Home Page

Feedback

Reprint Request

Introduction

The forensic literature should be consulted in more detail, for interested readers there is a reference list associated with this Current Topic which you might like to consult.

This Current Topic is based upon a more comprehensive publication by the author in Science & Justice.  Click here for a reprint request.  This article was co-authored by Dr. David Sweet.

Types of articles found

MedLine was used to find the articles to locate articles.  In total 1508 articles were found that contained the keyword “Forensic Dentistry”; 120 English language papers within this group were related to bitemarks.  Each of these papers was located and, using ISI Web of Science, a Science Citation Index value determined.  For comparison purposes, it should be noted that within the 1960 – 1999 MedLine database there are 1457 articles related to Forensic DNA, 60 related to Forensic Entomology, and 3538 related to Periodontology.

Due to the relatively small number of papers found, it was possible to identify trends within the entire search.  Approximately 20 papers have been published on the subject each year and the mid-eighties were the most productive period.  The type of publication is an important aspect to consider when evaluating the core research and scientific basis for bitemarks.  The graph below shows the distribution of the categories of papers among those identified in this research.  It is interesting to note that case reports were responsible for 28% of the total literature while empirical research from well-designed experimentation contributed only 15%. 

Bitemark analysis is based on two postulates: a) the dental characteristics of anterior teeth involved in biting are unique amongst individuals, and b) this asserted uniqueness is transferred and recorded in the injury (19).  A distinction must be drawn from the ability of a forensic dentist to identify an individual from their dentition by using radiographs and dental records and the science of bitemark analysis.  Dental identification, as opposed to bitemark identification, utilises the number, shape, type, and placement of dental restorations, root canal therapies, unusual pathoses, root morphology, trabecular bone pattern, and sinus morphology (2).

The debate over the uniqueness of human teeth is probably one of the fiercest in current forensic dental discourse.  Many forensic dentists, appellants, and lawyers have questioned the validity of dental uniqueness determination and demand to know from testifying experts the relative frequency of dental features identified in bitemarks.  An examination of the literature divulges the scientific evidence for this commonly held belief.  

The first article to consider the statistical nature of dental uniqueness was published by MacFarlane and Sutherland in 1974 (20).  The authors began by differentiating between “positive” and “negative” features of the dentition.  A positive feature was described as the presence of a tooth with a certain rotation or other individualising feature.  A negative feature was the absence of a tooth.  This study concentrated on the positive features that occurred on the anterior teeth (canine to canine, maxillary and mandibular).  Patients were selected from an outpatient clinic and in total 200 study casts (maxillary and mandibular) were produced.  The authors only studied the dental casts, not bitemarks that would have been produced by such casts.

The investigators noted the number and shape of each tooth, the presence of any incisal restoration, relationship of teeth to arch form, and tooth rotation (four categories). 

The study did not examine the presence or absence of spacing between teeth.  The assessments of each cast were entirely subjective.  Disappointingly, the authors elected not to publish a table of results.  Rather they presented images of typical casts and calculated, using their data, the frequency of the traits shown.  The authors noted that certain characteristics were not inter-related and thus the products of their incidences could be used to indicate an overall frequency.  However, certain features, such as mesio-palatal rotation of the upper central incisors were inter-related with a significance of p<0.001.  The authors stated that mesio-palatal rotation of the maxillary central incisors should therefore be taken as a single feature.  This demonstrated that the true frequency of such features was almost four times greater than the frequency when the rotations were considered as individual variables.

 In an example, MacFarlane et al concluded that a particular dentition would only be seen in eight people in 100,000 of the population with natural teeth.  The authors concluded that they had not confirmed the individuality of the human anterior teeth, nor had they considered the impact or representation of any of the features examined on a bitemark in human skin.  The highly subjective examination of the casts by multiple examiners and lack of tabulated results make this study weak, especially in light of the increased scientific scrutiny required by recent Court rulings.  However, a large (200) sample was used of a defined population and efforts were made to ensure that this sampling was randomised.

The considerable variation of bitemark presentations on human skin brings the accuracy of skin as a registration material into doubt.  While many studies have examined the accuracy of bitemarks on other substrates, such as cheese (5, 6), apples (7), sandwiches (8, 9), and soap (10), this review is restricted to human skin.  This represents both the most debated area of substrate accuracy and the most commonly bitten material (11).

Skin is a poor registration material (12) since it is highly variable in terms of anatomical location, underlying musculature, or fat, curvature, and looseness or adherence to underlying tissues (13).  Skin is highly visco-elastic, which allows stretching to occur during either the biting process or when evidence is collected.  In 1971, DeVore issued a preliminary report describing studies performed on the variability of bitemarks found on skin (14).  The experiment involved the inking of human skin (living volunteers) using a stamp with two concentrically placed circles with intersecting lines.  

Following the analysis of the photographs it was found that in all cases there was an expansion or shrinkage of the stamp, with a maximum linear expansion of 60% at one location (14).  The design of the stamp permitted the investigators to examine the distortion in both size and direction.  DeVore concluded that, due to the level of distortion found, photographic images of a bitemark in comparative analysis should be used only if the exact position of the body can be replicated.  The placement of a body in such a position is usually impossible, as the exact position of the body during an attack is rarely known. DeVore stated that further research to investigate the effect of postmortem changes on skin distortion were required. 

In 1974, researchers from the Bioengineering Unit of the University of Strathclyde examined the features of the biting process likely to impact upon the appearance of bitemarks on human skin (15).  They described the differing characteristics of skin from a variety of anatomical locations; e.g., Langer’s Lines represent directional differences in the degree of extensibility of skin.  Like DeVore, they emphasised the importance of body location during biting as the directional variations or tension lines will alter with movement.  The report also described distortion that can occur in skin after biting.  The oedematous response of skin to trauma is likely to stiffen the area, thus rendering it more stable.  However, the subsequent resorption of this fluid will cause a large amount of distortion.  They concluded that the changes in bitemark appearance are likely to be greater as the injury grows older.  This was found equally applicable to both living and dead victims.  The article concluded that forensic odontologists where “still ignorant... of the conditions during normal biting... considerable research is required [to address this]”.

Just how unique are these teeth?  In a select population, how many people would have teeth like this?

Analytical techniques

An essential component of the determination of the validity of bitemark analysis is that the techniques used in the physical comparison between suspect dentition and physical injury have been assessed and found valid.  One of the fundamental problems with this task is the wide variety of techniques that have been described in the literature.  Techniques using confocal, reflex and scanning electron microscopes, complex computer systems, typing of oral bacteria, special light sources, fingerprint dusting powder and overlays have all been reported (24 – 28)  It is a widely held belief that while methods that are more esoteric exist, the dominant technique for comparison of exemplars is transparent overlays. 

The lack of direction from the forensic dental organisations, both European and American, complicates this matter.  The American Board of Forensic Odontology (ABFO) has reported advice and guidance on many aspects of bitemarks and yet one of the most pivotal questions, i.e. what is the best comparison technique to use, has not been addressed (29, 30).  Should a Court wish to review the literature to ensure that a testifying expert is using generally accepted techniques they would find the task daunting and ultimately unrevealing. 

Transparent overlays utilise materials found in any dental office.  The vast majority of forensic dentists use techniques that utilise materials that are inexpensive and easily obtainable, hence the popularity of overlays.  There are numerous techniques for the fabrication of transparent overlays.  The only article that has assessed the accuracy of such overlays is that of Sweet and Bowers in 1998 (23).  This paper compared five common techniques of producing transparent overlays.  Of all the techniques, an examination of case reports and experiments reveals that the xerographic and radiographic techniques are the most popular. 

 Sweet and Bowers used 30 randomly selected study casts to examine the accuracy of overlays produced from each of the five techniques concerning tooth rotation and surface area.  The computer-generated overlays were the gold standard.   From these results, it can be seen that the computer technique represents the most accurate fabrication method with respect to representation of rotation and area of the biting edge.  The authors of the paper concluded that the fabrication methods that utilised the subjective process of hand tracing should not be used in favour of techniques that are more objective.  The use of computer-generated techniques was advised over any other method.

(C)2001 Forensic Dentistry Online

Dr. Rawson, a forensic dentist, two dental students, and a statistician wrote arguably the most cited and well-known bitemark paper describing an empirical experiment.  In an attempt to prove finally the uniqueness of the anterior segment of human teeth, Rawson examined 397 bites and applied a statistical probability theory to the results.  The significance of this paper warrants the comprehensive assessment of its validity which can be found in the main article.  A brief version is described here.

Twelve hundred wax bites were obtained from forensic odontologists in various geographic locations in the United States.  Each bite was made on a custom wax wafer 1-mm thick supported by a 1-mm hard cardboard backer.  The subjects were instructed to bite to the maximum depth of 1-mm.  This design removed the variation of incisal penetration found in the twin study.  A calibrated 1-cm scale was also impressed upon the wax. 

The bitemark indentations were filled with zinc powder and then radiographed using a technique designed to minimise any enlargement.  Following the exposure of one side of the wax the zinc was removed and the procedure repeated for the other side.  A study described earlier determined that the radiographic process for overlay production was relatively accurate but it found that hand-traced overlays were less accurate and generally unsuitable for use (23).  

Following the tracing of the biting edges, several elements of tooth position were assessed.  A center point for each tooth was determined and the x and y co-ordinates noted.  The angulation of each of the teeth was measured and all the data were entered into a computer for analysis.  It was determined that the minimum number of positions that a tooth can occupy is 150 and the greatest 239.9.  These figures were determined by multiplying the number of positions of x by y and by the angles observed.  The occurrence of fractions of positions (i.e. 239.9) is a reflection of this multiplication.  Rawson elected to use 150 as the number of possible positions for each tooth as this represented a conservative sample.  Using this premise, the article then stated that the probability of finding two sets of dentition with all six teeth in the same position was 1.4x1013.  With an assumed world population of 4 billion (4x109) Rawson stated that a match at five teeth on a bitemark would be sufficient evidence to positively identify an individual as the biter to the exclusion of all others.

One concern with this use of the product rule to multiply individual probabilities to establish an overall likelihood is that of independence of the variables.  The article assumed that the position of each of the teeth was entirely independent of the position of any others.  However, the independence of these features has not been established by this or any other study.  This has been shown incorrect; e.g., the dependence of mesio-palatal rotation described by MacFarlane (20).  It is likely that every tooth position influences another—intra-quadrant, intra-arch and between opposing arches.  This lack of independence renders Rawson’s certainties of individualisation invalid.  Rawson’s results also showed a possible sampling error, as evidenced by the data sets regarding possible tooth position for each unit.  Intuitively it should be anticipated that the left and right quadrants should represent a mirror image of each other in terms of possible tooth centre positions.  This was not the case.  The upper right lateral incisor was reported to have 239.9 possible locations while the upper left lateral incisor had 161.5 locations. 

 It can be argued that this paper, without the statistical treatment, confirms the anecdotal evidence of almost any practising dentist that the human dentition is unique.  It can be stated that, with an extremely high resolution of measurement, such as in this article, the minutia of the dentition can be described and proven unique.  The current authors would argue that this is the wrong question to ask.  It is the rendition of these asserted unique features on human skin that is the unknown quantity.  Rawson alluded to this point within his article:

 "... [the question is] whether there is a representation of that uniqueness in the mark found on the skin or other inanimate objects".                                                                 

Rawson has proven what his article claims, although perhaps not to the mathematical or statistical certainty expressed.  The article determined that the dentition is unique; however, when this paper is cited, authors often extend this conclusion to incorporate the uniqueness of bitemarks.  The question of bitemark uniqueness remains unanswered.