Welcome to the Urban Archaeology blog. Freelance archaeologist Chiz Harward provides a range of on and offsite services to the archaeological profession, including running and working on excavations, post-excavation services, training and development work, and illustration work. This weblog will carry news of projects as and when they happen as well as wider thoughts on archaeological issues, especially recording, stratigraphy and training.

Horse and Groom Inn excavation medieval farm reconstruction drawing

I've been working on the analysis and publication text for the LP Archaeology excavation at Horse and Groom Inn, Bourton-on-the-Hill, Gloucestershire. As well as prehistoric remains including Middle Iron Age burials, we excavated the well-preserved remains of a set of medieval masonry farm buildings set around a courtyard. 
Medieval farm buildings at Horse and Groom Inn, looking north west
As part of this post-excavation work I've been working on establishing what these buildings would have actually looked like -we have an almost complete ground plan of the buildings within the site, and taking into account what we know of the construction techniques, materials, and contemporary examples, we can reconstruct the buildings 'out of the ground'. 

This reconstruction is very important to the understanding of the buildings, as what may seem the obvious evolution of a building in plan, may be impossible once you take into account the gables and roofs, or when you think about how you actually get into a room, or up to the first floor. 

In this case we can establish that the main building had thick masonry walls, possibly supporting a second story, although one section of wall was significantly slighter and may have had a timber framed superstructure. The lack of any Cotswold tile fragments suggests the roofs were thatched -and that gives us a roof pitch of at least 45 degrees. We excavated one hearth, which would not have had a chimney -instead the smoke would have filtered through the thatch, or possibly a louvre in the ridge line. Later additions were partly built in timber and may have been open-fronted cattle byres or pig sties. Animal bones recovered from the site suggest a wide range of animals in the vicinity -dogs, cats, and geese as well as livestock. The buildings would have looked and smelled very much like a farmyard!

This process also provides a skeleton framework that can be used as the basis for a reconstruction drawing such as the one above. This is a draft that can be used to discuss the building, and further refine our ideas of the building's appearance. Whilst many elements will be a 'artist's impression' of the site, it will be firmly based on the excavated evidence -down to the chickens and dungheaps!

You can read more about the excavations at http://urban-archaeology.blogspot.co.uk/2013/11/construction-demolition-and-recycling.html, and elsewhere on this site

Pottery illustration at Horse and Groom Inn, and some thoughts on artefact illustration

Composite illustration of a late medieval Malvernian ware dripping dish from Horse and Groom Inn
The image is of a late medieval 'Malvernian ware' dripping dish from the Horse and Groom Inn site in the Cotswolds. The dish was found smashed on the floor of the medieval farmhouse, with other sherds found in a rubble deposit a few metres away. It may have been broken during the clearing out of the building before their demolition.

The dripping dish would have been used whilst roasting meats, and its presence, alongside fragments from other dripping dishes, suggests the buildings were a wealthy peasant or manorial site.

The dish would have had a handle on just one side, with spouts at each end. The dripping dish is in an orange fabric with a copper-speckled orange glaze on the inner surface; the outer surface is fairly heavily sooted from use. The dripping dish probably dates from the 15th century.
Traditional line and stipple illustration of the same late medieval Malvernian ware dish
I have prepared two illustrations of the dish, a traditional pen and ink line and stipple drawing - cleaned up and enhanced in CorelDRAW, a computer graphics package- and a composite image produced in CorelDRAW using photographs of some of the sherds, and using some of the line work of the previous drawing. 
Archaeological finds drawings aim to accurately show the artefact, allow comparison with artefacts from other sites, subtly highlight any particular features, as well as hopefully give some 'feel' of the object. As with many archaeological techniques there are conventions and standardised ways of presenting the artefacts and these conventions allow complex objects to be 'read' easily and consistently -once you know the rules! Clarity is essential in such drawings, and traditional pen and ink drawings offer this, especially when images are often reduced to quarter-size (or smaller) for publication. Pottery specialists will want to see clues as to whether the vessel is wheel-thrown or hand made, see sooting or other 'use' residues, all on an image perhaps only an inch high. Here the composite dripping dish drawing shows the fabric and glaze better than the 'traditional' version, but there is distortion from the photography due to the 3-D nature of the dish, and it can be hard to pick out the precise details of the form.

In this case the final publication drawing will be the traditional pen and ink version, however the composite illustration will be used for public lectures where it will be more immediately understood by those not used to archaeological conventions. 
The publication venue is of critical importance in considering what approach to take: what looks good on a computer screen at one size may look terrible on a printed page or at a larger scale on a site display board. The audience is also of great importance: archaeologists will expect standard conventions to be followed, but images produced for the archaeologist may be meaningless to an audience that is unused to cross-sections, tick lines and scales.
Composite illustration of a Middle Iron Age jar from Horse and Groom Inn; at this size the image shows the external appearance of the jar, however if reproduced at a quarter size, as is usual for many pottery illustrations, it may lack clarity and appear too diffuse. Partly this is due to the 'flat' appearance of the pottery, despite many attempts at accentuating the 3-D curve of the jar using raking lights and 'soft-boxes'.
Digital images are becoming increasingly common in archaeology, the ready availability of digital cameras means that many artefacts that would once have been drawn, are now photographed, however there are dangers with this approach, notably the distortion of more complex or 3-D objects which can be severe. The digital approach is often thought to be quicker, and therefore cheaper, than traditional drawing techniques, however for more complex objects there is often little difference in time spent, and the clarity of traditional drawings is often better, especially when objects are corroded and can look like an amorphous green lump! Whilst some flat artefacts can be easily photographed and checked for distortion, artefacts like broken pottery vessels can be a real nightmare to photograph, and often it just isn't possible.
Three Iron Age pottery vessels from Horse and Groom Inn; whilst the right hand jar <P3> could be reconstructed and photographed to create a composite illustration, the other two vessels couldn't, so a decision was made to use traditional techniques for all three vessels

For a 'traditional' drawing the pottery vessel is drawn at scale, to a format that shows both the external and internal aspects of the vessel. Cross-sections are used to show the thickness and form of the vessel, and of details like the handle and spout, their positioning is shown with 'tick-lines'. The drawing is initially done in pencil, and then traced over in pen on a plastic draughting film called 'permatrace'. Hand stippling and line work is added to emphasise the form and detail and enhance the three-dimensional quality of the drawing. The drawing is then scanned and loaded into CorelDRAW where outlines and tick-lines are redrawn using the computer pen tools -this allows you to select a specific line-width, and alter it to suit the final image- and the cross-sections are filled. The different component parts of the image can be moved around and the layout finalised.
Most of those stages are also needed for composite drawings, instead of the line and stipple detail the photographs are used, however nearly all other stages must be done to produce a drawing that is readable by pottery specialists who need to compare forms and sizes across multiple assemblages.

Composite illustration of a medieval Brill/Boarstall flagon from Horse and Groom Inn
One final aspect that is often overlooked is that with traditional drawings the illustrator has to look very closely at every artefact, from every angle, and may notice construction details that may be otherwise missed. These details can be accentuated and brought out in the finished drawing and also passed back to the artefact specialist.
Roman zoomorphic copper alloy key from Southwark, London. This traditional drawing shows every detail of the artefact, without distortion, in one drawing. With such a complex 3-D object photography cannot easily show all the details like the ring and dot decoration and the lion's mane in one image; tiny details can be 'lost' amongst the colour and shadows, whereas they can be accentuated and shown by hand.

As with everything, the final decision on the type of illustration should be made by actively considering the needs of both the individual artefact and the assemblage as a whole, the publication venue, scale and format. There is no simple rule, and both techniques are very useful and indeed compliment each other.

Post-excavation -going beyond the MoLAS Site Manual

Post-excavation work is well under way on the 100 Minories excavation dataset. The last of the finds are all being washed and the samples processed ready to go off to the specialists to prepare their reports, however I've been busy 'sub-grouping' the contexts, which is a job that involves being in the warm and dry. Perfect for this time of year.
The MoLAS Archaeological Site Manual -still the definitive work for on-site recording, but what do you do with the records?

We excavated and recorded several thousand individual contexts on the 100 Minories site, each context a separate 'event' in the archaeological sequence, each with a set of records and each potentially with associated artefacts and environmental material. Understanding all these thousands of contexts is a daunting task, so we aggregate the contexts into a series of larger groupings, working from context level through 'Sub-groups', 'Groups' and finally to the level of landscape 'Land-use' units like Buildings, Open Areas and Waterfronts. 
This process provides the structure behind a wider methodology that leads you through the post-excavation process, creating larger and larger units, working up through higher levels of understanding and interpretation, and eventually providing the basic structure for the final publication text. The system was developed at the Museum of London's Department of Urban Archaeology, and is designed to fit with the Single Context Recording system described in their manual -although it can be used for almost any logical recording system.

Subgrouping is the first step in this process, and is where we aggregate individual contexts into larger groupings (subgroups) based on their stratigraphic position and interpretation, working with the context matrix. Typically there may be 2 or 3 related contexts in a subgroup, but sometimes there is only one, and sometimes a lot more. Each subgroup consists of a context or contexts which are immediately related to each other and of the same date, so a subgroup might be the construction cut for a wall, the wall itself, and the construction backfill; a mortar bedding layer and its brick floor; or a row of posthole cuts. However the disuse backfill of a cellar cannot be in the same sub-group as the cellar wall as it is obviously not contemporary, this can make things tricky as generally the dating and other specialist data is not yet ready, so sub-groups are often kept small so we don't have to dismantle them later on.

The subgrouping process is carried out using a combination of tools and resources, at the centre is LP Archaeology's Ark browser based database: this contains all the context level information, as well as scanned copies of the context sheets, and is where we enter the new sub-group details and descriptions. Alongside this we use a QGIS project containing all the spatial data from the site -all the digitised plans, trench edges, section lines and other spatial elements; during sub-grouping we check all this spatial data and use the QGIS project to interrogate the contexts spatially. Finally there is the digital copy of the context level Harris Matrix, which is annotated with the subgroups as they are created, and a sub-group level matrix is then created.
LP Archaeology's ArkTools suite of plug-ins for QGIS allows querying, digitising, editing and manipulation of the dataset, including the site Harris matrix

The team at LP Archaeology have been busy building new Open Source tools to help in this process, building a set of tools to rapidly digitise the context plans and sections into our QGIS project, but also tools to check those drawings, and explore the spatial relationships between them and link to the database which contains bespoke forms for subgrouping allowing rapid inputting and querying. 

Alongside this work we are writing a work-flow based guide describing the sequence of tasks within the post-excavation process, using these new LP Archaeology QGIS tools and the Ark database, to guide users through the various interlinked tasks within a post-excavation programme. When completed this will cover post-excavation from on-site checking and data entry through the off-site processes and initial writing tasks; it will pick up where the MoLAS 3rd edition Archaeological Site Manual stops, and will provide a clear structure for post-excavation work on Single Context Recording sites.