4.6 Insect Remains by Mark Robinson

Cite this as: Robinson, M. 2015 Insect Remains, in M. Atkinson and S.J. Preston Heybridge: A Late Iron Age and Roman Settlement, Excavations at Elms Farm 1993-5, Internet Archaeology 40.

Wells and other potentially waterlogged archaeological features were excavated where encountered across the investigation area, in the Northern, Central and Southern Zones of the Roman settlement. However, a post-Roman rise in the water table created difficulties in reaching Roman waterlogged levels, so only a limited number of deposits in which insects were preserved were sampled. Assessment established that useful assemblages of insects were present in soil samples from four timber-lined wells (Table 168). The insects from these were found to have the potential to provide information on the intensity and character of the occupation of the site, which is relevant in establishing whether or not the settlement can be regarded as a town. It was decided to analyse samples from the bottom of each of the wells, plus a sequence of samples from Well 8188, in which many of the insects had probably been introduced with dumped refuse.

The contexts sampled from Wells 8188, 9421, 14984 and pit 13883 are described elsewhere in the stratigraphic narrative account. Sample weights, sample numbers, context numbers and dates of the well fills are given in Table 169 (.xlsx). A sequence of three samples was investigated from 2nd-century well 8188, with Sample 1848 at the bottom and Sample 1841 at the top. Insect remains were sparse and badly preserved in Sample 1843 from fill 8195, a layer of clay that separated the organic sediment of Sample 1841 from the organic sediment below, so the insects were not fully analysed from this sample.

Table 168: Insect - sample information
Well (feature no.) Sampled contexts Sample nos Area Period
Well 8188 (Group 788) 8196/8214 1841, 1845, 1848 E 3
Well 9421 (Group 772) 9895 498 D 3
Well 14984 (Group 710) 20034 812 L 4-5
Pit 13883 (Group 595) 18216 1084 I 4
Figure 685: Species Groups of Coleoptera from Elms Farm

Each sample was washed over onto a 0.2mm mesh to recover the organic fraction and subject to paraffin flotation to extract the insect remains from it. The flots were washed in detergent and sorted in water under a binocular microscope. The insect fragments were identified with reference to the collections of the Oxford University Museum of Natural History at magnifications of up to x100. The results are given in Table 169 and Table 170 (.xlsx), the nomenclature following Kloet and Hincks (1977) for Coleoptera. Some of the Coleoptera have been assigned to habitat-related species groups as an aid to interpretation following Robinson (1991, 278-91), except that the synanthropic group has been divided into general synanthropic species and serious pests of stored grain (Figure 685). The members of the groups are shown in Table 169 (.xlsx). Not all the Coleoptera have been classified into one of these groups.

The origin of the assemblages

With the exception of Sample 1084, from the bottom of Pit 13883, water beetles (Figure 685, Species Group 1) were at values below 5% when expressed as percentages of the total numbers of terrestrial Coleoptera. They were mostly small beetles, such as Helophorus cf. brevipalpis, which could have lived in stagnant water in the wells but they readily leave water and could also have been part of the 'background rain' of insects that fell into the well. The same species comprised around 15% when expressed as a percentage of the total terrestrial Coleoptera from Sample 1084, suggesting it was likely the well did support a population of these beetles. The sample also contained three individuals of Tanysphyrus lemnae, a weevil that feeds on the minute floating plant Lemna sp. (duckweed). It is likely this plant covered the surface of the water in the well. All the samples except Sample 498 from Well 9421 contained larval head capsules from Chironomidae and it is likely that these midges developed in the wells. A few examples of the beetles Lesteva longoelytrata and Platystethus cornutus gp. were present in some of the samples and it is possible that they were from splash zones around the tops of the wells. Otherwise almost all the insects were terrestrial species from the surrounding landscape.

Most of the insects in the samples from the wells probably entered through various natural agencies. Some could have been among organic refuse dumped in them but could equally have emerged from suitable habitats nearby. The proportion of synanthropic beetles in Samples 1848, 1845 and 1841 from Well 8188 (Figure 685, Species Groups 9a and 9b) was higher than from the other wells, as was the proportion of some other groups of Coleoptera from indoor habitats. It is very likely that organic refuse from inside a building, complete with its insect fauna, had been dumped into the well. The well did, however, also contain many insects likely to have fallen in from the surrounding landscape.

General site conditions

The landscape surrounding the site seems to have been relatively open. Although the wood and tree-dependent beetles of Species Group 4 were identified from two of the wells, they were mostly species that occur on dead wood, which could have been imported for fuel. A couple of examples of Leperisinus varius, a bark beetle of Fraxinus excelsior (ash), however, were as likely to have been from trees growing in the settlement as from recently felled trees that had been brought to the site. There was also a record of Chalcoides sp., a leaf beetle of Populus and Salix spp. (poplar and willow). Despite the present low-lying nature of the site and the proximity of the sea, there was no evidence for marshy conditions or brackish habitats. The main terrestrial habitats suggested by the insects were weedy waste ground, some areas of pasture, accumulations of decaying organic material, timber buildings and, in one instance, stored grain.

Well 9421, Fill 9895, Group 772, Area D: Sample 498

The main terrestrial habitat around the top of the well seems to have been weedy waste or neglected ground. There was an appropriate range of Carabidae (ground beetles) and Staphylinidae (rove beetles) for such a habitat although there was only a single individual, Harpalus rufipes, of the Carabidae of general disturbed ground and arable habitats (Species Group 6a).

Beetles that feed on Malvaceae, especially Malva sylvestris (common mallow), a waste ground plant that does not withstand grazing, included:

Podagrica fuscicornis	A. aeneum
Apion malvae		A. radiolus

Beetles that feed on Urtica dioica (stinging nettle) included:

Brachypterus urticae	Cidnorhinus quadrimaculatus
Apion urticarium

The clover and vetch-feeding weevils of Species Group 3 from the genera Apion and Sitona, which are favoured by hay meadow conditions, were quite well represented at 4.4% of the terrestrial Coleoptera. In the absence of any other evidence for hay meadow, it seems likely that these beetles were flourishing on the waste ground vegetation. There were probably some grassy areas because homopteran bugs from the genus Aphrodes, which feed on grasses, particularly A. bicinctus, were quite well represented and elaterid beetles with larvae that feed on the roots of grassland herbs (Species Group 11) comprised 2.5% of the terrestrial Coleoptera. Among them were members of the genus Athous: A. bicolor, A. haemorrhoidalis and A. hirtus. Other grassland insects included the weevil Gymnetron labile, which feeds on Plantago lanceolata (ribwort plantain). One of the most numerous insects from the sample were workers of the ant Formica cf. fusca, which nests in uncultivated land under stones and in old tree stumps.

The proportion of scarabaeoid dung beetles such as Aphodius spp., which feed on the droppings of large herbivores was, at 3.4% of the terrestrial Coleoptera (Figure 685, Species Group 2), only sufficient to indicate the presence of a few domestic animals in the vicinity. A much higher value would have been expected if the well had been adjacent to an enclosure used to confine stock.

Coleoptera of foul organic material were quite well represented. Species of Cercyon, Megasternum and Anotylus, which belong to Species Group 7, comprised 8.8% of the terrestrial Coleoptera. It is possible that they had been among decaying plant material that had been left near the edge of the pit or dumped in it. However, these beetles were no more abundant than is usual for rural settlements and very much higher values are often recorded from urban contexts.

The Lathridiidae (Species Group 8) also comprised 8.8% of the terrestrial Coleoptera, Lathridius minutus gp. being the most numerous. They feed on moulds on plant material such as old straw and hay. They are very much favoured by the habitats that occur in settlements, such as thatch, haystacks and animal bedding, as well as plant refuse. Their abundance suggested a strong presence of their habitats. Other beetles that occur in material such as old haystack refuse included Acritus nigricornis, Xylodromus concinnus, Aglenus brunneus and one of the general synanthropic species (Species Group 9a), Typhaea stercorea.

Beetles that infest structural timbers (Species Group 10), in particular Anobium punctatum (woodworm beetle), were, at almost 6% of the terrestrial Coleoptera, very well represented. They occur in seasoned, dry dead wood, which is not a common natural habitat, but they find the timber of buildings an ideal food source. The results certainly imply the proximity of a timber structure. General synanthropic beetles (Species Group 9a) were, however, poorly represented so it is unlikely that sweepings or refuse from inside a building had been dumped in the well. Grain beetles were absent.

Two other species of insect from the well are of interest. There were a couple of workers of Apis mellifera (honey bee), which raises the possibility that bees were being kept. There were also two individuals of the beetle Trox scaber, which feeds on dry animal skins and sinews on old bones. It is sometimes associated with animal-skin processing but it also occurs, for example, on dead birds in roof spaces and on debris in owls' nests.

Well 14984, Fill 20034, Group 710, Area L: Sample 812

The insects from this well gave even stronger evidence than those from Well 9421 for waste or neglected ground habitats, but there was much less evidence of human settlement. Beetles that feed on Urtica dioica (stinging nettle) comprised over 14% of the terrestrial Coleoptera. Ceutorhynchus pollinarius joined the nettle-feeding Coleoptera recorded from Well 9421, along with the nettle-feeding bug Heterogaster urticae. The weevils Apion aeneum and A. radiolus, which feed on Malvaceae (mallows) were again present. The most numerous ground beetles were Harpalus S. Ophonus sp., which flourish on weedy ground, while the ant Stenamma sp., which nests in shady and sheltered habitats, was well represented.

The Elateridae whose larvae feed on the roots of grassland plants (Figure 685, Species Group 11) were quite well represented and there was double the proportion of scarabaeoid dung beetles that feed on the droppings of domestic animals on pasture, in comparison with Well 9421 (Species Group 2). Although still not a very high value for Species Group 2, it was sufficient to show domestic animals were being kept somewhere nearby.

Coleoptera of foul organic material (Species Group 7) were no more abundant than they are in assemblages that form in deposits distant from human habitation. The Lathridiidae (Species Group 8), at 5% of the terrestrial Coleoptera, gave a value typical of rural settlements. The woodworm beetles of Species Group 10 and the synanthropic beetles of Species Groups 9a and 9b, however, were entirely absent.

Pit 13883, Fill 18216, Group 595, Area I: Sample 1084

The insect assemblage from this pit differed from those from the wells in that there was a significant autochthonous component of aquatic insects and insects which fed on water plants likely to have been living in the pit. The quantity of insects from the pit was also low, so interpretation of terrestrial conditions is less detailed than for the other wells. However, phytophagous insects again suggested a significant presence of waste ground with nettles. The most numerous insects were workers of the ant Tetramorium caespitum, which nests in sandy and gravelly soils, especially on heathland and near the sea. Its occurrence possibly reflected areas of shorter vegetation. There was only a low presence of scarabaeoid dung beetles.

Insects of accumulations of decaying organic material were sparse, although there was a significant occurrence of beetles associated with timber structures. Anobium punctatum (woodworm beetle) (Species Group 10) comprised 5% of the terrestrial Coleoptera. There was also an example of the general synanthropic beetle Ptinus fur and another beetle often associated with human habitation, Aglenus brunneus. Apis mellifera (honey bee) was again present.

Well 8188, Fill 8196, Group 788, Area E: Samples 1848 and 1845, Fill 8214, Sample 1841

The three samples that contained useful quantities of insect remains from the sequence from Well 8188 gave relatively similar results. A major proportion of the insects were beetles associated with organic refuse and indoor habitats. Many of the remainder can be found on weedy or grassy waste ground, including the nettle-feeding weevil Ceutorhynchus pollinarius and the weevil Mecinus pyraster, which feeds on Plantago lanceolata (ribwort plantain) and P. media (hoary plantain). Scarabaeoid dung beetles (Species Group 2) suggested a background presence of pasture but they were not abundant.

The beetles of general foul organic material, which belong to Species Group 7, ranged from 8 to 12% of the terrestrial Coleoptera (Figure 685). Most numerous was Megasternum obscurum but species of Cercyon, such as C. analis, were also well represented. Other species that tend to be associated with organic refuse in settlements, for example Oxytelus sculptus, were present. However, there were very few of the Diptera puparia, which occur in large dumps of stable refuse and manure heaps; for example there was only one puparium of Musca domestica (house fly) from the three samples. This suggests that although some foul organic refuse was present in the vicinity of the well, stable-type refuse containing fly puparia was not dumped into it.

The Lathridiidae (Species Group 8) ranged from 10.5 to almost 20% of the terrestrial Coleoptera, much higher values than from the other wells (Figure 685). Lathridius minutus gp. was again the most numerous. Either there was a major presence of old hay, straw or thatch nearby, or the insects had been among debris discarded into the well. They could easily have been derived from floor sweepings from inside a building. There was certainly a presence of Anobium punctatum (woodworm beetle) high enough to confirm the proximity of timber structures or to suggest that debris from a building had been incorporated into the well fill. A. punctatum, along with Lyctus linearis (powder post beetle), another beetle that attacks structural timbers and wooden artefacts, ranged from 4.8% to over 8% of the terrestrial Coleoptera from the well samples (Figure 685, Species Group 10). General synanthropic beetles (Species Group 9a), which were likely to have been living in indoor habitats, were well represented, ranging from 4% to over 6% of the terrestrial Coleoptera. Most numerous was Ptinus fur, but there were also examples of Stegobium paniceum, Tipnus unicolor, Typhaea stercorea and Tenebrio obscurus. Such a fauna might be found in a wide range of buildings with some slightly damp plant material on the floor and a little starchy material, perhaps a few spilt cereal grains or waste from food preparation in neglected corners, ranging from a stable or barn through to a dwelling house. They could alternatively have entered the deposits among floor sweepings.

Well 8188 was the only one to contain examples of beetles that are serious pests of stored grain (Species Group 9b). They varied from 0.5 to over 7% of the terrestrial Coleoptera in the three samples. They were mostly Oryzaephilus surinamensis but there were single individuals each of Cryptolestes ferrugineus and Sitophilus granarius. All are likely to have been Roman introductions to Britain that were able to flourish in above-ground grain stores. S. granarius is the only one that can readily attack intact grain in good condition, but once the surface of the grain has been damaged, it is vulnerable to attack by the other two species. Damp conditions that cause fungal growth can render grain liable to attack by O. surinamensis and once infestation has begun, sufficient warmth and moisture is generated by the metabolism of the beetle to cause the spoilage of further grain, making the infestation self-sustaining. While it is possible that spoilt grain had been discarded into the well, given the evidence of other beetles of indoor habitats from the well, it is also very likely that sweepings from the floor of a building, in which grain had been stored, had been thrown into the well.

In addition to the woodworm beetles there were a few other insects of dead wood, for example the beetle Bitoma crenata, which occurs under loose bark. It is possible that they had been living among firewood.


The insect results present an overall picture of a settlement with open areas supporting vegetation such as nettles between timber buildings. Grassland grazed by domestic animals seems to have been a background to the settlement, rather than the wells being situated next to paddocks for stock within the settlement. Analysis of insects from a Roman well at nearby Langford Road, Heybridge (Langton and Holbrook 1997), showed evidence for weedy, disturbed ground but few synanthropic species were present (D. Smith, pers. comm.).

Table 171: Summary of insect analysis
Context Feature Type Area Period Environs
9895 <498> 9421, Group 772 well D 3 Weedy waste/neglected land (not grazed), grassy, uncultivated. Only a few livestock, some rubbish dumped in well? Inc. straw. Timber structure nearby?
8196/8214 <1841> <1845> <1848> 8188, Group 788 well E 3 Lots of organic refuse (not foul), indoor habitats, + weedy/grassy waste ground. A few animals. Hay/straw/thatch, possible floor sweepings in/near well.
20034 <812> 14984, Group 710 well L 4-5 Waste/neglected ground, but little human settlement. Lots of nettles. Some animals?
18216 <1084> 5806, Group 639well I 4 More pond life. Waste ground with nettles, though shorter veg? Very few animals. Low waste, structures nearby.
Langford 241 well - (3-6) Predominantly arable land, no close habitation

The issue arises as to whether the settlement should be regarded, on the insect evidence, as urban or rural. There was certainly a strong indication from woodworm beetles for the presence of timber buildings. Anobium punctatum (woodworm beetle) and Lyctus sp. (powder post beetle) (Species Group 10) were several times more abundant from Elms Farm than from rural Roman settlements in the Upper Thames Valley, including the Barton Court Villa (Robinson 1981, 280-1), whereas they were similarly numerous from some Roman towns, e.g. Alcester (Osborne 1994). However, suitable deposits from the Meare West Iron Age Lake Village contained even higher proportions of A. punctatum (Girling 1979). The general synanthropic beetles of Species Group 9a, such as Ptinus fur and the members of the Lathridiidae (Species Group 8), such as Lathridius minutus gp., were likewise better represented than on the Upper Thames Valley rural Roman settlements, but again they were no more abundant than from the Meare West Iron Age Lake Villages.

The occurrence of serious pests of stored grain (Species Group 9b) including Oryzaephilus surinamensis and Sitophilus granarius in one of the Roman wells does probably have implications for the status of the site. These grain beetles appear to have been introduced and widely distributed in England during the 1st century AD as a result of the activities of the Roman army, for example being present at the fortresses of York (Kenward and Williams 1979) and Exeter (Straker et al. 1984). They tend to be found on military sites, in walled towns and on very large or high-status rural settlements (Robinson 1992, 59). They have so far proved absent from rural settlements in the Upper Thames Valley (e.g. Robinson 1981) and were not found at the small unwalled Roman towns of Tiddington, Warks, or Scole, Norfolk (Robinson forthcoming). Their occurrence was possibly related to the large-scale storage of grain in a fully cleaned (i.e. de-husked) state.

What the Elms Farm assemblage did not show, however, was the highly developed urban insect fauna that has been recorded from some Roman towns, for example York (e.g. Hall and Kenward 1990) and Lincoln (Kenward, pers. comm.). These towns have included a much wider range and higher concentrations of species associated with organic refuse, ectoparasites of humans and their domestic animals and far more other synanthropic species. The results therefore suggest that Heybridge had a character similar to that shown by some smaller Roman towns and large rural settlements but did not have the concentration of urban habitats shown by some of the major towns. There were probably large areas of weedy ground and open spaces between buildings.


Internet Archaeology is an open access journal. Except where otherwise noted, content from this work may be used under the terms of the Creative Commons Attribution 3.0 (CC BY) Unported licence, which permits unrestricted use, distribution, and reproduction in any medium, provided that attribution to the author(s), the title of the work, the Internet Archaeology journal and the relevant URL/DOI are given.

University of York legal statements