Roman Glass Research Project

I have now started on the huge task of going through the entire collection to record the details of the different types of waste on our database. I am sorting the different categories into separate boxes as I go, so that we can go back and look at the various groups in more detail later on. Meanwhile I am weighing everything and recording the dates and descriptions of the more complete, identifiable glass vessel fragments, generally those with rims or bases. There are over 2000 separate accessions so it may take some time!

Some of the bags are ‘bulk accessions’, containing many fragments. The largest of these are collections of vessel glass which was used as cullet, melted in the furnace to be used as raw material for blowing new vessels.

On one large piece you can see individual vessel fragments which have fused together in the heat. We have already sorted the glass into different colours and I am quickly going through the bags again to look for any rim fragments which we may have missed first time around. Most of the glass is in various shades of blue-green, the natural colour caused by the presence of iron oxides in the sand from which it was made originally, but quite a lot is better quality colourless glass. I am weighing all this glass, but life is too short to count every fragment!

I have been very lucky this week to have had some help with some final work on all those sieved glass residues. Monica, one of my colleagues, has done a brilliant and painstaking job of sorting out the smallest fragments into their various colours for which I am very grateful indeed.

John and I are writing a short booklet about the glass project. We have completed the first draft and are now sorting out the illustrations – lots of them. We hope that it will be published in the spring – something to look forward to for the New Year …

One of the important questions which we will try to answer is the scale of the glass production at Basinghall and this means that we will have to weigh the entire collection, and for some types of production waste, count every fragment.

I have started this week by looking at the tank metal, the raw molten glass from which vessels were blown. There is no evidence from London that glass was ever made from the raw materials. Instead, it was either imported from the Mediterranean in large blocks, or alternatively broken vessel and window glass, known as cullet, was collected for recycling. This was common practice throughout the Roman world where there appears to have been a trade in broken glass, rather like modern bottle banks, and the evidence suggests that it was the prime source of the basic material for the London glassblowers.

The glass was melted in a furnace, in rectangular tanks, hence the term ‘tank metal’ and at Basinghall some large fragments have straight edges, where they lapped the side of the tank. It looks as if the glass has been allowed to cool in the tank as some of it shows distinctive crystallisation, and this might mean that it was unused – perhaps thrown away when something went wrong with the furnace, or the workshop went out of use – there are so many possibilities and questions!

Most of the fragments are a natural blue-green in colour, but some of the large fragments are a much darker green. There is actually quite a lot of variation in colour, but some is due to the size of the fragments, so I have had to record their general size as well as the weight of the tank metal from each context (archaeological layer in the ground). Later in the project we will be looking at the chemical composition of some of the chunks of glass to see if they may have come from different sources.

After a day or so, I have recorded over 18.5k of tank metal from 125 contexts – that is a lot of glass!

This week I have finished sorting the residues from a single huge deposit of glass waste, selecting ever smaller fragments. The largest pile is naturally coloured blue-green vessel glass, which was collected for recycling, but I have found many more tiny fragments of glass thread, some less than 1mm in diameter. These were produced at various stages in the glass-working process, sometimes when testing the viscosity of the molten glass.

Strands and threads were also formed when a lump of hot glass, the gather, was removed from the furnace on the end of a blowing iron before it was inflated. Minute fragments fell on to the workshop floor and these are extremely fragile, which suggests that our dump of waste was from the working area itself.

When sorting out some colourless glass fragments I was very pleased to spot some distinctive glass with a fine crackle or crazing on the surface. This might be a faulty vessel, made in the workshop and thrown away, but it is too early to say yet – we shall have to look out for some more.

John and I have been looking at photos of modern glassblowers working in the Roman style with an experimental furnace, to see what sort of waste is produced by the various activities. This is a fascinating exercise and I am learning a lot. We hope to be able to classify our fragments by the production process to gain further insights into Roman glass-working techniques. Next week I shall begin the mammoth task of recording it all in greater detail.

First steps
So my first job is to complete the sorting of the glass residues which have all been sieved to retrieve the tiniest fragments. I have about 3kg of these, which were roughly sorted at assessment, but I am now going through them again, so that we can retrieve the maximum amount of information. The unsorted pile looks quite a jumble, but gradually this falls into more ordered heaps. You can see the different glass colours quite easily.

Roman glass was made by heating silica (sand), soda, which was used as a flux to reduce the melting temperature and calcium (lime) which acted as a stabiliser. Iron oxide, a naturally-occurring impurity in the sand produces a the characteristic blue-green shades of much Roman glass, but various minerals were added to produce coloured glass and also to make the clear colourless glass which became fashionable in the later 1st century AD. Most of the glass at Basinghall is blue-green, but much is colourless and there are a few fragments in dark blue and amber.

Sometimes, sorting these fragments into ever smaller heaps begins to feel endless, and I need a good light. Not a job for these dark evenings - its time to go home …..

Photos: a heap of unsorted Roman glass; Roman glass sorted into the different types of waste

The site

The site at 35 Basinghall Street lies on the western edge of the Walbrook

valley, on the fringes of a marginal area, away from the main focus of residential settlement in Londinium. Before the start of the excavation we already knew that the Walbrook valley was an area of industrial activity, with the site of a major pottery workshop on the eastern side of modern Moorgate. There was also evidence for leather and bone-working and for a 2nd century glass furnace on the edge of a canalised Walbrook tributary, suggesting that this part of the town was occupied by small workshops.

Discovery
When almost 70kg of glass waste was found in the south-west corner of the site at Basinghall we were faced with a colossal task. Usually, at MoLAS, every fragment of Roman glass is accessioned (given a unique number) and recorded on our database, but with tens of thousands of fragments, this was clearly impossible.

After initial recording the first stage in the post-excavation process is always an assessment of the material, looking at its significance in relation to the site, to Roman London as a whole and, as in this case, to the wider Roman world. Clearly we had to establish rapidly exactly what we were dealing with and how much of it there was in order to devise a programme of study and research which would lead to publication.

The scale of the problem
Most of the waste came from a single context, the fill of a pit, so we sorted this, by glass colour, into the different types of waste; raw materials in the form of molten glass from the tank, which was part of the furnace and cullet, broken vessel glass which was collected for recycling and production waste.

This took various forms which I shall discuss in later postings, but the most easily recognisable fragments were molten fragments and runnels from the furnace, which had fallen on to the workshop floor, and moils, little cylinders of glass which were left on the blowing iron after the vessel was removed. Each moil therefore represents a single glass vessel and normally these fragments would have gone back into the melting pot to be blown again.

At this first stage, we accessioned only the best examples of the moils and other types of waste, and all the vessels which could be recognised by form, bagging up most of the moils as ‘bulk’ accessions. Even so, we made out over 2000 record cards. Most glass-working sites in London have produced only about a dozen moils, but at Basinghall we estimate that we have over 3,000.

The questions
There are some fundamental research questions which we are going to try to answer during our work on this project and some of these can only be answered by analysis of the glass composition, a very expensive process. We hope to work out, for example, where the raw materials came from and how they were prepared, what the glass workers were making and the techniques which they used, when the workshop operated and for how long. We shall explore these and many other detailed aspects of the glass-working industry in London over the next few months.

Having devised and costed our programme of research, we presented it to the very supportive developers of the site, Stanhope plc, who have generously agreed to fund it. We are now ready to start the programme and hope that our initial estimates will prove to be accurate!