Table of Contents
- 1 Introduction
- 2 PAS object type to be used
- 3 Terms to use in the description
- 4 Roman buckles
- 4.1 Buckles of the first to third centuries AD
- 4.2 Late Roman buckles
- 5 Early Anglo-Saxon buckles
- 6 Middle Anglo-Saxon buckles
- 7 Late early-medieval buckles
- 8 11th- and 12th-century buckles
- 9 Medieval buckles
- 9.1 Sources
- 9.2 Buckle or brooch?
- 9.3 12th-century buckles
- 9.4 Buckles which begin in the 13th century
- 9.5 Buckles which begin in the 14th century
- 9.6 Buckles which begin in the 15th century
- 10 Spur buckles and other medieval buckles with specialist uses
- 11 Post-medieval buckles
- 12 Modern horse-harness buckles
- 13 Key references
A buckle is used for fastening two straps together (or two ends of a single strap, such as a belt). It consists of a frame through which the strap passes, and a pin which is pushed through a hole in the strap. A buckle frame may be directly fixed to the strap, or the buckle may have a plate attaching the frame to the strap.
N.B. There is no standard way round to photograph buckles. The strap may come in from the left, right or (occasionally) the bottom of the photo, depending on the form and decoration of the buckle. A mix of styles will be used in the photographs contained in this guide.
PAS object type to be used
Use BUCKLE for all parts of a buckle, even for a detached plate or pin. A buckle pin on its own isn’t a PIN, it is part of a buckle. Related to buckles are strap clasps, which are recorded using the term CLASP. Strap clasps also fasten straps together using frames, but do not use pins passing through holes in the strap.
PAS object classifications and sub-classifications to be used
Don’t use the classification field to flag up the elements of the buckle that are present – frame, plate or pin – as this field is used instead for certain types of buckle. See below under the period-specific sections for which types of buckle should be flagged up in the classification field. Here is a summary:
- Hawkes and Dunning type
- gaping mouth beast
- standing animal
- Limoges style
- M shaped
- hollow plate
- T shaped
- Meols type
- Cassels type
Terms to use in the description
In general, Egan and Pritchard’s conventions (1991, 51) are followed for buckles of all dates, with a few exceptions. This part of the guide will use medieval examples as these are the most common, but should be followed for buckles of all periods. So please don’t use non-standard terms such as spindle or chape.
The parts of a buckle are the frame, the pin and the plate. All buckles have a frame and a pin, but not all have a plate. The plate is normally attached to the strap with rivets.
A buckle frame can be single-loop or double-loop. The ‘loop’ is the space enclosed within the buckle frame; the frame is made up of a bar, an outside edge and perhaps – if useful in writing your description – a top and bottom edge.
A buckle frame will be assumed to be single-loop unless stated otherwise. So ‘oval buckle frame with offset bar…’ will be sufficient, rather than ‘oval single-loop buckle with offset bar…’.
The word bar refers to the area of the buckle frame, usually straight, where the strap or plate is attached and the pin is hinged. Don’t use ‘strap bar’, because it might have had a plate attached rather than a strap. Don’t use ‘spindle’, even for a separate iron bar, because we use this word for a spinning tool.
The rest of the frame can be divided into various parts; the top edge, bottom edge and outside edge.
The term outside edge is useful for that part of the buckle frame opposite the bar. The term ‘front’ should not be used for this part of the frame, because of the potential for confusion with ‘front’ and ‘back’ used in the sense of obverse and reverse. The term inside edge is used in Egan and Pritchard 1991 for the other end (by the strap) of a double-looped buckle frame, but in the absence of a strap, plate and/or pin it is usually impossible to tell which edge is which. For this reason, it is usually easier to consider a double-loop buckle frame as having two outside edges.
Where Egan and Pritchard need to talk about the edge of the buckle frame in its general sense, they use the word perimeter. The parts of the loop that are not the outside edge are called by Egan and Pritchard the ‘sides’, but we tend to use top and bottom edges for consistency with the outside edge. ‘Top’ and ‘bottom’ are relative terms; although the top and bottom edges are usually much the same, if they are not, you have to explain which way up you are holding the buckle.
The outside edge can have a central constriction for a separate sheet roller, or occasionally rollers can be carried on separate bars.
Use pin constriction for any narrowing on the bar where the pin is hinged. The wear of a plate’s two hinge loops can cause two narrowings on the bar, leaving a bump in the centre instead of a constriction. Alternatively, occasionally the bar was actually deliberately bent up in the centre where the pin was hinged, presumably to retain the plate or strap in position to either side.
Use pin rest for the area in the centre of the outside edge, opposite where the pin is hinged. The pin rest can project, or be pointed, with a point just in the centre; this is different to having a pointed outside edge, where the whole of the edge expands to the point (see illustration below). Don’t use the term ‘lipped’ as it is hard to understand. The pin rest can also be grooved. A frame with a pointed outside edge can be oval or rectangular internally.
Frame shapes for single-loop buckles
Here are six common frame shapes for single-loop buckle frames, and words you can use for them. Oval shades through to semi-circular and D-shaped and it can be difficult to know which word to use. On the whole, D-shaped should be used for those buckle frames with sharply right-angled corners, and oval for those with more rounded corners.
Shapes can be different internally and externally. The two buckles below are good examples; internally they are rectangular and oval, but their external shapes are quite complex and need careful description.
The photos below show some different shapes of loop, or outside edge, and bar, but all variants of the same basic oval or D-shaped frame.
The bar can be offset or narrowed, or both; it can have lobes which project at the ends. When the bar is offset it can be particularly hard to decide whether the frame is oval or D-shaped, but don’t stress too much about this, just opt for one or the other.
The outside edge can be what Egan and Pritchard call ‘ornate’; this means with a decorated moulding, but this can vary enormously in shape and decoration, so needs a full description. Otherwise an outside edge might be widened or thickened, or angled, to make a larger area for decoration.
It is all right to use jargon or shorthand terms such as ‘ornate’ or ‘kidney-shaped’, but you must also describe the shape precisely (e.g. ‘kidney-shaped, with an incurved outside edge’).
The two loops of double-loop buckles are divided by a bar. The loops may be identical, or may be different, leading to an asymmetric shape.
Egan and Pritchard 1991 contrasts ‘buckle frames with central bar’ and ‘double-loop buckle frames’ but there is no functional difference; both have two loops. The usual choice is ‘double-loop buckle frame’ unless it is easier to describe the particular frame as a whole and the bar separately. For example, an 18th-century shoe buckle seems quite clearly to have a frame and central bar, rather than two loops.
The word pin is preferable to ‘tongue’, as we quite often need to talk about brooch or buckle pins, and ‘brooch pin or buckle tongue’ would be unwieldy.
The parts of the pin are the loop and the shaft. They can be cast, or made from sheet (flat in cross-section) or wire (normally rounded in cross-section). It is useful to tell us how the pin was made, if you can work it out.
Cast pins often have a moulding (such as one or more ridges, or a raised rectangular panel) at the junction of the shaft and the loop. The word ‘moulding’ is imprecise, and it really helps to have a proper description.
Cast pins can have a closed or open loop. Wire or sheet-metal pins are bent around the frame leaving an open loop.
When a pin is made from sheet or wire, and the frame is cast, a different composition of metal can be used. More lead might be appropriate in an alloy for casting, and more copper in an alloy for cold working, hammering and bending. A mismatch between frame and pin might alternatively be due to a replacement pin, but it is hard to distinguish between an original pin that was made using a different technique, and a replacement pin.
Please use the word ‘plate’ for all components which attach the strap to the buckle – avoid using the word ‘chape’.
Buckle plates can either be made in one piece with the frame (‘integral’) or can be separate and attached by a hinge. Many shapes of plate are possible, but rectangular is perhaps the most common. For rectangular plates, it is helpful to state whether the rectangle is short (longer axis parallel to bar) or long (longer axis perpendicular to bar) – or, even better, give dimensions. Plates have an attachment end where the strap is attached, and (if hinged) a hinged end.
Don’t forget to describe the under-plate, if the buckle plate had one. Is it full-length or short? Full-width or narrowed?
Most (but not all) buckle plates have a space for the pin in the form of a pin slot or a pin hole. Pin slots are the norm on hinged buckle plates; pin holes are occasionally found, but generally on crude, perhaps home-made, plates. A pin hole is the only option for an integral buckle plate. Sometimes it is easier to describe two hinge loops instead of a pin slot and frame recesses.
It is possible to have a plate without space for a pin, if it is on a double-loop frame. Here the plate can be hinged around the outside edge of one loop and the pin can be on the central bar. Detached buckle plates of this shape are rare, and can be hard to identify.
There are no pre-Roman buckles recorded on the PAS database. The object type appears to have been brought to Britain in the early Roman period on military equipment, but only became popular as a method of fastening straps in the civilian world from the late Roman period onwards.
The Latin word for ‘buckle’ is buccula, which comes from the Latin word for ‘cheek’, bucca, and probably results from buckles having been first invented to fasten the chinstraps of helmet cheekpieces.
Buckles of the first to third centuries AD
Buckles are relatively rare (or relatively rarely recognised) in the first three centuries of Roman Britain. They appear to have been used exclusively by the military. There is good archaeological evidence for the use of lightweight buckles to fasten armour (lorica segmentata) and some evidence that heavier cast buckles were used to fasten belts.
Cast buckles (perhaps from belts)
Bishop and Coulston (2006, 106-8) discuss early military belts, but it is not always obvious how these were fastened. Examples of early Roman military belt-buckles (e.g. Bishop and Coulston 2006, 108, fig. 62.1 and 62.15) are not always easy to distinguish from the much more common later buckles.
Appels and Laycock (2007, 100-108) illustrate examples of cast copper-alloy buckles from the first to third century, but do not give findspots or explain how the dates were assigned. Similar examples from the PAS database (below, left and centre) have small integral plates with a slot above a second bar; a narrow strap would have been fixed through the slot and around this bar. There is also an integral bar for the pin between the curled-in ends of the frame.
The example on the right, YORYM-6ECD57, is of a form known as a ring-buckle; the pin was free to move around the frame. These are known from 3rd-century gravestones and altars, the earliest known depiction being on an altar from Abusina (now Eining in Bavaria) which dates from 211 AD (Bishop and Coulston 2006, 15). Very few are recorded on the PAS database and it may be that they were never common in Britain.
The pins from these buckles are simple and difficult to date when detached.
Buckles used to fasten armour
Small, lightweight buckles were also used to fasten armour of lorica segmentata type (e.g. Appels and Laycock 2007, 56; Bishop and Coulston 2006, 99, fig. 56; and a useful group from Camerton, Jackson 1990, nos. 16-27). These buckles have rectangular sheet plates ending in a single hinge loop, which articulates with two loops on another plate fixed directly on to one side of the armour. The buckle fastens a strap fixed to the other side of the armour, enabling the neck opening to be adjusted; the rest of the armour is fastened together down the front of the body using tied thongs rather than buckled straps.
These buckle frames have a separate bar on which the pin, plate and frame all hinge, but they are otherwise simple and undecorated. This makes them difficult to recognise, and we do not have many recorded on the PAS database.
Separate bars are not diagnostic of Roman buckles, unfortunately; they can also occasionally be found on later buckles, of early-medieval (see SF-A03892 and LIN-D9C500, both illustrated below), medieval and early post-medieval date (e.g. the Blossom’s Inn hoard (Margeson 1993, 25), SOM-ED83D2, IOW-3C4457 and NMS-9DC299). One characteristic which might help in distinguishing Roman buckles with separate bars is that the frames are often made from tall, rectangular-section strip, but there will always be hard cases (such as NMS-2CDDD6 and NMS-DF02D7) where it is very difficult to decide between a Roman and a medieval date.
Late Roman buckles
From c. 300 AD onwards, buckled belts become really popular in the Roman world, and most of the Roman buckles on the PAS database are dated to the 4th century. There are two main studies covering 4th-century buckles, by Hawkes and Dunning (1961) and by Simpson (1976). A comprehensive and reliable revision of 4th- and 5th-century material across Europe was produced by Böhme (1986) in the Jahrbuch des RGZM vol. 33. At the time of writing, this volume was not available online, but it is worth checking the RGZM’s website to see if it has now been digitised.
Buckles with zoomorphic decoration (Hawkes and Dunning types)
The largest group of Roman buckles on the PAS database is the late Roman buckle with zoomorphic decoration.
As with the earlier Roman armour buckles, these often have a separate bar on which the plate, frame and pin all hinge. This leads to a distinctive shape with hinge loops instead of a bar on the frame. As with the earlier cast belt buckles, many examples also have curled-in ends to the frame. Some also have ornate pins with side projections which fill the space within the frame.
The classic text defining late Roman buckles with zoomorphic decoration is Hawkes and Dunning 1961. This paper can be downloaded from the ADS as a pdf here. Hawkes and Dunning’s classification and vocabulary should be used in preference to that in Appels and Laycock 2007, and the Hawkes and Dunning type should be added to the classification field.
Hawkes and Dunning divided the buckles into four types, each with sub-types. The animals are often known as ‘dolphins’, but they have upturned noses and upright crests, so have sometimes also been identified as boars. Alternative or additional animals can include horses (on Type IB buckles) birds (e.g. BH-C63EA1 and SUR-886B81), fish (e.g. BERK-C6F061) or even humans (e.g. HAMP-34EBF6 and NLM-15A121). Hawkes and Dunning illustrate an example with both birds and humans (1961, fig. 18k, no. 24).
Type I is a simple oval or D-shaped frame with the outside edge having inward-facing dolphins (Type IA) and/or outward-facing horses (Type IB). If there is a plate, it is made from a long piece of sheet with stamped and/or engraved decoration (although see NMS-AEED18 for a Type IA frame with an oval plate). Detached long engraved plates can come from either Type IA or Type IB buckles.
Type IIA has the same confronted dolphins as Type I, but the tails are curved in at the base and the frame has hinge loops for a separate bar. There may be an integral pin bar (joining the incurved ends) as well; if so, it seems likely that these had a simple wire pin with open loop. Otherwise, cast pins with curving side projections were used. These projections tend curve backwards towards the loop; if you have one with projections that curve forwards towards the tip of the pin, it may be medieval (see NMS-4E41FB).
Plates for Type IIA are openwork, and often have edges and/or hinge loops decorated with small grooves or nicks running perpendicular to the edge. There can be a variety of patterns; most are based on rectangles, circles and U shapes, but there are also examples which imitate propeller-shaped belt-mounts, and sometimes animal or human heads.
Type IIB again has the inward-looking dolphins, but the frame and an openwork plate are cast in one piece. Most of these do not have incurved tails to the dolphins, but there are a few that do. These buckles seem to have been fragile, and we do not yet (2017) have any complete examples on the PAS database. Fragments of Type IIB are obviously hard to recognise, as some of both frame and plate have to survive.
Openwork plates can also be coupled with simple plain loops with no zoomorphic decoration, as in Appels and Laycock 2007, 229-31. The PAS database has a few of these, both large and small.
These buckles are easily confused with similar buckles with integral openwork plates, firstly of 7th-century date (see Marzinzik 2003, 53, 84, 319 and 472; Type II.26; also SUSS-C48534, and perhaps KENT-44A380) and secondly of high-medieval date, perhaps Romanesque (e.g. WAW-2F1657, GLO-5CC875, SF-87BBF6, NMS-5F04DC, NMS-D79682, HAMP1875 and the hinged plate of SOM-802172). Close inspection of the details of the frame, and any decorative techniques, will often be necessary to assign a date, as there was probably deliberate copying of Roman styles in both later periods.
Hawkes and Dunning Type IIC also has the incurving ends to the loop, but has two bars, one for the pin and one for the strap. Type IICs thus strongly resemble the first- to third-century cast examples above, but as they have animal decoration, they presumably represent a 4th-century development of the earlier buckles. Hawkes and Dunning‘s examples are both from early Anglo-Saxon cemetery sites, so offer no independent dating evidence. We have two on the PAS database at the time of writing (2017), BH-7FCB64 and HAMP-66F2F4. Neither have clear animal decoration, but both follow the shape of Hawkes and Dunning’s examples, with incurved elements or roundels between the bars.
Like the Type IIC, the Type III has an open-jawed animal head at either end of the bar. It can have a separate sheet or cast plate (IIIA), or an integrally cast plate (IIIB). Plates are short and can have stamped and/or engraved decoration.
Type III buckles are sometimes confused with the 11th- or 12th-century type with animal heads gripping the ends of the bar (e.g. YORYM-D2E525), but a useful rule of thumb is that the Roman examples tend to be larger and have a flat cross-section, and the 11th- or 12th-century type tends to be smaller and rounded in cross-section. At present, the buckles noted by Laycock and Marshall as their ‘crescent’ type (YORYM-FA4131 and BERK-AD5D33) are of uncertain date.
Lastly, Hawkes and Dunning’s Type IV has the exterior extended into a large plate. Although these are known from English graves, we do not have any recorded on the PAS database.
Buckles with no or simple decoration (Simpson types)
Simpson 1976 looks at late Roman buckles which fall outside Hawkes and Dunning’s classification. His Group I has frames with incurved outside edges and simple sheet plates, either rectangular or rounded in shape. Group II has D-shaped frames and simple sheet plate. Group III has curled-in ends to the loop like a Hawkes and Dunning Type IIA. Group III frames can have separate or integral bars, but Group I and II frames are cast in one piece.
It can be difficult to tell Simpson Group I and Group II buckles apart from similar buckles of other dates, but there are sometimes diagnostic features such as ribbed hinge loops on the plates.
Buckles with triangular plates
Buckles with triangular plates, usually cast in one piece, are known both from the late Roman world (4th century AD) and from the early Anglo-Saxon period (late 6th and 7th centuries AD). The two groups can be difficult to distinguish, but the late Roman ones tend to have semi-circular or D-shaped frames no wider than the plates, which are often openwork. The early Anglo-Saxon examples tend to have more oval frames, generally wider than the solid plates. Roman frames can have incurved outside edges, sometimes even pointed in the centre so that the frame is almost heart-shaped.
Böhme (1986, 485-6, Abb. 14) seems to be the only good scholarly source for these buckles. He gives a list and map of those known in 1986, and dates them on archaeological grounds to the 4th century AD.
See below for more on Anglo-Saxon buckles with triangular plates.
Early Anglo-Saxon buckles
Marzinzik’s classification is based on the presence or absence of a plate, which can be reliably inferred from grave finds but which is inconvenient for topsoil finds which are often incomplete. Marzinzik’s Type I has no plate; her Type II has a plate.
Oval or D-shaped frames
The simplest type of early Anglo-Saxon buckle is a small oval or D-shaped frame, with little or no distinction between the loop and the bar. The oval is often very narrow and the frame itself often has a chunky circular cross-section. These are mainly characteristic of the 7th century, although some earlier examples may exist; 6th to 7th century is probably a safe date-range. They are very much easier to identify if the pin survives, as the tip of the pin usually drapes gently over the outside edge of the frame. Although these buckles are well known from early Anglo-Saxon furnished graves, it should be noted that they can also be found on middle Anglo-Saxon settlement sites, and their date-range may well continue into the 8th or even 9th century (see below).
D-shaped versions of this simple type are Marzinzik Type I.10, whereas oval versions (with no angle between bar and loop, symmetrical about two axes) are Marzinzik Type I.11. If they have plates, they are classified according to the shape of the plate rather than the shape of the frame (see below).
Similar buckle frames are known with narrowed bars. They may originally have had plates, but plenty are known without plates, even as grave-finds. They can be very large or very small, or any size in between; larger buckles are more popular in the 6th century, small sizes in the 7th century and perhaps later. This is Marzinzik Type I.5.
The larger versions of this oval frame sometimes have a distinctive type of pin. This often has a cast-in-one decorative plate over the loop, traditionally called a ‘shield’.
The loop on the underside of the pin can either be closed (as on NMS-D541B1) or can be made from a bent-under spike, which is far easier to cast. Often the loop has disappeared entirely, and there is evidence (from solder or iron corrosion on the underside) that the loop was made separately, sometimes of iron and sometimes of copper alloy; this was clearly a point of weakness. The ‘shield’ can be a variety of shapes; the commonest is sometimes called ‘violin-shaped’ or ‘fiddle-shaped’, but this term doesn’t describe the shape fully, so you must go on to do this. The ‘shield-on-pin’ form comes into use at least by the early 6th century, and does not appear to outlast the early 7th century.
Hinged buckle plates
These oval or D-shaped frames are often found in graves together with hinged plates, some of which remain recognisable when detached from the frames. Rectangular plates with cast Style I (Marzinzik Type II.14) are obviously early Anglo-Saxon, but they are not always recognisable as buckle plates. This is because they are cast in exactly the same form as simple square or rectangular belt mounts, and converted to buckle plates either by being fixed to a sheet under-plate which has hinge loops, or being fixed to two separate sheet hinge loops (see Marzinzik 2003, plates 79.4 and 83.2 for clear examples, and HAMP-B1D3B7 for a possible PAS example with two separate sheet hinge loops). When detached, it is very hard to tell whether one of these plates is a buckle plate or a mount; as mounts are more common, it is probably best to record them as MOUNT.
Triangular buckle plates (Marzinzik Type II.23) are well known from prestigious Anglo-Saxon graves, but are rare among PAS finds. They divide into two types. Large hinged triangular plates tend to have thick, hollow plates with large domed rivets; they date to the late 6th and early 7th centuries. When found in graves, they are always the graves of men. Those recorded on the PAS database tend to be smaller than those excavated from graves.
The second type is small (c. 30mm long) and often cast in one piece; see below under ‘Buckles with integral plates’. Small buckles with triangular plates are also known in the Roman world (see above), and can occasionally be hinged (e.g. SUSS-9CA326); undecorated examples can be hard to tell apart.
Rectangular plates are known from the whole of the early Anglo-Saxon period; these are Marzinzik Type II.16 (‘long’), II.19 and Type II.24 (‘small’). Some can be dated further by their decoration, repoussé dots being commonest in the 5th or 6th centuries and ring-and-dot, openwork and notches or grooves in the edges all more common in the 7th (Geake 1997, 78). Simple rectangular plates are also known, often with chunky rivets, but these can be hard to date without a diagnostic frame or pin attached.
Other plate shapes are far less common, but include various rounded shapes which may have been imports from continental Europe (see, for example, SF-7339C4 and SF-118BD1). A group which Marzinzik describes as kidney-shaped (Type II.11), heart-shaped (Type II.12) and kite-shaped (Type II.13) are different in detail, but all have curved or pointed attachment ends and straight or double-curved hinged ends (see KENT-40A632). In difficult cases, the presence of pierced lugs rather than rivets to attach the plate to the strap is evidence for an early-medieval date.
Buckles with integral plates
So far, all the early Anglo-Saxon buckle plates illustrated in this guide have been hinged, but there are also many early Anglo-Saxon buckles with integral plates of all shapes – triangular, rectangular, rounded. Again, some have pierced lugs rather than rivets to attach them to the strap.
Small buckles (c. 30mm long) with integral triangular plates date to the mid to late 7th century and can be worn by both sexes (Geake 1997, 76-77). Small buckles with triangular plates are also known in the Roman world (see above), and undecorated examples can be hard to tell apart. D-shaped frames no wider than the plate tend to be Roman, and wider, more oval frames tend to be Anglo-Saxon; openwork plates tend to be Roman. If the buckle was attached to the strap using perforated lugs rather than rivets, it must be Anglo-Saxon rather than Roman.
Other frame shapes
Going back to frame shapes, oval is by far the commonest, but there are occasionally other shapes found. Rectangular frames (Marzinzik Type I.6) can be very variable (see illustration below). LIN-74D863 nods to oval exemplars; LIN-0C4953 has decoration of groups of transverse lines similar to those often found on oval frames. NMS-5B0BE5 has a short integral plate (Marzinzik Type II.17); it does not immediately look early Anglo-Saxon, but good parallels can be found in furnished graves (Marzinzik 2003, pl. 9.1 centre and right)
Incurved outside edges (Marzinzik Type I.7) also occur, but can be difficult to recognise when found without plate or pin. Finally, there are sometimes quite odd buckle frames which are only recognisable as early Anglo-Saxon from their art.
Middle Anglo-Saxon buckles
Buckles are rare finds from excavated middle Anglo-Saxon contexts, and it may be that they were hardly used during the period. The buckles from middle Anglo-Saxon settlement sites such as Hamwic and Flixborough are all small and undiagnostic (Hinton 1996, 6-8; Rogers and Ottaway in Evans and Loveluck 2009, 22-25) and it may be that the small simple buckles with pins drooping over the frames, familiar from early Anglo-Saxon contexts, continued in use during at least the 8th century. A buckle with small frame and drooping pin, from the fill of a mid 9th-century ditch at Flixborough, has a distinctively grooved plate with a good parallel on the PAS database at LIN-5AC340.
Decorated buckles occasionally give us some art-historical dating evidence. There is a well-preserved silver buckle with D-shaped frame, separate bar, and round-ended hinged plate with good Trewhiddle-style decoration on the PAS database (SF-A03892). The buckle from the Trewhiddle hoard itself, though, is small and undiagnostic, with a D-shaped frame and a simple rectangular undecorated plate.
Two buckles from Fishergate, an area of 8th- to early 9th-century occupation in York, had very distinctive shapes; the frames are pentagonal externally and rectangular internally, with solid pointed outside edges; the plates are integral. One had Trewhiddle-style ornament, the other a ‘fan-shaped’ motif similar to those on 9th-century strap-ends. These are paralleled by a few from the PAS database. NMS-346A95 and FAKL-A4D441, although both incomplete, have internally rectangular frames and possible fan-shaped motifs; SWYOR-CB35F3 is pentagonal externally and has an animal head on the end of the integral plate which, although worn and hard to see on the image, is again similar to those on 9th-century Anglo-Saxon strap-ends.
No strap-end of Thomas’s Class A (9th century) has ever been found in direct association with a buckle (Thomas 2000, 279).
Late early-medieval buckles
Buckles again become more common, and more recognisable, in the later part of the early-medieval period, but they have not yet attracted much research to help us in describing or dating them. Because of this, what follows is a brief impression of a few types, and should not necessarily be taken as authoritative.
It is thought that buckles in both the late Anglo-Saxon and Viking-age worlds may have been worn by both men and women, in sets with strap-ends (Kershaw 2013, 173-5) although, as strap-ends vastly outnumber the buckles, most belts must have been knotted or otherwise fastened without buckles (Thomas 2000, 265-269 and 279).
At Coppergate, only two copper-alloy buckles were found in Anglo-Scandinavian contexts (in contrast to over 100 iron buckles) and both were simple and undiagnostic (Mainman and Rogers 2000, 2568).
Some buckle types are datable from their Scandinavian-style art. The Borre style was in use from the late 9th to the late 10th century (Kershaw 2010a, 3). The Ringerike style dates perhaps to the very end of the 10th but mainly the 11th century, with the Urnes style belonging to the mid 11th to early 12th centuries (Kershaw 2010b, 5-6). Because there is often not much space on a buckle for the details, there are sometimes disagreements as to which art style is shown, but Borre-style and Ringerike-style art have both been identified on many buckles, and Urnes-style art on a few; the 10th-century Jellinge style does not seem readily applicable to buckle decoration, although it was widely used on brooches.
A very distinctive type of buckle was clearly made as part of a set with strap-ends of Thomas Class B, Type 4a (multi-headed). The plates are very similar to the strap-ends and the dating evidence for both is scant, because most if not all excavated examples seem to be residual (Thomas 2000, Appendix 3, nos.1005 and 1009; 204; Geake in Emery 2007).
The heads have rounded ears with drilled dots, and eyes also made from drilled dots. They have been identified as Borre-style, so on art-historical grounds they should date from the late 9th to the late 10th century. As NLM730 is complete with frame, plate and pin, it allows us to identify the frames; further examples of these frames include WILT-9FD11C and SWYOR-AC86A7.
There is a variety of similar frames with animal heads at the pin rest which may be in Borre style. One from Lincolnshire (LIN-D9C500) also has an animal head on the pin, which hinges on a separate bar. There are also frames with a second integral bar, such as DENO-98937A, and so these strictly speaking have a double loop. See WILT-ABB86B for an example which is more D-shaped than oval. Well dated parallels are hard to find; they may again be datable to the tenth century, or may perhaps be a little later.
BH-9B4B78 has a construction similar to NLM730, but a differently shaped plate; NMS-33CD71 has a comparable plate. This waisted plate shape is also found on buckles with integral plates which may be later (11th-century) such as SUSS-78A6E1 and HAMP-79DC46.
Other simpler D-shaped or sub-triangular buckles with Borre-style heads are known. Note the drilled dots for eyes, and the widely spaced, rounded ears that are sometimes described as like those of teddy bears. These heads are always seen facing the observer, or from above, and not in profile.
A large, usually well-made buckle frame which also appears to have Borre-style animal heads comes in a D-shaped version (with three animal heads) and also occasionally in a rectangular version (with four animal heads). One from Old Sarum (although not from a dated context) is in the Ashmolean Museum and was published by Hinton (1974, no. 32). The best-made examples have drilled dots for eyes on the animal heads, and the frames between are decorated with rows of dots or pellets, or engraved ladder pattern. Examples include WILT-952048, BERK-0BE984, NMS-35B8F3, SF-9F02E3 and SF7560 (all D-shaped) and KENT-A6FFF4, LVPL-99FBD2, NARC-709B85 and SUSS-438A54 (all rectangular). SF-76F478 retains a sheet copper-alloy plate.
Borre-style interlace is found on a few buckles frames with wide, flat outside edges. SWYOR-1F57BC, NMS-60B8C7, NMS-D68C64 and BUC-829D64 all have the distinctive loose double-strand knots which leave a small triangle for the pin rest. Thomas suggests that these buckles would make a fitting accompaniment to Class E, Type 4 strap-ends (2000, 281).
Ringerike-style animals are usually seen in profile. On well-made buckles they tend to be seen in relief; on cheaper buckles engraved outlines are used. There is at present a strong concentration of Ringerike-style buckles in Norfolk.
The commonest type of buckle with Ringerike-style decoration has a pair of profile animals with their noses abutting the bar, perhaps holding the ends of the bar in their mouths. These animals, when shown clearly, have turned-up ends to their noses and ‘lappets’ (perhaps ears or horns) emerging from the backs of their heads to end in curly scrolls. There is a more abstract pattern where the necks meet at the pin rest, sometimes resembling a trefoil or fleur-de-lis. An example of this type from Meols is discussed by Griffiths (in Griffiths et al 2007, 62, no. 307; pl. 8).
Less commonly, similar animal heads are turned around to face the pin rest. A more complex type is GLO-20D4FE, which also has a full-face animal at the pin rest. This time it is not in Borre style but instead has a bold, straight nose comparable to those seen on Class B stirrup-strap mounts. A more English (as opposed to Scandinavian) version of this comes from a late 10th- to 11th-century grave in Winchester (Hinton in Biddle 1990, no. 1101; also photographed in Backhouse et al. 1984, no. 97). Other buckle frames with a head between two animals, such as BERK-7A0563, may possibly be related, but this is a fairly common mythological motif which crops up in several cultures.
Because the Ringerike-style animal head often has both lappets and a turned-up end to the nose, both ends of the head can have small upward projections. It can therefore be hard to tell which way round the head is, and stylised animal heads can be difficult to describe. The buckle frames below are clearly related to the Ringerike-style frames above, but which way round they should be read can be uncertain. These very stylised buckles are probably still used in the 12th century; an example from Norwich with ring-and-dot decoration was found in a 12th- to late 13th-century context (Margeson 1993, 24-26, no. 128).
These Ringerike-style frames can also be found on buckles with integral plates, which tend to be narrower than the frames. See above for PUBLIC-67E0F6, which has a tiny stub of plate surviving. A more complete example, with animals in profile facing another animal at the pin rest, is DENO-0F2314; another, with two decorative holes, is NMS-A49756. The type with animals facing the bar, separated by an animal head, trefoil or fleur-de-lis at the pin rest, is more common; examples include WILT-60E406, HAMP-348BE5 and LANCUM-255E97.
There is also a small group of buckles with D-shaped frames with large moulded three-dimensional heads at the pin rest, and integral plates, sometimes with simple engraved decoration (e.g. NMS-B948D7, SUSS-A7C147, SF-D06B94, and the particularly fine example of HAMP-C9AB76). These have presumably developed from the Borre-style group with three-dimensional animal heads at the pin rest, and can be tricky to tell apart; they also come in a variant without plate (e.g. BUC-3D0F81).
An excavated example is known from St Mary’s Hospital in Chichester, where there was 10th- and 11th- century activity on site, but the buckle was residual in a 12th- to 14th-century context (Down and Rule 1971, 47, fig. 3.17, no. 17). Three much simpler examples came from 10th- to 11th-century contexts in Winchester and are illustrated by Hinton (in Biddle 1990, 512-14, nos. 1098, 1100 and 1106).
These animals (below) have angular lobes at the ears rather than the rounded ears of the Borre-style animals, but if you can’t see which art style is represented, don’t worry – just give your object a broad date-range of 10th to 11th century.
ESS-D5F027 has an animal head at the pin rest which has to be viewed from the side, and is therefore similar to the Ringerike-style animal heads on Thomas Class B, Type 6 strap-ends. Thomas dates these to the 11th century (2000, 205-6).
The Urnes style, unlike the other Scandinavian art styles, tends to produce ornament which is not symmetrical. The animals are often have limbs and other elements ending in thready tendrils. The heads can be more chunky, and detached heads can be very hard to assign to any one of the Scandinavian art styles.
Tendrils can be seen in cast relief on buckles with a D-shaped frame and expanded outside edge, such as SF-1623C7, SF-DA7852 and NLM-DC37C4. All of these appear to have identical motifs which can partly be cast in openwork.
Olwyn Owen has decoded the ornament from another of these buckles, which was excavated in Warwick (unfortunately from an unstratified context; Cracknell and Bishop 1994, 29-30, Fig. 15; Owen 2001, 214-15, pl. 11.6e). The outside edge is filled with a single Urnes-style animal in profile, with its head at one side of the bar and a foot at the other. The shoulder and hip joints occur at the corners; these can be seen clearly above, in SF-1623C7, where they are the typical Urnes-style spiral joints.
Another characteristic of the Urnes style, particularly in England, is the animal with a wide body that curves in a circular or spiral loop. This can be seen on the plate of WILT-BBDA52, which has a clear head at one end and interlacing limbs. Similar ornament can be seen on stirrup-strap mounts of Williams Class A, Type 10 (Williams 1997, especially no. 179). Cast relief decoration is also present on WILT-BBDA52’s frame, but it is harder to decode.
11th- and 12th-century buckles
There are a few buckle types which appear to be transitional between the late early-medieval buckles with Scandinavian-style art and the much more common medieval buckle types (for which, see below).
The first is a group of buckles with tiny integral plates which may be the first spur buckles. There is a large example on the PAS database (GLO-927993) dated to the 11th century by its Ringerike-style animal decoration, and an iron example with non-ferrous coating excavated from a late 11th- or 12th-century context at Winchester (Goodall in Biddle 1990, no. 1317). Most examples on the PAS database (e.g. BH-AFE45B, HAMP-E19CB1, DENO-31E468, SUR-FFC784) tend to be small and simple, with a pin hole in a lobe which projects a little into the frame. The plates are short, often tapering, with integral rivets.
The second is a group of buckles with animal heads gripping the bar, and a very rounded cross-section to the rest of the frame. They appear to date from the 11th and 12th centuries, with excavated examples from contexts of this date in Winchester (Hinton in Biddle 1990, no. 1110) and Norwich (Margeson 1993, no. 127). The Norwich example is ribbed on one face, a feature which is common on the better-made examples. A few, such as the Winchester example, have plates.
There is a brief discussion of this type, with further parallels, in the Meols report (Griffiths et al 2007, 62, nos. 311-314), where they are called ‘Anglo-Norman’.
Note that they are similar in design to the loops of swivel elements such as IOW-64F573, and fragments may be difficult to assign to a buckle or a swivel loop.
During the post-Conquest medieval period the range of buckles, and the quantities used, expands enormously. Despite (or perhaps because) of the huge numbers, there have not been many studies to help us date and understand medieval buckles.
Books that may be useful, with lots of examples from well-dated contexts, include Egan and Pritchard 1991 and Ottaway and Rogers 2002. Other general books of finds from well-excavated medieval towns and cities usually contain a reasonable selection of buckles – see Biddle 1990, Margeson 1993,
The catalogue of unstratified finds from Meols (Griffiths et al 2007) does not have any independent dating evidence, but does include well-informed and wide-ranging discussions of what dates are available from elsewhere. There is also a typology of simple shapes of single-loop buckle frame (Griffiths et al 2007, 84) which may be useful in descriptions.
A recent PhD thesis by Alex Cassels (Cassels 2013) is downloadable free from the EThOS website (find the thesis here and download it after logging in). This looks at buckles from 15 medieval towns and cities and includes a typology which is similar but slightly different to the Meols one. Both typologies have been designed around the specific (urban) data that their studies cover, and so not every PAS buckle will have a Meols or a Cassels type. The date-ranges of Cassels’s buckle types tend to be quite wide.
If your buckle falls into one of the Meols or Cassels types, then put this in the classification field – e.g. ‘Meols type 9’ or ‘Cassels type 1.1C’.
Ross Whitehead’s book Buckles 1250-1800 (the 2nd edition is Whitehead 2003) is widely used to find parallels and get a rough general idea of the likely date of a buckle. It is not always clear how the dates in this book have been established.
Buckle or brooch?
Sometimes, with a circular frame of medieval date, you may be unsure which object type to use. The general rule is that a circular frame with no pin constriction should be classified as a buckle; if there is a pin constriction, it should be recorded as a brooch (Egan and Pritchard 1991, 64-5). This works for nearly all examples, but it isn’t a hard and fast rule. Each object should be considered individually, and those with brooch-like features, such as highly decorated pins or frames, should be classified as brooches.
It is possible to date some highly decorated buckles to the 12th century on art-historical grounds. It seems likely that simple, utilitarian buckles were also used (perhaps small undecorated circular buckle frames) but it has so far been difficult to prove this with firm archaeological evidence.
Gaping mouth beast buckles
The ‘gaping mouth beast’ buckle (Ashley and Rogerson 2011) has been dated broadly to the 12th century on art-historical grounds. Ashley and Rogerson divide them into two types; Type 1 has the pin hinged in a hole in the base of the head, the tip resting in the centre of the straight outer edge, and Type 2 has the pin the other way round, hinged on a straight bar and resting in a groove on top of the animal’s head. Good examples of Type 1 include NMS-438193 and DENO-F09717, and good examples of Type 2 include NLM-B84B6D, NMS-3D5AA5 and SF-E1A115.
Put ‘gaping mouth beast’ in the classification field and, if you can, the Ashley and Rogerson type in the sub-classification field.
The earliest known gaping mouth beast buckle so far is PUBLIC-EC2A87, which has good 11th-century decoration. SF-E1A115, on the far right of the photograph above, has an offset bar and so is likely to be among the latest of the gaping mouth beast buckles, perhaps dating as late as the early 13th century.
Standing animal buckles
Another buckle type which has been dated to the 12th century on the basis of its animal art is the ‘standing animal’ type. This buckle has a D-shaped frame formed from a three-dimensional quadruped, with its head turned to look over its shoulder. The animal sometimes looks like a lion, with tail curled around the body and grooves on the neck indicating a mane, but on many examples the face also appears to be human (BH-3DEA56). Good examples include HESH-895616, KENT-B09477, LIN-6480BA (with nice ears) and those illustrated below; cheaper versions include LVPL-F733E3, LVPL-2338B5, SWYOR-88C3C3 and LIN-E0AE65.
It is possible that these buckles continued to be made into the 13th century, although the single example known from an excavated context may be residual. This comes from a late 13th- or early 14th-century context at Scott Street, Perth, (Cox 1996, 804-6, no. 1; see p. 745 for details of dating, and find the pdf of this article here) and joins a small group from Scotland (thanks to Mark Hall for information). The date-range cautiously quoted is usually c. 1100-c. 1300 AD.
Put ‘standing animal’ in the classification field.
Buckles which begin in the 13th century
The most common buckle form of all, the single-loop D-shaped or oval buckle frame with offset bar, seems to have begun in the late 12th or early 13th century and continues in use until the late 14th century. In London, the earliest of these oval buckles is generally found in ceramic phase 6 (c. 1150-c. 1200 AD) but the vast majority come from phases 7-11 (c. 1200-c. 1400 AD).
Sometimes these oval frames have a pointed pin rest (at the centre of the outside edge), and sometimes the whole outside edge is extended to a point.
Oval buckle frames with expanded and decorated outside edges of one form or another (called ‘ornate’ by Egan and Pritchard) also begin in the late 12th century and continue to the late 14th (Egan and Pritchard 1991, 76). These outside edges can be hard to describe; see above, under ‘Frame shapes’ for some guidance.
There are also very occasional double-loop buckle frames which look very like our 13th- or 14th-century single-loop buckles, so it’s likely that they are of the same date. These have pointed pin rests and/or offset inside edges which look very like the offset bars of the single-loop frames. It seems likely that the strap or plate would always have been fastened around the second loop, rather than around the central bar; in this, these early double-loop buckles are different from later medieval and post-medieval double-loop buckles, where the strap is always fixed around the central bar.
For the main series of double-loop buckles, see below under ’15th-century buckles’.
Buckle plates with cast openwork element (including the Marlowe Car Park F655 type)
This type is rarely found with a frame. The commonest form has a sub-triangular cast openwork element, riveted to a sheet backplate which includes the hinge loops.
A comparable example has been found in a late 13th-century pit at the Marlowe Car Park site in Canterbury (Blockley et al 1995, 1063, no. F655). Parallels are usually found by including the Marlowe Car Park small-find number (F655) in the Object Description field. Searches for ‘F655’ will then bring up all examples.
Variants of the type exist without the central struts. DENO-4F2386 and NLM-5A8007 consist of the cast openwork element only; SWYOR-9F5465 retains the sheet backplate with hinge loops and a central decorative rivet. This also has an excavated parallel, from a mid to late 13th-century context in Winchester (Biddle 1990, no. 1132).
Although a 12th-century date has been suggested for openwork sub-triangular buckle plates on the basis of the animal head, both of the excavated parallels were deposited in the 13th century, and David Hinton believes that the Winchester example is of 13th-century manufacture (in Biddle 1990, 509, 515-6, no. 1132).
A 13th-century date is made more likely by comparison with some undoubtedly 13th-century buckles with square or rectangular plates with similar cast openwork elements (such as NMS-F5DE61, SUR-47CA97, WMID-072880 and the highly complex BH-715066 and FAKL-AB1C25). A detached square plate of similar style came from an early13th-century context in London (Egan and Pritchard 1991, no. 1046).
Limoges is part of Aquitaine, and when the English king Henry II married Eleanor of Aquitaine in 1152 it became a possession of the king of England. It was constantly fought over until about 1300 when the English finally lost, but during that century and a half the English market for Limoges-style objects flourished.
The enamelling workshops of Limoges were making figurines, crosiers and caskets from at least 1150 AD, and by 1200, buckles were being produced (e.g. Boehm and Taburet-Delahaye 1996, 287, no. 91). Put ‘Limoges style’ in the classification field.
Limoges-style buckle frames are identified only by their decoration; their shape is normally oval with offset narrowed bar. The outside edge is widened and sometimes angled to give space for decoration. The sizes of these buckles vary widely.
Some Limoges-style buckles have a double loop but, as with other 13th-century double-loop buckles, the plate is attached to the inside edge rather than the central bar. Because of this, there is no need for a pin slot or pin hole, so detached plates can be hard to recognise. No complete example of a double-loop Limoges-style buckle has yet been recorded on the PAS database, but one with an openwork and enamelled plate is in the Kunsthistorisches Museum, Vienna (Boehm and Taburet-Delahaye 1996, no. 91; this link should take you to the page on Google Books). The Vienna example is dated to c. 1200-1220 AD.
The hallmark of the Limoges workshops is that the colours of enamel are not separated by metal lines but rather fade seamlessly into one another. This detail is not visible on many Limoges-style pieces, though, as often the enamel has completely disappeared.
N.B. The term ‘Limoges-style buckle’ implies that these were made in the style of the Limoges workshops. The term ‘Limoges buckle’, on the other hand, implies that the object was actually made in Limoges. We don’t know where these Limoges-style objects were made; they may well have been made in England.
Plates with fleur-de-lis terminal
These plates can be hinged (made separately to the frame) or integral (made in one piece). The integral plates often have openwork decoration and are rarely found complete. Dating evidence comes from several places. An incomplete openwork example mis-identified as a strap-end was found in a late 13th- or early 14th-century context in London (Egan and Pritchard 1991, no. 608). A hinged plate with rectangular buckle frame was found in a mid 13th-century context in Winchester (Hinton in Biddle 1990, no. 1122) and a detached hinged plate came from a 14th- or early 15th-century context in Norfolk (Margeson 1993, no. 141).
Buckles with M-shaped or bird-shaped plates
This type is confined to eastern England and has been dated on art-historical grounds to the 12th to 13th century. It has been defined and catalogued by Rogerson and Ashley (2018). Good examples include LIN-5FFC92, SF-5FF3F5, NMS-B73C92 and NCL-BAB662. ESS-F0D0B5 has an integral plate. NMS-4E41FB is the only example with a surviving pin, which has projecting arms which curve towards the tip. Similar detached pins (e.g. NMS-CBBE49) have often been identified as Roman; in fact, the projecting arms of Roman buckle pins curve backwards towards the loop.
Put ‘M shaped’ (with no hyphen) in the classification field.
Buckles which begin in the 14th century
Oval frames continue in use in the 14th century, when they are joined by buckle frames of rectangular, trapezoidal and long D-shaped form.
Oddly, archaeological evidence suggests that rectangular and trapezoidal buckle frames without elaborate plates don’t appear to come into fashion until the 14th century, and maybe even the late 14th century. Even very simple rectangular frames have not yet been found in contexts of the 12th and 13th centuries in York, Norwich and London (Ottaway and Rogers 2002, 2891; Margeson 1993, nos. 145 and 151; Egan and Pritchard 1991, nos. 425-427). This is despite their earlier occurrence with elaborate plates (see above under ‘Plates with fleur-de-lis terminal’ and ‘M-shaped or bird-shaped plates’).
The long D-shaped buckle frame with a terminal knop is another 14th-century type (Cassels type 1.5C, e.g. Egan and Pritchard 1991, no. 421, from a late 14th-century context). The same basic shape is also used for other strap-fittings, such as strap-slides (when it has an integral external rivet in the centre of the bar) or pendent loops. It is sometimes called ‘stirrup-shaped’, and although this is not very useful as a descriptive term, it does help in searching.
A common characteristic on 14th-century buckle frames is a pair of tiny ridges across the top and bottom edges of the frame. These can be found on square and rectangular buckle frames, and on the long D-shaped frames with knops, as well as on frames from strap clasps.
One of the commonest 14th- to 15th-century buckle types is the composite buckle. These normally have oval frames, usually with pointed pin rests, but can occasionally turn up as rectangular or trapezoidal variants. They presumably developed from 13th-century oval buckles with pointed pin rests.
Put ‘composite’ in the classification field.
The term ‘composite’ refers to the plate, which is made up of sheets sandwiched on either side of a forked spacer plate which is cast in one with the frame.
The sheets are attached to the spacer using solder, and the strap is attached to the sheets with a rivet. A strong strap is therefore fixed to a strong buckle frame by means of a flimsy soldered joint. This seems like a weak construction, but the design was extremely popular for nearly a century so must have been reasonably reliable.
The composite buckle was in use in London and Winchester from the early 14th to the early 15th century, and its decoration changes over time (Hinton in Biddle 1990, nos. 150, 158 and 159; Egan and Pritchard 1991, nos. 322-330). 14th-century examples have the ‘grooved aperture’ in their attachment end (see illustration above) but examples from 15th-century contexts lack this, having a straight or shaped end but with no aperture or groove (Egan and Pritchard 1991, 78-82).
Examples from York are all from 15th-century contexts, and all lack decoration of any kind, including the ‘grooved aperture’ (Ottaway and Rogers 2002, 2890-1). PAS finds from the north are currently showing the same lack of decoration, so it may be that the composite buckle was used in the north at a slightly later date.
Strap-ends with the same composite construction are known, and detached plates from these can be hard to distinguish from buckle plates. A useful tip is that the plates from buckles have a straight end to enable them to fit against the frame; the plates from strap-ends have a pointed or rounded end.
Large cast circular buckle frames (‘annular buckles’)
There is a distinctive group of large, heavy circular buckle frames with chunky pins, which appears to belong to the 14th century. The frames are 40mm or more in diameter and oval in cross-section. The pins often have decoration at the junction of loop and shaft, and sometimes have the tip worn or cut away on the underside to allow them to sit nicely on the frame.
These buckles are occasionally found in graves; where identifiable, these are the graves of men. The buckles are found, sometimes in pairs, at the top of the femurs. They are mainly found in the graves of monks, but this may be because monks were buried in their everyday clothes (their habits) whereas lay people tended to be buried in shrouds. The evidence is summarised in Gilchrist and Sloane (2005, 85-6).
Some authorities (Gilchrist and Sloane 2005, 86, quoting Russell-Smith 1956) suggest that these buckles are to fasten hose to a belt in a sort of suspender-belt arrangement. Russell-Smith goes into a good deal of detail about medieval men’s undergarments, and it is interesting to note that the medieval equivalent of underpants was the ‘breke’ or ‘breche’, held up in the absence of elastic by being wrapped over and tucked into a belt. But underpants are rarely shown on manuscript illustrations, and in any case the tucking-in obscures any buckles.
Standley (2013, 101-4) has re-examined the pictorial evidence and has concluded that the hose were attached to the breche-belt by ties rather than by a buckled leather strap. She also points out that such large buckles would have been uncomfortable on underwear. Instead, they may have been used on the belt fastening the monastic habit – but again, very few pictures seem to show habits fastened by buckled belts.
The presence of similar buckles in the mass graves from the Battle of Visby (fought in 1361) suggests that they may have been used to fasten armour in a similar way (Standley 2013, 101). Whatever their function, they always seem to date from the 14th century, probably the late 14th century (Egan and Pritchard 1991, 58 and 64-5). There is a good stratified example from a late 14th- or early 15th-century context in Winchester (Biddle 1990, no. 1245) and two from York (Ottaway and Rogers 2002, 2886-7, nos. 12880 and 14293).
NB A circular ring with oval cross-section may be a frame from one of these buckles, but may alternatively have had other uses. For example, hanging lavers of the 15th century were suspended from similar rings. It is probably safest to record frames with no pins as RING.
Recent work has confirmed that at least a third of the population died in the Black Death, which ravaged England in 1348-9. This population collapse is visible in pottery concentrations – there is far less pottery used after c.1350. It may be that this is also the reason that fewer 15th-century buckles are recorded than 14th-century ones.
Buckles which begin in the 15th century
At the start of the 15th century, girdles and belts become wider and the double-loop buckle frame becomes much more popular.
A wide variety of double-loop buckles are made, with rectangular, D-shaped and oval loops. The date at which they become popular may vary around the country; Egan and Pritchard give a date of 1350 onwards for London examples (1991, nos. 331-334), but in Norwich they come only from early post-medieval contexts (Margeson 1993, nos. 163-177). Many forms continue into the 16th century.
The commonest form has two identical oval loops. It is sometimes known as the ‘spectacle buckle’, but this term is not particularly helpful and a full description is better. The bars of these double-loop buckle frames can project beyond the frame; although this is often thought of as a post-medieval trait, it does occasionally occur in buckles from medieval contexts (e.g. Egan and Pritchard 1991, nos. 380 and 386; Biddle 1990, no. 1152).
Buckles with two identical rectangular loops are less common. There are also more unusual variants such as the type with two projections, and the rose-shaped type. This always has five petals, and so may allude to the Tudor dynasty, as their badge was a five-petalled rose.
There are also some more ornate double-looped buckle frames that can look much later than 15th century, but which do crop up in excavated contexts of that date.
The examples illustrated above can be compared (left to right) with three buckles from Norwich and three from London. HAMP-3E0A67 is similar to Margeson 1993, no. 152, from a context of c. 1400-1600. SF-BA1858 is similar to Margeson 1993, 158, from a late 15th-century context. SF-CB6C08 is similar to Egan and Pritchard 1991, no. 471, from a context of c. 1400-1450. NMS-842A53 is similar to Egan and Pritchard 1991, no. 447, from a context of c. 1350-1400 (but there is also one from a context of c. 1550-90 in Chelmsford (Cunningham and Drury 1985, fig. 26, no. 10) so it may have been a long-lived type). PUBLIC-4B301F is similar to Egan and Pritchard 1991, no. 472, from a context of c. 1400-1450; another of these distinctive asymmetric buckles was found in a 15th-century grave in the north transept at St Mary Merton (Gilchrist and Sloane 2005, 84-5).
Single-loop buckles with hollow plate (including ‘lyre-shaped’ buckles)
Large, ornate buckles still continue the single-loop form. There is a distinctive series of 15th-century belts with matching buckles and strap-ends, with comparatively small integral plates and large frames. The plates have the front, sides and (usually) the reverse all cast in one piece, and so are hollow. , The frames are often sub-triangular, with incurved sides; this shape is sometimes called ‘lyre-shaped’, but while the term is handy for searching, it excludes variants such as NMS-89BBC4 and SWYOR-753A86, both of which have oval frames. ‘Lyre-shaped’ can be used in a record, but please add ‘hollow plate’ in the classification field for all of these buckles and strap-ends, whatever shape the frames are.
Because the matching strap-ends hang down from the belt, they can be seen on occasional effigies and brasses which often give the date of death of the person depicted. This evidence has been used to give a date of between 1390 and 1410 to the strap-ends (Ward Perkins 1940, 268, fig. 84, nos. 15-17). Although Ward Perkins thought that this type then went out of fashion, it now seems that the buckles continued in use into the second half of the 15th century (Egan and Pritchard 1991, 106).
This type of buckle often breaks, usually across the join between the plate and the frame. If the pin hole breaks away, it can be impossible to work out if the object was originally a buckle or a strap-end. Note that two of the group illustrated above (LANCUM-1F98A7 and SUR-545EA5) have a black-letter M engraved on the plate, presumably for the Virgin Mary, and that this needs to be added to the inscription field.
Buckles with swivelling arm (‘locking’ buckles)
The ‘locking’ buckle is made up of three separate components which are quite easily recognised. The separate bar is cast in one piece with the curved ‘locking’ arm, which has a globular terminal often decorated with grooves. The pin is cast in one piece with a long tube which hinges on the bar. The frame is rectangular, with tall, shaped sides pierced at either side to hold the bar, and an outside edge which is L-shaped or grooved in cross-section. The curved arm swings down over the tip of the pin to rest in the groove or recess. Put ‘locking’ in the classification field.
Complete examples are shown in Ward Perkins 1940 (pl. LXXVII nos. 11 and 12), and six from London listed by him (1940, 279-80).
It is still a mystery how these buckles were used. Although it has been called the ‘locking’ buckle since at least 1940, the pin is never long enough to be held in position by the curved arm (or vice versa) and the arm is always free to swivel. It has been pointed out that the arm’s globular terminal is similar to the terminals on purse bars of Williams Class J (e.g. DEV-E87497) and it is possible that the arm was used as a hook, to suspend a purse.
The date has until now also been uncertain. Margeson (1993, 28, fig. 13, no. 139) gives a 14th-century date, perhaps thinking of a similarly shaped rectangular frame from a late 14th-century context in London (Egan and Pritchard 1991, 97, no. 445; see similar examples at SF-53FFD8, HAMP-15C9A4 and WILT-E177D5). On these there is no provision for a curved arm, and the two types may well be unrelated.
The small amount of direct dating evidence suggests a similar date to the purse bars, in the late 15th and 16th centuries (Williams 2018, 11-12). There are two stratified examples of ‘locking’ buckles from 16th-century contexts in London (Egan 2005, nos. 116-117) and one from a context of c. 1485-1600 at Sandal Castle, Wakefield (Mayes and Butler 1983, 231-3, fig. 1, no. 6). There is also an example clearly shown on a statue of St Roch in the Rijksmuseum in Amsterdam, which dates to c. 1500 AD (https://www.rijksmuseum.nl/en/collection/BK-NM-24).
Spur buckles and other medieval buckles with specialist uses
See above, under ‘11th- and 12th-century buckles‘ for the earliest spur buckles yet identified.
Spur buckles become common in the 13th century. In general we identify small buckles with short integral plates as spur buckles, from occasional examples found still in place on the spur, although medieval spurs with buckles still in place are, unfortunately, rare (HAMP844 is one of the few PAS records; also see Egan and Pritchard 1991, fig. 69, and Alexander and Binski 1987, 259-60, nos. 166-7). Plates with rivets seem to have been attached to the spur leathers, and plates with hooks seem to have been attached directly to the holes in the spur terminals.
These spur buckles are of varying sizes, with differently shaped frames, and with both integral and separate rivets. Some of the hooked buckles also have a separate rivet (see the bottom right example above, NMS-28D1ED). They are short and sturdy, just right for use on a spur. Buckles like the riveted ones, with bevelled edges to the plates, come from 13th- to 14th-century contexts in London (Egan and Pritchard 1991, 106-7).
Similar buckles with much longer plates ending in a hook also have at least one separate rivet; if the hooks fitted into the spur terminal, the use of the rivet is uncertain. As the plates become longer, some curve to fit around the foot, but others do not.
The long type with rivets and no hook are very common finds on the PAS database (although mysteriously absent from London and other urban contexts) and it is hard to imagine that every single one was used on a spur. Some are rather over-long and spindly, and also seem too flimsy for this use, so perhaps weren’t spur buckles at all. At present we are recording them as such, though, so add the word ‘spur’ to the classification field for all of these small buckles with integral plates, whether short or long, and whether hooked, riveted or both.
N.B. Although small flat buckles with a fleur-de-lis terminal are sometimes found with integral plates, these seem not to be spur buckles. See above, under ‘Plates with fleur-de-lis terminal’ in the section on 13th-century buckles. PAS examples include SF-29EFD6, NCL-C8B460, HAMP1875 and DOR-1E0744.
Another type of small buckle with integral plate is sometimes suggested as a spur buckle, but a function on a dog’s lead is more likely (see below). These buckles have been called T-shaped, keyhole and crossbow buckles. The name T-shaped is probably the best fit, so put ‘T shaped’ in the classification field. They divide into two basic shapes, both seen below; one has a keyhole-shaped hole and often an animal head at the outside edge of an oval frame, and the other has a more rectangular or sometimes D-shaped frame and a circular hole.
The zoomorphic examples (top left on the image above) have animal heads similar to those on strap clasps (such as LEIC-D6D2DC, DENO-508E28 and LVPL1490) which are normally dated to the late 13th to early 15th century. An example with a rectangular frame, similar to those on the right on the image above, was found in a mid to late 14th-century context at Ludgershall Castle, Wiltshire (Ellis 2000, 126-7, fig. 6.2, no. 10), giving archaeological confirmation for this dating.
This type of buckle looks very much as if it was not attached to a strap, and in fact a few have now been found attached to swivels. HESH-132E83 and BERK-703D40 are complex, well-made examples originally consisting of two T-shaped buckles ending in hinge loops rather than holes, flanking a central swivel. Another set (SF-3AD2BB and SF-86DE13) is made from gilded silver, which is an unusual material for a buckle of any sort. A similar central swivel has been found in a late 14th-century context in London, but misidentified as a tap handle and stem (Egan 1998, 242-3, no. 745).
Judging from the form of the buckles, though, something like NCL-5D3D67 (below) may have been more common; a simple T-shaped buckle ending in a hole, articulating with a long oval loop which also holds a swivel. Finds like this have suggested that these buckles may have been used on something like dogs’ leads; they are a little large for hawks’ jesses.
Complete sets like these, which are more than just a buckle, are usually recorded as LEASH (the British-English equivalent, LEAD, does not appear in the mda thesaurus).
Buckles with loop at either end
Another type of buckle with a specialised use is that with a loop at either end of an integral plate. We do not have many on the PAS database; most of the excavated examples are iron with a tin coating. Their function is uncertain; suggestions include use on armour or horse-harness. They are often bent in an S-curve, perhaps during use.
The plate always has three holes, but there are never two surviving pins. It is not always clear which hole was the pin hole. It seems likely that the other two holes were normally used for rivets. It is possible that the second loop was used as a strap-slide (to hold down the free end of the strap). Excavated examples from London date to the late 13th or 14th century (Egan and Pritchard 1991, 108-10, nos. 488-496).
A distinctive, decorative buckle is the type termed ‘disc-on-pin’ (Redmayne 2014). The frames are small, typically 19-30mm in diameter, and they resemble annular brooches; they are circular, with a constriction at one point to accommodate the pin. However, they also have a central bar attached, perpendicular to the orientation of the pin, set into grooves cut into the reverse of the frame, the edges of the grooves being hammered over to secure the bar. Due to the constriction on the frame it had previously been suggested that they were brooches, but examples are now known with a plate secured to the bar which makes it obvious that they are in fact buckles. The plates wrap around the bar and do not have any slot, as the pin is held on the inside edge of the frame.
As the name implies, the pin has an expanded central section, either a solid disc or an open ring. The disc is occasionally rather more lozengiform than circular. The pin continues beyond the disc to end in a short point. The front of the disc is often embellished with engraved patterns, dummy rivets, or other features. Redmayne 2014 gives a classification based on the form of the disc.
How these buckles worked, and why the bar was made separately, are still unclear, as is their precise date. The decorative details, including the dummy rivets, point to a date in the 13th and 14th centuries (compare, for example, the rivets with large heads on buckles such as Egan and Pritchard 1991, no. 314 (late 13th- and early 14th-century context) or Biddle 1990, no. 1170 (from a 14th-century context).
The short pointed end to the pin would have been perfectly functional, but in normal use the disc would have been hidden behind the strap. Tom Redmayne has suggested a way of folding the free end of the strap back on itself and securing it with a strap-slide, which would leave the decorative disc visible (Redmayne 2014, 5, fig. 12).
Fragments of disc-on-pin buckles can be difficult to identify. Detached pins, with open loops looking like hooks, are sometimes thought to be dress hooks; occasionally a frame with no bar or pin is thought to be from an annular brooch, because the slots on the reverse have been missed.
We record them as BUCKLE in the object type field, with ‘disc on pin’ in the classification field and the Redmayne type in the sub-classification field.
An assemblage of 28 copper-alloy buckles from the Mary Rose (Gardiner 2005, 99-105) is particularly useful for identifying 16th-century buckles, as they must all have been in use when the ship sank in 1545. But there were 415 men on the Mary Rose, and less than 40 survived the sinking; a minimum of 179 individuals were recovered during the excavation. 28 buckles is a small number for this many people, and shows that most of the men on the ship were probably not wearing a buckle.
Nine of the Mary Rose’s buckles were single-loop. Of these, one was D-shaped and eight trapezoidal. Decoration consists of oblique grooves, symmetrical about the pin rest. There are many parallels to these buckles on the PAS database.
Over half of the Mary Rose’s buckles were double-loop, most with identical oval loops; there were three with one loop rectangular and one D-shaped, and one with both loops rectangular. Decoration consists of oblique grooves, mainly symmetrical about the pin rest, and moulded rosettes. One pictured below (IOW-F01169) has traces of a black coating, often called ‘lacquer’, which is found occasionally on late 15th- and 16th-century items.
16th- and 17th-century sword-belt fittings
There is a wide range of 17th-century sword-belt fittings which include buckles and buckle-like objects used (without pins) as strap-slides. If you are struggling to identify a buckle, have a look at the guide to post-medieval sword-belt fittings.
Buckles appear to have risen in popularity during the 17th century. Secure contexts to help us date 17th-century buckles are annoyingly rare, and although Whitehead illustrates plenty of examples which seem likely to be of 17th-century date, definite proof is lacking. Here are some examples.
There are a few excavated examples which can back up these suggested dates. WMID-9CC53C is paralleled by an example from a 17th-century context at Pontefract Castle, which saw at least three sieges during the Civil War (Roberts 2002, 274, fig. 112, no. 14). WMID-BCEB64 is paralleled by an example from a context of c. 1690-1720 in Exeter (Allan 1984, no. 87, fig. 190).
The first shoe buckles (with separate bars)
Buckles began to be used on shoes in the second half of the century. They are often quoted as ‘post-Restoration’, but although the Restoration was not proclaimed until May 1660, Samuel Pepys famously recorded that ‘This day I began to put on buckles to my shoes’ in his diary entry for 4th January 1660. A date of 1650 onwards is probably safer for shoe buckles.
They can be recognised by their separate bars, usually made from iron, inserted into drilled holes in the frame. The bars hold the pins and also small plates, the earliest of which end in button-like discs or slender curving bars. These allowed the buckles to be easily taken on and off the straps of shoes, which had slits cut to accommodate the buckles. The buttons can be integrally cast or made separately, and the plates sometimes have makers’ marks on them. Two examples with buttons come from contexts of c. 1675-1700 in London (Egan 2005, 37-8, nos. 120-1).
Similar buckles were also worn at the knee to fasten knee-length breeches, which became fashionable in the first half of the 17th century and continued to be worn for two hundred years. Knee buckles may be distinguishable from shoe buckles – it’s been suggested that the anchor-like curving attachment bars may be from knee buckles rather than shoe buckles – but we need confirmation of this from fashion historians.
Shoe buckles continued in popularity throughout the 18th century, when the button-like attachment was largely superseded by the spiked plate. Plates with internal spikes are found in 18th-century contexts in Amsterdam (Baart 1977, 174, nos. 229-230). 18th-century shoe buckles tend to be larger than 17th-century examples and more steeply curved to fit around the foot.
Georgian shoe buckles have been collected as art objects for at least the last fifty years, and so have appeared in print with lots of non-standard terms for their parts. You will find words such as chape, spindle, pitchfork tongue, bridge, and roll; these are not easy to understand, so if you use them, please add a full description to explain what you mean.
The PAS does not routinely record objects less than 300 years old, and very few 18th-century buckles will be worth recording. If your buckle has a very interesting maker’s mark, or a historical message, or might be confused with another object type, then please record it – but otherwise, spend your time on something more important. For those who are interested in these buckles as art objects, the BEP Students’ Room at the BM has a very scholarly catalogue of the collection at Northampton Museum.
NB: if you have a maker’s mark on your buckle, don’t forget to add the letters to the Inscription field.
Similar buckles were also used on knee breeches and hats, and also on ‘stocks’ or fitted neck cloths, which are still used today in formal riding outfits (but with pins rather than buckles these days). Stock buckles are recognised from their plates, with studs that were pushed like buttons through embroidered eyelets on the stock.
Post-medieval spur buckles
17th-century spurs with double-loop buckles are occasional finds, but the buckles are no different to those found on other straps, and so cannot be recognised as such when detached from the spur. There is no standard way to illustrate a spur with its buckle; it may be clearer at right angles to the spur side, or parallel to the spur side.
Record combinations of spur and buckle as SPUR.
Modern horse-harness buckles
Modern (19th or 20th century) copper-alloy horse-harness buckles have recessed bars, presumably to accommodate thick straps, which makes them easy to recognise.
Biddle 1990 has a small number but a wide range of medieval and post-medieval buckles from well-dated contexts in Winchester.
Bishop and Coulston 2006 discuss Roman buckles as part of the history of military equipment.
Cassels 2013 looks at medieval buckles from a range of published and unpublished urban excavations. The range is limited, which is interesting in itself, and some of the date-ranges are very wide. Log into Ethos to download the pdf.
Egan and Pritchard 1991 (reprinted 2004) is the standard work for naming the parts of, and dating, medieval buckles. A huge number were recovered from dated contexts in London, together with some evidence for manufacturing.
Griffiths, Egan and Philpott 2007, the Meols report, has a huge number of buckles but no dated contexts. Geoff Egan used evidence from London to assign dates, but the Meols report remains useful for identification and finding parallels.
Marzinzik 2003 is the best source for all early Anglo-Saxon buckles.
Ottaway and Rogers 2002 illustrates a range of medieval buckles from dated contexts in York, and catalogues them at the back of the book. Although not all are illustrated, it is useful to have a group from a major city outside London. Find the pdf here.
Simpson 1976 looks at the simpler types of late Roman buckle.
Whitehead 2003 is widely used to find parallels to medieval and post-medieval buckles, and get a rough general idea of the likely date. It is not always clear how the dates in this book have been established, but for post-medieval buckles at any rate they seem reasonably accurate.