Sharp force trauma relating to cultural anthropology

Forensic Science International 299 (2019) 119–127

Sharp force trauma analysis in bone and cartilage: A literature review

Jennifer C. Love Office of the Chief Medical Examiner, 401 E Street, Washington, DC, 20024, United States

A R T I C L E I N F O

Article history: Available online 26 March 2019

Keywords: Sharp force trauma Saw mark Cut mark Tool mark

A B S T R A C T

Sharp force trauma (SFT) in bone and cartilage has been studied extensively. This literature review summarizes knife and saw mark research. Researchers have documented several features of cut surfaces and successfully associated them with various tool characteristics. Most study designs are based on light microscopic examination, but other technologies such as micro-computed scanning, scanning electron microscope, and epifluorescence microscopy have been investigated. Researchers have worked with human and non-human material, and found that the presentation of SFT differs between the two. Furthermore, they have designed studies to control the parameters surrounding SFT (e.g., tool angle, force, direction) as well as not to control these parameters (real-world scenario) and have found that the trauma produced in the two scenarios differ considerably. Researchers have attempted to calculate the error rate associated with cut and saw mark analysis and have reported very different results. Several high profile cases of successful SFT analysis have been published and are briefly reviewed. Expert testimony based on cut and saw mark analysis has been found admissible, but not in all cases. Unfortunately, researchers have not consistently used standard terminology, a list of terms gathered from the literature is provided. Despite the extensive research, more work is needed. Methods that mitigate potential sources of error that are not dependent on analyst’s experience must be developed.

© 2019 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Forensic Science International

journal homepage: www.elsevier.com/locat e/f orsciint

Sharp force trauma (SFT) is produced by a tool that is edged, pointed or beveled [1]. Traditionally, anthropologists have evaluated SFT observed on bone and cartilage through microscope analyses by describing features of the mark and associating them with a type or an individual tool. The majority of the research has focused on saw marks and knife marks and several articles are summarized here. Other beveled tools such as machetes have been minimally researched and are outside the scope of this article (for examples see Refs. [2–4]). Furthermore, anthropologists have investigated dismemberment and human mutilation and attempted to infer intent or state of mind of the perpetrator (for examples see Refs. [5–8]); this too is outside the scope of this article. The goal of this article is to succinctly summarize previous research in knife and saw mark analyses and to showcase gaps in the methodologies requiring additional research. Many terms may be new to readers; rather than define each term in the text, they are defined in Table 1.

Prior to summarizing the literature of knife and saw mark analyses, some fundamental principles should be discussed. First, the language used in the literature is not consistent. Symes et al. attempted to standardize the language [9], but not all researchers have followed suit. Three terms that appear to be

E-mail address: [email protected] (J.C. Love).

https://doi.org/10.1016/j.forsciint.2019.03.035 0379-0738/© 2019 Elsevier B.V. All rights reserved.

used interchangeably are cut mark, saw mark and tool mark. For this article, saw mark is defined as a mark made with a reciprocating action of the tool; cut mark is defined as a mark made with a single action of the tool. Usually, a cut mark is made with a knife, is V-shaped and has limited information; however, cut marks in cartilage may capture several features of the tool. Typically, a saw mark is made with a saw and tends to record more characteristics of the tool in the floor and walls of the mark. For this article, the term tool mark is reserved for the field of tool mark examination. Also, it is used as a general term referring to a mark that could be made by any type of tool in any type of medium.

Features of cut and saw marks are classified as class or individual characteristics. Class characteristics are indicative of a tool type, for example a serrated knife or alternating saw. Individual characteristics are indicative of a specific (individual) knife or saw. Some researchers have argued that a cut or saw mark can be matched to a specific tool using individual characteristics. This is a controversial area within forensic science. The Scientific Working Group of Anthropology clearly states in the Trauma Analysis guideline that identifying a specific tool from features of the cut or saw mark is an unacceptable practice [1]. Research that ventures into the identification of individual characteristics of a cut or saw mark is summarized below, but with reservation. The author does not advocate for matching a cut or saw mark with a specific tool.

Table 1 Saw and cut mark analysis terminology.

General terms Tool type A class of a tool such as serrated knife, alternating saw, hacksaw Class characteristic Feature observed in a cut or saw mark that is indicative of a tool type Individual characteristic Feature observed in a cut or saw mark that is specific to a tool Crime mark A cut or saw mark created with an unknown tool often the mark found in the decedent Test mark A cut or saw mark created with a known tool often created to compare to a crime mark

Tool characteristics Tooth distance The distance between teeth of a saw Tooth type The shape and beveling of each tooth of a tool (typically a saw) Power source Manual or mechanical (e.g., electric, gas, pneumatics) Tooth set How the teeth are placed along the blade; for example, an alternating saw has a pattern of a tooth bent to the right, then a tooth bent

to the left, and so on. Teeth per unit length (Tooth per inch (TPI))

Number of teeth per each unit length of the saw blade

Interserration distance Distance between two consecutive teeth along a serrated knife Degree of wear Wear of the tool from use

Cut/saw mark features Kerf Cut or saw mark Kerf wall Wall of a cut or saw mark Kerf floor Floor of a cut or saw mark False start Incomplete saw mark Kerf width The distance across the saw or cut mark typically measured at the minimum width Kerf wall shape Description of the saw or cut mark wall alignment when viewed from directly above the surface Kerf floor shape (also known as kerf class)

Shape of the kerf floor when viewed from the side of the cut surface

Tooth hop Waves observed in the kerf wall caused by the saw moving up and down as each tooth enters the material Blade drift (or cut surface drift) Directional change (drift) in the markings of the saw mark floor created by the back and forth motion of the saw as each tooth enters

the material, typically seen in marks created with alternating saw. Harmonics Distance between peaks observed three-dimensionally in the kerf wall Interstriation distance Distance between two striations measured on a cut surface. Micro-striation Striation that reflect individual defects in the tool's beveled edge typically less than 1mm in interstriation distance Pullout striations Grooves in the kerf wall that are created when a saw is lifted out of a saw mark (in a perpendicular direction) mid-cut. The distance

between two grooves is the distance between two teeth set to the corresponding side of the saw. Striation regularity Uniform placement of striations along the kerf wall indicative of power source. Cut mark shape Shape of the cut mark viewed perpendicular to the cut surface. For marks created with serrated and nonserrated knives, it is defined

as Y, T or V shape. Cut mark width Distance across a cut mark (similar to kerf width) Cut mark length Length of a cut mark along its long axis Wall angle (cut mark) Maximum angle between the two adjacent walls intersecting on the cut mark floor Serrated angle (only present in Y shape cut marks)

Maximum (obtuse) angle between the wall that does not intersect the floor and its adjacent wall which does

Floor radius Radius of the circle whose perimeter in tangential to the two adjacent walls intersecting on the cut mark floor Features reflecting direction of cut and progression of the tool

Direction of saw progress (Direction of cut)

Direction the saw or knife passes through the bone in regards to anatomic position

Breakaway spur Spur of bone at the endpoint of a complete saw cut Exit chipping Small divots in the margins of the kerf wall created as the saw teeth exit the bone Power stroke Push or pull of the saw that cuts material. Typically saw teeth are chiseled on one side and the saw only cuts when moving in one

direction. Passive stroke Pull or push of the saw that does not cut the material

120 J.C. Love / Forensic Science International 299 (2019) 119–127

1. Saw marks in bone

Bonte [10] published a seminal article in saw mark analysis in which he identified metric features of saw marks on bones (Fig. 1). Bonte’s analysis included the examination of the shape and pattern of striations observed in the wall (termed “kerf wall”) of a saw mark and showed that several class characteristics of the saw are recorded in the striation pattern. Bonte identified differences in the striations produced by the passive and power strokes. The power stroke, the movement that cuts through the material, created fine marks in the kerf wall resulting from each tooth. The passive, non- cutting, stroke created crude, deeper furrows resulting from all the saw teeth being pulled along a single level. The number of fine striations between two deep furrows reflected the number of saw teeth engaged in the stroke. Bonte found that the number of engaged teeth was approximately two-thirds the total number of teeth in the saw blade and that the number of deep furrows reflected the number of strokes completed during the cut. He also identified regularly spaced scratches, termed “pull-out striations”, which ran perpendicular to and crossed multiple layers of

striations. These marks were produced by grazing teeth when a jammed saw was lifted out of a saw mark. The distance between these perpendicularly oriented striations reflects the distance between the teeth set to the corresponding side of the saw blade.

Bonte’s work was followed by Andahl [11]. Andahl divided the cut mark into two components: the cut mark floor (termed “kerf floor”) and kerf wall. He found recorded in the kerf floor, both in false starts (incompletecuts)andonthebreakawayspurofcompletecuts,features that reflected the number of teeth per unit length, degree of wear, direction of cut and condition of the blade. He also showed that the placement and condition of the striations impressed into the kerf wall reflectedthetoothset. Andahlproposedathreestepmethodtoanalyze a saw mark in order to identify the potential saw type: (1) identify a possible saw type through detailed examination of the crime mark; (2) create test marks using the inferred saw type, or the suspect saw when available; and, (3) compare the characteristics of the test mark to the crime mark. Andahl recognized that the striation pattern observed in the kerf wall was complex and recommended creating a pencil rub of the cut surface using various amounts of pressure or pressing the cut surface on iodine impregnated paper to reduce the “noise”. Andahl

Fig. 1. Shown is a kerf wall of a saw mark in a femur. The measurements are taken from the peek to peek of a tooth hop. An eight teeth per inch alternating saw was used to make the mark.

J.C. Love / Forensic Science International 299 (2019) 119–127 121

stated that, when comparing crime and test marks, the accuracy of the analysiswasdependentontheamountofconsistencybetweenthetwo marks. He continued, “if the examiner is experienced and the crime mark contains sufficient detail then it will be found that only a blade withthesamefeatureswillyieldatestmarkcomparablewiththecrime mark” ([11]; p 40–41). Andahl stopped short of stating a specific weapon can be identified through a crime mark/test mark comparison, but stated “if the saw is damaged the evidential value of a positive comparison will be correspondingly higher” ([11]; p 44).

In 1992, Symes completed his dissertation titled Morphology of Saw Marks in Human Bone: Identification of Class Characteristics [5]. Using 26 types of saws and serrated knives, Symes made ten cuts with each tool in human bone. Through microscope analysis, he observed numerous features of each mark that reflected the class characteristics of the various tools. In his dissertation, Symes listed five features of each tool and six features of the corresponding mark made from each tool. The five characteristics of the tool were: tooth set; tooth per inch; tooth distance; tooth type; and, power source. The six characteristics of the saw mark were: kerf floor shape; minimum kerf width; blade drift; cut surface drift, exit chipping; and, harmonics. Symes did not weight the informative value of each feature or discuss how to proceed when features indicative of more than one tool were observed in a single cut.

Saville et al. [12] investigated the value of an environmental scanning electron microscope (ESEM) for saw mark analysis. With the higher magnification of the ESEM, the researchers recognized a new feature of a saw mark that they termed micro-striation. The authors found that the typical distance between deep furrows formed by two passive strokes was 1–4 mm, fine striations formed by a powered stroke was 30–400 mm, and micro-striations was less than 1 mm. They hypothesized that the micro-striations reflected individual defects in the tool’s beveled edge. To test this hypothesis, the researchers made several test marks and mock crime marks with four new and five used Jackplus Hardpoint Handsaws. The researchers compared the crime marks to the test marks and made a correct match for each comparison. Based on the results, the authors concluded that use of ESEM to analyze a saw mark expanded the discriminatory power of the analysis to the point that enabled individual tool identification.

Freas [13] compared the value of light microscope and scanning electronic microscope (SEM) while evaluating the effects of saw wear on the expression of saw mark features [13]. She examined consecutive cuts made in bone using a crosscut saw and hacksaw with a light microscope and SEM. The author found that the striations, especially the fine striations of the power stroke, became less defined with each consecutive cut and the associated wear of the saw teeth. The loss of detail was more significant in the marks made with the crosscut saw than the hacksaw. She hypothesized that the difference was due to the material hardness of the saw blade; a crosscut saw is designed to cut wood and a hacksaw is designed to cut metal. Despite the loss of detail, Freas identified tooth hop in the saw marks made with the worn saws, allowing her to conclude that a saw mark made with a worn saw was eligible for analysis. Contrary to Saville et al. [12], Freas found that the SEM was not analytically valuable for saw mark analysis on bone; the high powered magnification obscured the overall striation pattern important for class characteristic recognition.

In 2011, Bailey and van de Goot performed a study focused solely on the kerf width of false starts [14]. The researchers hypothesized that each saw type produces a different kerf width and kerf width could be used to accurately exclude saws as possible sources of a specific saw mark. To test their hypothesis, they used ten saw types, five manual and five mechanical powered, to make 100 false starts, 10 saw marks per saw. The kerf width of each false start was measured with dial calipers. Statistical cumulative logistic regression models were used to analyze the data. Based on the statistical model, 70–90% of the saws could be accurately excluded as possible saws for each saw mark based on the minimum kerf width. In a similar manner, Nogueira et al. [15] evaluated the kerf width, kerf wall shape and kerf floor shape of saw marks created in human and porcine bone using five saw types. The authors found that the three features were useful in identifying the class of the saw used to make the mark. Interesting, they found differences in the saw marks made in porcine bone and the saw marks made in human bone, indicating that porcine bone may not be a good model for human bone in some studies.

Pelletti et al. [16] examined the value of micro-computed (micro-CT) scans for analyzing saw marks. The researchers

122 J.C. Love / Forensic Science International 299 (2019) 119–127

examined 170 false starts experimentally created on porcine and human femora using five different saws. Examining the micro-CT scans, they recognized striae in the kerf floor, a feature missed with the stereomicroscope. The researchers concluded that the striae were informative to the tooth set. Further, the researchers found differences in saw marks made in human and porcine bones with a single saw, warning that bone source may affect the saw mark morphology.

Norman et al. [17] also investigated the value of micro-CT scans for saw mark analysis. The researchers conducted a study of 270 saw and cut marks made in human bone using eight tools (including two serrated knives) to answer four questions: (1) are micro-CT scans of saw marks valuable for saw mark analysis; (2) are saw marks created by different tools significantly different; (3) can kerf width be used to predict blade width; and, (4) do saw marks created under different conditions vary significantly? To answer these questions, saw marks were made with the saw controlled in a vice as well as made with an uncontrolled saw mimicking true crime marks. Also, some of the bones were defleshed and other bones had significant soft tissue at the time the saw marks were created. Two analysts scored the kerf wall and kerf floor shapes and measured five dimensions of each mark. Evaluating the two qualitative assessments and five quantitative assessments, the authors concluded that micro-CT is an appropri- ate tool for analyzing saw marks. The target variables were visible in the scans and the inter-observer measurement agreement was high. The minimum kerf width was the most diagnostic feature of the analysis with a significant difference between all tool types, except for the two knives and two of the saw types. Also, the researchers found a significant correlation between the kerf width and blade width and were able to predict the blade width from the kerf width with a high rate of accuracy when the bone was defleshed and the saw cut controlled. Finally, the authors found that various conditions (fleshed bone vs. defleshed bone and controlled cuts vs. uncontrolled cuts) affected the features of the saw mark indicating that controlled experimental saw marks may not reflect the variability observed in forensic case work.

Greer [18] used surface metrology to examine striations recorded in kerf walls of saw marks made with hacksaws and reciprocating saws. Surface metrology enables quantitative analy- sis of stria magnitude. The researcher evaluated the stria magnitude of 87 saw marks made in porcine bone using six reciprocating and eight hacksaws. Greer found a high level of variability in stria magnitude across a single kerf wall as well as between kerf walls created with different saw types. The analysis showed that the stria magnitude was a poor feature for distinguishing between hacksaw and reciprocating saw marks.

Berger et al. also examined saw mark features resulting from a reciprocating saw and hacksaw [19]. For the study, the researchers used five reciprocating saw blade types and one hacksaw to make saw marks in partially defleshed deer limbs. Each blade was used to make 10 complete saw marks and two false start kerfs. Multiple cut marks were also made in deer cervical vertebrae to mimic decapitation. Several unintentional false starts and deep scratches were created while sawing the bones resulting in a total of 89 marks available for analysis. Prior to analysis, the bones were macerated to remove the soft tissue. The saw marks were examined with a light microscope. The saw marks were examined for features following Symes [5] and [9] (described below). Of the traits examined, six traits were significantly different between blade types, and included: kerf width; false start shape; frequency of blade drift; harmonics; exit chipping; and, striation regularity. The researchers concluded that these features can be used to accurately distinguish between saw marks made with a recipro- cating saw and a hacksaw. Also, they concluded that different reciprocating blades created different features in the saw mark.

Several researchers have focused their attention on analyzing saw marks in compromised bone [20,21]. Pope and Smith examined experimentally induced saw marks in burned skulls (2004). The saw marks were easily recognized in the burned specimen. Marciniak also examined saw marks in burned bone (2009). She found that thermal destruction affected the visibility and recognition of saw mark striations, but that false starts were well-preserved in thermally damaged bones. She concluded that valuable class characteristics of a saw mark are present in burned bones [21].

In 2010, Symes et al. published a technical report for their National Institute of Justice grant funded project titled Knife and Saw Toolmark Analysis in Bone: A Manual Designed for the Examination of Criminal Mutilation and Dismemberment [9]. The objective of the study was to improve the scientific rigor of saw mark analysis in order to increase its admissibility in court. The project goals were (1) develop a manual of standardized terminology and illustrative images, (2) use the manual to train new analysts in saw mark analysis, and (3) test the accuracy of these new analysts. Symes et al. successfully developed the manual and used it to train analysts. To investigate the error rate of the new analysts, the researchers asked 12 analysts to examine 20 saw marks and classify each as resulting from a manual and mechanical powered saw. The analysts correct classification rate was approximately 70%. In addition to meeting the stated goals, Symes et al. proposed in their technical report a saw mark analysis website which would provide access to a comparative collection database consisting of numerous saw marks created with 27 saws. Unfortunately, it does not appear that the website is currently available; the authors do not provide a website address.

Rather than continuing on the path of evaluating the error rate associated with saw mark analysis, Love et al. [22] attempted to define potential sources of error and develop a method to mitigate the error. Borrowing heavily from the field of latent print examination, the authors presented four possible sources of error associated with saw mark analysis: human factor error; procedural error; error of variability; and, outcome error [22,23]. Human factor error and procedural error are failures directly related to the analyst. Human factor error is an error in performing an analysis such as mismeasuring a variable. Procedural error is a failure to follow a specific procedure such as a failure to examine and document all features associated with an analysis. These errors can be mitigated with appropriate training and proficiency testing, and were the focus of the work done by Symes et al. [9]. Error of variability is an error associated with intra-individual variability, e.g., differences between saw marks made with a single tool, and inter-individual variability, e.g., differences between saw marks made with two different tools. Outcome error is the reporting of the wrong event. For saw mark analysis, it is reporting a saw mark was created using an alternating saw when the mark was actually created with a rip saw. An ideal method utilizes features with low intra-individual variability and high inter-individual variability and limits or reports the potential for outcome error. Previously, no study attempted to measure the intra- or inter-individual variability of saw mark features or attempted to weight the informative value of the respective features. The traditional method provides no guidance on which features are more informative or how to handle conflicting features, features indicative of more than one saw type. The analyst depends on his experience to navigate these difficulties. Love et al. [22] tried to address these concerns using statistical modeling: tree classifica- tion; and, random forest classifier. The researchers created 30 complete and 30 false start saw marks using four saws. Three analysts examined each saw mark for 15 features following the training manual developed by Symes et al. [9]. Five of the 15 features were observed too infrequently to be included in the

Fig. 2. (A) Shown is a cast of pig costal cartilage cut with a serrated knife. The knife had a series of coarse and fine serrations creating a primary and secondary pattern of striations. (B) Shown is a cast of pig costal cartilage cut with a nonserrated knife. Note the absence of striations.

J.C. Love / Forensic Science International 299 (2019) 119–127 123

statistical modeling. Several classification and random forest classifiers were run on the data and floor shape and minimum kerf width were found to be informative variables for the specific data set. The classification tree enables the discriminatory value of each variable to be measured and constructs a pathway to an outcome based on the empirical data and not the experience of the examiner, reducing error of variability and outcome error.

2. Cut marks in bone and cartilage

Research in cut mark analysis includes the examination of marks on bone and cartilage. Cut marks in bone record limited characteristics of the tool; often kerf floor shape and minimum kerf width are the only variables recorded. However, cut marks in cartilage often record characteristics indicative of the tool type, including serrated or non-serrated, and if serrated, the serration pattern (Fig. 2). Research in cut mark analysis has focused on recording cut mark features and associating them to tool type, evaluating examination modalities, measuring inter- and intra- individual variability, and identifying error rate.

Bronte [41] published one of the first article describing cut marks in cartilage. In his article he very briefly describes how cut marks present in cartilage; then, proceeds by providing a long description of saw marks in bone. Little can be gleaned from the article, but it did set in motion a body of research critical to the development of cut mark analysis in cartilage. Watson [24] examined impression evidence made from two consecutively manufactured Buck knives and found that each produced a unique striation pattern. Rao and Hart [25] publish a case study in which a crime mark in human costal cartilage was matched to a test mark made with the suspect weapon. In order to facilitate the comparison, the cut mark in costal cartilage was cast with dental impression material. The suspect knife was used to make test marks in cellulose acetate butyrate, a yellow rubbery material. The crime mark and test mark were compared using a firearm comparison microscope. The major striations of the test mark matched the major striation of the crime mark. The analysts concluded a match between the crime and test marks.

Bartelink et al. [26], examining cut marks on bone, evaluated the discriminative value of minimum cut mark width measured with a SEM. The researchers used three nonserrated blades (scalpel blade, paring knife, and kitchen utility knife) to create 60 cut marks (20 per blade type) in a human long bone. They created controlled cut marks using a mechanical device that controlled for angle, direction and force and experimental cut marks by hand without control of the various parameters. The cut mark width was significantly different for all three blades, but it also was significantly different between the control and experiment cuts made with each of the blades. Furthermore, there was overlap of the cut mark widths between the three knife types limiting the discriminative value of the feature.

Cerutti et al. [27] also evaluated the relationship between the dimensions of a cut mark and the dimensions of the knife blade. To perform the experiment, the researchers created cut marks on porcine femoral diaphyses using 11 different knives and a mechanical device that controlled for blade angle and force. Ten cut marks were made with each blade held perpendicular to the bone surface and 10 cuts marks were made with each blade held at an angle to the bone surface. Depth, width and angle of the cut marks were measured and compared to the beveled angle of the blades. The measurements of the cut marks showed wide overlap between the various blades and did not correlate with the dimensions of the blade.

Norman et al. [17] investigated the value of micro-CT scans for cut mark analysis on bone. The researchers designed two studies. For the first study, they created 14 cut marks on dried porcine ribs

using two knives, one serrated and one nonserrated. The researchers used a mechanical device to create the cut marks controlling the level of force and blade angle. For the second study, the researchers used nine knives to create 273 cut marks in a porcine thorax. Two male participants created the cut marks

124 J.C. Love / Forensic Science International 299 (2019) 119–127

without controlling for force or angle in an effort to mimic a real- world scenario. The cut marks were examined with micro-CT scanning, optic laser scanning, SEM, and light microscope. The researchers recorded several features, many newly defined, of the cut marks, including: cut mark shape; length; wall angle; serrated angle; floor radius; and striations. The researchers found that they could easily score the various features on the micro-CT scans and that measurements were taken with high inter-observer reliability. SEM was required to observe the striations; the resolution of the micro-CT was not adequate for this purpose. However, striations were observed on few of the specimens. Furthermore, optic laser scanning did not capture the features of the cut marks well. The researchers found that knife type can be correctly identified from cut mark width and floor radius, and that blade edge thickness is correlated with cut mark width. Interesting, the cut marks made following the second study design (real-world scenario) were more variable than the cut mark made following the first study design resulting in greater inter-observer error associated with scoring cut mark shape.

Komo and Grassberger [28] also evaluated the value of micro- CT as well as macro photography and SEM for cut mark analysis in bone. Also, the researchers considered the effect of maceration on the cut mark dimensions. To perform their study, the team cut partially fleshed porcine ribs (skin and fat removed, muscle in place) with 12 knives (serrated and nonserrated) using a mechanical, guillotine-style device to control knife angle and force; a total of 27 cut marks were created. The specimens were scanned with the micro-CT prior to and after maceration; the ribs were photographed and examined with the SEM after maceration. The depth, width and length of the kerf were measured as well as several angles: opening angel; upper and lower slope angle; and, floor angle. The researchers found that maceration caused the kerf width to shrink by 8.6%. They also found that the micro-CT scan provided the most information of the cut marks. Finally, they found that measurement variance (e.g. kerf width) was greater between marks created with a single knife than cut marks created with different knives.

Capuani et al. [29] evaluated the value of epifluorescence microscopy for cut mark analysis in bone. The research team examined multiple cut marks made in partially defleshed human clavicles and ribs using a hatchet (n = 11), serrated knife (n = 23) and nonserrated knife (n = 24). Each specimen was macerated to remove remaining soft tissue; then, examined using an epifluor- escence microscopy. The kerf width, shape, and size of bone flakes in the kerf floor were evaluated for each cut mark. The researchers found that kerf width, shape, and lateralization (differences in the two walls of the kerf) are linked to blade shape. Furthermore, they found that the serrated knife consistently produced larger flakes of bone within the kerf. Based on the features, the researchers concluded that the kerf width is wider for serrated than nonserrated knives and cross-sectional shape of the kerf is different for the two blade types. They identified features that indicated the beginning and end points of the cut mark, informative to the direction the tool was moving as the cut mark was created. Also, they found slight differences between the marks made in the ribs and clavicles, and hypothesized that the differences were due to variations in cortical bone thickness and density. The researchers concluded that epifluorescence micros- copy is a valuable tool for cut mark analysis.

Sandras et al. [30] also evaluated the value of epifluorescence microscopy for cut mark analysis in bone. The stated goals of the study were to define the error rate and predictive values associated with matching a tool type with the cut mark features and to evaluate the importance of analyst experience. The researchers designed a three phase study. For each phase of the study, cut marks were made in partially defleshed human clavicles using five knives; three

nonserrated; and, two serrated. Three examiners with varying degrees of training participated in the study. For the first phase, eight cut marks were made with each knife and examined by the analyst with thegreatesttraining.Theanalystidentifiedsevenfeaturesof the marks: kerf thickness; kerf shape; edges; walls; edge vs. wall angle; profile shape; and, flakes. Unfortunately, the authors do not clearly define these features, but state that the features are qualitative. For the second phase, the same analyst examined 30 cut marks made with a single serrated knife to confirm that the chosen features were visible in a high number of cut marks. For the third phase, 50 cut marks were made with the five knives and the three analysts attempted to match each cut mark to the specific knife. The results of the study showed an accuracy rate of 74–94%, the accuracy of matching the cut mark to the tool was greater for serrated knives than nonserrated knives. The positive predictive value (56–90%) was variable depending on knife type; it was higher for serrated than nonserrated knives. The negative predictive value (93–97%) was consistently high for all analysts showing a low rate of excluding the correct knife. The researchers recognized that the accuracy rate of their study was higher than rates reported by others [9,22,31,32] and they hypothesized that the improvement was due to the inclusion of a high number of qualitative features. Of concern, the authors highlighted the importance of analyst experience in correctly performing the analysis, raising question of potential outcome error.

Pounder et al. [33] examined cut marks in bovine and porcine costal cartilage to identify class characteristics. The researchers created 136 cut marks with eight serrated knives and 72 cut marks with four nonserrated knives. Each cut mark was opened, separating the two cut surfaces of the wound track. The cut surfaces were photographed, dyed with food coloring and rephotographed, and cast with dental molding. Prominent striations were observed in the cut surfaces created with serrated knives. Subtle, much finer striations were observed on the cut surfaces created with the nonserrated knives. No robust striations were observed on the cut surfaces created with the nonserrated knives. The researchers hypothesized that the subtle striations reflected small defects in the beveled edge of the nonserrated knives. Finally, the cut marks created with coarsly serrated blades were distinguishable from cut marks made with finely serrated blades.

Puentes and Cardoso [34] examined the effects of blade penetration angle on cut mark characteristics as well as potential variation introduced by the quality of the victim’s costal cartilage. The researchers created 120 cut marks in six chest plates removed during autopsy (individuals aged 20–60 years) using three serrated knives with different serration morphologies. Each knife was used to make 40 cut marks, 20 cut marks per chest plate. Of the 20 cut marks, 10 were made with the knife passing through the cartilage in a front to back direction, and 10 were made with the knife passing through the cartilage in an up to down direction. The researchers’ questions were threefold: (1) is the interstriation distance of the cut surface consistent with the interserration distance of the knife blade; (2) what is the inter- and intra-observer error; and, (3) are cut surfaces made with the same knife in different individuals similar? The researchers found that when the cut mark was made with the knife passing through the cartilage in an up to down motion, the mean interstriation distance and mean interserration distance were not significantly different. To evaluate the inter-observer error, both researchers measured the interstriation distance of 10 cut surfaces; they found that the inter-observer reliability was high. To evaluate the intra-observer error, one researcher re-measured the interstria- tion distance of 10 cut surfaces and found that the intra-observer reliability was high as well. Finally, the interstriation distances of cut mark surfaces made with the same knife on two individuals were consistent,except foronecase.Inthatcase,one of thetwoindividuals was the oldest individual included in the study (60 years) with the

J.C. Love / Forensic Science International 299 (2019) 119–127 125

most calcified cartilage. The researchers hypothesized that the extent of calcification affects the recording of cut mark character- istics on the cut surface.

Pounder and Sim [35] evaluated the value of the micro-CT scans for analyzing cut marks in cartilage. The researchers created 12 stab marks in porcine costal cartilage using a serrated knife. They prepared the cut marks (wound tracks) for scanning following three separate methods: (1) coating the surfaces with osmium tetroxide solution; (2) opening the wound track slightly creating a space between the two surfaces; and, (3) completely opening the wound track. Examining the micro-CT scans of the wound tracks, the researchers identified regular striation patterns on the completely separated surfaces and the surfaces coated with osmium tetroxide solution. The scans of the wound tracks with slight distance between the two surfaces were not adequate for analysis. The researchers conclude that the use of micro-CT is an efficient, non-destructive alternative for examination of cut surfaces.

Love et al. [31] and Crowder et al. [36,32], both attempted to validate cut mark analysis in cartilage. Love et al. limited there study to evaluating the discriminative value of cut mark class character- istics observed in costal cartilage [31]. The researchers used three blade types (nonserrated, serrated with coarse serrationpattern, and serratedwith coarse and fine serrationpattern) to make 90cutmarks (30 per knife) in porcine cartilage. Three analysts examined each cut surface. The analysts documented the presence of striations, whether the striations occurred in an irregular pattern (indicative of a nonserrated knife), regular pattern (indicative of a serrated knife), or regular with a primaryand secondary pattern (indicative of a serrated knife with coarse and fine serrations). The researchers found a very high error rate (>65%) among the three analysts; cut marks made with all three knives were misclassified. When interstriation distance was included in the analysis, class classifica- tion accuracy increased slightly.

Crowder et al. [36,32], similarly evaluated the accuracy rate of cut mark analysis, but had very different results than Love et al. [31]. Crowder et al. designed a study using 14 knives with a range of blade characteristics including serrated, partially serrated, and nonserrated to make cut marks in porcine costal cartilage, deer long bone and casting wax. The researchers created two cut marks in the casting wax with each knife. One cut was made with the knife moving in a single down and back motion mimicking a stabbing action. The other cut was made with a sawing action. Also, they created one cut mark with each knife in the costal cartilage using the single down and back motion and one cut mark with each knife in the long bone using the sawing action. A total of 56 cut marks were created: 28 cut marks in wax; 14 cut marks in cartilage; and, 14 cut marks in bone. The analysts recorded coarse striations (observed with the naked eye), fine striations (observed with the aid of a light microscope), combination of coarse and fine striations, or none. Striations were further recorded as patterned (regularly spaced), unpatterned (irregularly spaced), or a combi- nation of patterned and unpatterned. Cut surfaces with coarse regular striations were scored as resulting from a serrated or partially serrated blade and cut surfaces with no striations or fine, unpatterned striations were scored as resulting from a nonserrated blade (Fig. 2). The average error rate, i.e., identifying the wrong blade type based on the striation pattern, was approximately 5% for the three analysts. Furthermore, the researchers found experience level affected the accuracy rate with the more experienced researcher being more accurate.

3. Sharp force trauma, medicolegal investigation, and the courts

Several publications have highlighted the use of saw and cut mark analyses in medicolegal investigations. In her 1998 text,

Forensic Osteology: Advances in the Identification of Human Remains, Reichs showcased six dismemberment cases [6]. She included a simple protocol for collecting and processing skeletal elements during autopsy for cut and saw mark analysis as well as lists of gross characteristics of cut and saw marks and how to interpret them. In the same textbook, Symes et al. present a high profile medicolegal case in which microscopic saw mark analysis was performed. Based on the saw mark analysis the authors identified a likely saw type (power circular saw), a saw type similar to the saw the perpetrator ultimately admitted using [42]. Rainwater [8] details three cases of dismemberment investigated by the New York City Office of the Chief Medical Examiner. In each case, numerous details of the tools used to dismember the victims were obtained from the saw and cut mark analyses. Symes alone has reported performing microscopic saw mark analysis in approxi- mately 200 dismemberment cases [37].

Testimony based on the use of microscopic saw mark and cut mark analysis has been admitted into evidence in criminal trials of numerous cases. The following two cases ([38] and [39]) showcase the potential scrutiny surrounding the testimony. In State of Texas v Timothy Wayne Shepherd, the court admitted testimony based on saw mark analysis during the trial. The facts of the case were that the defendant was accused of killing and dismembering his girlfriend and burning some of her remains using a barbecue grill located on his apartment balcony. While searching the apartment, the garbage disposal was removed from the kitchen sink and submitted to the Harris County Institute of Forensic Sciences for DNA analysis. The garbage disposal was disassembled and small fragments of bone and enamel were found within. The bone fragments were burned, ruled of no value for DNA analysis, and transferred to the forensic anthropologist. Two bone fragments were examined. The fragments were very small and lacked features needed to differentiation between human and nonhuman bone; however, there were clear saw marks identified on the fragments. The anthropologist testified that the marks resulted from a saw and the likely class of the tool. The defendant was found guilty of murder and sentenced to 99 years. The defendant appealed the ruling on five separate points. One of the points was improper admission of the tool mark testimony. According to Rule 702 of the Texas Rules of Evidence, expert testimony must be reliable and relevant and the expert must be shown to be qualified before the testimony may be admitted [40]. The appellant argued the testimony failed to meet all three requirements: the anthropolo- gist was not qualified to perform tool mark analysis; tool mark analysis was not reliable; and, it was not relevant to the case. The Texas Court of Appeals disagreed on all accounts and upheld the original ruling.

A very different outcome occurred in the case of Ramirez v State of Florida. The facts of the case are that a woman was found stabbed and beaten in her place of employment. The sharp force trauma included one stab wound to her costal cartilage. A police technician who was qualified as a tool mark examiner testified that the crime mark matched the test mark created with the suspect tool to the exclusion of all other tools. The Supreme Court of Florida questioned the reliability of the testing method which formed the basis of the witness’ conclusion and found insufficient scientific support for the analysis. The court suggested that the method used by the analyst in this case departed from the generally accepted method of tool mark examination significantly enough to make this particular method novel, and thus inadmissi- ble. The distinction between the structural composition of costal cartilage relative to the traditional media to which tool mark methods are applied, the lack of photographic and written records associated with the analysis, and the consequent inability to test the theory were considered significant departures from the generally accepted tool mark examination method that has been

126 J.C. Love / Forensic Science International 299 (2019) 119–127

upheld in the courts. The testimony was deemed inadmissible in subsequent appeals.

4. Future of sharp force trauma analysis

The research outlined in this article clearly shows that saw and cut mark analyses are based on a strong scientific foundation. Researchers have designed numerous studies using various tools and mediums, including human bone and cartilage, and have repeatedly shown that features observed in cut and saw marks correspond to tool characteristics. Researchers have made great stride in standardizing the language used in tool mark analysis and the appropriate steps to perform saw or cut mark analysis [9]. Yet, there is still much to do.

As argued by Love et al. [22], the error rate associated with cut and saw mark analysis is poorly defined. Repeatedly, researchers have published an error rate resulting from their study. These error rates are specific to the individual studies and cannot be extrapolated to cut or saw mark analysis in general. For example, Crowder et al. [36,32] measured the error rate for their study, which they defined as the ability for the observer to correctly classify the tool type based on the cut mark features. The researchers reported a 5–7% mismatched between the cut mark features and the tool type. This error rate is specific to the 14 tools included in the study, the medium types being analyzed and the three analysts and cannot be extrapolated to a crime mark that is potentially created by any number of tool types examined by an individual of undefined expertise.

To strengthen saw and cut mark analysis, large databases of cut and saw mark features collected from test marks made with a large variety of known tools must be created. From this large body of data, the intra-individual variability of marks made with the same tool and inter-individual variability of marks made from different tools can be evaluated. Statistical models (such as classification tree) that weight characteristics with high inter-individual variability and diminish the influence of characteristics with high intra-individual variability as well as report the statistic equivalent of a potential error rate are needed. Only with such methods can the potential error associated with cut and saw mark analysis be defined.

5. Conclusion

In summary, extensive research on cut and saw mark analysis in bone and cartilage using various types of microscopes and scanning technologies has been conducted. The research shows that knives and saws create features in cut and saw marks that are indicative of the tool type and in turn these features can be used to accurately identify the tool type. Currently, a method that produces a result (an association between cut or saw mark and tool type) with a known error rate (or an equivalent statistic) by weighing features with high inter-individual variability and limiting features with high intra-individual variability is not available. Future researchers should focus their attention on developing methods that meet these needs keeping in mind potential sources of error and ways to mitigate it.

Conflict of interest

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

CRediT authorship contribution statement

Jennifer C. Love: Conceptualization, Writing - original draft, Writing - review & editing.

References

[1] Scientific Working Group of Anthropology (SWGANTH), Trauma Analysis, (2011) (Assessed on 11/21/2018) https://www.nist.gov/sites/default/files/ documents/2018/03/13/swganth_trauma.pdf.

[2] B. Tucker, D. Hutchinson, M. Gilliland, T. Charles, H. Daniel, L. Wolfe, Microscopic characteristics of hacking trauma, J. Forensic Sci. 46 (2) (2001) 234–240, doi:http://dx.doi.org/10.1520/JFS14955J.

[3] V. Alunni-Perret, M. Muller-Bolla, J. Laugier, L. Lupi-Peégurier, M. Bertrand, P. Staccini, M. Bolla, G. Quatrehomme, Scanning electron microscopy analysis of experimental bone hacking trauma, J. Forensic Sci. 50 (4) (2005) 213–216, doi: http://dx.doi.org/10.1520/JFS2003213.

[4] P. McCardle, T. Lyons, Machete cut marks on bone: our current knowledge base, Forensic Res. Criminol. Int. J. 1 (2) (2015) 00007, doi:http://dx.doi.org/ 10.15406/frcij.2015.01.00007.

[5] S.A. Symes, Morphology of Saw Marks in Human Bone: Identification of Class Characteristics. Dissertation, University of Tennessee, Knoxville, TN, 1992.

[6] K.J. Reichs, Postmortem dismemberment: recovery, analysis, and interpreta- tion, in: K.J. Reichs (Ed.), Forensic Osteology: Advances in the Identification of Human Remains, Charles C. Thomas, Springfield, IL, 1998, pp. 353–388.

[7] S.A. Symes, J.S. Williams, E.Z. Murray, J.M. Hoffman, T.D. Holland, J.M. Saul, F.P. Saul, E.J. Pope, Taphonomic context of sharp-force trauma in suspected cases of human mutilation and dismemberment, in: W. Haglund, M. Sorg (Eds.), Advances in Forensic Taphonomy, CRC Press, Boca Raton, FL, 2002, pp. 403– 434.

[8] C. Rainwater, Three modes of dismemberment: disarticulation around the joints, transection of bone via chopping, and transection of bone via sawing, in: N.V. Passalacqua, C.W. Rainwater (Eds.), Skeletal Trauma Analysis: Case Studies in Context, John Wiley &Sons, Ltd., Hoboken, NJ, 2015, pp. 222–245.

[9] S.A. Symes, E.N. Chapman, C.W. Rainwater, L.L. Cabo, M.T. Myster, Knife and Saw Toolmark Analysis in Bone: a Manual Designed for the Examination of Criminal Mutilation and Dismemberment. National Institute of Justice Technical Report, National Institute of Justice, Washington, DC, 2010 Report no. 2005-IJ-CX-K016 2010.

[10] W. Bonte, Tool marks in bones and cartilage, J. Forensic Sci. 20 (2) (1975) 315– 325.

[11] R.O. Andahl, The examination of saw marks, J. Forensic Sci. Soc. 18 (1978) 31–46.

[12] P.A. Saville, S.V. Hainsworth, G.N. Rutty, Cutting crime: the analysis of the “uniqueness” of saw marks on bone, Int. J. Leg. Med. 121 (2007) 349–357.

[13] L.E. Freas, Assessment of wear-related features on the kerf wall from saw marks in bone, J. Forensic Sci. 55 (6) (2010) 1561–1569.

[14] J.A. Bailey, F.R.W. van de Goot, Statistical analysis of kerf mark measurements in bone, Forensic Sci. Med. Pathol. 7 (2011) 53–62, doi:http://dx.doi.org/ 10.1007/s12024-010-9185-6.

[15] L. Nogueira, G. Quatrehomme, C. Rallon, P. Adalian, V. Alunni, Saw mark in bones: a study of 170 experimental false start lesions, Forensic Sci. Int. 260 (2016) 123–130.

[16] G. Pelletti, G. Viel, A. Viero, S. Visentin, D. Miotto, M. Montisci, G. Cecchetto, C. Giraudo, Micro-computed tomography of false starts produced on bone by different hand-saws, Leg. Med. 26 (2017) 1–5.

[17] D.G. Norman, W. Baier, D.G. Watson, B. Burnett, M.A. Williams, Micro-CT for saw mark analysis on human bone, Forensic Sci. Int. 293 (2018) 91–100.

[18] S.Y. Greer, Quantifying variation in kerf wall striations created by hacksaws and reciprocating saws, Forensic Anthropol. 1 (3) (2018) 160–169.

[19] J.M. Berger, J.T. Pokines, T.L. Moore, Analysis of class characteristics of reciprocating saws, J. Forensic Sci. 63 (6) (2018) 1661–1672.

[20] E.J. Pope, O.C. Smith, Identification of traumatic injury in burned cranial bone: an experimental approach, J. Forensic Sci. 49 (3) (2004) 431–440.

[21] S.M. Marciniak, A preliminary assessment of the identification of saw marks on burned bone, J. Forensic Sci. 54 (4) (2009) 779–785.

[22] J.C. Love, S.M. Derrick, J.M. Wiersema, C. Peters, Microscopic saw mark analysis: an empirical approach, J. Forensic Sci. 60 (S1) (2015) S21–S26, doi: http://dx.doi.org/10.1111/1556-4029.2011..12650.

[23] K. Melissa, D.T. Kaye, T. Busey, M. Gische, G. LaPorte, C. Aitken, et al., Latent Print Examination and Human Factors: Improving the Practice through a Systems Approach. NIST Interagency/Internal Report (NISTIR), National Institute of Standards and Technology, Gaithersburg, MD, 2012 Report No.: 7842.

[24] D.J. Watson, The identification of tool marks produced from consecutively manufactured knife blades in soft plastics, AFTE J. 10 (1978) 43–48.

[25] V.J. Rao, R. Hart, Tool mark determination in cartilage of stabbing victim, J. Sci. 28 (3) (1983) 794–799.

[26] E.J. Bartelink, J.M. Wiersema, R.S. Demaree, Quantitative analysis of sharp- force trauma: an application of scanning electron microscopy in forensic anthropology, J. Forensic Sci. 46 (6) (2001) 1288–1293.

[27] E. Cerutti, F. Magli, D. Porta, D. Gibelli, C. Cattaneo, Metrical assessment of cutmarks on bone: is size important? Leg. Med. 16 (2014) 208–213.

[28] L. Komo, M. Grassberger, Experimental sharp force injuries to ribs: multimodal morphological and geometric morphometric analysis using micro-CT, macro photography and SEM, Forensic Sci. Int. 288 (2018) 189–200.

[29] C. Capuani, N. Telmon, J. Moscovici, F. Molinier, A. Aymeric, M.B. Delisle, C. Guilbeau-Frugier, Modeling and determination of directionality of the kerf in epifluorescence sharp bone trauma analysis, Int. J. Leg. Med. 128 (2014) 1059– 1066.

J.C. Love / Forensic Science International 299 (2019) 119–127 127

[30] A. Sandras, C. Guilbeau-Frugier, F. Savall, N. Telmon, C. Capuani, Sharp bone trauma diagnosis: a validation study using epifluorescence microscopy, Int. J. Leg. Med. 133 (2) (2018) 521–528.

[31] J.C. Love, S.M. Derrick, J.M. Wiersema, C. Peters, Validation of tool mark analysis of cut costal cartilage, J. Forensic Sci. 57 (2) (2012) 306–311, doi:http:// dx.doi.org/10.1111/j.1556-4029.2011.01977.x.

[32] C. Crowder, C.W. Rainwater, J.S. Fridie, Microscopic analysis of sharp force trauma in bone and cartilage: a validation study, J. Forensic Sci. 58 (5) (2013) 1119–1126, doi:http://dx.doi.org/10.1111/1556-4029.12180.

[33] D.J. Pounder, L. Cormack, E. Broadbent, J. Millar, Class characteristics of serrated knife stabs to cartilage, Am. J. Forensic Med. Pathol. 32 (2) (2011) 157–160.

[34] K. Puentes, J.F.V. Cardoso, Reliability of cut mark analysis in human costal cartilage: the effects of blade penetration angle and intra- and inter-individual differences, Forensic Sci. Int. 231 (2013) 244–248.

[35] D.J. Pounder, L.J. Sim, Virtual casting of stab wounds in cartilage using micro- computed tomography, Am. J. Forensic Med. Pathol. 32 (2011) 97–99.

[36] C. Crowder, C.W. Rainwater, J.S. Fridie, Microscopic Analysis of Sharp Force Trauma in Bone and Cartilage: a Validation Study. National Institute of Justice

Technical Report, National Institute of Justice, Washington, DC, 2011 Report no.2009-DX-BX-K238.

[37] D.W. Morton, A Cut Above: Sharpening the Accuracy of Knife and Saw Mark Analysis, Forensic Magazine, 2006 June/July.

[38] State of Texas vs. Timothy Wayne Shepherd Fourteenth Court of Appeals, (2011) (Accessed on 12/24/2018) https://law.justia.com/cases/texas/four- teenth-court-of-appeals/2011/87759.html.

[39] Ramirez v. State of Florida WL 1628609, 27 Fla. L. Weekly S18 Supreme Court of Florida (Dec. 20, 2001).

[40] Texas Rules of Evidence http://www.txcourts.gov/media/921665/tx-rules-of- evidence.pdf. (Accessed on December 28, 2018).

[41] W. Bonte, Considerations on the identification of notch traces from stabbing injuries, Arch. Kriminol. 149 (1972) 77–96.

[42] S.A. Symes, H.E. Berryman, O.C. Smith, Saw marks in bone: introduction and examination of residual kerf contour, in: K.J. Reichs (Ed.), Forensic Osteology: Advances in the Identification of Human Remains, Springfield, IL, Charles C. Thomas, 1998, pp. 389–409.