- Email: [email protected]
Loss and replacement of small particles on the contact surfaces of footwear during successive exposures
Forensic Science International 269 (2016) 78–88
Contents lists available at ScienceDirect
Forensic Science International journal homepage: www.elsevier.com/locate/forsciint
Loss and replacement of small particles on the contact surfaces of footwear during successive exposures David A. Stoney* , Andrew M. Bowen1, Paul L. Stoney Stoney Forensic, Inc., 14101-G Willard Road, Chantilly, VA 20151-2934, USA
A R T I C L E I N F O
Article history: Received 20 April 2016 Received in revised form 4 November 2016 Accepted 8 November 2016 Available online 16 November 2016 Keywords: Trace evidence Footwear Very small particles Soil minerals Transfer Persistence
A B S T R A C T
On the contact surfaces of footwear loosely, moderately and strongly held particle fractions were separated and analyzed in an effort to detect different particle signals. Three environmental exposure sites were chosen to have different, characteristic particle types (soil minerals). Shoes of two types (work boots and tennis shoes) were tested, accumulating particles by walking 250 m in each environment. Some shoes were exposed to only one environment; others were exposed to all three, in one of six different sequences. Sampling methods were developed to separate particles from the contact surface of the shoe based on how tightly they were held to the sole. Loosely held particles were removed by walking on paper, moderately held particles were removed by electrostatic lifting, and the most tightly held particles were removed by moist swabbing. The resulting numbers and types of particles were determined using forensic microscopy. Particle profiles from the different fractions were compared to test the ability to objectively distinguish the order of exposure to the three environments. Without exception, the samples resulting from differential sampling are dominated by the third site in the sequential footwear exposures. No noticeable differences are seen among the differential samplings of the loosely, moderately and strongly held particles: the same overwhelming presence of the third site is seen. It is clear from these results (1) that the third (final) exposure results in the nearly complete removal of any particles from prior exposures, and (2) that under the experimental conditions loosely, moderately and strongly held particles are affected similarly, without any detectable enrichment of the earlier exposures among the more tightly held particles. These findings have significant implications for casework, demonstrating that particles on the contact surfaces of footwear are rapidly lost and replaced. ã 2016 Elsevier Ireland Ltd. All rights reserved.
1. Introduction 1.1. Statement of the problem Very small particles are ubiquitous in our environment. These “VSP” are particle dusts which, as noted by Gross [1], are our “environment or surroundings in miniature,” and as noted by Locard [2] “may be formed of all the debris and all kinds of bodies . . . all the substances, organic or inorganic, existing on the earth.” Everywhere people walk, VSP transfer to and from their
* Corresponding author. E-mail addresses: [email protected] (D.A. Stoney), [email protected] (A.M. Bowen), [email protected] (P.L. Stoney). 1 Present address: U.S. Postal Inspection Service, 22433 Randolph Drive, Dulles, VA 20104, USA. http://dx.doi.org/10.1016/j.forsciint.2016.11.015 0379-0738/ã 2016 Elsevier Ireland Ltd. All rights reserved.
footwear. The mere presence at a crime scene requires this contact and transfer, and the particles are known to persistent for long periods of time [3,4]. Even though criminals necessarily track dusts to and from every crime scene, dust particles on a suspect’s shoes are very seldom used as evidence linking the accused to the crime. There is an extraordinary, untapped potential to exploit VSP found on footwear and in footwear impressions. At the same time, there are significant challenges to unlocking this potential. Most fundamentally, VSP on footwear evidence are invariably a mixture of materials that can originate before, during, or after any event or period of forensic interest [3,4]. Their usefulness depends on our ability to separate a reliable, relevant evidentiary “signal” from background noise (or signals from other exposures). As an additional practical challenge, the VSP mixture is composed of many different particle types, which must be collected and analyzed efficiently.
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
Methods have been developed to efficiently analyze VSP using either an iterative forensic approach [5] or a particle profiling approach [6–8]. An iterative forensic approach begins with a multidisciplinary screening of particle types. This is followed by assessment of possible contributions to case resolution that could result from specialist examinations. Choices of which particles to analyze are made based on this assessment. The results of these analyses are then used to re-assess the possible contributions of additional specialist examinations of additional particle types. A particle profiling approach proceeds through the simultaneous characterization of many particles. This results in a profile representing the population of particles in a specimen. These profiles allow the application of computational methods showing the potential to measure strengths of associations between VSP specimens and to determine linkages among items of evidence based on their adhering VSP [7,8]. The application of these methods to VSP on (for example) shipping containers, clothing, and improvised explosive devices (IEDs) has consistently involved comparing VSP at different locations or different “layers” as one opens an item. The outer layer typically has VSP from the most recent exposures. The innermost layer has VSP from earlier exposures. Intermediate layers have exposures whose relative timing depends on how and when the item was handled, assembled or opened. On footwear the problem is more complex, as there are not physically separated layers (as there are in layers of packaging, or an assembled device). However, research focused on the persistence of trace evidence generally [9–13], and on footwear specifically [4,14–16], strongly supports the hypothesis that, after transfer to an item, some particles are tightly held (and retained longer), while others are loosely held (and more rapidly lost). Morgan et al. [17] have specifically found that for sediments on footwear there is a “trend of two/three stage decay . . . , with subsequently less rapid loss . . . , followed by a period of much lower decay.” Importantly, we observe that this explanation implies that particles from earlier exposures will be more concentrated among the particles that are more tightly held. From this strongly supported supposition, we hypothesized that, if we use differential sampling of footwear (which separates loosely held, moderately held, and strongly held particle fractions) we will recover physically separated or enriched particle fractions that originated from different exposures. This project explored the use of differential sampling of VSP from the soles of footwear as a potential method for the separation for these signals. 1.2. Background and context 1.2.1. Footwear evidence generally Trace evidence examiners encounter footwear as part of clothing examinations. Trace evidence commonly found on footwear includes the major types of fragmentary material traces: glass, paint and fibers [18,19], as well as accumulations of soil. Trace evidence within shoe impressions is only rarely utilized as part of the comparison, although the potential to compare this form of trace evidence with footwear is well-recognized [20]. As methods in forensic geoscience have developed [3,21–25], such cases are being reported [26]. 1.2.2. Soil and dust on footwear Accumulations of soil on footwear or other items of evidence (such as digging tools and vehicles) have long been exploited for comparisons with reference samples of possible origin [27]. The long-standing focus has been on fairly large accumulations of soil that can reasonably be expected to be minimally mixed, or that are clearly layered, so as to allow physical separation of discrete samples. Only then can comparisons be reliably made using bulk
79
properties of soil (such as color, particle size distributions and elemental composition). In cases where significant mixture has occurred, analysis of soil evidence is frequently stopped short. This is because preliminary analyses indicate disparities in the bulk properties (e.g. color) that are typically used to screen for comparable specimens. Restriction of analyses to unmixed specimens severely restricts the numbers of applicable cases. The work of Morgan, Bull and co-workers [3,4,28–31] has addressed this limitation, setting forth a conceptual framework for forensic geoscience [3]. This framework includes specific emphasis on analytical methods that can recognize mixtures and that are applicable when mixtures are present. They describe these methods as “visual techniques” and have exploited quartz grain surface analysis for this purpose. Quartz occurs very widely in sediments and the physical appearance of quartz grains depends on fundamental geological mechanisms relating to their origin and transportation [32]. When two different sources of soil are mixed, expert quartz grain surface analysis can, with reasonable probability, detect this mixture. Comparisons of the different types of quartz grain surfaces can be made even though the sources are mixed. This approach need not be limited to a single mineral type, or specifically to mineral particles, and it need not depend on the presence of one particle type (e.g. quartz) in each of the mixed sources [33]. What is essential is that recognizable varieties of minerals (or other particles) be exploited efficiently. Visual microscopical techniques do this: different soil or dust samples will have different suites of VSP. The presence and variety of VSP is a character that is recognizable within a mixture and meets the fundamental requirements for “visual techniques” [3,4]. 1.2.3. Studies of particle transfer and persistence on footwear Morgan et al.’s approach to recognition and analysis of mixed soil samples has continued with applied research directed at understanding mechanisms of transfer, persistence and mixing of particles deposited on footwear [4,14]. Experiments have been conducted using test substances (Plasticine) as well as using specific particle types (pollen or quartz grains). Specific caserelated research has also been conducted [30]. These studies have demonstrated that (1) particles persist for a long period of time on footwear, (2) that mixing of particles from successive exposures routinely occurs on the soles of footwear, and (3) that following exposure, some particles are loosely held (and more rapidly lost), while others are tightly held (and retained longer). 1.2.4. Alternative sampling methods as opposed to differential sampling Staged, alternative sampling methods are often employed in trace evidence analysis [34–37]. One purpose is to employ an initial method (such as picking individual fibers or paint chips) to collect loosely held traces as they are recognized. These traces might otherwise be lost or redistributed as the examination proceeds. Another purpose is to preserve and document the location from which trace evidence was recovered (as in the regional taping of clothing in the recovery of fibers). Again, different methods may be used for alternative particle types (such as taping for fibers, followed by vacuuming to recover fine particles, or washing to recover pollen). However, there has not been a protocol for differential sampling and recovery of trace evidence with the express intention to fractionate loosely and tightly held particles, so that these populations can be compared and contrasted. 2. Materials and methods This project was designed to test the separation of particle signals on the contact surfaces of footwear by applying a series of
80
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
successively more aggressive sampling steps and contrasting the resulting types and quantities of particles. Project objectives were (1) to conduct suitable environmental exposures, (2) to develop differential sampling methods, and (3) to test the ability to separate small particle signals of interest. Three environmental exposure sites were chosen to have different, characteristic particle types (soil minerals). Shoes of two types (work boots and tennis shoes) were tested, accumulating particles by walking 250 m in each environment. Some shoes were exposed to only one environment; others were exposed to all three, in one of six different sequences. Sampling methods were developed to separate particles from the contact surface of the shoe based on how tightly they were held to the sole. Loosely held particles were removed by walking on paper, moderately held particles were removed by electrostatic lifting, and the most tightly held particles were removed by moist swabbing. The resulting numbers and types of particles were determined using forensic microscopy. Particle profiles from the different fractions were compared to test the ability to objectively distinguish the order of exposure to the three environments. 2.1. Site selection A range of candidate sites were selected in the state of Virginia, USA based on recorded differences in their surface geology, differences in the watershed (indicating different sources for sediments) and differences related to human activities (in the form of trail or roadside modifications). Candidate sites were also required to be conveniently assessable from roads and to have an apparently uniform path of at least 25 m (allowing the required 250 m walking exposures to be completed in 5 round trips). From a total of 37 sites originally evaluated, three were selected based on their distinguishability and ease of access. These were: (1) Piney River (PR), (2) Appalachian Trail (AT) and (3) Luck Stone Quarry (LQ). Site PR is an improved hiking trail in a wooded site along the Virginia Blue Ridge Railway Trail (Lat. 37.7078, Long. 79.0220). Site AT is an unimproved minor trail just off of the main Appalachian Trail leading along the Tye River (Lat. 37.8384, Long. 79.0220). Site LQ is along the edges of a public access road outside of the Luck Stone Quarry in Fredericksburg, Virginia (Lat. 38.2128, Long. 77.5488). 2.2. Footwear exposures Two types of footwear were used: athletic shoes with flexible rubber soles (Kirkland SignatureTM Men’s Athletic Shoes, Fig. 1, left) and work boots with hard rubber soles (Grabbers Black Steel Toe EH Non-Slip Work Boots, Fig. 1, right). Exposures to test sites were
made by walking a distance of 250 m along a route, achieved by ten transects of 25 m (five round trips). Eighteen pairs of each footwear type were exposed: 6 pairs for single-environment exposures (two duplicate pairs for each of the three test sites) and 12 pairs for sequential exposures to all three environments (2 duplicate pairs for each of the six alternative sequences). Surfaces were raked to remove large plant matter prior to exposures. The surfaces were dry and dusty. Exposures were conducted over a five day rain-free period during which the surfaces remained dry. Following each exposure the footwear was gently re-packaged in its original box, between folds of butcher paper. 2.3. Differential sampling In this study only the contact surfaces of the footwear soles were sampled (those surfaces coming in direct contact with a horizontal surface while walking). The arch area and recessed areas within the sole pattern were not sampled. The more loosely adhering particles were removed by walking: 18 firm, smooth walking steps on butcher paper (92 kg individual, US shoe size 10.5). Twelve steps were found to be sufficient to remove the most loosely held particles, as further steps recovered no perceptible additional particles. Loose particles were recovered from the paper by moist swabbing with pre-filtered 3% ethanol. Moderately adhering particles were removed using an electrostatic lifter (Sirche Electrostatic Dust Print Lifter Kit ESP900). Electrostatic lifting was conducted using a reverse procedure (foil side of lifting film down) and employing a piece of foil taped to the floor as a conductor [38]. With the full voltage setting, 6 smooth steps (92 kg individual, US shoe size 10.5) were found to be sufficient to remove the moderately held particles, as further steps recovered no perceptible additional particles. Particles were collected from the electrostatic film by moist swabbing with pre-filtered 3% ethanol. Direct moist swabbing of the contact surfaces of the footwear soles was conducted to remove and collect the most tightly held particles. Each of the three sampling methods resulted in particle suspensions in 3% ethanol within 1.5 mL microcentrifuge tubes. 2.4. Specimen processing Specimens were washed to remove clay and silt-sized particles. Distilled water was added to the particle suspensions to a height of 3 cm in the 1.5 mL microcentrifuge tubes, followed by mixing using a vortex mixer (Thermolyne Type 16700 Mixer) with occasional brief sonication (Fisher Scientific FS6 Ultrasonic Cleaner). After mixing, samples were allowed to settle in the water for approximately 9 s (3 s per cm of water height) allowing all
Fig. 1. The two types of footwear used in this study: athletic shoes with flexible rubber soles (left) and work boots with hard rubber soles (right).
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
particles with a density greater than 2.61 and diameter greater than 62.5 mm to settle [39]. After settling, the upper 2 cm of liquid (well above the settled particles) was carefully removed using a pipette, taking with it a portion of the clay and silt-sized particles. This process was repeated until the liquid above the settled particles was clear (usually from six to eight times). This resulted in a specimen of sand-sized particles, washed free of silt and clay. The size cut-off of 62.5 mm was chosen because that is the lower limit of the sand-sized fraction, as defined by the Wentworth scale commonly used by sedimentary geologists [40]. The density value of 2.61 was selected because it is close to the density of quartz. Due to its abundance in igneous, sedimentary and metamorphic rocks, quartz is the most common detrital mineral in sediments on the earth’s surface [41]. Quartz is also abundant in the three sites selected for this study. The sand-sized fraction was wet-sieved, sub-dividing it into portions greater 180 mm (the coarse sand fraction) and that less than 180 mm (the fine sand fraction). Wet sieving was performed using distilled water and an Endecotts, Ltd. brand stainless steel sieve with 180 mm openings. The fraction passing the sieve was washed into a glass petri dish and recovered by pipette. The coarse sand fraction was not examined further. This size selection was performed for convenience, as particles larger than about 180 mm are more difficult to mount under a coverslip and to examine at high magnification using light microscopy. Where practical, the entire fine sand fraction was transferred onto a microscope slide by direct pipetting and allowing it to air dry. Where use of the entire fraction was impractical, a representative subsample was prepared using the method of McVicar and Graves [42]. The air-dried samples were mounted under a coverslip using Cargille mounting medium (Series A) 1.540nD, 25 C. 2.5. Polarized light microscopy Polarized light microscopy was performed using a Leitz DMRP polarized light microscope and a Leica DMLP polarized light microscope, at 200 times magnification. Optical identification and characterization of minerals were based on comparisons to known samples and reference data [43,44]. Variations in morphological and optical features were determined as described by Bowen [33]. Categorization of some mineral grains, as viewed in single mounts, required (as a practical matter under examination conditions) a subjective judgement with respect to placing certain grains into a category. The number of grains requiring subjective placement into a category was small (less than a few percent) for the samples in this study. Quantitative determinations were made by point counting of particles using the ribbon method [45]. 3. Results 3.1. Raw point count data and data reduction Point count data for the 108 program specimens included 34 mineral varieties. These were consolidated into 22 primary mineral types. Thirteen of these occurred above 2% in one or more of the single-site exposures. Data were grouped into these 13 categories, with a 14th category of “other.” Table 1 shows these mineral categories and an example of the consolidation. 3.2. Retention of particles on athletic shoe vs. work boot soles Data relating to the effect of the alternative footwear types are illustrated in Table 2 which gives the overall mean fractions of mineral categories for single site exposures. When comparing the
81
results for athletic shoes and work boots exposed to the same sites, the recovered mineral fractions are very close and there is no evidence that the alternative sole types result in different accumulations of the different mineral types. Pearson correlations of 0.995, 0.989 and 0.998 were observed for the three sites, and paired two-sample t-tests (hypothesized difference of means = 0) resulted in P(t t) of nearly 1. Table 2 also shows that the three sites are clearly distinguished by their mineral classification proportions. 3.3. Differential sampling of sequential site exposures An example of results from differential sampling following sequential site exposures is given in Table 3 for work boots exposed to the test sites in the sequence AT, PR, LQ. Data are given for grain counts and mineral classification fractions for each of the three differential samplings (walking, electrostatic lifting, and swabbing) for each of the two duplicate sequential exposures (walking 1, walking 2, etc.). The proportions of mineral classifications for the three differential samplings are highly similar to one another, with each corresponding to the proportions seen in site LQ: the last of the sites to which the boots were exposed. This observation was consistent across each of the six exposure sequences, for both footwear types. For each of the 72 mineral classification profiles (2 footwear types 6 exposure sequences 3 sample differential samplings 2 replicates) a measure of distance from each of the test sites was calculated using the Bray–Curtis Dissimilarity (BCD) measure (Eq. (1)) [46]. Xn1 jyi;k yj;k j ð1Þ dði; jÞ ¼ Xk¼0 n1 jy þ yj;k j k¼0 i;k Table 4 shows an example of the calculation of the BCD for the electrostatic lifter sampling from one of the work boots exposed to the site sequence AT, PR, LQ. A value of %total BCD for the specimen (the relative BCD) is also calculated as a means to readily contrast the distances from each of the sites. Tables 5 and 6 give the BCD measures for each of the athletic shoe and work boot specimens. Note that for the sequential exposures, without exception, the BCDs are lowest for the third (final) exposure site. Relative BCDs are conveniently visualized and contrasted using ternary diagrams following the method of Graham and Midgely [47]. Fig. 2 explains the basic structure of these diagrams and Fig. 3 gives an example showing a plot of the relative BCDs for the athletic shoe single-site exposures. These data are also included on Fig. 4, which plots the relative BCDs for each of the athletic shoe specimens shown in Table 5. Fig. 5 shows the corresponding diagram for the work boot specimens shown in Table 6. As further explained in the text accompanying Figs. 2 and 3, the three sides of the triangle in Figs. 4 and 5 correspond to the overall fractions of mineral categories for single-site exposures (the mean percentages of the mineral categories given in Table 2). The plotted points represent the relative BCDs for each of the differential samples recovered from the footwear soles, from each of the test sites. Points close to only one of the sites are close the edge corresponding to that site. Points close to more than one site would appear in the central portion of the chart. The legend shows black circles corresponding to single-site exposures and separate markers for each of the six sequences of test site exposures. One point is plotted for each of the samples recovered from footwear, with six points for each exposure sequence (from each of two replicates and the three differential samplings). In Fig. 4 (for athletic shoes) all of the samples, whether representing loosely, moderately or tightly held particles, appear near the edge corresponding to the last of the sites to which the
82
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
Table 1 Example of raw data, consolidation of mineral varieties and classification into the 14 most abundant types (athletic shoes with sequential exposure AT, PR, LQ, replicate 1). Point counts with varieties
Point counts without varieties
Grouped into 14 categories
Grain type
Count
Grain type
Count
Grain type
Count
Quartz — clear Quartz — with inclusions Quartz — with bubbles Quartz — with iron coatings Hornblende Alkali feldspar — fresh Alkali feldspar — weathered Alkali feldspar — inclusions Alkali feldspar — microcline Alkali feldspar — Fe coating Plagioclase Biotite — brown Biotite — green Biotite — yellow Biotite — orange Highly altereda Epidote Titanite Lithic fragments Lithic fragments — feldspars Lithic fragments — other Opaques Iron oxides Pyroxene Amphibole — colorless Muscovite Apatite Carbonates Zircon High n clear/prismb High n polycrystallinec Garnet Yellow striatedd Rutile
125 8 10 4 71 19 6 1 2 2 66 7 3 0 0 12 14 4 0 0 2 6 0 0 0 1 1 4 0 0 0 0 0 0 368
Quartz Hornblende Alkali feldspar Plagioclase Biotite Highly altereda Epidote Titanite Lithic fragments Opaques Iron oxides Pyroxene Amphibole colorless Muscovite Apatite Carbonates Zircon High n clear/prismb High n polycrystallinec Garnet Yellow striatedd Rutile
147 71 30 66 10 12 14 4 2 6 0 0 0 1 1 4 0 0 0 0 0 0 368
Alkali feldspar Highly altereda Biotite Epidote High n Hornblende Iron oxides Lithic fragments Muscovite Opaques Plagioclase Quartz Titanite Other
30 12 10 14 0 71 0 2 1 6 66 147 4 5 368
a
Grains too highly altered to clearly observe optical properties, allowing no classification into another category. Colorless grains with low to moderate birefringence, high refractive indices (n), and without characteristic features placing them in another category. Grains having high refractive indices (n) and composed of multiple smaller crystals appearing to be a single mineral type, and without characteristic features placing them in another category. d Yellow polycrystalline grains exhibiting cleavage striations. b c
Table 2 Overall fractions of mineral categories for single-site exposures contrasting results for work boots and athletic shoes.
shoes were exposed. The plotted points mingle with those from the shoes exposed only to this site. This means that the samples are uniformly dominated by the third site in the sequential footwear exposures. Likewise, in Fig. 5 (for work boots) all of the samples are closest to the last of the sites to which the boots were exposed. 4. Discussion and conclusions This section proceeds with a discussion of the findings, followed by further discussion of their limitations and implications. 4.1. Findings Without exception, the samples resulting from differential sampling are dominated by the third site in the sequential footwear exposures. No noticeable differences are seen among the differential samplings of the loosely, moderately and strongly held particles: the same overwhelming presence of the third site is seen. It is clear from these results that the third (final) exposure results in the nearly complete removal of any particles that were transferred to the contact surfaces of the shoe from the first and second exposures. This occurs regardless of the exposure sequence and regardless of which specific site was used for the third exposure. It is also clear that under the experimental conditions loosely, moderately and strongly held particles are affected similarly,
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
83
Table 3 Mineral classification counts, fractions and comparisons for boots exposed to test sites in the sequence AT, PR, LQ.
without any detectable enrichment of the earlier exposures among the more tightly held particles. The hypothesis guiding this work was that distinguishable fractions would result from the differential sampling of the contact surface of footwear. That is, that by separating loosely held, moderately held, and strongly held particle fractions we would recover enriched particle fractions originating from different exposures. This hypothesis is rejected. Under the experimental conditions the contact surface of footwear was found to be overwhelmingly dominated by the most recent exposure: a walk of 250 m, on a dry soil surface, results in the virtually complete
removal and replacement of particles adhering to the contact surfaces from prior, similar exposures. Given that Morgan et al. [14] have shown in comparable studies that a generalized sampling of footwear soles (from both contact and recessed areas) shows the retention of particles from earlier contacts, the clear implication of the present research is that, although particles on the contact surfaces of footwear are removed and replaced, those that are present on the more recessed areas of the sole are not. This reasoning parallels results shown by Roux et al. [15] for the retention of automobile carpet fibers on footwear soles. In their
Table 4 Example of Bray–Curtis distance calculation using electrostatic lift sampling from boot #1 exposed to the site sequence AT, PR, LQ. Raw count Comparison to site AT
Alkali feldspar Alterite Biotite Epidote High index Hornblende Iron oxides Lithic fragments Muscovite Opaques Plagioclase Quartz Titanite Other Total
Comparison to site LQ
Comparison to site PR
Mean AT Observed Expected |Obs Exp|
Mean LQ Observed Expected | Obs Exp|
Mean PR Observed Expected |Obs Exp|
21 8 40 7 0 85 0 3
0.556 0.065 0.003 0.019 0.007 0.005 0.020 0.008
7 9 30 140 4 4 358
0.002 7 0.064 9 0.013 30 0.217 140 0.001 4 0.021 4 1.000 358 BCD = 0.616 %Total BCD = 0.447
21 8 40 7 0 85 0 3
199.1 23.4 0.9 6.8 2.5 1.8 7.1 2.7
178.1 15.4 39.1 0.2 2.5 83.2 7.1 0.3
0.085 0.025 0.056 0.031 0.002 0.219 0.001 0.016
21 8 40 7 0 85 0 3
0.5 23.0 4.6 77.8 0.2 7.6 358.0
6.5 14.0 25.4 62.2 3.8 3.6 441.4
0.006 7 0.015 9 0.068 30 0.452 140 0.010 4 0.012 4 1.000 358 BCD = 0.114 %Total BCD = 0.082
30.4 8.9 20.1 11.0 0.9 78.5 0.4 5.8
9.4 0.9 19.9 4.0 0.9 6.5 0.4 2.8
0.115 0.027 0.050 0.139 0.052 0.019 0.004 0.364
21 8 40 7 0 85 0 3
2.3 5.3 24.4 161.9 3.7 4.5 358.0
4.7 3.7 5.6 21.9 0.3 0.5 81.4
0.069 7 0.002 9 0.083 30 0.046 140 0.015 4 0.015 4 1.000 358 BCD = 0.649 %Total BCD = 0.471
41.2 9.8 17.8 49.9 18.6 6.8 1.3 130.3
20.2 1.8 22.2 42.9 18.6 78.2 1.3 127.3
24.5 0.7 29.8 16.4 5.5 5.4 358.0
17.5 8.3 0.2 123.6 1.5 1.4 465.0
84
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
Table 5 Bray–Curtis distances from mean single-site exposures for the athletic shoe specimens.
Table 6 Bray–Curtis distances from mean single-site exposures for the work boot specimens.
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
85
Fig. 2. Explanation of the ternary diagrams used to visualize and contrast the relative BCDs of a specimen from each of the three sites. The lines labeled S1, S2 and S3 represent zero distance from each of the three sites. Parallel lines positioned toward the center of the diagram (and beyond, though not illustrated) represent increased distances from each of S1, S2 and S3. Thus the circled position labeled “A” at the center of the diagram would represent a specimen showing equal distances from each of the three sites. The circled position labeled “B,” as an example, represents a specimen that has a distance of 0.20 from site S1, 0.50 from site S2 and 0.30 from site S3.
Fig. 3. Example of the use of the ternary diagram to visualize and contrast the BCDs of individual specimens. The lines labeled LQ, AT, and PR represent the overall fractions of mineral categories for single-site exposures of the athletic shoe comparisons (the mean percentages of the mineral categories given in the columns of Table 2 for shoes). The plotted data represent the six individual measurements. Each of the specimens differs from, but is close to, the overall fractions corresponding to the single site to which it was exposed. The chart allows visualization of the relative BCDs of each of the specimens from each of the three sites.
86
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
Fig. 4. Ternary diagram illustrating the distance of each of the athletic shoe samples from each of the test sites.
Fig. 5. Ternary diagram illustrating the distance of each of the work boot samples from each of the test sites.
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
studies a five minute walk was sufficient to remove all carpet fibers from the contact surfaces of footwear, leaving only small numbers of carpet fibers that were, “either physically caught in the rubber at the edge of the sole or were in recessed area of the sole that did not directly come into contact with the ground.” 4.2. Limitations arising from the scope and methodology of this study The findings in this work are based on the analysis of a limited size fraction of particles recovered from the contact surfaces of two types of footwear exposed to three specific dry, dusty environments. Accordingly, the conclusions should not be casually extended to apply to other particle size ranges, other portions of footwear, to footwear generally, or to other environments/ environmental conditions. In particular, this study did not include exposure to wet or muddy surfaces. It would be expected that moisture could affect the removal of particles from earlier exposures. Comparatively strong adhesion of particles could also result from a drying process and caked, muddy deposits would clearly require specific study. 4.3. Implications and significance Subject to the limitations discussed in the previous section, there are several significant implications from this study. 1. During successive exposures there is nearly complete,sequential displacement of particles from the contact surfaces of footwear. This means that the contact surfaces of recovered footwear will have traces from the most recent areas where the footwear was worn. The implications of this finding are important. For example, in cases where a body is found and may have been transported after death from one location to another, the contact surfaces of the footwear will retain unmixed small particle traces that are directly representative of the last location where the deceased walked. Comparison with the location where the body was found will determine whether or not the body was moved and, if so, the traces will provide clues helping to locate the area from which the victim was transported. Alternatively, for footwear associated with a suspect, it is clear that the traces to be compared with crime scene locations are not those on the contact surfaces; rather they are those from recessed surfaces (see point 3, below). 2. Methods for differential sampling of the contact surfaces of footwear may not need to be employed. Given the absence of differences among loosely, moderately, and tightly held particle populations on the contact surfaces of footwear, the differential sampling of these particles may represent an unnecessary step. For the specimens in this study, moist swabbing could have been directly employed to comprehensively recover the adhering particle traces. If this finding is found to be generally extendable it would simplify collection of these particle traces. 3. Recessed areas of footwear are likely responsible for the observed retention of particles from prior exposures. The mixing of particles on footwear, arising from activity before, during and after the crime itself, has been the major obstacle to the exploitation of this type of evidence as a means to test the association of an accused to a crime. Our research shows (again, under the specific experimental conditions) that the contact surfaces of the soles retain particles from only the more recent exposures. This suggests that the mixtures of particles seen on footwear may well arise from the recessed areas of footwear. If this is the case, it may be that these areas should be separately sampled for evidence of prior exposures. Removal of the fraction from contact surfaces may well reduce the complexity of the mixture
87
and could lead to alternative approaches to differential sampling (see point 4, below). 4. Research on differential sampling of footwear should continue, focusing on the difference between particle populations found on contact surfaces and those found on recessed areas. Recessed areas of footwear could be responsible for the mixtures of particles arising from activity before, during and after the crime itself. The results of our research allow the isolation of particles from the most recent exposures (by sampling the contact surfaces). Subtracting this background from the mixtures found within recessed areas of the footwear provides a likely means to separate the evidentiary particle “signal” from background noise. Exploration of this possibility remains an intriguing area for follow-on research. The computational and statistical aspects of any such approach remain undeveloped. They would undoubtedly incorporate methods including counting statistics, as opposed to any simplified subtraction of mean particle numbers or fractions. 5. Research on related computational and statistical methods to interpret mixtures of particles should continue, focusing on multivariate methods that take advantage of both qualitative and quantitative distinctions among traces and possible sources. Given that the current differential sampling approach (based on how tightly particles are held) cannot be conveniently adapted to recessed areas of the soles, an alternative means to separate mixtures of particle signals needs to be developed. Following the approach in point 4 above, computational and statistical methods will be needed to subtract the “signals” from the most recent exposures (found on the contact surfaces) from mixtures found in recessed areas. Authors contributions D.S. designed study, chose test sites, conducted footwear exposures, sampled footwear, interpreted data, wrote the first draft, revised manuscript. A.B. processed specimens, chose test sites, performed microscopical analyses. P.S. designed study, conducted footwear exposures, supervised program personnel, revised manuscript. All authors read and approved the final manuscript. Acknowledgments This project was supported in part by Award No. 2014-DN-BXK011 awarded by the National Institute of Justice, Office of Justice Programs, U.S. Department of Justice. The opinions, findings, and conclusions or recommendations expressed in this article are those of the author and do not necessarily reflect those of the Department of Justice. References [1] H. Gross, Criminal Investigation, Krishnamachari, Madras, India, 1906, pp. 187– 226. [2] E. Locard, The analysis of dust traces (part I), Am. J. Police Sci. 1 (1930) 276–298. [3] R.M. Morgan, P.A. Bull, Forensic geoscience and crime detection, Minerva Med. Leg. 127 (2007) 73. [4] R.M. Morgan, J. Freudiger-Bonzon, K.H. Nichols, T. Jellis, S. Dunkerley, P. Zelazowski, P.A. Bull, The forensic analysis of sediments recovered from footwear, in: K. Ritz, L. Dawson, D. Miller (Eds.), Criminal and Environmental Soil Forensics, Springer, New York, 2009 p. 253. [5] D.A. Stoney, A.M. Bowen, V.M. Bryant, E.A. Caven, M.T. Cimino, P.L. Stoney, Particle combination analysis for predictive source attribution: tracing a shipment of contraband ivory, J. Assoc. Trace Evid. Exam. 2 (13) (2011). [6] D.A. Stoney, P.L. Stoney, Use of Scanning Electron Microscopy/Energy Dispersive Spectroscopy (SEM/EDS) Methods for the Analysis of Small Particles Adhering to Carpet Fiber Surfaces as a Means to Test Associations of Trace Evidence in a Way That is Independent of Manufactured Characteristics, US Department of Justice, 2012 NCJ 239051.
88
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
[7] D.A. Stoney, A.M. Bowen, P.L. Stoney, Utilization of environmentally acquired very small particles as a means of association, Forensic Sci. Int. 254 (2015) 26. [8] D.A. Stoney, C. Neumann, K.E. Mooney, J.M. Wyatt, P.L. Stoney, Exploitation of very small particles to enhance the probative value of carpet fibers, Forensic Sci. Int. 252 (2015) 58. [9] P.A. Bull, R.M. Morgan, A. Sagovsky, G.J. Hughes, The transfer and persistence of trace particulates: experimental studies using clothing fabrics, Sci. Justice 46 (2006) 185. [10] J. Robertson, C. Roux, Transfer and persistence, in: J.A. Siegel, P.J. Saukko, G.C. Knupfer (Eds.), Encyclopedia of Forensic Sciences, Academic Press, London, 2000 p. 834. [11] C.A. Pounds, K.W. Smalldon, The transfer of fibres between clothing materials during simulated contacts and their persistence during wear: part I — fibre transference, J. Forensic Sci. Soc. 15 (1975) 17. [12] C.A. Pounds, K.W. Smalldon, The transfer of fibres between clothing materials during simulated contacts and their persistence during wear: part II — fibre persistence, J. Forensic Sci. Soc. 15 (1975) 29. [13] C.A. Pounds, K.W. Smalldon, The transfer of fibres between clothing materials during simulated contacts and their persistence during wear: part III — a preliminary investigation of the mechanisms involved, J. Forensic Sci. Soc. 15 (1975) 197. [14] R.M. Morgan, J. Ainley, K. Scott, P.A. Bull, Trace materials on footwear — science or ichnomancy? NIJ/FBI 2011 Trace Evidence Symposium (2011). [15] C. Roux, S. Langdon, D. Waight, J. Robertson, The transfer and persistence of automotive carpet fibres on shoe soles, Sci. Justice 39 (1999) 239–251. [16] S. Bennett, C.P. Roux, J. Robertson, The significance of fibre transfer and persistence — a case study, Aust. J. Forensic Sci. 42 (2010) 221–228. [17] R.M. Morgan, J. Freudiger-Bonzon, K.H. Nichols, T. Jellis, S. Dunkerley, P. Zelazowski, P.A. Bull, The forensic analysis of sediments recovered from footwear, in: K. Ritz, L. Dawson, D. Miller (Eds.), Criminal and Environmental Soil Forensics, Springer, New York, 2009 p. 264. [18] R.J. Davis, J.D. DeHaan, A survey of men’s footwear, J. Forensic Sci. Soc. 17 (1977) 271. [19] C. Roux, R. Kirk, S. Benson, T. Van Haren, C.I. Petterd, Glass particles in footwear of members of the public in south-eastern Australia — a survey, Forensic Sci. Int. 116 (2001) 149. [20] D.S. Hilderbrand, Footwear, the Missed Evidence, 2nd ed., Staggs, Wildomar, CA, 2007. [21] K. Pye, Geological and Soil Evidence: Forensic Applications, CRC Press, Boca Raton, FL, 2007. [22] K. Pye, D.J. Croft (Eds.), Forensic Geoscience: Principles, Techniques and Applications, The Geological Society, London, UK, 2004. [23] K. Ritz, L. Dawson, D. Miller, Criminal and Environmental Soil Forensics, Springer, Berlin, 2009. [24] A. Ruffell, J. McKinley, Forensic geoscience: applications of geology, geomorphology and geophysics to criminal investigations, Earth Sci. Rev. 69 (2005) 235. [25] D. Pirrie, A. Ruffell, L.A. Dawson (Eds.), Environmental and Criminal Geoforensics, Geological Society of London, Bath, UK, 2013. [26] P.A. Bull, A. Parker, R.M. Morgan, The forensic analysis of soils and sediment taken from the cast of a footprint, Forensic Sci. Int. 162 (2006) 6.
[27] R.C. Murray, Evidence from the Earth: Forensic Geology and Criminal Investigation, 2nd ed., Mountain Press, Missoula, MT, 2011. [28] R.M. Morgan, P.A. Bull, Data interpretation in forensic sediment and soil geochemistry, Environ. Forensics 7 (2006) 325. [29] R.M. Morgan, P.A. Bull, The philosophy, nature and practice of forensic sediment analysis, Prog. Phys. Geogr. 31 (2007) 43. [30] R.M. Morgan, J. Cohen, I. McGookin, J. Murly-Gotto, R. O’Connor, S. Muress, J. Freudiger-Bonzon, P.A. Bull, The relevance of the evolution of experimental studies for the interpretation and evaluation of some trace physical evidence, Sci. Justice 49 (2009) 277. [31] R.M. Morgan, J.C. French, L. O’Donnell, P.A. Bull, The reincorporation and redistribution of trace geoforensic particulates on clothing: an introductory study, Sci. Justice 50 (2010) 195. [32] P.A. Bull, R.M. Morgan, Sediment fingerprints: a forensic technique using quartz sand grains, Sci. Justice 46 (2006) 107. [33] A.M. Bowen, Individualizing minerals: a proposed approach to forensic soil comparison, The Microscope 55 (2007) 59. [34] J. Robertson, The forensic examination of fibres: protocols and approaches — an overview, in: J. Robertson (Ed.), Forensic Examination of Fibres, Elllis Horwood, Chichester, UK, 1992 p. 41. [35] R.D. Koons, J. Buscaglia, M. Bottrell, E.T. Miller, Forensic glass comparisons, in: R. Saferstein (Ed.), Forensic Science Handbook, vol. I, Prentice Hall, Upper Saddle River, NJ, 2002 p. 161. [36] N. Petraco, P.R. De Forest, N.D.K. Petraco, A guide to the analysis of forensic dust specimens, 2nd ed., in: R. Saferstein (Ed.), Forensic Science Handbook, vol. III, Prentice Hall, Upper Saddle River, NJ, 2010 p. 31. [37] SWGMAT, Trace evidence recovery guidelines, Forensic Sci. Commun. 1 (3) (1999) https://archives.fbi.gov/archives/about-us/lab/forensic-science-communications/fsc/oct1999/trace.htm. (Accessed 1 November 2016). [38] T.W. Adair, R. Tewes, Lifting shoe impressions from cylindrical objects: a simple method, Inf. Bull. Shoeprint/Toolmark Exam. 12 (2006) 7. [39] C.J. Lentfer, M.M. Cotter, W.E. Boyd, Particle settling times for gravity sedimentation and centrifugation: a practical guide for palynologists, J. Archaeol. Sci. 30 (2003) 149. [40] H. Blatt, G.V. Middleton, R.C. Murray, Origin of Sedimentary Rocks, 2nd ed., Prentice-Hall, Englewood Cliffs, NJ, 1980. [41] J.K. Zubillaga, W.W. Dickinson, A. Carranza-Edwards, Y. Hornelas-Orozco, Petrography of quartz grains in beach and dune sands of Northland, North Island, New Zealand, N. Z. J. Geol. Geophys. 48 (2005) 649. [42] M.J. McVicar, W.J. Graves, The forensic comparison of soils by automated scanning electron microscopy, Can. Soc. Forensic Sci. J. 30 (1997) 241. [43] W.A. Deer, R.A. Howie, J. Zussman, An introduction to the rock-forming minerals, Longman Scientific & Technical, 2nd ed., Wiley, New York, NY, 1992. [44] M.A. Mange, H.F.W. Maurer, Heavy Minerals in Colour, 1st ed., Chapman & Hall, London, New York, 1992. [45] J.S. Galehouse, Point counting, in: R.E. Carver (Ed.), Procedures in Sedimentary Petrology, Wiley Interscience, New York, NY, 1971 p. 385. [46] J.R. Bray, J.T. Curtis, An ordination of upland forest communities of southern Wisconsin, Ecol. Monogr. 27 (1957) 325–349. [47] D.J. Graham, N.G. Midgley, Graphical representation of particle shape using triangular diagrams: an excel spreadsheet method, Earth Surf. Process. Landf. 25 (2000) 1473.
Contents lists available at ScienceDirect
Forensic Science International journal homepage: www.elsevier.com/locate/forsciint
Loss and replacement of small particles on the contact surfaces of footwear during successive exposures David A. Stoney* , Andrew M. Bowen1, Paul L. Stoney Stoney Forensic, Inc., 14101-G Willard Road, Chantilly, VA 20151-2934, USA
A R T I C L E I N F O
Article history: Received 20 April 2016 Received in revised form 4 November 2016 Accepted 8 November 2016 Available online 16 November 2016 Keywords: Trace evidence Footwear Very small particles Soil minerals Transfer Persistence
A B S T R A C T
On the contact surfaces of footwear loosely, moderately and strongly held particle fractions were separated and analyzed in an effort to detect different particle signals. Three environmental exposure sites were chosen to have different, characteristic particle types (soil minerals). Shoes of two types (work boots and tennis shoes) were tested, accumulating particles by walking 250 m in each environment. Some shoes were exposed to only one environment; others were exposed to all three, in one of six different sequences. Sampling methods were developed to separate particles from the contact surface of the shoe based on how tightly they were held to the sole. Loosely held particles were removed by walking on paper, moderately held particles were removed by electrostatic lifting, and the most tightly held particles were removed by moist swabbing. The resulting numbers and types of particles were determined using forensic microscopy. Particle profiles from the different fractions were compared to test the ability to objectively distinguish the order of exposure to the three environments. Without exception, the samples resulting from differential sampling are dominated by the third site in the sequential footwear exposures. No noticeable differences are seen among the differential samplings of the loosely, moderately and strongly held particles: the same overwhelming presence of the third site is seen. It is clear from these results (1) that the third (final) exposure results in the nearly complete removal of any particles from prior exposures, and (2) that under the experimental conditions loosely, moderately and strongly held particles are affected similarly, without any detectable enrichment of the earlier exposures among the more tightly held particles. These findings have significant implications for casework, demonstrating that particles on the contact surfaces of footwear are rapidly lost and replaced. ã 2016 Elsevier Ireland Ltd. All rights reserved.
1. Introduction 1.1. Statement of the problem Very small particles are ubiquitous in our environment. These “VSP” are particle dusts which, as noted by Gross [1], are our “environment or surroundings in miniature,” and as noted by Locard [2] “may be formed of all the debris and all kinds of bodies . . . all the substances, organic or inorganic, existing on the earth.” Everywhere people walk, VSP transfer to and from their
* Corresponding author. E-mail addresses: [email protected] (D.A. Stoney), [email protected] (A.M. Bowen), [email protected] (P.L. Stoney). 1 Present address: U.S. Postal Inspection Service, 22433 Randolph Drive, Dulles, VA 20104, USA. http://dx.doi.org/10.1016/j.forsciint.2016.11.015 0379-0738/ã 2016 Elsevier Ireland Ltd. All rights reserved.
footwear. The mere presence at a crime scene requires this contact and transfer, and the particles are known to persistent for long periods of time [3,4]. Even though criminals necessarily track dusts to and from every crime scene, dust particles on a suspect’s shoes are very seldom used as evidence linking the accused to the crime. There is an extraordinary, untapped potential to exploit VSP found on footwear and in footwear impressions. At the same time, there are significant challenges to unlocking this potential. Most fundamentally, VSP on footwear evidence are invariably a mixture of materials that can originate before, during, or after any event or period of forensic interest [3,4]. Their usefulness depends on our ability to separate a reliable, relevant evidentiary “signal” from background noise (or signals from other exposures). As an additional practical challenge, the VSP mixture is composed of many different particle types, which must be collected and analyzed efficiently.
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
Methods have been developed to efficiently analyze VSP using either an iterative forensic approach [5] or a particle profiling approach [6–8]. An iterative forensic approach begins with a multidisciplinary screening of particle types. This is followed by assessment of possible contributions to case resolution that could result from specialist examinations. Choices of which particles to analyze are made based on this assessment. The results of these analyses are then used to re-assess the possible contributions of additional specialist examinations of additional particle types. A particle profiling approach proceeds through the simultaneous characterization of many particles. This results in a profile representing the population of particles in a specimen. These profiles allow the application of computational methods showing the potential to measure strengths of associations between VSP specimens and to determine linkages among items of evidence based on their adhering VSP [7,8]. The application of these methods to VSP on (for example) shipping containers, clothing, and improvised explosive devices (IEDs) has consistently involved comparing VSP at different locations or different “layers” as one opens an item. The outer layer typically has VSP from the most recent exposures. The innermost layer has VSP from earlier exposures. Intermediate layers have exposures whose relative timing depends on how and when the item was handled, assembled or opened. On footwear the problem is more complex, as there are not physically separated layers (as there are in layers of packaging, or an assembled device). However, research focused on the persistence of trace evidence generally [9–13], and on footwear specifically [4,14–16], strongly supports the hypothesis that, after transfer to an item, some particles are tightly held (and retained longer), while others are loosely held (and more rapidly lost). Morgan et al. [17] have specifically found that for sediments on footwear there is a “trend of two/three stage decay . . . , with subsequently less rapid loss . . . , followed by a period of much lower decay.” Importantly, we observe that this explanation implies that particles from earlier exposures will be more concentrated among the particles that are more tightly held. From this strongly supported supposition, we hypothesized that, if we use differential sampling of footwear (which separates loosely held, moderately held, and strongly held particle fractions) we will recover physically separated or enriched particle fractions that originated from different exposures. This project explored the use of differential sampling of VSP from the soles of footwear as a potential method for the separation for these signals. 1.2. Background and context 1.2.1. Footwear evidence generally Trace evidence examiners encounter footwear as part of clothing examinations. Trace evidence commonly found on footwear includes the major types of fragmentary material traces: glass, paint and fibers [18,19], as well as accumulations of soil. Trace evidence within shoe impressions is only rarely utilized as part of the comparison, although the potential to compare this form of trace evidence with footwear is well-recognized [20]. As methods in forensic geoscience have developed [3,21–25], such cases are being reported [26]. 1.2.2. Soil and dust on footwear Accumulations of soil on footwear or other items of evidence (such as digging tools and vehicles) have long been exploited for comparisons with reference samples of possible origin [27]. The long-standing focus has been on fairly large accumulations of soil that can reasonably be expected to be minimally mixed, or that are clearly layered, so as to allow physical separation of discrete samples. Only then can comparisons be reliably made using bulk
79
properties of soil (such as color, particle size distributions and elemental composition). In cases where significant mixture has occurred, analysis of soil evidence is frequently stopped short. This is because preliminary analyses indicate disparities in the bulk properties (e.g. color) that are typically used to screen for comparable specimens. Restriction of analyses to unmixed specimens severely restricts the numbers of applicable cases. The work of Morgan, Bull and co-workers [3,4,28–31] has addressed this limitation, setting forth a conceptual framework for forensic geoscience [3]. This framework includes specific emphasis on analytical methods that can recognize mixtures and that are applicable when mixtures are present. They describe these methods as “visual techniques” and have exploited quartz grain surface analysis for this purpose. Quartz occurs very widely in sediments and the physical appearance of quartz grains depends on fundamental geological mechanisms relating to their origin and transportation [32]. When two different sources of soil are mixed, expert quartz grain surface analysis can, with reasonable probability, detect this mixture. Comparisons of the different types of quartz grain surfaces can be made even though the sources are mixed. This approach need not be limited to a single mineral type, or specifically to mineral particles, and it need not depend on the presence of one particle type (e.g. quartz) in each of the mixed sources [33]. What is essential is that recognizable varieties of minerals (or other particles) be exploited efficiently. Visual microscopical techniques do this: different soil or dust samples will have different suites of VSP. The presence and variety of VSP is a character that is recognizable within a mixture and meets the fundamental requirements for “visual techniques” [3,4]. 1.2.3. Studies of particle transfer and persistence on footwear Morgan et al.’s approach to recognition and analysis of mixed soil samples has continued with applied research directed at understanding mechanisms of transfer, persistence and mixing of particles deposited on footwear [4,14]. Experiments have been conducted using test substances (Plasticine) as well as using specific particle types (pollen or quartz grains). Specific caserelated research has also been conducted [30]. These studies have demonstrated that (1) particles persist for a long period of time on footwear, (2) that mixing of particles from successive exposures routinely occurs on the soles of footwear, and (3) that following exposure, some particles are loosely held (and more rapidly lost), while others are tightly held (and retained longer). 1.2.4. Alternative sampling methods as opposed to differential sampling Staged, alternative sampling methods are often employed in trace evidence analysis [34–37]. One purpose is to employ an initial method (such as picking individual fibers or paint chips) to collect loosely held traces as they are recognized. These traces might otherwise be lost or redistributed as the examination proceeds. Another purpose is to preserve and document the location from which trace evidence was recovered (as in the regional taping of clothing in the recovery of fibers). Again, different methods may be used for alternative particle types (such as taping for fibers, followed by vacuuming to recover fine particles, or washing to recover pollen). However, there has not been a protocol for differential sampling and recovery of trace evidence with the express intention to fractionate loosely and tightly held particles, so that these populations can be compared and contrasted. 2. Materials and methods This project was designed to test the separation of particle signals on the contact surfaces of footwear by applying a series of
80
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
successively more aggressive sampling steps and contrasting the resulting types and quantities of particles. Project objectives were (1) to conduct suitable environmental exposures, (2) to develop differential sampling methods, and (3) to test the ability to separate small particle signals of interest. Three environmental exposure sites were chosen to have different, characteristic particle types (soil minerals). Shoes of two types (work boots and tennis shoes) were tested, accumulating particles by walking 250 m in each environment. Some shoes were exposed to only one environment; others were exposed to all three, in one of six different sequences. Sampling methods were developed to separate particles from the contact surface of the shoe based on how tightly they were held to the sole. Loosely held particles were removed by walking on paper, moderately held particles were removed by electrostatic lifting, and the most tightly held particles were removed by moist swabbing. The resulting numbers and types of particles were determined using forensic microscopy. Particle profiles from the different fractions were compared to test the ability to objectively distinguish the order of exposure to the three environments. 2.1. Site selection A range of candidate sites were selected in the state of Virginia, USA based on recorded differences in their surface geology, differences in the watershed (indicating different sources for sediments) and differences related to human activities (in the form of trail or roadside modifications). Candidate sites were also required to be conveniently assessable from roads and to have an apparently uniform path of at least 25 m (allowing the required 250 m walking exposures to be completed in 5 round trips). From a total of 37 sites originally evaluated, three were selected based on their distinguishability and ease of access. These were: (1) Piney River (PR), (2) Appalachian Trail (AT) and (3) Luck Stone Quarry (LQ). Site PR is an improved hiking trail in a wooded site along the Virginia Blue Ridge Railway Trail (Lat. 37.7078, Long. 79.0220). Site AT is an unimproved minor trail just off of the main Appalachian Trail leading along the Tye River (Lat. 37.8384, Long. 79.0220). Site LQ is along the edges of a public access road outside of the Luck Stone Quarry in Fredericksburg, Virginia (Lat. 38.2128, Long. 77.5488). 2.2. Footwear exposures Two types of footwear were used: athletic shoes with flexible rubber soles (Kirkland SignatureTM Men’s Athletic Shoes, Fig. 1, left) and work boots with hard rubber soles (Grabbers Black Steel Toe EH Non-Slip Work Boots, Fig. 1, right). Exposures to test sites were
made by walking a distance of 250 m along a route, achieved by ten transects of 25 m (five round trips). Eighteen pairs of each footwear type were exposed: 6 pairs for single-environment exposures (two duplicate pairs for each of the three test sites) and 12 pairs for sequential exposures to all three environments (2 duplicate pairs for each of the six alternative sequences). Surfaces were raked to remove large plant matter prior to exposures. The surfaces were dry and dusty. Exposures were conducted over a five day rain-free period during which the surfaces remained dry. Following each exposure the footwear was gently re-packaged in its original box, between folds of butcher paper. 2.3. Differential sampling In this study only the contact surfaces of the footwear soles were sampled (those surfaces coming in direct contact with a horizontal surface while walking). The arch area and recessed areas within the sole pattern were not sampled. The more loosely adhering particles were removed by walking: 18 firm, smooth walking steps on butcher paper (92 kg individual, US shoe size 10.5). Twelve steps were found to be sufficient to remove the most loosely held particles, as further steps recovered no perceptible additional particles. Loose particles were recovered from the paper by moist swabbing with pre-filtered 3% ethanol. Moderately adhering particles were removed using an electrostatic lifter (Sirche Electrostatic Dust Print Lifter Kit ESP900). Electrostatic lifting was conducted using a reverse procedure (foil side of lifting film down) and employing a piece of foil taped to the floor as a conductor [38]. With the full voltage setting, 6 smooth steps (92 kg individual, US shoe size 10.5) were found to be sufficient to remove the moderately held particles, as further steps recovered no perceptible additional particles. Particles were collected from the electrostatic film by moist swabbing with pre-filtered 3% ethanol. Direct moist swabbing of the contact surfaces of the footwear soles was conducted to remove and collect the most tightly held particles. Each of the three sampling methods resulted in particle suspensions in 3% ethanol within 1.5 mL microcentrifuge tubes. 2.4. Specimen processing Specimens were washed to remove clay and silt-sized particles. Distilled water was added to the particle suspensions to a height of 3 cm in the 1.5 mL microcentrifuge tubes, followed by mixing using a vortex mixer (Thermolyne Type 16700 Mixer) with occasional brief sonication (Fisher Scientific FS6 Ultrasonic Cleaner). After mixing, samples were allowed to settle in the water for approximately 9 s (3 s per cm of water height) allowing all
Fig. 1. The two types of footwear used in this study: athletic shoes with flexible rubber soles (left) and work boots with hard rubber soles (right).
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
particles with a density greater than 2.61 and diameter greater than 62.5 mm to settle [39]. After settling, the upper 2 cm of liquid (well above the settled particles) was carefully removed using a pipette, taking with it a portion of the clay and silt-sized particles. This process was repeated until the liquid above the settled particles was clear (usually from six to eight times). This resulted in a specimen of sand-sized particles, washed free of silt and clay. The size cut-off of 62.5 mm was chosen because that is the lower limit of the sand-sized fraction, as defined by the Wentworth scale commonly used by sedimentary geologists [40]. The density value of 2.61 was selected because it is close to the density of quartz. Due to its abundance in igneous, sedimentary and metamorphic rocks, quartz is the most common detrital mineral in sediments on the earth’s surface [41]. Quartz is also abundant in the three sites selected for this study. The sand-sized fraction was wet-sieved, sub-dividing it into portions greater 180 mm (the coarse sand fraction) and that less than 180 mm (the fine sand fraction). Wet sieving was performed using distilled water and an Endecotts, Ltd. brand stainless steel sieve with 180 mm openings. The fraction passing the sieve was washed into a glass petri dish and recovered by pipette. The coarse sand fraction was not examined further. This size selection was performed for convenience, as particles larger than about 180 mm are more difficult to mount under a coverslip and to examine at high magnification using light microscopy. Where practical, the entire fine sand fraction was transferred onto a microscope slide by direct pipetting and allowing it to air dry. Where use of the entire fraction was impractical, a representative subsample was prepared using the method of McVicar and Graves [42]. The air-dried samples were mounted under a coverslip using Cargille mounting medium (Series A) 1.540nD, 25 C. 2.5. Polarized light microscopy Polarized light microscopy was performed using a Leitz DMRP polarized light microscope and a Leica DMLP polarized light microscope, at 200 times magnification. Optical identification and characterization of minerals were based on comparisons to known samples and reference data [43,44]. Variations in morphological and optical features were determined as described by Bowen [33]. Categorization of some mineral grains, as viewed in single mounts, required (as a practical matter under examination conditions) a subjective judgement with respect to placing certain grains into a category. The number of grains requiring subjective placement into a category was small (less than a few percent) for the samples in this study. Quantitative determinations were made by point counting of particles using the ribbon method [45]. 3. Results 3.1. Raw point count data and data reduction Point count data for the 108 program specimens included 34 mineral varieties. These were consolidated into 22 primary mineral types. Thirteen of these occurred above 2% in one or more of the single-site exposures. Data were grouped into these 13 categories, with a 14th category of “other.” Table 1 shows these mineral categories and an example of the consolidation. 3.2. Retention of particles on athletic shoe vs. work boot soles Data relating to the effect of the alternative footwear types are illustrated in Table 2 which gives the overall mean fractions of mineral categories for single site exposures. When comparing the
81
results for athletic shoes and work boots exposed to the same sites, the recovered mineral fractions are very close and there is no evidence that the alternative sole types result in different accumulations of the different mineral types. Pearson correlations of 0.995, 0.989 and 0.998 were observed for the three sites, and paired two-sample t-tests (hypothesized difference of means = 0) resulted in P(t t) of nearly 1. Table 2 also shows that the three sites are clearly distinguished by their mineral classification proportions. 3.3. Differential sampling of sequential site exposures An example of results from differential sampling following sequential site exposures is given in Table 3 for work boots exposed to the test sites in the sequence AT, PR, LQ. Data are given for grain counts and mineral classification fractions for each of the three differential samplings (walking, electrostatic lifting, and swabbing) for each of the two duplicate sequential exposures (walking 1, walking 2, etc.). The proportions of mineral classifications for the three differential samplings are highly similar to one another, with each corresponding to the proportions seen in site LQ: the last of the sites to which the boots were exposed. This observation was consistent across each of the six exposure sequences, for both footwear types. For each of the 72 mineral classification profiles (2 footwear types 6 exposure sequences 3 sample differential samplings 2 replicates) a measure of distance from each of the test sites was calculated using the Bray–Curtis Dissimilarity (BCD) measure (Eq. (1)) [46]. Xn1 jyi;k yj;k j ð1Þ dði; jÞ ¼ Xk¼0 n1 jy þ yj;k j k¼0 i;k Table 4 shows an example of the calculation of the BCD for the electrostatic lifter sampling from one of the work boots exposed to the site sequence AT, PR, LQ. A value of %total BCD for the specimen (the relative BCD) is also calculated as a means to readily contrast the distances from each of the sites. Tables 5 and 6 give the BCD measures for each of the athletic shoe and work boot specimens. Note that for the sequential exposures, without exception, the BCDs are lowest for the third (final) exposure site. Relative BCDs are conveniently visualized and contrasted using ternary diagrams following the method of Graham and Midgely [47]. Fig. 2 explains the basic structure of these diagrams and Fig. 3 gives an example showing a plot of the relative BCDs for the athletic shoe single-site exposures. These data are also included on Fig. 4, which plots the relative BCDs for each of the athletic shoe specimens shown in Table 5. Fig. 5 shows the corresponding diagram for the work boot specimens shown in Table 6. As further explained in the text accompanying Figs. 2 and 3, the three sides of the triangle in Figs. 4 and 5 correspond to the overall fractions of mineral categories for single-site exposures (the mean percentages of the mineral categories given in Table 2). The plotted points represent the relative BCDs for each of the differential samples recovered from the footwear soles, from each of the test sites. Points close to only one of the sites are close the edge corresponding to that site. Points close to more than one site would appear in the central portion of the chart. The legend shows black circles corresponding to single-site exposures and separate markers for each of the six sequences of test site exposures. One point is plotted for each of the samples recovered from footwear, with six points for each exposure sequence (from each of two replicates and the three differential samplings). In Fig. 4 (for athletic shoes) all of the samples, whether representing loosely, moderately or tightly held particles, appear near the edge corresponding to the last of the sites to which the
82
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
Table 1 Example of raw data, consolidation of mineral varieties and classification into the 14 most abundant types (athletic shoes with sequential exposure AT, PR, LQ, replicate 1). Point counts with varieties
Point counts without varieties
Grouped into 14 categories
Grain type
Count
Grain type
Count
Grain type
Count
Quartz — clear Quartz — with inclusions Quartz — with bubbles Quartz — with iron coatings Hornblende Alkali feldspar — fresh Alkali feldspar — weathered Alkali feldspar — inclusions Alkali feldspar — microcline Alkali feldspar — Fe coating Plagioclase Biotite — brown Biotite — green Biotite — yellow Biotite — orange Highly altereda Epidote Titanite Lithic fragments Lithic fragments — feldspars Lithic fragments — other Opaques Iron oxides Pyroxene Amphibole — colorless Muscovite Apatite Carbonates Zircon High n clear/prismb High n polycrystallinec Garnet Yellow striatedd Rutile
125 8 10 4 71 19 6 1 2 2 66 7 3 0 0 12 14 4 0 0 2 6 0 0 0 1 1 4 0 0 0 0 0 0 368
Quartz Hornblende Alkali feldspar Plagioclase Biotite Highly altereda Epidote Titanite Lithic fragments Opaques Iron oxides Pyroxene Amphibole colorless Muscovite Apatite Carbonates Zircon High n clear/prismb High n polycrystallinec Garnet Yellow striatedd Rutile
147 71 30 66 10 12 14 4 2 6 0 0 0 1 1 4 0 0 0 0 0 0 368
Alkali feldspar Highly altereda Biotite Epidote High n Hornblende Iron oxides Lithic fragments Muscovite Opaques Plagioclase Quartz Titanite Other
30 12 10 14 0 71 0 2 1 6 66 147 4 5 368
a
Grains too highly altered to clearly observe optical properties, allowing no classification into another category. Colorless grains with low to moderate birefringence, high refractive indices (n), and without characteristic features placing them in another category. Grains having high refractive indices (n) and composed of multiple smaller crystals appearing to be a single mineral type, and without characteristic features placing them in another category. d Yellow polycrystalline grains exhibiting cleavage striations. b c
Table 2 Overall fractions of mineral categories for single-site exposures contrasting results for work boots and athletic shoes.
shoes were exposed. The plotted points mingle with those from the shoes exposed only to this site. This means that the samples are uniformly dominated by the third site in the sequential footwear exposures. Likewise, in Fig. 5 (for work boots) all of the samples are closest to the last of the sites to which the boots were exposed. 4. Discussion and conclusions This section proceeds with a discussion of the findings, followed by further discussion of their limitations and implications. 4.1. Findings Without exception, the samples resulting from differential sampling are dominated by the third site in the sequential footwear exposures. No noticeable differences are seen among the differential samplings of the loosely, moderately and strongly held particles: the same overwhelming presence of the third site is seen. It is clear from these results that the third (final) exposure results in the nearly complete removal of any particles that were transferred to the contact surfaces of the shoe from the first and second exposures. This occurs regardless of the exposure sequence and regardless of which specific site was used for the third exposure. It is also clear that under the experimental conditions loosely, moderately and strongly held particles are affected similarly,
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
83
Table 3 Mineral classification counts, fractions and comparisons for boots exposed to test sites in the sequence AT, PR, LQ.
without any detectable enrichment of the earlier exposures among the more tightly held particles. The hypothesis guiding this work was that distinguishable fractions would result from the differential sampling of the contact surface of footwear. That is, that by separating loosely held, moderately held, and strongly held particle fractions we would recover enriched particle fractions originating from different exposures. This hypothesis is rejected. Under the experimental conditions the contact surface of footwear was found to be overwhelmingly dominated by the most recent exposure: a walk of 250 m, on a dry soil surface, results in the virtually complete
removal and replacement of particles adhering to the contact surfaces from prior, similar exposures. Given that Morgan et al. [14] have shown in comparable studies that a generalized sampling of footwear soles (from both contact and recessed areas) shows the retention of particles from earlier contacts, the clear implication of the present research is that, although particles on the contact surfaces of footwear are removed and replaced, those that are present on the more recessed areas of the sole are not. This reasoning parallels results shown by Roux et al. [15] for the retention of automobile carpet fibers on footwear soles. In their
Table 4 Example of Bray–Curtis distance calculation using electrostatic lift sampling from boot #1 exposed to the site sequence AT, PR, LQ. Raw count Comparison to site AT
Alkali feldspar Alterite Biotite Epidote High index Hornblende Iron oxides Lithic fragments Muscovite Opaques Plagioclase Quartz Titanite Other Total
Comparison to site LQ
Comparison to site PR
Mean AT Observed Expected |Obs Exp|
Mean LQ Observed Expected | Obs Exp|
Mean PR Observed Expected |Obs Exp|
21 8 40 7 0 85 0 3
0.556 0.065 0.003 0.019 0.007 0.005 0.020 0.008
7 9 30 140 4 4 358
0.002 7 0.064 9 0.013 30 0.217 140 0.001 4 0.021 4 1.000 358 BCD = 0.616 %Total BCD = 0.447
21 8 40 7 0 85 0 3
199.1 23.4 0.9 6.8 2.5 1.8 7.1 2.7
178.1 15.4 39.1 0.2 2.5 83.2 7.1 0.3
0.085 0.025 0.056 0.031 0.002 0.219 0.001 0.016
21 8 40 7 0 85 0 3
0.5 23.0 4.6 77.8 0.2 7.6 358.0
6.5 14.0 25.4 62.2 3.8 3.6 441.4
0.006 7 0.015 9 0.068 30 0.452 140 0.010 4 0.012 4 1.000 358 BCD = 0.114 %Total BCD = 0.082
30.4 8.9 20.1 11.0 0.9 78.5 0.4 5.8
9.4 0.9 19.9 4.0 0.9 6.5 0.4 2.8
0.115 0.027 0.050 0.139 0.052 0.019 0.004 0.364
21 8 40 7 0 85 0 3
2.3 5.3 24.4 161.9 3.7 4.5 358.0
4.7 3.7 5.6 21.9 0.3 0.5 81.4
0.069 7 0.002 9 0.083 30 0.046 140 0.015 4 0.015 4 1.000 358 BCD = 0.649 %Total BCD = 0.471
41.2 9.8 17.8 49.9 18.6 6.8 1.3 130.3
20.2 1.8 22.2 42.9 18.6 78.2 1.3 127.3
24.5 0.7 29.8 16.4 5.5 5.4 358.0
17.5 8.3 0.2 123.6 1.5 1.4 465.0
84
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
Table 5 Bray–Curtis distances from mean single-site exposures for the athletic shoe specimens.
Table 6 Bray–Curtis distances from mean single-site exposures for the work boot specimens.
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
85
Fig. 2. Explanation of the ternary diagrams used to visualize and contrast the relative BCDs of a specimen from each of the three sites. The lines labeled S1, S2 and S3 represent zero distance from each of the three sites. Parallel lines positioned toward the center of the diagram (and beyond, though not illustrated) represent increased distances from each of S1, S2 and S3. Thus the circled position labeled “A” at the center of the diagram would represent a specimen showing equal distances from each of the three sites. The circled position labeled “B,” as an example, represents a specimen that has a distance of 0.20 from site S1, 0.50 from site S2 and 0.30 from site S3.
Fig. 3. Example of the use of the ternary diagram to visualize and contrast the BCDs of individual specimens. The lines labeled LQ, AT, and PR represent the overall fractions of mineral categories for single-site exposures of the athletic shoe comparisons (the mean percentages of the mineral categories given in the columns of Table 2 for shoes). The plotted data represent the six individual measurements. Each of the specimens differs from, but is close to, the overall fractions corresponding to the single site to which it was exposed. The chart allows visualization of the relative BCDs of each of the specimens from each of the three sites.
86
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
Fig. 4. Ternary diagram illustrating the distance of each of the athletic shoe samples from each of the test sites.
Fig. 5. Ternary diagram illustrating the distance of each of the work boot samples from each of the test sites.
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
studies a five minute walk was sufficient to remove all carpet fibers from the contact surfaces of footwear, leaving only small numbers of carpet fibers that were, “either physically caught in the rubber at the edge of the sole or were in recessed area of the sole that did not directly come into contact with the ground.” 4.2. Limitations arising from the scope and methodology of this study The findings in this work are based on the analysis of a limited size fraction of particles recovered from the contact surfaces of two types of footwear exposed to three specific dry, dusty environments. Accordingly, the conclusions should not be casually extended to apply to other particle size ranges, other portions of footwear, to footwear generally, or to other environments/ environmental conditions. In particular, this study did not include exposure to wet or muddy surfaces. It would be expected that moisture could affect the removal of particles from earlier exposures. Comparatively strong adhesion of particles could also result from a drying process and caked, muddy deposits would clearly require specific study. 4.3. Implications and significance Subject to the limitations discussed in the previous section, there are several significant implications from this study. 1. During successive exposures there is nearly complete,sequential displacement of particles from the contact surfaces of footwear. This means that the contact surfaces of recovered footwear will have traces from the most recent areas where the footwear was worn. The implications of this finding are important. For example, in cases where a body is found and may have been transported after death from one location to another, the contact surfaces of the footwear will retain unmixed small particle traces that are directly representative of the last location where the deceased walked. Comparison with the location where the body was found will determine whether or not the body was moved and, if so, the traces will provide clues helping to locate the area from which the victim was transported. Alternatively, for footwear associated with a suspect, it is clear that the traces to be compared with crime scene locations are not those on the contact surfaces; rather they are those from recessed surfaces (see point 3, below). 2. Methods for differential sampling of the contact surfaces of footwear may not need to be employed. Given the absence of differences among loosely, moderately, and tightly held particle populations on the contact surfaces of footwear, the differential sampling of these particles may represent an unnecessary step. For the specimens in this study, moist swabbing could have been directly employed to comprehensively recover the adhering particle traces. If this finding is found to be generally extendable it would simplify collection of these particle traces. 3. Recessed areas of footwear are likely responsible for the observed retention of particles from prior exposures. The mixing of particles on footwear, arising from activity before, during and after the crime itself, has been the major obstacle to the exploitation of this type of evidence as a means to test the association of an accused to a crime. Our research shows (again, under the specific experimental conditions) that the contact surfaces of the soles retain particles from only the more recent exposures. This suggests that the mixtures of particles seen on footwear may well arise from the recessed areas of footwear. If this is the case, it may be that these areas should be separately sampled for evidence of prior exposures. Removal of the fraction from contact surfaces may well reduce the complexity of the mixture
87
and could lead to alternative approaches to differential sampling (see point 4, below). 4. Research on differential sampling of footwear should continue, focusing on the difference between particle populations found on contact surfaces and those found on recessed areas. Recessed areas of footwear could be responsible for the mixtures of particles arising from activity before, during and after the crime itself. The results of our research allow the isolation of particles from the most recent exposures (by sampling the contact surfaces). Subtracting this background from the mixtures found within recessed areas of the footwear provides a likely means to separate the evidentiary particle “signal” from background noise. Exploration of this possibility remains an intriguing area for follow-on research. The computational and statistical aspects of any such approach remain undeveloped. They would undoubtedly incorporate methods including counting statistics, as opposed to any simplified subtraction of mean particle numbers or fractions. 5. Research on related computational and statistical methods to interpret mixtures of particles should continue, focusing on multivariate methods that take advantage of both qualitative and quantitative distinctions among traces and possible sources. Given that the current differential sampling approach (based on how tightly particles are held) cannot be conveniently adapted to recessed areas of the soles, an alternative means to separate mixtures of particle signals needs to be developed. Following the approach in point 4 above, computational and statistical methods will be needed to subtract the “signals” from the most recent exposures (found on the contact surfaces) from mixtures found in recessed areas. Authors contributions D.S. designed study, chose test sites, conducted footwear exposures, sampled footwear, interpreted data, wrote the first draft, revised manuscript. A.B. processed specimens, chose test sites, performed microscopical analyses. P.S. designed study, conducted footwear exposures, supervised program personnel, revised manuscript. All authors read and approved the final manuscript. Acknowledgments This project was supported in part by Award No. 2014-DN-BXK011 awarded by the National Institute of Justice, Office of Justice Programs, U.S. Department of Justice. The opinions, findings, and conclusions or recommendations expressed in this article are those of the author and do not necessarily reflect those of the Department of Justice. References [1] H. Gross, Criminal Investigation, Krishnamachari, Madras, India, 1906, pp. 187– 226. [2] E. Locard, The analysis of dust traces (part I), Am. J. Police Sci. 1 (1930) 276–298. [3] R.M. Morgan, P.A. Bull, Forensic geoscience and crime detection, Minerva Med. Leg. 127 (2007) 73. [4] R.M. Morgan, J. Freudiger-Bonzon, K.H. Nichols, T. Jellis, S. Dunkerley, P. Zelazowski, P.A. Bull, The forensic analysis of sediments recovered from footwear, in: K. Ritz, L. Dawson, D. Miller (Eds.), Criminal and Environmental Soil Forensics, Springer, New York, 2009 p. 253. [5] D.A. Stoney, A.M. Bowen, V.M. Bryant, E.A. Caven, M.T. Cimino, P.L. Stoney, Particle combination analysis for predictive source attribution: tracing a shipment of contraband ivory, J. Assoc. Trace Evid. Exam. 2 (13) (2011). [6] D.A. Stoney, P.L. Stoney, Use of Scanning Electron Microscopy/Energy Dispersive Spectroscopy (SEM/EDS) Methods for the Analysis of Small Particles Adhering to Carpet Fiber Surfaces as a Means to Test Associations of Trace Evidence in a Way That is Independent of Manufactured Characteristics, US Department of Justice, 2012 NCJ 239051.
88
D.A. Stoney et al. / Forensic Science International 269 (2016) 78–88
[7] D.A. Stoney, A.M. Bowen, P.L. Stoney, Utilization of environmentally acquired very small particles as a means of association, Forensic Sci. Int. 254 (2015) 26. [8] D.A. Stoney, C. Neumann, K.E. Mooney, J.M. Wyatt, P.L. Stoney, Exploitation of very small particles to enhance the probative value of carpet fibers, Forensic Sci. Int. 252 (2015) 58. [9] P.A. Bull, R.M. Morgan, A. Sagovsky, G.J. Hughes, The transfer and persistence of trace particulates: experimental studies using clothing fabrics, Sci. Justice 46 (2006) 185. [10] J. Robertson, C. Roux, Transfer and persistence, in: J.A. Siegel, P.J. Saukko, G.C. Knupfer (Eds.), Encyclopedia of Forensic Sciences, Academic Press, London, 2000 p. 834. [11] C.A. Pounds, K.W. Smalldon, The transfer of fibres between clothing materials during simulated contacts and their persistence during wear: part I — fibre transference, J. Forensic Sci. Soc. 15 (1975) 17. [12] C.A. Pounds, K.W. Smalldon, The transfer of fibres between clothing materials during simulated contacts and their persistence during wear: part II — fibre persistence, J. Forensic Sci. Soc. 15 (1975) 29. [13] C.A. Pounds, K.W. Smalldon, The transfer of fibres between clothing materials during simulated contacts and their persistence during wear: part III — a preliminary investigation of the mechanisms involved, J. Forensic Sci. Soc. 15 (1975) 197. [14] R.M. Morgan, J. Ainley, K. Scott, P.A. Bull, Trace materials on footwear — science or ichnomancy? NIJ/FBI 2011 Trace Evidence Symposium (2011). [15] C. Roux, S. Langdon, D. Waight, J. Robertson, The transfer and persistence of automotive carpet fibres on shoe soles, Sci. Justice 39 (1999) 239–251. [16] S. Bennett, C.P. Roux, J. Robertson, The significance of fibre transfer and persistence — a case study, Aust. J. Forensic Sci. 42 (2010) 221–228. [17] R.M. Morgan, J. Freudiger-Bonzon, K.H. Nichols, T. Jellis, S. Dunkerley, P. Zelazowski, P.A. Bull, The forensic analysis of sediments recovered from footwear, in: K. Ritz, L. Dawson, D. Miller (Eds.), Criminal and Environmental Soil Forensics, Springer, New York, 2009 p. 264. [18] R.J. Davis, J.D. DeHaan, A survey of men’s footwear, J. Forensic Sci. Soc. 17 (1977) 271. [19] C. Roux, R. Kirk, S. Benson, T. Van Haren, C.I. Petterd, Glass particles in footwear of members of the public in south-eastern Australia — a survey, Forensic Sci. Int. 116 (2001) 149. [20] D.S. Hilderbrand, Footwear, the Missed Evidence, 2nd ed., Staggs, Wildomar, CA, 2007. [21] K. Pye, Geological and Soil Evidence: Forensic Applications, CRC Press, Boca Raton, FL, 2007. [22] K. Pye, D.J. Croft (Eds.), Forensic Geoscience: Principles, Techniques and Applications, The Geological Society, London, UK, 2004. [23] K. Ritz, L. Dawson, D. Miller, Criminal and Environmental Soil Forensics, Springer, Berlin, 2009. [24] A. Ruffell, J. McKinley, Forensic geoscience: applications of geology, geomorphology and geophysics to criminal investigations, Earth Sci. Rev. 69 (2005) 235. [25] D. Pirrie, A. Ruffell, L.A. Dawson (Eds.), Environmental and Criminal Geoforensics, Geological Society of London, Bath, UK, 2013. [26] P.A. Bull, A. Parker, R.M. Morgan, The forensic analysis of soils and sediment taken from the cast of a footprint, Forensic Sci. Int. 162 (2006) 6.
[27] R.C. Murray, Evidence from the Earth: Forensic Geology and Criminal Investigation, 2nd ed., Mountain Press, Missoula, MT, 2011. [28] R.M. Morgan, P.A. Bull, Data interpretation in forensic sediment and soil geochemistry, Environ. Forensics 7 (2006) 325. [29] R.M. Morgan, P.A. Bull, The philosophy, nature and practice of forensic sediment analysis, Prog. Phys. Geogr. 31 (2007) 43. [30] R.M. Morgan, J. Cohen, I. McGookin, J. Murly-Gotto, R. O’Connor, S. Muress, J. Freudiger-Bonzon, P.A. Bull, The relevance of the evolution of experimental studies for the interpretation and evaluation of some trace physical evidence, Sci. Justice 49 (2009) 277. [31] R.M. Morgan, J.C. French, L. O’Donnell, P.A. Bull, The reincorporation and redistribution of trace geoforensic particulates on clothing: an introductory study, Sci. Justice 50 (2010) 195. [32] P.A. Bull, R.M. Morgan, Sediment fingerprints: a forensic technique using quartz sand grains, Sci. Justice 46 (2006) 107. [33] A.M. Bowen, Individualizing minerals: a proposed approach to forensic soil comparison, The Microscope 55 (2007) 59. [34] J. Robertson, The forensic examination of fibres: protocols and approaches — an overview, in: J. Robertson (Ed.), Forensic Examination of Fibres, Elllis Horwood, Chichester, UK, 1992 p. 41. [35] R.D. Koons, J. Buscaglia, M. Bottrell, E.T. Miller, Forensic glass comparisons, in: R. Saferstein (Ed.), Forensic Science Handbook, vol. I, Prentice Hall, Upper Saddle River, NJ, 2002 p. 161. [36] N. Petraco, P.R. De Forest, N.D.K. Petraco, A guide to the analysis of forensic dust specimens, 2nd ed., in: R. Saferstein (Ed.), Forensic Science Handbook, vol. III, Prentice Hall, Upper Saddle River, NJ, 2010 p. 31. [37] SWGMAT, Trace evidence recovery guidelines, Forensic Sci. Commun. 1 (3) (1999) https://archives.fbi.gov/archives/about-us/lab/forensic-science-communications/fsc/oct1999/trace.htm. (Accessed 1 November 2016). [38] T.W. Adair, R. Tewes, Lifting shoe impressions from cylindrical objects: a simple method, Inf. Bull. Shoeprint/Toolmark Exam. 12 (2006) 7. [39] C.J. Lentfer, M.M. Cotter, W.E. Boyd, Particle settling times for gravity sedimentation and centrifugation: a practical guide for palynologists, J. Archaeol. Sci. 30 (2003) 149. [40] H. Blatt, G.V. Middleton, R.C. Murray, Origin of Sedimentary Rocks, 2nd ed., Prentice-Hall, Englewood Cliffs, NJ, 1980. [41] J.K. Zubillaga, W.W. Dickinson, A. Carranza-Edwards, Y. Hornelas-Orozco, Petrography of quartz grains in beach and dune sands of Northland, North Island, New Zealand, N. Z. J. Geol. Geophys. 48 (2005) 649. [42] M.J. McVicar, W.J. Graves, The forensic comparison of soils by automated scanning electron microscopy, Can. Soc. Forensic Sci. J. 30 (1997) 241. [43] W.A. Deer, R.A. Howie, J. Zussman, An introduction to the rock-forming minerals, Longman Scientific & Technical, 2nd ed., Wiley, New York, NY, 1992. [44] M.A. Mange, H.F.W. Maurer, Heavy Minerals in Colour, 1st ed., Chapman & Hall, London, New York, 1992. [45] J.S. Galehouse, Point counting, in: R.E. Carver (Ed.), Procedures in Sedimentary Petrology, Wiley Interscience, New York, NY, 1971 p. 385. [46] J.R. Bray, J.T. Curtis, An ordination of upland forest communities of southern Wisconsin, Ecol. Monogr. 27 (1957) 325–349. [47] D.J. Graham, N.G. Midgley, Graphical representation of particle shape using triangular diagrams: an excel spreadsheet method, Earth Surf. Process. Landf. 25 (2000) 1473.