Holocene environmental changes in the Lower Thames Valley,London,UK: Implications for understanding the history of Taxus woodland

Abstract

A radiocarbon-dated multiproxy palaeoenvironmental record from the Lower Thames Valley at Hornchurch Marshes has provided a reconstruction of the timing and nature of vegetation succession against a background of Holocene climate change, relative sea level movement and human activities. The investigation recorded widespread peat formation between c. 6300 and 3900 cal. yr BP (marine ‘regression’), succeeded by evidence for marine incursion. The multiproxy analyses of these sediments, comprising pollen, Coleoptera, diatoms, and plant and wood macrofossils, have indicated significant changes in both the wetland and dryland environment, including the establishment of Alnus (Alder) carr woodland, and the decline of both Ulmus (Elm; c. 5740 cal. yr BP) and Tilia (Lime; c. 5600 cal. yr BP, and 4160–3710 cal. yr BP). The beetle faunas from the peat also suggest a thermal climate similar to that of the present day. At c. 4900 cal. yr BP, Taxus (L.; Yew) woodland colonised the peatland forming a plant community that has no known modern analogue in the UK. The precise reason, or reasons, for this event remain unclear, although changes in peatland hydrology seem most likely. The growth of Taxus on peatland not only has considerable importance for our knowledge of the vegetation history of southeast England, and NW Europe generally, but also has wider implications for the interpretation of Holocene palaeobotanical records. At c. 3900 cal. yr BP, Taxus declined on the peatland surface during a period of major hydrological change (marine incursion), an event also strongly associated with the decline of dryland woodland taxa, including Tilia and Quercus, and the appearance of anthropogenic indicators.

Keywords

archaeology climate change Holocene relative sea level change vegetation history

Introduction

There remains a paucity of information on the vegetation history of southeast England, which is surprising given its geographical and historical importance with respect to our understanding of Holocene plant migration and succession, and human impact on the landscape and environment. Data that exist are derived primarily from alluvial and peat sequences located in coastal and river valley wetlands (e.g. Long et al., 2007; Waller and Hamilton, 2000; Waller and Schofield, 2007). These have made a valuable contribution to studies of vegetation history (e.g. Thomas and Rackham, 1996; Sidell et al. 2000; Wilkinson et al., 2000), set against a background of sea level change (e.g. Devoy, 1979; Sidell, 2003), cultural developments and land-use history (e.g. Carew et al., 2009; Meddens, 1996). This paper presents the findings from a multiproxy palaeoenvironmental study of a coastal wetland site at Hornchurch Marshes, situated in the Lower Thames Valley. The investigation provided an opportunity to enhance understanding of the factors influencing vegetation change in southeast England during the Holocene with a particular emphasis on the history of Taxus baccata (L.; Yew).

The Lower Thames Valley extends from central London to Tilbury, Mucking and Stanford Le Hope in the east (Figure 1a). Hornchurch Marshes are located approximately in the middle of this area, and on the north bank of the Thames near Dagenham (National Grid Reference: TQ 51100 82400; elevation: c. 1 m a.s.l.; Figure 1b). The Marshes are located on the valley floor of the River Thames (floodplain) in the estuarine reach of the river, and bounded on their north side by an area of higher ground known as the East Tilbury Marshes Gravel (Gibbard, 1995). These gravel deposits are a terrace remnant of a former braided river system deposited during the middle Devensian (c. 45,000 to 30,000 years ago). The valley floor is underlain by sands and gravels of late-Devensian age (Shepperton Gravel) that appear originally to have formed a surface differentiated in part at least by upstanding bars and intervening channels. Fine-grained, and sometimes organic (e.g. peat), estuarine and fluvial sediments have progressively buried this topography during the Holocene (Devoy, 1979, 1982). These near-surface sediments comprise ‘an interbedded sequence of alluvial sediments, dipping and thickening markedly from west to east, collectively referred to as the Tilbury Alluvium’ (Gibbard, 1995: 33).

Figure 1.

(a) The Lower Thames Valley and Hornchurch Marshes; (b) Hornchurch Marshes showing the location of the study area and nearby sites at Frog Lane (presence of Taxus wood macrofossils recorded in peat), Manser Works (dryland Bronze Age human activity with a burnt mound, postholes, pit and hearth) and Bridge Road (wetland Bronze Age human activity with a brushwood trackway) (Meddens, 1996; MoLAS, 1997; Newham Museum Service, 1995; Potter, 2003); (c) study area showing distribution of Taxus wood macrofossils in test pits and archaeological trench (Densem, 2001).

Palaeoenvironmental data (publicly or in the form of commercial archaeological reports) on these near-surface sediments indicate that following the end of the last glaciation, the lower reaches of the Thames Valley and its tributaries were inundated by the sea, and marine and estuarine sediments accumulated. Since that time, the evidence suggests that sea level continued to rise at a much slower rate as a response to either glacio-eustatic or sedimentary processes. Fluctuations in the relative height of sea level have occurred from time to time, for reasons such as long-term subsidence of southeast England and climate change (Devoy, 1982). Exposure of parts of the Lower Thames Valley during periods of lower or stable relative sea level (‘regression’) created more terrestrial environments, resulting in the formation of peat deposits and occasionally soils. These peat deposits formed the focus of research at Hornchurch Marshes.

Of particular significance were wood macrofossils of Taxus (L.), recovered from the peat deposits during test pit excavations in advance of an archaeological evaluation of the site, which provided unequivocal evidence for in situ growth of Taxus woodland in a peatland context (Figure 1c; Densem, 2001). Taxus has been recorded in other middle-Holocene palaeoecological records from northwest European peatlands, including the Somerset Levels (Beckett and Hibbert, 1979; Orme and Coles, 1989) and East Anglian Fens (Godwin, 1940, 1975; Godwin et al., 1935; Waller, 1994a) in the UK, as well as Ireland (Delahunty, 2002; O’Connell and Molloy, 2001) and the Belgian Coastal Plains (Deforce and Bastiaens, 2007). However, the reason(s) for its colonisation and decline in these wetland ecosystems is often unclear. The current study in the Lower Thames Valley therefore provided an opportunity to further our knowledge and understanding of the colonisation and decline of Taxus woodland in coastal wetlands during the middle Holocene.

Methods

The lithostratigraphic sequence within the trench recovered in column and core samples was systematically cleaned and described (Troels-Smith, 1955), and the heights above mean sea level noted (British Ordnance Datum – m OD; Figure 2). Continuous bulk samples were also recovered for macrofossil analysis. The organic matter content was determined using the loss-on-ignition method (Bengtsson and Enell, 1986).

Figure 2.

Radiocarbon-dated lithostratigraphy, age–depth model, organic matter content (%) and reconstructed depositional environment at Hornchurch Marshes.

Pollen grains and spores were extracted following standard procedures (Branch et al., 2005), and identified using type collections and the following sources of keys and photographs: Moore et al. (1991), Reille (1992). Plant nomenclature follows the Flora Europaea as summarised in Stace (1997). A total of 300 pollen grains (excluding aquatics and spores) were recorded for each sample. The results are expressed as a percentage of total land pollen (trees, shrubs and herbs). The percentage pollen diagram was subdivided into three pollen assemblage zones using numerical methods (CONISS) carried out within TGView (Grimm, 2004).

Diatom preparation from the mineral-rich lithostratigraphic units (Figure 2) followed standard techniques with two sets of slides prepared to confirm the absence of diatoms evident in a number of samples (Battarbee, 1986). Several diatom floras and taxonomic publications were consulted to assist with diatom identification, including Hendey (1964). Diatom species’ salinity preferences were classified using the halobian groups of Hustedt (1953, 1957: 199).

The plant macrofossils (seeds, fruit bodies and wood) were extracted by subsampling a standard volume of sediment (5000 ml), dispersing the sample in hot water, and sieving of the sample through 4 mm, 1 mm, 500 μm and 250 μm mesh sizes. The plant macrofossils were identified using type collections, and keys and photographs e.g. Berggren (1981). The wood specimens consisted mostly of fragments of roundwood. A large quantity of loose bark was also present, some of which was thick (i.e. >10 mm) and had clearly originated from wider trunk or branch fragments than were present in the samples. The sample specimens were subsampled and at least six fragments were selected from each for examination. The selection was based on diameter (to represent roundwood from across the range) and morphology (to include fragments with different external characteristics). Thin sections were prepared using standard methods (Gale and Cutler, 2000) to show the transverse, tangential and radial surfaces. These were mounted in glycerol, and the wood structure matched to reference slides. Plant nomenclature follows the Flora Europaea as summarised in Stace (1997). The results are presented as total counts of all specimens recovered.

Coleopteran subfossils were obtained from a vertical series of samples taken in 10 cm units through the peat. They were recovered by wet sieving of dispersed samples (each 2000 ml) over a 300 μm mesh, concentrated by paraffin flotation (Atkinson et al., 1987).

Six radiocarbon determinations were made on identified wood macrofossils extracted at key lithostratigraphic horizons through the main peat sequence (Figure 2). The radiocarbon determinations were calibrated using the maximum intercept method (Stuiver and Reimer, 1986), OxCal version 4.0.1 (Bronk Ramsey, 1995, 2001, 2007), and the internationally agreed data set for terrestrial samples from the Northern Hemisphere (Reimer et al., 2004). The full age range (rounded to 10 years) is quoted as ‘cal. yr BP’. An age–depth model was created using OxCal version 4.0.1 with a P-sequence algorithm and k factor of 100 (Blockley et al., 2007; Bronk Ramsey, 2007).

Lithological interpretation and results of the radiocarbon dating

Between −4.15 and −3.83 m OD, a clay-rich sedimentary unit was deposited probably on the margins of a river channel (floodplain). This unit is probably indicative of a rising water-table due to marine transgression during the early Holocene (Devoy, 1979; Sidell, 2003). Between −3.83 and −3.54 m OD, organic-rich mineral sediments were deposited, representing increasing stabilisation of the landscape and the formation of a more terrestrial rather than fluvial environment (Figure 2). This could be due to either: (a) the abandonment or moving away of a nearby active river channel, or (b) a regional reduction in the water-table resulting from a lowering in height of relative sea level (see Devoy, 1982; Sidell, 2003).

Between −3.54 m OD and −1.74 m OD, the formation of wood peat represents the creation of more terrestrial conditions at the site. The results of the radiocarbon dating indicate that peat formation commenced at c. 6300 cal. yr BP, and ceased at c. 3900 cal. yr BP (Figure 2; Table 1). This corresponds to a regional reduction in the rate of relative sea level rise from 2.6 mm/yr to 0.8 mm/yr from c. 8000 BP, which led to widespread peat formation in the Lower Thames Valley between approximately 6800 and 5800 cal. yr BP (Sidell, 2003). The formation of wood peat was ‘interrupted’ on two occasions by the deposition of silt-rich fluvial sediments representing much wetter conditions. These occurred between −3.50 to −3.47 (c. 6100 cal. BP) and −2.81 to −2.68 m OD (centred on c. 5400 cal. yr BP). Sometime after c. 3900 cal. yr BP, between −1.74 and −1.36 m OD, the deposition of a clay-rich sedimentary unit represents the return of fluvial, possibly estuarine, conditions at Hornchurch Marshes. The results indicate, therefore, that the entire sequence spans the late Mesolithic to middle Bronze Age cultural periods.

Table 1.

Description of the local pollen assemblage zones and radiocarbon dates from Hornchurch Marshes.

Local Pollen Assemblage Zone	Description	Radiocarbon dates
MOY3 (−1.74 to −1.36m OD): Alnus, Herbaceous taxa	Values of non-arboreal pollen exceed 30% during this zone and include a diverse range of taxa such as Poaceae (20%), Chenopodium type (6%), Cyperaceae (11%), Lactuceae (6%), Plantago lanceolata (<5%) and Cereale type (<5%). Arboreal pollen taxa are dominated by Alnus (60%) with Pinus (10%), Corylus (9%) and Salix (<5%) also present. Aquatic and spore taxa include Filicales (18%), Pteridium (5%), Typha latifolia (<5%) and Pre-Quaternary spores (<5%).
MOY2 (−2.41 to −1.74m OD): Alnus, Quercus, Taxus	High percentage values of arboreal pollen dominate this zone. Alnus (75%), Quercus (25%) and Taxus (12%) are the main taxa, although Tilia (<5%), Ulmus (<5%) and Corylus (5%) are present throughout. Herbaceous pollen taxa include Cyperaceae (<10%), Poaceae (<5%), Ranunculus type (<5%), Rumex (<5%) and Apiaceae (<5%). Aquatic and spore taxa include Utricularia (5%), Filicales (<10%) and Polypodium (<5%).	Beta-156200: Wood (Alnus); −2.46 to −2.36m OD; 4370 ±60 uncal yrs BP; −29.0 δ¹³C (‰); 5275-4836 cal yrs BP (95.4%)Beta-154303: Wood (Alnus); −2.11 to 2.09m OD; 4640 ±110 uncal yrs BP; −29.1 δ¹³C (‰); 5591-4980 cal yrs BP (95.4%)Beta-156199: Wood (Fraxinus); −1.86 to −1.76m OD; 3590 ±80 uncal yrs BP; −26.9 δ¹³C (‰); 4144-3688 cal yrs BP (95.4%)
MOY1 (−3.75 to −2.41m OD): Alnus, Quercus	The zone is characterised by high pollen percentage values of Alnus (96% to 35%) and Quercus (35%). Other arboreal pollen taxa, such as Betula (<5%), Tilia (<5%), Ulmus (<5%) and Corylus (5%) are less well represented. Herbaceous pollen taxa include Cyperaceae (<10%), Poaceae (<5%) and Apiaceae (<5%). The assemblage of aquatic pollen taxa and spores includes Filicales (16%), Polypodium (7%), Typha latifolia (<5%) and Utricularia (<5%).	Beta-156302: Wood (Salix/Populus); −3.65 to −3.63m OD; 5670 ±40 uncal yrs BP; −27.8 δ¹³C (‰); 6560-6320 cal yrs BP (95.4%)Beta-156202: Wood (Alnus); −3.41 to −3.31m OD; 5260 ±70 uncal yrs BP; −27.8 δ¹³C (‰); 6263-5906 cal yrs BP (95.4%)Beta-156201: Wood (Alnus); −2.81 to −2.71m OD; 4860 ±70 uncal yrs BP; −25.5 δ¹³C (‰); 5745-5331 cal yrs BP (95.4%)

Results and interpretation of the palaeobotanical analyses

Pollen analysis

The pollen-stratigraphy has been divided into three local pollen assemblage zones (LPAZ) reflecting the main changes in percentage values (Table 1; Figure 3). LPAZ MOY1 indicates that between −3.75 and −2.41 m OD (c. 6600–4900 cal. yr BP) the dryland vegetation cover at Hornchurch Marshes was dominated by mixed Quercus-dominated (Oak) deciduous woodland with Tilia (Lime), Ulmus (Elm) and Corylus (Hazel), and lesser amounts of Taxus, Betula (Birch) and Fraxinus (Ash). The most notable change in the composition of the dryland vegetation occurs from c. 5700 cal. yr BP with a reduction in Ulmus woodland.

Figure 3.

Percentage pollen and constrained cluster analysis diagram, incorporating the lithostratigraphic descriptions, associated radiocarbon dates and age–depth model, from Hornchurch Marshes.

The wetland woodland composition was analogous to present-day fen (‘carr woodland’) and may be found in low-lying areas that are waterlogged for much of the year. The fen was dominated by Alnus glutinosa (Alder) woodland, a tree that is commonly found in wetland ecosystems. The mixed wetland woodland at Hornchurch Marshes would also have consisted of Betula, Fraxinus, Salix (Willow), Frangula alnus (Alder Buckthorn), Hedera helix (Ivy) and Prunus (e.g. Blackthorn). The pollen assemblage indicates that the ground flora would have consisted of the following herbs and ferns: Cyperaceae (Sedges), Filicales (e.g. Buckler Fern), Urtica (Nettle), Filipendula (Meadowsweet), Ranunculus (Buttercup), Valeriana officinalis (or dioica; e.g. Marsh Valerian), Sparganium erectum (Bur-reed) and Thalictrum (Meadow-rue). LPAZ MOY1 also provides evidence, albeit circumstantial, for vegetation succession. At the base of the sequence, the high pollen percentages of Cyperaceae may indicate the presence of sedge swamp at Hornchurch Marshes prior to the establishment of Alnus. Second, the reduction in pollen values of Alnus during the zone corresponds to an increase in a diverse range of herbaceous pollen as well as Fraxinus and Hedera. These data suggest that a minor change in local wetland vegetation (c. 50–150 m for arboreal taxa; see Bunting et al., 2004, 2005) composition allowed the colonisation of more light-demanding plant taxa, and the increased representation of pollen from nearby dryland. There is no pollen-stratigraphic evidence to indicate that these changes in vegetation composition were induced by human activities, such as clearance of woodland, but instead probably represent the natural creation of temporary woodland glades/gaps.

The transition to LPAZ MOY2 is characterised by the colonisation of Taxus woodland on the fen surface either as a tall shrub or tree (c. 4900 cal. yr BP; Figure 3). The new vegetation composition would have consisted therefore of Alnus carr woodland with Taxus. The absence of a diverse assemblage of herbaceous pollen taxa supports this interpretation and implies that there was very little penetration of light to the understorey. The presence of Taxus also implies that the peat surface may have been drier and ‘more terrestrial’ than in LPAZ MOY1 because it seems unlikely that a very wet peat surface that was permanently waterlogged would have supported Taxus woodland (see Thomas and Polwart, 2003).

The upper part of LPAZ MOY2 records two important changes in the local vegetation cover. The first was a deepening of water on the fen surface indicated by the presence of Utricularia pollen (e.g. Utricularia major; Bladderwort), and second the slight decline of arboreal pollen values, principally Alnus and Taxus. These changes indicate that the fen surface was becoming wetter, perhaps permanently waterlogged, and hence less suitable for the sustained growth of the main tree taxa. The dryland vegetation cover continued to be dominated by mixed Quercus-dominated deciduous woodland with Tilia, Ulmus and Corylus.

LPAZ MOY3 provides a record of significant changes in pollen-stratigraphy (c. 3900 cal. yr BP): (1) the decline in arboreal pollen values of the principal tree taxa – Alnus, Tilia, Quercus and Taxus; (2) the expansion of pollen taxa indicting secondary woodland colonisation, especially Fraxinus and Corylus; (3) an increase in the diversity of herbaceous pollen indicating increased penetration of light to the ground surface (including pollen taxa of the Compositae family (Daisy), Poaceae (Grass), Potentilla type (e.g. Potentilla erecta; Cinquefoil), Rumex sp. (Docks and Sorrels) and Sinapis type (e.g. Charlock)); (4) an increase in pollen of light demanding shrubs, including Hedera helix and Juniperus (Juniper); and (5) an increase in pollen taxa indicative of aquatic vegetation communities, such as Typha latifolia (Reedmace; Figure 3).

The reduction in dryland and wetland woodland, coupled with clear changes in the vegetation composition and structure, coincide with two important events: (1) a change in lithology to mineral-rich sedimentary deposits, and (2) the presence of direct pollen-stratigraphic indicators of human activities – cereal cultivation. The pollen-stratigraphic record provides unequivocal evidence that the deposition of mineral-rich sediments resulted in significant changes in the local vegetation cover. The presence of Alnus together with Poaceae, Typha latifolia and Sparganium type pollen indicates an open freshwater wetland. However, the high pollen values of Chenopodium type (e.g. Saltwort) may indicate that local environmental conditions were brackish rather than exclusively freshwater. Plants of the Chenopodiaceae family may be split into two broad groups, those associated with brackish and marine environments such as Salsola kali (Prickly Saltwort), and those commonly found in waste places and the edges of arable fields on dryland, such as Chenopodium album (Fat Hen). Unfortunately, it is not possible to separate these two groups using pollen analysis, although some support for the presence of brackish water conditions may be the single occurrence of Armeria maritima pollen at the top of LPAZ MOY2.

The decline of arboreal pollen and the increase in non-arboreal pollen values may also be accounted for by Bronze Age human activity. The presence of cereal pollen, albeit in small quantities, together with a range of herbaceous pollen types commonly associated with human modification of the vegetation cover, such as Plantago lanceolata (Ribwort Plantain), indicates that cultivation was probably taking place on dryland to the north of Hornchurch Marshes. The decline of Quercus, Tilia and Ulmus pollen may signify clearance of dryland woodland for cultivation. It seems likely therefore, that changes in vegetation cover were occurring on the wetland and dryland at broadly the same time because of the combined effects of human activities and natural environmental change.

Plant macrofossil and wood analyses

Plants commonly found within fen carr woodlands dominated the plant macrofossil record from Hornchurch Marshes (Figure 4). Although overall the concentration and preservation of plant macrofossils was poor, with some samples having no macrofossil preservation, the presence of Alnus glutinosa, Salix sp., Rubus sp. (e.g. Blackberry) and Prunus sp. indicates a mosaic of woodland with an understorey of shrubs. The abundance of Salix remains between −1.58 and −1.38 m OD (sometime after c. 3800 cal. yr BP) is the most significant change in the plant macrofossil record and identifies a transition from Alnus to Salix dominated wet woodland. This transition occurs during a period of mineral-rich sediment deposition and suggests that local conditions became less suitable for the growth of Alnus woodland. The wood macrofossil record complements these data with all samples composed of roundwood ranging from twiggy material to wider stems/ branches (Figure 4). Loose bark was also present and the dimensions of many of these fragments indicated an origin from wider branches or trunks than were represented by the wood fragments. A high proportion of the bark was morphologically similar to that of Alnus glutinosa but, occasionally, some thick and deeply fissured pieces were present (but remain unidentified). Alnus was dominant, while other taxa were rarely present. Fraxinus excelsior occurred in samples −3.08 to −3.18 m and −1.78 to −1.88 m OD. Corylus avellana roundwood and nutshells occurred sporadically throughout and, in addition, Sambucus (Elder) and possibly Rhamnus cathartica (Buckthorn) occurred in the lower part of the sequence, while Cornus or Viburnum (Dogwood or Wayfaring Tree) was identified from the uppermost part. Non-Alnus roundwood appeared to be less frequent in the upper part of the sequence and included Fraxinus excelsior, Prunus spinosa, and possibly Rhamnus cathartica.

Figure 4.

Plant and wood macrofossil data incorporating the lithostratigraphic descriptions, associated radiocarbon dates and age–depth model, from Hornchurch Marshes.

Diatom analysis

Diatom valves were poorly preserved throughout most of the mineral-rich sediment units (Figure 2). The only sample for which it was possible to make a diatom count was from −1.66 m OD (sometime after c. 3800 cal. yr BP) that had a moderately high diatom concentration with moderately good diatom preservation (Table 2). The diatom assemblage is dominated by brackish water (mesohalobous) diatoms (34%) whilst halophilous species comprise over 4% and brackish-marine species (polyhalobous to mesohalobous) represent 9% of the flora. Marine (polyhalobous) species comprise 21% of the diatom assemblage and freshwater diatoms represent 18% of the valves present. Typical of diatom assemblages from the Thames Estuary the dominant species is the planktonic estuarine species Cyclotella striata (17%), along with a high percentage of another brackish water species Nitzschia navicularis (9%) which in contrast to the former species is a benthic diatom associated with epipelic (mud surface) habitats. The outer-estuary marine species Paralia sulcata is also abundant (11%). With the exception of Cyclotella striata, the freshwater and brackish diatom life forms are predominantly non-planktonic whilst polyhalobous taxa are predominantly planktonic. It is therefore likely that marine diatoms represent an allochthonous input from the outer estuary whilst the non-planktonic brackish water and freshwater species represent aquatic conditions in the immediate environment along with Cyclotella striata from the open water of the tidal river. Amongst the brackish and freshwater species present, epipelic (mud surface), epipsammic (sand surface) and epiphytic habitats are all represented by the species present.

Tabel 2.

Diatom analysis from Hornchurch Marshes (sample at −1.66 m OD).

Taxon	Count	%
Amphora sp.	2	0.8
Diploneis sp.	14	5.6
Gomphonema sp.	2	0.8
Gyrosigma sp.	5	2.0
Navicula sp.	6	2.4
Nitzschia sp.	2	0.8
Stauroneis sp.	1	0.4
Surirella sp.	1	0.4
Unknown	3	1.2
Unknown Salinity Total		14.3
Campylosira cymbelliformis(A. Schmidt) Grun. ex Van Heurck 1885	1	0.4
Cymatosira belgicaGrun. in Van Heurck 1881	5	2.0
Paralia sulcata sulcata(Ehrenb.) Cleve 1873	28	11.1
Podosira stelligera(J.W. Bail.) Mann 1907	2	0.8
Rhaphoneis amphicero(Ehrenb.) Ehrenb. 1844	6	2.4
Rhaphoneis sp.	3	1.2
Rhaphoneis surirella(Ehrenb.) Grun. in Van Heurck 1881	2	0.8
Thalassionema nitzschioides(Grun.) Grun. ex Hustedt 1932	2	0.8
Trachyneis aspera aspera(Ehrenb.) Cleve 1894	3	1.2
Polyhalobous Total		20.6
Actinoptychus undulatus(J.W. Bail.) Ralfs in Pritch. 1861	2	0.8
Cocconeis scutellum scutellum Ehrenb. 1838	1	0.4
Coscinodiscus sp.	1	0.4
Diploneis smithii smithii (Breb. ex W. Sm.) Cleve 1894	11	4.4
Navicula marina	2	0.8
Pseudopodosira westii(W. Sm.) Sheshukova-Poretzskaya 1964	5	2.0
Thalassiosira decipiens(Grun.) E. Jorg. 1905	1	0.4
Polyhalobous to Mesohalobous Total		9.1
Caloneis westii(W. Sm.) Hendey 1964	2	0.8
Campylodiscus echeneisEhrenb. 1840	1	0.4
Cyclotella striata striata (Kutz.)Grun. in Cleve & Grun. 1880	43	17.1
Diploneis didyma (Ehrenb.) Cleve 1894	3	1.2
Navicula digitoradiata (Greg.) Ralfs in Pritch. 1861	1	0.4
Navicula peregrina peregrina (Ehrenb.)Kutz. 1844	1	0.4
Nitzschia granulata Grun. 1880	8	3.2
Nitzschia navicularis (Breb. ex Kutz.)Grun. in Cleve & Grun. 1880	23	9.1
Synedra pulchella pulchella Ralfs ex Kutz. 1844	4	1.6
Mesohalobous Total		34.1
Navicula mutica mutica Kutz. 1844	7	2.8
Navicula slesvicensisGrun. in Van Heurck 1880	1	0.4
Halophilous Total		3.2
Rhoicosphenia curvata (Kutz.)Grun. 1860	2	0.8
Surirella ovata ovata Kutz. 1844	1	0.4
Halophilous to Oligohalobous Indifferent Total		1.2
Amphora libyca Ehr.	5	2.0
Aulacoseira sp.	1	0.4
Caloneis bacillum bacillum (Grun.) Cleve 1894	1	0.4
Cocconeis disculus (Schum.) Cleve 1896	7	2.8
Epithemia adnata adnata (Kutz.) Rabenh. 1853	1	0.4
Fragilaria construens venter (Ehrenb.)Grun. in Van Heurck 1881	4	1.6
Fragilaria pinnata pinnata Ehrenb. 1843	7	2.8
Gomphonema angustatum angustatum (Kutz.) Rabenh. 1864	1	0.4
Hantzschia amphioxys amphioxys (Ehrenb.) Grun. 1877	1	0.4
Navicula elginensis elginensis (Greg.)Ralfs in Pritch. 1861	2	0.8
Navicula rhynchocephala rhyncocephala Kutz. 1844	5	2.0
Nitzschia amphibia amphibia Grun. 1862	7	2.8
Pinnularia major major (Kutz.) W. Sm. 1853	1	0.4
Pinnularia subcapitata subcapitata Greg. 1856	1	0.4
Oligohalobous Indifferent Total		17.5
Total Diatom Count	252	100

Results and interpretation of the coleopteran analysis

The coleopteran fauna from the peat was rich in the variety of species but poor in their numerical frequency (Table 3). Altogether 98 beetle taxa were recognised of which 57 could be determined to species. The fauna suggests that there was no significant environmental change during the accumulation of the peat and for that reason it will be treated here as a single assemblage.

Table 3.

List of Holocene Coleoptera from the Hornchurch Marshes. The nomenclature and taxonomic order follows that of Lucht (1987). The numbers in the sample columns and opposite each species indicate the minimum numbers of individuals of that species in each sample.

Sample Depth (m OD)	−3.38 to −3.48	−3.28 to −3.38	−3.18 to −3.28	−3.08 to −3.18	−2.88 to −2.98	−2.68 to −2.78	−2.58 to −2.68	−2.48 to −2.58	−2.28 to −2.38	−2.18 to −2.28	−1.88 to −1.98	−1.68 to −1.78	−1.48 to −1.58
Coleoptera
Carabidae
Elaphrus cupreus (Duft.)		1			1		1
Loricera pilicornis (F.)	1		1
Dyschirius sp.						1
Bembidion doris (Panz.)		2	1
Bradycellus sp.			1
Pterostichus diligens (Sturm.)		1				1		2				1	1
Pterostichus minor (Gyll.)	1	1	3	2	2	2	2	1		1
Agonom fuliginosum (Panz.)	1					1	1
Dromus quadrimaculatus (L.)	1
Haliplidae
Haliplus sp.			1					2
Dytiscidae
Hygrotus inaequalis (F.)	1	1		1
Hygrotus decoratus (Gyll.)	1		1
Hydroporus palustris (L.)				1					3
Hydroporus spp.	1	2			1		1		1	2
Graptodytes bilineatus (Sturm,)				1	1
Agabus sturmi. (Gyll)							1
Agabus undulatus (Schrk.)				1
Ilybius sp.	1		1		1			1
Nartus grapei (Gyll.)	1		1						2
Acilius sp.	1								1
Dytiscus sp.				1				1
Hydraenidae
Hydraena testacea Curt.	2	8	9	4	1	3	2	2	6	1	2	1	1
Hydraena sp.	1		2
Ochthebius minimus (F.)	3	7	10	13	3	12	9	5	10		2		1
Limnebius aluta Bedel,	6	5	12	4	1	5	4	1	9			1
Limnebius truncatellus Thoms.				1
Hydrochus elongatus (Schall.)	1		3	3		1	1	1
Hydrochus brevis (Hbst.)									1
Helophorus sp.					1
Hydrophilidae
Coelostoma orbiculare (F.)	1	1	2	1	2		1	1
Cercyon ustulatus (Preysel.)				1
Cercyon convexiusculus Steph.	2	1	4	1	2	3	1	1	2
Cercyon sternalis Shp.	9	2		3	3	3	5	1				1
Hydrobius fuscipes (L.)	1	1	3	2	2	2	3	2	2
Anacaena sp.	1		2			1
Laccobius sp.		1	1		2
Enochrus sp.		1	1	1			3		2
Cymbiodyta marginella (F.)	1		1		1			1	1
Histeridae
Plagaderus dissectus Er		1
Silphidae
Phosphuga atrata (L.)			1
Liodidae
Agathidium sp.								1
Scydmaenidae
Stenichnus sp.					1
Orthoperidae
Corylophus cassidoides (Marsh.)	1	3	3		2
Orthoperus sp.	1		1				1
Ptiliidae
Ptenidium spp.	1	1			1	1	1		1
Acrotrichis sp.	2				1	1
Staphylinidae
Proteinus sp.		1
Eusphalerum sp.				1	1
Oxytelus inustus Grav.				1
Stenus juno (Payk.)		2		1		1
Stenus spp.	1	2	1	1	2	1	2	2	3
Lathrobium terminatum Grav.		1
Lathrobium spp.	1	1				1	1	1	1
Quedius sp.	1
Tachyporus sp.	1
Aleocharinae Gen. et sp. indet.	1	5	5	1	4	1	1	2	3	1			1
Pselaphidae
Euplectus sp.					1
Trichonyx sulcicollis (Reichb.)		1
Bryaxis sp.		1	1	1						1
Cantharidae
Rhagonycha testacea (L.)			1				1		1
Rhagonycha lignosa (Müll.)			1
Elateridae
Ampedus sp.		1
Agriotes sputator (L.)		1
Athous sp.											1
Eucnemidae
Melasis buprestoides (L.)	1
Helodidae
Gen. et sp. indet.	12	24	21	6	4	8	12	3	10	1	1
Dryopidae
Dryops sp.		1	1	1	1
Byrrhidae
Limnichus pygmaeus (Sturm,)							1
Nitidulidae
Meligethes sp.	1		1
Cucujidae
Laemophloeus bimaculatus (Payk.)							1
Phalacridae
Olibrus sp.						1
Lathridiidae
Lathridius sp.							1
Corticaria sp.	1
Colydiidae
Colydium elongatum (F.)								1
Cerylon histeroides (F.)		1					1
Anobiidae									1
Ernobius sp.		1					1
Anobium punctatum (Geer,)			1	1					1
Ptinidae
Ptinus fur (L.)		1			3
Scarabaeidae
Aphodius sp.						1				1	1
Phyllopertha horticola (L.)						1
Lucanidae
Dorcus parallelopipedus (L.)						1
Sinodendron cylindricum (L.)								1
Cerambicidae
Pogonocherus hispidus (L.)							1
Chrysomelidae
Plateumaris sericea (L.)					1
Plateumaris braccata (Scop.)	1
Prasocuris phellandrii (L.)					1		1	1
Melasoma aenea (L.)	1	1	3	2	1	1	1	1	3	1			1
Agelastica alni (L.)				1	1		1	1	1
Psylliodes sp.	1
Scolytidae
Hylastes spp.	2		1
Hylesinus crenatus (F.)		2					1
Hylesinus oleiperda (F.)		1				1
Taphrorychus bicolor (Hbst.)			1
Curculionidae
Apion sp.	1
Caulotrupodes aeneopiceus (Boh.)	1	1
Tanysphyrus lemnae (Payk.)	6	24	41	3	12	3	6	1	5
Notaris scirpi (F.)						1
Thryogenes sp.	1					1

The most outstanding feature of this assemblage were the large numbers, both in specific diversity and in their abundance, of aquatic species that live in stagnant, well-vegetated, eutrophic fresh water. For the specific habitat requirements of these water beetles see Balfour Browne, (1950) and Nilsson and Holmen (1995) for the dytiscids, and (Hansen, 1987) for the hydraenid and hydrophilid species.

The marginal vegetation comprised Carex (Sedges) and Scirpus (Clubrush or Bulrush), the principal food plants of Plateumaris sericea, Notaris scirpi and Thryogenes. Plateumaris braccata feeds largely on Phragmites. Prasocuris phelandrii feeds chiefly on various aquatic Apiaceae such as Oenanthe phellandrium, Circuta virosa, and Sium latifolia. The soldier beetles Rhagonycha testacea and Rhagonycha lignosa are predators on other insects and are frequently found on the flower heads of Apiaceae. Their larvae live on the ground where they are predators mainly on slugs and snails (Harde, 1984). Limnichus pygmaeus feeds on algae on the muddy margins of ponds. The abundance of the minute weevil Tanysphyrus lemnae is particularly significant here because it feeds almost exclusively on the duckweed Lemna (Duckweed), though a variety of this species feeds on Caltha (e.g. Marsh Marigold). In this context it is most likely that this weevil was living on the duckweed Lemna.

Several carnivorous or scavenging species lived in these habitats adjacent to the water. Thus Bembidion doris and Elaphrus cupreus are found in swampy conditions often which are strongly shaded. Pterostichus minor and Agonum fuliginosum that live in swamps where the vegetation is rich and where the ground surface is shaded. They have their maximum frequency in swampy forests where they are especially associated with Alnus (Lindroth, 1992). Loricera pilicornis also lives in damp shaded habitats usually near small bodies of stagnant water. The staphylinid beetles in this assemblage live in damp leaf litter where they predate other small arthropods and worms. The minute species Corylophus cassidoides is also carnivorous and lives in detritus Phragmites, Carex and grasses. The large number of individuals of Helodidae are also indicative of moist habitats since, although the adult beetles live in damp leaf litter, their larvae are aquatic. Of particular environmental significance is the complete absence from this assemblage of any beetle species that require saline habitats and the presence of several species that strictly avoid brackish conditions.

Many of the coleopteran species in this assemblage indicated the local presence of trees either directly as a food source or more indirectly as an essential component of their habitat. Thus, the abundant presence of Melasoma aenea and, to a lesser extent, Agelastica alni indicated that Alnus dominated the surrounding swampy area since both species feed almost exclusively on the leaves. Agelastica alni is a very rare species in Britain at the present day. The scolytid beetles in this assemblage tunnel under the bark of both healthy and sickly deciduous trees where the larvae feed in the cambium layer. Thus Hylesinus crenatus and Hylesinus oleiperda attack various broadleaved trees but particularly prefer Fraxinus. Taphrorhychus bicolor attacks a wide spectrum of deciduous trees. Similarly, the larvae of the long-horned beetle Pogoncherus hispidus develop in tunnels in a wide variety of deciduous wood (Koch, 1992). Many of the species in this assemblage live in rotting wood, almost exclusively of deciduous trees. Thus, the larvae of Melasis buprestoides live in rather dry rotten wood of various deciduous trees in sun-exposed places. Sinodendron cylindricum and Dorcus parallelipipedus (the Lesser Stag Beetle) are large beetles that have larvae that excavate deep tunnels into rotten, crumbling wood, taking several years to mature (Jessop, 1986). The larva of Anobium is the familiar ‘woodworm’ which attacks wood of all sorts, particularly if it has been already infested with fungus. The elaterid Ampedus lives in rotten wood of either deciduous or coniferous trees.

This assemblage also included species that use trees more indirectly as habitats rather than as food sources. Thus, the carabid species Dromius quadrimaculatus is a predator that hunts for other insects amongst the branches of various deciduous trees and conifers (Lindroth, 1992). Plagioderus dissectus feeds on small insect larvae under the bark of dead deciduous trees (Halstead, 1963). Trichonyx silcicollis is also a predator living under dead bark, often of old elm stumps (Pearce, 1957). Laemophloeus bimaculatus similarly lives under dead bark where it probably feeds on fungi. The curiously shaped Colydium elongatus enters the burrows of various wood boring beetles where it preys upon the larvae there.

This beetle fauna, viewed as a whole, indicates that the woodlands in the Lower Thames Valley at this time included an abundance of moribund trees with local accumulations of dead and decomposing timber. There is only scant evidence from the Coleoptera for the local presence of drier, more open country. Oxytelus inustus is a predatory species that lives amongst vegetable debris in warm meadow-like habitats. Agriotes sputator lives in grassland where its larvae feed on the roots of plants. Similarly, Phyllopertha horticola (the Garden Chafer) has larvae that feed underground on the roots of grass and herbaceous plants. The rarity of these species suggests that these habitats were either locally rare or at some distance from the site.

Some species have particular habitats not covered above. Thus, Phosphuga atrata is a predator on snails. Ptinus fur is found naturally in old trees and bird’s nests but at the present day has taken to living as a stored product pest. Apart from this very tenuous synanthropic connection, there is no evidence from this assemblage for the presence of any human activity in the area.

Climatic implications of the coleopteran assemblage

All the species may be found living today in southern England and some species are restricted to the south suggesting that the climate was similar to that of the present day. However there are no species present that have geographic ranges exclusively to the south of Britain, indicating that the climate was probably not warmer than now.

It is possible to quantify the thermal climate at this time using the Mutual Climatic Range (MCR) method (Atkinson et al., 1987). Altogether 16 species of carnivorous and general scavenging species in this assemblage are also present on the MCR data base. These species were used in order to avoid any phytophagous species whose ranges may be determined by that of their preferred food plants. Thus, the MCR method was designed to provide an estimate of the thermal climate that was independent of any similar estimate derived from the plants. Using the MCR programme the following figures were obtained:

TMAX (mean temperature of the warmest month) 16°C–22°C

TMIN (mean temperature of the coldest months) −11°C to 6°C

These figures indicate that the monthly mean temperature lay somewhere between these limits and not that the mean monthly temperatures ranged between these limits. Here again these estimates of the thermal climate indicate that temperatures at the time need not have been any warmer than those of the present day in southern England.

Discussion

Environmental change and the history of Taxus woodland in the Lower Thames Valley

The pollen-stratigraphic record indicates that the main expansion of Taxus woodland at Hornchurch Marshes occurred at −2.41 m OD, which has been radiocarbon dated to c. 4900 cal. yr BP, with a decline recorded at c. 3900 cal. yr BP. The absence of supporting wood macrofossil evidence for Taxus in the samples analysed was somewhat surprising given the unequivocal evidence for the presence of Taxus wood in eight out of the 18 test pits excavated at Hornchurch Marshes (Figure 1c; Densem, 2001). This suggests that although Taxus was growing on the peat surface it was not the dominant component of the woodland cover. It is perhaps less surprising however that the coleopteran analysis has not detected the presence of Taxus because of the known chemical resistance (antifeedant activity) of Taxus to insects (Daniewski et al., 1998). The combined proxy data nevertheless indicate that Taxus formed a woodland community with Alnus for which there are only a few detailed European palaeobotanical records (e.g. Beckett and Hibbert, 1979; Deforce and Bastiaens, 2007; Delahunty, 2002; Godwin, 1975; O’Connell and Molloy, 2001; Orme and Coles, 1989; Waller, 1994a) and, significantly, for which there is no known modern analogue in the UK (see Seel, 2001).

This interpretation is supported by pollen and plant macrofossil analyses carried out elsewhere in the Lower Thames Valley (Figure 5; Batchelor, 2009, unpublished data, 2009), with Taxus macrofossils recorded within the peat stratigraphy as far west as East India Docks (Pepys, 1665), and Aveley Parish and Erith Forest in the east (Seel, 2001; Wilkinson and Murphy, 1995). It is possible that Taxus grew beyond these margins and current research by one of the authors (Batchelor) as part of a programme of doctoral study in the Lower Thames Valley is enhancing our understanding of the precise timing and nature of the rise and fall of Taxus. Known records suggest that Taxus grew on the north and south bank of the River Thames, from the interface between the peatland and dryland, to a position proximal to the present day river channel. In addition, despite unequivocal evidence for Taxus growth on the peat surface, it is highly likely that Taxus also colonised the adjacent dryland.

Figure 5.

Map of the distribution of Taxus within the Lower Thames Valley (based on Batchelor, 2009, unpublished data, 2009): (1) Joan Street (Sidell et al., 2000); (2) Union Street (Sidell et al., 2000); (3) Preston Road (Branch et al., 2007); (4) Silvertown (Wilkinson et al., 2000); (5) Greenwich Industrial Estate (Morley, 2003); (6) East Beckton District Centre (Jarrett, 1996); (7) Royal Albert Dock (Batchelor, 2009); (8) Woolwich Trade Park (Batchelor, 2009); (9) Crossness Sewage Works (Batchelor et al., 2007a); (10) North Bexley (Branch et al., 2004); (11) Imperial Gateway (Batchelor et al., 2009); (12) Corinthian Quay (Corcoran and Lamb, 2002); (13) Spine Road Development/Bronze Age Way (Sidell et al., 1996); (14) Aveley Marshes South (Batchelor, 2009); (15) Tilbury Fort (Batchelor, 2009); (16) Blackwall Docks (Pepys, 1665); (17) East India Docks (Pepys, 1665); (18) Fort Street (Wessex Archaeology, 2000); (19) Beckton Tollgate (Tamblyn, 1994); (20) Albert Dock (Spurrell, 1889); (21) Beckton Alp (Truckle and Sabel, 1994); (22) East Ham FC (Scaife, 2001); (23) Crossness Sewage Works (Batchelor et al., 2007b); (24) Frog Lane (Newham Museum Service, 1995); (25) Aveley Parish (Wilkinson and Murphy, 1995); (26) Royal Docks Community School (Holder, 1998); (27) Beckton Nursery (Divers, 1995a); (28) Golfers’ Driving Range (Batchelor, 2009); (29) A13 Woolwich Manor Way (Gifford and Partners, 2001); (30) Ferndale Street (Divers, 1995b); (31) Beckton Sewage Works, Divers, 1995c); (32) Hays Storage (Divers, 1996); (33) Norman Road (Batchelor et al., 2008); (34) Slade Green (Bates and Williamson, 1995); (35) Erith Forest (Seel, 2001); (36) Aveley Marshes North (Batchelor, 2009); (37) Wennington Marsh (Sidell, 2003).

The reason or reasons for its colonisation of the peat surface are uncertain, and this scientific issue forms the focus of Batchelor’s research (Batchelor, 2009). Here, we propose five possible reasons for its colonisation, although others may emerge. First, the growth of Taxus on the peat would require relatively dry and stable surface conditions, which were certainly present at Hornchurch Marshes from c. 5300 cal. yr BP according to the combined litho- and biostratigraphic data, following a period of much wetter surface conditions. Indeed changes in peatland hydrology at c. 3900 cal. yr BP marked by the deposition of fine-grained mineral sediments and renewed flooding of the peat surface were also coincident with the decline of Taxus woodland. Flooding, due to a rise in the water-table, would have created an unsuitable environment for Taxus rooting and growth, especially under saline (estuarine) conditions (Buschbom, 1968; Seel, 2001; Thomas and Polwart, 2003). Recorded throughout the Lower Thames Valley, including Hornchurch Marshes, these sediments are commonly associated with marine/brackish diatom taxa, reflecting the increasing influence of estuarine conditions (Sidell, 2003). Depending upon the rate of flooding, this may have led to a catastrophic decline in Taxus populations.

Second, variations in relative sea level would have strongly influenced water-table height and peat formation, and corresponding changes in vegetation cover. The colonisation of Taxus at Hornchurch Marshes occurred however approximately 1200 years after peat initiation, although notably during a period of reduced flooding (as noted above). Therefore, although there are clear relationships between relative sea level change, peat formation and Taxus colonisation, at present it is uncertain whether this was the main forcing factor behind the growth of Taxus on the peatland. However, as noted above, flooding of the peat surface because of a rise in relative sea level (transgression) would have created unstable and unfavourable conditions for Taxus growth, and would have initiated a decline. The records from Hornchurch Marshes provide strong support for this interpretation, and are in agreement with the current sea level model for the Lower Thames Valley (Sidell, 2003).

Third, there is unequivocal evidence for human activity in the Lower Thames Valley wetlands during the Bronze Age, and to a lesser extent during the Neolithic (e.g. Carew et al., 2009; Meddens, 1996), and impact on the peatland surface may have created gaps in the woodland cover creating optimal conditions for colonisation of Taxus. However, there is no direct biostratigraphical or archaeological evidence from Hornchurch Marshes to support this interpretation.

Fourth, the period of Taxus colonisation at Hornchurch Marshes coincides with a period of climate change according to combined terrestrial, marine and ice core proxy records, with evidence for enhanced precipitation-evaporation in the UK from 4200 to 4000 cal. yr BP (e.g. Barber et al., 2003; Charman, 2010; Hughes et al., 2000; Langdon et al., 2003; Mayewski et al., 2004). Taxus baccata certainly grows best in the high humidity of mild oceanic climates, with mild winters and abundant rainfall (Hageneder, 2007; Thomas and Polwart, 2003; Tittensor, 1980). Therefore, whilst climate change is unlikely to have been the primary driver, a favourable climate probably aided colonisation of Taxus at Hornchurch Marshes. Finally, it is possible that a different species or subspecies of Taxus colonised the peat surface at Hornchurch Marshes, with ecological preferences quite different from Taxus baccata. The existence of a different genotype adapted to growth on peatland may certainly explain differences in ecology displayed by Taxus through time. However, it is of note that Taxus has been recorded on peat in the UK during previous interglacials, which suggests that its growth at Hornchurch Marshes was most likely related to favourable local conditions linked to hydrology and possibly climate (Andersen, 1975; West, 1962).

Ulmus and Tilia declines in the Lower Thames Valley

In addition to the changes in wetland vegetation, the biostratigraphic records have also permitted a broad reconstruction of the dryland vegetation cover during the middle Holocene at Hornchurch Marshes, which has implications for our understanding of vegetation succession in the Lower Thames Valley. Throughout the sequence, Quercus woodland was probably dominant with Ulmus, Tilia and Corylus. The low percentages of Ulmus pollen suggest that elm was only a minor component of the woodland cover. However, given the evidence for closed Alnus woodland on the wetland surface, it is possible that Ulmus pollen is under-represented because of ‘filtering’ of extra-local and regional pollen by the dense local vegetation cover. Nevertheless, the decline in Ulmus pollen values at 5740–5460 cal. yr BP, although possibly a minor event at Hornchurch Marshes, is of some significance because it coincides with the widely recognised ‘Ulmus decline’ in NW European pollen records (Parker et al., 2002) and those from the Lower Thames Valley (e.g. Devoy, 1979; Wilkinson, 1988; Wilkinson et al., 2000).

There is now an exhaustive literature that discusses the possible causes of the Ulmus decline: climate change to cooler conditions (e.g. Smith, 1981), soil deterioration (Peglar and Birks, 1993), competitive exclusion (e.g. Huntley and Birks, 1983; Peglar and Birks, 1993), human influence (e.g. Lamb and Thompson, 2005; Scaife, 1988), and disease (Girling, 1988; Perry and Moore, 1987). Although a multicausal model is generally advocated, mainly involving human interference during the Neolithic cultural period and disease (e.g. Parker et al., 2002), the new data from Hornchurch Marshes do not provide a means of testing this hypothesis. Unfortunately, there is no biostratigraphical evidence for human activity, such as cereal pollen, although dung beetles recorded later in the stratigraphic record might indicate animal husbandry (Table 3). In addition, despite the palaeoentomological study at Hornchurch Marshes, there is no evidence for Scolytus scolytus or S. multistriatus, the beetles responsible for carrying the fungus Ceratocystis (Ophiostoma) ulmi that causes disease. These Scolytus species have been recorded at Hampstead Heath, London (Girling, 1988) and Red Moss of Candyglirach, Aberdeen (Clark and Edwards, 2004), and more recently at Old Seagers Distillery, London, and Horton Kirby, Kent (Batchelor et al., unpublished data, 2012; Elias et al., 2009). Therefore, there is a small but growing body of sites, including those in the Lower Thames Valley, providing evidence for the key beetle taxa associated with the Ulmus decline, which provides stronger support for disease as a possible cause. Whether disease, or another factor, was responsible for the Ulmus decline at Hornchurch Marshes remains unknown but the record nevertheless provides a contribution to the growing data base of information on the elm decline in the Lower Thames Valley.

The history of Tilia woodland in the UK, and especially the cause(s) of its decline from the second half of the middle Holocene onwards, have been comprehensively reviewed by Grant et al. (2011). At Hornchurch Marshes, the decline in Tilia woodland commenced c. 5600 cal. yr BP, with a notable reduction at 4160–3710 cal. yr BP, which is consistent with that reported by Grant et al. (2011). Given that the lime decline occurred over a prolonged time period (from c. 5600 cal. yr BP) and during a period of widespread peat formation in the Lower Thames Valley, the cause may be equated with paludification Type I of Grant et al. (2011; see also Waller, 1994b). It is also highly likely that the further decline in Tilia at 4160–3710 cal. yr BP was due to the onset of mineral-rich (alluvium) sediment accumulation causing an apparent decline in lime woodland because of changes in pollen recruitment (see Grant et al., 2011). The low percentage values of Tilia pollen throughout the sequence are considered to be the combined effects of pollen production and recruitment (i.e. entomophily or alluvial sedimentation) and the ‘filtering’ of extra-local and regional pollen by the closed Alnus woodland on the wetland surface (see Bunting et al., 2004, 2005).

The Tilia decline is a widely recognised event in the Lower Thames Valley, occurring between c. 5000 and 3000 cal. yr BP. At Bramcote Green, for example (Branch and Lowe in Thomas and Rackham, 1996), pollen-stratigraphic evidence indicated that it occurred at the transition from peat formation to inundation of the floodplain surface by estuarine sediments, which is consistent with the further decline at 4160–3710 cal. yr BP at Hornchurch Marshes. However, at Hornchurch Marshes and Bramcote Green there is also evidence for human activity at this time, notably cereal pollen (although the possibility that these large Poaceae pollen grains were derived from wild, coastal grass species should be noted). Given the archaeological evidence throughout the Lower Thames Valley for Bronze Age human activities on both the wetland (e.g. wooden trackways and platforms) and dryland (e.g. settlement) at this time there remains a possibility that human interference may have contributed to the decline of Tilia in some parts of the Valley (see Carew et al., 2009; Meddens, 1996; Potter, 2003). This suggestion is consistent with the findings of Grant et al. (2011), who indicate that 56% of the 164 Tilia declines recorded in the UK can be attributed to human activity (see also Turner, 1962).

Conclusions

The radiocarbon-dated palaeoenvironmental records from Hornchurch Marshes indicate that peat initiation started at c. 6300 cal. yr BP during a period of marine regression, and ceased at c. 3900 cal. yr BP with the onset of estuarine sedimentation. These data enhance our understanding of the timing and nature of sea level change in this part of the Lower Thames Valley.

During the period of peat accumulation, the wetland was composed of Alnus carr woodland, sedge and grass swamp and open water according to the combined pollen, plant macrofossil and insect evidence. At c. 4900 cal. yr BP, Taxus colonised the peat surface, forming a plant community that has no known modern analogue in the UK. The precise reason, or reasons, for this event remain unclear, although changes in peatland hydrology linked to sea level seems most likely. At c. 3900 cal. yr BP, Taxus declined during a period of major hydrological change, an event also strongly associated with the decline of other woodland taxa, including Tilia, and the appearance of anthropogenic indicators in the pollen data. Therefore, human activity and hydrological change may both have been broadly contemporary, contributory factors.

The new pollen data have also provided evidence for a decline in Ulmus woodland at 5740–5460 cal. yr BP. Unfortunately, there was no clear evidence for competition, disease, climate change or human activity in the multiproxy records from Hornchurch Marshes; the main causes proposed for the elm decline. This is disappointing given the growing body of evidence for both disease and human activity in the Lower Thames Valley. Therefore, the cause of the decline at Hornchurch Marshes remains uncertain. In contrast, the marked decline in Tilia woodland at c. 5600 cal. yr BP, and especially from 4160 to 3710 cal. yr BP, was linked to paludification and pollen recruitment (respectively), although human activity may also have been a causal factor.

The palaeoenvironmental investigation at Hornchurch Marshes has therefore provided an important contribution to our understanding of the Holocene environmental history of the Lower Thames Valley. The study has demonstrated that a multiproxy approach in coastal wetland contexts enables a detailed reconstruction of the regional and local environment permitting a clearer differentiation between vegetation succession on the wetland and dryland, and improved evaluation of the processes influencing vegetation change. Ongoing research in the Lower Thames Valley into the history of Taxus, as well as Ulmus and Tilia, by several of the authors will provide, in due course, a much clearer picture of the timing and duration of vegetation changes across a wider geographical area, enabling detailed comparison with palaeoecological records from other parts of the UK and Europe.

Footnotes

Acknowledgements

CRB and NPB would like to thank Dr Jane Sidell (University College London and English Heritage) and Professor Martyn Waller (University of Kingston) for helpful comments on the Taxus research project. NPB and RD would like to thank Dr Ian Tyers for identification of the Taxus wood macrofossils. The authors would like to thank the anonymous reviewers for helpful comments on the manuscript.

Funding

These investigations were undertaken with the support and funding from Compass Archaeology Limited (site code: MOY00) and Duncan Hawkins of CgMs Consulting. CRB’s doctoral research was jointly funded by CgMs Consulting and ArchaeoScape (Royal Holloway, University of London).

References

Andersen

S-Th

(1975) The Eemain freshwater deposit at Egernsund, South Jylland, and the Eemian landscape development in Denmark. Danmarks Geologiske Undersogelse Arbog 1974: 49–70.

Atkinson

Briffa

Coope

(1987) Seasonal temperatures in Britain during the past 22,000 years, reconstructed using beetle remains. Nature 325: 587–592.

Balfour Browne

(1950) British Water Beetles Vol. 2. London: Ray Society.

Barber

Chambers

Maddy

(2003) Holocene palaeoclimates from peat stratigraphy: Macrofossil proxy climate records from three oceanic raised bogs in England and Ireland. Quaternary Science Reviews 22: 521–539.

Batchelor

(2009) Middle Holocene environmental changes and the history of yew (Taxus baccata L.) in the lower Thames Valley. Royal Holloway, University of London Unpublished PhD Thesis.

Batchelor

Branch

Austin

(2007b) Crossness Sewage Works, Crossness, London Borough of Bexley: Environmental Archaeological Analysis of Waterlogged Wood Remains (Site Code: CXS07). ArchaeoScape (Royal Holloway, University of London).

Batchelor

Branch

Elias

. (2007a) Thames Water Utilities LTD, Tidal Thames Quality Improvements, Crossness, London Borough of Bexley: Environmental Archaeological Analysis (Site Code EAW06). ArchaeoScape (Royal Holloway, University of London).

Batchelor

Elias

Young

. (2009). Former Borax Works, Norman Road, Belvedere, London Borough of Bexley: Environmental Archaeological Analysis (Site Code: NNB07). Quaternary Scientific (University of Reading).

Bates

Williamson

(1995) A Report on the Stratigraphy, Palaeoenvironmental and Archaeological Significance of the Slade Green Relief Road Site. Geoarchaeological Service Facility, Department of Human Environment, Institute of Archaeology, University College London.

10.

Battarbee

(1986) Diatom analysis. In: Berglund

(ed.) Handbook of Holocene Palaeoecology and Palaeohydrology. Chichester: John Wiley, pp. 527–563.

11.

Beckett

Hibbert

(1979) Vegetational change and the influence of prehistoric man in the Somerset levels. New Phytologist 83: 577–600.

12.

Bengtsson

Enell

(1986) Chemical analysis. In: Berglund

(ed.) Handbook of Holocene Palaeoecology and Palaeohydrology. Chichester: John Wiley, pp. 423–451.

13.

Berggren

(1981) Atlas of Seeds and Small Fruits of North-west European Plant Species with Morphological Descriptions: Part 3 Salicaceae–Cruciferae. Sweden: Berlings, Arlöv.

14.

Blockley

SPE

Blaauw

Bronk Ramsey

. (2007) Building and testing age models for radiocarbon dates in Lateglacial and early Holocene sediments. Journal of Quaternary Science 26(15–16): 1915–1926.

15.

Branch

Batchelor

Elias

. (2007) Preston Road, Poplar High Street, Poplar, London Borough of Hamlets (Site Code: PPP06): Environmental Archaeological Analysis. ArchaeoScape (Royal Holloway, University of London).

16.

Branch

Canti

Clark

. (2005) Environmental Archaeology: Theoretical and Practical Approaches. London: Edward Arnold.

17.

Branch

Silva

Swindle

(2004) An Environmental Archaeological Assessment: North Bexley Drainage Improvements, Belvedere, Kent (EXY01). ArchaeoScape (Royal Holloway, University of London).

18.

Bronk Ramsey

(1995) Radiocarbon calibration and analysis of stratigraphy: The oxcal program. Radiocarbon 37(2): 425–430.

19.

Bronk Ramsey

(2001) Development of the radiocarbon program oxcal. Radiocarbon 43(2a): 355–363.

20.

Bronk Ramsey

(2007) Deposition models for chronological records. Quaternary Science Reviews 27(1–2): 42–60.

21.

Bunting

Armitage

Binney

. (2005) Estimates of ‘relative pollen productivity’ and ‘relevant source area of pollen’ for major tree taxa in two Norfolk (UK) woodlands. The Holocene 15: 459–465.

22.

Bunting

Gaillard

M-J

Sugita

. (2004) Vegetation structure and pollen source area. The Holocene 14: 651–660.

23.

Buschbom

(1968) Salzresistenz oberirdischer Sprobteile von Holzgewächsen I. Chlorideinwirkungen auf Sproboberflächen. Flora 157: 527–561.

24.

Carew

Meddens

Batchelor

. (2009) Human–environment interactions at the wetland edge in East London: Trackways, platforms and Bronze Age responses to environmental change. London and Middlesex Archaeological Society 60: 1–34.

25.

Charman

(2010) Centennial climate variability in the British Isles during the mid–late Holocene. Quaternary Science Reviews 29: 1539–1554.

26.

Clark

SHE

Edwards

(2004) Elm bark beetle in Holocene peat deposits and the northwest European elm decline. Journal of Quaternary Science 19(6): 525–528.

27.

Corcoran

Lam

(2002) Land at Project Alice, the Former British Gypsum Site, Corinthian Quay, Church Manor Way, Erith: A Report on the Geoarchaeological Evaluation. Museum of London Archaeology Service.

28.

Daniewski

Gumulka

Anczewski

. (1998) Why the yew tree (Taxus baccata) is not attacked by insects. Phytochemistry 49(5): 1279–1282.

29.

Deforce

Bastiaens

(2007) The Holocene history of Taxus baccata (yew) in Belgium and neighbouring regions. Belgian Journal of Botany 140(2): 222–237.

30.

Delahunty

(2002) Religion, war, and changing landscapes: An historical and ecological account of the yew tree (Taxus baccata L.) in Ireland. University of Florida unpublished PhD thesis.

31.

Densem

(2001) Report on Archaeological Prospecting Pits at Manor Way/Consul Avenue, Hornchurch Marshes, Dagenham, London Borough of Havering. Compass Archaeology Ltd.

32.

Devoy

RJN

(1979) Flandrian sea-level changes and vegetational history of the lower Thames estuary. Philosophical Transactions of the Royal Society of London B285: 355–410.

33.

Devoy

RJN

(1982) Analysis of the geological evidence for Holocene sea-level movements in southeast England. Proceedings of the Geologists’ Association 93: 65–90.

34.

Divers

(1995a) Archaeological Excavation of the Former Beckton Nursery. Newham Museum Service.

35.

Divers

(1995b) Archaeological Evaluation at Ferndale Street, Cyprus, London E6. Newham Museum Service.

36.

Divers

(1995c) Archaeological Evaluation at the Site of the Proposed East London Sludge Incineration Plant, Beckton Sewage Treatment Works, London E6. Newham Museum Service.

37.

Divers

(1996) Archaeological Investigation of Hays Storage Services LTD, Pooles Lane, Ripple Road, Dagenham, Essex. Newham Museum Service.

38.

Elias

Webster

Amer

(2009) A beetle’s eye view of London from the Mesolithic to the Bronze Age. The Geological Journal 44: 537–567.

39.

Gale

Cutler

(2000) Plants in Archaeology: Identification Manual of Vegetative Plant Materials Used in Europe and the Southern Mediterranean to c. 1500. Otley: Westbury Academic and Scientific Publishing.

40.

Gibbard

(1995) Palaeogeographical evolution of the lower Thames valley. In: Bridgland

Allen

Haggart

(eds.) The Quaternary of the Lower Reaches of the Thames. Durham: Quaternary Research Association, pp. 5–34.

41.

Gifford and Partners Ltd (2001) A13 Thames Gateway Phase III Archaeological Works: Freemasons Road Underpass Assessment Report on Detailed Excavation. Gifford and Partners Ltd.

42.

Girling

(1988) The bark beetle Scolytus scolytus (Fabricius) and the possible role of elm disease in the early Neolithic. In: Jones

(ed.) Archaeology and the Flora of the British Isles. Committee for Archaeology Monograph 14. Oxford University, pp. 34–38.

43.

Godwin

(1940) A Boreal transgression of the sea in Swansea Bay: Data for the study of post-glacial history VI. New Phytologist 39(3): 308–321.

44.

Godwin

(1975) The History of British Flora: Second Edition. Cambridge: Cambridge University Press.

45.

Godwin

Clifford

(1935) Controlling factors in the formation of fen deposits, as shown by peat investigations at Wood Fen, near Ely. Journal of Ecology 23: 509–535.

46.

Grant

Waller

Groves

(2011) The Tilia decline: Vegetation change in lowland Britain during the mid and late Holocene. Quaternary Science Reviews 30: 394–408.

47.

Grimm

(2004) Tilia2 and TGView. Springfield: Illinois State Museum.

48.

Hageneder

(2007) Yew a History. Gloucestershire: Sutton Publishing Ltd.

49.

Halstead

DGH

(1963) Coleoptera: Histeroidae. London: Royal Entomological Society.

50.

Hansen

(1987) The Hydrophiloidea (Coleoptera) of Fennoscandia and Denmark. Fauna Entomologica Scandinavica 18. Leiden: Scandinavian Science Press.

51.

Harde

(1984) A Field Guide in Colour to Beetles. London: Octopus.

52.

Hendey

(1964) An Introductory Account of the Smaller Algae of British Coastal Waters, Part V: Bacillariophyceae (Diatoms). Ministry of Agriculture Fisheries and Food, Series IV.

53.

Holder

(1998) An Archaeological Excavation Assessment and Updated Project Design for Royal Docks Community School Site, Prince Regent Lane, Newham. Museum of London Archaeology Service.

54.

Hughes

PDM

Mauquoy

Barber

. (2000) Mire-development pathways and palaeoclimatic records from a full Holocene peat archive at Walton Moss, Cumbria, England. The Holocene 10(4): 465–479.

55.

Huntley

Birks

HJB

(1983) An Atlas of Past and Present Pollen Maps of Europe: 0–13,000 Years Ago. Cambridge: Cambridge University Press.

56.

Hustedt

(1953) Die Systematik der Diatomeen in ihren Beziehungen zur Geologie und Okologie nebst einer Revision des Halobien-systems. Svensk Botanisk Tidskrift 47: 509–519.

57.

Hustedt

(1957) Die Diatomeenflora des Fluss-systems der Weser im Gebiet der Hansestadt Bremen. Abhandlungen Naturwissenschaftlicher Bremen 34: 181–440.

58.

Jarrett

(1996) Archaeological Evaluation at the East Beckton District Centre. Newham Museum Service.

59.

Jessop

(1986) Dung Beetles and Chafers (Coleoptera: Scarabaeoidea). London: Royal Entomological Society.

60.

Koch

(1992) Die Käfer Mitteleuropas, Ökologie, 1. Krefeld: Goecke and Evers.

61.

Lamb

Thompson

(2005) Unusual mid-Holocene abundance of Ulmus in western Ireland – Human impact in the absence of a pathogen? The Holocene 15(3): 447–452.

62.

Langdon

Barber

Hughes

PDM

(2003) A 7500-year peat-based palaeoclimatic reconstruction and evidence for an 1100-year cyclicity in bog surface wetness from Temple Hill Moss, Pentland Hills, southeast Scotland. Quaternary Science Reviews 22: 259–274.

63.

Lindroth

(1992) Ground Beetles (Carabidae) of Fennoscandia, a Zoogeographic Study, Part I. Andover: Intercept.

64.

Long

Waller

Plater

(2007) Dungeness and the Romney Marsh: Barrier Dynamics and Marshland Evolution. Oxford: Oxbow Books.

65.

Lucht

(1987) Die Käfer Mitteleuropas, Katalog. Krefeld: Goecke and Evers.

66.

Mayewski

Rohling

Stager

. (2004) Holocene climate variability. Quaternary Research 62: 243–255.

67.

Meddens

(1996) Sites from the Thames Estuary Wetlands, England and their Bronze Age use. Antiquity 70: 325–334.

68.

MoLAS (1997) Union Railways, Phase 4, Area 3, 10, Rainham, Purfleet, West Thurrock, London Borough of Havering: Archaeological Watching Brief. Museum of London Archaeology Service.

69.

Moore

Webb

Collinson

(1991) Pollen Analysis. Oxford: Blackwell Scientific.

70.

Morley

(2003) Greenwich Industrial Estate, Bugsby’s Way, Charlton, London SE7, a Geoarchaeological Investigation. Museum of London Archaeology Service.

71.

Newham Museum Service (1995) West End of Frog Lane, A13 Extension (HO-LI95): Archaeological Watching Brief. Newham Museum Service.

72.

Nilsson

Holmen

(1995) The aquatic Adaphega (Coleoptera) of Fennoscandia and Denmark II. Dytiscidae. Fauna Entomologica Scandinavica 32: 1–188. Leiden: E.J. Brill.

73.

O’Connell

Molloy

(2001) Farming and woodland dynamics in Ireland during the Neolithic. Biology and Environment: Proceedings of the Royal Irish Academy 101B(1–2): 99–128.

74.

Orme

Coles

(1989) Prehistoric woodworking from the Somerset Levels: 2. Species selection and prehistoric woodlands. Somerset Levels Papers 15: 6–24.

75.

Parker

Goudie

Anderson

. (2002) A review of the mid-Holocene elm decline in the British Isles. Progress in Physical Geography 26(1): 1–45.

76.

Pearce

(1957) Coleoptera (Pselaphidae). London: Royal Entomological Society.

77.

Peglar

Birks

HJB

(1993) The mid-Holocene Ulmus fall at Diss Mere, south-east England – Disease and human impact? Vegetation History and Archaeobotany 2: 61–68.

78.

Pepys

(1665) Samuel Pepys Diary September 1665. Available at http://www.pepysinfo/1665/1665sep.html accessed 20 April 2007.

79.

Perry

Moore

(1987) Dutch elm disease as an analogue of Neolithic elm decline. Nature 326: 72–73.

80.

Potter

(2003) Former Manser Works, 137–139 New Road, Rainham, Essex, LB Havering: Interim Archaeological Post-excavation Assessment. Compass Archaeology Ltd.

81.

Reille

(1992) Pollen et spores D’Europe et D’Afrique du Nord. Marseille: Laboratoire de Botanique Historique et Palynologie.

82.

Reimer

Baille

MGL

Bard

. (2004) IntCal04 terrestrial radiocarbon age calibration, 0–26 cal kyr BP. Radiocarbon 46(3): 1029–1058.

83.

Scaife

(1988) The elm decline in the pollen record of south-east England and its relationship to early agriculture. In: Jones

(ed.) Archaeology and the Flora of the British Isles. Committee for Archaeology Monograph 14. Oxford University, pp. 21–33.

84.

Scaife

(2001) East Ham FC, Pennyroyal Avenue, Beckton, London E6, an Environmental Assessment. Hertfordshire Archaeological Trust Report 903.

85.

Seel

SPS

(2001) Late prehistoric woodlands and wood use on the lower Thames floodplain. University College London, Unpublished PhD Thesis.

86.

Sidell

(2003) Relative sea-level change and archaeology in the inner Thames Estuary during the Holocene. University College London, Unpublished PhD Thesis.

87.

Sidell

Scaife

Wilkinson

. (1996) Spine Road Development, Erith, Bexley (Site 2649): A Palaeoenvironmental Assessment. Museum of London Archaeology Service.

88.

Sidell

Wilkinson

Scaife

. (2000) The Holocene Evolution of the London Thames. Museum of London Archaeology Service Monograph 5.

89.

Smith

(1981) The Neolithic. In: Simmonds

Tooley

(eds) The Environment in British Prehistory. London: Duckworth, pp. 125–209.

90.

Spurrell

FCJ

(1889) On the estuary of the Thames and its alluvium. Proceedings of the Geologists’ Association 11: 210–230.

91.

Stace

(1997) New Flora of the British Isles. Cambridge: Cambridge University Press.

92.

Stuiver

Reimer

(1986) A computer program for radiocarbon age calculation. Radiocarbon 28(2B): 1022–1030.

93.

Tamblyn

(1994) Archaeological Evaluation of Beckton Areas A and B Tollgate Road, Beckton, London Borough of Newham. Newham Museum Service.

94.

Thomas

Rackham

(1996) Bramcote Green, Bermondsey: A Bronze Age trackway and palaeoenvironmental assessment. Proceedings of the Prehistoric Society 61: 221–253.

95.

Thomas

Polwart

(2003) Taxus baccata L. Journal of Ecology 91: 489–524.

96.

Tittensor

(1980) Ecological history of yew Taxus baccata L. in southern England. Biological Conservation 17: 243–265.

97.

Troels-Smith

(1955) Characterisation of unconsolidated sediments. Danmarks Geologiske Undersøgelse, Raekke IV(3): 38–73.

98.

Truckle

Sabel

(1994) Beckton Alps Ski Slope, Newham Way, East Ham E6, London Borough of Newham. Newham Museum Service.

99.

Turner

(1962) The Tilia decline: An anthropogenic interpretation. New Phytologist 61: 328–341.

100.

Waller

(1994a) The Fenland Project, Number 9: Flandrian Environmental Change in Fenland. Cambridge: Cambridgeshire County Council.

101.

Waller

(1994b) Paludification and pollen representation: The influence of wetland size on Tilia representation in pollen diagrams. The Holocene 4: 430–434.

102.

Waller

Hamilton

(2000) Vegetation history of the English chalklands: A mid-Holocene pollen sequence from the Caburn, East Sussex. Journal of Quaternary Science 15: 253–272.

103.

Waller

Schofield

(2007) Mid to late Holocene vegetation and land-use history in the Weald of south-eastern England: Multiple pollen profiles from the Rye area. Vegetation History and Archaeobotany 16: 367–384.

104.

West

(1962) A note on Taxus pollen in the Hoxnian Interglacial. New Phytologist 61: 189–190.

105.

Wessex Archaeology (2000) Fort Street (West) Silvertown, London, E16, Archaeological Excavation Assessment Report. Wessex Archaeology.

106.

Wilkinson

(1988) Archaeology and Environment in South Essex: Rescue Archaeology along the Grays by-pass, 1979/80, Number 42. East Anglian Archaeology.

107.

Wilkinson

Murphy

(1995) The Archaeology of the Essex Coast, Volume 1: The Hullbridge Survey. East Anglian Archaeology.

108.

Wilkinson

Scaife

Sidell

(2000) Environmental and sea-level changes in London from 10,500 BP to the present: A case study from Silvertown. Proceedings of the Geologists’ Association 111: 41–54.