This report contains an assessment of the palaeo-environmental samples taken from the excavation. The report was written by Mike Allen of Allen Environmental, but has been edited for web by Chaz Morse and Guy Hunt. For a full index of the assessment reports, please see the assessment report index.

Summary of archive

Class Quantity
Bulk samples 293 Environmental samples


This document provides an assessment of the charred plant and charcoal remains, waterlogged plant remains, insects from the site as well as the geoarchaeology of the soils, context (979) and (959).


Samples were taken by the excavator from deposits considered to be of high potential for the recovery of environmental remains. A preliminary appraisal of these samples was undertaken those considered to be of significant archaeological importance were assessed. A total of 293 samples were taken from the site, of these 152 were considered for assessment.

A large suite of 293 bulk samples were taken from the excavation. These included samples for general plant macrofossil recovery, as well as samples from cremation burials for human bone and charcoal/charred plant remains recovery/assessment, and samples from inhumations for the recovery of smaller human bones (i.e. phalanges etc). In addition there were a small number (10) of potentially waterlogged samples. These were appraised on the basis of context security and priority.

Processing and flotation methods

Samples deemed as of low priority (and void) were weeded out were not processed in the first instance (i.e. 56+4), leaving 233 high and medium priority samples. All of the 233 high and medium priority samples were processed by standard flotation methods with normally 40-litre samples being fully processed and the flots retained on a 0.5mm mesh and residues on a 1mm/2mm mesh by S. Campion (L – P : Archaeology). Flot sizes varied from 5ml to 200ml- though the majority were large at about 190ml. After flotation sorting was conducted according to recovery aims; i.e. samples from skeletons taken for the recovery of phalanges etc; cremation-related samples were processed to facilitate recovery of residue bone as well etc). The coarse residues (>4mm) of all samples have been extracted. The numbers of bulk samples are, therefore, better summarised material type for which they were taken. The flots and coarse fractions of the samples from inhumations were sorted for human bone and do not form part of the palaeo-environmental assessment.

Pre-assessment appraisal

The samples (flots and residues) were rapidly appraised by S. Campion and M. Bamforth (L-P Archaeology) to aid in defining the necessity for processing, and aid in removing samples not worth progressing to assessment.


Of the 88 cremation samples, 82 produced flots, which were provided for assessment for charcoal and charred plant remains. Six samples produced no charred remains in the flots.

Inhumation Samples

No material from samples taken for recovery (sieving) of human bone was provided for palaeo-environmental assessment of the charred plant and charcoal remains, as material in grave backfill is largely incidental to the burial, and better samples are provided from these periods from more suitable contexts.

Charred plant and charcoals.

All of the high and medium priority samples (67 flots) were provided for assessment of the charred plants and charcoal remains.

Waterlogged Plant Remains

The waterlogged samples were subsampled into waterlogged plant remains (1-2 litres) and insects (10 litres). The remaining sample was processed by flotation for artefact recovery, following standard methods. The waterlogged flots and residues were retained on 250μm and 0.5mm mesh sieves. Of the ten waterlogged samples 9 were high priority and 1 low priority – only 4 were waterlogged. All high priority samples were sent for assessment.

A series of 11 samples were deemed as potentially from waterlogged deposits on site. Subsamplesof 1-2 litres were processed by standard flotation methods with the flot recovered on a 250μm sieve and residues on 1mm mesh sieve. The flots and residues were rapidly appraised by S. Campion and 6 samples were deemed not waterlogged.


The environmental project conforms to relevant sections of the Environmental Archaeology: a guide to the theory and practice of methods, from sampling and recovery to post-excavation (English Heritage 2002).

Charred and Waterlogged Plant Remains – By Alan J. Clapham

A total of 152 samples from 145 contexts were assessed for charred plant remains. The majority, 82 samples from 76 contexts were from Roman cremations. Sixty-five samples from 64 contexts were from non-cremation contexts of which 2 were pre-Roman, 49 were Roman, and 13 post-Roman in date. Five samples from five waterlogged contexts all post-Roman in date were also assessed.

The aims of the assessment were to determine the state of preservation, type, and quantity of environmental remains recovered, from the samples and information provided. This information will be used to assess the importance of the environmental remains.

Assessment method

The flots were scanned using a low power stereo light microscope and plant remains identified using modern reference collections maintained by the Worcestershire Historic Environment and Archaeology Service, and a seed identification manual (Cappers et al, 2006). Nomenclature for the non-cereal plant remains follows (Stace 1997).

Assessment results

Overall, plant remains either charred or waterlogged were found in 85 of the 152 samples assessed. Of the 82 Roman cremations, 30 produced plant remains. Of the 65 non-cremation contexts, 50 contained plant remains. These non-cremation contexts can be divided into pre-Roman, Roman and post-Roman. The two pre-Roman contexts did not produce any plant remains. Of the 50 Roman contexts, 40 produced remains and 10 of the 13 post-Roman context contained plant remains. The five waterlogged contexts all produced plant remains. The preservation of the charred and waterlogged plant remains was good enough in most cases to permit identification to species where possible. The waterlogged contexts were richer than the charred samples although in those contexts assessed for charred remains also contained uncharred seeds, these are most likely to be waterlogged and of a similar date to the charred material rather than modern intrusive seeds.

Fragments of charcoal were also noted in many of the samples that contained plant remains but in most cases the fragments were too small to permit identification. Larger fragments were extracted at the appraisal stage to be assessed elsewhere.

Roman cremation burials

Thirty of the 82 cremation contexts contained plant remains and these were quite sparse. Some cereal remains in the form of wheat (Triticum sp.) and barley (Hordeum sp.) grains were identified from some of the contexts but again they were in limited numbers. No cereal chaff was recovered from any of the contexts. Other plant remains identified included weed seeds, some of which may be associated with crop processing activities but it is more likely that they represent a ‘background flora’. Other species identified may be more significant and may represent offerings. Possible hemp (cf. Cannabis sativa) was recorded from contexts (1170) and (1174), and opium poppy (Papaver somniferum) was found in cremation urn (1499). Broad bean (Vicia faba) was identified from context (1173) and context (1526).

A charred duckweed (Lemna sp.) seed was found in (1170) and as this is an aquatic species it may suggest that water containing the seeds was used to douse the funeral pyre. Other aquatic or wetland species such as spike-rush (Eleocharis sp.) and sedges (Carex spp.) were also identified from some of the contexts.

Several fragments of an amorphous mass were identified from several of the contexts, (1170), (1173), (1331), (1631), (2112), and (2131). These fragments could be of badly distorted charcoal, parenchyma (such as a tuber or rhizome), or possibly burnt bread. Another possibility is that they may be a product of the cremation process itself.

Uncharred seeds were found in some of the contexts and these have been interpreted as being waterlogged and being of a similar date to the charred remains. Several contexts ((1814), (1839), (1999)) contained waterlogged examples of possible hemp seeds with flax seed (Linum usitatissimum) also being found in (1814). Other species such as mallow (Malva sp.), elderberry (Sambucus nigra) and smooth sow-thistle (Sonchus oleraceus) were also recovered in small numbers.

Non-cremation contexts

Sixty-five samples from 64 contexts were assessed for plant remains. Two were of pre-Roman date, 50 Roman and 13 post-Roman. A variety of features were represented including pits, linear features, layers and other miscellaneous features. Of the 65 samples 50 produced plant remains.


Two contexts (875) and (876) were pre-Roman in date and consisted of two fills from a palaeochannel. No plant remains were recovered.


Fifty Roman contexts from non-cremation features were assessed of which 40 produced plant remains, both charred and waterlogged. Charred plant remains were not common and consisted mainly of cereals grains such as wheat and barley. No cereal chaff was identified from these contexts. Weed seeds present were those usually associated with crop processing or may have been part of a ‘background flora’. There were several species which were of interest. Possible hemp seeds were identified from contexts (1199), (1201), and (1163), and broad bean from box drain fill (1502), apple (Malus sp.) pips were also found in this context. A charred flax seed was found in the pit fill (1621).

Waterlogged material was found in the majority of the samples and mainly consisted of plants that may have been growing around the site. Of interest were the larger numbers of fig (Ficus carica) seeds found in contexts (1502), (1605), (1621), (1969), (2001), (2003), (2011), (2012), and (2090). Grape seeds (Vitis vinifera) were found in box drain [1406] fill (1502) and quarry pit[2004] fill (2003). Broad bean remains were also found in box drain [1406] fill (1502). Peas (Pisum sativum) were found in pit fill (1163) the fill if cut [1164] (context (1163)) and linear (1603) the fill if cut [1605]. It is possible that these remains represent cess.

Duckweed seeds were present in pit fill (1183) the fill if cut [1184] suggesting that there may have been some local flooding and the presence of blinks (Montia fontana ssp. chondrosperma) and sedges may indicate damp areas.


Ten samples from thirteen contexts assessed contained plant remains. Very little charred plant material was present in these contexts and consisted of cereal grains (wheat and barley). No cereal chaff remains were recovered. The small numbers of weed seeds found most likely represent a ‘background flora’. Waterlogged plant remains were more common and were dominated by weed seeds which may have either have been growing locally or may be from the remains of crop processing activity. The most common seed in the contexts was fig which was found in pit fills (1009), (1011), (1012), (1342). A wetland component to the assemblages may be represented by the sedge nutlets.

Waterlogged contexts

Five samples from waterlogged contexts produced waterlogged plant remains. All five samples were rich but those from quarry pit (642), (885), and (879) the fills of [639], [890], and well [506]. Cereal remains in the form of wheat glume bases were found in quarry pit [639], from fill (506). The majority of the remains were mainly of weed seeds which are most likely associated with crop processing, although some are indicative of wetland areas. Other crops present include flax seeds in well [506] (context (879)) and possible hemp in quarry pit [639] (context (642)). The large number of fig seeds in soakaways [1286] and [1683] (contexts (1282) and (1686)) and to a lesser extent in quarry pit [890] (context (885)) may suggest the presence of cess material. This is further supported by the large number of raspberry (Rubus idaeus), bramble (Rubus Sect Glandulosus) and strawberry (Fragaria vesca) seeds in soakaways [1286] and [1683] (contexts (1282) and (1686)). Other fruit remains include walnut nutshell fragments (well [506], context (879)), plum (Prunus domestica ssp domestica) in soakaway [1683](context [1686]), cherry (Prunus avium/cerasus) in quarry pit [639] and soakaway [1683] (contexts (642) and (1686) respectively), apple in well [506], (context (879)) and grape in quarry pits [639], [890], and soakaways [1286] and [1683] (contexts (642), (885), (1282) and (1686) respectively). One unusual find is that of tomato (Lycopersicon esculentum) seeds in well [506] (context [879]). This may represent modern day contamination. Cotton thistle (Onopordium acanthium) in quarry pit [890] (context (885)) may represent a garden element of the assemblage.


Of the 152 samples, 85 produced identifiable plant remains whether charred or waterlogged. The most significant contexts of this assessment were the 82 Roman cremation samples of which 30 produced plant remains. The majority of the 30 contexts contained sparse plant remains and may represent a ‘background flora’, although a few contexts did produce some interesting remains. Cremation burials [1169] and [1174] (contexts (1170) and (1173)) produced possible hemp seeds and opium poppy was found in cremation urn [1499] (context (1501)) and broad bean in cremation burial[1174] (context (1173)(1173)) and context (1526). It is possible that these plant remains may indicate offerings made at the time of cremation. Waterlogged possible hemp was also found in the cremation contexts. An amorphous material which may be either bread or parenchyma may also indicate some ritual associated with cremation practices.

The non-cremation contexts again produced very little in the way of plant remains although two of the Roman contexts (in [1199] and [1164]) did contain possible hemp seeds (contexts (1201) and [1163]) and box drain [1406] (context (1502)) contained broad bean remains. Peas were found in pit [1164] and box drain [1406](contexts (1163) and (1502)).

The post-Roman contexts also contained very few plant remains and mostly consisted of weed seeds that may have been growing in the vicinity of the features.

The post-Roman waterlogged samples most likely represent cess and the presence of various fruit species, especially fig support this idea. The presence of a few tomato seeds in well [506], from context (879), may indicate some modern contamination.

Although there are quite a large number of weed seeds present in the contexts, the lack of cereal chaff suggests that there is little if no crop processing occurring on site. This may be a reasonable conclusion given the fact that the majority of the area consisted of cremation deposits. It is most likely that the weed seeds were growing on the site and are part of the natural background flora.

The remains from the post-Roman deposits suggest that foodstuffs were imported onto site.


The most significant aspect of this assessment is the presence of possible funerary offerings in the Roman contexts, especially in the cremation deposits. Archaeobotanical evidence for funerary offerings in the United Kingdom are limited and therefore these results provided additional information to that already obtained.

Charcoal – By Alan J. Clapham, Michael J. Allen & Steven Campion

The charcoal remains from the 152 samples for assessment were rapidly appraised (by M.J. Allen and S. Campion) with regards to size and quantity of fragments. The charcoal fragments from the flots were mostly finely comminuted with the larger fragments being retained in the residue. Overall, the charcoal is well preserved and should permit identification of the majority of pieces. Large quantities of charcoal were recovered from over half of the cremation contexts assessed. In the non-cremation contexts there were fewer contexts with large numbers of identifiable charcoal fragments. From a brief scan of the charcoal it appears that the charcoal from the cremations is mostly of heartwood with some smaller twiggy material. Some of the charcoal from the cremations was vitrified and distorted suggesting that the wood was exposed to high temperatures. The charcoal fragments from the non-cremation deposits is more variable with some contexts having mostly heartwood, and others consisting of small branch wood, indicating a possibility of woodland management.

Roman Cremation contexts

The six contexts selected for full charcoal analysis mostly came from cremation pits, although [1507] is from a dump and [1526] is an unknown feature type. Two of the samples selected, [1170] and [1173] also had possible charred hemp (Cannabis sativa) seeds present as indicated in the plant macrofossil assessment.

Roman Features

The seven non-cremation contexts selected for full charcoal analysis are from a variety of feature types including pits, box drain [1406] and dumps.

Post Medieval

The two post Medieval contexts are from a pit and a quarry pit.


The selection of contexts for full charcoal analysis will address the following questions:

Is there was any evidence for the selection of certain tree species for the cremations? This may help determine pyre technology.

Is there any evidence for woodland management which may have influenced tree species selection?

Are there any specific properties i.e. sweet smelling when burnt, that may have influenced selection for the cremation deposits?

If there was no selection of specific woods for the cremations is it possible to determine the make-up of the local woodland?

Are the charcoal species in contexts (1170) and [1173] any different from the other cremation assemblages?

Is there any difference between the wood species identified in the cremation deposits and those from the Roman non-cremation deposits?

Is there a wider range of wood taxa in the Roman non-cremation deposits than there is in the cremations?

Is there any difference in size of wood used in the cremation and non-cremation deposits?

Is it possible to reconstruct the make-up of the local woodland from the Roman non-cremation deposits?

Is there any difference between the taxa used in the Roman occupation and that used in the post Medieval?

Is there any evidence for woodland management in the post Medieval deposits?

Is it possible to detect any changes in the woodland composition between the Roman and post-medieval deposits?

Coleoptera – By Damian Evans and Michael J. Allen

A series of five samples were deemed to be potentially waterlogged as a result of the site conditions, and rapid appraisal of the flots and residues from samples processed for plant remains. All of these samples were from post medieval contexts and included a well [506] quarry pits [639] and [890] and soakaways [1286] and [1683].


Subsamples of 2-4litres of the five samples were processed by standard paraffin flotation methods (Shackley 1981) by Laura Evans; samples were wet sieved through a mesh of 300μm, and then paraffin floated and insects decanted onto 300μm mesh sieve. The processed samples were scanned under a stereo-binocular microscope by Damian Evans; three of the five samples fully sorted, the remaining two were partially sorted.


The three fully sorted samples produced few, small and fragmented remains of Coleoptera and a few fly pupae. The two late post medieval samples that were only partially sorted (sample 10 from well [506] and sample <149> from soakaway (1683)) contained only few remains, and rapid scan and partial sorting and did not record any immediately identifiable beetle fragments. The few beetle remains that are present were small and fragmented Curculionidae spp. and Carabid spp. but indicate open country / garden-type habitats. There are a few fly pupae which may suggest occupation debris, but there are not enough present to be able to suggest for farm or coprolite remains.


The species present in open country to garden/waste ground habitats are all generalist species and little further interpretational information is likely to be gained. Further all deposits are from post Medieval contexts, and these clearly cannot address the environment and habitats of the main period of interest; the Romano-British period.

Geoarchaeological Assessment – By Damian Evans and Michael J. Allen

An ‘agricultural soil’ was reported to occur across the site and is believed to be late to early post Roman or Medieval/pots medieval in date (Hunt & Morse 2006). Excavation has roduced truncated Romano-British cremation burial urns, but very few artefacts within the body stratigraphy; just a few sherds of Romano-British pottery and possibly some medieval sherds . This stratigraphy was exposed in section, and examined, described and sampled. The stratigraphic sequence and buried ‘agricultural soil’ was describes and characterised. On-site interpretation of the deposits and a working hypothesis for the sequence was given (Allen 2008). A series of four samples of undisturbed soil / sediment were removed in kubiena tins (descriptions below), and these were accompanied by five small (100g) bulk samples (List of samples, below) which could facilitate analysis of particle size, soil chemistry y (Po, Pi, MS etc), heavy metals.

Pedological and sedimentological record of the ‘agricultural soil’ and associated deposits are given below where nomenclature follows that outlined in Hodgson (1976).

From on descriptions and examination and limited rapid examination of the sequence, a clear buried humic soils (Bah2) was discerned overlying the brickearth (context (979)). This clearly has been disturbed and mixed by anthropgenic activity and includes typical occupation detritus such as marine shell (largely oyster) and Romano-British pottery. The yellowish green hues and patches suggest the inclusion of cess material. This was sealed by a developed soil agricultural soil profile (context (979)), probably derived from the accumulation and tillage, of humic debris, dumped soil material and colluviated soil. The upper part of this profile presents aclear ploughed topsoil, while the lower profile is developed soil (bB/C or bB/Rw). this in turn was sealed by the upcast and dumping of brickearth-type parent material sealing and preserving the agricultural soil. Thus the following sequence of events can be postulated:-

The ‘natural’ comprises sorted weakly bedded gravel and pockets of silty sand loam ‘brickearth. This is the parent material / source of many of the deposits, including the ‘agricultural soil, (979) and the overlying deposit (852), and deposits yellowish brown (‘orange’) deposits to the south of the described and sample location.

Surface gravels form a hard standing, and the addition of dumped organic and humic matter including cess and fragmented marine shell (middening and waste) enabled some pedogenesis (soil formation) to occur. There seems to be evidence of trampling and mixing, possibly by animals: -> context (959).

‘Brickearth’ sensu lato material is dumped over this courtyard and humic ‘soil’ surface, providing the parent material for the cultivtion soil.

Addition of humic matter (but not artefacts), and mixing facilitating limited pedogenesis (soil formation), and increasing humic content. This incipient soil shows signs of physical mixing (?spading, cultivation or ploughing for ?cultivation or horticulture): -> context (979).

Elsewhere deep physical mixing did not encompass the full depth of the deposited parent material, and this can be seen in patches in the section (to the south of the sampled location), and these might represent uncultivated baulks or pathways between field or horticultural plots, or may just be isolated stands of unmixed deposit. Soil formation is limited, and only incipient pedogensis seems present.

Further stonefree ‘brickeath’ was dumped on the ‘cultivation soil’ possibly as upcast from numerous pit digging in the vicinity. Initially this was a thin shallow spread at the location described and sampled, as some pedogenic reworking of this deposit can be seen into the active ‘cultivation soil’; -> base of context (852).

Further massive dumping of mixed stonefree ‘brickearth’ and topsoil upcast. Some but limited physical mixing present suggesting dumping of large volumes of material. This may not have created uniform layer across the area. This deposit (852) does not constitute a soil or an agricultural deposit.

At some considerable period later a humic Ah hrosion, possible of a ‘garden soil’ formed over the dumped brickearth (852). Worms and roots penetrated through the redeposited and dumped brickearth bringing sandier and humic material down worm and root voids (as recognised in context (852).

The four kubiena samples taken for consideration for soil micromorphological analysis have been prepared as soil micromorphological slides. The accompanying impregnated blocks will serve as a permanent archive.


Analyses of soil/sediment micromorphology accompanied with a limited suite of soil chemistry (i.e. Po, Pi, MS etc) would enable the following questions to be addressed and considered of the sequence.

Basal humic yard ‘soil’ (959)
What anthropogenic activity is associated with the formation of the yard deposit (959)?

Is there evidence of animal trampling, stabling etc.?

What is the origin of the highly humic nature of the deposit?

Is there evidence of soil formation in this humic deposit?

‘Agricultural soil’ (979)
What is the parent material of this soil? Has it been dumped over(959) as an essentially single mass?

Is this an old land surface/buried soil? Was it ploughed / disrupted and is this ploughing, arding or spading?

What has been added to the soil? Is this domestic maddening, domestic waste, stabling waste?

Upcast brickearth and soil (852)
What is the character of this deposit? Can analysis confirm its origin?

What is the mode of deposition, and any post deposition anthropogenic events?


The analytical recommendations are listed below based on the assessment of the palaeo-environmental and geoarchaeological material, the archaeological context and the project priorities.

Charred and waterlogged plants

Further analysis of the Roman contexts which contain the possible hemp seeds and other possible funerary offerings, these are: (1170), (1173), (1501), (1526) and also (1814), (1839), (1999) from the cremations in order to elucidate funerary practice.

Further analysis of the Roman non-cremation contexts which also contain possible hemp these are: (1201), (1163) and (1502) for the finds of broad bean.

No further analysis needs to be carried out on the post-Roman material as the assessment has provided enough detail on the nature of these deposits.


Samples for full analysis then selected on the basis of suitable quantities of charcoal and the archaeological significance of the features. A total of 15 samples were selected for further analysis. Six of the samples were from the cremation deposits and ten were from the non-cremation deposits, of which seven are Roman in date and two post Medieval.


No further analytical work is necessary. The information reported in the assessment should be included in the reporting of the relevant features or period as relevant.


The four soil micromorphological slides should be analysed (Dr. R.I. Macphail) to address the questions outlined above.

In addition five disturbed samples should be analysed for basic soil chemistry properties, in order to quantify better amounts of organic matter, inputs of phosphate from food and faecal waste, and any additions of burned materials that may relate to manuring or middening practice and general use of space through time during the accumulation.