CHANGES IN ASSEMBLAGES J.B. Riding, B.G. Rawlins, and SOIL K.H. Coley:POLLEN Changes in soil pollen assemblages on footwear worn at different sites ON FOOTWEAR WORN AT DIFFERENT SITES

135

JAMES B. RIDING BARRY G. RAWLINS British Geological Survey Kingsley Dunham Centre Keyworth Nottingham NG12 5GG United Kingdom e-mail: [email protected]; [email protected] KIRSTIN H. COLEY Department of Geography Royal Holloway University of London Egham Surrey TW20 0EX United Kingdom

Abstract The application of palynology to forensic investigations relies on the similarity of pollen assemblages from forensic items, such as footwear, with control samples from a crime scene. The pollen from material adhering to footwear is likely to reflect some combination of pollen from the locations where the boots/shoes have been worn most recently. This study investigated the changes in pollen assemblages on footwear that had been worn at different sites. Six rural sites in the East Midlands of England, United Kingdom were visited wearing pristine boots (i.e. no mixing), and boots that were previously worn at other localities (i.e. potential mixing). Samples of adherent soil from these items of footwear, and control samples, were analysed palynologically in order to assess the degree and significance of mixing of the pollen assemblages. With the exception of one sample, the pollen adherent to footwear or in the soil samples from each of the six sites (no mixing) had a characteristic signature. This supports the general distinctiveness of pollen from individual sites, the concept of widespread palynological heterogeneity, and the utility of palynology in forensic geoscience. The data from this study show that when mixing occurs from wearing footwear at different sites, the pollen/spore content of the boots etc. dominantly reflects that of the last site. This was evident from a visual examination of the raw data, and was confirmed using detrended correspondence analysis applied to the eleven dominant taxa. These data showed clustering of samples based on the last site visited. The more abundant the pollen/spores, the closer the samples were clustered. The clustering was less convincing at localities that yielded relatively sparse palynomorphs. However, sample material from footwear that was potentially contaminated with soil from previous localities typically exhibited some subtle differences. These were normally slight increases in diversity, and small variations in certain pollen types. The relative insignificance of these differences means that they would be difficult to discern consistently and quantify. It is thus critical that, in relevant forensic investigations, footwear belonging to suspects is seized as soon as possible after a crime is committed.

Key words: forensic palynology; soil analysis; provenance determination; multivariate statistics.

INTRODUCTION Forensic palynology has been used in many criminal cases to associate the pollen/spore assemblage from clothing, fabrics, or footwear belonging to a suspect with a crime scene, or other locations associated with an investigation such as a body deposition site (Milne et al., 2005; Bryant and Jones, 2006). Palynology is the study of pollen, spores and other organic remains (palynomorphs) that can be

Palynology, 31 (2007): 135–151 © 2007 by AASP Foundation

ISSN 0191-6122

either modern or fossil (Moore et al., 1991; Jansonius and MacGregor, 1996; Traverse, 2007). Palynomorphs are abundant, chemically/mechanically robust, and small; they are therefore relatively ubiquitous. By far the most important palynomorphs in forensic studies are the two main terrestrial groups, pollen and spores. Forensic palynology generally assumes that due to contact between a suspect and the ground, other surfaces, or vegetation, an adherent pollen assemblage from a forensic

136

sample will be distinctive for a particular location. This is because every locality apparently has a distinctive palynological profile due to the huge variability in vegetation spectra and taphonomic factors, both of which affect the distribution of pollen and spores (Wiltshire, 2006). Comparison of control samples from the site of investigation and other locations with forensic sample(s) ought, therefore, to establish the probability that a forensic sample came from a specific crime scene. Despite its use in criminal investigations, few studies have been undertaken to test some of the basic assumptions upon which the application of forensic palynology are based. One study tested whether pollen/spores in soil samples from a relatively small (localized) area, i.e. comparable to a crime scene, exhibited significant variation (Horrocks et al., 1998). The palynological comparability of surface soil samples from other localized areas of similar vegetation type was also tested (Horrocks et al., 1998). In this instance, surface soil samples from the principal control site were dominated by grass pollen and bracken spores, and overall these samples had similar pollen/spore contents, which was demonstrated statistically (Horrocks et al., 1998). However, the pollen/spore associations from the control site differed considerably from other sites with similar vegetation types. In another study, soil samples were collected from consecutive footprints made by clean shoes within a localized area and their pollen content analyzed (Horrocks et al., 1999). The resultant data were compared with pollen and spore associations from the shoes that made the prints. In this experiment, soil samples from and between the prints, and from two soil samples from the shoes indicated a homogenous pollen/spore assemblage. It was clearly demonstrated that ‘perfect matches’ do not occur, because minor differences within this sample set were present (Horrocks et al., 1999). One of the potential limitations to the application of forensic palynology is the dilution or mixing of the pollen assemblage from a crime scene with that from sites visited both before and after the alleged crime. It is inevitable that the pollen association taken from a shoe or boot will never perfectly match any specific locality because of the effectiveness of footwear at picking up pollen grains (Wiltshire, 2006). Unless items of clothing or footwear are seized from a suspect immediately after a crime has been committed, the adherent pollen assemblage from the crime scene, and any pollen that was present before the crime scene was visited, will be mixed with pollen in material from any sites subsequently attended. However it has been suggested that pollen is efficiently retained on footwear over considerable periods of time, even if the items are cleaned (Wiltshire, 2006). This mixing of pollen will depend on numerous

PALYNOLOGY, VOLUME 31 — 2007

characteristics of the clothing/footwear, and the ground or surfaces at each site (e.g. moisture content and soil texture), and the nature of the contact between them. It is possible that pollen/spores from several localities will adhere to footwear or clothing. An ideal, but unlikely, scenario would be for a criminal to wear a new pair of boots, with deep tread, at a damp site with clay-rich soil that adheres easily. It has been assumed that the pollen on clothing or footwear will predominantly reflect the assemblage at the site that was visited last, but this has not been scientifically tested to date. In this contribution, results are presented from an investigation into the changes in the pollen assemblages of material adhering to footwear that has been worn at sites with differing pollen assemblages. Pristine footwear was worn at each of six rural sites, and the adherent material subsampled before the footwear was worn at subsequent sites, and further subsamples taken. Control soil samples were also collected from each site to determine its typical pollen assemblage. Then each of these samples were analysed for their pollen assemblage. In doing so, the effect of mixing pollen/spores from soil at different sites over time, with the pollen/spore assemblages from individual sites were investigated. The results were analysed statistically in order to assess the degree of similarity between the pollen/spore assemblages at the six sites, and potential mixtures between these sites. The implications of these findings for the application of forensic palynology are discussed. MATERIAL AND METHODS General Strategy and Sampling Six rural sites in the East Midlands of England, United Kingdom were visited, where pristine boots were worn (Text-Figure 1; Appendix 1). With each new locality, boots were also used that had been worn at previous localities within two ‘pyramid’ structures (Text-Figure 2). Each ‘wearing’ comprised one of us (JBR) walking normally, and in random directions, for one minute at the specific locality within a small area (ca. 9 m3). The items of footwear were not worn between localities. Composite control soil/surface samples were also taken from each locality. Five subsamples were collected from a ca. 3 m transect at each locality, and thoroughly mixed. This subsampling strategy aims to provide a representative sample by avoiding the skewing of a single sample due to the proximity of a localized, particular highly polleniferous plant. The subsampling procedure followed the recommendations of Adam and Mehringer (1975). The control subsamples were collected by carefully scraping the up-

J.B. Riding, B.G. Rawlins, and K.H. Coley: Changes in soil pollen assemblages on footwear worn at different sites

20

N

60

40

80

5 00

137

4 00

10km 80

60 Nottinghamshire

d

40

Wo llato n drain

Industria l Wollaton Museum Hall 6 Wo llaton Park

Thompson’s Wood The Lake Beesto n

100m

100m allotment s

3

allotment s Ruddington 4

Museum Lake 5 Rushcliffe Country k Par

Works

2

e 20

Rhododendron Hill The Rough Nanpantan Charn Hill wood Quarry (Granite) 1 Five Tree Plantation Rose Cottag e 100m

Roe’ sn Plantatio

c Leicestershire

3 00

100km

a

80

b

Wollaton Rhododendron Hill Industrial Museum

The Rough

Wollaton Hall

ain

Charnwood Quarry (Granite)

dr

Nanpantan Hill

1

6 Wollaton Park

Thompson’s Wood

Five Tree Plantation Rose Cottage

The Lake

Beeston

Roe’s Plantation

100m

c

100m

d 100m

allotments

allotments

Ruddington

4

Museum

3 Works

The Lake

2

5 Rushcliffe Country Park

e

Text-Figure 1. Locality map of sample locations 1 to 6 within the East Midlands of England, U.K. a – a small scale map of England with the East Midlands indicated. b – a map of the counties of Nottinghamshire and Leicestershire illustrating the three areas where samples were taken. c – a detailed map illustrating location 1 near a wood, southwest of Nanpantan, Leicestershire. d – a detailed map illustrating location 6 at the margin of Wollaton Park Lake, Nottingham. e – a detailed map illustrating locations 2, 3, 4, and 5 at Ruddington Country Park, Nottingham.

138

Text-Figure 2. Pollen dilution triangle for sites 1 to 4 illustrating the pure (bulk) soil samples in the left column, and the boot samples in the triangle to the right. The triangle indicates that the potential pollen dilution via contamination increases to the left and towards the base. Sample code 1234B (number 12) is hence potentially the most palynologically heterogeneous. The numbers refer to the sites visited; B denotes boot and S denotes soil.

permost 5 mm of surface materials using a clean knife (see also Appendix 2). The samples of adherent soil from the items of footwear and the control samples were analysed palynologically to attempt to determine if the pollen/ spore content becomes diluted by wearing the footwear at other sites. The control samples provide essential data for comparison, even with adherent material taken from the pristine footwear, which should have a very similar composition. The collecting and sampling protocols are described in detail above, and also in Appendix 2. A cumulative strategy was adopted within two ‘pyramid’ structures. In the first, and larger, ‘pyramid’, only the first pair of boots was worn at locality one, the first and second pairs were worn at locality two, the first, second and third pairs were worn at locality three, and the first, second, third and fourth pairs of footwear were worn at locality four (Text-Figure 2). In the second, smaller, ‘pyramid’, the fifth pair of boots were

PALYNOLOGY, VOLUME 31 — 2007

worn at both sites. Samples of the adherent soil on the footwear were taken as appropriate, together with control samples (Text-Figure 2; Appendices 1, 2). The point of this strategy is to determine the magnitude of the dilution or mixing of the pollen/spore signals, assuming that palynological dilution is a real phenomenon. The footwear used comprised sturdy ‘outdoor’ boots with a maximum tread depth of between 7 and 9 mm deep. The six pairs used were all new, and were thoroughly cleaned using a surfactant before the fieldwork. Similar footwear was used throughout this study. Representative examples of the footwear used are illustrated in Text-Figure 3. In the sampling/subsampling from the footwear, the adherent mud/soil was carefully scraped away into a new sampling bag using a clean scalpel. No attempt was made to differentiate a stratigraphy of the adherent materials, all the mud/soil down to the sole was taken using the scalpel blade. In some cases, several subsamples were taken from each boot. The subsamples were taken from discrete areas of the tread that are described in Appendix 1. It was assumed that the pollen content of the adherent materials is relatively homogenous over the entire sole of the boot, and that any bias introduced during subsampling was minimal. The times between the initial wearing, sampling/ subsampling, and re-wearing where appropriate are also

Text-Figure 3. Photographs of two pairs of clean boots used in this study, illustrating the treads. The left hand photograph is of pair 3, ‘Dr Martens’ black 7-eye safety boots. The right hand photograph is of pair 4, ‘Century’ wellington type boots.

J.B. Riding, B.G. Rawlins, and K.H. Coley: Changes in soil pollen assemblages on footwear worn at different sites

documented in Appendix 1. These data are important because these parameters may determine how much drying there was, and whether any material became detached from the footwear. It is acknowledged that, in forensic cases, several different techniques are used for obtaining pollen from footwear exhibits. These include washing or wiping the pollen from the upper part of the boot or shoe, and washing pollen from the laces separately. Pollen from the uppers and/or laces is often a good indicator of the last place visited, and also if the owner has walked through crops, grass, herbs, other vegetation, or in an open area. Furthermore, it is possible to dissect specific layers of adherent material. This experiment simply used all the adherent materials from the soles of the footwear used, and washing was not undertaken. The reason for this strategy is that the primary objectives were concerned with testing the transference of pollen in soil/ surface materials taken from footwear and their dilution potential, and that the resources available were limited. There is clearly much scope for future research in this area. Sample Preparation and Study The 18 samples were prepared in the British Geological Survey palynology laboratory using standard palynological preparation procedures (Moore et al., 1991). They were demineralized using hydrochloric acid (HCl) and hydrofluoric acid (HF), the residual mineral grains were removed using heavy liquid separation, and the residues subjected to acetolysis. The heavy liquid used was zinc bromide with a specific gravity of 2.45–2.52. The microscope slides were mounted using the permanent mounting medium Elvacite. This is a permanent mounting medium and key grains can easily be given a coordinate on the coverslip and reexamined. The disadvantage of a permanent mountant is that problematic pollen grains cannot be moved in order to help identifications. If the mounting medium is viscous, e.g. glycerine or silicone oil, the grains can readily be manipulated. However the grains can move in a viscous mounting medium and relocation can be a significant problem. In a legal case, a key pollen grain crucial to the investigation must be readily relocatable so that it can be verified by both the legal teams. Sufficient slides were produced in order to allow statistically significant pollen counts to be made. The palynomorphs were studied using an Olympus CH2 transmitted light microscope. Pollen grains were identified using standard European keys (e.g. Moore et al., 1991), and the pollen reference collection in the Department of Geography, Royal Holloway, University of London, United Kingdom. The palynomorphs identified in this study are listed in Appendix 3.

139

Statistical Analysis From the full dataset, only those pollen types that accounted for more than 1% of the total pollen count were selected and these eleven types are highlighted in Table 1, and listed in Table 2. The matrix of 18 samples with a subset of 11 pollen types was analysed using Detrended Correspondence Analysis (DCA). This overcomes the arch/ horseshoe effect. This phenomenon, and the tendency to concentrate/compress the end of the ordination axis, are the major problems associated with Correspondence Analysis (CA), which is often used for the analysis of ecological data (Hill and Gauch, 1980). Detrended Correspondence Analysis compensates for this by stretching and straightening the data plots. The DCA was undertaken using the ‘vegan’ package in an R-mode statistical environment (Oksanen et al., 2007). The minor (<1%) components were not subjected to DCA because, if large numbers of trace taxa were included, the results would be skewed such that this technique would not produce coherent plots and hence would be of limited use. It should be borne in mind however, distinctive pollen types with low frequencies are potentially crucial in a legal sense. These may be of critical importance in providing key evidence that a person visited a specific locality. RESULTS The 18 samples are listed in Appendix 1 and the resultant pollen data is presented in Table 1 and Text-Figure 4. Minor pollen types are those which account for less than 1% of the total pollen count in all samples; these are subsequently referred to as trace components. Site 1, Scrubland Near Wood, Southwest of Nanpantan, Leicestershire, U.K. (samples 1 and 2) This rural locality (Text-Figures 1c, 5) is an area of rough scrubland and has a flora of small birch and oak trees, ferns, brambles, and herbs including grasses, nettles, thistles, and other weeds. Stands of mixed deciduous woodland surround the locality. Pollen from samples 1 and 2 (i.e. pure locality 1) is of relatively low diversity (13 and 16 taxa respectively) and is dominated by Poaceae, with common Lactuceae, Pinus, and Cyperaceae (Text-Figure 4). Trace components include Ambrosia, Aster type, Betula, Caryophyllaceae, Cedrus, cf. Centaurea type, Picea, Plantago type, Quercus, Tilia, undifferentiated broken saccate pollen, and undifferentiated trilete spores (Table 1). The pollen is rich in anemophilous (wind-pollinated) forms such as Pinus and Poaceae, and entomophilous (insect-pollinated) types such

Table 1. The pollen/spore raw data for the 18 samples in this study. The palynomorph taxa are set out in columns and the samples are in rows. The order of taxa is the same as in Appendix 3. The numbers in the cells refer to the numbers of grains counted. Three dots (…) indicates that the respective taxon is absent. The numbers for the eleven pollen types which are in bold font are the ones which account for over 1% of the total pollen/spore count (see Table 2).

140 PALYNOLOGY, VOLUME 31 — 2007

J.B. Riding, B.G. Rawlins, and K.H. Coley: Changes in soil pollen assemblages on footwear worn at different sites

Table 2. Pollen type (not species or sample) scores from the first two ordinates of the Detrended Correspondence Analysis, and summary statisticsfrom the analysis.

141

in sample 2 than sample 1; the remaining rare pollen types occur in similar numbers. The diversity of rare pollen types is somewhat higher in sample 2; Ambrosia, Cedrus, cf. Centaurea type, and undifferentiated trilete spores were recorded only in this sample. The only unique occurrence in sample 1 was a single grain of Tilia pollen. These results, unsurprisingly, support the expectation that material adhering to footwear that has only been worn at one locality will have a virtually identical pollen signature as the locality itself. This is further supported by multivariate statistical analysis (Text-Figure 6). The dominance of Poaceae at site 1 does not uniquely characterize the site. However of all the samples investigated as part of this study, the dominance of Poaceae together with common Lactuceae, Pinus, and Cyperaceae, in the complete absence of Alnus and the rarity of Plantago type is a characteristic signature. The rare components in both samples are also relatively similar. Site 2, Elder/Hawthorn Woodland at Ruddington Country Park, Nottingham, U.K. (samples 3, 4 and 5)

as Lactuceae. This assemblage reflects the natural vegetation at and around this site, which is a mixture of typical scrubland trees and herbs. Sample 1 yielded more Lactuceae and Cyperaceae, and less Pinus and Poaceae than sample 2 (Text-Figure 4; Table 1). Quercus pollen is more common

This locality is adjacent to a footpath in a large country park (Text-Figure 1e). The flora comprises deciduous trees and shrubs including elder and hawthorn with brambles and stands of nettles (Text-Figure 7). The surrounding areas are fields.

Text-Figure 4. Percentages of pollen in each sample for the 11 dominant taxa (i.e. 1% and above of the total pollen in all samples). The samples are presented in sequence, i.e. 1-18. In the sample codes, the numbers refer to the sites visited in the sequence shown; B refers to boot, and S refers to soil. Therefore 123B means the sample was taken from a boot that had visited sites 1, 2, and 3. The control soil sample taken from site 6 is 6S.

142

Text-Figure 5. Site 1, scrubland near wood, southwest of Nanpantan, Leicestershire, U.K.

This locality was the sparsest palynologically (Table 1). The reason for this is not clear; it may be due to the churning of the soil by animals such as rabbits and badgers, thereby promoting the mixing and rapid oxidation of the soil. The pollen floras from samples 3, 4 and 5 are substantially similar; the diversity is moderate, with 17, 17 and 21 taxa recorded respectively. These are overwhelmingly dominated by Pinus and Poaceae (Text-Figure 4; Table 1). Samples 4 and 5 respectively produced less and more Poaceae than sample 3 (Text-Figure 4). Moderate levels of Cyperaceae and Lactuceae are present, however other forms are present in extremely low proportions (Text-

PALYNOLOGY, VOLUME 31 — 2007

Figure 4; Table 1). Trace levels of Acer, Alnus, Artemisia, Aster type, Brassicaceae, Caryophyllaceae, Chenopodiaceae/Amaranthaceae, Corylus, Fraxinus, Hedera helix, cf. Littorella, Picea, Plantago type, Polypodiaceae, Quercus, Ranunculus type, Salix, Sinapis cf. arvensis, Trifolium, undifferentiated broken saccate pollen, and Ulmus were recorded (Table 1). The diversity of rare pollen types is somewhat higher in sample 5. Unique occurrences in sample 5 are Artemisia, Corylus, cf. Littorella, and Sinapis cf. arvensis. Fraxinus and Salix are only present in sample 4. Rare forms found in all three samples are Caryophyllaceae, Quercus, and Polypodiaceae. The pollen is dominated by anemophilous plant types such as Pinus and Poaceae; entomophilous types such as Aster type, Lactuceae, and Ranunculus type are subordinate. This assemblage reflects the natural vegetation at this site and its environs, which comprise scrubland trees with some herbs. The pollen/spore associations from pure samples from site 2 (i.e. samples 3 and 4) are significantly similar in both taxonomic spectrum and relative proportions. As at site 1, this observation supports the expectation that soil adhering to footwear only worn at one locality will have a virtually identical pollen signature to the parent soil. This is supported by multivariate statistical analysis (Text-Figure 6). This site is characterized by prominent Pinus and Poaceae, low numbers of Alnus, Cyperaceae, Lactuceae, Polypodiaceae, Quercus, and Ranunculus type, and the complete absence of Plantago type and cf. Primula. Both the prominent and rare components are similar in the footwear sample and the control.

Text-Figure 6. Sample scores from the first two ordinates of the detrended correspondence analysis. In the sample codes, the numbers refer to the sites visited in the sequence shown; B refers to boot, and S refers to soil. Therefore 1234B means the sample was taken from a boot that had visited sites 1, 2, 3, and 4. The control soil sample taken from site 4 is 4S.

J.B. Riding, B.G. Rawlins, and K.H. Coley: Changes in soil pollen assemblages on footwear worn at different sites

Text-Figure 7. Site 2, elder/hawthorn woodland at Ruddington Country Park, Nottingham, U.K.

Sample 5 potentially includes material from both sites 2 and 1. The gross similarity to samples 1 and 2 (Text-Figure 4; Table 1) indicates that the majority of the material is from site 2, i.e. the last site visited. However there are some differences, which may be attributable to input (i.e. contamination) from site 1. For example, the diversity is slightly higher and there is more Acer, Brassicaceae, Lactuceae, Picea, Plantago type, Poaceae, Quercus, and undifferentiated broken saccate pollen in sample 5 than in samples 3 and 4 (Text-Figure 4; Table 1). Furthermore, at this site, the occurrences of Artemisia, Corylus, Sinapis cf. arvensis, and cf. Littorella are unique to sample 5. Sample 5 yielded significantly elevated levels of several pollen types that are relatively common at site 1 and therefore may be derived from there. These are Lactuceae, Picea, Plantago type, Poaceae, Quercus, and undifferentiated broken saccate pollen. It is considered that these differences in both diversity and numbers recorded indicate input from both sites 2 and 1 in sample 5.

143

levels of Aster type, Cyperaceae, Lactuceae, Pinus, and undifferentiated broken saccate pollen are present. Furthermore Alnus, Apiaceae, Brassicaceae, Betula, Caryophyllaceae, Chenopodiaceae/Amaranthaceae, Lycopodium, Malva, Myriophyllum, Picea, Plantago type, Quercus, Ranunculus type, Sinapis type, and Tilia were recorded in low proportions (Text-Figure 4; Table 1). The diversity of rare pollen types is low in sample 6 in comparison to samples 7, 8, and 9. Unique occurrences are that of Tilia in sample 7, Quercus, Lycopodium, and Myriophyllum in sample 8, and Caryophyllaceae and Ranunculus type in sample 9. The only rare form found in all three samples is Alnus. The dominance of Poaceae pollen reflects the presence of abundant grasses in the plantation, and in the surrounding park. Pinus is also unsurprisingly present, together with a variety of other tree and herb pollen. This association is more diverse than would have been predicted from the overall flora. This probably reflects significant transportation of pollen into this site. The pollen/spore associations from the two pure samples from site 3 (i.e. samples 6 and 7) are remarkably similar in diversity, taxonomic spectrum, and relative proportions. As before, this phenomenon demonstrates that soil on pristine footwear normally yields a virtually identical pollen assemblage to the parent soil; this is also indicated by statistical analysis (Text-Figure 6). The site is characterized by dominant Poaceae, with subordinate Alnus, Aster type, Betula, Cyperaceae, Brassicaceae, Lactuceae, Picea, Pinus, and undifferentiated broken saccate pollen. Both prominent and rare components are similar, and Acer, Hedera helix, and cf. Primula are entirely absent. Samples 8 and 9 are susceptible to contamination from material from sites 2 and 1. The overall similarities

Site 3, Pine Plantation at Ruddington Country Park, Nottingham, U.K. (samples 6, 7, 8 and 9) This locality is a pine plantation in a large country park (Text-Figure 1e). The flora is dominated by pine trees and grasses, with some other herbs/weeds (Text-Figure 8). All four samples from site 3 yielded abundant pollen/spore associations (Text-Figure 4; Table 1). The floras from this site are all substantially similar; in particular they are all overwhelmingly dominated by Poaceae pollen (Text-Figure 4). However, less Poaceae pollen was counted in samples 8 and 9 in comparison to numbers 6 and 7. The diversity is generally relatively low; samples 6, 7, 8 and 9 produced 10, 12, 14 and 16 taxa respectively. Moderate

Text-Figure 8. Site 3, pine plantation at Ruddington Country Park, Nottingham, U.K.

144

PALYNOLOGY, VOLUME 31 — 2007

between the pollen/spores in samples 8 and 9 to those from samples 6 and 7 (Text-Figure 4; Table 1) strongly suggests that the majority of the material was derived from the last locality visited, i.e. site 3. However some differences were noted. The diversity in samples 8 and 9 is slightly higher than in samples 6 and 7; furthermore there are more Alnus, Aster type, Betula, Brassicaceae, Cyperaceae, Lactuceae, and Pinus in samples 8 and 9 than in samples 6 and 7 (Text-Figure 4; Table 1). Moreover, the occurrences of Plantago type, Quercus, Lycopodium, and Myriophyllum were only recorded in sample 8, and Caryophyllaceae, Ranunculus type, and Sinapis type, are unique to sample 9. Because samples 8 and 9 produced higher diversities and elevated levels of certain taxa that are present at sites 1 and 2, these compositional differences probably represent input from sites 1 and 2. For example, the relatively high levels of Cyperaceae and Lactuceae in sample 8 probably are, at least in part, derived from site 1. Likewise the presence of Quercus in sample 8 may be derived from sites 1 and 2, and the occurrence of Caryophyllaceae and Ranunculus type in sample 9 could have come from site 2. The Picea and Pinus pollen in samples 8 and 9 is also likely to have derived from sites 1 and 2. By contrast, undifferentiated broken saccate pollen is, somewhat surprisingly, confined to samples 6 and 7. The bar diagram of the most common pollen/spore types (Text-Figure 4) clearly illustrates the difference between pure site 3 soil, and the boots that had visited sites 1 and 2. Site 4, Mixed Deciduous Woodland at Ruddington Country Park, Nottingham, U.K. (samples 10, 11, 12, 13 and 14) This site is a stand of dense mixed woodland bordering a large country park (Text-Figures 1e, 9). The five samples from this site produced relatively abundant pollen/spore associations of moderate diversity (Text-Figure 4; Table 1). The diversities in samples 10 to 14 are 17, 18, 17, 15 and 15 respectively. The pollen/spore assemblages are all substantially similar and are dominated by Alnus, Aster type, Cyperaceae, Lactuceae, Pinus, and Poaceae. By far the most prominent taxa are Alnus and Poaceae (TextFigure 4). Minor levels of Betula, Brassicaceae, Caryophyllaceae, Chenopodiaceae/Amaranthaceae, Frankenia cf. laevis, cf. Littorella, Lycopodium, Malva, Picea, Plantago type, Quercus, Ranunculus type, Rhynchospora, Sinapis type, Tilia, and undifferentiated broken saccate pollen were also recorded (Table 1). Unique occurrences are those of cf. Littorella and Tilia in sample 10, Frankenia cf. laevis, Quercus, Ranunculus type and Sinapis type in sample 11, and Malva and Rhynchospora

Text-Figure 9. Site 4, mixed deciduous woodland at Ruddington Country Park, Nottingham, U.K.

in sample 14 (Table 1). The only rare forms found in all five samples are Brassicaceae, Chenopodiaceae/ Amaranthaceae, Picea, and undifferentiated broken saccate pollen. The prominence of Alnus and Poaceae reflects the common alder trees in this area of woodland, and the grass of the nearby parkland respectively. Other trees include Betula and Quercus, and herbs are represented also. The pollen/spore associations from the two pure samples of soil from site 4 (i.e. numbers 10 and 11) are similar in diversity, taxonomic make up, and relative proportions. This again demonstrates that soil from new footwear normally yields virtually identical pollen assemblages to the parent soil (Text-Figures 4, 6). Pollen assemblages from the two samples are characterized by dominant Alnus and Poaceae, with lesser proportions of Aster type, Cyperaceae, Lactuceae, and Pinus. Both the common and rare components are similar, and Acer, Hedera helix, and cf. Primula are entirely absent. Samples 12, 13 and 14 are prone to contamination from material from sites 3, 2 and 1. The similarities between the pollen/spores in samples 12, 13 and 14, and those from samples 10 and 11 indicate that the majority of the material was derived from the last locality visited, i.e. site 4. The bar diagram of the most common pollen/spore types (TextFigure 4) clearly illustrates the similarities between all five samples from site 4. However some relatively minor differences were observed. Unlike at sites 2 and 3, the diversity of the three samples from multiple sites is lower than that of pure soil from site 4 (Table 1). Malva and Rhynchospora were only recorded rarely in sample 14; no unique taxa were recovered from samples 12 and 13. More Poaceae and undifferentiated broken saccate pollen are present in samples 12, 13, and 14 than the pure soil samples 10 and 11. It is not

J.B. Riding, B.G. Rawlins, and K.H. Coley: Changes in soil pollen assemblages on footwear worn at different sites

clear if these increases represent contamination, and if so, from which site(s) they were derived. By contrast, several taxa diminish in numbers in the multiple site samples 12, 13, and 14; these are Aster type, Betula, Cyperaceae, and Picea (Text-Figure 4; Table 1). The numbers of Alnus pollen in sample 13 are significantly higher than the other samples at site 4. This increased level apparently represents natural variability at site 4 because Alnus proved extremely rare at sites 1, 2 and 3. Similarly the numbers of Pinus in sample 12 is relatively high; this may also be due to natural variations (Text-Figure 4).

145

This site is a low, grassy area next to a lake with nearby stands of mixed shrubs and trees, including birch trees (Text-Figures 1e, 10). The most prominent components are common grasses, low-level herbs/weeds, and the common reed (Phragmites communis – Poaceae). Pollen from samples 15 and 16 (i.e. pure locality 5) are of moderate diversity (18 and 17 taxa respectively). Sample 15 is relatively abundant, and is dominated by Alnus, Betula, Poaceae, and Sinapis cf. arvensis. Lower numbers of Aster type, Chenopodiaceae/Amaranthaceae, Lactuceae, Picea, Pinus, Quercus, and Sinapis type are also present (TextFigure 4; Table 1). Trace components observed include Brassicaceae, Cyperaceae, Isoëtes, cf. Lobelia, cf. Primula, and cf. Urtica. This assemblage reflects the occurrence of alder and birch trees, grass and herbs. By contrast, the assemblage from sample 16 is relatively sparse in pollen/spores; only 155 grains were counted from three microscope slides. This association is overwhelmingly dominated by Poaceae (Text-Figure 4). Lower pro-

portions of Aster type, Cyperaceae, Lactuceae, Picea, Pinus, Pediastrum, Plantago type, Quercus, and Sinapis type are also present. Trace levels of Alnus, Betula, Chenopodiaceae/Amaranthaceae, and Typha were observed. The area is grassy and this explains the preponderance of Poaceae. The occurrence of Pediastrum in this sample is entirely consistent with a lakeside environment; this is a widespread genus of freshwater colonial algae (Batten, 1996). Unusually, the pollen floras of these two samples are significantly different. Of the four dominant taxa in sample 15, sample 16 produced extremely low numbers of Alnus, Betula, and Sinapis cf. arvensis. Despite its abundance in sample 15, no representatives of Sinapis cf. arvensis were observed in sample 16. By contrast, the proportion of Poaceae in sample 16 is significantly higher than in sample 15 (Text-Figure 4). The proportions of Aster type, Cyperaceae, Lactuceae, Picea, Pinus, Quercus, and Sinapis type are relatively similar. Except for Pediastrum and Plantago type in sample 16, the only unique occurrences of identifiable pollen/spores are those that are relatively rare. These include Brassicaceae, Isoëtes, cf. Lobelia, cf. Primula, and cf. Urtica in sample 15, and Typha in sample 16 (Table 1). The palynomorph associations from samples 15 and 16 are hence substantially different, despite both being pure soil/lakeside mud from the same site. This is a highly unusual situation, and demonstrates that substantial natural variability in pollen profiles does exist. This phenomenon may also reflect the fact that this is a man-made lake which is approximately 15 years old. It may be that the palynological heterogeneity observed reflects the abundant levels of transported material during the landscaping process. The substantial statistical distance between these samples is shown in Text-Figure 6. Normally, sites are characterized

Text-Figure 10. Site 5, edge of lake at Ruddington Country Park, Nottingham, U.K.

Text-Figure 11. Site 6, margin of Wollaton Park Lake, Nottingham, U.K.

Site 5, Edge of Lake at Ruddington Country Park, Nottingham, U.K. (samples 15 and 16)

146

PALYNOLOGY, VOLUME 31 — 2007

by a small number of abundant taxa. In this case, the two samples from site 5 have similar proportions of Aster type, Cyperaceae, Lactuceae, Picea, Pinus, Quercus, and Sinapis type, which are of low to moderate abundance (TextFigure 4; Table 1). Site 6, Margin of Wollaton Park Lake, Nottingham, U.K. (samples 17 and 18) This locality is the immediate margin of a large lake close to a path which is next to extensive grassed parkland with occasional shrubs and trees (Text-Figures 1d, 11). Unlike site 5, the pollen floras from samples 17 and 18 at site 6 are substantially similar in terms of both the spectrum of taxa present, and their relative proportions (Text-Figure 4; Table 1). This is confirmed by multivariate statistical analysis (Text-Figure 6). The diversity is moderate, with 19 and 18 taxa recorded respectively. The two samples are dominated by Poaceae, with significant proportions of Alnus, Lactuceae, Pinus, and Quercus (Text-Figure 4). Other forms are present in relatively low proportions. These include Aster type, Betula, Brassicaceae, Caryophyllaceae, Cedrus, Chenopodiaceae/Amaranthaceae, Cyperaceae, Malva, Picea, Pediastrum, Plantago type, cf. Saxifraga, Sinapis type, Tilia, undifferentiated broken saccate pollen, and cf. Urtica (Table 1). Sample 18 yielded substantially less Pinus and Quercus than sample 17; otherwise the associations from these two samples are broadly comparable (Text-Figure 4). This site is characterized by prominent Poaceae, with common/moderate numbers of Alnus, Lactuceae, Pinus, and Quercus. In particular, the relatively high proportions of Quercus, together with the moderate levels of Alnus, is highly characteristic. The dominant pollen reflects the grassed park with stands of alder and oak trees, with some weeds/herbs. The presence of Pediastrum in sample 17 is consistent with a lakeside setting (Batten, 1996). Sample 18 potentially includes material from sites 5 and 6, however the striking similarity of the two pollen/spore floras indicates that the majority of the material is from site 6, i.e. the last site visited (Text-Figure 4; Table 1). In any case, contamination from site 5 would be difficult to discern due to the markedly disparate nature of the floras from site 5 (see above). However, it is possible that the slightly raised levels of Alnus and Poaceae, and the occurrences of rare Chenopodiaceae/Amaranthaceae and Plantago type, in sample 18 may be, at least in part, derived from site 5. It is also possible that this putative contamination may have diluted the levels of Cyperaceae, Pinus, and Quercus in sample 18 (Text-Figure 4). It must be stressed that these differences are relatively minor, and well within normal levels of natural variability.

Detrended Correspondence Analysis The first two ordinates explain most of the variation in the data and the site scores for them are plotted in TextFigure 6. The species scores are given in Table 2. A notable feature of the site scores (Text-Figure 6) is the clustering of both ‘boot’ and ‘soil’ samples according to the site visited last. For example, all the ‘boot’ samples with material derived from sites 1, 2, 3 and 4, and their respective soil samples, are enclosed by the dashed ellipses drawn around them. The pollen composition for the ‘soil’ from site 5 appears to be anomalous; the score on the first ordinate is far larger than those for the ‘boot’ sample from this site, and they are not closely associated. The only explanation for the large observed differences in pollen composition from this site is that it comprises transported material (made ground) which has not been thoroughly mixed. However, the pollen composition from the ‘boot’ which had been worn at sites 5 and 6 does occur in close proximity to both the ‘boot’ sample from site 5 and the ‘soil’ sample at site 6. DISCUSSION The raw pollen/spore data from the 18 samples is displayed in Table 1, and the relative proportions of the eleven dominant taxa illustrated in Text-Figure 4. From both Table 1 and Text-Figure 4, it is clear that each of the six sites has an extremely distinctive pollen/spore association. However, site 5 is unusual in that the two samples produced substantially different palynofloras. The distinctiveness of the palynofloras from these sites was also demonstrated statistically using detrended correspondence analysis (TextFigure 6). It is clear from Text-Figure 6 that the samples from sites 1–4 and site 6 cluster closely together, demonstrating that the statistical distance between them is low. There are two principal conclusions emanating from this study. Firstly, that the six individual localities have distinctively different pollen/spore signatures. This is due to the nature of the virtually infinite variability of local vegetation spectra, and to the multiplicity of complex taphonomic processes which affect the pollen and spores deposited. Some examples of the many pertinent factors are altitude, nutrient availability, prevailing wind, soil type, and topography. This apparent palynological heterogeneity confirms the potential usefulness of palynomorph associations in forensic studies in order to establish provenance, and hence to link people or items to a specific location. The analysis herein shows that adherent materials on footwear dominantly reflect the last site that was visited, even if several sites have been attended previously (Text-Figures 4, 6). In other words, palynomorph associations derived from the

J.B. Riding, B.G. Rawlins, and K.H. Coley: Changes in soil pollen assemblages on footwear worn at different sites

last site are not significantly diluted by the pollen from previous sites. Detrended correspondence analysis supports this (Text-Figure 6). However, footwear samples that were potentially contaminated with materials from previous localities normally exhibited some subtle differences. These are typically slight increases in diversity, and small variations in certain pollen types. The subtlety of these differences means that they would be difficult to discern and/or quantify in a ‘real’ forensic investigation. Differences in the pollen load on footwear which has been worn at other sites is to be expected because shoes/boots etc. are known to pick up palynomorphs, especially if the sites are on soft/natural ground, as opposed to in the built environment (Wiltshire, 2006). The findings of this study show that in forensic investigations, it is imperative that footwear from suspects is seized as soon as possible, since the palynological signature on the shoes/boots will probably reflect the last site visited. If several polleniferous sites are visited after the wearing of footwear at a crime scene, the crime scene pollen signature will normally be diluted, perhaps profoundly so. If there is a delay in seizing footwear, any mud etc. on the boots/shoes should be carefully studied and dissected prior to preparation. This is in order to attempt to establish a pollen chronology or stratigraphy for the material entrained on the tread of the footwear. One possible pertinent scenario is where a crime scene is characterized by highly unusual, probably exotic, pollen type(s). If the unusual (?non-native) pollen was found in a background of abundant, natural (?native) pollen picked up from other localities, the former could link a suspect to a crime scene which was visited prior to other localities. In such a case, even small numbers of the unusual pollen that had persisted on the footwear could be critical to the case. ACKNOWLEDGMENTS This research was funded by projects managed by Drs Michael G. Petterson (BGS, Keyworth) and J. Russ Evans (BGS, Edinburgh). James B. Riding and Barry G. Rawlins publish with the approval of the Executive Director, British Geological Survey (N.E.R.C.). This contribution received helpful and perceptive reviews from Drs. Vaughn M. Bryant Jr. (Texas A&M University) and Stewart G. Molyneux (BGS, Keyworth).

References Cited ADAM, D.P., and MEHRINGER, P.J. Jr. 1975 Modern pollen surface samples – an analysis of subsamples. U.S. Geological Survey Journal of Research, 3(6): 733–736.

147

BATTEN, D.J. 1996 Chapter 7C. Colonial Chlorococcales. In: Jansonius, J., and McGregor, D.C. (eds.). Palynology: principles and applications. American Association of Stratigraphic Palynologists Foundation, Dallas, 1: 191–203. BRYANT, V.M., and JONES, G.D. 2006 Forensic palynology: Current status of a rarely used technique in the United States of America. Forensic Science International, 163: 183–197. HILL, M.O., and GAUCH, H.G. Jr. 1980 Detrended Correspondence Analysis: an improved ordination technique. Vegetatio, 42: 47–58. HORROCKS, M., COULSON, S.A., and WALSH, K.A.J. 1998 Forensic palynology: variation in the pollen content of soil surface samples. Journal of Forensic Sciences, 43(2): 320–323. 1999 Forensic palynology: variation in the pollen content of soil on shoes and in shoeprints in soil. Journal of Forensic Sciences, 44(1): 119–122. JANSONIUS, J., and McGREGOR, D.C. (eds.) 1996 Palynology: principles and applications. American Association of Stratigraphic Palynologists Foundation, Dallas, 3 volume set, 1330 p. MILNE, L.A., BRYANT, V.M. Jr., and MILDENHALL, D.C. 2005 14. Forensic Palynology. In: Coyle, H.M. (ed.). Forensic Botany. Principles and applications to criminal casework. CRC Press LLC, Boca Raton, Florida, U.S.A, 217–252. MOORE, P.D., WEBB, J.A., and COLLINSON, M.E. 1991 Pollen Analysis. Second Edition. Blackwell Scientific Publications, Oxford, 216 p. OKSANEN, J., KINDT, R., LEGENDRE, P., and O’HARA, R.B. 2007 vegan: Community Ecology Package version 1.8-5. http://cran.r-project.org. TRAVERSE, A. 2007 Paleopalynology. Second edition. Springer, Dordrecht, The Netherlands, 813 p. WILTSHIRE, P.E.J. 2006 Consideration of some taphonomic variables of relevance to forensic palynological investigation in the United Kingdom. Forensic Science International, 163: 173–182.

148

PALYNOLOGY, VOLUME 31 — 2007

APPENDIX 1. List of samples. Note that the geographic coordinates are in meters, and refer to the British National Grid. The data were captured using a standard global positioning system (GPS) unit. Location 1: Scrubland near wood, southwest of Nanpantan, Leicestershire, U.K. (Text-Figures 1c, 5) Rough scrubland dominated by shrubs. GPS coordinate: 449774 Easting, 316764 Northing; elevation 157 m. Wearing and subsampling: 15.20h; 30th November 2005. Boot type: new pair of ‘Dr Martens’ 6-eye black boots – ‘pair 1’. Soil type: surface soil, dark brown/black slightly sandy soil that is relatively dry and well drained; much leaf/vegetable matter. Local geology: no outcrops visible; Charnian. Local flora: dominated by ferns/bracken. Some grasses and herbs/weeds such as nettles and thistles. Some small trees especially birch, hazel, and oak; mixed mature deciduous woodland nearby (ca. 6 m). SAMPLE #1 – bulk soil sample. Collection number MPA 54840. SAMPLE #2 – pure locality 1 soil sample taken immediately from the left-hand boot sole of pair 1. This was scrapings from the infilled treads all over the boot sole. Collection number MPA 54841.

Location 2: Elder/Hawthorn woodland, Ruddington Country Park, Nottingham, U.K. (Text-Figures 1e, 7) South side of park, along a footpath/cycle track. GPS coordinate: 458090 Easting, 332140 Northing; elevation 46 m. Wearing and subsampling: 10.30h; 1st December 2005. Boot type: new ‘Caterpillar’ safety boots – ‘pair 2’. Soil type: surface soil, dark brown slightly sandy soil that is relatively dry and well-drained. Much leaf/vegetable matter, largely hawthorn leaves with some red hawthorn berries. Abundant rabbit droppings were observed and a badger sett is present nearby. The samples were taken to the south side of the path, among elder trees. Local geology: no outcrops visible; till on Triassic mudstone. Local flora: trees are dense stands of elder and coppiced hawthorn; there are other (minor) mixed deciduous trees/shrubs. Some brambles and stands of nettles were observed. There were no grasses in the immediate area, but grassed fields are present nearby. SAMPLE #3 – bulk soil sample. Collection number MPA 54844. SAMPLE #4 – pure locality 2 soil sample taken immediately from the left-hand boot sole of pair 2. This was scrapings from the infilled treads all over the boot sole. Collection number MPA 54845. SAMPLE #5 – soil sample taken immediately from the righthand boot sole of pair 1. This was scrapings from infilled treads from the entire heel sector of the boot sole (mapped as ‘sector no. 1’). This was scraped off with a clean scalpel, and the scrapings are theoretically material from locations 1 and 2. There was 19 hours, 10 minutes between the visits to locations 1 and 2. No unequivocal mixing was observed. Collection number MPA 54848.

Location 3: Pine plantation - Ruddington Country Park, Nottingham, U.K. (Text-Figures 1e, 8) Pine woodland, close to children’s play area. GPS coordinate: 457043 Easting, 302226 Northing; elevation 41 m. Wearing and subsampling: 11.40h; 1st December 2005. Boot type: new pair of ‘Dr Martens’ black 7-eye safety boots – ‘pair 3’. Soil type: surface soil, mid/dark brown slightly sandy soil that is relatively dry and well drained. Much leaf/vegetable matter, largely pine needles is present. The samples were taken several meters from the edge of the woodland. Local geology: no outcrops visible; till on Triassic mudstone. Local flora: pine trees. Occasional immature birch trees; some thistles. Grasses and low herbs/weeds on ground; close to grassed areas. SAMPLE #6 – bulk soil sample. Collection number MPA 54849. SAMPLE #7 – pure locality 3 soil sample taken immediately from the left-hand boot sole of pair 3. This was scrapings from infilled treads all over the boot sole. Collection number MPA 54850. SAMPLE #8 – soil sample taken immediately from the righthand boot sole of pair 1. This was scrapings from infilled treads from the central sector of the boot sole (mapped as ‘sector no. 2’). This was scraped off with a clean scalpel, and the scrapings are theoretically material from locations 1, 2 and 3. There was 19 hours, 10 minutes between the visits to locations 1 and 2, and 1 hour, 10 minutes between the visits to locations 2 and 3. No definite mixing was observed. Collection number MPA 54853. SAMPLE #9 – soil sample taken immediately from the righthand boot sole of pair 2. This was scrapings from infilled treads from the heel sector of the boot sole (mapped as ‘sector no. 1’). This was scraped off with a clean scalpel, and the scrapings are theoretically material from locations 2 and 3. There was 1 hour, 10 minutes between the visits to locations 2 and 3. No definite mixing was observed. Collection number MPA 54854.

Location 4: Mixed deciduous woodland - Ruddington Country Park, Nottingham, U.K. (Text-Figures 1e, 9) Mixed deciduous woodland, close to the north part of the perimeter road. GPS coordinate: 457147 Easting, 332518 Northing; elevation 26 m. Wearing and subsampling: 13.20h; 1st December 2005. Boot type: new ‘Century’ wellington type boots - ‘pair 4’. Soil type: surface soil, black, sandy loam that is relatively dry and well drained. Much leaf/vegetable matter, largely sycamore and birch leaves. Samples were taken several meters from the edge of the woodland, close to the perimeter road. Local geology: no outcrops visible; till on Triassic mudstone.

J.B. Riding, B.G. Rawlins, and K.H. Coley: Changes in soil pollen assemblages on footwear worn at different sites

Local flora: coppiced mixed deciduous woodland dominated by sycamore with some birch and oak. Grass and herbs present at forest floor. Fungi (puffballs) also relatively common. The site is close to grassed areas and allotment gardens over the road. SAMPLE #10 – bulk soil sample. Collection number MPA 54855. SAMPLE #11 – pure locality 4 soil sample taken immediately from the left-hand boot sole of pair 4. This was scrapings from the infilled treads all over the boot sole. Collection number MPA 54856. SAMPLE #12 – soil sample taken immediately from the righthand boot sole of pair 1. This was scrapings from infilled treads from the entire upper part of the boot sole (mapped as ‘sector no. 3’). This was scraped off with a clean scalpel, and the scrapings are theoretically material from locations 1, 2, 3 and 4. There was 19 hours, 10 minutes between the visits to locations 1 and 2, 1 hour, 10 minutes between the visits to locations 2 and 3, and 1 hour, 40 minutes between the visits to locations 3 and 4. Some definite mixing was observed. Under the darker soil was grey mud/soil, possibly from locality 1. Collection number MPA 54859. SAMPLE #13 – soil sample taken immediately from the righthand boot sole of pair 2. This was scrapings from infilled treads from the entire upper part of the boot sole (mapped as ‘sector no. 2’). This was scraped off with a clean scalpel, and the scrapings are theoretically material from locations 2, 3 and 4. There was 1 hour, 10 minutes between the visits to locations 2 and 3, and 1 hour, 40 minutes between the visits to locations 3 and 4. No definite mixing was observed. Collection number MPA 54860. SAMPLE #14 – soil sample taken immediately from the righthand boot sole of pair 3. This was scrapings from infilled treads from the whole of the boot sole. This was scraped off with a clean scalpel, and the scrapings are theoretically material from locations 3 and 4. There was 1 hour, 40 minutes between the visits to locations 3 and 4. No definite mixing was observed, although some possible stratification was vaguely present. Collection number MPA 54861.

149

Location 5: Edge of lake at Ruddington Country Park, Nottingham, U.K. (Text-Figures 1e, 10) Grassed area close to lake edge. GPS coordinate: 457464 Easting, 332010 Northing; elevation 34 m. Wearing and subsampling: 15.45h; 20th November 2005 Boot type: new ‘Dr Martens’ 10-eye, black safety boots - ‘pair 5’. Mud type: wet, dark red-brown mud/soil at margin of lake. Local geology: no outcrops visible. Local flora: the most prominent components are the common reed (Phragmites communis – Poaceae), common grass and low-level herbs/weeds. Stands of mixed trees/shrubs close by. These include birch trees. SAMPLE #15 – bulk mud/soil sample. Collection number MPA 54833. SAMPLE #16 – pure locality 5 mud/soil sample taken immediately from the left-hand boot sole. This was scrapings from the infilled treads all over the boot. Collection number MPA 54834.

Location 6: Edge of Wollaton Park Lake, Nottingham, U.K. (Text-Figures 1d, 11) Lakeside margin, close to footpath. GPS coordinate: 452915 Easting, 338914 Northing; elevation 41 m. Wearing and subsampling: 13.20h; 23rd November 2005. Boot type: pair 5 Mud type: wet, very dark red-brown mud/soil at margin of lake. Local geology: the lake bank comprises clay-rich till with common erratic pebbles. Local flora: grasses, herbs/weeds, and reeds. Stands of trees (oak), catkin-bearing tree and shrubs are present nearby. SAMPLE #17 – bulk mud/soil sample. Collection number MPA 54835. SAMPLE #18 – mud/soil sample taken immediately from the right-hand boot sole of pair 5. This was scrapings from infilled treads from the bottom-left sector of the boot sole (mapped as ‘sector no. 1’). This was scraped off with a clean scalpel, and the scrapings are theoretically material from locations 5 and 6. There was 69 hours, 7 minutes between the visits to locations 5 and 6. Collection number MPA 54837.

150

PALYNOLOGY, VOLUME 31 — 2007

APPENDIX 2. Detailed sampling protocol Each visit to a site involved one of us (JBR) wearing the appropriate footwear and walking normally (i.e. not using exaggerated force in order to deliberately entrain material into the boot tread), in random directions, for one minute. The wearing of the respective footwear was confined to a small area of approximately 9 m3 at the specific locality. The items of footwear were not worn anywhere else, i.e. between localities. All the localities were on natural surfaces, i.e they were more or less vegetated. At each locality a composite sample of soil/surface materials was taken as a control. Five subsamples were collected from an approximately 3 m transect at each site, and thoroughly homogenized. This strategy of subsampling was undertaken in order to avoid the potential biasing of a single sample because of the proximity of specific plants. This should to lead to the collection of composite samples which are representative of the local pollen rain (Adam and Mehringer, 1975). The control samples were obtained by carefully scraping the uppermost 5 mm of surface material into a new sample bag using a clean knife. Furthermore, GPS readings were recorded, notes were made on the local vegetation, geology, and soil moisture conditions, and some photographs taken. ‘Pyramid 1’ – localities 1–4 (see Text-Figure 2). 1 At the first locality, a pair of new boots (‘Pair 1’) was worn. A sample of adherent material was collected from a mapped sector of the sole of the left boot of Pair 1; this sample was labelled ‘Pair 1, Left - Locality 1’. The right boot of Pair 1 was left intact. A surface/soil sample was taken as a control; theoretically this should be the same material as the adherent material sample from the left boot of Pair 1. 2 At the second locality, another pair of new boots (‘Pair 2’) was worn. A sample of adherent material was taken from a mapped sector of the sole of the left boot of Pair 2; this sample was labelled ‘Pair 2, Left - Locality 2’. The right boot of Pair 2 was left intact. A surface/soil sample was taken as a control; theoretically this should be the same material as the adherent material sample from the left boot of Pair 2. Pair 1 was worn again and a sample of adherent material was taken from a mapped sector of the sole of the right boot; this sample was labelled ‘Pair 1, right - Localities 1+2’. 3 At the third locality, another pair of new boots (‘Pair 3’) was worn. A sample of adherent material was taken from a mapped sector of the sole of the left boot of Pair 3; this sample was labelled ‘Pair 3, left - Locality 3’. The right boot of Pair 3 was left intact. A surface/soil sample was taken as a control; theoretically this should be the same material as the adherent material sample from the left boot of Pair 3. Pairs 1 and 2 were worn again and samples of adherent material were taken from mapped sectors of the soles of the right boots; these samples were labelled ‘Pair 1, right Localities 1+2+3’ and ‘Pair 2, right - Localities 2+3’ respectively.

4 At the fourth locality, another pair of new boots (‘Pair 4’) was worn. A sample of adherent material was taken from a mapped sector of the sole of the left boot of Pair 4; this sample was labelled ‘Pair 4, left - Locality 4’. The right boot of Pair 4 was left intact. A surface/soil sample was taken as a control; theoretically this should be the same material as the adherent material sample from the left boot of Pair 4. Pairs 1, 2 and 3 were worn again and samples of adherent material were taken from mapped sectors of the soles of the right boots; these samples were labelled ‘Pair 1, right Localities 1+2+3+4’, ‘Pair 2, right - Localities 2+3+4’ and ‘Pair 3, right - Localities 3+4’ respectively.

‘Pyramid 2’ – localities 5 and 6. 5 At the fifth locality, a pair of new boots (‘Pair 5’) was worn. A sample of adherent material was collected from the sole of the left boot of Pair 5; this sample was labelled ‘Pair 5, Left - Locality 5’. The right boot of Pair 5 was left intact. A surface/soil sample was taken as a control; theoretically this should be the same material as the adherent material sample from the left boot of Pair 5. 6 At the sixth locality, Pair 5 were worn. A sample of adherent material was taken from a mapped sector of the sole of the right boot of Pair 5; this sample was labelled ‘Pair 5, Right - Locality 6’. A surface/soil sample was taken as a control.

J.B. Riding, B.G. Rawlins, and K.H. Coley: Changes in soil pollen assemblages on footwear worn at different sites

151

APPENDIX 3. Listing of the botanic and common names of the palynomorph taxa observed in this study. Note that unidentified/undifferentiated forms could not be identified because they were either broken, or diagnostic features were not visible. Palynomorph taxa

Common name

I POLLEN Acer Alnus Ambrosia Apiaceae (Umbelliferae) Artemisia Aster type Betula Brassicaceae (Cruciferae) Caryophyllaceae Cedrus cf. Centaurea type Chenopodiaceae/Amaranthaceae Corylus Cyperaceae Frankenia cf. laevis Fraxinus Hedera helix Lactuceae cf. Littorella cf. Lobelia Malva Myriophyllum Picea Pinus Plantago type Poaceae cf. Primula Quercus Ranunculus type Rhynchospora Salix cf. Saxifraga Sinapis type Sinapis cf. arvensis Tilia Trifolium Typha Ulmus cf. Urtica Undiff. broken saccate pollen Unidentified pollen

maple/sycamore alder ragweed carrot family, e.g. celery, parsley, hogweed mugwort daisies birch cabbage family, e.g. mustard, watercress carnation family, e.g. chickweed, stichwort cedar cornflower, knapweed, star thistle etc. goosefoot hazel sedge family sea heath ash ivy dandelion tribe shoreweed lobelia mallow water milfoil spruce pine plantain grass family primrose/cowslip oak buttercups, crowfoots, and spearworts star sedge willow saxifrage charlock/mustard charlock/wild mustard lime clover bulrush elm nettle N/A N/A

II SPORES Isoëtes Lycopodium Polypodiaceae Undifferentiated trilete spores

quillwort clubmoss polypod fern family N/A

III MISCELLANEOUS Pediastrum

colonial freshwater alga

152

PALYNOLOGY, VOLUME 31 — 2007