REPORT

CRIME SCENE EVIDENCE
(1) Original Autopsy Photograph
(A) General discussion.
I have reviewed a series of black and white computer generated prints showing a cutaneous injury on a breast. I will assume the original conventional photograph was taken during an autopsy examination and the images were digitized at a later date. The multiple copies of this conventional print are high resolution and some are of increased magnification. No pixellation is visibly present in any of them. See Figure 1 for general anatomical orientation of the original photograph.


(B) Anatomical Orientation of the Bitemark.

Figure 1. The victim’s head is toward the top of the picture. This does not imply the bitemark is properly oriented with the injuries above the nipple caused by upper teeth.

Figure 2. In fact, the proper bitemark orientation could be reversed.

Discussion about whether Figure 1 or Figure 2 is correct occurs on page 3 under Issue Two.
<b> (C) Physical Findings
Visual inspection of the injury indicates:

1) It is a human bitemark.
2) The bitemark is on a curved surface.

(D) Photographic and Forensic Identification Findings from the Original Autopsy Photograph
</b> Issue One: What is the proper orientation of the bitemark?
This next question is also called: "how do you orient the bitemark" or "where are the big upper front teeth and where are the smaller lower teeth?" Please refer to Figure 3 and the discussion that follows.

A

B

Figure 3. The two possible orientations A and B

(1) Definition of terms:

Tooth width is the longest distance along its biting surface [mesial to distal]. Tooth thickness [lip to tongue; or labial to lingual] is the distance at right angles to the width.
Jaw width is the distance, in the same jaw, from one side to the other. The cuspids (eyeteeth) are the usual landmarks for this measurement.

(2) The fundamental step in bitemark analysis is the determination of which teeth made specific marks. This is based on the following general human dental anatomy or TOOTH CLASS CHARACTERISTICS:

A. Front teeth are seen as the primary biting teeth in bitemarks. There are three incisor types: centrals, laterals and cuspids.

1. Shape differences of the six upper front teeth. The 2 upper central incisors are wide and lateral incisors are narrower. The upper cuspids are cone shaped.
2.Shape differences of the six lower front teeth. The 2 lower centrals and 2 laterals are uniform in width and total 4. The lower cuspids are cone shaped.

B. The upper jaw is wider than the lower jaw.
C. A bitemark showing the upper front teeth and the lower front teeth will show a total of 12 teeth marking in the skin.

(3) The next step in bitemark analysis is the determination of which marks were made from upper teeth and from lower teeth. This is based on the following general factors or BITEMARK CLASS CHARACTERISTICS:

A. The upper 4 front teeth make rectangular marks. The central being wider than the laterals.
B. The upper cuspids make round or ovoid marks.
C. The lower 4 front teeth make rectangular marks that are similar in width.
D. The lower cuspids make round or ovoid marks.
E. Gaps seen between marks that of themselves clearly show specific teeth occur from four possibilities:

1. The suspect has no tooth present.
2. The tooth is shorter due to its normal shape or previous breakage.
3. There was an object (i.e. clothing) that blocked the tooth (sometimes more than one tooth) from contacting the skin.
4. Hypothetical scenarios that talk about tissue movement or biting mechanisms.

F. Areas between known biting teeth that show significantly fainter bruising or three-dimensional indentations are attributed to teeth that did not impact the skin due to some feature present on the tooth. Difference is tissue contours might be another cause. This would be clearly seen in the bitemark photograph. The typical reason is the edge of the tooth is chipped or the tooth is shorter than the adjacent teeth.

Conclusion: Bitemark Orientation from the Autopsy Photograph
Conclusive (positive) bitemark orientation is not possible from the autopsy photograph. It is ambiguous due to its incomplete representation of teeth. More information is needed.

Issue Two: What is the forensic identification value of the bite mark injury?
The autopsy photograph shows a skin injury that varies in terms of edge definition, is incomplete in dental arch width and has few specific tooth characteristics. The typical "U" or "C" shaped arrangement of a full set of upper or lower teeth marks is not seen. The absence of extraordinary dental features present in this injury makes the value of the bite mark "low," or "weak" or "slight." Historically in the United States, this type or category of "low," "weak" or "slight" bite mark injury, when brought to court as definitive evidence of the biter’s identity, is supported by ambiguous and contradictory statements from multiple dental experts. The judicial records of past U.S. cases generally show that primary legal strategy in the face of such dental ambiguity is to personally attack the credentials, honesty and level of experience of the opposing expert. I am sure this case will not be any different due to the minimal identification value of the tooth marks.
This particular bitemark’s clearly observable features and non-features are:
1. There is a laceration that is wider than any one front tooth of Mr. Torgersen (ambiguous information). See Figure 4.

Figure 4.

2. No jaw width determination of upper or lower teeth is possible (absence of information).
3. There is no "C" or "U" shaped curvature that clearly defines upper teeth from lower teeth (absence of information)
4. 4 (four) marks that clearly break the skin. See Figure 5.

Figure 5. Four marks penetrate the skin.

Figure 6 is another view of the autopsy bitemark photograph with alternative groupings called Region One and Region Two.

Figure 6. Red outlines demarcate Region One and Region Two.

Figure 7 is a close-up of Region One that shows a laceration (#1) and one ovoid mark (#2) and a shallow curvature between #1 and #2.

Figure 7. Region One is immediately to the side of the nipple. It contains the laceration (#1) and an ovoid mark (#2). There is a shallow curvature present between #1 and #2.

Figure 8 is a close-up of Region Two that shows a shallow curvature between #1 and #3 and three marks.

Figure 8. Region Two combines the final three ovoid marks. There is a shallow curvature present between #1 and #3.

Conclusion: Positive orientation for the bitemark using the alternative regions shown in the autopsy bitemark photograph is impossible. More information is needed.
In light of this fact, the 1957 cast of the breast has to be analyzed.
(2) 1957 Model of the Breast.
The resin model of the original impression taken from the breast in 1957 proved to be capable of answering the issue of bitemark orientation. See Figure 9 and 10.

Figure 9. 1957 Model of Breast

.

Figure 10. Close-up of Figure 9 showing class characteristics of lower teeth.

Conclusion:
Region One contains marks from upper teeth 13, 12, 11.
Region Two contains marks from lower teeth 31,41,43.
Issue Three: Is this photograph an accurate representation of the bitemark?
The two-dimensional (flat photograph) shows the position of the injury patterns (tooth marks) on the three-dimensional curved surface of the breast. This is called parallax distortion. This is caused by the marks being different distances from the camera lens and not being parallel to the camera lens. There is no means of photo lab or digital correction for this physical situation. It is called Type II Distortion in Digital Analysis of Bite Mark Evidence, which was written by Dr. Raymond Johansen and myself (1). This renders unusable any photograph simultaneously showing the upper and lower areas of the bitemark. Proper means of photographing the original evidence would have been to take separate photographs, each containing a properly placed scale and with specific marks parallel to the camera lens.
I used the life digital image of the bitemark and a life size digital image of the 1957 cast of the breast to determine the physical presence of parallax distortion.

Methods:
Adobe Photoshop 5.0 was used on a PC desktop computer. The digital photographs and computer scans of the all models and photographs were performed either by Dr. Senn or myself. I use an Epson Expression XL scanner capable of 300dpi resolution.

Testing:
1. Region One

A. Digital Image: 1957br~1 (Original Autopsy Digital Image)

Digital Methods:
This image was rotated, made life-size (1:1), and inverted (the color values changed from white to black and black to white). The resultant image then became "1957br~1 life-size invert."
A centroid mark was placed on #13 mark. A mark was placed on a dark spot on the extreme distal area of #12. These locations were chosen as the deepest areas of the marks. A measure line was drawn connecting the two dots. This line was measured at .70 cm. See Figure 11.

Figure 11. Distance Measurements from #12 landmark and #13 centroid on 1957 photo of breast.

B. Digital Image: 1_6 life-size (Digital image of 1957 model of the breast).

The same landmarks were used as in section A. and measured. The measurement between #12 and #13 was .79 cm.

<b>

Figure 12. Distance between teeth #12 landmark and #13 centroid (Region One)

</b> Conclusion: The original autopsy photo of the top of the bite mark is smaller than the digital photos by 13%. The digital image of Region One was taken under controlled circumstances that eliminated Type II distortion and is photographically accurate. The autopsy photograph is inaccurate in Region One.

2. Region Two

A. Digital Image: 1_3 life-size invert (1957 cast of breast showing area below the nipple)

This image clearly shows three defined marks (#43, #41, and #31). A dot was placed in the centroids of #41 and #43. The distance was calculated using the "Measuring Tool." The value was 1.12 cm.

B. Digital Image: 1957br~lifesize.invert (1957 photograph of the breast).

Centroid landmarks were developed on "1957br~lifesize.invert" establishing the distances between teeth centroids of # 41 and #43. The value was 0.82 cm.

Figure 14. 1957 autopsy photo: 0.82 cm is the distance between centroids of #41 and #43.

Conclusion:
The original autopsy photo of Region Two is smaller than the digital photo by 25%. The digital image of Region Two was taken under better circumstances and is photographically accurate. The autopsy photograph is inaccurate in Region Two.
Summary of Findings:





1) The 1957 autopsy photograph of the breast is inaccurate. It is smaller than the 1957 model of the actual breast.
2) The upper arch is 13% smaller and the lower is about 25% smaller. Since the distortion values are different (13% and 25%), this proves the bitemarks in Region One and Region Two are not on the same plane. They are on differing curves of the breast.
3) There is significant Type II distortion of the bite mark in the original autopsy picture. It should not be used for direct analysis with Mr. Torgersen’s teeth.
4) The bitemarks in Region One and Region Two cast of the 1957 breast must be analyzed independently with Mr. Torgersen’s teeth.

Comparison of the Bitemark on the 1957 Model to Mr. Torgersen’s Teeth.

The resin casts of Mr. Torgersen’s teeth models from 1957 were scanned by me and used to create a computer- generated overlay of the biting surfaces of his upper and lower front teeth. I call this overlay a "compound overlay" because it contains the image of the edges of the teeth as well as the biting perimeter of the teeth. This overlay was made life-size to correspond to images of the bitemark that had the same magnification. The bitemark images used were made by Dr. Senn who digitally photographed Region One and Region Two of the 1957 bitemark model . Each image contained an ABFO No. 2 ruler. Each image was evaluated for off-angle camera distortion (TYPE I distortion) and found to be accurate representations.
NOTE: NO ATTEMPT HAS BEEN MADE TO MAKE ANY IMAGES IN THIS REPORT LIFE-SIZE (1:1).

Figure 14. 1957 model of Mr. Torgersen's upper teeth

Figure 16. Compound overlay of Mr. Torgersen's teeth. This image is flipped horizontally onto a life-size image of the 1957 bitemark bitemark. The scale is used to verify dimensional accuracy of the image. The edges of his teeth are moderately worn and show flattened surfaces with incisal grooving. The arch shape is normal with no rotations (crooked teeth).

Figure 17. Compound overlay of Mr. Torgersen's lower teeth. His lower four front teeth are not perfectly aligned as his upper teeth. NOTE: The magnification of this figure is greater than Figure 16.

Comparison Methods:

1. I spatially compared computer prints of each bitemark Region and the overlays. This means I placed the overlay (after flipping horizontally) onto the image of the bitemark.
2. Computer assisted comparison of the same images in section 1. was performed using Photoshop and my desktop computer.

Findings:

1) The bitemark shows teeth #31 and #41 in the exact OPPOSITE physical relationship seen in Mr. Torgersen’s mouth.

Figure 18. Mr Torgersen’s teeth. #41 is in front of #31. Note: this image is flipped horizontally.

Figure 20. The wax bitemark made by Mr. Torgersen's teeth duplicates the features of his teeth #31 and #41.

Figure 21. 1957 model of breast. #31 and #41 are the reverse of Mr. Torgersen's teeth.

2) The bitemark shows a continuous, level surface on the mesial (midline toward the nose) edges of #31 and #41. Mr. Torgersen does not have this relationship. His tooth #31 is curved where it immediately abuts #41.

Figure 22. 1957 model of breast. #31 and #41 are inside the white lines.

Figure 23. Image of models of Torgersen’s lower front teeth. Tooth #31 has lost the corner that is next to tooth #41.

3) Mr. Torgersen has #42 with the same height as the two adjacent teeth.

See Figure 24. The bitemark shows #42 having no contact with the tissue. The adjacent teeth #41 and #43 clearly mark. See Figure 25.

Figure 24. Mr. Torgersen’s 1957 model showing #41, #42 and #43 biting edges.

This is significant because in the original autopsy photograph, there is a faint bruise that shows in the area between #43 and #41. See Figure 26.

Figure 26. Brightness and contrast enhanced image. The circle shows the area of #42 that is seen as a faint bruise in the 1957 autopsy photograph.

The #41 and #43 marks actually break the skin but #42 is NOT present on the 1957 cast of the breast. It is only slightly present in the original autopsy photograph as a bruise. The tissue contour between #42 and #43 is smooth and continuous. No hypothesis nullifying this major feature is anything more than speculation. If there were dynamic tissue distortion in the area of #42 it would render Region Two worthless for any analysis.

Conclusions:
There are numerous physical discrepancies between Mr. Torgersen’s teeth and the bitemark. Both methods 1 and 2 confirmed the findings. These discrepancies cannot be ignored considering the weak forensic identification value of the bitemark. The following section discusses the methods I used and the Guidelines controlling my opinion.

Recent Information on the Accuracy of Hand Drawn Outlines of Teeth

I will assume that there are odontologists in the Torgersencase that will use hand drawn overlays. The Court should know that this method is not the most accurate available to them.
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. This paper compared five common techniques of producing transparent overlays.
Sweet and Bowers used 30 randomly selected study casts to examine the accuracy of overlays produced from each of five techniques concerning tooth rotation and surface area. The study tested the ability of the methods to create accurate AREA and ROTATION duplications of a suspect’s teeth. AREA means the biting edges of the teeth. ROTATION means the angular relationships between teeth in the upper and lower jaw.


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 fabrication methods that utilized the subjective process of hand tracing should not be used in favor of techniques that are more objective. The use of computer-generated techniques was advised over any other method.

The Scientific Limitations of Bite Mark Testimony
The following section describes the major scientific issues involving bite mark analysis. .
The Accuracy of Skin as a Substrate for Bitemarks
The threshold variable in bitemark analysis is the fact that, in cases of physical assault having skin injuries, the anatomy and physiology of the skin, and the position the victim affects the detail and shape of the bitemark. There is one article from the early 1970's from Britain by Devore showing how the shape of the test bite (actually it was an inked circle) on a bicep varied whether the arm was flexed or pronated. No significant tests have been published on this subject since the 1971 paper. The Torgersen case involves a breast which highly variable in shape due to position and mechanical forces.
What is significant in casework analysis is the need to experimentally control or establish the amount of positional variation in an actual bitemark case. This is not possible in the Torgersen case. It is, however a favorite topic of speculation by odontologists. In the face of the impossibility to duplicate the actions that occurred during the actual biting, some dental experts develop hypotheses to explain features in the bitemark injury that exclude or weaken the case against the defendant. They want to excuse physical findings in the bitemark injury that conflict with their ultimate opinion. The rebuttal to this forensic advocacy is to ask the question, " what are other possible explanations are there for this feature" and "why do you discount them?"

What also is an issue is the amount of dental detail present in the bitemark. Skin is a poor impression material. The injury may be a combination of reddened bruises combined with faint and irregularly shaped bruises. The presence of vague and ambiguous skin injuries should create in the reasonable person a question as to the limitations to the evidence.
Bruising and Other Considerations
Recognition of the fact that bruising is actually subcutaneous bleeding demands that the investigator not assume that the reddened areas that appear to be teeth are an accurate representation of individual teeth. The presence of three-dimensional indentations in skin, as seen in Torgersen, will give rise to whether the indentations are accurate or just an artifact (phantom). I anticipate the choice of which is correct will vary among the dental experts in this case and probably fall one way for the State and the other way for Mr. Torgersen.
My forensic training tells me the actual physical features are the only basis for an opinion. Speculations and explanations about why one feature (e.g. an indentation or a faint mark) should be interpreted as something it is not (e.g. in order to support an opinion that the tooth really is longer, or in a different position) smacks of fervent soothsaying.

Arguments for the Individuality of Human Teeth

The foundation of dental identification and bite mark analysis is that the total arrangement of a person's dentition creates a dental "profile" of sorts. An attempt to prove this was in 1982 when a small study of five identical twins was done at UCLA. This investigation had nothing to do with dental restoration shape and position that are commonly used in identification of missing and unknown individuals. It was an attempt to test the differentiation of tooth shape and position within this small sample. The test involved pressing dental stone tooth models into a plaster of Paris substrate. There are experimental problems in this study involving the degree of depth necessary in a test bite mark to separate one twin from another twin. In bite mark cases having only two-dimensional discoloration (i.e bruising), the conclusion of this 1982 study is irrelevant. In addition, the authors published an opinion that conclusively establishes the uniqueness of the human dentition in their final statement, "(it is) clear that in terms of occlusal arch form and individual tooth positions, even so-called identical twins are in fact not dentally identical." This is a seriously over-stated conclusion that has been perpetuated in the odontology literature.

Figure 27. Outlines of the same set of teeth. The different perimeter shapes depend on how far the teeth are pressed into the test substrate.
The determination of uniqueness of the human dentition has a history in bite mark analysis that covers over four decades. The demonstration of uniqueness is a blend of art and opinion. The probability of more than just one person producing a similar or identical bite mark in a specific case is the challenge in every bite mark case and some identification cases. The common leap of dental examiners, based on their "knowledge, experience, and training," in testifying the low or insignificant probability of a false positive identification ignores common sense. Uniqueness will never be proven through statistics or studies like the one described above. Rather, the incidence of "false positives" or the "likelihood of a another set of matching teeth" is the direction to take. The forensic DNA community NEVER uses the word "unique" in describing the genetic profiles generated from both a suspect and crime scene evidence. They do state the chance of another’s profile being the same in statistical terms based on serious, prolonged and multi-reviewed studies.
Legal commentaries from the 70's and 80's attacked bite mark analysis as a "new and unfounded science of identification" and that the "uniqueness of the human dentition hasn't been established." The odontology response then and now is to hang its hat on a minima of journal articles, which wobble in their attempts to prove the assumption without much scientific rigor.
A frequently used statement from experienced bite mark experts is "the controversy seems to hinge on how closely we look at the teeth and teeth marks." This describes the odontologist's rule of thumb protection against false positive bite mark identifications. The weight given to a conclusion is based on the number of characteristics seen in the injury. Probability of a positive match is often the result of how many tooth marks are seen in the injury, not in the uniqueness value (indeterminable, see supra) of each individual characteristic of either the defendant or the bite mark injury. Proof of uniqueness of specific dental characteristics, however, is unavailable in the scientific literature.
The heavy use of probability theory is seen in the seminal bite mark articles of the last four decades. Their implication is that so many "points" of similarity or concordance between the bite mark(s) and a suspect's teeth assumes the fact that the dental features seen in the bite mark are unique to one person. .
There is vague reference to the literature regarding what number of points of concurrence satisfies scientific standards for a bite mark identification. Keiser-Neilsen was never talking about bite marks when he introduced the use of the "product-rule " for dentists in 1960. It was purely a theoretical use of simple probability theory assuming that each dental feature was independent of the next and the sum frequency of occurrence for a "profile" could be used to establish the frequency of all the features occurring at once. He was actually talking about missing and filled teeth, not bite marks. Rawson et al's 1984 paper on the "Statistical Evidence for the Individuality of the Human Dentition" was a study utilizing teeth impressions in wax, which were then hand, traced and computer analyzed. Bite marks in skin were not the target of the 397 sets of wax teeth marks he choose to investigate out of a larger cohort of over 1200. This study is irrelevant regarding bite marks, a fact he alludes to within the body of the paper. The common use of this paper is to justify dental examiners’ positive identification statements regarding one person (the defendant) and a bite mark in skin.
In order to achieve an accurate estimate of the frequency of the occurrence for a complex event, each feature (in this case the shape and position of teeth) has to be independent of each other. This is the basis for DNA population studies regarding genomic frequencies of various polymorphic loci. Each allele present at these loci (or tooth in this case) has to be proven to occur independently before being included in the calculations of the "product rule." Suffice it to say that Rawson et al. study did not achieve or even attempt to prove this. The scatter data available in his paper shows that the distribution of tooth positions was not proportional. Certain anterior teeth positions occurred less than others. The argument regarding the lack of independence of dental values (shape and position) was reported as a finding in a much earlier British paper that has been totally ignored since it was written. In essence, the use of uniqueness of the human dentition and probability theory regarding bite mark analysis is flawed.



Not all bite marks have the level of forensic value necessary to identify just one individual. Toolmark terminology was adopted early on by odontologists regarding the types of dental features seen in bite marks and the human dentition. A characteristic within a bite mark or in a person's dentition is a distinguishing feature, trait, or pattern. A class characteristic reflects a feature of generic value to a large population. Each human tooth has shape and position features common to the human species. The diagnosis of a human bite mark is dependent on the presence of these class characteristics being present in the injury.
Individual dental characteristics are reported to be features that are unique to an individual variation within a defined group. The presence of worn, fractured or restored teeth is valued as unique features. If a bite mark possesses the reflection of such a feature(s), the degree of confidence in a match increases. The odontological literature is silent regarding the frequency of these traits. It is actually counter-intuitive to assume enamel chips, fractures, and dental restorations are ALWAYS inherently unique. The shape of human teeth is quite constant in nature and their changes over time are based on common events. The chance occurrence of more than one person having a crooked front tooth is quite large. That is why orthodontists have such large practices.
In a law text published in 1997, Rawson stated that a concordance between the biter and the injury is assured by the presence of the following:

" Missing teeth, pattern of rotation, angulation or position of each tooth…
if the frequency of certain positions is known for the general population ."

This has not been accomplished by the forensic literature.
Rawson continues in this vein by adding up the hypothetical frequencies of occurrence of more rare or in his terms "uncharacteristic features." He then arrives at the use of "product rule" probability per Keiser-Neilsen, based on the assumption that these features of higher forensic value are independent of one another. This is conceptually flawed and has never been proven to be true.
Incumbent to this process is the total assumption that human uniqueness determination of dental characteristics is possible. As stated supra, this author is of the mind that the real test is how often someone else's teeth could match the pattern seen in a bite mark.
Not all bitemarks have the level of forensic value necessary to identify just one individual.
In every bitemark case, the arguments exist regarding the forensic weight (value) that these characteristics possess. There is available orthodontic data on general dental characteristics such as racial variations in skeletal anatomy, jaw width, front tooth position, twisting of teeth (rotation), etc. They are seldom used in forensic cases and there are scant studies from the forensic point of view on this topic. There is no experimental foundation to an opinion that states only one person made a bitemark. The American Board of Forensic Odontology Standards and Guidelines for Bitemark Analysis state the following must occur in a case having a "positive identification."

Definition: Reasonable Medical Certainty:
· Highest order of certainty that suspect made the bite.
· The investigator is confident that the suspect made the mark.
· Perpetrator is identified for all practical and reasonable purposes by the bitemark.

· Any expert with similar training and experience, evaluating the same evidence should come to the same conclusion of certainty.

· Any other opinion would be unreasonable.

This definition should be considered as the evidence standard or prima facie case for the State. It must be supported by clear evidence without reasonable doubt. Any significant divergence of dental opinion might raise the question in the Court’s mind to "who should we believe?" The balance of any debate in a criminal case should favor the accused until proof is submitted that erases any reasonable doubt. This case is filled with reasonable doubt both from the inaccuracy of the original autopsy photograph and the physical findings derived from comparisons using the 1957 cast of the breast and 1957 model of Mr. Torgersen’s teeth.

__________________