SECTION 4.4 THE INVERTEBRATE
AND WATERLOGGED PLANT REMAINS
|
961 (Trench 14) |
Sub-soil hollow. Grey clay with patches of fine gravel and sand beneath a layer of gravel. |
965 (Trench 14) |
Lower fill of a Late Glacial channel. Grey Chara marl with a few waterlogged plant remains evident. Underlay Sample 464/A. |
464/A (Trench 14) |
Upper fill of a Late Glacial channel. Top 0.05 m of a layer of dark brown strongly-laminated peat. Above Sample 965. |
371/D/3 (Trench 14, Phase 1, SB20) |
Lower fill of a middle Iron Age (c.3rd to 1st century BC) enclosure ditch on Island 3. Dark grey silty clay with iron-pan flecks and a little limestone gravel. |
153/C/5 (Trench 8, Phase 1, E3) |
Lower fill of a middle Iron Age enclosure ditch on Island 2. Grey- brown clay loam with limestone gravel. |
577/A (Trench 13, Phase 2c, E16) |
Bottom layer of a mid-late 1st century AD enclosure ditch in Longdoles Field. Dark brown, somewhat organic, gravelly loam. |
577/N/4 |
Mid-late 1st century AD. Dark brown, somewhat organic, gravelly loam. |
1528/A/3 (Trench 13, Phase 2c, E17) |
Lower fill of a mid-late 1st century AD enclosure ditch in Longdoles Field. Dark brown organic silt with some gravel. |
1704/C/5 (Trench 13, Phase 2c, E17) |
Bottom layer of a mid-late 1st century AD enclosure ditch in Longdoles Field. Dark brown somewhat gravelly organic loam. |
2490 (Trench 13, Phase 2) |
Bottom layer of an early 1st - early 2nd century AD enclosure ditch in Longdoles Field. Grey organic silty clay with lenses of gravel and sand. |
6 (Trench 1, Phase 1/2, E1) |
Bottom layer of a 1st century BC - 1st century AD enclosure ditch on Island 1. Dark brown organic gravelly loam. |
962/A |
Water-hole or sump cut by Roman trackway ditch between Islands 2 and 3. Brown organic silt with some sandy lenses. Cut thorny twigs of Crataegus or Prunus sp. were conspicuous. Stratigraphically this feature could be late 1st century AD or early Roman. However, on the basis of the plant and insect remains it has been thought more likely to belong to the late 1st century AD phase. In particular, there was a high proportion of scarabaeoid dung beetles and a presence of seeds from a group of weeds of nitrogen-rich disturbed ground, including Chenopodium ficifolium, which were very abundant in some of the 1st century AD samples but almost absent from the early Roman samples. |
766/6 (Trench 13, Phase 3b/c) |
Organic deposit from the bottom of an early Roman (mid 2nd to late 3rd century AD) well in Longdoles Field. Grey gravelly silt with brown lenses of organic material. |
2160/B/5 (Trench 13, Phase 3) |
Organic deposit from an early Roman pit in Longdoles Field. Dark grey-brown organic silt with calcareous sand. Hazel nuts were much in evidence. |
1202/15 (Trench 17, Phase 3) |
Organic deposits from the bottom of an early Roman water-hole or well in Longdoles Field. Brown organic silt with some lenses of sand and grit. |
1318/C/3 (Trench 17, Phase 3/4) |
Layer of organic material from an early-mid Roman hole pit or water-hole in Longdoles Field. Compressed dark brown laminated peat mostly of monocotyledonous plant material. |
2839/A/5 (Trench 29, Phase 3) |
Bottom layer of an early Roman well or water-hole in Longdoles Field. Dark brown organic silt with some gravel. |
2867/A/6 |
Bottom layer of an early Roman well or water-hole in Longdoles Field. Dark brown organic silt with some gravel. |
2906/4 (Trench 29, Phase 3b/c) |
Organic material around a cone of stone pine (Pinus pinea) in an early Roman water-hole or well in Longdoles Field. Dark grey-brown organic gravelly silt. |
1487/A (Trench 32) |
Bottom layer of the early Roman roadside ditch west of the main settlement. Brown organic silt with leached limestone grit and sand. |
2495/A/2 (Trench 34, Phase 3/4) |
Peat layer from an early-mid Roman field ditch about 100 m west of the main settlement. Dark brown mossy peat with leached shell fragments. This feature could have been part of either the early Roman or the late Roman ditch system. However, there was a similarity shown by some elements of the hay meadow flora from this sample with the hay assemblages from the early Roman settlement. Seeds of the rare plant Baldellia ranunculoides were found in this sample and Sample 1318/C/3 but have not been discovered in any other sample from the Upper Thames Valley. It is therefore thought likely that this deposit was of early Roman date. |
502/7 (Trench 13, Phase 3d/4a) |
Bottom layer of a late Roman (4th century AD) stone-lined well in Longdoles Field. Dark grey organic gravelly and gritty loam. |
700/S/3 (Trench 13, Phase 4c) |
Organic layer from the late Roman (4th century AD) villa enclosure ditch in Longdoles Field. Dark grey-brown organic silty clay. |
1989/B (Trench 13, Phase 4a/b) |
Bottom layer of a late Roman tank in Longdoles Field. Dark grey-brown organic clay loam with box leaves evident. |
2721/6 (Trench 13, Phase 4d) |
Bottom layer of a late Roman pit or well in Longdoles Field. Brown organic silt. |
2721/3 (Trench 13, Phase 4d) |
Organic layer above 2721/6 in a late Roman pit or well in Longdoles Field. Grey-brown organic silty clay with some sandy gravel lenses and a fragment of an oak plank. |
‹3›-‹1› (Feature 501, Trench 13, Phases 4d-5) |
Sequence of samples through the late Roman villa enclosure ditch in Longdoles Field. |
Sample 1 |
Bottom sample of the sequence, being the top of the Roman gravelly fill of the villa ditch. Dark grey gravelly silty clay. |
Sample 2 |
Middle sample of the sequence from the villa ditch. Probably post-Roman. Dark grey-brown silty clay with a little eroded limestone gravel. |
Sample 3 |
Uppermost sample of the sequence from the villa ditch. Medieval. Buff shelly alluvial silty clay. This deposit occurred extensively in the top of Roman and earlier features, also being represented by Sample 456. It contained pottery of medieval date. |
456 (Trench 14, Phase 5) |
Part of the same medieval alluvial deposit as Sample ‹1› from the edge of Island 3. Yellowish-buff shelly silt. |
696/3 (Trench 13, Phase 5) |
Bottom layer of a stone-lined medieval well (a coin of 1473-77 was recovered from its fill). Dark brown somewhat gravelly organic loam. Below Sample 696/2. |
696/2 (Trench 13, Phase 5) |
Organic layer above 696/3 in a stone-lined medieval well. Gritty grey-brown organic silt. |
The procedures used generally followed those described in Lambrick and Robinson (1979, 79-80). Samples for waterlogged macroscopic plant remains only and those for plant remains, insects and molluscs were water-sieved down to 0.2 mm and sorted under water with the aid of a binocular microscope. A wash-over technique was used so that the light organic fraction could be passed through the stack of sieves and sorted before any mineral material from the sample was sieved, thus reducing the damage to organic items. The contents of all the sieves were sorted down to an aperture size of 0.5 mm. A tenth sub-sample of the fraction between 0.5 mm and 0.2 mm was sorted for macroscopic plant remains and the numbers of these items were multiplied by ten for inclusion in the tables of results. After sorting, all the residue from this fraction was then added to the insect sub-sample. Samples for insects were washed over a 0.2 mm sieve and then the residues subjected to paraffin flotation. After washing with hot water and detergent, the flots were sorted for insect remains. Samples for molluscs only were water-sieved down to 0.5 mm, dried and sorted.
Sorted macroscopic plant remains and insects were stored in 70% ethanol, mollusc shells were dried to await identification. Specimens were identified with reference to the various collections housed in the University Museum, Oxford.
The identifications from the samples have been listed in Tables 4-32.
The results of Dr Turner's analyses are given in Tables 8, 14 and 20. She reports that the type X pollen grains from Sample 502/7 were in poor condition and appeared to be contemporaneous with the rest of the assemblage. Type X could not be matched with anything in her reference collection. She describes it as tricolpate, eureticulate and slightly longer than broad, eg (25 x 20μm) (20 x 18μm) (22 x 20μm). The luminae are under 1μm and the muri consist of a single row of columellae. (This could only be seen in the very best-preserved grains). The large amount of pollen type X might indicate a more local origin for it than the rest of the assemblage.Fig 1 Pollen diagram
The results of Dr Greig's analyses are given in Table 27. He reported preservation to be good.
Nomenclature in the Tables follows Clapham et al. (1987). A summary pollen diagram is given by period in Fig. 4.4.1: Summary pollen diagram .
The results for the identification of waterlogged seeds are given in Tables 5, 44, 9, 15, 21 and 28. Nomenclature follows Clapham et al. (1987).
The results for the identification of other waterlogged plant remains are given in Tables 6, 55, 10, 16, 22 and 29. In some cases the number of items has been given, in others only presence or absence has been indicated. Nomenclature follows Clapham et al. (1987). Wood was present in some of the samples but has not been identified.
Miss Tomlinson examined some of the compressed, laminated plant-material which comprised Sample 1318/C/3. She reported that the first fragment of vegetative material which she examined has six papillae overlapping each stomata and rows of larger papillae. She believes it to be from a member of the Cyperaceae. Other items in the material included fragments of Gramineae and some possible dicotyledon stems.
Dr F Rose identified the moss which formed a significant part of Sample 2495/A/2 as Calliergon giganteum (Schimp.) Kindb.
Minimum numbers of individuals represented by the fragments identified from the samples are given in Tables 11, 17, 23 and 30. Coleopteran remains were present in Samples 965 and 464/A but were not identified. Coleopteran remains were absent from Sample 2721/3. Nomenclature follows Kloet and Hincks (1977). A Coleoptera diagram by period is given in Fig. 4.4.2: Coleoptera diagram by period , species groups following Robinson (1991, 278-81).
Minimum numbers of individuals represented by the fragments identified from the samples are given in Tables 12, 18, 24 and 31. Other insect remains were present in Samples 965 and 464/A but were not identified and were absent from Sample 2721/3. Nomenclature follows Kloet and Hincks (1964) for the small orders and Hemiptera, Kloet and Hincks (1978) for the Hymenoptera and Kloet and Hincks (1976) for the Diptera.
Minimum numbers of individuals of Mollusca from each sample are given in Tables 4, 66, 7, 13, 19, 25, 26 and 32. Mollusc shells were present in Sample 965 but not identified and absent from Samples 1704/C/5, 1318/C/3 and 2721/6. Nomenclature follows Kerney (1976) for freshwater molluscs and Waldén (1976) for land snails.
Many of the flots from the middle Iron Age samples which had been processed for carbonised plant remains contained abundant mollusc shells. However, the flotation method used to extract the carbonised remains does not give sufficiently reliable recovery of shells for fully quantitative investigations and not all the shells were contemporaneous with the archaeological deposits. Some of the flots contained fragments of tufa and shells of aquatic molluscs, particularly Valvata piscinalis, encrusted in tufa. They had apparently been derived from the Pleistocene gravels. Many of the samples also contained shells of Candidula gigaxii and Cernuella virgata, both of which are regarded as medieval introductions (Evans 1972, 179). Post-depositional contamination is confirmed by the presence in most of the flots of modern seeds, especially Atriplex spp., Fallopia convolvulus and Polygonum spp.
These results serve as a warning that dry fills of archaeological features on the Thames gravels ought not be considered as entirely sealed contexts. Intrusive items up to at least 5 mm in diameter can be present. They probably entered them by such means as falling down cracks and earthworm burrows. In the case of the middle Iron Age settlement at Claydon Pike contamination does not seem to have presented a major problem with the carbonised plant remains because there has not been any post-Iron Age occupation on that part of the site to result in the presence of potentially intrusive carbonised remains.
Pinus pinea L. (stone pine). An intact cone of P. pinea was found in Sample 2906/4 from a phase 3 Roman well. It is a tree of Mediterranean origin but cones have been found on other sites in Roman Britain (Kislev 1988, 76). Its large, globular cones cannot be mistaken for those of any other European, Western Asiatic or North African pine.
Brassica species. Seeds of Brassica rapa L. ssp. sylvestris (L.) Jan. (wild turnip) were found in waterlogged contexts belonging to phases 2, 3 and 4. The seeds were black, about 1.6 mm in diameter and had a coarse reticulum which included oblong cells. While the leaves of this species would be edible, it is a common native annual weed of disturbed ground which grows around settlements. Seeds of Brassica nigra (L.) Koch (black mustard) were found in Sample 502/7 from a late Roman (late 3rd/4th century AD) well. The seeds were dark brown, about 1.5 mm in diameter with a reticulum of irregular high ridges forming a network of more or less regular polygonal cells. The high ridges of the reticulum presented a rather ragged appearance. B. nigra was cultivated in Roman times for its edible seeds which are one of the sources of culinary mustard but it is also a native annual of river banks which is common in waste places too. A third species of Brassica was represented by a single seed in Sample 502/7. The seed was black, larger than the other two species, being about 2.0 mm in diameter and had a fine reticulum of regular polygonal cells. It resembled seeds of the Brassica cultivars B. oleracea (cabbage), B. napus L. (rape and swede) and B. rapa L. ssp. rapa L. (turnip). B. oleracea includes the wild cabbage, a maritime cliff plant which is possibly native to Britain, as well as the cultivated cabbage (Clapham et al. 1987, 71). However, it is thought implausible that the maritime wild cabbage was growing at Claydon Pike. It is therefore argued that the third species of Brassica was a cultivar, even though it is not possible to determine to which of the cultivated species it belonged. A similar conclusion was reached for waterlogged Brassica seeds from the Bancroft Roman Villa, where the two native species which occur inland, B. rapa ssp. campestris and B. nigra, were present alongside seeds of a third species which resembled those of the cultivars (Pearson and Robinson 1994, 574-8). Brassicas, including cabbage, are known to have been cultivated in Roman Italy (Apicius III, ix, xiii).
Buxus sempervirens L. (box). B. sempervirens leaves were identified from Sample 1318/C/3, from a mid to late Roman water hole in trench 17, Sample 1989/B, from a late Roman pit and Sample 2721/6, from a late Roman drainage sump. Sample 1989/B also contained seeds, flower buds, fruit fragments and twigs of B. sempervirens, providing strong evidence that it was grown on the site. A leaf and fruit fragment of B. sempervirens are illustrated from another Roman settlement in the Thames Valley at Farmoor (Lambrick and Robinson 1979, 101).
Pyrus pyraster Burg. (pear). Three seeds of P. pyraster were identified from Sample 2160/B/5, a phase 3 pit in trench 13. They were distinguished from Malus sp. (apple) on the basis of a layer of small circular cells overlying the coarse elongate cells which in Malus sp. comprise the surface layer. The seeds also had the more elongate shape which characterises P. pyraster.
Scandix pecten-veneris L. (shepherd's needle). Seeds of S. pecten-veneris were found in Sample 766/6 from a mid 2nd to late 3rd century Roman well and Sample 502/7 from a late Roman (late 3rd/4th century AD) well. The elongate seeds (mericarps) of this plant are very distinctive. However, difficulties with identification had been experienced because the plant growing under that name in several British botanic gardens from which reference material had been obtained proved to be the related Scandix australis L. The unidentified umbellifer mericarp illustrated from a late Roman well at Barton Court Farm Villa (Miles 1986, 34) is S. pecten-veneris. This plant is a weed of cultivation possibly introduced from Southern Europe during the Roman period.
Baldellia ranunculoides (L.) Parl. (lesser water plantain). Seeds of B. ranunculoides were identified from two phase 3/4 contexts, Sample 1318/C/3 from a water hole in trench 17 and Sample 2495/A/2 from a ditch in the far western trench 31. They were identified from the fine white ribs which form a cage around the dark achene. B. ranunculoides has a localised distribution in Britain. This waterside plant has not been recorded from any other archaeological sites in the Upper Thames Valley.
Pterostichus madidus (F.). Four individuals of this beetle were found in Sample 502/7 from a late Roman (late 3rd/4th century AD) well and it was abundant in Samples 696/2 and 696/3, the two samples from the medieval well. Although now a very common species, there has been some doubt expressed about its status in Roman Britain (Kenward 1982, 72). The Roman examples from Claydon Pike were from a stratigraphically secure deposit at the bottom of a well. The remains identified included pronota, with their distinctive rounded hind angle, and elytra, with their rounded shoulders.
Aphanus rolandri(L.). A pronotum belonging to a lygaeid bug from Sample 2906/4, the fill of a phase 3 well in trench 29 and six further examples from Sample 502/7 from a late Roman well were tentatively attributed to A. rolandri. This bug is on the edge of its range in England and is now only known from some south-west and south-east coastal counties (Southwood and Leston 1959, 100-1).
Apis mellifera L. (honey bee). Heads of workers of A. mellifera were found in Samples 502 and 1989/B from late Roman deposits and Sample 696/3 from a medieval well. They were identified using the same characters shown by an Iron Age example from Mingies Ditch Oxon (Allen and Robinson 1993, 118).
Melophagus ovinus (L.) (sheep ked). A puparium of M. ovinus was identified from Sample 1202/15 from a phase 3 water hole in trench 17. The adult, which is an ectoparasite of sheep, is viviparous and extrudes a fully grown larva which immediately pupates on the wool of the host (Edwards et al. 1939, 123-4). The puparium is truncate, oval in cross-section and has rows of shallow depressions along either side.
The vast Iron Age and Roman crop-mark complex at Claydon Pike is situated upon the first gravel terrace of the Thames near the confluence of the Rivers Thames and Colne. The site itself is drained by the Colne. The gravel is predominately oolitic limestone pebbles with some sand. The first terrace dates from some time in the Devensian, and, at its full height, stands about 2 metres above modern ordinary river level. At Claydon Pike, however, an ancient system of channels had dissected the terrace into a series of low islands before the end of the last glaciation.
An organic Chara marl, (context 965), had accumulated in a deeper part of one of these channels within Warrens field. Not surprisingly, the assemblage of plant remains from layer 965 was dominated by Chara oospores and calcified thallus fragments but otherwise it contained a sparse flora (Tables formerly 5 and 6, not 4 and 54 and 5) indicating cold, open conditions. The only shrub species present was Betula nana, represented by its narrowly winged fruits, leaves and distinctive female cone scales. It is a low-growing arctic/highland-moor shrub. B. nana was abundant throughout England in the Late Glacial, the period to which the deposit can be assumed to belong. The other plants were all aquatic and waterside species which probably fringed the channel. The Mollusca from layer 965 were mostly aquatic species, Planorbidae and Lymnaea sp., but neither plants nor molluscs suggest flowing water in the channel.
A few clay-filled involutions, with lenses of fine gravel and sand, occurred on the gravel islands. Most of them had suffered leaching of calcium carbonate, and molluscs were absent, but in layer 961 within trench 14 of Warrens field, the buffering effect of limestone gravel resulted in the survival of shells. The Mollusca, listed in Table formerly 4, not 66, suggested cold, sub-arctic conditions. Pupilla muscorum was the most abundant species and the shells were from very large specimens, which is typical of late Devensian contexts (Evans 1972, 150). Their abundance indicates open, terrestrial conditions. Succinea or Oxyloma sp. was also well represented. In contrast, they are species of marshes and emergent water plants.
Samples 961 and 965 probably provide a good picture of the environment at Claydon Pike in the Late Glacial (Late Devensian). On the dry gravel islands were small marshy hollows. Between the islands, the former channels had become swampy areas with a series of still-water pools in their deeper parts. The landscape was open with some low shrub of Betula nana. There were probably large expanses of gravel with only a sparse herb cover, freezing and thawing making the gravel surface unstable. The many small bodies of water were fringed by sedges but their emergent vegetation was not dense enough to shade out the light-demanding submerged Chara sp.
Above the Chara marl of layer 965 there had formed a layer of peat, (layer 464), in which many seeds of Menyanthes trifoliata were evident. Fruits of a tree species of Betula had replaced remains of B. nana. Bud scales from a tree or shrub species of Salix were also present. The vegetation of the pool and its margin seems to have become richer, but despite the development of some tree cover, the herb seeds present show that the landscape remained partly open. The flora suggests some amelioration of climatic conditions and would be more characteristic of the Windermere Interstadial (Late Devensian Zone II) or the very early Post-Glacial (Flandrian Early Zone I) rather than the Loch Lomond Stadial (Late Devensian Zone III).
The soil which developed on top of the gravel islands during the Early Post-Glacial had not survived anywhere in an undisturbed state. Recent ploughing extended down to the surface of the terrace and gravel had been incorporated into the profile. However, on the basis of the fills of the middle Iron Age features in Warrens field, the soil seems to have been a clay loam. The depth of soil over the islands was probably similar to the modern depth of 0.25 - 0.30 m because little ploughsoil had been eroded into the hollows between the islands. The soil was probably a stone-free analogue of the Badsey series, shallow calcareous stony loams to clay loams which are at present extensive on the first gravel terrace of the Thames in the Lechlade area (Jarvis 1973, 114-15, 179-180). The underlying gravel means that these soils are normally well-drained when the water table is low and sometimes experience problems of drought.
The floodplain hollows between the islands contained a clay to clay loam beneath medieval alluvium which would have been ill-drained where it overlay the peaty channel fills but rather better drained under conditions of low water table where it overlay gravel.
Woodland succession followed by clearance presumably occurred at Claydon Pike prior to settlement of the site during the middle Iron Age but deposits containing biological remains from these periods were absent.
Although the molluscan assemblages from Samples 153/CC/5 (island 2) and 371/D/3 (island 3) suggest that many of the deeper middle Iron Age (Phase 1) ditches held water, waterlogged organic material did not survive in them. The environmental evidence for this period is therefore limited.
Slum aquatic molluscs, particularly Aplexa hypnorum, Lymnaea truncatula and L. peregra, predominated in the above two samples. There was also an open-country faunal element including such dry-ground species as Pupilla muscorum, which probably reflects conditions on the islands during the middle Iron Age.
It is difficult to use the molluscan evidence to ascertain whether the islands were experiencing flooding during the middle Iron Age. Many of the flots from the samples taken for carbonised plant remains contained flowing water aquatic molluscs such as Valvata piscinalis and Bithynia tentaculata. However, they were normally encrusted with tufa and fragments of tufa were also present in the samples. These shells had almost certainly been derived from bands containing tufa fragments within the gravels themselves, which in turn had probably been reworked from even earlier sediments. The only flowing water mollusc from Samples 153/CC/5 and 371/D/3 was a single specimen of Valvata cristata which could have been derived from the gravel. The flot from Sample 416 (island 3), however, contained many specimens of Carychium sp., Lymnaea truncatula, Anisus leucostoma, Vallonia pulchella and Trichia hispida gp., along with rather fewer specimens of Valvata cristata, V. piscinalis and Bithynia tentaculata. These shells were not encrusted with tufa and were in better condition than the encrusted shells. This assemblage of terrestrial, amphibious and flowing water molluscs was very similar to the molluscan assemblages from the late Saxon / early medieval alluvium at Claydon Pike. Sample 416 was taken from the earlier phase of middle Iron Age linear boundary 472, located at the edge of gravel island 3. It indicates the limit of Iron Age flooding. However, it is possible that this sample had been contaminated with the later alluvium. Alluvial sediments were not noted in any of the Iron Age features.
Four water-logged samples were investigated from the bottoms of early 1st to early 2nd century AD enclosure ditches (E 16/17) in Longdoles Field (Samples 577/A, 577/N/4, 1528/A/3, 1704/C15), one sample from a ditch just to the south of these enclosures (2490), one sample from an isolated enclosure (Sample 6) and one sample from a water-hole or sump on the floodplain in Warrens field (Sample 962/A), which may have been early Roman in date. The features all seem to have held stagnant water which supported various water plants such as Ranunculus S. Batrachium sp., Nasturtium officinale, Apium nodiflorum and Lemna sp., small water beetles particularly Helophorus cf. brevipalpis and Ochthebius minimus, and slum aquatic molluscs. The waterlogged plant and invertebrate remains which had their origins beyond the limits of these features mostly seem to have entered the deposits via various natural agencies although Sample 1528/A/3 also contained imported plant material. The enclosures were sufficiently small that there was a strong element within the assemblages reflecting conditions beyond the immediate environs of the enclosures as well as that reflecting more local conditions.
The evidence from the pollen and Coleoptera suggests that the landscape of the first gravel terrace at Claydon Pike during the 1st/early 2nd century AD was predominantly grassland. As well as showing high values for Gramineae, the pollen analyses also gave high values for Cyperaceae, suggesting wet areas with abundant sedges. Seeds of the Juncus effusus group (the tussock rushes J. inflexus, J. effusus, and J. conglomeratus) were exceptionally abundant, suggesting that Holcus lanatus - Juncus effusus rush pasture (MG 10 of the National Vegetational Classification of Mesotrophic Grassland, (Rodwell 1992)) grew on the wetter areas of the site. Other possible members of this wet grassland community that were well represented by their seeds included:
Ranunculus cf. repens |
Juncus articulatus sp. |
||
Potentilla anserina |
Carex sp. |
||
Mentha cf. aquatica |
Eleocharis palustris |
||
Prunella vulgaris |
There is evidence from other sites for a rising water table on the floodplain of the Upper Thames Valley during the Iron Age with some sites experiencing flooding (Robinson and Lambrick 1984). This rise in the water table would probably have created conditions on the area of floodplain at Claydon Pike which, if combined with heavy grazing, would have favoured tussock rush pasture.
There was much less evidence for drier grassland as might have grown on the top of the gravel islands themselves, although it is possible that Potentilla cf. reptans, Prunella vulgaris and Leontodon sp. were members of such a community. However, the presence of the 1st century AD enclosures from which most of the samples came in Longdoles Field would have restricted the development of a drier grassland flora. Sample 6 contained a single seed of Sanguisorba minor and pollen of Scabiosa columbaria was identified from Sample 1704/C/5. It is possible that these plants represent dry calcareous grassland growing on the unoccupied gravel islands.
The Coleoptera included a significant element which feed on roots in grassland, the chafers and elaterids of Species Group 11 comprising 7.5% of the terrestrial Coleoptera. Some of them would have been able to live in ill-drained pasture but others, such as Agrypnus murinus, require well-aerated soil. The Coleoptera from a deposit tend to have been derived from a larger catchment than the seeds provided that they have entered as a result of the usual natural agencies, and in this case are probably providing more information on the other gravel islands than the seeds. The abundance of scarabaeoid dung beetles which feed on the dung of large herbivores on pastureland (Species Group 2), at 19.5% of the terrestrial Coleoptera, indicates a concentration of domestic animals on the site. The most numerous of these beetles was Aphodius granarius, which is one of the more common species in cattle and horse dung on the Thames floodplain at present. Vetch and clover-feeding weevils from the genera Apion and Sitona (Species Group 3) which tend to be numerous in hay meadows were no more abundant than might be expected from extensive pastureland.
The pasture in the vicinity of the enclosures seems to have experienced overgrazing and damage from trampling. In places the sward seems to have been churned into mud enriched with dung. Seeds of Juncus bufonius gp., annual rushes of mud and wet, broken ground were very numerous. The following members of the Alliance Chenopodion fluviatile of the Bidentetea Tripartitae were particularly well represented by their seeds: Chenopodium ficifolium, C. cf. rubrum and Polygonum lapathifolium. It is an annual terrestrial community of drying mud and, less often, old compost heaps or piles of manure (Silverside 1977, 231-235). Its members are either nitrophiles or can tolerate high nitrogen levels. Another likely member of this mud community, though represented by rather fewer seeds, is Rumex maritimus. These plants, with the exception of J. bufonius, are not particularly resistant to trampling or grazing. This would suggest that stock did not have unrestricted access to the entire site throughout the summer.
There were some differences between the seed assemblages. Sample 962/A contained few seeds of the annuals mentioned above but many seeds of some of the plants of wet pasture such as Potentilla anserina and Eleocharis palustris. This water-hole on the floodplain seems to have been away from the main areas of disturbance. The isolated enclosure (Sample 6) seems to have experienced wetter conditions than the enclosures within trench 13. Seeds of Juncus bufonius gp. and Chenopodium rubrum, which both favour wet mud, were very abundant whereas Chenopodium ficifolium, a nitrophile which favours somewhat drier conditions than C. rubrum, was absent from this sample alone. Sample 6 also contained more seeds from the Juncus effusus gp. than any of the other samples.
There was no firm evidence from either the botanical or entomological remains for the presence of arable fields in the vicinity of the site, although cultivation would probably have been possible. A few waterlogged cereal remains were present but they could have been imported.
Annual weeds of drier habitats than those occupied by communities of the Bidentetea Tripartitae were also represented by their seeds. Most of them, such as Brassica rapa ssp. sylvestris and Urtica urens, occur in various communities of the Order Polygono-Chenopodietalia of the Stellarietea Mediae (Silverside 1977, 236-245). Two of the weeds mentioned as possible members of the Chenopodion fluviatile, Chenopodium ficifolium and Polygonum lapathifolium, are also members of the Fumario-Euphorbion, one of the more basophilous alliances of the Chenopodietalia which occurs on heavy soils. The various alliances of the Chenopodietalia occur on nitrogen-enriched disturbed ground around settlements, in gardens and amongst spring-sown cereals. Suitable habitats probably occurred within some of the enclosures on top of the gravel islands.
Seeds of Urtica dioica were abundant in most of the samples and nettle-feeding beetles, including Apion urticarium, were also present. It is likely that nettles became established on some of the disturbed areas, particularly where the soil had a high nitrogen and phosphorus content from the incorporation of dung, once the frequency of disturbance had been reduced.
There were also a few seeds from the following weeds: Anthriscus caucalis, Hyoscyamus niger and Onopordum acanthium. These plants are no longer common in the region but have frequently been recorded from Iron Age sites in the Upper Thames Valley (Robinson 1981, 274-275). The samples, however, did not contain seeds from any of the weeds which seem to have arrived in the region at about the time of the Roman conquest, such as Agrostemma githago, Conium maculatum and Anthemis cotula. Seeds of all three species were identified from phase 3 contexts at the site.
The Coleoptera included many species of Carabidae and Staphylinidae which occur in a variety of open habitats. They ranged from very hygrophilous species such as Chlaenius nigricornis or nitidulus, which occurs on wet mud, through to Brachinus crepitans, the bombardier beetle, which occurs in sunny places on dry, short-turfed, grassland.
Tree pollen, particularly of Quercus sp., comprised 9% of the total identified pollen from the phase 2 samples. Wood and tree-dependent Coleoptera (Species Group 4) however, were entirely absent from these samples. Since pollen tends to have a wider catchment area than Coleoptera, this would suggest that any woodland was distant from the site, indeed it was probably growing beyond the river gravels.
There was only a slight presence of shrub pollen and macroscopic remains of woody taxa from the enclosure ditch samples were few. There seems to have been little if any scrub in the vicinity of the enclosures and their boundaries had not been reinforced with hedges. Sample 962 contained some cut thorny twigs of Prunus/Crataegus type. However, there was little other evidence of the proximity of scrub or hedges and it seems more likely that they had been brought to the site.
Many of the beetles from the samples are members of various decomposer communities. Some of the dung beetles have already been considered. Another group of beetles, certain Sphaeridiinae and Oxytelinae (Species Group 7), as well as occurring in dung on pasture also occur in manure heaps and various other sorts of foul vegetable material. They were however, no more abundant than might be expected given the concentration of grazing animals around the site. Similarly, the Lathridiidae (Species Group 8), mould feeders which can be particularly abundant in settlements, living on old hay, thatch etc., were no more abundant than might be expected in a grassland landscape. There was, however, evidence for plant material being brought to the site. Sample 2490 contained frond fragments of Pteridium aquilinum and spores of Pteridium were well represented in the pollen sample from this context. P. aquilinum is a plant of well-drained acid soils and did not therefore come from the immediate vicinity of the site
There were very few possible indoor beetles from the phase 2 samples. There were only a couple of woodworm beetles (Species 10 Group), insufficient to indicate the presence of buildings, and the synanthropic groups of beetles (Species Groups 9a and 9b) were not represented.
The 1st/early 2nd century AD enclosures at Claydon Pike had much in common with Iron Age settlements on the Thames floodplain and will be compared with them. The enclosures were used for the herding of domestic animals that were grazed on the floodplain and islands of gravel terrace. The increasing wetness of the floodplain was probably exacerbated by the trampling of the stock impeding the drainage of the soil even on the areas of gravel terrace. Similar conditions, of ill-drained pastureland with rush tussocks and disturbed areas with nutrient-rich mud supporting Chenopodium rubrum, Rumex maritimus etc. also existed around Iron Age enclosures elsewhere on the Thames floodplain at Port Meadow and Farmoor Enclosure 3 (Lambrick and Robinson 1988, 65-71) as well as nearby at Thornhill Farm (Jennings et al 2004). The evidence from the dung beetles does not indicate which species of domestic animal were grazed at Claydon Pike. However, wet conditions on the site and the presence of the snail Lymnaea truncatula, which is the intermediate host of the sheep liver fluke, would suggest that it is more likely that cattle or horses were the main stock rather than sheep. This is corroborated by the generally high proportion of cattle remains within the faunal assemblage (see Sykes, section 3.2).
The Coleoptera provide little evidence of the presence of human settlement or buildings on the site. However, the ‘intensity’ of human occupation on other Iron Age sites that have been investigated in the Upper Thames Valley was insufficient to be reflected by the beetle evidence. There was no evidence that the enclosures experienced flooding during the late Iron Age or early Roman period. The islands would therefore have been suitable for permanent habitation. In this way Claydon Pike is more similar to Port Meadow, where the Iron Age settlements were also on top of gravel islands on the floodplain, rather than Farmoor, where the settlements were on the floodplain and experienced flooding. However, the frequent re-alignments shown by the enclosure ditches and the absence of any more permanent boundaries in the form of hedges suggests that the life of each phase of an enclosure was short.
A very small quantity of waterlogged spelt wheat chaff was identified from the 1st century AD deposits, but the carbonised plant remains provide better evidence for the use of cereals on the site. It is possible that they had been imported from elsewhere. Bracken was brought to the site, perhaps for use as bedding. While it is by no means certain what the bracken was used for, the importation of bracken seems to have been a normal activity on Iron Age sites in the Upper Thames Valley (e.g. Robinson 1981, 261). Unlike the subsequent Roman phases of the site, there was no evidence for horticultural crops from these samples. This too seems usual for Iron Age sites in the region. However, various wild plants were present which could have been used as green vegetables, for example water cress (Nasturtium officinale).
Seven waterlogged samples were investigated from water-holes or wells assigned to the phase 3 settlement in Longdoles Field, and two samples from ditches beyond the main settlement (Table 2):
The water-holes, pits and wells all seem to have held water which supported low populations of small water-beetles such as Helophorus sp., Ochthebius minimus and Hydraena testacea. Three of these features seem from the seed evidence to have had aquatic plants growing in them. Feature 2160 (Trench 13) seems to have had Callitriche sp. growing in it. The aquatic flora of Feature 2867 (Trench 29) included Ranunculus S. Batrachium sp., Callitriche sp. and Veronica S. Beccabunga sp., while Lemna sp. covered the surface of the water in Feature 2906 (Trench 29). Remains of the weevil Tanysphyrus lemnae, which only feeds on Lemna spp., confirms the presence of that plant in Feature 2906 and adds it to the species from Feature 2160.
The interpretation of the waterlogged plant and invertebrate remains from the phase 3 water-holes, pits and wells is a more complex problem than for the phase 2 deposits. Human transport had probably been a factor in the introduction of waterlogged biological remains in all these features. Three of the samples, Samples 766/2, 1202/15 and 1318/C/3, had a major component of hay and, in the case of Sample 766/2, straw as well. With this imported plant material probably came some of the insects. Interpretation is also difficult because the area of the settlement was so large that some of the remains particularly the seeds, which tend to have a small radius of origin when they have not experienced human transport, are giving little information on environmental conditions beyond the limits of the settlement. It was possible to see the environment of the 1st century AD enclosures as a part of the more general landscape without these problems.
The two ditch samples from west of the main settlement (1487, 2495) contained a rather higher proportion of remains from plants and invertebrates which lived in them than did the waterlogged deposits inside the settlement. They will be considered below. All the plant and invertebrate remains seem to have entered these deposits through natural agencies.
Sample 2495/A/2 was from a shallow field-ditch where it ran through one of the wetter areas of the floodplain in trench 31 (Fig. 2.2.1). The peaty deposits in the ditch extended beyond the edge of the ditch onto the floodplain. Standing water in the ditch seems, on the basis of the seeds present, to have supported a floating-leaved community of Ranunculus S. Batrachium sp. and Potamogeton cf. coloratus, and an emergent community of:
Ranunculus flammula |
Baldellia ranunculoides |
||
Nasturtium officinale |
Alisma sp. |
||
Apium nodiflorum |
Iris pseudacorus |
||
Mentha aquatica |
Carex spp. |
The deposit also contained numerous fronds of the fenland moss Calliergon giganteum. This emergent community probably extended onto any areas of floodplain that retained standing water into the early summer. Elsewhere it probably graded into a fen meadow community on the peaty areas as described by Wheeler (1980,762-764) the species represented by their seeds including:
Caltha palustris |
Pedicularis palustris |
||
Thalictrum flavum |
Valeriana cf. dioica |
||
Lychnis flos-cuculi |
Carex spp. |
Thalictrum tp. pollen was abundant. Seeds from meadowland species as might have grown on the somewhat better-drained mineral soils were also present:
Ranunculus cf. acris |
Leucanthemum vulgare |
||
Rhinanthus sp. |
Centaurea cf. nigra |
The pollen evidence from Sample 2495/A/2 would be consistent with the presence of extensive Alopecurus pratensis - Sanguisorba officinalis (MG4) or Centaurea nigra - Cynosurus cristatus (MG5) haymeadow (Rodwell 1992). Rhinanthus tp. pollen was present and the value for Plantago lanceolata pollen, at 13%, was high. A single grain of Scabiosa columbaria pollen perhaps reflects dry calcareous grassland growing on one of the unoccupied gravel islands.
The evidence from the insect remains supports the evidence from the plant remains for the conditions in and around Ditch 2495. Trichopteran larvae lived in the ditch and the water beetles included amphibious species, such as Coelostoma orbiculare and Dryops sp., which also occur amongst wet plant debris on the ground. The phytophagous Coleoptera included:
Plateumaris sericea |
on |
Carex spp. |
|
Nasturtium officinale |
on |
Iris pseudacorus |
|
Apium nodiflorum |
on |
Ranunculus spp. |
|
Mentha aquatica |
on |
Plantago lanceolata and P. media |
Clover and vetch-feeding weevils from the genera Apion and Sitona (Species Group 3) comprised 13% of the terrestrial Coleoptera, which is high enough to suggest meadowland conditions, although they also occur at lower levels in other grassland types.
The sample from the phase 2 deposits contained abundant remains of pastureland plants and beetles. These were rare or absent from Sample 2495/A/2. For example, there were only 80 seeds of the tussock rushes of the Juncus effusus gp., whereas values from similar sized samples from the 1st century AD features ranged from 220 to circa 11000. These rushes cannot readily withstand mowing or competition from taller-growing herbs, whereas they are not readily eaten by stock because they are tough. There was only a single seed of Eleocharis palustris and the scarabaeoid dung beetles of Species Group 2, which are particularly good indicators of pasture, were entirely absent.
Unfortunately the assemblage from the roadside ditch, Sample 1487, almost entirely comprised plants and invertebrates that were living in the stagnant water of the ditch itself and it gives little information on the surrounding landscape. Even those waterlogged deposits from the settlement which did not contain substantial quantities of dumped plant material must be interpreted cautiously for information on the surrounding landscape. However, the Coleoptera from Sample 2906/4, a water-hole or well on Platform 3, does seem to include a significant proportion from beyond the boundaries of the settlement. The chafers and elaterid beetles which feed on roots in grassland (Species Group 11) comprised 4% of the terrestrial Coleoptera while the abundance of the clover and vetch-feeding weevils of the genera Sitona and Apion, at 9% again hints at the presence of meadowland. However, 7% of the terrestrial Coleoptera were scarabaeoid dung beetles (Species Group 2), indicating that at least some grazing of domestic animals was taking place, although not at the intensity of the earlier phase. The other sample that seemed to have a significant proportion of beetles naturally derived from beyond the edge of the settlement was Sample 2160/B/5, from a pit on Platform 1, which had a value of 8% for scarabaeoid dung beetles.
It is thought probable that seasonal flooding was occurring on at least the lowest part of the floodplain at Claydon Pike during the 2nd and 3rd centuries AD. Sample 2495/A/2 contained shells of flowing water molluscs, including Bithynia tentaculata, with their periostraca intact (and therefore unlikely to be reworked). Parts of the floodplain were certainly very wet and there is evidence from elsewhere in the Upper Thames Valley for extensive flooding on the Thames floodplain throughout the Roman period (Robinson and Lambrick 1984).
Sample 2495/A/2 contained a few seeds from annual weeds and the sample from Feature 1487, the roadside ditch, contained a few seeds from wayside plants as might be expected along a road. However, there was no evidence for extensive areas of disturbed ground outside the settlement.
It is clear that significant quantities of arable products, particularly cereals but also flax, were brought to the site. It is possible that some were cultivated on the more distant gravel islands, indeed flax benefits from being grown on soils where the water table is close to the surface. However, there is no evidence for Roman cultivation of the gravels in this area.
The level of tree pollen from Sample 2495/A/2, at 14%, was somewhat higher than might have been anticipated. Wood and tree-feeding beetles (Species Group 4) were absent. The average percentage of tree pollen from all the phase 3 samples, however, at 8.7% was a very similar value to that from the 1st/early 2nd century AD contexts. Quercus was again the most abundant tree pollen type, but Fraxinus pollen was also present. There was little shrub pollen from the samples and few macroscopic remains of woody taxa from Samples 1487/A and 2495/A/2. Sample 1487/A did contain a single seed of Viburnum opulus. While seeds of V. opulus were not abundant in any of the Roman samples, they were encountered in rather more of the Roman samples from Claydon Pike than is usual on sites in the Upper Thames Valley. It is possible that V. opulus was quite common along boundaries at Claydon Pike during the Roman period. It is unusually common in hedgerows at Claydon Pike at present.
Some of the samples contained annual weed seeds characteristic of the Centauretalia cyani Order of the Class Stellarietea Mediae. They had probably been introduced to the site along with cereal remains. There were also a few seeds from annual weeds belonging to the Order Polygono-Chenopodietalia of the Stellarietea, such as Brassica rapa ssp. sylvestris and Urtica urens which, along with such more generalist annual weeds as Stellaria media gp. and Atriplex spp., probably grew around the settlement (Silverside 1977, 236-240). However, annual weeds do not seem to have been very abundant growing inside the settlement.
Those areas of the settlement that were not bare ground seem to have supported one of, or an intermediate between, the following three communities of neglected ground. The first of these is Arrhenatherum elatius grassland. It is an ungrazed grassland community, particularly of calcareous soil, which readily becomes established on ungrazed grassland or former disturbed ground. It is characterized by tall umbellifers such as Anthriscus sylvestris, Pastinaca sativa and Heracleum sphondylium. Their seeds were conspicuous in Sample 2906/4. The second is a tall herb community, as grows on the shady side of hedges or on ungrazed land under clusters of trees, which includes the following plants:
Anthriscus sylvestris |
Stachys cf. sylvatica |
||
Urtica dioica |
Alliaria petiolata |
All these seeds occurred in Samples 2160/B/5 and 2906/4. Finally, seeds of Conium maculatum, Urtica dioica and Sambucus nigra were present in most of the samples and abundant in Sample 2839/A/S. These species can form a community on disturbed then neglected nitrogen- and phosphorus-enriched soil, as often occurs around settlements.
The importance of areas of bare ground, with perhaps no more than a scattering of such trample-resistant plants as Coronopus squamatus, Polygonum aviculare spp. and Plantago major, ought not be underestimated. Many of the ground beetles from the samples such as Pterostichus melanarius and Harpalus spp. readily leave the cover provided by vegetation to range over bare ground.
The phase 3 drainage ditches seem to have been very effective. The mud community of the Bidentetea Tripartitae had disappeared almost without trace. Several of the samples contained a few seeds of Myosoton aquaticum, which was perhaps growing around the top of the water-holes on soil moistened by splashes.
Macroscopic remains of woody taxa were reasonably well represented in most of the waterlogged samples from the phase 3 settlement. Wood and tree- dependent beetles comprised 1.4% of the terrestrial Coleoptera from these samples. There seems to have been a significant presence of trees and shrubs in or around the settlement. A few Crataegus/Prunus type thorns were found in three of the samples, but these were insufficient to suggest thorn hedges along the boundaries. The two woody species (excluding Sambucus nigra) best represented by their macroscopic remains were Salix sp., represented by buds and in one case leaves, and Fraxinus excelsior, represented by fruits and bud scales. The Salix leaves belonged to S. viminalis or a S. viminalis hybrid. They were commonly grown as osiers. Remains of Salix sp. and F. excelsior were identified from most of the samples from the settlement. The wood and tree-dependent Coleoptera included the following more host-specific species:
Phyllodecta vitellinae |
on |
Salix and Populus spp. |
|
Chalcoides sp. |
on |
Salix and Populus spp. |
|
Magdalis ruficornis |
on |
Rosaceous trees and shrubs esp. Prunus and Crataegus spp. |
|
Hylesinus crenatus |
mostly on |
Fraxinus excelsior |
|
Leperisinus varius |
mostly on |
Fraxinus excelsior |
It is possible that some of the boundaries on the site were lined with osiers and ash trees. If so, they would almost certainly have been cut for withies and poles. It is also possible that there was a withy bed at the site or there were clumps of trees and bushes on the platforms. Various other shrubs probably grew amongst them in lesser numbers. The following were represented by their seeds:
Rhamnus catharticus |
Corylus avellana |
||
Rosa sp. |
Sambucus nigra |
||
Crataegus cf. monogyna |
Viburnum opulus |
There were also remains from horticultural trees and shrubs which will be discussed later.
Remains of trees and shrubs were particularly abundant from Water-Hole 2160 in trench 13. This deposit also contained seeds of the tall herb community of shady places mentioned above. It seems very likely that this part of trench 13 was overhung with F. excelsior (common ash) trees and various shrubs. However, the open-country pollen flora and insect assemblage from Sample 1318/B/5 suggests that these conditions were very local.
Sample 1318/C/3 in trench 17 mostly comprised sedge hay. However, it also contained a few leaf fragments of Buxus sempervirens. Box leaves have now been found on many Roman sites in England where waterlogged deposits have been investigated including a rural settlement at Farmoor, further down the Thames (Lambrick and Robinson 1979, 121, 127). As well as growing box for ornamental hedges, the Romans seem to have used its shoots for making garlands for ritual purposes.
Sample 2906/4 in trench 29 contained a cone of Pinus pinea, the stone pine. It is naturally a tree of the Mediterranean region and Portugal, but it has been cultivated in Britain for at least the past four centuries although young trees are apt to be cut back by severe frosts (Bean 1914, 188). The distribution map given by Krüssmann (1985, 231) shows P. pinea as being cultivated in England south of a line from the Bristol Channel to the Wash and also on the Cheshire Plain. However, it does not grow in central France, Germany or other parts of Northern Europe. A mature tree in the Royal Botanic Gardens, Kew, fruits well. There are several archaeological records of P. pinea from Roman London and it was identified from Verulamium (Kislev 1988, 76). Charred cone fragments of P. pinea were found in a Roman temple of Mithras on Hadrian's Wall at Carrawburgh, where they had been used as altar fuel (Richmond and Gillam 1951, 6-7,20,86-87). There are not many other finds of P. pinea from rural Roman sites, although single cones were identified from the Bancroft Villa, Bucks (Pearson and Robinson 1994) and the Chew Park Villa, Somerset (Rahtz and Greenfield 1977). It is possible that a stone pine tree grew at Claydon Pike. The top of one of the gravel islands would probably have been well-enough drained and it has been fruited in England 14 years after sowing seed (Lloyd 1970, 423). However, it is also possible that the cone had been imported, indeed a Roman wreck in the Mediterranean was found to contain a cargo of stone pine cones (Kislev 1988, 76).
As quite commonly occurs on Roman sites, the waterlogged deposits contained remains of various horticultural crops. They included orchard fruit, herbs, spices and possible vegetables. None of the remains was abundant, but all could have been grown at the settlement. Two of the fruits, Prunus domestica cf. ssp. insititia (bullace or damson) and Prunus avium (cherry) have been found on other Roman sites in the region, for example Farmoor (Lambrick and Robinson 1979, 120-121). On present evidence, it seems possible that both these fruits were Roman introductions to the British Isles. All the pre-Roman Flandrian records cited by Godwin (1975, 197) for these species, however, are in some way suspect (Moffett et al., 1989, 246). The third tree fruit, Pyrus pyraster (pear), has rarely been recorded from Roman sites in Britain. At this date, many varieties of pear were known in Italy (Roach 1985, 117-118).
The herbs, spices and vegetables comprised Coriandrum sativum (coriander), Apium graveolens (celery) and Anethum graveolens (dill), with Papaver somniferum (opium poppy), an unidentified species of Brassica and Daucus carota (carrot) as possible cultivars which could also have been growing as weeds on the site. All have been identified from other sites on the gravels in the Upper Thames Valley and they seem to be part of a Romanisation of the diet that spread throughout the populace of the region (Robinson and Wilson 1987, 54).
There was a rather different balance of decomposer communities of beetles on the site in phase 3 compared with phase 2. Species Group 7, which comprises beetles of various sorts of foul vegetable material including manure heaps, was somewhat less abundant than previously. However, the Lathridiidae (Species Group 8) at 7.7% of the terrestrial Coleoptera from the phase 3 samples, were more than three times as abundant. The increase in one of the members of the group which is particularly favoured by such habitats as old hay and mouldy thatch, Lathridius minutus gp., was even more dramatic. Similarly, there was a considerable rise in woodworm beetles (Species Group 10) and the appearance of various other synanthropic beetles (Species Group 9a). This evidence for the presence of timber buildings is very much supported by the archaeological evidence for trench 13.
Species Gp 9a / 9b
The synanthropic beetles comprised the following species:
Tipnus unicolor |
Typhaea stercorea |
||
Ptinus fur |
Tenebrio molitor |
||
Mycetaea hirta |
All of these beetles have somewhat esoteric habitats in Britain away from human influence, such as birds' nests, but nowadays are much more common in neglected houses and farm buildings. Most flourish in granary waste but none is a true grain beetle. Serious pests of stored grain (Species Group 9b), such as Sitophilus granarius were absent. These beetles might be expected living together in, for example a byre or stable with hay or straw litter on the floor. One of the species, Typhaea stercorea, often occurs in haystack waste.
There was plenty of evidence from the waterlogged botanical remains for the importation of plant material, particularly hay. Sample 1318/C/3 mostly comprised sedge hay. The seeds from the deposit included almost all of the taxa noted as occurring in Sample 2495/A/2, the field ditch some distance from the settlement. They ranged from the amphibious taxa through the fen meadow community through to the reasonably well-drained hay meadow community:
Ranunculus acris |
Valeriana cf. dioica |
||
R. flammula |
Leucanthemum vulgare |
||
Thalictrum flavum |
Centaurea cf. nigra |
||
Nasturtium officinale |
Baldellia ranunculoides |
||
Pedicularis palustris |
Alisma sp. |
||
Rhinanthus sp. |
Carex sp. |
Of particular significance is the occurrence of Baldellia ranunculoides in both samples. This plant is now very rare or possibly absent from the region and has not previously been found in archaeological deposits in the Upper Thames Valley. Its presence, together with the overlap of species, strongly suggests that some of the hay in Sample 1318/C/3 had been cut from the same or a very similar catchment to that for the plant remains in Ditch 2495.
Other members of the hay meadow communities of circumneutral soils (MG4 and MG5) represented by their seeds were:
Linum catharticum |
Plantago cf. lanceolata |
||
Medicago lupulina |
Leontodon sp. |
||
Prunella vulgaris |
The sample also contained a high percentage of Centaurea nigra tp. pollen, numerous flowers of Trifolium sp. and some fragments of Vicia or Lathyrus pods.
Not all the plant remains from Sample 1318/C/3 were obvious hay meadow plants. A few, such as some box leaves, clearly had other origins. In addition, however, there were very many seeds of Juncus effusus gp. and Eleocharis palustris. When the sample was being disaggregated, some of the J. effusus gp. seeds were in clusters, suggesting that they had entered the deposit as capsules on cut stems rather than as stray seeds. Neither J. effusus gp. nor E. palustris thrive under hay meadow conditions. J. effusus gp. tussocks cannot withstand repeated cutting and E.palustris cannot readily compete against other tall-growing herbs in meadowland (Walters 1949). However, pastureland can occasionally be left ungrazed from late winter onwards and a hay crop taken in the summer without any major long term changes in the flora provided it is usually grazed in the spring. It seems, therefore that some of the hay had been cut from what was usually pastureland, with a flora similar to the rush pasture which grew on the floodplain during phase 2.
The beetle assemblage from Sample 1318/C/3 was restricted, comprising Cryptophagidae and Lathridius minutus gp., which probably lived in the material when it was damp hay and a fauna of foul compost, with Omalium rivulare, Gyrohypnus fracticornis and Oxyomus sylvestris which probably flourished in Feature 1318 before all the organic material had become compacted below the water table.
The two other samples with a significant component of hay were Samples 766/2 and 1202/15. Both contained Vicia and or Lathyrus pods and seeds of other familiar hay meadow plants.
Rhinanthus sp. |
Centaurea nigra |
||
Leucanthemum vulgare |
Sample 766/6 contained seeds of Scabiosa columbaria and Picris hieracioides. There was also a high value, 11%, for Centaurea scabiosa pollen from this sample. All three are plants of dry calcareous grassland. Perhaps they were growing in hay meadow on high areas of gravel terrace some distance from the site, but it also possible that hay was imported from limestone grassland growing in the foothills of the Cotswolds.
Samples 766/6 and 1202/15, unlike Sample 1318/C/3, also included cereal remains. Both contained waterlogged spikelets of Triticum spelta and there were a few rachis fragments of Hordeum sp. and capsule fragments of Linum usitatissimum in Sample 1202/15. Sample 766/6 contained seeds from various possible arable weeds and also two weeds strongly tied to arable agriculture: Agrostemma githago and Scandix pecten-veneris. A major component of Sample 766/6 in addition to hay seems to have been cereal straw, although it was difficult to extract recognisable culm nodes.
There was only a small assemblage of beetles from Sample 766/6 but it was very distinctive. Anobium punctatum, the woodworm beetle, was the most abundant. Other beetles present included Ptinus fur and Tenebrio molitor. It clearly comprises an indoor fauna as might have entered hay and straw stored in a barn. There is no evidence from the insect remains of any addition of foul organic material as would have happened if the hay and straw had been used as litter for animals.
It is difficult to arrive at a conclusion as to the main purpose of the phase 3 settlement at Claydon Pike. One of the major activities at the site would seem to have been the management of hay meadows on the floodplain and gravel terrace and the collection of the hay at the settlement. There is possible evidence for Roman hay meadows elsewhere in the Upper Thames Valley but not as firm as that from Claydon Pike. The insect evidence from Appleford, Oxon, hinted at the presence of meadowland although this was not supported by the botanical evidence (Robinson 1984, 37-39). It is now realised that Sample 1060/2 from an early Roman well at Farmoor, Oxon included a hay meadow assemblage. It contained seeds of Rhinanthus sp., Leucanthemum vulgare and Centaurea cf. nigra and Vicia or Lathyrus pod fragments. There was a large beetle assemblage as summarized below:
Species Group 2 |
Pasture / Dung |
1.5% |
|
Species Group 3 |
?Meadowland |
14.2% |
|
Species Group 8 |
Lathridiidae |
30.3% |
|
----- |
|||
Total number of Terrestrial Coleoptera |
393 |
The high value for Species Group 3 and the low value for Species Group 2 might suggest that meadowland surrounded the well at Farmoor. However, the very high value for Species Group 8 suggests that it was old hay that entered the deposit and it is possible that the meadowland weevils of Species Group 3 had been transported to the site in hay. Claydon Pike thus provides the first good evidence for the presence of meadowland on the floodplain and first terrace of the Upper Thames during the Roman period.
The entomological evidence is consistent with the storage of some hay at the site. However, those groups of insects which might be expected to flourish in old haystack bottoms etc., such as the Lathridiidae, although very much more abundant than during phase 2 at Claydon Pike were not as abundant as at the Roman sites of Farmoor and Barton Court on the Thames gravels (Robinson 1981, 280-281).
Neither the archaeological nor entomological evidence suggests large scale storage of grain at the site. However, cereals were certainly brought to the site, and the carbonised plant remains provide plenty of evidence for activities concerning cereals. It is possible that threshing waste was mixed with the hay for fodder and it is even possible that it was brought to the site for this purpose.
While it is by no means certain that the management of hay meadow and the collection of fodder was the main activity at the site, there is little environmental evidence for much else. There does not seem to have been a great concentration of domestic animals at the settlement, although the evidence from the dung beetles suggests that they were by no means absent. There was evidence that some of the hay had been cut from what was normally pastureland and the aftermath of the hay meadows was probably grazed. If hay meadows are not grazed following mowing, various tall coarse umbellifers such as Heracleum sphondylium become established and they were not evident either from the samples containing cut hay or the sample from the field ditch in the hay meadow. Either sheep or wool had been brought to the site because a puparium of Melophagus ovinus, the sheep ked, was found in Sample 1202/15. This wingless fly is a bloodsucking ectoparasite of sheep which does not survive for more than a few days if removed from its host (Edwards et al. 1939, 123-124). However, there was no evidence for any large scale activity involving sheep.
The ‘intensity’ of human occupation of the site as reflected by the beetle assemblages was greater than in phase 2 and there was evidence for timber buildings. There was no evidence for flooding of the the main excavation area during phase 3. Indeed the settlement would have been well situated as a dry place for rick-yard to which hay was brought from the floodplain. In botanical terms the site appears fully ‘Romanized’. The familiar range of horticultural crops was present, perhaps grown on the site, and box hedges seem to have been present. The weed seeds included some species regarded as Roman introductions.
There was little evidence from the waterlogged remains that the main excavation areas were put to different uses during phase 3, although deposits of hay were only found from trenches 13 and 17.
Five waterlogged samples were successfully investigated from features within the late Roman (Phase 4, early to late 4th century AD) villa enclosure in Longdoles field. A sixth sample was investigated from Well 697, but the preservation of waterlogged remains in it was unsatisfactory so work was abandoned. A column of molluscan samples was examined through the outer enclosure ditch.
These waterlogged features all seem to have held standing water with the usual fauna of small water-beetles, especially Helophorus and Ochthebius spp. Aquatic plants do not seem to have lived in the waters of the features, although the seeds from Sample 700/S/3 suggest that Filipendula ulmaria, Lythrum salicaria, Sium latifolium and Urtica dioica grew along the sides of the enclosure ditch.
The majority of the remains in the waterlogged deposits seem to have entered them through various natural agencies although a small quantity of agricultural debris was present in some of them. There were also some riverine beetles in Sample 1989/B which seemed to have been brought to the site (see below).
Most of the macroscopic plant remains probably had their origins within the villa enclosure, but pollen and insects came from a larger catchment. A grain of Aesculus sp. (horse chestnut) pollen was identified from the pollen sample from Ditch 700. It was clearly a contaminant, because Aesculus sp. is a relatively recent introduction to Britain. Curiously, Aesculus is an insect-pollinated plant which does not scatter its pollen very far and no trees of it were noticed in the vicinity of the site. There is no reason to suspect any of the other pollen from the sample to have had a recent origin.Sort out Aesculus muddle
The evidence of the pollen and Coleoptera suggests that grassland continued as a major aspect of the landscape at Claydon Pike into the late Roman period. It is uncertain whether much of it remained as hay meadow. The pollen from Sample 700/S/3 did include traces of Rhinanthus sp. and Centaurea nigra sp. pollen and the weevils which are favoured by meadowland conditions (Species Group 3) were only slightly less abundant than in the earlier Roman samples. However, seeds of possible hay meadow plants were limited to a few seeds of Leucanthemum vulgare. There was no evidence of hay brought to the site.
Scarabaeoid dung beetles (Species Group 2) at 11.5% of the terrestrial Coleoptera, were almost twice as abundant as in the previous period. They suggest a significant presence of domestic animals around the site. There was not, however, the trampled and waterlogged soil that occurred on the site during phase 2. Seeds of the Juncus effusus group, which are produced prolifically by their parent plants and so small that they can be dispersed by the wind, were not well represented. The phytophagous beetles included various species which feed on grassland herbs such as Gymnetron labile and G. pascuorum, both of which feed on Plantago lanceolata. There were also various Carabidae and Staphylinidae which often occur in grassland, although most will also occur on weedy disturbed ground including arable. One species, Pterostichus madidus, is of particular interest. It is now a particularly common Carabid beetle of various open country habitats but there are very few Roman records of it from Britain and no prehistoric records (Kenward 1982, 72). The four Claydon Pike specimens were from Sample 502/7, which came from the bottom of a stone-lined well. The pottery assemblage from the deposit was primarily 4th century AD in date and some of the biota from the sample, such as waterlogged glumes of Triticum spelta, would have been most unusual finds from a post-Roman context.
There was no evidence for any arable fields in the vicinity of the site although there were a few waterlogged crop remains from the villa and, as has already been noted, cultivation would have been possible on top of the gravel islands.
The exceptional number of seeds of Urtica dioica suggests that stinging nettles grew over large areas of the interior of the villa enclosure. Nettle-feeding insects were also prominent in the samples:
Heterogaster urticae |
Cidnorhinus quadrimaculatus |
||
Brachypterus urticae |
Ceutorhynchus pollinarius |
||
Apion urticarium |
Other tall herbs represented by their seeds that probably grew amongst the nettles included:
Malva sylvestris |
Rumex obtusifolius |
||
Anthriscus sylvestris |
There was only a single seed of M. sylvestris, but the presence of the following mallow-feeding weevils suggests it to have been more abundant:
Apion malvae |
A. aeneum |
||
A. rufirostre |
A. radiolus |
In areas that were disturbed more frequently this vegetation probably gave way to weeds such as Sisymbrium officinale, Stellaria media gp. and . Atriplex sp. which in turn probably graded into almost bare trampled areas with only a scattering of Polygonum aviculare agg. It is also possible that there were grassy areas with plants such as Ranunculus cf. repens and Prunella vulgaris. Splashing of the water around the top of some of the features again provides the only explanation for why Myosoton aquaticum seems to have grown on the site, its seeds being present in several of the samples.
The average percentage of tree pollen from the late Roman samples, at 8.7% was at the same level as from the earlier Roman samples. Fruits of Fraxinus excelsior (Common Ash) were identified from a couple of samples and the scolytid beetle Leperisinus varius, which mostly feeds on ash, was present. There were probably a few ash trees within the villa enclosure or growing along its boundary but otherwise any woodland seems to have been relatively distant from the site.
There was little evidence to suggest the presence of thorn hedges. There were only a few thorny twigs in the samples. Stones of Crataegus sp. and Prunus spinosa were entirely absent. However, buds of Salix sp. and seeds of Viburnum opulus were found in most of the samples and it is likely that these shrubs grew along some of the boundaries.
Samples 1989/B and 2721/6 contained leaves of Buxus sempervirens (Box). Box seeds and fruit fragments were also present in Sample 1989/B. It seems that box hedges or bushes were also cultivated on the site during the late Roman period.
There were again seeds from horticultural crops or possible horticultural crops which could have been grown within the villa enclosure. The only tree fruit was Prunus domestica cf. ssp. insititia (bullace or damson). However, there was also a seed from another fruit, Fragaria vesca (wild strawberry), which could have been cultivated or grown wild. Classical Roman authors refer to the collection of strawberries from the wild rather than their cultivation (Roach 1985, 275).
The herb and vegetable seeds again included Coriandrum sativum (coriander) and Apium graveolens (celery) with Papaver somniferum (opium poppy) and Daucus carota (carrot) as either cultivars or weeds. There was also a single seed of Foeniculum vulgare (fennel). Although there are other records of fennel from Roman Britain (eg. Willcox 1977), its seeds are not nearly as frequently found as are, for example, coriander seeds, and this is the first record for the Upper Thames Valley.
Sample 502/7 contained seeds of Brassica rapa ssp. sylvestris, B. nigra and a third species of Brassica. The first species is wild turnip, a weed of arable and waste land although its young leaves can be cooked and eaten. B. nigra is black mustard and it is possible that it was cultivated for its seeds, it also grows as a weed on waste land. The third species is almost certainly a cultivar of some sort. It could not be identified with certainty, although it closely matched seeds of a cultivated cabbage (B. oleracea). The reason for suspecting it to be a cultivated vegetable is B. rapa ssp. sylvestris and B. nigra are the only wild Brassica species likely to have occurred inland during the Roman period. The wild B. oleracea is a plant of maritime cliffs and the remaining species, apart from the cultivars that have been recorded from Britain, are alien casuals from Asia or the Mediterranean region (Clapham et al., 1987, 70-73). Cabbages and other Brassica cultivars were certainly well known to the Romans (eg. Apicius III, ix, xiii).
The Sphaeridiinae and Oxytelinae which comprise Species Group 7 slightly increased their abundance to 11.2% of the terrestrial Coleoptera. These beetles occur in dung on pasture, manure heaps and foul vegetable material. Their increase was perhaps related to the increased presence of domestic animals on the site in the late compared to the earlier Roman period. The percentage of Lathridiidae (Species Group 8), in contrast, fell to 3.7% of the terrestrial Coleoptera, which may well be due to there no longer being large quantities of hay brought to the site. However, the level of synanthropic beetles (Species Group 9a), at 1.6%, remained much the same. They comprised the same species as in the earlier Roman period except that Tenebrio molitor was absent. Woodworm beetles (Species Group 10) were again present, with Lyctus linearis joining Anobium punctatum. These results are consistent with the archaeological evidence for the presence of buildings on the site.
There was little evidence from the waterlogged plant remains for the importation of plant material. There were no deposits of hay and waterlogged crop remains were limited to a few glumes of Triticum spelta and a seed of Linum usitatissimum. A few arable weed seeds also seem to have been brought to the site including Agrostemma githago and Scandix pecten-veneris, but the waterlogged remains do not suggest any large scale storage of cereals or accumulation of crop-processing waste.
The late Roman villa seems, in large part, to have been concerned with the grazing of domestic animals. Although parts of the floodplain were probably experiencing flooding, there was no evidence from the molluscan samples for flooding extending onto the edge of the villa platform during the Roman period. Grazing on the floodplain seems to have been sufficiently well managed not to have resulted in damage to the sward in the wet areas around the edge of the platform.
It is uncertain to what degree the site was involved in arable agriculture. At one extreme it is possible that crops were imported from elsewhere for consumption at the site. However, it is also possible that the villa estate included some higher ground that was used for arable.
The insect remains suggest two other economic activities which could have taken place at the villa. A total of three heads of worker honey bees (Apis mellifera) were identified from Samples 502/7 and 1989/B. This suggests a nest of honey bees in the vicinity of the site and raises the possibility of beekeeping.
Sample 1989/B contained a total of six Elmidae belonging to the species Elmis aenea, Esolus parallelepipedus and Limnius volckmari. They live in clean flowing water, clinging to stones and aquatic plants. They do not occur in stagnant water, ditches or slowly flowing rivers. At present in the Upper Thames system they are mostly restricted to clean, fast-flowing tributary streams as is, for example, Esolus parallelepipedus (Walker 1911,8) Elmidae were absent from all other waterlogged samples from the site and Feature 1989, a rectangular pit cut below the water table, would have provided no more suitable a habitat for them than any of the other features on the site. The other water beetles from Sample 1989/B, such as Helophorus brevipalpis gp., Ochthebius sp. and Limnebius nitidus, can live in stagnant water. One of the snails from the sample, however, Planorbarius corneus, which was absent from the other samples from the site, is a species of permanent bodies of water which flourishes in ornamental ponds. The Elmidae could have been transported by floodwater, but there was no evidence for flooding and Feature 1989 was on a high part of the platform. A more satisfactory explanation is that Feature 1989 was a tank used for the temporary live storage of fish, and the Elmidae were accidentally introduced with them. One of the items discovered in Feature 1989 was an open wickerwork basket, scoop or fish trap. Elmids would certainly crawl onto a fish trap or keep basket, if it were put into one of the small rivers near the site and would not readily let go if it were lifted out of the water. If the trap or basket were brought back and put into the tank, this would provide a ready means for the introduction of the beetles.
The evidence from the Coleoptera for the ‘intensity’ of occupation of the site and presence of timber buildings is as would be expected for a Roman villa. The evidence for horticultural crops and box hedges is also what might be expected for a villa.
Two waterlogged samples were investigated from a medieval (Phase 5) well cut through the late Roman ruins in Longdoles field, Samples 696/3 and 696/2. Two samples of alluvium of medieval date from the top of Roman ditches were investigated for molluscs (Samples 456 and 1). Water plants do not seem to have grown in the well but various small water-beetles, particularly Helophorus and Ochthebius spp., lived in its waters. Almost all the other plant and invertebrate remains seem to have entered the well through natural agencies. The molluscs from the alluvium comprised both riverine aquatic species which had been transported in floodwaters, and amphibious and terrestrial species which mostly lived on the site.
Pollen analysis (by J. Greig) of the well samples revealed rich open grassland floras. The Coleoptera also comprised rich grassland assemblages. Chafers and elaterid beetles which feed on roots in grassland (Species Group 11) had increased in abundance since the late Roman period to 6.9% of the terrestrial Coleoptera. However, the proportion of Scarabaeoid dung beetles (Species Group 2), at 2.3% of the terrestrial Coleoptera, was very low. This suggests that domestic animals were not concentrated in the vicinity of the well and that the grassland was no more than lightly grazed. The clover and vetch-feeding weevils of the genera Apion and Sitona, which tend to be more prolific in meadowland than pastureland (Species Group 3) were, at 7.8% of the terrestrial Coleoptera, rather abundant. The macroscopic plant remains were almost entirely from grassland plants with a strong hay meadow element. However, the remains were not cut hay which had been brought to the site, they had mostly probably blown into the deposit from the surrounding vegetation. There were so few seeds from non-grassland plants that the well seems to have been set in an expanse of species-rich meadowland. Seeds of the hay meadow plants Rhinanthus sp., Leucanthemum vulgare and Centaurea cf. nigra were conspicuously present, but in addition to seeds, C. nigra was represented by its pectinate bracts and Rhinanthus sp. by capsules. The meadowland environment enabled various herbs which also occur in pastureland to set seed profusely, such as Ranunculus cf. acris, Rumex acetosa and Plantago cf. lanceolata. The grassland seeds from the medieval well included some, such as Cardamine pratensis, Silaum silaus and Primula cf. veris, which are rarely encountered in archaeological deposits. Seeds from the following pastureland species. Potentilla anserina, Plantago major and Eleocharis cf. palustris were absent and there was only a very slight presence of the Juncus effusus gp.
Much of the medieval meadowland at Claydon Pike would have experienced seasonal inundation. The upper fill of the Roman features around the edge of the platforms was silty clay alluvium of medieval date, although alluvium was not recorded from the top of the platforms. Late Saxon and early medieval alluviation in the Upper Thames Valley seems to have extended further than Roman alluviation (Robinson and Lambrick 1984). The molluscan assemblages from the alluvium at Claydon Pike were of a sort which is characteristic of floodmeadow rather than pasture (Robinson 1988). The molluscan assemblages can be divided into three parts. Firstly, there are aquatic species, such as Bithynia tentaculata and Valvata cristata, which would have lived in the river from which the floodwater came. Secondly there are amphibious species, particularly Lymnaea truncatula and Anisus leucostoma, which are able to live on the floodplain and retreat into cracks during dry periods. Finally, there are terrestrial taxa which commonly occur in wet grassland or marshes:
Carychium sp. |
Vallonia pulchella |
||
Succinea or Oxyloma sp. |
Limax or Deroceras sp. |
||
Cochlicopa sp. |
Trichia hispida gp. |
||
Vertigo pygmaea |
This latter group thrives in modern riverside floodmeadows but these species are very rare or absent from closely grazed flood pasture in the Upper Thames Valley.
The seeds suggest some variation in the vegetation. Seeds of Carex spp. were numerous and there were possibly stands of Carex spp. in the wetter parts of the meadowland. The thin dry soil over the Roman ruins perhaps supported Daucus carota and Crepis capillaris.
The medieval meadowland could have belonged to either Alopecurus pratensis - Sanguisorba officinalis flood meadow (MG4) or Cynosurus cristatus - Centaurea nigra meadow (MG5) of the National Vegetation Classification (Rodwell 1992). The former is the characteristic community of alluvial floodmeadows in the Upper Thames Valley that are still managed for hay by traditional methods. It has, however, been argued that in the absence of certain characteristic species amongst the botanical remains, the medieval assemblages show a closer similarity to MG5. Both of these categories are shut up in late winter/early spring, mown in June or July and the aftermath grazed.
The insects do not add much information on the composition of the grassland, although they comprised a very full meadowland fauna. There were numerous cicadellids from the genus Aphrodes which feed on grasses and chrysomelids from the genus Longitarsus which feed on meadowland herbs. There were also various beetles which tend to congregate on meadowland flowers such as Cantharis rustica, Rhagonycha fulva and Oedemera lurida. The carabids from the samples included many specimens of Pterostichus madidus, showing that by the medieval period this species had become very well established in the region.
There was little evidence for open habitats apart from meadowland. Some seeds of Cardamine cf. hirsuta had perhaps come from plants growing between the stonework of the well pit. Otherwise, there were only a few stray seeds from plants of disturbed ground. Curiously, there was a single seed of Centaurea cyanus, a weed that is closely tied to arable agriculture. None of the phytophagous Coleoptera is restricted to feeding on weeds of disturbed ground.
Tree pollen was very sparse, comprising 0.8% of the total identified pollen. Wood and tree-dependent Coleoptera (Species Group 4) were absent. The medieval landscape seems to have been very open indeed. The trees or woodland that were giving values of just under 10% for tree pollen from the 1st century AD to the late Roman period had been lost. There was only a slight presence of shrub pollen and a couple of seeds of Sambucus nigra from the samples. Any scrub or hedges seem to have been a minor or distant feature of the landscape.
Woodworm beetles (Species Group 10) were absent and the low values for the other groups of beetles associated with various sorts of accumulated organic material (Species Group 7-9) are consistent with the absence of any medieval settlement on the site. The values for these groups are mo more than would be expected for grassland. However, the occurrence of a couple of individuals of Typha stercorea might hint at a small accumulation of old hay.
Little remains to be said about site activities and the use of the site because the medieval evidence is for a block of meadowland rather than an occupation site surrounded by a somewhat varied landscape. Hay meadow does seem, on the basis of documentary and molluscan evidence, to have been one of the major uses of the floodplain of the Upper Thames Valley during the medieval period (Robinson 1988).
This work was done in the University Museum, Oxford and funded by HBMC (England). I am extremely grateful to Dr J R A Greig for pollen analysis of the medieval samples, Dr J Turner for pollen analysis of the 1st century AD and Roman samples, Dr F Rose for the identification of moss, and Miss P R Tomlinson for work on vegetative plant material. I would like to thank Miss V R Straker for help with identification of some of the seeds.
Allen, T G and Robinson, M A, 1993 The prehistoric landscape and Iron Age enclosed settlement at Mingies Ditch, Hardwick-with-Yelford, Oxon. Oxford: Oxford University Committee for Archaeology
Apicius, The art of cooking (Flower, B and Rosenbaum, E (trans), 1958 London: Harrap)
Bean, W J, 1914 Trees and shrubs hardy in the British Isles, 2.1 edn. London: John Murray
Clapham, A R, Tutin, T G and Moore, D M, 1987 Flora of the British Isles, 3 edn. Cambridge: University Press
Edwards, F W, Oldroyd, H and Smart, J, 1939 British blood-sucking flies. London British Museum (Natural History)
Evans, J G, 1972 Land snails in archaeology. London: Seminar Press
Godwin, H, 1975 The history of the British Flora. Cambridge: University Press
Jarvis, M G, 1973 Soils of the Wantage and Abingdon district, Memoirs of the Soil Survey of Great Britain: England and Wales, Harpenden
Kenward, H K, 1982 ‘Insect communities and death assemblages, past and present’, in Hall, A R, and Kenward, H R (eds) Environmental archaeology in the urban context (Counc Brit Archaeol Res Rep 43). London: Counc Brit Archaeol, 71-8
Kerney, M P, 1976 A list of the fresh and brackish-water Mollusca of the British Isles, J. Conchology Lond. 29, 26-28
Kislev, M E, 1988 ‘Pinus pinea in agriculture, culture and cult’, in Küster, H.-J. (ed) Der prähistorische Mensch und seine Umwelt. Stuttgart: Konard Theiss Verlag, 73-79
Kloet, G S and Hincks, W D, 1964 A check list of British insects, 2nd edition (revised): small orders and Hemiptera (Royal Entomological Society of London; Handbook for the Identification of British Insects 11, pt 1). London: Royal Entomological Society
Kloet, G S and Hincks, W D, 1976 A check list of British insects, 2nd edition (revised): Diptera and Siphonaptera (Royal Entomological Society of London; Handbook for the Identification of British Insects 11, pt 5). London: Royal Entomological Society
Kloet, G S and Hincks, W D, 1977 A check list of British insects, 2nd edition (revised): Coleoptera and Strepsiptera (Royal Entomological Society of London; Handbook for the Identification of British Insects 11, pt 3). London: Royal Entomological Society
Kloet, G S and Hincks, W D, 1978 A check list of British insects, 2nd edition (revised): Hymenoptera (Royal Entomological Society of London; Handbook for the Identification of British Insects 11, pt 4). London: Royal Entomological Society
Krüssmann, G, 1985 Manual of cultivated conifers, 2 edn. London: Batsford
Lambrick, G H and Robinson, M A, 1979 Iron Age and Roman riverside settlements at Farmoor, Oxfordshire (Counc Brit Archaeol Res Rep 32). London: Counc Brit Archaeol
Lambrick, G H and Robinson, M A, 1988 ‘The development of floodplain grassland in the Upper Thames Valley’, in Jones, M K (ed) Archaeology and the flora of the British Isles) (University Committee for Archaeology Monograph 14). Oxford: University of Oxford, 35-75
Lloyd, C, 1970 The Well-Tempered Garden. London: Collins
Miles, D, 1986 Archaeology at Barton Court Farm, Abingdon, Oxon (Counc Brit Archaeol Res Rep 50). London: Counc Brit Archaeol
Moffett, L, Robinson, M A and Straker, V, 1989 ‘Cereals, fruit and nuts: charred plant remains from Neolithic sites in England, Wales and the Neolithic economy’, in Milles, A, Williams, D and Gardener, N (eds)) The Beginnings of Agriculture, (British Archaeological Reports International Series 496). Oxford: BAR, 243-61
Pearson, E and Robinson, M A, 1994 ‘Environmental evidence from the villa’, in Williams, R J and Zeepvat, R J, Bancroft, a late Bronze Age / Iron Age settlement, Roman villa and temple / mausoleum 2, Buckinghamshire Archaeological Society 7, 565-84
Rahtz, P A and Greenfield, E, 1977 Excavations at Chew Valley lake, DoE Archaeol Rep 8
Richmond I A and Gillam J P, 1951 The temple of Mithras at Carrawburgh, Archaeologia Aeliana 29, 1-92
Roach, F A, 1985 Cultivated fruits of Britain. Oxford: Blackwell
Robinson, M A, 1981 ‘The Iron Age to Early Saxon Environment of the Upper Thames Terraces’, in Jones, M and Dimbleby, G (eds) The Environment of Man: the Iron Age to the Anglo-Saxon period (Brit Archaeol Rep 87). Oxford: BAR, 251-86
Robinson, M A, 1983 ‘Arable/Pastoral Ratios from Insects’, in Jones, M (ed) Integrating the Subsistence Economy (Brit Archaeol Rep S181). Oxford: BAR, 19-55
Robinson, M A, 1988 Molluscan evidence for pasture and meadowland on the floodplain of the Upper Thames Basin, in The exploitation of wetlands (ed P. Murphy and C. French), (Brit Archaeol Rep 186). Oxford: BAR, 101-112
Robinson M A and Lambrick, G H, 1984 Holocene alluviation and hydrology in the Upper Thames Basin, Nature 308, 809-814
Robinson, M and Wilson, R, 1987 ‘A Survey of Environmental Archaeology in the South Midlands’, in Keeley, H C M (ed) Environmental Archaeology: a Regional Review Vol 2 (Historic Buildings and Monuments Commission for England Occasional Paper, 1). London: HBMCE, 16-100
Rodwell, J S (ed), 1992 British plant communities 3. Grasslands and montane communities, Cambridge: University Press
Silverside, A J, 1977 ‘A phytosociological survey of British Arable-weed and related communities’, University of Durham phD thesis
Southwood, T R E and Leston, D, 1959 Land and water bugs of the British Isles, London: Warne
Waldén, H W, 1976 A nomenclatural list of the land mollusca of the British Isles, J. Conchology Lond 29, 21-25
Walker, J J, 1911 Preliminary list of the Coleoptera observed in the neighbourhood of Oxford from 1819 to 1907 : third supplement, Ashmolean Natural History Society of Oxfordshire Report, 45-54
Walters, S M, 1949 Eleocharis L., J. Ecology 37, 192-206
Wheeler, B D, 1980 Plant communities of rich-fen systems in England and Wales III. Fen meadows, fen grassland and fen woodland communities and contact communities, J. Ecology 68, 761-88
Willcox, G H, 1977 Exotic plants from Roman waterlogged sites in London, J. Archaeol. Sci. 4, 269-282