Forensic Geoscience and Crime Detection Identification, Interpretation and Presentation in Forensic Geoscience

MINERVA MED LEG 2007;127:73-89

Forensic geoscience and crime detection Identification, interpretation and presentation in forensic geoscience

R. M. MORGAN, P. A. BULL

Forensic geoscience is based upon well establi- Oxford University Centre for the Environment shed ideas and analytical techniques developed University of Oxford, Oxford, UK throughout the 20th century. Recent advances in geological and chemical analytical techniques have enabled, for the first time, detailed soil and sediment assay which has been taken up particularly by British scientists. A conceptual cope with the ever increasing complexities of framework for forensic geoscience is reiterated data analysis and interpretation in order for which supports the positive development in the the results to be presented in an unbiased and future of forensic geoscience, soil and sediment analyses. However, this paper also seeks to hi- clear way to both judge and jury in a court of ghlight the pitfalls encountered in recent years: law. These problems are compounded in the use of analytical techniques that are depen- forensic geoscience by additional complica- dent upon one another; the dependence on the tions in data acquisition, analysis, interpreta- exotic components of a sample, whether they tion and presentation which are individual to are representative or not; the confusion between descriptive; exclusionary and diagnostic tech- this particular, rapidly expanding, field of niques, and the failure to comprehend the re- forensic enquiry. Inspection of the post-2000 quirement to exclude rather than to match com- scientific published literature confirms the ex- parison samples. Furthermore, the use of auto- panding nature of forensic geoscience,1 but mated, computer driven, multiple-sample analy- we must ask ourselves whether this burgeon- tical machinery is questioned when there is no visual assessment by an operator included in ing of geological applications is but another the procedure. Avoiding these pitfalls and adop- false dawn in a series of false starts seen ting the philosophical framework specific to fo- throughout the 20th century. Certainly, if we do rensic geoscience will ensure a role for the ap- not learn from mistakes and do not take heed plication of forensic geoscience in the field of cri- of comments and advice given in the past, minal investigations. then this current resurgence in the use of geo- Key words: Forensic geoscience - Philosophy - science applications to forensic problems will Soil - Geologic sediment. once again fail to reach its full potential. This paper introduces the recent developments in forensic geoscience set within the context of ne of the most pressing issues in foren- problems which beset the discipline. Osic science at the present time is how to What is forensic geoscience? Address reprint requests to: R. M. Morgan, Oxford University Centre for the Environment, University of Oxford, South Parks Road, Oxford OX1 3QY, UK. Forensic geoscience is a field of enquiry E-mail: [email protected] that utilises techniques developed in the geo-

Vol. 127, N. 2 MINERVA MEDICOLEGALE 73 MORGAN FORENSIC GEOSCIENCE AND CRIME DETECTION

TABLE I.—Macro and micro geoforensic applications.

Scale Technique Purpose Reference

Macro Resistivity and electrical tomography Detecting burial sites France et al. 1992,2 Buck 2003,3 Scott and Hunter 2004 4 Electromagnetic surveying (EM) and Ground- Detecting burial sites Nobes 1999, 2000 5, 6 Penetrating Radar (GPR) GPR Detecting historic mass Davis et al. 2000,7 burial sites Chen et al. 2001 8 Plastic landmine detection Forensic geomorphology (landscape interpretation) Detecting burial sites Owsley 1995 9 Physical probe Detecting buried objects Owsley 1995 9 Forensic remote sensing (aerial photography, satellite Environmental forensic Brilis et al. 2000, imagery and global positioning systems) applications 2001, 10, 11 Grip et al. 2000 12 Geographic Information System (GIS) Spatial analysis of crimi- Canter 2003, 13 nal behaviour and crimi- Zhivotovsky et al. 2001 14 nal ancestry Micro Physical characteristics of Binocular microscopy Comparison of samples Murray 2004,15 a soil/sediment sample (mineralogy) and/or assessment of the Morgan et al. 2006 16 provenance of samples. 17 X-Ray Diffraction (XRD) Brown 2006, (mineralogy) Ruffell and Wiltshire 2004 18 QemSCAN Pirrie et al. 2004 19 Quartz grain surface Bull and Morgan 2006 20 texture analysis Particle size analysis Morgan and Bull in press.21 Colour Janssen et al. 1983,22 Sugita and Marumo 1996, 23 Junger 1996.24 Chemical characteristics ICP-MSa and ICP-AESb Comparison of samples Rawlins and Cave 2004,25 of a soil/sediment sam- (elemental chemistry) and/or assessment of the Jarvis et al. 2004 26 ple AAS/Dionexc (elemental provenance of samples. Morgan et al. 2006, 16 chemistry) Bull et al. 2006 27 FTIRd (organic content) Cox et al. 2000 28

Biological characteristics Pollen Comparison of samples Bruce and Dettmann of a soil/sediment sam- and/or assessment of the 1996,29 Bryant et al. 1996, ple provenance of samples. 30 Szibor et al. 1998,31 Eyring 1996,32 Horrocks et al. 1998,33 Horrocks and Walsh 1998, 1999,34, 35 Horrocks 2004,36 Miller Coyle 2005,37 Mildenhall et al. 2006,38 Montali et al. 2006,39 Wiltshire 2006.40 Bacterial DNA Horswell et al. 2002 41 Plant wax signatures Dawson et al. 2003 42 a ICP-MS: Inductively Coupled Plasma-Mass Spectrometry; b ICP-AES: Inductively Coupled Atomic Emission Spectroscopy; c AAS: Atomic absorption Spectrometry; d FTIR: Fourier Transform Infra Red spectroscopy.

74 MINERVA MEDICOLEGALE Giugno 2007 FORENSIC GEOSCIENCE AND CRIME DETECTION MORGAN

sciences (such as geology, geomorphology, anthropogenic components such as glass, botany, biology and statistics) for application paint fragments, metallic particles etc. Thus, to civil and criminal judicial proceedings. a sample of soil/sediment recovered from Forensic geoscience may include both macro clothing, a vehicle or a crime scene has “…a and micro-scale applications 2-42 (Table I). large, almost limitless number of characteristics The macro-scale applications of forensic which make it unique but which relate to the geoscience have been reviewed elsewhere, 43 climate, geology or land use of the site from thus this review considers specifically the which the soil was derived”,54 indeed there role of forensic geoscience in crime detection are “…an almost unlimited number of soil at the micro-scale. At this micro-scale the varieties and soils change rapidly from place geo-forensic application falls within the aegis to place even over short distances”.55 The va- of trace evidence analysis, which has devel- lue of soil/sediment analysis in providing oped its own principles and practice since useful evidence in forensic enquiries lies with the utilisation of the microscope in forensic the ability of the forensic practitioner to ma- enquiry in the 19th century. ke comparisons between soil/sediment The well established principles and pro- samples and, on occasion, to infer prove- tocols of trace evidence studies as applied nance. to hair, paint fragments, glass particulates, fi- The concept of facies and sedimentary bres, metals etc. have now amalgamated units is a keystone to the analysis of soils within them the more recent developments of and sediments in forensic enquiry. A “…sed- forensic geoscience (incorporating studies of imentary facies is defined as any aerially re- soils and minerals, pollen and micro-fossils). stricted part of a designated stratigraphic unit Whilst such a grouping is seemingly logical, which exhibits characteristics significantly the well developed protocols of the former different from those of other parts of the group of trace evidence analyses do not take unit”.56 It is, in other words, distinctive as a re- into account the nature, philosophy and prac- sult of the similarities within the unit and its tice specific to these geoforensic studies. distinctiveness from other units. Accordingly, Conversely, the geoforensic trace materials it follows that facies are discrete and, there- are often analysed and interpreted by a foren- fore, restricted in extent both stratigraphical- sic geoscientist in a totally obverse manner in ly and geographically. 57 Thus there is a dis- comparison to that of the properly estab- cernable difference between sedimentary lished trace material analyst. units in the spatial dimension. The most im- Geoforensic science has relevance to oth- portant application of this, therefore, is that er forensic disciplines. Medical applications sediments are not to be found in a melange include the identification of trace materials ad- of indiscrete units but rather the inter-rela- hering to bodies, under fingernails, in nasal tionships of sedimentary environments, and passages, in lung and stomach contents and hence of facies, are not chaotic or random, also in the blood system (diatoms;44-48 but subject to controls such as climate, tec- pollen;49, 50 entomology 51). tonics and geological setting.58 Therefore, soils and sediments have been argued to pos- Soil/sediment as geoforensic evidence sess the “…clearest expressions of landscape complexity”, their “…spatial variability is high It is gradually being recognised that and intricate over short distances, whether “…much potentially useful information is in profile or in plan”.59 As no two similar en- locked up in even small amounts of soil” 52 vironments are exactly identical,60 this char- and this can be attributed not only to its acteristic of soils and sediments enables the “…prevalence at crime scenes and its trans- use of comparison methods in the analysis of ferability between the scene and the criminal” trace soil/sediment samples to be used to 53 but also to the fact that soils/sediments are great effect in forensic enquiry. comprised not only of naturally occurring However, it is not only the identification of rocks, minerals, flora and fauna but also of the components of a soil/sediment sample

Vol. 127, N. 2 MINERVA MEDICOLEGALE 75 MORGAN FORENSIC GEOSCIENCE AND CRIME DETECTION

that enables the use of such evidence; it is of physical evidence that cannot be deemed to great importance that the interpretation of be unique (in the same sense as a fingerprint such analysis and their presentation to the or DNA profile). Rather, soil/sediment evi- court are accurate and meaningful. Indeed, dence should be viewed as being of a prob- more powerful analytical instruments are con- abilistic nature.65 It is, therefore, not possible stantly being developed providing the po- to look for a “match” or an association tential for greater and greater discriminatory between two samples that are being com- detail of soil/sediment samples in a forensic pared. This attitude contrasts with the ap- context to be achieved.55 It is not the accuracy proach in the more traditional geosciences of such techniques that is questionable as where comparisons are made between the “…physical evidence cannot be wrong; it soil/sediment samples in order to identify cannot perjure itself; it cannot be wholly ab- similarities so as to classify, date or infer sent…” rather “…only in its interpretation provenance of the samples in question.17, 66, can there be error”.61 Such error has the po- 67 tential for untenable consequences in the The geological techniques which are used forensic arena; the courtroom. to assess the characteristics of soil/sediment samples in forensic geoscience are deemed to The philosophical framework of forensic geo- be predominantly descriptive techniques. In science certain cases these descriptive techniques may prove to be exclusionary and more Whilst forensic geoscience and the more rarely, diagnostic. One of the biggest prob- traditional geosciences may share a Popperian lems that needs to be addressed by the geo- approach to falsification 62, 63 and indeed forensic community is that of the analysis of utilise some of the same techniques and in- soil/sediment samples taken from pertinent formation, “…the philosophical approach of exhibits comprised of a mixture of soil/sed- the two could not be more contrasting”.1 The iments from a number of different sources development of forensic geoscience in the (for example from a vehicle or a shoe). These 20th century has been documented 43, 64 and it “anthropogenic” samples are likely to include is clear that many of the major contributions materials transferred pre-, syn- and post- have come from the geosciences. Whilst the de- forensic event. It is, therefore, problematic velopment of the analysis of different compo- to undertake analysis of these samples us- nents of soils/sediments in geological studies ing an analytical technique which requires by a number of analytical techniques has the homogenisation of the sample prior to greatly contributed to the advance of foren- analysis. Such a step renders it impossible to sic geoscience, the transposition of tech- assess whether the results derived exclude niques from one discipline to another is not the sample from the having a similar prove- simple and should not be done unless con- nance to the crime scene sample, or whether sideration of the contrasting approaches is the exclusion is a false negative result (due to taken into account. the presence of materials from other sources We consider that there are four major facets in addition to the crime scene material). of the philosophical framework within which Examples of this have been documented by geoforensic practitioners should operate:1 Morgan et al. 16 and Morgan and Bull.1, 21 1. The requirement to exclude rather than There is general consensus amongst the to match comparison samples. geoforensic community that a number of an- 2. The nature of analytical techniques (de- alytical techniques should be employed in scriptive, exclusionary and diagnostic). the analysis of soil/sediment samples.54, 68, 69 3. The necessity of employing a number of However, whilst some practitioners advocate independent forensic techniques. the use of certain suites of analytical tech- 4. The balance between the “exotic” and niques, 70, 71 we contend that it is of greater the “ubiquitous” component of a sample. importance that the third facet of the geo- Forensic geoscience deals with a form of forensic philosophical framework is adhered

76 MINERVA MEDICOLEGALE Giugno 2007 FORENSIC GEOSCIENCE AND CRIME DETECTION MORGAN

to. A suite of techniques that measure char- deed truly exotic, or whether it is far more acteristics of the soil/sediment sample that common than is thought. The presence of are dependent upon each other may well an element or combination of elements iden- provide corroborative results, but such re- tified by high powered analytical techniques sults can not be deemed to strengthen the may not have been routinely sought before interpretation of the original analysis.72, 73 The and in consequence no detailed micro-scale colour, elemental composition and particle database exists to confirm or reject this “ex- size distribution of a soil/sediment sample otic” designation. will all be dependent to some extent on the Such a philosophical framework has been mineralogy of that sample. To present the presented by Morgan and Bull,1 but it is im- findings of these four different analytical tech- portant to note that this is a very recent ad- niques to a court and assert that the agree- dition to the published literature. Many geo- ment in the findings adds weight to the forensic practitioners (often with a geoscience strength of that evidence is not a tenable po- background or training) continue to approach sition. Techniques that can be considered to geoforensic case work with a geoscience be more independent may include, for ex- mind set. There is currently a debate amongst ample, mineralogy, pollen, quartz grain analy- different laboratories and scientists, but it is sis, and anthropogenic additions such as paint undeniable that it is of great importance that and fibres (a number of examples are in- a framework is agreed upon in order that the cluded in Lombardi,68 Hunter et al.,69 Brown,54 field of geoforensics may become a robust Bull et al. 27 and Morgan et al. 16). scientific discipline. The decision to look for the ubiquitous or “exotic” component of the soil/sediment sam- The conceptual framework of forensic ple must be reached with care and will de- geoscience: the order of events pend upon the particular geoforensic case in question. The “exotic” component of a Given the nature of forensic geoscience soil/sediment (such as an anthropogenic ad- and the methods by which analysis of dition, or a rare or highly place-specific pollen soils/sediments is approached, it is now nec- grain; see Morgan et al. 16) may prove to be essary to consider the order of events which a diagnostic feature that enables the exclusion lead to sample characterisation, analysis, in- of particular samples from the enquiry. terpretation and presentation. We present Caution must be taken, however, to ensure here a conceptual framework, that builds up- that when very small quantities of samples are on the earlier work of Inman and Rudin 74 being compared, the “exotic” component is within which all forensic investigations gen- representative of the parent sample and the erally operate. There are six fundamental original source. It is possible that due to their tenets upon which the process of forensic very nature, “exotic” components may only science (including forensic geoscience) rests. be found in very low concentrations in a They are: sample assay. There is a possibility that such 1. Division of matter and transfer. “exotic” components may not be represent- 2. Persistence and tenacity. ed in two samples taken from the same 3. Collection. source site, which may lead to the problem 4. Analysis (identification). of an untestable false-negative exclusion. 5. Interpretation. Conversely, whilst the ubiquitous component 6. Presentation. of a soil/sediment is highly likely to be rep- It will only be possible to analyse and iden- resented in all samples taken from the same tify trace physical evidence (tenet 4) if the source, it is possible that due to its ubiquity it evidence has previously divided from its par- can only provide descriptive evidence (rather ent material and transferred (tenet 1), per- than being diagnostic and/or exclusionary). To sisted and exhibited sufficient tenacity to per- complicate the matter further, we must also assist under the specific circumstances of a par- sess whether the “exotic” component is in- ticular crime event (tenet 2), and been col-

Vol. 127, N. 2 MINERVA MEDICOLEGALE 77 MORGAN FORENSIC GEOSCIENCE AND CRIME DETECTION

TABLE II.—The conceptual framework of forensic geoscience: the order of events.

Order of events Considerations Relevant precepts

1. Transfer Sample character ac- — Primary, secondary, one and two- Division of matter.74 quisition way — Pre-, syn and post-forensic event Locard 76 “Every contact leaves a tra- transfers.1 ce”.

2. Persistence — Persistence of trace physical evi- Fibres: Pounds and Smalldon 77-79 and Tenacity dence; 2-stage decay (Bull et al.) 75 Robertson and Roux;80 — Trace evidence can persist up to Glass: Koons et al. ,81 Curran et al.,82 the order of days/weeks. Brewster et al. ;83 — Many forms of trace evidence Foam: Wiggins et al. ;84 (soil, glass, fibres) have demon- Hair: Dachs et al. 85 strated significant tenacity: this is Morgan et al. 86 good for sample availability, but pre-, syn- and post-forensic events may confuse matters.

3. Collection Sampling — Comparable samples are requi- Gilbert and Pulsipher,87 Warren,88 red. Nocerino et al. ,89 Morgan et al. ,16 — Samples taken must be represen- McKinley and Ruffell 90 tative of the parent material.

4. Analysis Type of investigation — Compare and Exclude (Morgan et Seek-and-Find (Provenance) al. 16, Morgan and Bull,1 Ruffell (Rawlins et al. 91, Wiltshire,40, 50 and McKinley) 43 Brown 54) 4a. Type of sample Anthropogenic (mixed) vs natural (not mixed) 4b. Size of sample Case by case variation 4c. Which techniques? Techniques for comparison Techniques which are able to identify — If samples could have mixed pro- provenance venance it will not be possible to — Geographical indicators such as use techniques which require sam- pollen, mineralogy, quartz grain ple homogenisation for analysis. surface textures. — Exotic components may be use- — Independent forms of analysis ful (mineralogy, pollen, quartz).64 needed. — Independent analytical techniques are needed.

5. Interpretation 5a. Is there a database? Localised databases constructed from Large scale micro-detail databases hi- the analysis of samples from exclu- ghly desirable. sion locations may be adequate. 5b. Is it possible to ap- Some statistics may be possible pro- Not possible until comparator sam- ply statistics? vided that the conditions of the mul- ples are provided. tivariate statistical test are met (e.g. On occasion it may be possible to un- CDF) dertake comparison with large-scale database samples. 5c. Elimination of bias? — A number of different operators working independently on the analysis of samples (independent analysis of the same technique, independent analysis by way of different techniques). — Coding samples enable analysis without any reference to their source. 5d. Role of experimen- — Crucial to establish the nature, transfer, persistence, tenacity and method tal studies of collection (tenets 1-4) of trace evidence in order to be able to carry out appropriate analysis, interpretation and presentation (tenets 5-7). — Reproducibility of results – it is important to be able to establish the reproducibility of analysis and interpretation of a particular form of trace evidence.

6. Presentation 6a. Role of graphical — This is an ethical question as well as a methodological one. On some oc- representation of casions graphics derived from statistical analysis can provide compelling findings. confirmation of the findings.1 However, the jury are not experts and such graphics can provide a simplification of reality which may unduly sway a ‘non-expert’ jury.1

(Continued)

78 MINERVA MEDICOLEGALE Giugno 2007 FORENSIC GEOSCIENCE AND CRIME DETECTION MORGAN

TABLE II.—To be continued.

Order of events Considerations Relevant precepts 6b. Explanation to the — The results should be explained to the court with an awareness of the li- court. mitations of the analytical techniques employed. Statistical data and derived graphics will only be as good as the original data but may still over- simplify reality. — Introduction of the defenders fallacy 92, 93 must be avoided. — The expert witness must remain dispassionate when providing an opinion derived from such analysis.94 6c. The influence of The CSI effect? 95 evidence on the Negatively: Positively: court. — Can influence the jury with regard — Forensic evidence may clarify con- to the strength and reliability of tradictions of personal testimony. forensic evidence presented. — Jury may expect a greater degree — Can influence the jury with regard of evidence before declaring a ver- to the absence of forensic evi- dict. dence. — The jury may be conversant in the — Jury may not be aware of the li- perils of statistics due to media mitations of both the forensic evi- coverage of disgraced expert wit- dence presented or the fallibility of nesses (e.g. Dyer 96) and so may the expert witness presenting it. not be swayed by persuasive — May mean the jury demand un- sounding figures and arguments. reasonable levels of physical evidence in trials. lected in an appropriate manner (tenet 3). for the court. This is of course the idealised The interpretation and presentation of the scenario. In reality, there are significant com- analysis (tenets 5 and 6) will, in turn, rely on plexities concerning the transfer, persistence appropriate and accurate analysis and iden- and tenacity of trace evidence. tification. The sequence of these events are The transfer of evidence is founded upon summarised in Table II 1, 16, 40, 43, 50, 54, 64, 74, 96 the central thesis evoked by Edmond Locard and are expanded upon in the following sec- 76 that “every contact leaves a trace”. Locard tions. expounded upon this original premise by stating that “…whenever two objects come in- Acquisition of sample characteristics (tenets to contact, there is always a transfer of ma- 1 and 2) terial. The methods of detection may not be sensitive enough to demonstrate this, or the The division of matter (the fractionation decay rate may be so rapid that all evidence of sediments from the parent body) is nec- of transfer has vanished after a given time, essary for the generation of physical evi- Nonetheless, the transfer has taken place”.64 dence.74 In the forensic context it is then nec- These contacts may involve one-way, or in- essary for the transfer of evidence to take deed, two-way transfer; the latter may occur place from the forensic event site to either an- where, for example, evidence from the per- other location, or clothing or objects associ- petrator is deposited in a room and evidence ated with the perpetrator (shoes, clothing, from the room is deposited on the perpetra- vehicles etc.). It is then not only important for tor (for a recent example see Bull et al. 27). the evidence to persist upon the personal Furthermore, such contacts leading to the items associated with the perpetrator, but al- transfer of trace evidence may be both pri- so for the evidence to be recognised and col- mary and secondary. Primary transfer occurs lected. Finally, an appreciation of the tenac- when, for example, the perpetrator directly ity of this transferred evidence (normally re- makes contact with a particular source of ev- garded as trace evidence) is required before idence. Secondary transfer may occur if the an interpretation statement can be provided perpetrator, who has transferred evidence

Vol. 127, N. 2 MINERVA MEDICOLEGALE 79 MORGAN FORENSIC GEOSCIENCE AND CRIME DETECTION

upon his person from a primary transfer, the provenance of the soil/sediment sam- makes a secondary contact and transfers ev- ple (Table II). idence collected from the primary transfer In a comparative investigation the forensic onto another object or person (for example geoscientist needs to compare samples in order see Grieve and Biermann 97). that they can be excluded from having derived The persistence of trace evidence has long from the source of a comparator sample. It is been recognised as an important property very difficult to state that a sample has, or most and much work has been undertaken to de- likely has, derived from a forensic scene as termine the decay curves of various forms geo-forensics is one of the more probabilistic of trace evidence over time. The seminal sciences.65 Kiely 99 summarises the situation works of Pounds and Smalldon 77-79 and that exists in the United States of America. He Robertson et al.,98 investigating the persis- identifies the terms allowed by the courts to tence of fibres on clothing, is paralleled by support the identification of a crime scene item work undertaken on a number of different with a sample from the defendant as such: forms of trace evidence in a whole series of — match (usually limited to fingerprints studies published over the last ten years (see and ballistics); Table II under persistence precepts for a brief — compatible with; summary). The persistent nature of these dif- — consistence with; ferent forms of evidence, and particularly — similar in all respects; soils and sediments 75 can present problems — not dissimilar; for the forensic geoscientist. The problem is — same general characteristics; simply that the longer trace evidence can re- — identical characteristics; main on an object, the greater the chance — could have originated from; that there will be an amalgamation of trace — can not be excluded. evidence from pre-, syn- and post-forensic In the United Kingdom some practitioners event transfers. The inability to separate two provide a scale of 1 to 10 where 1 shows no or three phases of evidence introduction will scientific evidence and 10 shows extremely severely curtail the use of that trace evidence strong scientific evidence. Unfortunately this in the forensic context unless the multiple quantification is at best subjective and at provenances can be identified. worst, as we shall see later, biased. It is perhaps worthy to note here that the forensic Identification of sample characteristics (te- geoscientist should really present his/her nets 3 and 4) findings in the form of either “I can exclude these samples from having any association NATURE with the comparator sample” or “I can not In order to identify the characteristics of a exclude these samples from having an asso- sample, it is necessary to establish the philo- ciation with the comparator sample”. When sophical standpoint to be adopted and the the second statement is corroborated by a nature of the geoforensic investigation. In series of independent analytical techniques terms of the philosophical approach, it is at imparted upon the sample, the strength of this point in the order of events (Table II) that evidence is increased. The matter of the that the relevant philosophies of geoforen- influence of such a series of statements on the sic analysis and geological analysis diverge. jury is discussed further on. The nature of the geoforensic investigation Perhaps one of the most difficult tasks that will generally take one of two distinct forms; face the forensic geoscientist is that of the those that seek to compare and exclude so-called “seek and find” investigation. This samples that have been collected from a type of analysis involves the study of the suspect, crime scene and/or victim and those sample characteristics in order to attempt to where only samples from the suspect have provide an indication as to the provenance of been collected which require a “seek and a sample of unknown origin (for a recent ex- find” approach to provide an indication of ample see Brown 54). This is often accom-

80 MINERVA MEDICOLEGALE Giugno 2007 FORENSIC GEOSCIENCE AND CRIME DETECTION MORGAN

plished by the identification of the miner- victim to avenge, no guilty person to con- alogical and geological nature of the soil in vict and no innocent person to save. You addition to its botanical components which must bear testimony within the limits of sci- may include pollen, spores, macro and micro ence.”94 fauna (Table II). The characteristics of that In undertaking the analysis of soil/sedi- soil are then used to exclude areas which ment samples in the context of a forensic in- could not possibly be the source of the ma- vestigation it is essential that the geoforensic terial in order to refine the area for initial practitioner “remains a man of science” and searches to be made. Guesses (albeit in- adheres to an appropriate geoforensic philo- formed – but they are still guesses) concern- sophical framework (as discussed previous- ing the type of vegetation, geology, geo- ly). Once the nature of the investigation has morphology, land-use, presence or absence been identified and appropriate samples pro- of water features and nearness to roads, are vided, it will also be important to assess the all used to hone down the search area to a types of samples that have been provided number of more likely sites. These sites are for analysis. In general, there will be two then investigated one by one and soil/sedi- types of geoforensic sample; those that have ment samples collected. Exclusion principles a “natural” origin (such as those taken from are then applied until comparator samples a grave site, or a crime scene) which have in- are found and analysed that cannot be ex- variably been taken from a single source, cluded. This process has become known as and those that are “anthropogenic” in nature “environmental profiling”.38, 100 (such as samples taken from vehicles, footwear and clothing) which have the po- PRACTICE tential to be derived from a number of provenances and may include soil/sediment from It will always be important to assess each pre-, syn- or post-forensic event sources. case on its own merits, taking into account In any investigation where there is a pos- the specific auxiliary factors that are perti- sibility that anthropogenic samples may be nent. The samples that are provided and the present, it is of great importance that appro- questions asked of those samples will differ priate forms of analysis are undertaken (as and thus the practice of identifying sample will be discussed more fully later). In addition characteristics will vary accordingly at both to the type of sample presented for analysis, the collection and analysis stages. in each geoforensic case it is also important Collection.—It is crucial for the success of to assess the size or quantity of the samples. subsequent geoforensic analysis, interpreta- In order to make accurate and meaningful tion and presentation that the collection of comparisons between samples, it is crucial soil/sediment samples is carried out accu- that the size of the smallest sample is taken rately, appropriately and effectively. Not on- into account. The amount of sample avail- ly must the sample taken be representative of able will have a direct bearing upon the num- its source, but it must also be of a form that ber of techniques that it will be possible to ap- is directly comparable with samples taken ply to that sample and should also dictate from a suspect, the belongings of a suspect the order in which techniques are undertaken (Palenik’s sequence as documented in or a victim (Table II). The scope of the geo- 55 forensic analysis will be reduced if incom- Murray and Solebello ). If a particular analytical technique is destructive to the sam- parable samples are provided for analysis ple, it must be undertaken after other tech- such as fine grained materials recovered from niques have been applied to the samples in clothing for comparison with a bulk sample question. Once these factors (generally par- taken from a crime scene (for a recent ex- ticular to each individual case) have been ample see Morgan et al. 16). taken into account, it is then possible to plan Analysis.—”If the law has made you a wit- which forms of analysis are appropriate for ness, remain a man of science; you have no the samples which may also yield useful in-

Vol. 127, N. 2 MINERVA MEDICOLEGALE 81 MORGAN FORENSIC GEOSCIENCE AND CRIME DETECTION

formation, and the order in which to carry duce more problems. It is possible to analyse them out. the concentrations of vastly different element The analytical techniques that will be ap- suites; some analysis may be undertaken us- propriate and yield useful information will ing 20 or 25 elements or compounds, some vary depending upon the aim of the investi- 50 or more and in some cases even more gation (Table II). For “seek and find” inves- than that. There are no direct micro-detailed tigations it is important to carry out tech- databases which can readily be used in a niques that will yield useful information re- geo-forensic context and meaningful forensic garding the provenance of those samples. In interpretation can not be assured. It must al- a “compare and exclude” investigation there ways be remembered that these machines may be a wider range of techniques that can were developed as an analytical tool for be applied to achieve the exclusion of sam- chemists, geologists and other natural scien- ples which do not necessarily infer prove- tists; none of the machines were developed nance. for forensic analysis and hence, have inher- In its broadest sense there are two types of ent drawbacks. A recent development, how- analytical test that can be undertaken in geo- ever, that of the QemSCAN, utilises the visual forensic soil/sediment analysis. The first in- qualities of an electron microscope together volves some degree of direct visual analysis with the real-space resolution of elemental by the operator and the second involves au- and mineralogical analyses.19 The problems tomatic scanning of a sample or aliquot. This of representativeness of analysis and the latter form of analysis is popular at present, problems of the designation of “exotic” ver- because it enables rapid sequential analysis sus the ubiquitous without the aid of a per- of many samples, automated to the point that tinent database still exist. The machine, how- operator presence is not necessary. The prob- ever, does lend itself well to the possibilities lem, however, is encountered if the sample of exclusion. preparation for such analyses requires homogenisation of the samples either in pow- Interpretation – soil evidence as a probabili- 1 der, pellet or solution form. This may, there- stic science (tenet 5) fore, go some way to explain why the microscope (light or electron) is a popular “look “...we must always remind ourselves that see” method of analysis. Some of the more our system of criminal justice resides in a sophisticated modern analytical techniques world of probability”.99 enable visual selection of sample areas for The large and impressive databases which “spot analysis” and this is surely preferable to exist for fingerprints and DNA comparisons the “black-box” approach of full automation. do not, unfortunately, exist in geoforensic However, even this concession provides new soil/sediment analysis. Typically a localised problems to consider; sample analysis of a database is constructed for a particular case small area of, for example a quartz sand grain, with perhaps nine or more samples taken may provide chemical detail of adhering ce- around an incident site and many others then ment coatings. Before this can then be used taken for exclusion purposes. These are then as a comparative exclusion technique with compared to materials recovered from a su- other sand grains taken from other samples, spect or belongings of a suspect. The result a sufficient and representative area of the may be that of 60 or 70 samples taken, the grain must be analysed and all of the grains only ones that cannot be excluded from ha- from the original sample must also be ving derived, at least in part, from the foren- analysed. Far too often the occasional spot sic scene in question may be those associa- analysis, unrepresentative of the whole, is ted with a suspect. The forensic geoscientist used to exclude (or even worse, include) can go no further than to seek to exclude by samples. using as many relevant independent analyti- Multielemental analyses by automated ma- cal techniques as possible. Larger scale da- chines such as ICP or electron probe pro- tabases produced by soil surveys or geolo-

82 MINERVA MEDICOLEGALE Giugno 2007 FORENSIC GEOSCIENCE AND CRIME DETECTION MORGAN

gical surveys serve only to provide broad li- Further problems exist when applying what neation and exclusion of samples in most in- amounts to being black-box analyses to fo- stances, but can serve to concentrate analy- rensic data sets (cluster analysis, canonical sis in a particular spatial area due perhaps discriminant function analysis, principle com- to the unusual or distinctive mineral assem- ponents analysis etc.). The derived graphi- blages, pollen types etc. Smaller scale data- cal representations often look very compelling bases at a micro-detail level are only recen- for the jury but since the graphs are produced tly being constructed, but even these are of in two-dimensional format and often represent limited spatial extent.20 three-dimensional relationships these facts Once analysis has been completed, the da- can be lost at the jury level. Indeed the very ta set produced is often sought to be sum- groupings and classifications produced of- marised or given a statistical relevance by ten demonstrate fragility and are open to the application of multivariate statistical te- many vagaries as found within geological chniques. The drive for such statistical samples.101, 102 comfort stems in large part from the expe- Statistical summaries and analyses can be riences of DNA analysis.65 Here probability ru- used, however, wittingly or unwittingly, to les; many cases have fallen as a result of the influence the interpretation of a data set de- inaccurate terminology used to summarise pending on whether one is prosecuting or the probability of an association of the DNA defending. This becomes extremely perti- results. This has an unfortunate knock-on ef- nent in the British and American systems fect for forensic geoscience. Whilst we ac- where trial by jury (with the occasional ex- cept the summary statistics may be useful ception of Diplock cases) is the normal sta- and simple correlations may help the jury to te of affairs. The classic case of the Defender’s understand the significance of the results, fallacy is as pertinent to forensic geo-scienti- even these simple parametric and nonpara- sts as it is to any other form of forensic data metric tests are open to much abuse. We re- interpretation.92 member occasions in court when standard Another very important aspect of inter- deviation – should it be 2 standard devia- preting the analytical data derived from tions or 3 standard deviations was the argu- geoforensic analysis is that of bias. Whilst it ment – was determined for three samples. is widely accepted that “…an expert witness We remember commenting to the advocate should employ sound and methodologically that if there were only 2 samples the rela- rigorous reasoning processes avoiding bias tionship could have been expressed as a strai- and partisanship no matter who is paying their ght line! fees”,94, 103 in reality, truly achieving such a sta- Some forensic geoscientists utilise Pearson’s te can be difficult. It is beneficial to code correlation coefficients for vast sets of che- geoforensic samples prior to analysis and in- mical data and disregard those associations terpretation of the results. It may also be ad- that are not statistically relevant enough vantageous for not only other independent highlighting only the successes. Of course, forms of analysis to be undertaken, but for this is not good science and tends to stress both similar analyses and independent forms matching rather than excluding and with it all of analysis to be carried out by independent la- the inherent probabilities discussed pre- boratories. Corroborative results from similar viously. The lack of independent variables and independent analytical techniques will within the same data set does not appear to add weight to the conclusions reached from the deter the use of Pearson’s correlation coeffi- resultant data. This is, however, not always cient. Such is the situation in the criminal law possible in reality, but it should perhaps be court because the advocates, when alerted higher on the agenda. to these problems, choose not to challenge on Finally, the role of experimental studies in the basis of statistical relevance or rigour aiding the interpretation of geoforensic analy- lest they lose the interest and attention of the sis should be highlighted. In order to make jury. accurate and meaningful interpretations, it is

Vol. 127, N. 2 MINERVA MEDICOLEGALE 83 MORGAN FORENSIC GEOSCIENCE AND CRIME DETECTION

crucial that the nature, transfer, persistence stimonies. As discussed previously, the use of and tenacity of a form of trace evidence (te- multivariate statistics and their associated nets 1 and 2) and the most efficient methods graphical packages have the potential to aid of sample collection (tenet 3) are recogni- the interpretation of results of geoforensic sed. Indeed trace sediment analysis lends it- analysis.101 It is important, however, to be self quite properly to repeated experimental aware of the impact such graphical presen- reconstruction and much work has been un- tations of findings may have on a jury. Whilst dertaken in this area (Table II). on certain occasions they may corroborate A note of caution must be introduced re- the results of the analysis and provide a com- garding a more recent trend in experimental pelling image of the exclusions derived, in forensic geoscience. There has been a proli- another case, where similar forms of analysis feration of papers identifying the develop- were employed and similar exclusions deri- ments and capabilities of particular analytical ved, the same multivariate statistics and as- techniques intended, in the first instance, for sociated graphical presentation may not pro- geologists or chemists. The remit of many of vide such a clear and compelling image. It is, these papers is to select a number of samples, therefore, pertinent to ask whether or not some from distinctly different environments such graphical images should be used to pre- or provenances, and then analyse those sam- sent the results of geoforensic analysis to a ples using the relevant state of the art ma- jury who may not be well versed in the chine to show that the machine can indicate, strengths, weaknesses and limitations of a from the results produced, that the samples particular analytical technique and the sub- are different. However, the problems of sam- sequent statistical packages used to compu- ple representativeness, homogenisation and te the images. mixing outlined throughout this paper are The so-called “CSI effect” may have a bea- still pertinent, not withstanding the fact that ring on how the analysis and findings of fo- the original sample collection procedure or rensic geoscience are presented to the court. experimental design do not mimic the reality Whilst there have been a number of reports of a forensic case where far more subtle va- that juries may now demand unreasonable le- riations and differences exist to complicate vels of physical evidence to reach a verdict, the matter. these claims are tempered by those that sug- gest that the jury are better informed and able to understand the intricacies of the analy- Presentation (tenet 6) sis of trace physical evidence.95 The presen- “…you cannot separate, for trial purposes, tation of evidence to a court must take into forensic evidence from the testimony of fo- account the nonexpert nature of the jury all rensic experts”.99 the while being sensitive to any factors or The role of an expert witness and the evi- methods of presentation that may unduly dence that they provide in a court room is not sway or influence them. only a methodological question, but also an ethical one. It is not for the forensic geoscientist to proffer any opinion as to the guilt So what can go wrong? or otherwise of the defendant, rather it is to present their findings as impartially and So far we have provided a review of the clearly as possible.104 The issues surrounding philosophical approach necessary when dea- the presentation of evidence to a jury are ling with geoforensic collection, analysis, in- presented in Table II. terpretation and presentation of evidence in The major hurdle to overcome is how to a criminal case file. The American courts make the evidence precise and correct whil- operate a strict policy of the admissibility of st all the time making it comprehendible for evidence within a tested scientific peer-re- the jury. This is no different a problem than viewed context following the US Supreme that encountered by other forensic expert te- Court 1993 ruling of Daubert vs Merrell Dow

84 MINERVA MEDICOLEGALE Giugno 2007 FORENSIC GEOSCIENCE AND CRIME DETECTION MORGAN

TABLE III.—A replica of evidence submitted to the jury in a recent murder case in the United Kingdom.

Strength of soil and scientific evidence

10 Extremely strong Physical fit and chemical/textural match 'Exotic' particle assemblage + chemical and biological (e.g. pollen, diatoms) match.

9 Very strong Chemical, textural, colour and biological match. Or, 'exotic' particle assemblage in common.

8 Strong to very strong Chemical, mineralogical, textural and colour match.

7 Strong Chemical and biological match.

6 Moderately strong Very good chemical match.

5 Moderate evidence Very good biological match from discrete soil samples. Or, detailed chemical analysis of soil grain coatings.

4 Weak to moderate Fairly good chemical or biological match; data obtained from washed mixtures.

3 Weak Fair chemical or biological match; data obtained from washed mixtures, Or, sand grain shapes and surface textures >50 grains from discrete soil samples.

2 Very weak Quartz grain surface textures (<50 grains) or from washed samples.

1 Extremely weak Pollen or sand grain textures with small numbers of grains (<50 grains).

0 No scientific evidence Pollen or quartz grain textures with <10 grains

Pharmaceuticals. This has come to be known pendent prosecution witnesses related to pol- as the Daubert ruling. In England and Wales len analysis and quartz grain surface texture the judicial system is somewhat less deman- analysis (although a whole raft of other te- ding and a certain degree of discretion is al- chniques had been employed) in order that lowed for expert witnesses (accredited and exclusion could be established. Indeed, pol- unaccredited). len and surface texture analysis occupied the We provide here in Table III a copy of a weakest four categories of the defence ex- handout given to the jury by an expert wit- pert’s chart. It would not be surprising to ness in order to clarify the geoforensic evi- learn that the analysis undertaken by the de- dence that had been presented to the court fence expert involved the analysis of multiple by the prosecution and then challenged by chemical assays of homogenised materials. that expert in his defence capacity. It appears These figured exclusively in the strongest fi- pertinent to make some comments: firstly, ve of his designated categories. Fifthly, the inspection of Table III will show that the con- chemical analysis undertaken made no ac- cept of excluding a sample rather than mat- count for the requirement of homogenisa- ching a sample was completely ignored. The tion of the samples prior to analysis and thus word “match” appears in points 3, 4, 5, 6, 7, no account could be taken of pre- or post- 8, 9 and 10. Nowhere does the principle of murder event soil mixing on clothing, shoes exclusion occur. Secondly, the scale of 0-10 and vehicles of the defendant. Sixthly, the is arbitrary and meaningful only to that one assertion in category 8 (strong to very strong) expert. Thirdly, the descriptions and termi- of a mixture of chemical, mineralogical, tex- nology employed for each of the scale num- tural and colour match techniques fails to ac- bers is value laden and lacks scientific re- count for the possible interdependency of view. Fourthly, it is quite evident that the these analytical techniques upon one another. pertinent evidence provided by two inde- Seventhly, the use of exotic particles in ca-

Vol. 127, N. 2 MINERVA MEDICOLEGALE 85 MORGAN FORENSIC GEOSCIENCE AND CRIME DETECTION

tegories 9 and 10 fails to appreciate the re- sons and exclusions. The United States of presentativeness of these exotic particles and America has in contrast been somewhat tem- their relative rarity within a sample. There is pered in its adoption of geoforensic analysis thus an untestable false-positive or false-ne- due in part to the Daubert ruling. The deve- gative association between samples. Eighthly, lopment of geoforensic scientific analysis in in category 5 the detailed chemical analysis Australasia has been led more by biological of soil grain coatings, where spot analyses than geological applications. Perched equal- using XRD of adhering materials on individual ly between these different methods and ki- sand grains was nonrepresentative of the netics of uptake are the general European grain (only one or two areas were tested) geoforensic studies. All those employing fo- and, therefore, nonrepresentative of the sam- rensic geoscience analysis can learn from so- ple (whether it was less than 10 grains or me of the more over exuberant interpreta- greater than 50 grains). Comparison, the- tions provided by some British forensic scien- refore, between samples using this criterion tists and indeed there are potentially great only would be meaningless. Ninthly, refe- benefits of the technological revolution that rence to numbers of grains to be analysed is upon us. (categories 0-3) flies in the face of peer re- Perhaps an underlying tenet for forensic viewed published literature. Tenthly, and fi- geoscientists should be that evoked by Kirk nally, a seeming designation of the strength 61 that physical evidence and its analysis can of soil and related scientific evidence based not be wrong, it is only the interpretation upon whether a sample is washed or not that can introduce error. It would appear that seems puzzling and inappropriate in a scien- the most efficient methods of comparison tific context. still involve the physical observation by a These “ten ways to get it wrong” were scientist. Microscopes were used to great ef- taken from an actual murder case and portray fect in forensic studies in the 19th and 20th the disregard that can occur when geoscien- centuries. They will play a pivotal role in fo- tists enter forensic geoscience without caution rensic geoscience in the 21st century. In sum- and due regard to the published literature in mary, therefore, it is important that geofo- the field. There seems, in this case, to be de- rensic analysis concerns itself with the fol- fence bias against the prosecution argument lowing tenets: provided independently by two expert wit- — when undertaking comparison of sam- nesses who, without corroboration and using ples, exclusion should be sought rather than independent techniques, arrived at the same a match; general conclusions. The jury found the de- — it is important to identify the nature of fendants “guilty”. the analytical technique; that is whether it is descriptive, exclusionary or diagnostic; — it is necessary to employ a number of Synthesis independent techniques; — there is a balance sought between the The intention of this review paper was to analysis of the exotic and the ubiquitous com- highlight the development of forensic geo- ponent of a sample. science particularly within England and “…Old theories are old for a reason. They Wales. The reason this area was chosen was are robust, flexible. They have an uncanny because of the rapid expansion of publica- correspondence to reality. They may even tions and practical applications of analytical be true.”105 techniques in real crime scene situations. The This review paper has indeed presented legal jurisdiction of England and Wales is an old theories but is hopeful that it has shown ideal area for forensic geoscientific adapta- them still to be pertinent in the present day. tion; the varied geology and the delineation New applications will always build upon old between natural and anthropogenic sediment theories but must do so within the philo- inputs makes it ideal for specific compari- sophical framework of forensic geoscience.

86 MINERVA MEDICOLEGALE Giugno 2007 FORENSIC GEOSCIENCE AND CRIME DETECTION MORGAN

Riassunto tenna for ground penetrating radar applications. J Appl Geophys 2001;47:309-16. 9. Owsley DW. Techniques for locating burials, with Geoscienza forense e rivelazione del crimine. emphasis on the probe. J Forensic Sci 1995;40:735-40. Identificazione, interpretazione e presentazione nel- 10. Brilis GM, Gerlach CL, van Waasbergen RJ. Remote la geoscienza forense sensing tools assist in environmental forensics: Part I. Digital tools – traditional methods. Environ Forensics La geoscienza forense si basa su concetti ben sta- 2000;1:63-7. biliti e su tecniche analitiche sviluppate durante il 11. Brilis GM, van Waasbergen RJ, Stokely PM, Gerlach XX secolo. Recenti progressi nelle tecniche analiti- CL. Remote sensing tools assist in environmental fo- che geologiche e chimiche hanno permesso, per la rensics: Part II. Digital tools. Environ Forensics 2001;2: prima volta, analisi dettagliate del suolo e del sedi- 223-9. 12. Grip WM, Grip RW, Morrison R. Application of aerial mento raccolte particolarmente da scienziati britannici. photography and photogrammetry in environmental fo- È fondamentale una struttura concettuale per la geo- rensic investigations. Environ Forensics 2000;1:121-9. scienza forense che supporti lo sviluppo positivo nel 13. Canter D. Mapping murder: the secrets of geographi- futuro della materia, delle analisi del suolo e del se- cal profiling. London: Virgin Books; 2004. dimento. Comunque, questo articolo cerca di evi- 14. Zhivotovsky LA, Ahmed S, Wang W, Bittles AH. The fo- denziare le “trappole” incontrate in anni recenti: l’u- rensic DNA implications of genetic differentiation between endogamous communities. Forensic Sci Int tilizzo di tecniche analitiche dipendenti l’una dall’al- 2001;119:269-72. tra, la dipendenza da componenti esotici di un cam- 15. Murray RC. Evidence from the earth: forensic geology pione se sono rappresentativi o meno, la confusione and criminal investigation. Missoula, MT: Mountain tra tecniche descrittive, di esclusione o diagnostiche, Press Publishing; 2004. e l’incapacità di comprendere l’esigenza di esclude- 16. Morgan RM, Wiltshire P, Parker AG, Bull PA. The role of forensic geoscience in wildlife crime detection. re piuttosto che di accoppiare campioni di confron- Forensic Sci Int 2006;162:152-62. to. Inoltre, l’utilizzo di macchinari automatizzati, com- 17. Brown AG. The use of forensic botany and geology in puterizzati, che analizzano molteplici campioni è di- war crimes investigations in NE Bosnia. Forensic Sci Int battuto quando non vi è una valutazione visiva di un 2006;163:204-10. operatore inclusa nella procedura. L’evitare queste 18. Ruffell A, Wiltshire, P. Conjunctive use of quantitative and qualitative X-ray diffraction analysis of soils and trappole e l’adottare la struttura filosofica specifica rocks for forensic analysis. Forensic Sci Int 2004;145: per la geoscienza forense garantiranno un ruolo per 13-23. l’applicazione della geoscienza forense nel campo 19. Pirrie D, Butcher AR, Power MR, Gottlieb P, Miller GL. delle investigazioni criminali. Rapid quantitative mineral and phase analysis using automated scanning electron microscopy (QemSCAN); Parole chiave: Geoscienza forense - Filosofia - Suolo - potential application in forensic geoscience. In: Pye Sedimenti geologici K, Croft DJ editors. Forensic Geoscience - Principles, Techniques and Applications. Geological Society Special Publication No. 232. Bath: Geological Society Publishing House; 2004. p. 123-36. References 20. Bull PA, Morgan RM. Sediment fingerprints: a forensic technique using quartz sand grains. Sci Justice 1. Morgan RM, Bull PA. The philosophy, nature and prac- 2006;46:107-24. tice of forensic sediment analysis. Progr Phys Geogr 21. Morgan RM, Bull PA. The use of particle size analysis 2007;31:43-58. of sediments and soils in forensic enquiry. Sci Justice 2. France DL, Griffin TJ, Swanburg JG, Lindemann JW, 2007. In press. Davenport GC, Trammell V et al. A multi-disciplinary 22. Janssen DW, Ruhf WA, Prichard WW. Use of clay for soil approach to the detection of clandestine graves. J colour comparisons. J Forensic Sci 1983;28:773-6. Forensic Sci 1992;37:1445-58. 23. Sugita R, Marumo Y. Validity of colour examination 3. Buck SC. Searching for graves using geophysical te- for forensic soil identification. Forensic Sci Int 1996;83: chnology: field tests with ground penetrating radar, 201-10. magnetometry, and electrical resistivity. J Forensic Sci 24. Junger EP. Assessing the unique characteristics of clo- 2003;48:5-11. se-proximity soil samples: just how useful is soil evi- 4. Scott J, Hunter JR. Environmental influences on resi- dence? J Forensic Sci 1996;41:27-34. stivity mapping for the location of clandestine graves. 25. Rawlins BG, Cave M. Investigating multi-element soil In: Pye K, Croft DJ editors. Forensic Geoscience - geochemical signatures and their potential for use in fo- Principles, Techniques and Applications. Geological rensic studies. In: Pye K, Croft DJ editors. Forensic Society Special Publication No. 232. Bath: Geological Geoscience: Principles, Techniques and Applications. Society Publishing House; 2004. p. 33-8. London: Geological Society; Special Publications 5. Nobes DC. Geophysical surveys of burial sites: a case 2004;232:197-206. study of the Oaro urupa. Geophysics 1999;64: 357-67. 26. Jarvis KE, Wilson HE, James SL. Assessing element va- 6. Nobes DC. The search for ‘yvonne’: a case study of riability in small soil samples taken during forensic in- the delination of a grave using near-surface geophysi- vestigation. In: Pye K, Croft DJ editors. Forensic cal methods. J Forensic Sci 2000;45:715-21. Geoscience: Principles, Techniques and Applications. 7. Davis JL, Higginbottom JA, Annan AP, Daniels RS, London: Geological Society; Special Publications Berdal BP, Bergan T et al. Ground penetrating radar sur- 2004;232:171-82. veys to locate 1918 Spanish flu victims in permafrost. 27. Bull PA, Parker AG, Morgan RM. The forensic analysis J Forensic Sci 2000;45:68-76. of soils and sediment taken from the cast of a foot- 8. Chen CC, Rama R, Lee R. A tapered-permittivity rod an- print. Forensic Sci Int 2006;162:6-12.

Vol. 127, N. 2 MINERVA MEDICOLEGALE 87 MORGAN FORENSIC GEOSCIENCE AND CRIME DETECTION

28. Cox RJ, Peterson HL, Young J, Cusik C, Espinoza EO. 50. Wiltshire PEJ. Hair as a source of forensic evidence in Forensic analysis of soil organic by FTIR. Forensic Sci murder investigations. Forensic Sci Int 2006;163:241-8. Int 2000;108:107-16. 51. Campobasso CP, Introna, F. The forensic entomologi- 29. Bruce R, Dettmann M. Palynological analyses of st in the context of the forensic pathologist’s role. Australian surface soils and their potential use on fo- Forensic Sci Int 2001;120:132-9. rensic science. Forensic Sci Int 1996;81:77-94. 52. Gallop A, Stockdale R. Trace and contact evidence. In: 30. Bryant VM, Jones JG, Mildenhall DC. Studies in foren- White P editor. Crime scene to court: the essentials of sic palynology. In: Jansonius J, McGregor DC editors. forensic science. Cambridge: Royal Society of Chemi- Palynology: principles and applications. Dallas, TX: stry; 1998. p. 56-81. American Association of Stratigraphic Palynologists 53. Saferstein R. Criminalistics: an introduction to forensic Foundation; 1996. vol. 3, p. 957-9. science. 6th ed. Upper Saddle River, NJ: Prentice Hall; 31. Szibor R, Schubert C, Schoning R, Krause D, Wendt, U. 2004. Pollen analysis reveals murder season Nature 1998; 54. Brown T. Going to ground. Police review. 4th ed. 2000 395:449-50. 55. Murray RC, Solebello LP. Forensic examination of soil. 32. Eyring MB. Soil pollen analysis from a forensic point of In: Saferstein R editor. Forensic Science Handbook. view. Microscope 1996;44:81-97. Volume I. Upper Saddle River, NJ: Prentice-Hall; 2002. 33. Horrocks M, Coulson SA, Walsh KAJ. Forensic paly- p. 615-33. nology: variation in the pollen count of soil surface 56. Moore RC. Meaning of facies. Sedimentary facies in samples. J Forensic Sci 1998;43:320-3. geologic history. The Geological Society of America 34. Horrocks M, Walsh KAJ. Forensic palynology: assessing Memoir 39. Baltimore, MD: Waverley Press Inc; 1949. the value of the evidence. Rev Palaeobot Palynology p. 1-34. 1998;103:69-74. 57. Blatt H, Middleton GV, Murray RC. Origin of 35. Horrocks M, Walsh KAJ. Fine resolution of pollen pat- Sedimentary Rocks. Englewood Cliffs: Prentice-Hall; terns in limited space: differentiating a crime scene 1972. from an alibi scene seven metres apart. J Forensic Sci 58. Nichols G. Sedimentology and stratigraphy. Oxford: 1999;44:417-20. Blackwell Science Ltd.; 1999. 36. Horrocks M. Subsampling and preparing forensic sam- 59. Thomas M. Landscape sensitivity in time and space- an ples for pollen analysis. J Forensic Sci 2004;49:1024-7. introduction. Catena 2001;42:83-98. 37. Miller Coyle H. Forensic botany: principles and appli- 60. Selley RC. Ancient sedimentary environments and their cations to criminal casework. Boca Raton: CRC Press; sub-surface diagnosis. 4th ed. London: Chapman and 2005. Hall; 1996. 38. Mildenhall D C, Wiltshire PEJ, Bryant VM. Forensic 61. Kirk PL. Crime Investigation. 2nd ed. New York: Wiley; palynology: why do it and how it works. Forensic Sci 1974. Int 2006;163:163-72. 62. Popper K. The logic of scientific discovery. New York: 39. Montali E, Mercuri AM, Grandi GT, Accorsi CA. Towards Basic Books; 1959. a “crime pollen calendar”- Pollen analysis on corpses throughout one year. Forensic Sci Int 2006;163:211-23. 63. Rhoads BL, Thorn CE. The scientific nature of geo- 40. Wiltshire PEJ. Consideration of some taphonomic va- morphology. New York: Wiley; 1996. riables of relevance to forensic palynological investi- 64. Murray RC, Tedrow JCF. Forensic geology: earth scien- gation in the United Kingdom. Forensic Sci Int 2006; ce and criminal investigation. New Brunswick, NJ: 163:173-82. Rutgers University Press; 1975. 41. Horswell J, Cordiner SJ, Maas EW, Martin TM, 65. Broeders APA. Of earprints, fingerprints, scent dogs, cot Sutherland BW, Speir TW. Forensic comparison of soils deaths and cognitive contamination – a brief look at the by bacterial community NDA profiling. J Forensic Sci present state of play in the forensic arena. Forensic 2002;47:350-3. Sci Int 2006;159:148-57. 42. Dawson LA, Towers W, Mayes RW, Hillier S, Fraser A, 66. Pye K, Croft DC. Forensic geoscience: introduction Craig J et al. Use of plant wax signatures in under- and overview. In: Pye K, Croft DJ editors. Forensic standing soils. Proceedings of the Clay Minerals Group geoscience: principles, techniques and publications. of the Mineralogical Society and the Forensic Science London: Geological Society; Special Publications Society - Trace Metals, Isotopes and Minerals in Forensic 2004;232:1-5. Science 30th October 2003, London. 67. Pye K. 2004 Forensic Examination of rocks, sediments, 43. Ruffell A, McKinley J. Forensic Geoscience: applica- soils and dusts using scanning electron microscopy tions of geology, geomorphology and geophysics to cri- and X-ray chemical microanalysis. In: Pye K, Croft DJ minal investigations. Earth Sci Rev 2005;69:235-47. editors. Forensic geoscience: principles, techniques 44. Funayama M, Aoki Y, Sebetan IM, Sagisaka K. Detection and publications. London: Geological Society; Special of diatoms in blood by a combination of membrane fil- Publications 2004;232:103-22. tering and chemical digestion. Forensic Sci Int 1987;34: 68. Lombardi G. The contribution of forensic geology and 175-82. other trace evidence analysis to the investigation of 45. Karkola K, Neittaanmaki H. Diagnosis of drowning by the killing of Italian Prime Minister Aldo Moro. J investigation of left heart blood. Forensic Sci Int 1981; Forensic Sci 1999;44:634-42. 18:149-53. 69. Hunter JR, Brickley MB, Bourgeois J, Bouts W, 46. Gruspier KL, Pollanen, MS. Limbs found in water: in- Bourguignon L, Hubrecht F et al. Forensic archaeo- vestigation using anthropological analysis and the dia- logy, forensic anthropology and human rights in tom test. Forensic Sci Int 2000;112:1-9. Europe. Sci Justice 2001;41:173-8. 47. Piette MHA, De Letter EA. Drowning: still a difficult 70. Croft DJ, Pye K. Multi-technique comparison of sour- diagnosis. Forensic Sci Int 2006;163:1-9. ce and primary transfer soil samples: and experimen- 48. Singh R, Singh R, Kumar S, Thakar MK. Forensic analy- tal investigation. Sci Justice 2004;44:21-8. sis of diatoms – a review. Anil Aggrawaal’s Internet 71. Pye K, Blott S, Croft DJ, Carter JF. Forensic comparison Journal of Forensic Medicine and Toxicology 2006;7. of soil samples: assessment of small-scale spatial va- 49. Mildenhall DC. An unusual appearance of a common riability in elemental composition, carbon and nitrogen pollen type indicates the scene of the crime. Forensic isotope ratios, colour, and particle size distribution. Sci Int 2006;163:236-40. Forensic Sci Int 2006;163:59-80.

88 MINERVA MEDICOLEGALE Giugno 2007 FORENSIC GEOSCIENCE AND CRIME DETECTION MORGAN

72. Morgan RM, Freudiger-Bonzon J, Bull PA. Elemental 87. Gilbert RO, Pulsipher BA. Role of sampling designs in analysis of soil samples for forensic purposes by in- obtaining representative data. Environ Forensics ductively coupled plasma spectrometry – precision 2005;6:27-33. considerations. A Comment. Forensic Sci Int. In press. 88. Warren J. Representativeness of environmental sam- 73. Morgan RM. Forensic geoscience: sedimentary mate- ples. Environ Forensics 2005;6:21-5. rials in forensic enquiry. Thesis. Oxford University 89. Nocerino JM, Schumacher BA, Dary CC. Role of la- Centre for the Environment. Oxford: University of boratory sampling devices and laboratory subsam- Oxford; 2006. pling methods in representative sampling strategies. 74. Inman K, Rudin N. The origin of evidence. Forensic Sci Environ Forensics 2005;6:35-44. Int 2002;126:11-6. 90. McKinley J, Ruffell A. Contemporaneous spatial sam- 75. Bull PA, Morgan RM, Sagovsky A, Hughes GJA. The pling at scenes of crime: advantages and disadvanta- transfer and persistence of trace particulates: experi- ges. Forensic Sci Int. In press. mental studies using clothing fabrics. Sci Justice 91. Rawlins BG, Kemp SJ, Hodgkinson EH, Riding JB, 2006;46:185-95. Vane CH, Poulton C et al. Potential and pitfalls in esta- 76. Locard E. Analyses of dust traces parts I, II and III. Am blishing the provenance of earth-related samples in fo- J Police Sci 1930;1:276-98, 401-18 and 496-514. rensic investigations. J Forensic Sci 2006;51:832-45. 77. Pounds CA, Smalldon KW. The transfer of fibres 92. Thompson WC, Schumann EL. Interpretation of stati- between clothing materials during simulated contacts stical evidence in criminal trials. Law Hum Behav and their persistence during wear. Part I: fibre tran- 1987;11:167-87. sference. J Forensic Sci Soc 1975;15:17-27. 93. Aitken CG. Statistics and the evaluation of evidence for 78. Pounds CA, Smalldon KW. The transfer of fibres forensic scientists. Chichester: John Wiley and Sons Ltd; between clothing materials during simulated contacts 1995. and their persistence during wear. Part II: fibre persi- 94. Hiss J, Freund M, Kahana T. The forensic expert wit- stence. J Forensic Sci Soc 1975;15:29-37. ness – An issue of competency. Forensic Sci Int. In 79. Pounds CA, Smalldon KW. The transfer of fibres press. between clothing materials during simulated contacts 95. Houck MM. CSI: Reality. Scie Am 2006;206:84-9. and their persistence during wear. III. A preliminary in- 96. Dyer C. Professor Roy Meadow struck off. BMJ vestigation of the mechanisms involved. J Forensic Sci 2005;331:177. Soc 1975;15:197-207. 97. Grieve MC, Biermann TW. Wool fibres – transfer to 80. Robertson J, Roux C. Transfer and persistence. In: vinyl and leather vehicle seats and some observations Siegel JA, Saukko PJ, Knupfer GC editors. Encyclopedia on their secondary transfer. Sci Justice 1997;37:31-8. of Forensic Sciences. London: Academic Press; 2000. p. 98. Robertson C, Kidd BM, Parkinson HM. The persistence 834-8. of textile fibres transferred during simulated contacts. 81. Koons RD, Buscaglia J, Bottrell M, Miller ET. Forensic J Forensic Sci Soc 1982;22:353-60. glass comparisons. In: Saferstein R editor. Forensic 99. Kiely TF. Forensic evidence: science and criminal Science Handbook. 2nd ed. Englewood Cliffs: law.2nd ed. Boca Raton: CRC Press; 2006. Regents/Prentice Hall; 2002. 100. Wiltshire PEJ. Environmental profiling and forensic 82. Curran JM, Hicks TN, Buckleton JS. Forensic interpre- palynology: background and potential value to the tation of glass evidence. London: CRC Press; 2000. criminal investigator. Handbook for the National Crime 83. Brewster F, Thorpe J, Gettinby G, Caddy B. The re- and Operations Faculty with The British Association for tention of glass particles on woven fabrics. J Forensic Human Identification. 1993. Sci 1985;30:798-805. 101. Morgan RM, Bull PA. Data interpretation in forensic se- 84. Wiggins KG, Emes A, Brackley LH. The transfer and diment geochemistry. Environ Forensics 2006;7:325-34. persistence of small fragments of polyurethane foam 102. Isphording WC. Statistics in court: the right and wrong onto clothing. Sci Justice 2002;42:105-10. ways. In: Pye K, Croft DJ editors. Forensic geoscien- 85. Dachs J, McNaught LJ, Robertson J. The persistence of ce - principles, techniques and applications. Geological human scalp hair on clothing fabrics. Forensic Sci Int Society Special Publication No. 232. Bath: Geological 2003;138:27-36. Society Publishing House; 2004. p. 281-8. 86. Morgan RM, Bull PA, Cohen J, Murly-Gotto J, O’Connor 103. Frecklton I. Regulating forensic deviance: the ethical R. The persistence and tenacity of trace particulates as responsibilities of expert report writers and witnesses. found on clothing and motor vehicle seats. An expe- J Law Med 2004;12:141-9. rimental approach using glass fragments, fibres, soils 104. Barnett PD. Ethics in forensic science: professional and anthropogenic materials. The 4th European standards for practice of criminalistics. London: CRC Academy of Forensic Sciences 2006 Conference, 2006, Press; 2001. June 13-16, Helsinki. 105. Horgan J. The end of science. London: Abacus; 1998.

Vol. 127, N. 2 MINERVA MEDICOLEGALE 89