Kay Hartley Mortarium Archive Project

Kay F. Hartley, Ruth Leary, Yvonne Boutwood, 2022. https://doi.org/10.5284/1090785. How to cite using this DOI

Digital Object Identifiers

Digital Object Identifiers (DOIs) are persistent identifiers which can be used to consistently and accurately reference digital objects and/or content. The DOIs provide a way for the ADS resources to be cited in a similar fashion to traditional scholarly materials. More information on DOIs at the ADS can be found on our help page.

Citing this DOI

The updated Crossref DOI Display guidelines recommend that DOIs should be displayed in the following format:

Sample Citation for this DOI

Kay F. Hartley, Ruth Leary, Yvonne Boutwood (2022) Kay Hartley Mortarium Archive Project [data-set]. York: Archaeology Data Service [distributor] https://doi.org/10.5284/1090785

Data copyright © Kay F. Hartley, Ruth Leary, Yvonne Boutwood unless otherwise stated

This work is licensed under a Creative Commons Attribution 4.0 International License.
Creative Commons License

Historic England logo

Primary contact

Kay Hartley Mortarium Project

Send e-mail enquiry

Resource identifiers

Digital Object Identifiers

Digital Object Identifiers (DOIs) are persistent identifiers which can be used to consistently and accurately reference digital objects and/or content. The DOIs provide a way for the ADS resources to be cited in a similar fashion to traditional scholarly materials. More information on DOIs at the ADS can be found on our help page.

Citing this DOI

The updated Crossref DOI Display guidelines recommend that DOIs should be displayed in the following format:

Sample Citation for this DOI

Kay F. Hartley, Ruth Leary, Yvonne Boutwood (2022) Kay Hartley Mortarium Archive Project [data-set]. York: Archaeology Data Service [distributor] https://doi.org/10.5284/1090785


Figure 1 Rubbing of a stamp of the potter Albinus
Figure 1 Rubbing of a stamp of the potter Albinus

Overview | Table 1 | Table 2 | Table 3

Verulamium, Mancetter-Hartshill and Lincolnshire Industries

Since 1956, Kay Hartley has compiled a large archive covering all aspects of mortarium studies including rubbings of stamps, records of their findspots, fabric etc. The archive includes the rubbings, cards recording data for each stamped sherd, drawings of some stamps and many of the vessels. The initial task of the project was the bulk scanning of the card index cards which record the details of each stamp Kay has seen over the decades of research. These comprise three collections – the Romano-British potters' stamps, stamps of Continental potters found in Britain, and many found on the Continent, and a summary of the stamps by site.

Kay has taken rubbings of all the stamps she has handled and these number in the thousands. To test the feasibility of the project and the proposed methodology, the team decided to concentrate first on the Verulamium region. The project staff members, Kay Hartley and Yvonne Boutwood, started selecting the best rubbings to illustrate for each die of each potter (fig. 1). These were then scanned and information relating to them on the potters' card index was then entered into a database.

The potter/stamp distribution dataset

The card index archive was compiled over some 56 years and during this time approaches to pottery studies have altered. The cards record details of the stamp identification, the findspot and excavation details, any known publication references, the fabric description, the position of the stamp on the vessel, the presence of scoring on the vessels, concentric disposition of the trituration grits, the diameter of the vessel, the estimated vessel equivalence, the Museum, any other comments- notably burning, evidence of kiln wastage, associated counterstamp and whether a drawing was made of the vessel. Each recorded stamp was usually given a K number. Combined with the place and museum, the excavation or site, this gave each stamp a unique identified. In practice, some of the stamps entered lacked the K number and in some cases K numbers were duplicated. The card data except the fabric description details was typed into the stamp distribution dataset to form the basis of the distribution maps by potter.

Kay recorded the details of the fabrics of each stamp she handled on the relevant card but these do not readily correlate with the National Roman Fabric Reference collection fabrics used today (Tomber and Dore 1998) nor with the regional fabric series now in existence. The fabric descriptions occupy a significant part of the data on the cards but are omitted in the stamp distribution dataset. The fabric descriptions are available in the scanned index cards for anyone who wants to consult them but, in the dataset, they have been replaced by ware groups, NRFRC fabric codes, where these exist, or source attributions by Kay. This part of the dataset is in need of further work but this could not form part of the present project, the aim of which is to make the present, historic archive available rather than improve it.

Similarly, some attributes recorded on the card indices are not consistently recorded such as vessel diameter, estimated vessel equivalent, die number and die reading. These were therefore omitted from the dataset presented here but available in the downloadable scans of the record cards.

The stamp distribution dataset presented here records the potter identification. This is usually a name but could be a numbered trademark, herringbone or incomplete. If recorded on the cards or one of the selected rubbings, the die number will be noted but these were not always recorded on the index cards and time did not permit them to be added at this juncture. The Museum and/or excavation details are recorded, usually with an accession number and/or context details. The presence/absence of concentric trituration grits and scoring are recorded, the position of the stamp, which way it is facing and further comments by Kay are recorded. Bibliographic details are given if possible and the latitude and longitude have been added using commercial geocoding tools. These are based on placenames rather than OS grid references so are not as accurate as one might like but are sufficient for regional distribution patterns. The source and fabric codes were added by the team during the compilation of the dataset and are not always present on the archive cards. In some cases, these are uncertain and need further study but such study could not form part of this project’s brief. On the record cards there is often full fabric descriptions by Kay which can be consulted.

The card index data was not completely up to date and does not include stamps Kay has not seen and made rubbings of. It is anticipated that colleagues will add these and other new stamps to the datasets periodically as new material is identified.

The stamp and die dataset

To make up for these omissions, a second dataset (Potter record) was compiled from scratch which lists the attributes of every stamp/die combination giving the reading, letter variants and symbols used, the source and the fabric(s) used, the border classification details and other notes considered significant by Kay. This dataset was compiled by Yvonne Boutwood (borders) and Ruth Leary (drawing on data recorded by Kay and Yvonne during the selection of the rubbings) and Kay then edited and corrected this dataset. This is searchable and forms the basis of the mortarium stamp query facility which can be used by colleagues to identify a stamp they have excavated or, at least, narrow down the possible matches. Roger Tomlin was asked to give readings based on the rubbings and make any comments on the onomastics and these are included in this dataset.

In this dataset, the Regions follow those used in the Mortarium Bibliography for Roman Britain, the fabrics follow the National Roman Fabric Reference collection fabrics collection fabrics (Tomber and Dore 1998 with the addition that MAH WH1 indicates the earlier fabric with mixed trituration grits and MAH WH2 the later fabric with argillaceous trituration grits) and the industries include regional groups such as the Lincolnshire potteries and large inter-related industries such as that around St Albans and the related industry in London (VRW and VRW London respectively) and at Mancetter-Hartshill (MH)

Many of the record cards for stamped sherds were compiled before the publication of Kay’s reports so limited efforts were made to track down the publications. This has been only partially successful and some reports are likely to remain unpublished or have not yet been identified. The bibliographic references were added to the stamp distribution dataset and the bibliography is available.

Stamp and die characteristics

The stamp and die characteristics are grouped by potter/stamp and die. The overall date range for the potter is given and for each die a more precise date range is given if there is evidence for such dating. For each die a reading is given and it is possible to search for elements of that reading should a researcher be seeking matches for a partial stamp fragment. Roger Tomlin studied all the rubbing scans and photographs and provided readings and comments on the onomastics. Sometimes the reading is not certain and possible alternatives are also given. Letters in the stamp can sometimes be impressed retrograde or ligatured and these are listed and any special letter or motifs included in the stamps are also given such as differences in letter forms or leaf motifs. Two-line stamps are also noted and these should be distinguished from cases where a single stamp has been used stamped close together giving the impression of a two-line stamp. Stamps which have occurred with the stamp are listed. These are usually counterstamps such as FECIT stamps but occasionally the stamps of other potters' names occur e.g. The Verulamium regions potters Gissus and Lallaius. This dataset also gives the industry or industries with which the potter and/or die is associated and the fabrics used. All these characteristics can help identify the potter and die of a stamped sherd.

Border classification

Y. Boutwood

As well as ensuring the preservation of the archive and its availability for other scholars to consult, the team were aware that future researchers would want to interrogate the data to find matches for their own mortarium stamps. Intuitively one first attempts to read the letters and decipher the potter's name, but other design elements such as border patterns, within both name and counterstamp impressions, can be quite distinctive and aid identification. So, another aspect of the project developed a classification of borders, evaluated from the selected rubbings for each potter's die and devised a simple recording system. The latter separately records borders occurring at the top and bottom, left and right sides, surrounding the name or letter panel. Borders in these four areas may differ within one stamp and this methodology allows a flexibility in retrieval, particularly if trying to identify a partial stamp.

  • 1.1 Roman Mortaria stamps and counterstamps have border patterns incorporated into their design within the stamp impression.
  • 1.2 This classification initially described the main classes for potters associated with Verulamium Region industries (VRW) and is subsequently broadened to include Mancetter-Hartshill (M-H), Alcester and Lincolnshire (Table 1). Examples are illustrated using photographs of stamp rubbings from Kay Hartley's archive (Table 2). A comparison of borders occurring across the four evaluated groups are summarised (Table 3).
  • 1.3 Linear borders occur at the top and bottom, left and right sides surrounding the area containing the stamp, or counterstamp letters (name or letter panel). Patterns on these four areas may differ within one stamp and can be classified separately.
  • 1.4 The left and right borders are often poorly impressed, particularly when coinciding with the rim bead or flange end. When incompletely impressed ‘Not Known' (NK) is recorded. When left and right borders are clearly impressed they may reveal a pattern that can be classified.
  • 1.5 The name or letter panel is commonly divided from the surrounding border pattern by a frame line, but when no frame line occurs, which is rarer, it is specifically recorded as Frame-None (Class 2.1).
  • 1.6 Decorative motifs (e.g. leaf) within the name or letter panel are not part of this classification of borders, but linear patterns within it, parallel to the frame line, are included (e.g. Class 11.0, row of dots).
  • 1.7 Patterns are divided into broad Classes (e.g. Class 4.0) and significant variations form subdivisions (e.g. Classes 4.1 to 4.7).
  • 1.8 More than one pattern can occur together within a single border, forming a composite pattern (Class 12.0).

Stamp identification query

The potters' stamp and die dataset is searchable using the query form. Start with evaluating the elements that you are most confident and certain of, so the order may vary, be that a two-line stamp, border patterns, reading letters within the name stamp, or determining the industry/fabric. Remember you may be holding the stamp upside down. Look to see exactly what letters or other motifs are present. You may find ligatured or retrograde letters, even if the name as a whole is not retrograde and this may take you to one or two possibility only. Reading through the free text description of unusually formed letters may also aid you narrow your search within the results returned by the search query. The search query will return the possible matches and you can then compare your stamp with the rubbings displayed.

If you find no matches this may be a new die or a new part of a die. For example, part of the border hitherto unknown may be on your sherd. For example, for SEPTVMINVS Die2, we only have borders surviving to top and bottom- Herringbone-Linked and Chevron respectively. If you found this stamp with, say, bar-diagonal to the left or right, this die would not be selected. You would need to try not selecting for the right and left borders and looking at the list returned to see if you have found a new element of a known die. Before assuming you have a new die, try just filling in some, but not all, of the attributes you can see on your sherd and check it does not match one of the stamps listed.

Until data for all the industries are uploaded, your stamp may, of course, not be present.

The scanned rubbings

Kay and Yvonne selected the rubbings which best illustrated each die. In some cases, dies had only one rubbing but in most, there were many to choose from. In rare cases, rubbings of stamps from the Continent were used as these were the most complete examples. The team chose the best rubbings to represent the whole stamp and these were scanned individually and laid out in a group to show how they related to one another. Both the individual and group scans can be viewed and downloaded from the website.

Printing the rubbings

The scanned rubbings can be downloaded and printed at 1:1. To achieve this, after downloading, insert the tiff or jpeg into a Word document, right click on the file, select size and position then click on reset. This ensures the rubbing will print at 1:1 enabling you to compare a rubbing you have of a stamp directly with a possible match.


The initial task was to do a trial on the data for the Verulamium industry. While further funding was being sought, the team added data for the Mancetter-Hartshill industries and the Lincolnshire industries and together these are being made available online in the first stage of the project. All the record cards have been scanned and are available on the website. Rubbings for all the potters working in the three industries have been selected and scanned and the data relating to the characteristic of each die and the distribution of stamps of each potter have been Historic England have agreed, in principle, to fund further work to add the remaining data for potters working in England.


Ansatefrom the Latin ansa (handle) is used to denote stamp-labels which have projecting triangular or balloon-like ends to their frames.
Counterstampstamp used in conjunction with another stamp either side of the spout. Often these say FECIT for made
Dietool used to impress the potter’s stamp on unfired mortarium. Evidence for these being made of clay or wood.
Die 1an individual stamp die used by a potter which can be distinguished from other stamps used by the same potter
Die 1aa counterstamp associated with die 1 of a potter
Incusestamped sherd where the letters are sunken due to use of a die with raised letters
Ligaturedtwo or more letter conjoined, often to save space
Manuby the hand of, usually followed by the potter’s name
Retrofor retrograde
Retrogradename has been written on the stamp die the right way round so the stamp comes out backwards
Surmoulagedie made by taking a moulding of a stamp impressed on a fired pot
]stamp incomplete at the beginning
]stamp incomplete at the end
( )illegible letter (s)

Select abbreviations found on the cards

AA or Arch Ael:Archaeologia Aeliana
AJAntiquaries Journal
BMBritish Museum
B of EBank of England
CILCorpus Inscriptionum Latinarum
CWTransactions of the Cumberland & Westmorland Architectural and Archaeological Society
GMGuildhall Museum
JBAAJournal of the British Archaeological Association
JRSJournal of Roman Studies, London
HUNT MHunterian Museum
NMANational Museum of Antiquities, Scotland
NMWNational Museum of Wales
NUMNottingham University Museum
PSALProceedings of the Society of Antiquaries of London
PSASProceedings of the Society of Antiquaries of Scotland
RIBRoman Inscriptions of Britain
SACSussex Archaeological Collections
YMYorkshire Museum York
VCHVictoria County History


The team received help in the form of advice, supportive references, as well as grants for which we would like to record our thanks and appreciation. The work was funded by grants from the Haverfield Bequest, The Roman Research Trust, Historic England and the Study Group for Romano-British pottery. Prof. Mike Fulford, on behalf of the Haverfield Bequest and, later the Roman Research Trust, was particularly supportive and helpful from the beginning of the project, giving sound and timely advice and encouragement. Colleagues in the Study Group for Romano-British pottery have been supportive both in awarding grants towards initial purchase of equipment and the archive scans and also in their unstinting support of the project. Rob Perrin and Hilary Cool kindly supported our application to the Roman Research Trust for funding and generously helped with the application process. The Committee of the Study Group for Romano-British Pottery have maintained an interest throughout the project and given support and encouragement to the team at key points during the project. Duncan Brown from Historic England gave generous help through the process of applying for funding.

ADS logo
Data Org logo
University of York logo