Database description

krosno ciężarkowe, rys. M. Jagodzińska, pieczęć z grobu tolosowego A w Aghia Triada, przerys M. Jagodzińska za CMS II.1. 64 a


The ‘Textiles and Seals’ project team would like to express their particular thanks to Prof. Diamantis Panagiotopoulos and the CMS Archive in Heidelberg for generous consent for the use of data from the CMS casts, including their photographs made by Agata Ulanowska, as well as the CMS drawings and photos in the T&S database. We also wish to express our thanks to Dr Maria Anastasiadou, who kindly agreed to the use of drawings from her 2011 monograph in the Iconography of textile production module, as well as to Prof. Pietro Maria Militello, who kindly agreed to the use of his drawings and photographs of tools from Agia Triada and Phaistos in the Seal impressed and marked textile tools module.

Data in the module Iconography of textile production

Courtesy of the CMS Archive in Heidelberg, the majority of seals and sealings collected in the Iconography of textile production module come from the old version of the CMS Arachne database. The CMS Arachne is the largest and most important database of Aegean seals, which provides options for complex and multiple queries on seal iconography (more about the CMS Arachne can be found here). However, due to a different classification method of motifs that have been recognised as textile production-related (see below), it cannot be used in a straightforward way for the searches posited by the ‘Textiles and Seals’ project. The CMS seals are referred in the ‘Object references’ field as CMS with number.
The next source of large number of data comprises Maria Anastasiadou’s monograph The Middle Minoan Three-Sided Soft Stone Prism: A Study of Style and Iconography, published in 2011 as CMS Beiheft 9. If they are not in the CMS Arachne, the seals from Anastasiadou 2011 are referred in the field ‘Object references’ as ‘A’ with a number corresponding to the catalogue number in her book. The abbreviation ‘CHIC #’ with a number denotes data from the Corpus Hieroglyphicarum Inscriptionum Cretae authored by Jean-Pierre Olivier and Louis Godart (Études crétoises 32, 1996). The annotations ‘P.TSK’ and ‘PTSK’ with numbers denote seals from Petras discussed in 2012 and 2017 by Olga Krzyszkowska in two papers (‘P.TSK’–‘Seals from the Petras cemetery: a preliminary overview’, in: M. Tsipopoulou (ed.) 2012 Petras, Siteia – 25 Years of Excavations and Studies. MoDIA 16, and ‘PTSK’ – ‘Further seals from the cemetery at Petras’, in: M. Tsipopoulou (ed.) 2017 Petras, Siteia. The Pre- and Proto-Palatial Cemetery in Context. MoDIA 21). HMs is a reference to the inventory number of the Heraklion Archaeological Museum, here used to denote two seal impressions from Kato Syme: HMs 1106 and 1107, discussed by Ingo Pini in his 2006 article ‘Die Siegel und Siegelabdrücke aus Gefässhenkeln aus dem Heiligtum von Symi’ from Athenische Mitteilungen 121.
The data collected in the Iconography of textile production module of the T&S database do not comprise all possible references to textile production on Aegean seals.

Identification of textile production-related motifs – methodology

Textile production-related motifs have been recognised on the basis of a number of factors. These include textile motifs that have already been identified on Aegean seals, the characteristics of Bronze Age textile production together with experimental and experience archaeology (specifically hands-on experience in spinning and weaving). Proposed identifications have been derived, at first, from a general visual resemblance of a motif to actual fibrous plants, woolly animals, textile tools and tasks. Such identifications were then corroborated by examining whether specific features defining real-world references to textile production could be detected in the motifs. The defining features are the unique physical characteristics of plants and animals (e.g., shape of a plant stem and crown, characteristic shape of horns and tails in woolly animals), features of functional importance (e.g., a heddling loom mechanism) and technical gestures of textile workers (e.g., hand gestures in spinning, body posture in weaving). Next, the new identifications were cross-checked with previous ones or terms traditionally used to denote discussed motifs. To test the validity of the proposed interpretations, iconographic comparanda in other arts and cultures were sought, especially small-scale depictions, e.g., on Mesopotamian seals. The recognised ‘textile motifs’ are classified as follows:

  • potential references to raw materials, represented by: ‘Flax’, ‘Woolly animals’: with ‘Sheep’, ‘Goat’ and ‘Lamb’ motifs, animals’ ‘Heads in profile’, ‘Protomes’ and animals in combinations with human figures, ‘Silk moths’;
  • potential references to processing of fibres and spinning: ‘Combs’ and ‘Combers’ (also a potential reference to weaving), ‘Spindle with whorl’ and ‘Spinners’ motifs;
  • potential references to dyestuffs and dyeing: ‘Murex shell’ motif;
  • potential references to weaving and woven fabrics: ‘Loom weights’, ‘Weavers’, ‘Human figures with ‘weft beaters’, ‘Warp-weighted loom’, ‘Rigid heddle’, ‘Weft-beater’, ‘Fabric with fringes’ and ‘Interlaced band’ motifs;
  • potential symbolic references to textile production: ‘Spiders’.

Motifs such as ‘Flax’, ‘Head in profile of ‘woolly’ or possibly ‘woolly animal’, ‘Spindle with whorl’, ‘Rigid heddle’ and ‘Fabric with fringes’ have also been recognised in the graphic forms of some signs of the Cretan Hieroglyphic script.
As a note of caution, it should be stated that the new identifications cannot be seen as definite; several remain tentative or do not fully adhere to the applied principles. Moreover, several previous identifications are, graphically and logically, no less reliable than those newly proposed.

Data in the module Impressions of textiles on clay

Courtesy of the CMS Archive in Heidelberg, textile impressions published in the Impressions of textiles on clay module come from the modern casts stored in the CMS Archive that were taken from a selected number of the undersides of clay sealings by the CMS team. 

‘Textile impressions’ is a broad terminological umbrella that covers imprints of actual textiles, i.e. fine threads, cords and fabrics, as well as mats, baskets, wickerwork containers, thongs, as well as leather and skin pieces. The recorded evidence can also be termed ‘technical textiles’, due to their function in administrative and sealing practices in the Aegean during the Bronze Age. Technical textiles were used to wrap and tie objects, such as doors, pegs or knobs of the sides of chests, openings of jars, wickerwork containers, packets of folded skin or parchment, etc., prior to their sealing.

So far, the collected data come from Aghia Triada, Chania, Knossos, Malia and Phaistos on Crete, Lerna and Geraki on the Greek Mainland, and the islands of Kea (A. Irini) and Lemnos (Myrina). Fully recorded are the data from Lerna (52 casts), Geraki (34 casts) and Phaistos (145 casts), while the comprehensive study of textile impressions on all casts in the CMS Archive in Heidelberg is to be continued. Textile imprints from the undersides of clay sealings that were not recorded in the CMS are not included in the T&S database. 

In the CMS volumes and in the CMS Arachne database, the examined casts are referred to according to the inventory number of the clay sealing, preceded by an abbreviation of the museum name or an abbreviation of the archaeological site. For example, the sealings stored in the Heraklion Archaeological Museum are recorded with HMs annotation followed by a specific number, while the sealings from Lerna, stored in the Archaeological Museum of Argos, are recorded with a letter L and a dot. In the T&S database, we consider the modern casts stored in the CMS Archive as different objects from the clay sealings from which they were taken. Therefore, we refer to them by using exclusively the abbreviations of the names of the sites or the islands the sites are located on, followed by inventory numbers of clay sealings according to the scheme presented in Table 1.


Agia Triada AT
Akrotiri, Thera AKR
Chamaizi CHAM
Chania CH
Geraki GER
Seraglio, Kos KOS
Agia Irini, Kea KEA
Kato Zakros KATZ
Knossos KN
Skarkos, Ios IOS
Myrina, Lemnos LEM
Lerna LER
Malia MAL
Palaikastro PAL
Petras PET
Phaistos PH
Table 1. Abbreviations for find places in ‘Object reference‘ and ‘Inventory number‘ fields.

Three modern casts from the CMS Archive, with known place of origin but without inventory numbers, are recorded in the T&S database as: PH No no. 1, PH No no. 2 and LER No no. respectively. Two casts from Phaistos with the same inventory number are annotated in the CMS Archive as PH 676 and HMs 676. To differentiate them in the T&S database, the first one is recorded as PH 676, while the second one, associated with a specific seal impression on the front side, is PH HMs 676.

Identification of textile imprints on the modern casts of the undersides of clay sealings – methodology

Textile imprints on clay are a useful source for analysing the properties of actual spun fibres, textiles and other organic products. In textile archaeology, these properties are examined by recording the diameter of threads and cords, their twist direction and angle, to see their structure, e.g. if they were single spun or twisted, or spliced, or plied of two or more elements. In the case of textiles, the weaves or stitches and the number of threads (warps and wefts) per cm (t/p/cm) are recorded. However, the quantity and quality of information that can be retrieved from a specific impression varies due to its state of preservation, clearness, size and the properties of both textile and clay fabrics, as well as due to the adopted methodology.

Products made of raw materials that are hard and feature low elasticity, such as plant fibres, leave more legible impressions than, e.g., soft and elastic wool. Production technique matters too, since more three-dimensional objects, such as e.g., twined textiles, leave deeper and clearer impressions. These factors affect measurements taken from the imprints, which might have been further altered due to the shrinkage of wet clay and deformations in firing process and in the sealing practices, e.g. when clay sealings were removed while clay was still damp. According to Agata Ulanowska’s experiments, the measured diameters and twist angles of the impressed yarns and cords can be smaller and larger in comparison to those of actual textiles. Yet, the differences are not substantial, resulting in a maximum 0.37 mm difference between the diameter of yarn measured from a tabby textile (1.12 mm) and its negative imprint (1.49 mm). Since the CMS casts have been taken from the negative clay impressions, they document the positive copies of the impressed products and the twist direction is not altered on the casts and their photographs published in the T&S database.

Textile impressions from the CMS casts have been analysed using a Dino-Lite AM4515ZT digital microscope with magnification from c. x16 to x74.5 and additional side lighting. Additionally to the measurements, microscopic analysis has revealed a range of various visual appearances of microstructure of the impressed products. In the ‘Textiles and Seals’ project and the T&S database, imprints are classified into eight material groups, reflecting characteristics and diagnostic features of their microstructures, such as smoothness, section or an ending of the impressed fibres and skins. The classification of the material groups is as following:

  • Long strands of fibres, characterised by tightly aligned, smooth surfaces with long, narrow or broader strips of fibres with relatively high relief in the microstructure. Ending and use-wear representative for this group are preserved, e.g., on LER 4.458. They show single straight strands of fibres in the ending and cracks on the surface. Diagnostic features suggest plants with fibrous stems (if their procurement resulted in strips of fibres), bulrush, and tree bast (e.g. limebast), as well as guts (e.g. pig guts) as potential raw materials.
  • Broad strands of fibres, characterised by broad, band-like strands of fibrous microstructure. Possible use-wear and an ending can be observed on PH 721, use-wear – on PH 789. Diagnostic features suggest entire stems of plants, such as bulrush, possibly tree bast and palm-like leaves.
  • Long individual fibres, characterised by a smooth or coarse surface with long twisted or loosely twisted fibres seen individually in the microstructure. No endings and use-wear are preserved on the analysed CMS casts from this group. Diagnostic features suggest fibres from variously procured plants with fibrous stems, such as flax, nettle, hemp and, potentially, sinews.
  • Short individual fibres, characterised by loosely aligned and loosely twisted individual fibres with single fibres of various diameters protruding in the microstructure, as on PH 920 and, possibly, PH 693. Diagnostic features suggest wool and goat hair or fibres from plants with fibrous stems, such as, for example, nettle obtained from the dried stem.
  • Wicker-like, characterised by stick-like, straight or slightly bent elements of smooth, wood-like microstructure. A possible ending is preserved on PH 800. Impressions of wickerwork techniques provide additional support for the proposed identification. Diagnostic features suggest plants, such as willows, reeds and rushes.
  • Leather thong-like, characterised by a smooth or porous, grained microstructure and band-like appearance. Occasionally, both rectangular (MAL 1072) and oval sections (PH 720) can be recognised, as well as sharp edges of folds, as observed on KN 425. No endings are preserved on the CMS casts examined so far, but possible use-wear, characterised by cracks on the surface, can perhaps be visible, e.g. on PH 845 α’P. No attribution to a specific animal is possible.
  • Leather piece-like, characterised by irregular, porous and grainy microstructure with occasional long vein-like thickenings, e.g. as on CH 1559. Fine leather-like pieces of smooth or slightly grained surface have already been identified as parchment (for the previous research, please see the Outreach page).
  • Fur-like, characterised by short, overlying and irregularly placed fibres, covering the entire or a large part of the impressed surface, e.g. as on KN 1180. 


  • Non recognisable, a category which comprises all examples with no clear diagnostic features.

Data in the module Seal-impressed and marked textile tools

All textile tools presented in the Seal-impressed and marked textile tools module have been published. The collected tools provide the first overview of seal-impressing and marking practices on textile tools and do not comprise all seal-impressed and marked tools from Bronze Age Greece. 

Courtesy of the CMS Archive in Heidelberg, the majority of seal-impressed textile tools in this module (48 out of 74 records) come from the CMS publications and the CMS Arachne database. The remaining seal-impressed and all marked tools come from a range of publications or mentions that are referenced in exported Excel files, in column E=‘Arachne link or bibliographic reference’. Except for the CMS, the largest number of data come from works of J. Weingarten (17 seal-impressed loom weights from Palaikastro): chapters in BSA 86 (1991) and Πεπραγμένα Η′ Διεθνούς Κρητολογικού Συνεδρίου (2000); M. Tsipopoulou’s contributions on potter’s marks from Petras (altogether 15 tools): in Kadmos (1990) and in her 2016 monograph Petras, Siteia I; and the study by J.-P. Olivier and L. Godart (1978) Fouilles exécutées à Mallia. Le Quartier Mu I. Écriture hiéroglyphique crétoise, Études Crétoises 23 (13 tools). Fifteen incised textile tools from Agia Irini, Kea have been mentioned in †J.E. Cutler’s 2021 monograph Crafting Minoanisation, Ancient Textile Series 33. 

The data collected so far come from Agia Irini on Kea, Akrotiri on Thera, Seraglio on Kos, Skarkos on Ios, Chamaizi, Chamalevri, Kato Zakros, Knossos, Malia, Palaikastro, Petras, Phaistos, Sitia on Crete and Lerna on the Mainland. Each find place is reflected in the object reference by a site abbreviation (for the list of abbreviations, see Table 1) or its full name. The object reference number also comprises the inventory number of a museum or excavation’s storeroom, or a catalogue number from the publication if the first two numbers are not available. Tools that are stored in the museums with no inventory numbers are marked as ‘No no.’ A reference to the seal-impression is annotated in square brackets as the CMS number. For example, the object reference: MAL HMp 17176 [CMS II,6 175] informs the reader that the recorded tool comes from Malia on Crete, it is stored in the Heraklion Archaeological Museum (HMp) under the inventory no. 17176, while the seal-impression on this tool is published in CMS II,6 under the number 175.

Parameters of textile tools and remarks on their measurements and weights

In the T&S database, textile tools are classified as ‘Spindle whorls’, ‘Loom weights’ and ‘Spool-like’ implements. Due to their potential multifunctionality, spool-like tools are considered a separate class of tools, however, they could have also been used as loom weights. Geometric forms of tools are defined as ‘Conical‘, ‘Cuboid’, ‘Cylindrical’, ‘Discoidal’, ‘Irregular’, ‘Pyramidal truncated’, ‘Semi-spherical’, ‘Spherical’ and ‘Spool-like’, regardless of the tool function. The choice of forms corresponds to the collected tools; more forms can be added with new tools.

In textile archaeology, textile tools are documented by recording their dimensions: diameter, thickness, height and width, as well as weight and use-wear. The main functional parameters, i.e. the parameters that have an impact on a quality of product made using a particular tool, are: weightdiameter and height in case of the spindle whorls and weight and thickness in case of the loom weights and spoolsUse-wear marks attest that the tool was in use, they may also indicate a specific technique of use, e.g. a specific manner in which a loom weight was suspended in order to tension the warp threads.

In older publications, however, information about weight of tools and their potential use-wear has often been missing. In the T&S database, all missing data are displayed as: ‘No data’ in short description of records. Since the collected tools have not been examined in person by the ‘Textiles and Seals’ project team, ‘No data’ is also applied to describe a preservation percentage, unless there is a remark about tools’ preservation in the publication. No attempt is made to estimate the preservation percentage or use-wear on the basis of photos.

The understanding of what is diameter, thickness, height and width may not be straightforward for some tool forms. Figure 1 shows how the specific tool dimensions are understood in the T&S database.

Figure 1. Dimensions of tools in the T&S database

In case of pyramidal truncated loom weights, we assume that thickness is represented by the side of the tool without a perforation, since this would be the side by which the pyramidal loom weights were touching each other when aligned in a row on a warp-weighted loom. When there is no differentiation between the width and the thickness in a publication, we consider the published ‘width’ as our ‘thickness’, assuming that the available data is the only one that reflects a functional parameter of the tool. Cuboid weights of dimensions published in an abbreviated way, e.g.: 61 x 58 x 50 mm, are displayed in the Textile tools module as ‘No data’, since we do not know which dimensions have been represented by individual values. 

Cross-referencing of records in all modules of the T&S database

The data which may appear in all modules of the T&S database and, therefore, can be crossed-referenced, are seal faces with textile motifs that have been impressed on textile tools or on the front sides of clay sealings bearing textile impressions on their undersides.

In the search engine of the module Iconography of textile production, cross-referenced records can be found by choosing ‘Yes’ in the fields ‘Sealings with textile impressions on the underside’ and ‘Tools with textile impressions’.

In the search engine of the module Impressions of textiles on clay, cross-referenced records are marked as active links to the respective CMS numbers – each link is re-directing the user to a record with a textile motif in the module Iconography of textile production. Alternatively, cross-referenced records can be found by choosing ‘Yes’ in the field ‘Sealings with textile motifs’.

In the search engine of the module Seal-impressed and marked textile tools, cross referenced records are displayed in a description under the ‘Seals’ label, as links to the respective CMS numbers – each link is re-directing the user to a record with a textile motif in the module Iconography of textile production. Alternatively, cross-referenced records can be found by choosing ‘Yes’ in the field ‘Impressed with textile motif’.

For more detailed information, see the Outreach page.
Author: A. Ulanowska