__—_‘.
_——__
__—_‘
____——
_—_____.
._—___——
__—_—
_______—

4m
Io):
(IMO—A

 

go 7 L1 301 Q 6

[WWWWII lWWlW ll IWWW WIWWIIWWIW

‘W- 31293 00533 8631

lIBRARY
Michigan State

University

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the

thesis entitled

Individualization of Toolmarks in Bone: A
Scanning Electron and Light Incident Microscopy Study of
Metal Knife Dynamics in Bone

presented by

Max Michael Houck

has been accepted towards fulfillment
of the requirements for

Master of Arts degree in Anthropology

 

 

Date \3217 01%}-

0-7639 MS U is an Ajflrmative Action/Equal Opportunity Institution

 

 

 

 

 

 

MSU RETURNING MAIERIALE:
Place in book drop to
LIBRARJES remove this checkout from
w your Y‘PCOY‘d. {JEEE will

be charged if book is
returned after the date

stamped below.

7, ——._ - _ —. ‘- A.- 1:-lw--..-c——- .-..- -- «fi— r-----—-—-~‘—-———~— .

OCT 08 2008
- 'T' 5131.5 333

 

 

 

 

 

 

 

INDIVIDUALIZATION OF TOOLMARKS IN BONE: A
SCANNING ELECTRON AND LIGHT INCIDENT MICROSCOPY STUDY OF
METAL KNIFE DYNAMICS IN BONE.

BY

Max Michael Houck

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS

Department of Anthropology

1988

ABSTRACT

INDIVIDUALIZATION OF TOOLMARKS IN BONE: A
SCANNING ELECTRON AND LIGHT INCIDENT MICROSCOPY STUDY OF
METAL KNIFE DYNAMICS IN BONE.

BY
Max Michael Houck

The forensic discipline of toolmark analysis has focused on
the identification of toolmarks produced on inanimate media.
While some attention.has been given to the process of
identification of toolmarks made in bone, it.has been limited to
implements with gross defects which allowed expedient
individualization, ‘Use of the usual methodology of striation
pattern matching to evaluate a mark by class and individual
characteristics has not been attempted.

Applying techniques of paleoanthropologists and taphonomists
of stone tool cutmarks on.bone, and developing new standardized
techniques, human and non-human bone was affected with
commercially available knives. The resulting cutmarks were
viewed with a light incident microscope and a scanning electron
:microscope. The two microscopic techniques were compared, Using
traditional methods of striae pattern matching, the cutmarks were
evaluated for class and individualizing characteristics.
Significant criteria for matches, inconclusive comparisons and
exclusions based upon class and individual characteristics is

discussed.

To Janet Patricia and Max W. Houck,
for making it possible

and

to Nancy Ruth Tomboulian,
for making it worthwhile.

ii

ACKNOWLEDGMENTS

I would like to thank.the following individuals and institutions,
in no particular order, who made my research.both possible and
pleasurable: Dr. Roger'Haut and.Tom.Kremenski of the Department
of Biomechanics, Michigan State University, for use of the.Trauma
Biomechanics Laboratory and.Workshop; Terry Donovan and.Tony Owens
of Gatan, Inc. for instruction and beam-time on the CamScan Series
4; the faculty and staff of the Center for Electron Optics,
Michigan State University, for uncommon advice and support;
Richard Slocum, Edward E. SparroW'Hospital, and.Van Alstine's
Packing Company, Inc., for sample specimens; the Staff of the
Department of Anthropology's main office for their patience and
professional advice; the Department itself for financial support
during my research; my committee, Dr. Norman Sauer, chair, and Dr.
William Lovis of the Department of Anthropology, and Dr. Jay
Siegel, Department of Criminal Justice, for their guidance and
opinions; finally Nancy Tomboulian, for her rapier-like editing
and.her renderings of Figures 1 and.2 and.my parents for their

financial and emotional support.

iii

'EBIEOFW

ListofTables.... ....... ........ . ................. .........iv
Listof Figures....... ..... ........ .. ......... ..........v
I. Introduction ..... ...1
II. literamreReviewardBaclgrumd” .................. .......4
IV. Matches, Nm-matdmarflfibaclusions......... ....... .......22
V. Results ............. ...........30
VI. StmmaryandOonclusiors ....... ........ ...... ..44

iv

Table 1:
Table 2:
Table 3:
Table 4:

Table 5:

LIST ormsms

Scanning Electron Microscopy Fstiuates..........32
SEN [Eta Matrix.................................32
Accuracy Percentages For SEM33
Formula Matrix For Fisher's Exact Test ...... ....33

Light IncidentMicroscope Estimatesw ..... 42

LIST OF FIGJRFS

Figure 1: Inpact Trauma Dev10e17
Figure 2: cross Section of Specimen.Base....................18
Figure 3: venn Diagram......................................27
Figure 437
Figure 5.....................................................37
Figure 638
Figure 738
Figure 8.....................................................39
Figure 939
Figure 1041
Figure 11....................................................41

Fim 12....0....0.00000000000000000000000000000000000.00.0043

vi

Although Sinpson (1982) probably did not have the forensic
sciences inmindwhenhepointedoutthat all historical
scienceshaveastheirgoalretmdiction, speakingaboutevents
in the past, the statanent nevertheless directly applies to
that investigative field. Forereic science may be considered
an historical science since, like archaeology and
paleoanthropology, it attanpts to explain past behavior within
a specific context, although it is more similar to the former
thanthelatter. 'Ihisisbecause, incontrasttothe
aforanentioned disciplines, the time period between mission
ofaneventofirrterestardtheacmalsmdyofthateventis
exceedingly brief. Unlike paleoanthropology, the forensic
sciencesdonotaddzesstheenormityofhtmanprehistoryand
the growth of civilization. They are, however, able to be more
specific cancer-rung the actions, participants and items
involved in the studied situation.

As with all historical sciences, certain limitations and
constraints are inposed on the methodologies appropriate for
addressing questions about the past behavior under study. For
paleoanthropologists the explicit underlying theoretical
justification for the methods used is a variation on the
principle of uniformitarianism. Originally proposed by James
mittonin1785andexpardedonbylyell inthe1930's, the

1

2

principle states that processes which are not currently
observableintheworldcmmotbethaightofasinstnimental in
the earth's formation (Knudsen, 1978). Currently doservable
geologic processes may be aployed to explain past geologic
eventssincetheseprocessesareasamedtooperateina
uniform way through time. Paleoanthropologists apply this idea
toallpiysicalprocesses31fmaitabonewithaflinttool
today, itwill yieldaresultcmparabletoabonewtwitha
flint tool 2 million years ago. ‘me forensic sciences,
however, use this theoretical mflerpirming implicitly and, for
the nest part, unconsciously. Historical sciences rely on this
conceptualization of the primary causal mechanisms, and their
concanitant effects, by which past events of interest are
manifested in contextually-oriented information and use it to
deduce affinities between these and corresponding present-day
events. In the forensic sciences, the past events under study
mayhaveoowredmlyafavmorrthsordaysaqo, asopposedto
millions of years.

Cautimmstbeexercisedintheinterpretatimofthe
dataandtheersuingoonclusicre, however, sothata'iedoesnot
overstate the case. Oaaclusions of this nature should have a
structurethat allows foradistinctimbetweenstatanentssuch
as "this windowsill displays characteristics that best
oonespordwithmarksproduoedbyatool withaflat, chiseled
edge 1/2" in width" rather than "this windowsill was gauged by
a a 1/2" flathead screwdriver". As Shipnan (1981) points out,

3
given the intrinsic constraints of historical sciences, this is
an acceptable and, in fact, thimal level of resolution.
Otherwise, thecbservermayunintentionallybiasthenethods
used or results obtained or even exclude fran consideration
potential avenues of research.

This thesis applies techniques of paleoanthropology and
taphoncmyofstonetool cutmarksonbonetotheforensic
discipline of toolmark examination. To this erri, it stands as
a link between these disciplines and it is hoped will initiate
continued interdisciplinary research with traditionally "non-
allied" disciplines. New standardized methods and devices were
developed to affect human and non-Imman bone with calmercially
available metal knives. The resulting cutmarks were viewed
with a light incident microscope and a scanning electron
microscope. 'mequalityofthetimsystemstoimagethe
specimens was carpared. “me cutmarks were evaluated with
traditional forensic methods of class and individual
characteristic carparison. What constitutes significant
criteria for matches, inconclusive cauparisons and exclusions
based upon class and individual characteristics and the
scientific and legal ramifications of these evaluative
designations are discussed.

IIITERA'IUREREVIEWANDBACIGUJND

The analysis of toolmrks and firearm are often
consideredanalogousprccesses;whilethismaybetrueina
theoretical sense, it is not necessarily so in practice. Mien
twomlletsofthesamecaliberaresequentially firedfrcmthe
sane gun, they should exhibit characteristics which are very
similar in quality and expression. The width, depth and
generalcharacterofthestriationstransferredfranthebanel
will not differ significantly, unless extenuating circumstances
arise. This "standardization" is due to the mechanical and
physical limitations inposed upon the bullet/gun interaction by
the precise nature of the chanical charge, the gun's particular
dimersions, etc. Whaiatoolisused, Ixawever, toproducetwo
setsofsequentialtoolmarks, theresultsmaybevery
different. The angle of the tool at impact, the force with
midlthetool isamlied, themediumusedtoreccrdthetcol's
surfacearritheactimofthetcolupcnthemedium, whetherit
creates an imprint or slides thus transfering striations, all
vary with each interaction. These variables and others lead
inevitably to difficulty in replicating toolmarks; thus, a
practical distinction exists between the production of firearms
and toolmark test standards.

5

To use informationinterdiangibly between these two
precessescanclarithisissuemidiinthesanewaythata
infrared spectrummaybe referredtoas a "fingerprint" of a
particular substance. milethespectrmnindeedmaybe
indicative ofasubstance, it innowaybearsresanblanceto
fingerprint identification. It is a correlation intended for
descriptivepurposes; thetwosysteneofccnparismare
conceptually distinct and discrete. In practice, toolmark
examination and the analysis of fired bullets are not
equivalentmethodsbecausethe formernnstaccolmt forcertain
traits associated with respect to their productim. For
firearm analysis, this matter is mostly m.

vmen, inthecomcseofthisthesis, areferencewhichhas
firearms examination as its basis is cited, it is used only
withinatheoretical franemrkwhidicatparestwoormresets
of transferred tooling and/or usage marks. Furthermore, a
practical distinction is recognized between the production of
teststarriardsforfirearmsandtoolmarksardmcalparism
between the two processes is made in that regard.

The nejor concern of the forensic subdiscipline of
toolmark analysis is to individualize nerks nude by a specific
tool or object involved in the cannissim of a crime (Flynn,
1957). This involves the recognitim and netching of class and
individual characteristics of the "inpressions and striations
inpartedbyasuspectedtoolwiththenerkingsmrelevant
objects found at the crime scene" (Bisbing, et al., 1988:

6
Flynn, 1957: 95). Class diaracteristies are the measurable
details of a specimen which indicate membership in a restricted
group or group source (Davis, 1958). Individual
characteristics are those which have structures or canbinations
ofstmcturesthatare'miqueanddistinctiveofjustone
specific inplanent" (Bard and Gilmore, 1968: 390). Identifying
toolmarksas-havingbeennedebyaparticulartool requiresa
significant correspondence between the suspected tool and the
toolmark fran the crime scene (31rd and Gilmore, 1968) together
with "the absence of significant differences which cannot be
explained" (Bard and Greene, 1957: 309) by subsequent
alteration. Individual characteristics are randan in
production and generally result frcm manufacturing devices,
grinding and finishing pm (Bord and Gilmore, 1968:
MetalsI-Iandbook, 1964), andnormalwearandusage (Birdand
Gilmore, 1968; Cassidy, 1980). These processes produce surface
pherunena, called striations, that are characteristic of only
that tool: each tool's particular production and utilization
history places it in a category of die, e.g., individualization
(Cassidy, 1980) .

The nest cannon method of individualizing toolnarks is by
carparingtestmarksmadeusingthesispecteditenwiththe
markings fran the crime scene (Flynn, 1957). Omparison,
categorization and subsequent matching of a toolmark with a
particular tool is contingent upon two dynamic factors. First,
the ability of the substrate to receive and record the surface

7
phermnaofthetcol, and, second, thequantityand
significance or magnitude of the phernnena on the tool itself
(Rao and Hart, 1983) .

As a substrate, bone denaetrates cmpositional
regularities similar to inanimate materials (Bonte, 1959) such
that instances of tools leaving both class and individualizing
criteriauponmmghardtisaeshavebeendoamented (Bonte,
1959; Bird am Greene, 1957; Mittleman and Wetli, 1982), thus
satisfying the first criteria for potential analysis (Rao and I
Hart, 1983). pas and Hart (1983) and Bonte (1959) state that
most cases of identifications involving cut bone almost always
irwolveatoolthathasgrossdefectsofthearttingedge. Rao
and Hart also mention, however, that cartilage, because of its
softer texture, may record characteristics that bone might not
and mld ease identification of tools with "less prominent
individualities" (1983: 798). In contrast, Bonte (1959: 321)
states that in his experience, "bone...shows traces (of tool
marks) better than wood". For exanple, rills produced by
toothedimplements, suchasacerpenter'ssaw, maybefoundin
the bottan of an incaupletely saved portion of a bone. With
this informationonemaybeabletodeterminetheinterhtooth
distance, themmberofengagedteethardestinetethelength
of the saw blade (Baths, 1959).

Bonte (1959: 315) also mentions that knife wounds in rib
cartilage "offer excellent opportunities for tool

classification". Burd and Greene (1957) note that in one case,

8
testmarksweremademparaffinanithesemarksverecmpared
successfully with a section of skull from the nurder victim.

As early as 1927, in the sensational Beernan nurder case,
DeRichter identifiedthemlrderveapon, ahanmer, ashaving
beenthetoolwhichcausedthecranialdepressionfractureof
the victim. He arrived at this conclusion by matching
striations (DeRichter, 1929 cited in 'lhcmas, 1967). Mittleman
and Wetli report that "(w)hen used as bludgecns, [threaded
metal pipes, bolts and dowels] may leave a characteristic
pattern injury on the skin and underlying bone" (1982: 567) and
recognition of this pattern could categorize the veapm to
class or even possibly individual status (1982) .

In forensic tool mark examination, the direction of the
tool'sprogressisoftenapparentduetoaridgeorspiculeof
theaffectedmedimnwhidlispishedaheadcfthetccl (Birdand
Greene, 1957) . Indications of foreign obj est/tissue
interactionssichasthesemaybeusedindeterminationof
directioninsofttisslewounds. AsnotedbyDixon, "'Iheso-
called 'skin tags' located along lateral margins of the...wound
trough point toward the weapon,...in a direction opposite the
path of the projectile" (1980: 275) . In a carpariscn
situation, observations and interpretations of such trauma are
facilitated if the experimntal marks are made at as nearly the
original angle of incidence as possible.

Itissignificanttomteinthiscartextthataitmarks

on bone or, by association, cartilage are definitive and direct

9
evidence of human involvement or interference with the
substance (Shipnan and Rose, 1983). This is because "(n)o
processhasyetbeendiscoveredwhidiproducesnarksthatmimic
slicing, dropping or scraping nerks m a microscopic level"
(Potts and Shipnan, 1981: 577).

Within the field of paleoanthropology, several researchers
havesuriiedtheproblanoftaplnmy, thestudyofthe
cirunnstareesardeventsinvolvirganorganismfrmthetineof
its death until it is examined (Sinpson, 1983), in a way that
is methodologicelly relevant to the forensic discipline of tool
markexamination. Workhasbeendonetodistinguishavariety
oftaptermicnerksmbcne, suchascarnivoretoothnerks,
stone tool cutmarks, preparator's marks (Potts and Shipnan,
1981: snipman, 1981; Shipnan and Rose, 1983), directionality of
cutmarks (Branage and Boyde, 1984), sequence of clenarks
(Shipnan, 1981) and effects of abrasion on forming bone
surfaces (Branage, 1984).

Tooth marks, stone tool wtmarks, are preparator's nerks
nayappearsimilaratagrosslevel ofvislal examimtimare
therefore only the distinctive characteristics of each
tarhcrmic process should be considered diagnostic (ShiFman,
1981): it is critical to distinguish the desired traits fran
the extraneous backgramd information. Tooth scratches, nest
oftenmadebyacarnivore'scaninetoothbeingdraggedacross
the bone surface, are described as "elongate grooves that may
varyfranV—shapedtoU—shapedincross—sectimwiththebottcm

10

of the groove being aieoth" (Shipnan, 1981: 365) and generally
have a single groove at the nadir (Shipnan and Rose, 1983).
Occurring singly, as sets of parallel or subparallel marks, or
asgroupsofmarks, toothscratchesneyvarywidelyintheir
orientaticm to ead1 other (Shipnan, 1981) .

'niefmetionalequivalentofatoothscratdl, stonetool
cut nerks produce elcmgate grooves, V-shaped in cross-section,
whichmayornayretbenarrmerthanatoothmark. Although
oreethoughttobeausefuldemarcatorofwtmarks (Pottsand
Shipnan, 1981: Dunn, 1981), cross-sectional shape is now
considered to be a poor criterion (Shipnan, 1983) . 'Ihe
distinguishing criteria of tool scratches are fine parallel
striations within the original groove (Shiman, 1981) producing
numerous distinct "tracks" at the bottan (Shiman and Rose,
1983). "Ihese fine striations are drag marks or tracks made by
the fine projectiore that deviate to one side or the other of
the edge of the artifact" (Shipnan, 1981: 365) . Shipnan and
Rose (1983: 66) found that "without exception, all...tool
slicing marks shoved the typical, nultiple, fine striatims
withinardparalleltotl'ielongaxisoftl'iemaingroove". As
mightbeexpected, prehistoriccutnnarksshowthesane
microscopic traits as nedern experimental crhnarks (Shipnan and
Rose, 1983). In their exterimental production and subsequent
examinatim of cutmarks, Shiman and Rose (1983: 70) also found
that "manyoftheapparentfeaturosofanysinglecrmarkare
actually feamresmtheperiosteumratlerthanonthebae

11

itself". Interestingly,oneofShirnenarriRcse'sreseard1
conclusicms was that little difference exists between the first
marknedewimastaetoolardmarksnedeaslateinthe
seriesastheZSOthnark. 'Ibthemthissuggeststhat"the
microscopic features of crhnarks might be used to identify the
particllartoolthatmadethen"ifthetoolitselfhadnotbeen
significantly altered (Shipnan and Rose, 1983: 74-75) .

Inordertomoreacrtelyevaluatethetarhormyofcarcass
processingbyhaninids, BranageandBoyde (1984) did
experimentalresearchintothedeterminatimofthedirectim
inwhidlacumarkismade. Icoldmatstudiesofsnearsin
dentaltissue,BranageareBoydeieticedthatthesmearswere
liftedintheoppositedirectimofthewtandafracture
pattern similar to that seen in surface abrasion of glass
(Gordon, etal., 1959). 'Ihispattern istermeda "Hertzian
fractureccne",wheretheapexoftheconefacesthedirectim
oftheabrasivenevement,ardisalsoevidentinvear
striationsonteeth(Gordon, 1984). Totesttheserhernnenain
relationtotoolmarksonbones, BrarageandBoyde (1984)
experinentallyproducedoverzoowtmarksonbovinebcneswith
stonetools. 'Iheyfom'rithatbonesmears,similartodental
emmelsnears,vereliftedintheoppositedirectimofthe
cut. Further, oblique faults of bone adjacent to slicing marks
hadthemedialportioreofthemark,tresenearestthecenter,
displaced in the direction of the alt ("forward", relative to

the wt) . These latter artifacts are hypothesized to result

12
from the tool "chattering" building up pressure and then
releasing it repetitively across the bone surface (Brunage and
Boyde, 1984) .

The sequence of overlapping calmerk production can be
determinedmriercertainconditiore. "Ifthedeprthsofthe
intersecting marks are reasonably similar, then the fine
striations of the later mark will overlie and obsclre those of
the earlier mark" (Ship'nan and Rose, 1983: 89) . Marks of
different depths are difficult to evaluate in this context, as
are poorly preserved cltmarks. In a forensic investigation,
this type of evidence, accurately interpreted, could lead to
the designation of primary musal trauma and the resultant I
hierarchy of trauma incidence, to an individual.

The successful application of the aforenentioned
evaluative criteria rests upon the assunption that the bone
surfacehasnotdegeneratedorbeenobliteratedtoa
significant degree. After only brief contact with abrasive
agents, stonetoolarebonetoolmarksmaylosesaneoftheir
microscopic traits (Shipnan and Rose, 1983) . Bone abrasion,
theresultofanyagentthaterodesthebonesurfacethralgh
the application of mysical force, has been defined and
categorized by Branage (1984). Any taphonanic inquiry into an
osteologicel assemblage or forensic case where the depositional
historyismflmomneedstoconsiderthepossibilityof
abrasively altered or obliterated data.

Cutmarksonbone, however, maypersistforaverylong

13
period of time. 'lhe research on haninid butchering practices
deals with cutmarks fran the late Pliocene and Pleistocene
epochs, which date to fran abort 2 million years ago (Jolly and
Plog, 1987) . An historic incident of toolmarks persisting in
bone is an eighteenth century hanicide at Fort Ouiatenm in
Indiana (Sauer, et al., 1988). Under a scanning electron
microscope, striations fran a metal axe are clearly present on
the cut surface of the victim's rib (Sauer, et al., 1988:
figure 8, page 71) and, with restorative work, could possibly
allow individual categorization. This knowledge could be
crucial to the application of the proposed technique in cases
of prolonged interment.

Pivotal to the success of these palecanthropclogical and
taphonanicstudiesofhasbeentheuseofthescarmingelectron
microscope (SEM). Useofthesmhasseveraladvantagesmen
canpared to conventional light microscopes, including superior
resolution of three-dimensional structures, increased depth-of-
field, and the capability for higher magnification of
specimens, up to 200,000 x (Hayat, 1978; Watt, 1985). Althodgh
the SIM does represent sheer brute magnification superiority
when compared to a conventional light microscope, at a working
distareeofaborthnmfronthespecimen, aconventionalSEM
achieves a resolving power of abort 4 nm while a light
microscope achieves only abort 200 nm (Watt, 1985) . In

practical application, this means,

14
"(i)nspection of the sane specinen with a light
microscopeandtheSEMhasshomthatthelatter

often reveals features that are unclear or invisible
under the light microscope even when the

m’fications am m sane" (Shipnan, 1981: 360:
original alphasis) .

Altholghtl'iesmhasbeenusedwidelyinotl'erareasof
forensic science (Bdrm and Bdnn, 1983: Katterwe et al., 1980:
Pfister, 1982: Taylor, 1973: Wong, 1982), and has shom great
success in toolmark analysis (Devaney and Bradford, 1970: Grove
et al., 1972; Haas et al., 1975; Judd et al., 1974; Madbnell
and Pnflen, 1971; Singh and Aggarwal, 1984) little has been
doneintheareaof interpretingtheresultsofassaultive
and/or taphononic processes on human reneins in situations of
legal concern. Clearly, electron microscopy has the potential
to be an extrenely helpful technique in the individualization
of toolmarks on bone. 'mis is especially true since
microscopic characteristics are the nest reliable means of
categorizing marks of uncertain origin (Shipnan and Rose, 1983)
andthereisresetamomtofdiaracteristieswhidiconstiwte
a positive identification of a tool mark (Davis, 1958) .
Maximizatimofdataisparanomttotoolmarkanalysisandthe
useofasmcanincreasettequantityandqualityof
recoverable data fron an affected material . Despite such
advantages, application of this technique to the usual
methodology of striation pattern matching to evaluate a mark by
class and individual characteristics has not been
systematically attarpted on bone tralmla.

MEDDIBANDMATERIAIS

Intleprqaaratimandproductionoftleexperimental
equipnenttobeusedinthissudy,severalfactorsoftte
knife-boredynamicneededtobeaddressed. fleangleatwhidl
the knife was applied mist be duplicated, force of application,
orwork,andthedegreeofhardnessoftheaffectedmaterials
hadtobeaddressed(Flyrm,1957),otterwiseitmigntrethave
beenpossibletosortouttheparticular effects.
Additioaally,theprocesshadtobereproducibletoinsureboth
intra-andinter-observerstandardizatim.

Morecatparableresultsareadlievedwl‘entleoriginal
materialthatwasaffectedbythetoolinqlestimis
duplicated; a less resistant material will also suffice (Bard
ardGreene,1957). Tothisend,bov1ret1blalshaftswere
obtailedfrunalocalbutdlerardwerequarteredlengtlmiseto
maximizeusablesurfacearea. Asnuchadherentmaterialas
possible was removed without damaging the periosteum. Half of
ttetibiaequarterswereplacedinaventinghoodatroan
temperature for72-80hourstodry. 'Iherenainderofthe
tibiae quartersvere kept refrigerated at 34°Funtil needed.
Forinterhspeciescotparason,ah1manribsectionrereved
duringautopsyatEdwardE.SparrovHospitalwasalso

refrigerated for later use.
15

16

Toinducethetraumainthebone, aguillotine-like
machine, the Inpact Tratmla Device (ITD) , was built using the
facilities at the Trauma Bianechanics laboratory at Michigan
State University (Figure 1). lbs I'ID consists of an anodized
almimlmdroptouerwithasingleoenterpistmguidedbytwo
sets of 'lhmpson precision bearings. The knives were placed in
aspeciallymadedmckatthebottanofthepistmand
positioned with six centering bolts. The bones were placed
horizontally, periosteal side towards the knife blade
(upwards) , in plastelene, a non-drying synthetic clay, on a
sample basemade of Iexan (Figure 2). ‘me samplebasewas
hingedsoitcolldbesetatanangletangentialtothecenter
piston using supportive insets to produce the desired angle.
ItebaseoftleI'I'Dhasguideholeseveryl.25ansothesanple
basecanbeirerenentedlirearlybereaththecenterlireof
inpact in order to produce sequential ortmarks.

Originally, four cannercially available lmives were
erchasedforthissuriy: oneofthem, knife 2, provedtobe
unsuitable forusewithtlechuckontheI'IDbecausetheblade
wastoonarrow. Iterefore, itwascrnittedfrunthesoldy.
Fifteen coded specimen interactions each with for
diaracteristieswereproducedusingtherardonmmberfmctim
of a Hewlitt Packard 33C calollator. The characteristics were
chosen for ease of reproducibility, production of necessary
couponents, and sinplicity of interpretation. This last factor
wasconsideredtobeofusebecauseofthepreliminarynature

*4

\J

G—Weight

Piston

 

 

/Beorings

 

 

”LEE?

 

Bearings

 

 

 

Knife
Chuck

 

 

 

C 0
0° v.\
C O
O
O O .

A. [Z
/

 

Figure l. Impact Trauma Device

 

 

 

‘
'

 

 

47

Sample
Base

Guide Holes

18

Bone

 

 

   

Plostelene

 

 

 

Figure 2. Cross-section of Specimen Bose

19
of this research: it aided in sorting out "backgromd noise"
from potentially important observations .

Each coded interaction contained four mmbers which
carried information regarding the specific parameters. The
first nunber indicated the species of the osteologicel sample:
a "1" or a "2" means the specinen was bovine and a "3" meant
thespecinenwashuman. 'melmifeusedtocutthebore, 1, 3,
or4, isrepresentedbythesecondntmrber. 'Ihethirdnumber
indicated the angle at which the knife contacted the bone, 45°
or 90°. The fourth number represents the anemt of veight on
the center piston, .68 kg (1.5 lbs.) or 1.6 kg (3.5 pounds).
Fifteen cuts were made following the specificatioe in the
coded sanple interaction information. If, when making a cut, a
specification was not followed or the piston release allowed a
non-uniform release, that out was removed and discarded. The
I'I'Dwasthenresetadtheortattaiptedagain. All inpacts
weremadeatadistareeofBOcnbetweenthesanpleboreand
the knife's edge. A spring-release system was enployed to
insure exact piston release. The same section of each knife
bladewasenployed forallcutstorestricttheamountofcut
surface needed to be searched for striations. After each oat
wasmade, thatportionofthebonewiththemarkonitwas
removed and nunbered for specimen preparation.

After inducing the trauma, the neist refrigerated
specimenswereplacedinaventilatedhoodatroontcrperamre

for 72-80 bone to dry. All specinens were sonicated for 1

20
mimltetoreneveadherentdebris. Inretrospect, soniceting
forupto3mimrtesmildbeadvisibleduetocmsequent
problanswithparticlesbuildingupachargemfiertheelectron
gold for folrmimrtesusinganImscopebbdelleOOSprtter
coater. Carbon-based paint, such as is used on television
tubes, wasappliedatttepointofcomectimbetheenthe
specimenarrithemmtingstubtoirereasetteconductivityof
the specimens.

All cutmarks were evaluated for criteria which would
constitute class menbership and individualization of a toolmark
to a particular tool (Biasotti and olpertino, 1964: (assidy,
1980: Davis, 1958; Burd ard Gilnere, 1968; Haas, et al., 1975;
Judd, et al., 1974: Singh and Aggarwal, 1984:) using an Olympus
nedelHilight ireidentmicroscopeandaCanScanSeries4sm.

'Ihelight incidentmicroscopeandtheSIMverechosen for
reasons specific to each instrunent. The light incident scope
was chosen because it is ubiquitols in forensic science labs
anditistheinstrmnentthatnesteasilylenis itselftothe
task of toolmark analysis ('Iedeschi, et al., 1977) and because
sanples require re special preparation or envimment. 'Ihe
SIM, while requiring qoecial preparation and environmental
control, namely a gold-coated specinen and a high vacuum,
provides a death of field and resolution not possible in a
light microscope (Hayat, 1978: Watt, 1985). Studies have shown
that the utility of the SIM in toolmark analysis far outweighs

21

any special coreiderations its use nay necessitate (Haas, et
al., 1975; Judd et al., 1974; Matricerdi, et al., 1974; Singh
and Aggarwal, 1984: Specter and anisorge, 1973) .

'meuseoftheCanScanSeriesllwasespeciallyhelpful
because of its unique corparison abilities. A variable sized
imagewiniowcanbedefiredono'escreenandtheimage
cortairedwithinthiswirdowissubstiurtedwithintresame
equivalent locationintheinagemtheotherscreen. This
ereblesadirectcorparisontobemadebetweenthetwosanples.
'Ihis study exployed a large stage capable of holding for
specinens in the dianber at one time: thus, nultiple conjoint
comparisons vere facilitated. In addition, a netorized stage
was used which allowed for precise manipulation of the

specinen.

MA'IU'IES, W-MA‘ICHES, AND EXCLUSIQIS

Mostexpertsintoolmarkorfirearnsanalysisagreethat
whatccnstitutesamatchbetweentwocorparedspecinensis
largely dependent upon the qualitative and quantitative
criteria chosen by each individual examiner, gained throlgh
practical professional experience (Bisbing, et a1. , 1988:
Biasotti and Cupertino, 1964) . No rigid evaluative or
statistical standards exist, such as prcper magnifications
(Flynn, 1957) or the recessary number of characteristics used
to determire a match (Davis, 1958) , and the specific techniques
arployedwillvarydeperdingmtheclass oftool involved, the
orientation and physical traits of the mark and other factors
(Burd and Greene, 1957). Amatch is a conclusion drawn fron
relevantobservationsbyapersonwhoisconsideredtobean
expert in the field. The criteria of a matd1 is generally
understood to be a sufficient correspondence of individual
characteristics (Biasotti and alpertino, 1964; Bisbing, et al.,
1988; Ber ard Gihrere, 1968; Burd and Greene, 1957; CaSSidy,
1980: Flynn, 1957; Singh and Aggarwal, 1984) between the test
starriardanithequestioredsanplewithortthepresereeof
significant non-corresporiing traits.

No single arbitrary standard for vhat consitiurtes

"sufficient" criteria exists for toolmark analysis. Each
22

23
individual examiner decides what quantity and quality of traits
eddlibitedonapartiollarspecinenisenolghtowarrantthe
resultingstataient. ‘Ihedesignationoftheobserverasan
expert,tterefore,ispivotaltothecertaintyoftte
conclusion. 'neexpertneedstobeawareoftherangeof
variationoftletoolmarksinquestimandwhatcanorcamet
besaidaborttlerelatimshipbetueentleloewnandtte
questionedsanples. 'Ihesoundnessoftheresultsrestsin
largepartupontheacquiredexperienceoftheexamirer.

It shouldnotbe surprising, then, to find that
researdersdisagreeastothetypeofcoelusiorethatcanbe
drawnfromtoolmark cotpariscms. BmdardGreeneu957) list
the following four conclusions:

I. "Noopinionorconclusioncanbereadieddueto
alteratimofeitherthetcolmarkorthetool, which
hasoconredsireetleconnissionofthecrine"
(1957: 308),

II. "'Ihequestionedtoolcanbeeliminatedashavingbeen
responsible for producing the tool mark under study"
(1957:309),

III. "'Ihetool mayhavemadethemarkbutaconclusive
identification is ret justifi " (1957: 309),

IV. "The tool did produce the mark in questim" (1957:
309) .

'IheauthorsnotethatinthecaseofconclusionII, this

decisimcanbereadiedontlebasisofrm-correspondeieeof
the class-level characteristics of the test standard and the
questionedspecinenor, iftheclasstraitsdocorrespmd, on
the "lack of identifying detail...which it appears shoild be

24

present if this particular tool did produce the mark" (1957:
309). Rowe (1988), on the other hand, lists only three
possible outcanes of striations cotparisons:

1. The questioned itan node the mark (identification)

2. 'Ihequestioned itemdidnotmakethemark (ren-

identification)
3. 'nneresultsoftheconparisonsareinconclusive
(Rowe, 1988: 430).

Rowe explains that conclusion 1 is based an observation of
similar class features and matching striation patterns on the
teststarriardanithequestionedspecinen, mileconclusion3
hingesonthecorrespondenceofclasscharacteristicsbutnot
on traits of an irndividual nature. Conclusion 2 indicates that
the questioned specimen has different class-level
characteristiesthanthosefondontheteststaniard. While
it is gernerally accepted that a non-correspondence of class or
genus characteristics recessarily inpl ies exclusion (Biasotti
and leertjno, 1957: Bisbing, et al., 1988: Burd and Gilmore,
1968: Bird and Greene, 1957: Cassidy, 1980: Flynn, 1957: Robe,
1988), nestresearchers inthe fieldarehesitanttomakea
specific statanent abort a non-correspondence of
individualizing traits. This level of conclusion, however, has
been implied. In 1930, Goddard stated that

"tmannsoftlnesarrecaliberardmakewill

exhibit sufficiently pronoureed characteristics

of their own, .. .to make possible a determination

as to which of them fired a given bullet. 'Ihis

isbecauseabulletintraversingabarrel
acquires on its surface the characteristic

25

W“ peglia: 1_:g that barrel alone (alphasis
added, 1930: 6)".

Buxton (1930: 212) also mentions that "if the (rifling) grooves
correspondbuttheotherdistinguishingmarksdonot, itis
shown that the two (bullets) were fired fromldifferent firearms
of the sam make". Watson found that, after canparing two
different knives, "no correspondence was observed in the
pattern of accidental [read individual] characteristics
present" and "the identifying elanents form a carbination the
coexistenoeofwhidiishighlyinprobableinatoolmark
produced by another knife" (1978: 45). In an SEN study of wire
cutters, SinghaniAggarwal statethata "rm-correspoaflence
ofthesurfacestmcturesmtheartfacesproducedby
different pliers affirms that they are characteristic of the
cutting tool" (1984: 121) . 'Ihese studies clearly indicate that
scmeexpertsconsider itpossibletoexcludeatoolmarkbased
on the carparison of individual characteristics. Amlying sane
of the more basic aspects of formal logic will facilitate a
betterunderstarriingoftheargmnentsusedtobolsterthiskind
of opinion.

Using a Venn diagram (Figure 3), "A" will represent all
possible straight-edged knife cutnerks, "B" will represent all
Granarksnedewithlmifemmberl, whichhas
astraightedge, and"C" will representalloxtmarksmadewi‘th
all straight-edged knives other than B.

Fran this relationship we can inmediately infer that while

either B or C are sufficient conditiors for A, they are not

26
necessary. 'meconditiors forAcouldbemetbyeitherBorC.
therefore, A is aneoessarybu‘tnot sufficient

 

 

 

 

 

 

Venn Diagram (Figure 3)

condition for B or C. In relation to each other, the
S‘tata'lerrts "If A, then B" ani "If A, then C" are contraries;
they offer alternate competing resporses to a situation (Enmet,
1981) . ‘Ihis is a similar situation to ccuparing "'Ihat tree is
a birch" ard "'Ihat tree is a pine": at mast, one statement is
true. Itisinpor‘tanttoranemberthatitisentirely
possible for both statanents to be false; the tree in reality
maybeanoak. 'Ihisiswhyphrasingisinpor‘tanttothe
interpretation of "historical" data: if a set of tool marks is
identified as having been nade by a screwdriver, all other tool
marks are eliminated fran the researdier's consideration. If,
ontheotherhard, thestatenerrtispirasedtocmn‘ntethatthe

marksarecmsistentwithhavingbeenproducedbya

27

screwdriver, the possibility of their production by sane other
doject renains available to investigation.

WemaysaythatthetermsBardCarecmtradictoryin
thatoneistrueandoneisfalseandtheymaybeneithertrue
togethernor falsetogether (Ehuet, 1981). Inthis sense, B is
equivalen‘tto "rot-C" aniCisequivalen‘tto "not-B". Wienwe
say "IfeitherBorC, thenA", eitherBorC, hrtnotboth,
are sufficient but not necessary criteria for A. We are
therefore not entitled to infer fran this statement, "If A,
then B" or "If A, then C". We may, however, infer "If not-A,
then not-B" or "If not-A, then not-C" (Emmet, 1981); this is
the logical basis for a class exclusion. In a more manageable
format, our first statenent mild read like this,

Ifthismarkhastraits similartoknifelorsane

otherstraight-edgedknife, thenitwasmadebya
straight-edged knife.

Andthevalidinferencesweccnlddrawfrunitnmldbeas
follows,
Ifthisnarkwasmtnadewithastraight-edged
knife, thenitwasnotmadewithknifel (a
straight-edgedknife):
Ifthisnarkwasmtmadewithastraight—edged

knife, thenitwasnotnedewithanyother
straight-edged knife.

Individual exclusionshavebeenbasedonanon-
corresparlence or lack of individual characteristics (Biasotti
and Cupertino, 1964; Bisbing, et al., 1988: Burd and Gilmore,

1968). 'Ihisccuparativeprooess isbaseduponthedisjunctive

28

syllogistic form:

Either B or C

Not-B

Therefore, C.
'IletermsBandCarecalledthedisjunctsoftlesyllogismarfl
are its canporent propositiore (Oopi, 1982). If one prauise is
adisjunction, inthiscasetlefirstlire, andtheotheris
ttedenialorcmtradictimofmeofttetwodisjmicts, the
secondlire, then'toinferthetmthoftheotherdisjmct is
valid. Disjunctive syllogislus are valid only when the
categorical pranise contradicts are disjunct of the disjunctive
premdseandtleconclusionaffirmstleotherdisjtmctoftte
disjunctive premise (Copi, 1982) . ‘Ihis argunent expressed in
words would be:

Either this mark has traits similar to knife 1 or sane

other straight-edged knife
It does not have traits 5.1m '1ar to m'fe 1,

'Iherefore, this mark has characteristics similar to sure
other straight-edged knife.
'Ihis is the sane systematic evaluation made in fingerprint
analysis; assmihgthatmtec‘rmical errorsweremadeinthe
production of the test standard,

"sl'iould an urexplained dissimilarity occur, as
for exanple the appearance of a clearly defined
ridge characteristic in a latent print which does
not exist in the inked impression, the corelusion
isirescapablethattteprintswerenotnedeby
the sane finger" (menssens, et al., 1986).

More prosaically, this process is analogous to outparing a
motographofaperson's facewithammberofphotograrhsof
people of a similar class category, for instance, white males

29
betweenZSandBSyearsofage. Anobserverwouldbeableto
statewtetherornotthepersonintlereferencephotogramwas
in tte pile of canparison photographs. If the person's
photograrhwasnot, tl'entleobserverwouldhaveexcludedthe
subject on the basis of persmal or individual characteristics.

The validity of non-matches as a possible conclusion
permitscertaincategories ofcbservationstobeusedina
statistically legitimate manrer. Since the populatim of
observations is finite, this allows for a greater range of
statenentsabouttleacwracyandindependenceoftle
observations. Morewillbesaidaboutthisintherext
chapter-

For purposes of this thesis, "exclusion" will defire a
sioiationwheretwoitansdonotbelongtothesameclass
category. "Non-exclusion" denotes a irdeterminable, possibly
similar, class relationship. The word "match" will mean a
sufficientmmber, tobedeterminedbytheexamirerof
corresponding individual characteristics and the phrase "rm-
match" will indicate a m1-correspordence or lack of otherwise
individualizing striations patterns: a nan-match therefore
recessarily implies exclusion based upon individualizing
characteristics. "Not possible" denotes that ireufficient
criteriaexist formkingareasonable judganentaboutthe
class or individual relationship of the specinen in question.

Fifteen specimens were produced which allowed for a
population of 105 indepenient-pair cmparism observations,
afterranovingsamesanplearrireflexivecmparisaeaftertle
formla used by Flynn (1957):

a + l

S=-----xn (1)
2

wheresistheannoftleterms,aistlefirstterm,listhe
lasttermandnistlemmberofterns.

1+15
S=-—----x15
2
16
S= x15
2
S= 8x15
S = 120 (-15 reflexive comparisons)
S= 105

'Ihe breakdown of the actual relationship of these 105
carparisons is that 50 were class exclusions, 24 were non-

matdies, and 31 were matches.
30

31

my Electron Him

With the CanScan Series 4 SEN, the estimation of the
population's make-up was divided into exclusive and non-
exclusive categories. Of the 105 comparisons, 39 were
excludable by class characteristics alone while the other 66
were assigred mikmwn class category relationships. This
latter group were then subjected to microscopic analysis to
determire if striations sufficient to evaluate a match or a
non-match were present. 'Ihis type of categorization made no
allowance for an equivocal response; observations were a match,
a non-match or, if either of these designations were not
possible, left at the class level of information.

0f the 66, 39 were determired to have striations which
would allow individualizing canparisme: the other 27 had
insufficient characteristics for this type of observation.

0f the 39 which were usable, 26 were classified as matches
andlBasnon—matches. 'Ihreeoftle26matcheswereinfact
false positives or a type II error (B): that is, they were
actually non-matches (Table 1). Type II errors carry greater
ethioalconsequencesthandctypeI (a) errorssincetypeII
errors would result in wrongly incriminating evidence whereas
an inconclusive canparison does not reoessarily inply either
innocence or lack of association (Gaudette, 1987) .

Amajorconcernofthewrrentthesisistle
relatiorehip between the accuracy of the technique and the
nature oftheobservationsuponwhichtheaccuracyis

32

Soaming Electron Microscopy Estimates (Table 1)

 

Class
Match Non-match Exclusion
Cbrrect 23 13 39
W 0 __0
Totals 26 13 39

Insufficient Characteristics = 27

predicated. 'Ihese topics were explored with Fisher' s Exact
Testof Independence, wl'erethemrJdnghypothesis isthat
whetheraobservatimaltmitisasucoessorafailmeis
dependent on its classification as an observational unit
(Bradley, 1976) . 'Ihe results with the light microscope do not
meet the requirexents for statistical analysis and merit
discussion only. The scamu'ng electron microscope data matrix
set-up is seen in Table 2: "C" indicates a correct estimate,
"I" indicates an incorrect estimate, "M"

5m Data mtrix (Table 2)

 

 

c I | T
M 23 3 : 26

13 o : 13
'r 36 3 : 39

standsforanestimatedmatcharfiW'representsanestimated
non-match or individual exclusion. "T" indicates the totals

33
for each column and row. Using this tabular
array, the accuracy percentages were derived by division of
totalsintotheestimatesandareshowninTable3. Whileit
isapparentthattheseaccuracyvaluesshowthemethodtobe

quite accurate, they would be meaningless unless it

Accuracy Percentages For SEN (Table 3)

 

 

c I | T
M .885 .115 i 1.0
N 1.00 0.00 : 1.0
T .923 .077 : 1.0

conldbedanonstratedthattheestimationsareindependentof
eadnother:whetheranestimateisaancosssorafailure
shouldhavenothingtodowiththetypeofestimatethatwas
made.

'Iheformula for Fisher'stestisobtairedfrcmthematrix
showninTable4:observationnmitsAandBare

Fornula Matrix for Fisher's Exact Test (Table 4)

 

 

s F | T
A r n-r : n
B R-r (N-R) - (n-r) : N-n
T R N-R : N

eitherasucoessSorafailureF. Arandansanpleofn
observations was drawn frun the infinite subpopulation of units

34

that are A's and a randan sarple of N-n observations frun the
infinite subpopulation of units that are B's. airpose that r
ofthenn'nitsintheAsanpleanfiR-rofthemnitsintheB
sanple are successes; when canbined, a pooled sanple of N
units, Rofwhid‘naresucoessesisdrtaired.

'Iheworkinghypothesis forthistestusingthesnni
estimates is as follows: the success or failure of an
observationisdepenientuponhowtheobservationalunitis
classified, in this case as either a netch or ton-match, with
alpha set a .05. If this hypothesis is rejected, the
subpopnlationsoanndBarehcmogereousinternsofthe
proportion of successes. Whether a sample unit was a match or
arm-matd'nwouldhaveno influenceuponitssucoessor failure
(Bradley, 1978) . If the hypothesis is confirmed, then the
observational categories would be preferentially correct or
incorrect due to their classification: for exanple, matches
wouldbecorrectnnoreoftenthannon-matdnes. Because itisa
starndard value and a 95% accuracy rate mild be acceptable,
alpha is set a .05.

'Iodeterminetheprobabiltyoftheteststatisticr, the

hypergeanetric probability law is used:

R N-R

r n-r

P(I‘) = (2)
N
n

 

By substitnrtingthevalues inTable 3, the fornula appears as:

35

26 3
23 3
P(r)

 

39
36
4.032926

2.585222
P(r)

 

2.039846

3.719941

15 , 600

 

P(r) =
54,834

P(r) = 0.28449

Rounding up, the value of P(r) is .285. Using Fisher's Exact
as a test statistic (Bradley, 1978), if the test value is less
thanthesetlevel ofalpha, thentheobservationalunitsare
dependentuponeachother: ifgreater, theunitsare
independent (Bradley, 1978). Since .285 is greater than .05,
the test rejects the null hypothesis; the success or failure of
amatdnoranon-matdnofalmifemarkisixdmendentoftheir
classification as a match or a mun-match. For these
observatioretobedependent, thealynalevelwouldneedtobe
set at .286 or higher: clearly this would not be a useful
statistic. Since the observational units are independent, the
accuracyvaluescanbeacoeptedwith confidence thatnonon-
observer biases are intrinsic to the method.
Oreofthestrengthsofthistednniqueistheclaritywith

36
which the visual characteristics present themselves . In Figure
4, thebackgronndisakrmncitmarkandthetwovertical
wirndovsarethequestioedspecimen. Asimilarconparisonat
50): also yields a positive matdn (Figure 5) . Fracmre lines
perpenflionlartothedirectionofcrt (Figure6) arepresentin
nnostcuts, althoughtheirexactcauseismtmderstood.'1hese
may be related to the "dnattering" oblique fractures observed
by Bronage and Boyde (1984). [he to the I'ID's single inpact,
as opposed to the continual build-up of a slicirng mark, oe or
two severe fractures might be expected rather than multiple
fracodresatregularintervals. EVenatlSOx, asshownin
Figure 7, matches were possible: in fact, increased
magnification was necessary for the arts made by the serrated
knnife. Individualizing dnaracteristies were visible only at
higher magnifications. 'Ihe direction of the knife cut was
discernnibleasisshominFigure 8: asthebladepasses
through the hoe, it
crushesthenascentsurfaceoftheortandproduces "lifts" in
theoppositedirection. ‘Ihisisthephenouenonnotedby
BronageandBoydeintheirreseardnonofdirectionalityof
stone tool cutmarks (Bronage and Boyde, 1984). In Figure 9,
thelmifeistravelingfronthebottontothetopofthe
photograrh. Heretheliftsarenoreclearlyseen. In
conjunction with gross visual examination and anatonnical
orientation of a cinnark, microscopic determination of ortmark
directionality could be of great assistarnoe in victim/assailant

37

r I

 

Figure 4

 

 

2321 dry

'dnrefttom of
60x

2122 dry

direction 0,
cut tron bottou
-: Photo_to up ..

 

40
positioning at the tine of the assault. Also, directionality
cannarrowtheexaminner'srargeoftestmarksforconparisonby
eliminating fron conparison irpossible or difficult stab wound
trajectories.

Of special concern to forensic anthropologist and
pathologistsisthequestionofwhethercrtmarksshoildbe
associated with the victim peri- or post-Horton. Inn a broad
tinefrane,thistypeofestimationispossibleusingthesm.
FigurelOshowsaortmarknadenfinentreborewasstillfresh:
the bore was subsequently air-dried over a period of 72-80
hoursatroontorperature. 'Ihelireofcut,thelocationwhere
thelmifefirstcontactedttebone,isseparatefronthelire
wheretheperiosteumwasfirstcnrt. Astheperiosteumdried
arndshrank, it pulled back frontheinnitial line of cut. This
iscontrastedinFigurellwhereacutmadeaftertheboehad
alreadybeenair-dried. 'Ihelireofcutforttebonean'xithe
periostemnareequal:nopost-cutshrinkageoftheperiosteum
hastakenplace. 'nneseresultscouldbeofuseininstances
whereagereraltinneorientationoftramnaandorderofevents

Li I ! I i: ! 11'
Using the conparison light incident microscope, of the 105
conparisons in the actual population, 33 were excluded on the
basis of class characteristics. 0f the ranaininng 72, 63 had no
individualizing characteristics that were visible under the

41

O :i [I ['2— 1'] '2' 0PM -—I———J

 

Figure 11

42

light microscope. Only 3 positive identificatio'e annd 6 non—
matches were made (Table 5) . Directionality was not possible
todetermine, althoughwhethertheboehadbeenontwhenwet
or dry was evident in certain instances. Overall, the light
incident microscope performed poorly relative to the SEM. 'Ihis
isdueinlargeparttothelightmicroscope's shorterdepthof
field arnd weaker resolution (Figure 12) . Although

Light Inncidennt Microscopy Estimates (Table 5)

 

Class
Match Non-mtch Ecclusion
Correct 3 6 33
W o __0
Totals 3 6 33

Insufficient Characteristics = 63
possible, achieving definitive results with a light incident
microscope is greatly dependent upon the specinen cutnnnark's
characteristiessuchasangle, depthandforceusedtomakethe
cut. Onts which are deep and vertical are more difficult to
evaluate than are shallow, more obliquely angled cuts. Making
a replica might be of assistance with diffionlt to observe
crtmarks.

43

 

M AND CDNCLUSIONS

Theresultsofthisstuiyshonldbetakenaspreliminary
but very positive. Borrowing techniques frun
paleoanthropological annd taphononic studies of stoe tool
cutmarks on bone, and developing new standardized techniques, a
nnovel nethod of identifying krnifemarks made in hoe was
developed. Cutmarks were made on human and non-human hoe with
commercially available knives. The resulting onzmarks were
prepared for and viewed with a light incident microscope and a
scanning electron microscope. The ability of the two
microscopictechniquestoimagetheontmarkswasconpared.
Using nethods cannon to the forensic discipline of toolmark
annalysis, the cutmarks were evaluated for class and
individual izing characteristics . A working definition for an
inclusion (match) and occlusion (non-match) based upon
individual traits was developed annd implemented. What
constitutes significant criteria for drawing various
conclusions concerning the relationship betweenn two sets of
toolmarks was discussed.

Toolmarks fronlmivesdotransfertoboeandttmsmatches
and exclusions are possible with the nethodology developed in

this thesis. No visible difference exists between affected
44

45

bovine or human hoe at the magnnifications ennployed in this
study. Non-humanboe isanadequatenedium forresearchof
this type. The use of the SEM provides quantitatively annd
qualititatively superior data with which to make conparisons .
Theaccaracyratewiththecoditionssetforthinthisthesis
was 89% for matches and 100% for non-matches. Only 3 errors
veremadeusingtheSEMandthesewere falsepositives. ‘nnese
may be attributed to difficulty in aligning the two specinens
due to differences in the sample angle. Four samples were
placedintheonamberatatineandeachwasclanpedintoplace
onthestubholder. Therefore, asthestubholderturred, so
did the inndividual specimens; they retained their position
relative to oe annother. 'Ihis neant that unless each specimen
position was changed, the observer would always view the sane
side of the specinnen. Although nnot a significant problan in
relation to the sample unnder study, since the relationships
weresetandthepopulationfinite, atypeIIerrorcancneate
serious problons in a realdworld situation within the realm of
thelegal syston. Thispotentialsouroeoferrorcould
coceivably be corrected for by an alternate mounting netted
which would allow for a wider range of specinnen manipulation.
Nobias cocerningtheaccuracywasattrihntabletothe
designations of observational unnits as matches or non-matches:
the success or failure is independent of tleir classification.

In application, a researcher would do well to examine
potential cutmarks with a light incidennt microscope before

46

utilizing an SIM. This "screening" could save tine annd
difficultyinnarrowingdowntleactuallmifeedgeareastobe
searched. Since it is possible to corpare ontnarks on a light
microscope, thisstageofanalysiscouldobviatetheuseofan
SEMforthat specimen. Theprocsssoutlinedhereispossible
withansmthatdoesrcthavetheconparisonwirdow
capabilities of the CannScan SFM: microphotographs are conpared
innstead of images annd documentation is of greater cocern.
Samplepreparationrequirennents fortheSfldneybegreateras
conpared to the light microscope: the results are, however, of
a mudn higher level of resolution.

Tomingtothelegalacoeptanceofthisstudy, that is,
havingtheresultsofsuchconparisonsastleseacoeptedas
evidenceinacourtoflaw, severalprecedentsexist. The
Washingtonaanemeoonrtronerked"tleedgeonoeblade
differs frontheedgeofanotherbladeasthelinesonoe
human hand differ form the linnes on another." (State v Clark,
1930) . In Counonwealth v. Bartolinni, a cleaver and testimony
thatmarksonthevictinn's skull couldhavebeonmadebysucha
cleaver were admitted into evidence (1938) . Also, in State v.
Crnurdnill, aknifewasconnectedtomarksonahcmicide
victim's eternal cartilage (1982). Similar analyses have been
perforned with knifonnarks in cartilage (Bonte, 1959: Galan,
1986: 1987: Rao annd Hart, 1983). And in 1927 an examiner was
requiredtoidentifytheweapon, ahanmer, whichwas
responsible for a depression fracture of the skull. He

47
achieved his goal by the matching of striations DeRichter
(1929) . With further research and experimentation, to produce
a nnore standardized and reliable test standard replication
technique, itseemsreasonabletoenqnectthatthisnethodcould
became a reliable, legally acceptable metncd of idenntifying
knnives used as honicide weapons.

49

BIBLICXRAHIY

Biasotti, A.A., and Cupertino, M.S. 1964. Tie Principles of
Evidence Evaluation as Applied to Firearms and Tool Mark

Identificntim- W 9: 428-433-

Bisbing, R. E., Stolorow, M. D., Gerakaris, J., and McKasson,
S. C. 1988. Meandering Over Fractal Surfaces: Physical
Matching in Archaeology and Criminalisties. Paper
Presented at tre Anerican Acadouy of Forensic Sciences,
February 15-20, 1988, Philadelfi‘nia, PA.

Bohnn, E. and Bohm, I. 1983. Bibliography of Scanning Electron
Microscopy Application in Forensic Medicine. gagging

W Part 1: 305-309-

Bonte, W. 1959. Tool Marks in Bone and Cartilage. W
W. 5: 315-325-

Bradley. J- V- 1978. W
Prentice Hall: Englewood Cliffs.

Bronage, T. G. 1984. Interpretation of Scanning Electron
Microscopic Images of Abraded Forming Boe Surfaces.
WW 64:161-178-

Bronage, T.G. and Boyde, A. 1984. Microscopic Criteria for
the Determination of Directionality of unmarks on Boe.

mican Journal of Phfiigal And-mlgy, 65:359-366.

Bunnnn, H. T. 1981. Archaeological evidence for meat-eating by
Plio-Pleistooene honinids fron Koobi Fora and Olduvai

Gorge. Na_tn_n;§, 291: 574-577.

Bird, D.Q. and Gilmore, A.E. 1968. Individual and Class

Characteristics of Tools. gournal of Forensic Eggnog,
13: 390-396.

50

Rnxton, J. 1930. The Science of Ballistics: Judicial

Applications. wig Jgu_rna_L of Police Serenge' , 3:
211-219.

cassidy, F.H. 1980. Domination of Tool Marks fron
Sequentially Manufactured T'ogue-and-Groove Pliers.
Weighing. 25: 796-809.

Oomcnnwealth v. Bartolini. 1938. 299 Mass. 503, N.E. 2d 382.

Oopi. I- M- 1982. W. Macmillan
Publishing, 00., Inc.: New York.

Davis, (LE. 1958. In ' 1 Marks °
the Stria . C.C. Thonas Publ.

DeRichter, G. 1929. Affaire de Beernem. Identification du
marteau ayant servi d'instrmnent du crime. may];

W11: 1121-1136-

Devaney, J.R. and Bradford, L.W. 1971. Applications of
Scanning Electron Microscopy to Forennsic Science at the
Jet Propulsion laboratory, 1969-1970. W
W911: 561-568.

Dixon, D. S. 1980. Determination of Direction of Fire fronn
Graze GmshotWounds. W: 25:

272-279.
Emmet, E. R. 1981. W. Helix Books:
Totowa, NJ.

Flynn, E.M. 1957. Toolnnark Identification. w
W. 2: 95-107.

Galan, J .I. 1986. Identification of a Knife Wound in Bone._

WM.
18: 72-75.

Galan, J.I. mrsonal communication. Nov. 1987.

51

Gaude‘tte, B.D. 1987. Chapter 5: The Forennsic Aspects of

Textile Fiber Examination. mic Science Mk
Volune II. R. Saferstein, ed. Prentice Hall, Englewood
Cliffs.

Goddard, C. 1930. Firearnns asEvidence. Angler—icar_n_Jon.nrr___nal91.=
Police $9313, 1: 3-7.

Gordon, J. E., Marsh, D. M., andParratt, M. E. 1959. anthe

strength and structure of Glass. With—camel
W 249A: 65-72-

Gordon, K. R. 1984. Microfraconre Patternns of Abrasive Wear
Striations on Teeth Indicate Directionality. Alarm
J of ical 1 , 63: 315-332.

Grove, C.A., Jndd, G., Horn, R. 1974. EValuation of SEM
Potential in the Domination of Shotgnm and Rifle Firing
Pin Inpressions. arm of Forensic Sciences, 19: 441-
447.

Haas, M.A., Camp, M.J., and Dragen, R.F. 1975. A Oonparative
Study of the Applicability of the Scanning Electron Micro-
scopeandtheLightMicroscopeintheExaminationof
Vehicle Light film. W. 20:
91-102.

Hayat, 14- A. 1978. W
W. University Park Press: Baltinnore.

Jolly. C- J. and Plog, F. 1987. mm
W. Alfred Knopf: New York.

Judd, G., Sabo, J., Hamilton, W., Ferries, S. and Horn, R.
1974 . 8m Microstriation Characterization of mllets and
Contaminant Particle Identification. W
m, 19: 798-811.

Katterwe, H., Goebel, R., Grooss, K.D. 1982. The Oonparison
Scanning Electron Microscope Within the Field of Forensic
Science. MW. Part 2: 499-504-

52

Knnudsen, S. J. 1978. W. RanndNally
College Rnblishing: Chicago.

thricardi, V.R., Clark, M.S., and DeRonja, F.S. 1975. The
OonparisonofBrokenSurfaes: AScanningElectron
Microscopic Study. Mi of Forensic Scigg, 20: 507-

523.

Mayr. E- 1982- mm. Bellmap
Press: Cannbridge, MA.

Metals Handbook, Volune 2. 1964. t t' Cl ' and
Finishing' . 8th ed. Anerican Society for Metals. Metals
Park, Ohio.

Mittlonnan, R.E. and Wetli, C.V. 1982. The Threaded Bolt

Injury Pattern. Journal of Forensic Scim, 27: 567-
571.

Moenssene, A. A., Inbau, F. E., andStarrs, J. E. 1986.

W 'Ihchundatim
Press. Mineola, NY.

Pfister, R. 1982. The Use of the Scanning Electron Microscope
and Associated Techniques in Forensic Sciences (A Bibliog-
raphic Update) W. Part 3:
1037-1042. .

Potts, R. and Shipnan, P. 1981. Cu‘tmarks Made by Stoe Tools
on Boes fronn Olduvai Gorge, Tanzania. Nature, 291:577-
580.

Rao, V.J. and Hart, R. 1983. Toolnnark Determination in
Cartilage of Stabbing Victim- W
m, 28: 794-799.

Rose, J. J. 1983. A Replication Technique for Scanning
Electron Microscopy: Applications for Anthropologists.

WW 62: 255-261-

53

Rowe, W. F. 1988. Firearms Identification. Chapter 8 in,

W R. Saferstein, ed-
Prentice Hall: Englewood Cliffs.

Sauer, N. J., Dmlap, S. S., and Simson, L. R. 1988.
Medicolegal Investigation of an Eighteenth Century
Hamicide- W
m, 9: 66-73.

simsm, G. G. 1982. W
Scientific American Library, W.H. Freanan and Co.: New
York.

Shipmn, P. 1981. Applications of Scaming Electrcm
Microscopy to Tapaonanic Problems. In, W
Mtials of Mlgical Museum Collections, Annals

of the flew Yorg Lagdgm Q: Sgigxgg, Vol. 376. A.M.
Cantwell, J.B. Griffin, and N.A. Rotschild, eds.

Shipnan, P. 1983. Early haninid lifestyle: Hunting and
gathering or foraging and scavenging? EILLSD
W. in Press

Shiman, P. and Rose, J.J. 1983. Early Haninid Hurting,
Butchering, and Carcass-Processing Behaviors: Approaches
to the Fossil Recond- W
may. 2357-98-

Singh, R.F. and Agganval, H. R. 1984. Identification of Wires
ardtheQrtting'I’oolbySwmingElectrmMicroscope.

Wm 26: 115-121.

Speeter, D., Ohnsorge, J. 1973. Rasterelektraaernnikroshop—
isd1erfafibare$dezeid1enamKnod1emW

W 73: 137-143.

State V. W. 1982. 231 Kan. 408, 646 P. 2d 1049.
State V. Clark. 1930. 156 Wash. 543, 287 P.18.

Taylor, M.E. 1973. Scaming Electron Microscopy in Forersic

Science. MW. 13: 269-
280.

54

rI‘edeschi, C. G., Eckert, W. G., 'Iedeschi, L.G. 1977. m;
m: Volume I, Mechanical Trauma. W. B. Saunders
Oo.: Philadelphia.

'Ihanas, F. 1967. Garments on the Discovery of Striation
Matching and on Early Contrthtiors to Forensic Firearm
Identification. Journal of Fogic giences, 12: 1-7.

Wong, Y. S. 1982. Forensic Applicatiors of Scanning Electrm
Microscopy/Energy Dispersive X-ray Analyser in Hong Kong.
WW Part 2: 591-597-

”TIE 11111111111111“