

available AM data.


Crew personal effects date to the time period of the wreck and are consistent

with that of an English crew. This category is scored a 1.

Location

The location of the wreck site is consistent with the AM data. This category is

scored a 1.

55


No AM data is available relative to the condition of the shipwreck and site, so

this category scored a 0.



found at Beaufort Inlet yields a final score for the candidate ship for identification, La

Concorde, of 2. This rating is no conclusion: The historical and archaeological data in

all or part of the categories where comparisons are possible are not consistent in

sufficient detail to establish that they belong to the proposed ship.

Table 4: La Concorde Shipwreck Identification Matrix

Shipwreck Identification Matrix Candidate Ship for Identification La Concorde AM to PM Shipwreck Identification Categories Data

Consistency Scoring

Dating 2




Location 1



56

Conclusions

Case Study 2 used the HSIM to determine if the wreckage discovered at Beaufort

Inlet was that of La Concorde. The AM and PM data were gathered and compared for

consistency. The dating, vessel construction and rigging, and cargo and cargo handling

equipment categories all indicated that the consistency between the AM and PM data

was not sufficient enough in detail to identify the Beaufort Inlet shipwreck as La

Concorde.

57

Case Study 3: The Bark Cortland


This case study involves applying the HSIM to a shipwreck that was discovered

in Lake Erie, north of Avon Point Ohio, USA on 30 July 2005 by the Cleveland

Underwater Explorers (CLUE) (Figure 6). The wreck was discovered while searching an

area determined through historical research to have a high probability of containing the

wreck of the bark Cortland. Since the wreck was discovered in the area where Cortland

supposedly sank, Cortland is the candidate ship posited for identification of the

shipwreck found.

Figure 6: Location of the Bark Cortland

(Map: David M. VanZandt)

58

Shipwreck Site

The wreck is located approximately 20 miles (32 km) northwest of Cleveland,

Ohio and approximately 10 miles (16 km) north of Avon Point, Ohio (Figure 6). A high-

resolution sidescan sonar survey conducted at the wreck site revealed that the shipwreck

is comprised of three distinct sections (Figure 7).

Figure 7: Sidescan Image of Cortland Wreck Site

(Sidescan Image: David M. VanZandt)

The three sections were identified during preliminary dives as the bow, a piece of

decking, and the stern. The midship section of the wreck was not seen and is thought to

have collapsed due to the weight of the cargo it was carrying. The stern section is lying

on its starboard side almost completely buried leaving only a small portion of the port

side exposed. It is square constructed and several frames are visible.

The hull at the bow is intact from the stem to just aft of the forecastle deck and

windlass. A bowsprit is present. The bow rests on its starboard side at an angle of

approximately forty-five degrees and is below the lake bottom. It is kept exposed by the

scouring action of the current that flows over the wreck site. Though the starboard

gunwale is almost completely buried the port gunwale rises approximately 8 feet (2.5 m)

above the soft mud bottom. The bowsprit is partially pulled out of the bow and extends

approximately 20 feet (6 m) forward with the end being broken off. Under the bowsprit

is a distinct scroll head (Figure 8).

59

Figure 8: Scroll Head of the Bark Cortland

(Photo: David M. VanZandt)

There is a three-foot (one m) high, slatted, forecastle deck on the bow section. A

capstan, two catheads and a pawl bit with rocker assembly are found on this deck. Below

the rocker are two holes in the deck where drive rods for actuating a bilge pump would

have been placed. Along the gunwales on either side are line chocks for anchor cables or

dock lines. A windlass, still wrapped with anchor chain, is located just aft and below the

forecastle deck. The ship’s bell (size 4), which had detached and fallen from the bell

holder on the pawl bit during or subsequent to the sinking, was found sitting on top of

the windlass spool. The bell was subsequently raised and conserved at the Great Lakes

Historical Society by marine archaeologist Carrie Sowden. Further aft of the windlass

the bow section abruptly ends in a chaos of broken timbers. In this mass of broken

timbers at the rear of the bow section is what appears to be a tapered yard, broken on one

end.

60

The stern, like the bow, is almost completely buried in soft, oozy mud. It lies on

its starboard side at an approximate forty-five degree angle with only a small portion of

the transom and gunwales exposed. The stern is square and has a carved lip visible along

the top of the transom. A wooden cleat and line chock are mounted in the corner of the

transom. A portion of the cabin coaming is intact along the port side and some decking

is still present. Moving forward from the transom and past the portion of cabin combing

the stern breaks up into massive frames and then disappears into the soft, mud bottom.

Planking on the port side of the stern is of carvel construction, square joined, and

fastened with two bolts and two spikes per frame


There is limited archaeological information from this wreck site due to its recent

discovery and only a few preliminary dives performed during the initial reconnaissance

survey. The archaeological data collected to date is tabularized in Table 5:

Table 5: Cortland PM Archaeological Data

Cortland PM Archaeological Data Wooden construction Sail powered Carvel built Square plank joining Planking fastened with two spikes and two bolts per frame

Capstan on forecastle deck

Heavily kneed construction Square transom Scroll head Bell with foundry marking showing bell

was manufactured by Rumsey & Company, Seneca Falls, New York and was size 4

Rocker activated forward pump Bow and stern sections resting at approximately 45 degrees to starboard in soft mud.

Midships section assumed collapsed and silted over

Location about 20 miles (32 km) NW of Cleveland, Ohio and about 10 miles (16 km) N of Avon Point, Ohio

61

AM Historical Research Data on Cortland

Master builder, Albert G. Huntley, of Sheboygan, Wisconsin, constructed the

clipper-style bark Cortland in 1867 (Sholes 1867, p. 1). Cortland was built at the request

of Asahel P. Lyman Esquire, a respected Sheboygan businessman in the mercantile field

(Sholes 1867, p. 1). At that time it was the largest schooner ever built in Sheboygan and

became the pride of Lyman’s fleet (The Sheboygan Press 1930, p. 3). Mr. Lyman spared

little expense in its construction which cost approximately $50,000.00, a considerable

sum in 1867 (Milwaukee Sentinel 1867, p. 3). Captain and master of the vessel, James

W. Louden, supervised the construction from the initial laying of the keel to the final

rigging of the sails (Milwaukee Sentinel 1867, p. 3). Captain Louden, a thorough and

experienced seaman, had been in Mr. Lyman’s employ for over seven years (Milwaukee

Sentinel 1867, p. 3).

On the morning of 21 June 1868, Cortland, under the command of Captain

Louden, was downbound the Lakes from Sheboygan, Wisconsin (Cleveland Leader

1868a, p. 3). It carried a crew of 11, one passenger, and a cargo of 891 gross tons (982

tonnes) of iron ore (Cleveland Leader 1868a, p. 3; McDonald 1958, pp. 311-3). The ore

had been loaded at Escanaba, Michigan and was destined for the steel mills of

Cleveland, Ohio (Cleveland Leader 1868a, p. 3).

The evening of the same day the steamer Morning Star, under the command of

Captain E. R. Viger, departed Cleveland Harbor at half past ten o’clock on its usual

Cleveland to Detroit passenger and freight run (Cleveland Leader 1868b, p. 3). Its

departure time was unusually late due to a last minute load of freight. Morning Star

carried a crew of thirty-eight, cabin passengers and emigrants along with a cargo of 80

tons (72 tonnes) of iron and a small quantity of miscellaneous freight (Cleveland Daily

Herald 1868b, p. 3; Cleveland Leader 1868c, p. 3). The Cleveland Leader described the

weather conditions that evening as dark and rainy with ‘a peculiar mist penetrating the

air’ (Cleveland Leader 1868b, p. 3).

62

Around midnight on 21 June 1868 Andrew Brown, a seaman on watch standing

on the ‘Gallant fore Castle’ of Cortland, saw the lights of an approaching steamer

(McDonald 1958, pp. 311-3). He reported the sighting to the first mate, who, after going

aft and returning with some glasses (binoculars), confirmed that a large steamboat was

approaching (McDonald 1958, pp. 311-3). Cortland’s position was approximately 26

nautical miles (42 km) above Cleveland, running on a port tack and ‘bearing east by

north half north with the wind north by east, a fresh breeze’ (Cleveland Leader 1868a, p.

3). Morning Star was on its normal run from Cleveland to Detroit via Pelee Passage

traveling on a northwest bearing at a speed between ten and twelve miles per hour (16 -

19 kph) (Cleveland Leader 1868b, p. 3). At nearly 15 minutes of one o’clock in the

morning Captain Viger of Morning Star was on watch along with the second mate, a

lookout, and the helmsman when they heard Cortland’s bell (Cleveland Daily Herald

1868a, p. 3). Before Captain Viger had time to issue stopping orders, Morning Star

collided with Cortland ‘near the mizzen rigging on her starboard side’ (McDonald

1958, pp. 311-3).

The force of the crash was enormous and was described by a reporter from The

Plain Dealer (The Plain Dealer 1868, p. 3):

… while Captain Viger was still on deck, two bells from a sail vessel were

heard and before the engines of the Morning Star could be stopped she

struck the bark Cortland with her full force. The bark was laden with iron

ore, yet such was the force of the Star that she passed nearly through her. As

the two vessels collided such was the force of collision that the two anchors

of the Star were thrown from their position on the lower deck across the

bark one of them passing completely over the vessel, holding her closely

confined, brought her stern directly against the wheel house of the steamer.

The force of the waves soon beat the wheel house and the wheel to pieces,

rendering it completely useless.

63

The bow of Morning Star was ripped completely open and it sank in

approximately fifteen minutes. Cortland, once free of Morning Star’s anchors, drifted

away and sank about an hour and a half later (Cleveland Daily Herald 1868b, p. 3).

The final death toll was never determined but it has been estimated that at least

30 people lost their lives2. Following the common practice of the time, only paying

adult passengers, not children or emigrants, were recorded on the passenger manifest of

Morning Star. This practice, coupled with the fact that many of the bodies were never

recovered, leaves the final toll in human lives unknown.

The cause of the accident between Morning Star and Cortland was determined

later to be faulty navigational lighting on Cortland. About twenty minutes prior to the

collision the mate of Cortland removed the green lantern, which was burning dimly,

from the mizzen rigging and took it inside to trim the wick (McDonald 1958, p. 311-3).

Just as he returned the lamp to the rigging, Morning Star struck Cortland at that precise

point killing the mate on impact (McDonald 1958, p. 311-3).

Cortland’s official enrollment shows the vessel measured 173.6 feet (52.9 m) in

length, 34.4 feet (10.5 m) in breath and 13.8 feet (4.2 m) in depth (Figure 9) (Sholes

1867, p. 1). Its cargo capacity was 636.99 tons (702.15 tonnes) under deck and 39.14

tons (43.14 tonnes) on deck, endowing it with a total cargo carrying capacity of 676.13

tons (745.30 tonnes) (Sholes 1867, p. 1). This made it one of the largest vessels ever

built on the Great Lakes during this time period (The Sheboygan Press 1930, p. 3). It

was constructed with a square stern and well adorned with a scroll head (Sholes 1867, p.

1). It had one deck, three masts, and its official enrollment number was 15 (Sholes 1867,

p. 1). Cortland’s enrollment was duly signed by A. L. Sholes, Collector of Customs, on

21 August 1867 (Sholes 1867, p. 1).

2 Author’s calculation from various newspaper accounts

64

Figure 9: Cortland Official Enrollment Document

(Image: Jim Paskert Collection)

65

When the newly-launched Cortland arrived in the Port of Milwaukee its large

size and graceful lines caused quite a stir along the dock. The Milwaukee Sentinel

reported, ‘A large number of our citizens visited the Cortland yesterday and all united in

pronouncing her one of the most complete vessels ever built on the lakes’ (Milwaukee

Sentinel 1867, p. 3).

The Milwaukee Sentinel (Milwaukee Sentinel 1867, p. 3) further described the

details of Cortland’s construction:

Her hull is of the greatest possible strength, the frames being 12x15 at the

bottom, 12x8 at the top, and 22 inches from center to center; stanchions 8 inches

square; outside plank 4 inches thick; garboard streaks respectively 8, 7 and 6

inches thick, scarfed and keyed; the planking is square fastened, with two bolts

and two spikes in each frame, and the ceiling, besides, is edge bolted

throughout. Her main keelson is 18x40, sister keelsons 10x22, pocket pieces

18x36. There are two center boards, one 24 and the other 22 feet, made of 12, 10

and 8 inch plank. She has five breast-hooks forward, secured with beams, which

in turn are kneed off thoroughly; three similar hooks aft add greatly to the

strength of the stern. Her deck frames are kneed off and nailed in a manner

similar to those of the Homer, while two lines of beam supporters extend the

entire length of the vessel, just outside the combings of the hatches. The

windlass pits, tow posts and snubbing posts are extra heavy and kneed off, and

the shifting boards in the hold are let in between stanchions. Her spars are

towering and massive, yet well proportioned and graceful in appearance.

Asahel P. Lyman, Esquire must have been very proud of Cortland because he

commissioned a photograph of the vessel, a very unusual event at this time since

photography was in its infancy and very expensive (Figure 10). This photographic

evidence confirms the description of the vessel recorded on the enrollment paper and

printed in the Milwaukee Sentinel in addition to showing that Cortland was

‘barquentine’ rigged with the foremast being square rigged and the main and

mizzenmasts fore-and-aft rigged.

66

Figure 10: Only Known Picture of the Bark Cortland

(Photo: Center for Archival Collections, Bowling Green State University)

AM to PM Data Evaluation

Dating

So far the only artifact examined that has a dating potential is the ship’s bell. The

manufacturer’s information cast on the bell shows that it was manufactured between

1863 and 1890. This is the time period that the manufacturer, Rumsey & Co, had a

facility in Seneca Falls, New York, USA. Although this data puts the bell in the same

time period as Cortland, this category is scored a 0 because the time period is too broad

and provides insufficient evidence on which to base a positive identification.

67


The AM/PM data comparison (Table 6) is shown as follows:

Table 6: Cortland AM/PM Data Comparison

AM Ship Data PM Archeological Data

Two capstans, one mounted forward and one mounted amidships, and square stern

Capstan mounted forward on forecastle deck and square stern

Scroll head Scroll head Bell mounted at forecastle deck Bell mounted at forecastle deck Planking square fastened with two bolts and two spikes per frame

Planking square fastened with two bolts and two spikes per frame

Two pumps, one amidships, one forward Forward pump on and below forecastle deck

Location: approx 20 miles (32 km) NW of Cleveland, Ohio and approx 10 miles (16 km) N of Avon Point, Ohio

Location: approx 20 miles (32 km) NW of Cleveland, Ohio and approx 10 miles (16 km) N of Avon Point, Ohio (Locational data has been left approx as the coordinates of the wreck have not been publically released)

Although at this time the archaeological data is sparse, the available data

comparison is very consistent and the category is scored a 1.


No cargo or cargo handling equipment has been found to date. There is

circumstantial evidence, assuming that the midship section of the wreck is collapsed,

that the cargo was very heavy and possibly iron ore (conjecture). The AM data indicates

that Cortland had cargo shifting boards in the holds between stanchions (Milwaukee

Sentinel 1867, p. 3). The shifting boards would have held the cargo in place putting a

tremendous strain on the timbers with Cortland resting at an angle on the bottom

(Milwaukee Sentinel 1867, p. 3). This was probably the direct cause of the collapsed

68

midship section of the wreck. However since this has yet to be proven as fact the

category is scored a 0.


No crew personal effects have been found to date. This category is scored a 0.

Location

The location of the wreck is consistent with the calculated and reported position

ascertained from the AM data. A search of shipwreck databases shows no other ships of

this type known to be lost in this area. This category is scored a 1.


The condition of the shipwreck and site is consistent with the AM data. Cortland

was reportedly struck on its starboard side and had its centerboards down (McDonald

1958, p. 313). The damage incurred on the starboard side with the centerboards down

would cause the ship to sink and come to rest on its starboard side on the bottom. The

wreck currently is lying on its starboard side in a soft mud bottom. The divers that

salvaged the rigging from Cortland reported, ‘…the bottom of the lake covered with

several feet of soft mud, in which the vessel is imbedded, and it is thought to be almost

an impossibility to ever raise her to the surface. She now lies nearly on her side with the

mud nearly or quite over her bulwarks’ (Cleveland Daily Herald 1868c, p. 3).

Since the AM/PM data comparison is consistent, this category is scored a 1.

69



found yields a final score for the candidate ship for identification, Cortland, of 1. This

rating is identification: The historical and archaeological data in all categories where

comparisons are possible are consistent in sufficient detail to establish that they belong

to the proposed ship and there are no irreconcilable discrepancies. The shipwreck is the

ship proposed.

Table 7: Cortland Shipwreck Identification Matrix

Shipwreck Identification Matrix Candidate Ship for Identification Cortland AM to PM Shipwreck Identification Categories Data

Consistency Scoring

Dating 0




Location 1



70

Conclusions

Case Study 3 applied the HSIM to determine if the wreckage discovered

approximately 20 miles (32 km) NW of Cleveland, Ohio and approximately 10 miles

(16 km) N of Avon Point, Ohio was that of Cortland. The AM and PM data were

gathered and compared for consistency. The vessel construction and rigging, location,

and condition of the shipwreck categories all indicated a consistent match between the

AM and PM data. The consistency of the data match was sufficient enough in detail to

establish the identification of the wreckage as Cortland.

71

Case Study 4: The Sidewheel Steamer Anthony Wayne


This case study involves applying the HSIM to a shipwreck discovered in Lake

Erie, north of Vermilion, Ohio, USA on 16 September 2006 by CLUE member Tom

Kowalczk (Figure 11). The wreck was discovered while searching an area determined

through historical research to have a high probability of containing the wreck of the

sidewheel steamboat Anthony Wayne. Since the wreck was discovered in an area where

Anthony Wayne supposedly sank it is the candidate ship posited for identification of the

shipwreck found.

Figure 11: Location of Anthony Wayne

(Map: David M. VanZandt)

72

Shipwreck Site

The wreck is located 16.7 miles (26.9 km) from the mouth of Sandusky Bay and

7.3 miles (11.7 km) north of Vermilion, Ohio (Figure 11). It has been designated site

33ER556 by the State of Ohio. A high-resolution sidescan sonar survey was conducted

of the wreck site and revealed that the shipwreck was in two distinct sections and

possibly silted over (Figure 12).

Figure 12: Sidescan of Anthony Wayne Wreck Site

(Sidescan Image: Tom Kowalczk)

Preliminary dives determined that the shipwreck site consists of two exposed

sections, the midship section and the bow section. The wreck lies with the bow facing

approximately south and the aft section facing approximately north. The midship section

consists of the paddlewheels, mechanical drive system, and some remains of the hull.

73

The starboard sidewheel was found stripped of its outer arms with just an

exposed hub outboard. The port sidewheel is more intact with a nearly full set of

undamaged buckets on the lower portion of the wheel. Each bucket is divided into three

sections with two center radial supports. The total width of each bucket is 8 feet (2.4 m)

with the center section about 5 feet (1.5 m) long. The inner and outer sections of the

bucket to the housing edges measured 3 feet (1 m) each. The outside radius of the port

sidewheel measures 12 feet (3.7 m) from the center hub to the outer bucket, and 20 arms

are counted around the outside of the hub. The port sidewheel is 24 feet in diameter (7.3

m). Each sidewheel has an independent shaft with a crank that connects at the center of

the ship to a horizontal Pittman arm. There is approximately 2 feet (0.6 m) of hull

remaining above the sidewheel shafts. At this point from the interior of the port side of

the hull to the interior of the starboard side of the hull measures 25 feet (7.6 m). From

the interior of the starboard side of the hull to the centerline between the two shafts

measures 13 feet (4.0 m). The sidewheel shafts measure 1 foot (0.3 m) in diameter and

the cranks measure approximately 3 feet (1 m) in length to the pin joint. The pin joint is

complicated and appears to have an extra linkage connecting the crank to the Pittman

arm on the port side (Figure 13). The Pittman arm itself measures 20 feet (6.1 m) long

before disappearing into the silt. The cross section of the Pittman arm near the pin joint

is 6 inches (15 cm) wide by 12 inches (30 cm) high. There is a cylinder, possibly a steam

expander, standing near the end of the Pittman arm; that measures 28 inches (71 cm) in

diameter and stands 38 inches (97 cm) high. The hole in the top of the cylinder measures

9 inches (23 cm) in diameter. No linkages are associated with this vertical cylinder and

there is a small open space underneath it. To the east side of the pin joint of the Pittman

arm is another standing cylinder with a cross head linkage at the top and with piping

coming off its side. It is thought that this is an air pump. The cylinder measures 18

inches (46 cm) in diameter.

74

Figure 13: Pittman Arm Linkage of Anthony Wayne


There are two connecting links that commence at the sidewheel shaft on the

starboard side and extend forward before turning 90 degrees downwards and

disappearing into the silt. These are control linkages that run from the cams on the shaft

to the valves on the steam engine. All of the equipment observed is consistent with a

horizontal steam engine and the engine and boilers are likely present but buried low in

the hull underneath the silt forward of the Pittman arm.

The bow consists of the stem and some railing located approximately 75 feet (23

m) forward from the midship section (Figure 14). There are two wood-stock anchors still

attached to either side of the bow almost completely buried in the silt. Emerging from

the silt approximately 10 feet (3 m) behind the stem are two small knightheads.

75

Figure 14: Bow of Anthony Wayne


A preliminary site plan of Anthony Wayne, produced by Brad Kruger of Texas A

& M University, is presented in Figure 15.

Figure 15: Anthony Wayne Site Plan

(Image: Bradley A. Krueger)

76

Excavation of the wreck site began in the spring of 2009 headed by Bradley A.

Krueger of Texas A & M University with support provided by the Great Lakes

Historical Society (GLHS) and the Cleveland Underwater Explorers (CLUE). The

purpose of the excavation was to determine and assess the remains of the wreck that

were buried beneath the silt. The initial excavation began on the port side of the wreck

forward of the paddlewheel’s axis to follow the hull side to the turn of the bilge to

determine the depth and condition of the bottom of the hull. This effort continued down

approximately 10 feet (3 m) until the excavation effort was refocused and moved to the

center of the wreck forward of the paddlewheel’s axis to determine if the engine and

other mechanical equipment were still present. The excavation of this area uncovered the

Pittman arm to crosshead linkage (Figure 16), four poppet steam valves for engine

operation (Figure 17), and a high pressure, horizontal, crosshead steam engine (Figure

18).

Figure 16: Anthony Wayne Crosshead Linkage


77

Figure 17: Anthony Wayne Poppet Steam Valve and Linkage


Figure 18: Anthony Wayne High Pressure Steam Engine Cylinder Head


78


Archaeological information from this wreck site is limited due to the wreck’s

recent discovery with only a few preliminary dives performed during the initial

reconnaissance survey and excavation just beginning. The archaeological data collected

to date is listed in Table 8.

Table 8: Anthony Wayne PM Archaeological Data

Anthony Wayne PM Archaeological Data Wooden construction Two wood-stock anchor Carvel built Steam power, high pressure, horizontal

engine Square plank joining Steam expander (speculation) Two side paddlewheels Crosshead air pump (speculation) Paddlewheel diameter: 24 feet (7.3 m) (measured 12 feet (3.7 m) from hub to end of arm)

Approx beam: 25 feet (7.6 m), interior measurement at the paddlewheel axles

Paddlewheel axles, cranks, and pin joint connection to Pittman arm


The sidewheel steamer Anthony Wayne (also know as ‘Mad Anthony’ and

‘General Wayne’) was built in 1837 at Perrysburgh, Ohio by master carpenter, Samuel

L. Hubble, for the Perrysburgh & Miami Steamboat Company (Thunder Bay National

Marine Sanctuary 2009a). Heyl states that it was ‘first known as the GENERAL WAYNE,

she was enrolled at Buffalo in early 1839 as ANTHONY WAYNE; in newspaper notices

she is usually listed just as WAYNE’ (Heyl 1964, p. 99). There is some uncertainty as to

the original name of the vessel since all the official vessel enrollments from the Port of

Miami, Ohio have been lost, but the enrollments from 28 September 1839 on list its

name as Anthony Wayne (Stubbins 1839, p. 1).

79

Anthony Wayne’s official enrollment from 28 September 1839 lists the vessel as

measuring 156 feet (47.5 m) in length, 25 feet 9 inches (7.8 m) in breath and 10 feet 3

inches (3.1 m) in depth and measures 390 and 46/95 ‘tuns’ (354.3 tonnes) (Stubbins

1839, p. 1). It was constructed as a steam boat with three decks, one mast, no galleries,

and a scroll head (Stubbins 1839). It was powered by a 120 horsepower, vertical

crosshead steam engine manufactured by Hathaway & Company in 1836 (US Treasury

Department 1838, p. 339). The vertical crosshead engine frame is clearly visible in the

wood cut of Anthony Wayne made shortly after its launch (Figure 19) and in a scaled

hand drawing made by Heyl (Heyl 1964, p. 99) (Figure 20).

Figure 19: Wood Cut of Anthony Wayne

(Wood Cut: J. W. Orr, Mariners Museum, Newport News, VA)

80

Figure 20: Scaled Drawing of Anthony Wayne

(Heyl 1964, p. 99)

Anthony Wayne began its career carrying passengers and freight between

Perrysburgh and Toledo, Ohio, and Buffalo, New York (Cleveland Daily Herald &

Gazette 1838, p. 2). It continued this service until 1847 when it ‘was seriously damaged

in an accident on Lake Erie’ (Heyl 1964, p. 99). It was towed to Monroe, Michigan,

stripped of its engine and associated mechanical equipment and sold to Charles Howard

Esquire who had it towed to Detroit to be converted into a sailing ship (Cleveland

Weekly Herald 1847, p. 3). The engine from Anthony Wayne was put into the steamer

Baltimore, newly constructed at Monroe, Michigan in 1847 (Cleveland Weekly Herald

1847, p. 3).

The hull was subsequently towed to Trenton, Michigan where it was extensively

rebuilt ‘from the keel up in 1848 and 9’ (The Daily Sanduskian 1850c, p. 2). It was

rebuilt by D. W. Donahue as a steamer not as a sailing vessel as originally planned (The

Daily Sanduskian 1850c, p. 2). New boilers manufactured by Wolcott & Co. at Detroit,

Michigan were installed in the spring of 1849 (The Daily Sanduskian 1850c, p. 2). The

engine installed in the rebuilt Anthony Wayne was a high pressure, 150 horsepower,

horizontal, crosshead, steam engine salvaged from the steamer Columbus (Heyl 1956, p.

93; Sandusky Clarion 1849, p. 2; Thunder Bay National Marine Sanctuary 2009b).

81

Anthony Wayne was re-enrolled at the Port of Detroit on 21 April 1949

(Nawmond 1849, p. 1). The new enrollment lists the rebuilt vessel as measuring 155 feet

(47.2 m) in length, 27 feet 4 inches (8.3 m) in breath, 10 feet (3.0 m) in depth and

measuring 400 and 80/95 tons (363 tonnes) (Nawmond 1849, p. 1). The enrollment

further indicates that the rebuilt vessel had one deck, one mast, no galleries, and a plain

head (Nawmond 1849, p. 1). The new, rebuilt Anthony Wayne was significantly different

from the original Anthony Wayne. Some of the major changes included a steam engine

change from a vertical style engine to a horizontal style engine and a reduction in the

number of decks from the original three deck design to a design that featured only one

deck.

Anthony Wayne re-entered passenger and freight service in 1849 running on the

Toledo line (Toledo to Sandusky to Cleveland, Ohio, to Buffalo, New York and return)

(Buffalo Morning Express 1849, p. 2). It continued in that service until 28 April 1850,

when some seven miles (11 km) north of Vermilion, Ohio when its boilers exploded

(The Daily Sanduskian 1850a, p. 2).

Anthony Wayne had arrived in Sandusky, Ohio the previous morning from

Toledo, Ohio on its normal Toledo to Buffalo run (Mansfield 1899, p. 660). It left

Sandusky at 10:00 pm bound for Cleveland when, about two and a half hours later, north

of Vermilion, the starboard boilers exploded (Mansfield 1899, p. 660). The explosion

was so violent that it blew the two under-deck starboard boilers up through the main

deck where they landed perpendicular to the hull (Mansfield 1899, p. 660). The cabins

above the boilers were torn away and the hull badly damaged (Mansfield 1899, p. 660).

The ship sank within ten minutes of the explosion (The Daily Sanduskian 1850b,

p. 2). ‘When the Wayne went down, she was on fire, and the flames were just bursting

out…’ (The Janesville (Wis.) Gazette 1850, p. 4) and ‘…the hurricane deck separated

from it and remained afloat, held to the boat by the rudder chains, the mast and the

shrouds’ (The Daily Sanduskian 1850b, p. 2).

82


Dating

No artifacts have been recovered or dated. This category is scored a 0.


The AM/PM data comparison (Table 9) is shown as follows:

Table 9: Anthony Wayne AM/PM Data Comparison

AM Ship Data PM Archeological Data

Sidewheel steamer style construction Sidewheel steamer style construction Horizontal steam engine Horizontal steam engine Distance from bow stem to paddlewheel axis calculated from iconography: 80.75 feet (Mariners Museum image), 92.5 feet (Heyl scaled image)

Distance from bow stem to paddlewheel axis measured: 92 feet

Wooden construction Wooden construction Carvel Built Carvel Built Location of accident: about 18 miles (29 km) from the mouth of Sandusky Bay and about 8 miles (13 km) north of Vermilion, Ohio

Location of accident: 16.7 miles (26.9 km) from the mouth of Sandusky Bay and 7.3 miles (11.7 km) north of Vermilion, Ohio

Severe damage due to boiler explosion and fire: upper works, cabins, and hurricane deck blown off during explosion and sinking

No remains of the upper works present. Only remains, excluding paddlewheels, from paddlewheel axles and below at midship section of wreck and stem and below, minus decking, at bow section

Although at this time the archaeological data is sparse, the available data

comparison is consistent and the category is scored a 1.

83


No cargo or cargo handling equipment has been found to date. This category is

scored a 0.


No crew personal effects have been found to date. This category is scored a 0.

Location

The location the wreck site is consistent with the calculated and reported position

ascertained from the AM data. A search of shipwreck databases show no other ships of

this type are known to be lost in this area. This category is scored a 1.


The sidewheel steamer Anthony Wayne suffered a massive boiler explosion

which threw the two starboard boilers ‘into a perpendicular position, tearing away the

steerage cabin above, and shattering the hull badly’ (Mansfield 1899, p. 680). ‘When

the boat sank the hurricane deck separated from it and remained afloat’ (The Daily

Sanduskian 1850b, p. 2). The condition of the wreck, missing all of the upper works,

cabins and decking, is consistent with a massive explosion.

The condition of the shipwreck and site is consistent with the AM data. This

category is scored a 1.

84



found yields a final score for the candidate ship for identification, Anthony Wayne, of 1.

This rating is identification: The historical and archaeological data in all categories

where comparisons are possible are consistent in sufficient detail to establish that they

belong to the proposed ship and there are no irreconcilable discrepancies. The shipwreck

is the ship proposed.

Table 10: Anthony Wayne Shipwreck Identification Matrix

Shipwreck Identification Matrix Candidate Ship for Identification Anthony Wayne AM to PM Shipwreck Identification Categories Data

Consistency Scoring

Dating 0




Location 1



85

Conclusions

Case Study 4 used the HSIM to determine if the wreckage discovered 16.7 miles

(26.9 km) from the mouth of Sandusky Bay and 7.3 miles (11.7 km) north of Vermilion,

Ohio was that of Anthony Wayne. The AM and PM data were gathered and compared for

consistency. The vessel construction and rigging, location, and condition of the

shipwreck categories all indicated a consistent match between the AM and PM data. The

consistency of the data match was sufficient enough in detail to establish the

identification of the wreckage as Anthony Wayne.

86

Chapter 5 Discussion of Results

Case Study 1

This case study applied the HSIM to a shipwreck that was known not to be the

one posited for identification. The shipwreck was previously identified as ‘a variation of

a Mackinaw boat’ by Hundley (Hundley 1984, p. 40). Hundley’s PM data was used in

the application of the HSIM in this case study. Comparing the application and results of

the HSIM to the application and results of Hundley’s prior identification effort will serve

to demonstrate the validity of the method and, at the same time, underscore the

importance of good AM data collection techniques and research to the identification

process.

There was ample PM data available to Hundley for comparison to the AM data,

but the AM data, specifically locational information and vessel construction, was not as

thoroughly researched as it could have been. Hundley based much of his AM data on the

articles written by Murphy and relied on Murphy’s and others’ work with primary

source documents available for Griffon instead of researching the primary source

documents directly (Hundley 1984, p. var.). This is not necessarily a bad research

practice if the primary source documents are not readily available. It does, however,

introduce the possibility for overlooked, omitted, misinterpreted, and/or misstated

information leaving the value of the author’s work ultimately reliant on the best and,

quite possibly, the worst of another’s intentions and efforts.

The location of a wreck plays a key role in the identification process if adequate

and believable AM, locational data is available. At the extremes, this data can either

eliminate a posited candidate or add to the confidence level of a vessel posited for

identification. Believable data is data that does not have an ulterior motive associated

with it. This is often very hard to determine but comes into play in cases where the

87

people reporting the loss are not truthful for self-serving reasons such as scuttlings for

insurance purposes, smuggling, or simple avoidance of potential blame or liability.

Location contributed significantly to the conclusion in Case Study 1 that the

Griffon Cove wreck was not Griffon. While Hundley came to the same conclusion, his

case study did not use location as a parameter for determining the identification of the

wreck. While this did not significantly alter his conclusion, using locational information

certainly would have strengthened his position. Even though location played no role in

his analysis, Hundley presented several possible locations in his Historical Background

section stating that Griffon had been boarded by Indians and burned five or six leagues

from its anchorage and that the crew mutinied and sailed it to the west (Hundley 1984, p.

6). With just this AM information one could have concluded that Griffon probably sank

in Lake Michigan not in Griffon Cove, Lake Huron. If Hundley had researched the AM

data more thoroughly for location information he would have discovered several key

pieces of locational evidence including eyewitness accounts that saw Griffon struggle in

a storm and wreck near the Huron Islands, Lake Michigan (Anderson 1901, pp. 45-7)

and the subsequent discovery of some wreckage associated with the ship (Marshall

1879, p. 287). This data, when compared to the PM location of the wreck site, would

have bolstered his conclusion, as it did this author’s, that the Griffon Cove Wreck, found

in Lake Huron, was not Griffon.

A vessel’s construction also plays an important role when acquiring AM data for

the HSIM. Hundley’s AM construction data for Griffon was based on the primary source

data research of others. While this AM data was sufficient enough for Hundley to make

a comparison to the PM data, he overlooked several available AM data points used by

this author in Case Study 1 that would have further enhanced his analysis. These data

points are the iconograph of Griffon depicted in an early map of the Niagara area from

1688 (Figure 2) (Remington 1891, p. 43) and the possible rigging of the vessel from an

article in the Marine Review and Marine Record (Figure 3) (Joy 1903, p. 24). This data,

missed by Hundley, strengthened the conclusion that the Griffon Cove wreck is not

Griffon.

88

This case study highlights the importance of the Location and Vessel

Construction/Rigging categories to the identification process and emphasizes the

associated need for good AM data collection techniques and research. While the

comparisons of Hundley’s work to this author’s may imply that Hundley’s AM data

research and collection efforts were less than thorough, it must be recognized that the

existence and availability of historical information is no doubt much easier to ascertain

in 2009 than it was in 1984 when Hundley performed his analysis. Although all of the

AM data used by this author in this case study existed in 1984, Hundley certainly was at

a comparative disadvantage relative to locating it. When consulting, considering, and/or

using the work of ‘those that came before’ it is important to remember that such work

utilized nothing more than ‘the best information available at the time’.

Case Study 2

This case study applied the HSIM to the Beaufort Inlet shipwreck that has been

tentatively identified as La Concorde, which was renamed Queen Anne’s Revenge by

Blackbeard the pirate. Although there are thousands of artifacts and remains that

‘suggest’ that the wreckage is La Concorde, Case Study 2 highlights the importance of

data evaluation transparency that the HSIM fosters. This allows for an ‘objective’, rather

than ‘subjective’, evaluation of the AM to PM data. It also reinforces the importance of

having adequate AM data for the evaluation process.

The HSIM identified three categories that raise doubts as to the wreck’s identity.

These are: Dating, Vessel Construction and Rigging, and Cargo and Cargo

Handling Equipment. Evaluation of the first category, Dating, showed that a question

remained regarding the date of cannon number three which possibly dates to 1730

(Lusardi 2006, p. 202), well past the 1718 wrecking date of La Concorde. If this date is

resolved and is determined to be prior to 1718, the category could be reevaluated and

scored differently. This shows that the HSIM follows the ‘scientific method’ permitting

89

new or subsequent data to be added to the original hypothesis thus providing a possible

different outcome based on a different data set upon reevaluation.

The second category, Vessel Construction and Rigging, was evaluated and no

conclusion could be made due to inconsistencies in the AM/PM data comparison. This

can possibly be attributed to the lack of AM data available for La Concorde prior to

1710. The earliest historical records indicate that the frigate La Concorde was acquired

in 1710 by Montaudouin during the War of Spanish Succession (Lawrence 2009a).

Whether it was of Spanish, French or English construction is not known. This makes the

comparison of AM to PM data based solely on the available AM data as of 1710. If it

was of Spanish construction, the Spanish bell and pump sieve would tend to support this.

If, on the other hand, La Concorde was of English or French construction, the Spanish

bell and pump sieve would tend to contradict this and possibly rule out La Concorde as

the identity of the subject wreck.

The evaluation of the third category, Cargo and Cargo Handling Equipment,

also yielded several inconsistencies. The first of these is the origin of the various

cannon. If this was a French frigate why was it armed with English-manufactured

cannon when the French and English were at war with each other? Was this an English

or Spanish frigate that was captured during the war? The lack of AM data relative to the

early history of the ship makes the answers to such questions purely speculative.

Without adequate AM data it is impossible to make a valid comparison to the PM data

and derive any logical, related conclusions.

The second anomaly in this data is the flatware artifacts that have so far been

recovered and analyzed (The QAR Project 2009i). The analysis of these artifacts points

to the ship being of English origin (The QAR Project 2009i). Again, not enough AM

data is available regarding the ship’s early history to permit a more detailed comparison.

Interestingly and as previously reported in Chapter 4, the flatware does not bear any

markings to associate it with La Concorde or any other vessel. The artifacts recovered

from Whydah and Henrietta Maria, both dating from the same period of time as La

90

Concorde, bear the initials of the ships name leading one to expect La Concorde’s

flatware to be similarly inscribed (Lusardi 2006, p. 217).

The final inconsistency is the lack of slave trade goods. La Concorde had been

involved in the slave trade for a number of years and was transporting slaves when

captured by Blackbeard. La Concorde wrecked less than a year later yet no significant

evidence of the slave trade has been found at the site. The absence of slave trade goods

compared to the volume of such artifacts found on both Whydah (Hamilton 2007, p.

144), a former slaver captured by pirates, and Henrietta Maria (Moore & Malcom 2008)

leads one to believe the wreck found had not been involved in the slave trade. This

category was scored as no conclusion because of these inconsistencies in the AM/PM

data comparison.

In addition to the above there is one, third-hand and therefore dubious, report that

Blackbeard may have switched his flag from the French La Concorde to an English ship

prior to heading for the east coast of North America (Wilde-Ramsing 2006, p. 192). If

that was the case then La Concorde is certainly not the wreck found at Beaufort Inlet

suggesting the need for additional and detailed AM data research to systematically

determine the wreck’s identity.

Several other possibilities warrant consideration. The wreck that was found may

be the Spanish snow El Salvador. It was of Spanish construction and would have had a

Spanish bell and Spanish pump sieve. El Salvador was also reported to have sunk in the

same area and its wreckage may also be present on the site comingled with wreckage

from another vessel. This could be pure coincidence or the result of numerous hurricanes

that have struck the wreck site over the years. These hurricanes could have transported,

‘collected’, and consolidated artifacts, wreckage, and almost anything else in their paths

into various underwater ‘junkyards’.

While the outcome of Case Study 2 regarding the identification of the subject

shipwreck as La Concorde was inconclusive, the case study served to demonstrate the

91

importance and value of both AM and PM data. It becomes readily apparent that the

accumulation and analysis of quality data is necessary to arrive at a confident

conclusion. At the same time, however, this quality data may actually have the opposite

effect by creating more questions than providing good answers.

Case Study 3

This case study applied the HSIM to a shipwreck that was discovered based on

locational information. The posited identity of the wreck was postulated to be the bark

Cortland, a ship that was involved in a collision with the sidewheel steamer Morning

Star on Lake Erie. This study shows the importance of the Location and Condition of

Shipwreck and Site categories in the identification process when adequate AM data is

available. It also proves that identification is possible without detailed PM data.

This wreck was found using the AM locational data obtained through historical

research. This research yielded information that allowed the approximate position of the

wreck to be calculated and a search area determined. A subsequent search followed and

a wreck was discovered at that approximate location. Further research of the historical

record and wreck databases determined that this was the only ship of its type known to

be lost in that general area. This illustrates that location can be a very strong indicator of

a vessel’s identity if adequate AM data for the vessel is available.

The condition of the shipwreck and the wreck site conditions are also key pieces

of evidences if adequate AM data is available. In this case, salvage divers working on

the wreck after its loss described the wreck’s condition and position on the bottom in

addition to the bottom conditions at the wreck site (Cleveland Daily Herald 1868c, p. 3).

This provided enough AM data to allow a meaningful comparison to be made to the PM

data that was gathered from the wreck site. The comparison of the AM to PM data was

vary favorable and led to rating the category as being consistent with all available data.

92

The historical research conducted while gathering AM data uncovered a very

detailed description of the vessel’s construction (Milwaukee Sentinel 1867, p. 3). This is

normally not the case for vessels early in the historical record. The description allowed

the vessels identity to be determined early in the survey process by identifying certain

key features with just a basic reconnaissance underwater survey. As a result, future

detailed survey work can be focused on the vessel and its contents as no further survey

work is necessary relative to its identity and origins.

Case Study 4

This case study applied the HSIM to a shipwreck that was also discovered based

on locational information. The posited identity of the wreck was postulated to be the

sidewheel steamboat Anthony Wayne which caught fire and sank in Lake Erie after its

boilers exploded. The case study shows the importance of the Location, Vessel

Construction and Rigging, and Conditions of the Shipwreck and Site categories in

the identification process when adequate AM data is available. It also demonstrates that

identification is possible without detailed PM data.

As in Case Study 3 this wreck was found using the AM locational data obtained

through historical research. The research allowed the approximate position of the wreck

to be calculated and a search area determined. A wreck was discovered in the search

area near the approximate location during a subsequent search. Further research of the

historical record and wreck data bases determined that this was the only ship of its type

known to have been lost in that general area. Again, as in Case Study 3, this illustrates

that location can be a very strong indicator of a vessel’s identity if adequate AM data for

the vessel is available.

The evaluation of the vessel’s construction and rigging played a key part in the

identification as well. Extensive AM research efforts focused on the vessel’s

construction details and yielded enough specific information regarding the configuration

93

of the vessel at the time of loss to make this possible. Anthony Wayne was originally

constructed in 1837 with a vertical, crosshead, high pressure, steam engine (US Treasury

Department 1838, p. 339). Interestingly, had the original and oft-cited construction data

been used in this case to evaluate this category, the comparison would have excluded the

wreck from being Anthony Wayne. The originally observed paddlewheel drive

mechanisms along with the horizontal position of the Pittman arm indicated a vessel

powered by a horizontal steam engine. Subsequent excavation of the wreck and the

discovery of a high pressure, horizontal, crosshead, steam, engine proved this to be the

case. Further research into Anthony Wayne’s history uncovered an accident in 1847

(Heyl 1956, p. 99) that left Anthony Wayne valuable only for its parts. The original

engine, boilers, and other machinery were removed (Cleveland Weekly Herald 1847, p.

3) and installed in other vessels. The hull was eventually sold and completely rebuilt

(The Daily Sanduskian 1850c) with a high pressure, horizontal, crosshead, steam, engine

salvaged from yet another vessel (Heyl 1956, p. 93; Sandusky Clarion 1849, p. 2;

Thunder Bay National Marine Sanctuary 2009b). Had this extensive AM research not

been conducted the outcome of the identification would have been entirely different.

The condition of the shipwreck was also a key piece of evidence in determining

the wreck’s identification. Although there were no reports available from the historic

record on the condition of the wreck, there was enough indirectly related information in

the accounts of the accident to allow intelligent speculation regarding the condition of

the shipwreck. This speculative AM data, when compared to the PM data, shows that the

vessel’s speculated condition is consistent with the actual condition of the wreck on the

bottom.

As stated previously, the sidewheel steamer Anthony Wayne suffered a massive

boiler explosion which threw the two, starboard boilers ‘into a perpendicular position,

tearing away the steerage cabin above, and shattering the hull badly’ (Mansfield 1899, p.

680). ‘When the boat sank the hurricane deck separated from it and remained afloat’

(The Daily Sanduskian 1850b, p. 2). The condition of the wreck, missing all of the upper

works, cabins and decking, is consistent with a massive explosion. This circumstantial

94

evidence based on the AM data was compelling enough to rate this category consistent

with the PM data observed and recorded.

In this case study the identification posited was determined to be a positive

identification. The AM data associated with the location, vessel construction, and

condition categories was more than sufficient in quantity and quality, when compared to

the minimal amount of PM data, to allow such a determination to be made with a high

degree of confidence.

Conclusions

Chapter 5 contains detailed discussions on the application and thought processes

that went into applying the HSIM to the four different case studies. Case Study 1 points

out the importance of thorough AM data research along with the significance of

locational data when trying to determine the identification of a shipwreck. Case Study 2

emphasizes the point that a detailed history of the vessel is almost paramount in the

identification process as without it the comparison between AM and PM data becomes

less meaningful and there is a high probability that the results will be inconclusive. Case

Studies 3 and 4 further reinforce the concept of good locational data being a key to

identification when little PM data is available.

95

Chapter 6 Conclusions

The development and application of the HSIM demonstrates that the ‘traditional’

forensic science approach to identifying deceased individuals can be adapted

successfully to the identification of shipwrecks. Utilizing the principles of the scientific

method, this method provides a systematic approach to achieve the goal of identification

provided sufficient historical and archaeological data are available for analysis. The

method relies on a structured comparison and analysis of the acquired PM

archaeological data to the AM historical data and attempts to remove as much

subjectivity from the identification process as possible.

The HSIM is by no means an infallible method for determining a shipwreck’s

identification and it certainly has its limitations. As with any method, the outcome is

influenced by the quality and critical evaluation of the data used during its application. If

limited or bad data is evaluated, the conclusion drawn will be without substance or value

and will be either incorrect or erroneously correct, although this will not be apparent.

Sufficient and good data are requisites for the method. No method, no matter how

simple or complex, can compensate for insufficient, erroneous, or misinterpreted data.

One of the strong points of the systematic approach to data analysis used in the proposed

method is that errors caused by the insufficient or incorrect evaluation of data are more

apparent.

The HSIM was developed using five major categories for evaluation of the

AM/PM data: Dating, Vessel Construction and Rigging, Cargo and Cargo Handling

Equipment, Crew Personal Effects, Location, and Condition of Shipwreck and Site.

These categories were chosen in an attempt to globally capture all the possible data that

could be expected from the AM and PM research efforts. These categories are not

necessarily set in stone and can be modified to suit any particular shipwreck where the

adjustment or addition of other major categories (i.e. armament, cooking, passengers,

etc) appears appropriate. When necessary, the major categories can be broken down into

sub-categories without affecting the outcome of the data analysis. This may be needed

96

when there is an abundance of data, as in the case of La Concorde, to allow a more

detailed evaluation of certain subsets. It is important to note that recognizing the value

and sufficiency of both AM and PM data is certainly not as simple as it sounds and the

parameters for such change from case to case.

The HSIM is not limited to a certain type of shipwreck; steamers, schooners, or

dugout canoes can be tested. It is applicable to any wreck where sufficient historical

information is available for direct comparison to the archaeological data obtained. The

four case studies presented herein show that the method can be successfully applied to

shipwrecks of various types, from different time periods, and in varying states of

preservation to produce an identification or exclusion from identification for a vessel

that has been posited for identification.

The HSIM provides an application framework that will enable the archaeologist

to systematically attempt to identify a shipwreck in a straight forward, methodological

fashion. It removes as much subjectivity from the identification process as possible and

provides a clearer understanding of how the identification result was derived.

97

Appendix 1: AM Historical Data Worksheet for Wooden Ships

98

AM Historical Data Worksheet for Wooden Ships

Vessel Name: Name of ship: __________________________________________

Previous Names (List chronologically starting from original build): ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

Source(s): __________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

Original Construction Data: Enrollment Number: __________________________________________

Type of Vessel:

___ Scow ___ Schooner ___ Bark

___ Brig ___ Barge ___ Scow-schooner

___ Schooner-barge ___ Tug ___ Propeller

___ Tanker ___ Steam ___ Dredge

___ Whaleback ___ Ferry

___ Paddlewheel Steam Boat

___ Other: __________________________________________________

Date Built: _________________________________________________

Where Built: ________________________________________________

Enrollment Date: ____________________________________________

Enrollment Port: _____________________________________________

Home Port: _________________________________________________

99

Official Number: _____________________________________________

Builder: ____________________________________________________

Owner: ____________________________________________________

Captain: ___________________________________________________

Length: ____________________________________________________

Beam: _____________________________________________________

Number of Decks: ___________________________________________

Number of Masts: ____________________________________________

Net Tonnage: _______________________________________________

Gross Tonnage:______________________________________________

Number of Galleries: _________________________________________

Type of Head:

___ Plain ___ Figure ___ Scroll ___ Billet ___ None

Other: _____________________________________________________

Hull Construction:

Wood: ___ Carvel ___ Clinker ___ Arched

___ Composite: ______________________________________

___ Fastening Method: ________________________________

___ Other: __________________________________________

Metal: ___ Iron (Riveted) ___ Steel (Riveted)

___ Steel (Welded) ___ Aluminum

___ Other: __________________________________________

Hull Sheathing:

___ Copper ___ Wood ___ Tin ___ Lead ___ Steel

___ Other: __________________________________________

Hull Style:

___ Scow ___ Double ender ___ Plumb Stem

___ Clipper Stem ___ Counter Stem ___ Transom Stern

___ Round Stern ___ Square Stern ___ Center Board

___ Fixed Keel ___ Flat Bottom ___ Round Bottom

___ V Bottom

___ Other: _______________________________________________

100

Number of Hatches and dimensions:______________________________

Number of Center Boards and dimensions:_________________________

Propulsion:

___ Sail ___ Steam ___ Gas ___ Diesel ___ Other: _______________________________________________

Sail:

Number of Masts: _________________________________________

Standing Rigging: _________________________________________

Shrouds:

___ Rope ___ Wire

Shroud Attachment:

___ Deadeye ___ Turnbuckle

Steam:

Number of Engines: _________

Type of Engine:

___ High Pressure ___ Low Pressure

___ Vertical ___ Horizontal

___ Steeple ___ Inverted

___ Cross Head ___ Single Cylinder

___ Compound ___ Triple expansion

___ Walking beam ___ Side-lever

___ Grasshopper ___ Oscillating

___ Inclined ___ Trunk

___ Condensing ___ Non-Condensing

___ Other: __________________________________________

Number of Boilers: ___________________________________________

Boiler Material: ______________________________________________

Other Boiler Information: ______________________________________

___________________________________________________________

___________________________________________________________

101

Gas/Diesel/Other:

___ Description: _________________________________________

_________________________________________

_________________________________________

_________________________________________

Paddlewheels:

Dimensions: ________________________________________________

___ Sidewheel ___ Sternwheel ___ Fixed Pitch

___ Variable Pitch

___ Other: _________________________________________________

Propellers:

Number: ___________________________________________________

Dimension: _________________________________________________

Number of Blades: ___________________________________________

Type: ______________________________________________________

Material: ___________________________________________________

Other: _____________________________________________________

Deck Equipment:

___ Capstan(s) Number: ____

Manufacturer’s Data: __________________________________________

___ Windlass(s) Number: ____


___ Anchors Number: ____ Weight: _____________

Anchor Chain Data: __________________________________________

___ Sheet Winches(s) Number: ____


___ Bell(s) Number: ____


___ Pump(s) Number: ____


102

Other Information: ___________________________________________

___________________________________________________________

___________________________________________________________

___________________________________________________________

___________________________________________________________

103

Rebuild Data: Re-enrollment Number: _______________________________________

Type of Vessel:

___ Scow ___ Schooner ___ Bark

___ Brig ___ Barge ___ Scow-schooner

___ Schooner-barge ___ Tug ___ Propeller

___ Tanker ___ Steam ___ Dredge

___ Whaleback ___ Ferry

___ Paddlewheel Steam Boat

___ Other: __________________________________________________

Date Built: _________________________________________________

Where Built: ________________________________________________

Enrollment Date: ____________________________________________

Enrollment Port: _____________________________________________

Home Port: _________________________________________________

Official Number: _____________________________________________

Builder: ____________________________________________________

Owner: ____________________________________________________

Captain: ___________________________________________________

Length: ____________________________________________________

Beam: _____________________________________________________

Number of Decks: ___________________________________________

Number of Masts: ____________________________________________

Net Tonnage: _______________________________________________

Gross Tonnage:______________________________________________

Number of Galleries: _________________________________________

Type of Head:

___ Plain ___ Figure ___ Scroll ___ Billet ___ None

Other: _____________________________________________________

104

Hull Construction:

Wood: ___ Carvel ___ Clinker ___ Arched

___ Composite: ______________________________________

___ Fastening Method: ________________________________

___ Other: __________________________________________

Metal: ___ Iron (Riveted) ___ Steel (Riveted)

___ Steel (Welded) ___ Aluminum

___ Other: __________________________________________

Hull Sheathing:

___ Copper ___ Wood ___ Tin ___ Lead ___ Steel

___ Other: __________________________________________

Hull Style:

___ Scow ___ Double ender ___ Plumb Stem

___ Clipper Stem ___ Counter Stem ___ Transom Stern

___ Round Stern ___ Square Stern ___ Center Board

___ Fixed Keel ___ Flat Bottom ___ Round Bottom

___ V Bottom

___ Other: _______________________________________________

Number of Hatches and dimensions:______________________________

Number of Center Boards and dimensions:_________________________

Propulsion:

___ Sail ___ Steam ___ Gas ___ Diesel ___ Other: _______________________________________________

Sail:

Number of Masts: _________________________________________

Standing Rigging: _________________________________________

Shrouds:

___ Rope ___ Wire

Shroud Attachment:

___ Deadeye ___ Turnbuckle

105

Steam:

Number of Engines: _________

Type of Engine:

___ High Pressure ___ Low Pressure

___ Vertical ___ Horizontal

___ Steeple ___ Inverted

___ Cross Head ___ Single Cylinder

___ Compound ___ Triple expansion

___ Walking beam ___ Side-lever

___ Grasshopper ___ Oscillating

___ Inclined ___ Trunk

___ Condensing ___ Non-Condensing

___ Other: __________________________________________

Number of Boilers: ___________________________________________

Boiler Material: ______________________________________________

Other Boiler Information: ______________________________________

___________________________________________________________

___________________________________________________________

Gas/Diesel/Other:

___ Description: _________________________________________

_________________________________________

_________________________________________

_________________________________________

Paddlewheels:

Dimensions: __________________________________________

___ Sidewheel ___ Sternwheel ___ Fixed Pitch

___ Variable Pitch

___ Other: _______________________________________________

106

Propellers:

Number: ________________________________________________

Dimension: ______________________________________________

Number of Blades: ________________________________________

Type: ___________________________________________________

Material: ________________________________________________

Other: __________________________________________________

Deck Equipment:

___ Capstan(s) Number: ____

Manufacturer’s Data: _______________________________________

___ Windlass(s) Number: ____


___ Anchors Number: ____ Weight: _____________

Anchor Chain Data: _______________________________________

___ Sheet Winches(s) Number: ____


___ Bell(s) Number: ____


___ Pump(s) Number: ____


Other Information: ________________________________________

________________________________________________________

________________________________________________________

________________________________________________________

________________________________________________________

Cargo Data: Type, Owner, Consignee: ____________________________________

_________________________________________________________

_________________________________________________________

_________________________________________________________

107

Incidents: accidents, groundings, sinkings, fires, etc. (give dates and sources)

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

Final Sinking: Where: ________________________________________________________________

________________________________________________________________

When: ________________________________________________________________

Why: ________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

Sources: ________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

Crew: Names of crew and their positions

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

108

Official Report Data: Government, insurance, coroners, judicial, etc. (Give dates and sources)

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

Construction Drawing Data: Construction drawings found ___ Yes ___ No

________________________________________________________________

________________________________________________________________

________________________________________________________________

Imagery: ___ Photos ___ Drawings

Sources: ________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

Inventory List: (Give dates and sources)

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

109

Manifest: (Give dates and sources)

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

Provisions: (Give dates and sources)

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

Personal belongings: (Give dates and sources)

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

110

Models: (Give location, dates, and sources)

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

________________________________________________________________

Survivor and eyewitness accounts: (Give dates and sources)

_______________________________________________________________

_______________________________________________________________

_______________________________________________________________

_______________________________________________________________

_______________________________________________________________

_______________________________________________________________

_______________________________________________________________

_______________________________________________________________

_______________________________________________________________

_______________________________________________________________

_______________________________________________________________

_______________________________________________________________

_______________________________________________________________

**Note: List all information sources, enrollments, newspaper articles, books, official reports and correspondence, official documents or other sources that the information was obtained.

111

Appendix 2: Shipwreck Identification Matrix with Rating Key

112

Rating Key:

Data Consistency Scoring:

0 - Historical and/or archaeological data is not available or is insufficient to perform a meaningful comparison 1 - The historical and archaeological data are consistent in sufficient detail to establish that they belong to the proposed ship and there are no irreconcilable discrepancies 2 - The historical and archaeological data are not consistent in sufficient detail to establish that they belong to the proposed ship 3 - The historical and archaeological data are clearly inconsistent

Final Scoring:

1 - Identification: The historical and archaeological data in all categories where comparisons are possible are consistent in sufficient detail to establish that they belong to the proposed ship and there are no irreconcilable discrepancies. The shipwreck is the ship proposed 2 - No conclusion: The historical and archaeological data in all or part of the categories where comparisons are possible are not consistent in sufficient detail to establish that they belong to the proposed ship. 3 - Exclusion: The historical and archaeological data in all or part of the categories where comparisons are possible are clearly inconsistent. The shipwreck is not the ship proposed.

Shipwreck Identification Matrix Candidate Ship for Identification Ships Name Here AM to PM Shipwreck Identification Categories Data

Consistency Scoring

Dating




Location


Final Scoring (Highest number in right hand column)

113

References

Ahlström, C 1997, Looking for Leads: Shipwrecks of the past revealed by contemporary documents and the arachaeological record, First edn, Humaniora, The Finnish Academy of Science and Letters, Helsinki.

Anderson, MBT 1901, Relation of the Discoveries and Voyages of Cavelier de La Salle from 1679 to 1681: The Official Narrative, Chicago: The Caxton Club, 1901, viewed November 6 2007, <http://www.americanjourneys.org/aj-122/index.asp>.

Auer, J 2004, 'Fregatten Mynden: a 17th-century Danish Frigate Found in Northern Germany', International Journal of Nautical Archaeology, vol. 33, no. 2, pp. 264-80.

Baraybar, JP 2008, 'When DNA is Not Available, Can We Still Identify People? Recommendations for Best Practice', Journal of Forensic Sciences, vol. 53, no. 3, pp. 533-40.

Beltrame, C & Gaddi, D 2002, 'Report on the first research campaign on the Napoleonic brick, Mercure, wrecked off Lignano, Udine, Italy in 1812', The International Journal of Nautical Archaeology, vol. 33.1, pp. 60-73.

Buffalo Morning Express 1849, 'The Wayne', Buffalo Morning Express, 27 April, 1849, p. 4.

Carnes-McNaughton, L 2008, Ceramic Assemblage Analysis from Shipwreck 31CR314 Queen Anne's Revenge Site, QAR-R-08-03, Underwater Archaeology Branch, Office of State Archaeology, Department of Cultural Resources, State of North Carolina.

Carnes-McNaughton, L & Myers, SG 2007, Beads Recovered from Shipwreck 31CR314 Queen Anne's Revenge Site, QAR-B-07-02, Underwater Archaeology Branch, Office of State Archaeology, Department of Cultural Resources, State of North Carolina.

Carnes-McNaughton, L & Wilde-Ramsing, MU 2008, Preliminary Glassware and Bottle Analysis from Shipwreck 31CR314 Queen Anne's Revenge Site, QAR-R-08-02, Underwater Archaeology Branch, Office of State Archaeology, Department of Cultural Resources, State of North Carolina.

Cleveland Daily Herald 1868a, 'The Morning Star Disaster', Cleveland Daily Herald, June 24, 1868, p. 3.

Cleveland Daily Herald 1868b, 'Terrible Disaster', Cleveland Daily Herald, June 22, 1868, p. 4.

114

Cleveland Daily Herald 1868c, 'The Wreck of the Cortland', Cleveland Daily Herald, July 11, 1868, p. 3.

Cleveland Daily Herald & Gazette 1838, 'Perrysburg and Buffalo Daily Line of Steam Boats', Cleveland Daily Herald & Gazette, 7 March 1838, p. 4.

Cleveland Leader 1868a, 'The Collision', Cleveland Leader, June 23, 1868, p. 4.

Cleveland Leader 1868b, 'The Morning Star Disaster', Cleveland Leader, June 26, 1868, p. 4.

Cleveland Leader 1868c, 'Terrible Lake Disaster', Cleveland Leader, June 22, 1868, p. 3.

Cleveland Weekly Herald 1847, 'Arrival', Cleveland Weekly Herald, 17 November 1847, p. 4.

The Daily Sanduskian 1850a, 'Coroner's Inquest', The Daily Sanduskian, April 30, 1850, p. 2.

The Daily Sanduskian 1850b, 'Frightful Diaster!' The Daily Sanduskian, April 29, 1850, p. 2.

The Daily Sanduskian 1850c, 'Steam-Boat Gen, Wayne', The Daily Sanduskian, May 2, 1850, p. 4.

Hamilton, CE 2007, 'The Pirate Ship Whydah', in RK Skowronek & CR Ewen (eds), X Marks the Spot: The Archaeology of Piracy, University Press of Florida, Gainesville, Florida, pp. 131-59.

Hennepin, L 1903, A New Discovery of a Vast Country in America, Chicago: A.C. McClurg & Co., viewed November 6 2007, <http://www.americanjourneys.org/aj-124/index.asp>.

Heyl, E 1956, Early American Steamers, vol. II, VI vols., Erik Heyl, Buffalo.

Heyl, E 1964, Early American Steamers, vol. III, VI vols., Erik Heyl, Buffalo.

Horrell, CE 2009, 'The British Merchant Ship Western Empire: A Lesson from the Archives', Mineral Management Service (MMS).

115

Hundley, PF 1984, 'The Griffon Cove Wreck: a case study in archaeological reconstruction of hull timber remains', Australian Institute for Maritime Archaeology, no. Special Publication No. 2.

The Janesville (Wis.) Gazette 1850, 'Steamboat Explosion', The Janesville (Wis.) Gazette, May 9, 1850, p. 4.

Joy, RP 1903, 'The Griffin', Marine Review and Marine Record, vol. 28, no. 8, February 12, p. 24.

Kennard, J 2009, 'Discovery of the Steamer Homer Warren', Inland Seas, Quarterly Journal of the Great Lakes Historical Society, vol. 63, no. 2, pp. 126 - 35.

Laughton, JK 1898, Teach, vol. LVI, Dictionary of National Biography, Smith, Elder, & Co., London.

Lawrence, RW 2008, An Overview of North Carolina Shipwrecks with an Emphasis on Eighteenth-Century Vessel Losses at Beufort Inlet, QAR-R-08-01, Underwater Archaeology Branch, Office of State Archaeology, Department of Cultural Resources, State of North Carolina.

Lawrence, RW 2009a, French Archival Research, viewed 4 April 2009, <http://www.qaronline.org/history/french.htm>.

Lawrence, RW 2009b, Preliminary Observations on British and American Documents Concerning the Activities of the Pirate Blackbeard, March 1717 to June 1718, The QAR Project, viewed 4 April 2009, <http://www.qaronline.org/history/timechart.htm>.

Lawrence, RW & Wilde-Ramsing, MU 2001, 'In Search of Blackbeard: Historical and Archaeological Research at Shipwreck Site 0003BUI', Southeastern Geology, vol. 40, no. 1, pp. 1-9.

Levin, JA 2006, 'Western Empire: The Deep Water Wreck of a Mid-Nineteenth Century Wooden Sailing Ship', Masters thesis, Texas A&M.

Lusardi, WR 2006, 'The Beaufort Inlet Shipwreck Artifact Assemblage', in RK Skowronek & CR Ewen (eds), X Marks the Spot: The Archaeology of Piracy, University Press of Florida, Gainsville, Florida, pp. 196-218.

Mansfield, JB 1899, History of the Great Lakes, 1st, reprint Freshwater Press, Cleveland, Ohio, 1972 edn, vol. 1, 2 vols., J. H. Beers & Co., Chicago.

116

Marshall, OH 1879, The Building and Voyage of the Griffon in 1679, vol. 1, Bigelow Brothers, Buffalo.

McDonald, WA 1958, 'The Morning Star and the Courtland', Inland Seas, Quarterly Journal of the Great Lakes Historical Society, vol. 14, no. 4, pp. 311-3.

Milstead Post, SA 2007, 'Definig Her Kind: An Historical and Archaeological Investigation of the Composite Built Gunvessel HMS Ready', East Carolina University.

Milwaukee Sentinel 1867, 'Barque Cortland', Milwaukee Sentinel, August 20, 1867, p. 3.

Moore, DD 2005, 'Technical Comments Relating to 'Ruling Theory' and the Identification of Beaufort Inlet Wreck', International Journal of Nautical Archaeology, vol. 34, no. 2, pp. 335-9.

Moore, DD & Malcom, C 2008, 'Seventeenth-Century Vehicle of the Middle Passage: Archaeological and Historical Investigations on the Henrietta Marie Shipwreck Site', International Journal of Historic Archaeology, vol. 12, pp. 20-38.

Morange, J 1718, Deposition upon Returning to Martinique, Serie Colonies C8A 24 fol. 125, Archieves Nationales, Paris.

Murphy, R 1955, 'Discovery of the Wreckage of the Griffon (Part I)', Inland Seas, vol. 11, no. 4, pp. 232-42.

Myers, FA 1956, 'The Manitoulin Griffon vs. the Tobermory Griffon', Inland Seas, Quarterly Journal of the Great Lakes Historical Society, vol. 12, no. 1, p. 300.

Co Customs 1849, Enrollment Anthony Wayne, by Nawmond, SJ, Port of Detroit.

Newsom, LA & Miller, RB 2009, Wood Species Analysis of Ships Timbers and Wooden Items Recovered from Shipwreck 31CR314, Queen Anne's Revenge Site, QAR-R-09-01, Underwater Archaeology Branch, Office of State Archaeology, Department of Cultural Resources, State of North Carolina.

O'Shea, JM 2004, 'The Identification of shipwreck sites: a Bayesian approach', Journal of Archaeological Science, vol. 31, pp. 1533-52.

The Plain Dealer 1868, 'Horrible Steamboat Accident', The Plain Dealer, June 22, 1868, p. 3.

The QAR Project 2009a, Anchors, viewed 28 March 2009, <http://www.qaronline.org/artifacts/anchors.htm>.

117

The QAR Project 2009b, Artifacts, viewed 18 March 2009, <http://www.qaronline.org/artifacts/artifacts.htm>.

The QAR Project 2009c, Artillery, viewed 28 March 2009, <http://www.qaronline.org/artifacts/artillery.htm>.

The QAR Project 2009d, Bells, viewed 19 September 2009, <http://www.qaronline.org/artifacts/bell.htm>.

The QAR Project 2009e, Cufflinks, viewed 28 October 2009, <http://www.qaronline.org/artifacts/cufflinks.htm>.

The QAR Project 2009f, Divider, viewed 29 March 2009, <http://www.qaronline.org/artifacts/divider.htm>.

The QAR Project 2009g, Gun Barrel, viewed 29 March 2009, <http://www.qaronline.org/artifacts/gunBarrel.htm>.

The QAR Project 2009h, Medicinal, viewed 29 March 2009, <http://www.qaronline.org/artifacts/medicinal.htm>.

The QAR Project 2009i, Pewter Flatware, viewed 29 March 2009, <http://www.qaronline.org/artifacts/pewterFlatware.htm>.

The QAR Project 2009j, Pipes, viewed 29 March 2009, <http://www.qaronline.org/artifacts/pipe.htm>.

The QAR Project 2009k, Shackle, viewed 29 March 2009, <http://www.qaronline.org/artifacts/restrainingDevice.htm>.

The QAR Project 2009l, Stemware, viewed 29 March 2009, <http://www.qaronline.org/artifacts/stemware.htm>.

The QAR Project 2009m, Structural Components, viewed 28 March 2009, <http://www.qaronline.org/artifacts/structuralComponents.htm>.

The QAR Project 2009n, Tools, viewed 29 March 2009, <http://www.qaronline.org/artifacts/tools.htm>.

Remington, CK 1891, The Ship-Yard of the Griffon, J. W. Clement, Buffalo.

118

Rogers, BA, Richards, N & Lusardi, WR 2005, ''Ruling Theories Linger': Questioning the Identity of the Beauford Inlet Shipwreck', The International Journal of Nautical Archaeology, vol. 34, no. 1, pp. 24-37.

Sandusky Clarion 1849, 'The Steamer Anthony Wayne', Sandusky Clarion, May 7, 1849, p. 4.

The Sheboygan Press 1930, 'Model of Ill-Fated Lake Ship Has Been Built By Sheboygan Craftsman', The Sheboygan Press, January 21, 1930, p. 3.

Co Customs 1867, Permanent Enrollment, Cortland, by Sholes, AL, US Government.

Staniforth, M & Vickery, L 1984, The Test Excavation of the William Salthouse Wreck Site, An interim report, Special Publication No. 3, Australian Institute for Maritime Archaeology.

1839, Enrollment, Anthony Wayne, by Stubbins, AQ, Port of Buffalo Creek.

Thunder Bay National Marine Sanctuary 2009a, Anthony Wayne, viewed 19 April 2009, <http://www.greatlakesships.org/vesselview.aspx?id=153573>.

Thunder Bay National Marine Sanctuary 2009b, 'Columbus'.

US Treasury Department 1838, Steam Engines, by US Treasury Department, US Government Printing Office.

Wilde-Ramsing, MU 2006, 'The Pirate Ship Queen Anne's Revenge', in RK Skowronek & CR Ewen (eds), X Marks the Spot: The Archaeology of Piracy, University Press of Florida, Gainesville, Florida, pp. 160-95.

Wilde-Ramsing, MU 2007, Bronze Bell Recovered Upon Discovery from Shipwreck 31CR314, Queen Anne's Revenge Site, QAR-B-07-01, Underwater Archaeology Branch, Office of State Archaeology, Department of Cultural Resources, State of North Carolina.

Withers, SD 2002, 'Quantitative methods: Bayesian inference, Baysian thinking', Progress in Human Geography, vol. 26, no. 4, pp. 553-66.