Embracing the Provinces: Society and Material Culture of the Roman Frontier Regions

Embracing the Provinces is a collection of essays focused on people and their daily lives living in the Roman provinces, c. 27 BC-AD 476. The main aim is to showcase the vibrancy of Roman provincial studies and suggest new directions, or new emphasis, for future investigation of Roman provincial world. It capitalizes on a wealth of data made available in recent decades to provide a holistic view on life in the Roman provinces by analyzing various aspects of daily routine in the frontier regions, such as eating, dressing, and interacting. The contributors, who are acknowledged experts in their fields, make use of innovative interpretations and modern approaches to address current issues in the study of the provinces and frontiers of the Roman Empire.

Twenty-one essays are cohesively structured around five themes, encompassing studies on the female and juvenile presence on Roman military sites, Roman provincial cooking, and Roman cavalry and horse equipment. For the first time in the Roman provincial scholarship the volume has a special section on the subject of Roman leather, providing a much-needed overview of the current stance of work. A few papers deal also with experimental archaeology. The essays reflect a wide geographical and chronological range, while retaining thematic consistency, and will be of great interest to those working in Roman archaeology and provincial studies.

• Language: English
• Publisher: Oxbow Books
• Release date: Aug 16, 2018
• ISBN: 9781789250169

Book preview

Part 1

It’s a man’s world

Chapter 1

Two pieces of cavalry helmet from the province of Gelderland

Annelies Koster

The largest number of Roman helmets found in the Netherlands have been from dredging operations in the river Waal, near Nijmegen and to a lesser extent further east in the vicinity of Millingen and Pannerden (Klumbach 1974). More helmets and helmet parts have also come to light in Nijmegen, found in archaeological excavations and by members of the public, i.e. from the Kops Plateau and the St. Josephhof (van Enckevort and Willems 1994; Junkelmann 1996, 93, O83–6; Meijers and Willer 2007, 21–30; van Enckevort 2007). In recent years several more helmet parts were found following large scale earth removal and dredging operations, carried out in the eastern river area during the Room for the River program. Despite detailed archaeological research preceding such large scale operations, some finds may still come to the surface after the excavations have taken place. This is the unfortunate reality of the methods employed by these earth removal operations. Sometimes amateur archaeologists are permitted to search the dumping ground or the gravel depot with metal detectors. However, they do not always report their finds.

This was not the case for an extraordinary bronze face mask from a Roman cavalry helmet, which the Museum Het Valkhof gratefully received in 2015, as a long term loan from the territorial development firm K3Delta (inv. no. 2015.135). The mask had been found in 2014 by two metal detectorists searching the gravel depot belonging to K3Delta in De Steeg (gem. Rheden). It is thought that the face mask may have been dredged up with the sand and gravel from the river bypass at Nijmegen. However, it is also possible that the mask comes from earth removal and dredging operations in the vicinity of Rees on the Rhine (DE), as the sand and gravel from both locations were stored together in the depot. So there is no absolute certainty as to the find location.

The bronze mask, which was somewhat dented when it was found (Fig. 1.1), has been restored to its original shape by the archaeological restorer Johan Langelaar from Amersfoort (Fig. 1.2 and Plate 1).

The face mask (H: 170 mm W: 160 mm) was cut from a sheet of copper alloy of 1–2 mm thick. At the tip of the nose the sheet bronze was hammered extremely thin, allowing it to damage easily. The tip of the nose has been dented, deformed and cracked and seems to have been repaired on the inside in Roman times with some tin. The dent on the nose could possibly have been the result of a slap in the face during a fight.

Figure 1.1 Face mask as it was found, before restoration. Photo: Johan Langelaar, Amersfoort

Figure 1.2 Front, back and side view of the mask after restoration. Photos: Museum Het Valkhof, Ronny Meijers, Nijmegen

The forehead of the mask is extremely low and the missing bowl of the helmet must have protected a large part of the forehead. A hinge above the nose originally connected the face mask to the bowl of the helmet.

The face looks stern and expressionless. The eyebrows are thin and arched and the cut-out eyes are narrow and oval. They are quite like those from most other face masks, and are pronounced by the stylised eyelids. The nose is strong, long and narrow with large nostrils. The mouth is small and well proportioned. It is also open, allowing the wearer of the mask to breathe.

The holes on both sides of the round chin were probably designed to provide additional connecting points between the face mask and the helmet. Some face masks have bronze studs in these holes, to which leather connecting straps could be fastened. For instance, the face mask from Vechten (Rijksmuseum van Oudheden in Leiden (NL), inv. no. VF*1047) also has holes on both sides of the chin (Klumbach 1974, 64, no. 54; also see for fastening the mask to the helmet, Hanel and Willer 2009, 211, abb. 2). But most face masks of this type do not have holes under the chin. This could mean that the masks were held in place by the cheek pieces instead.

In addition to the two holes, a bronze loop, fixed with two rivets, was attached under the chin of the face mask (Fig. 1.3). Maybe the fastening with the bronze studs alone did not function well enough. Through this loop a leather thong could be guided across the cheek pieces to the back of the helmet and fastened there.

The face mask is part of a Pseudo Attic cavalry helmet of the Kalkriese type. It is characterised by the low forehead with hinge, by which the mask was fastened to the helmet, and also by the absence of ears. This means that the helmet should have had cheek pieces with preformed ears (Born and Junkelmann 1997, 18–21). The most well-known and best dated face mask of this type was found in Kalkriese near Osnabrück in 1989/90. It was most likely lost in the battle between the Roman troops and the Germanic tribes in AD 9 (Franzius 1993, 131–5). The Kalkriese face mask was of iron and originally covered with silver foil, parts of which have been preserved under the bronze border that frames the mask (Franzius 1993, 131; Willer and Meijers 2007, 31–50; Hanel and Willer 2009).

Figure 1.3 Chin of the face mask, before restoration. Photo: Johan Langelaar, Amersfoort

Two masks of the Kalkriese type are known in the Netherlands: one of iron was found on the Kops Plateau in Nijmegen and dates to sometime before AD 70, based on dates for the military presence there (Willems 1991, 12; on loan in the Museum Het Valkhof, inv. no. 2017.2). The other iron mask, covered with sheet bronze, was found during the construction of fort Vechten between 1867 and 1869. Klumbach (1974, 64, no. 54) dates this mask to the second half of the first century, on obscure grounds however.

Meanwhile about ten pieces of the Kalkriese type face mask are now known. Born and Junkelmann (1997, 18) mention seven of these pieces, but recently more pieces have come to light in the art trade and also in private collections (Humer 2006, 103–4, no. 340, abb. 39; Christies Sale 14230, Antiquities, 5 July 2017, London, lot 133). Some were found together with their helmet bowls and/or cheek pieces. One was a bronze mask from a helmet of the Weisenau type (Junkelmann 1996, 93, O88), a type of helmet from the first quarter of the first century AD, which was most likely also used by horsemen of the Celtic auxiliary cavalry (Junkelmann 2000, 42). There is also an iron face mask, covered with sheet brass, which was found together with an iron helmet bowl, covered with richly decorated sheet bass in relief. This helmet is of the Weiler/Koblenz-Bubenheim type (Junkelmann 1996, 84–8). Another bronze face mask was found together with two cheek pieces with ears, like the ones of the Weisenau and Weiler types (Humer 2006, 103–4, no. 340, abb. 139).

Figure 1.4 Both sides of the brow plate. Photos: Museum Het Valkhof, Ronny Meijers, Nijmegen

It can be proven that these face masks were actually part of cavalry helmets by the inscription TVR (ma) PAVLI FVSCI (cavalry squad of Paulius or Paulus, possession of Fuscius or Fuscus) written on the side of an iron face mask, covered with a tinned sheet of brass (Junkelmann 2000, 189–90, taf. xxi, from unknown findspot).

Junkelmann sketched the typological development of these face mask helmets from the late Augustan period, in which the Kalkriese type – the type without ears on the mask and with a hinge low on the forehead – seems to belong at the beginning of the series. This type of mask seems to have been in use until the middle of the first century, as part of the helmets of the Weisenau and Weiler types.

The mask could be lifted upwards, without the helmet falling apart in two pieces. The helmet bowl always remained firmly fixed to the head. Experiments have also made it clear that the helmets may also have been used in cavalry battles. The relatively large openings in the mask provided enough sight and also allowed enough fresh air for the rider to engage in mounted combat (Born and Junkelmann 1997, 18–21, 31).

Above all, research by the LVR-Landesmuseum in Bonn and the Museum Het Valkhof, has demonstrated that the combination of iron and bronze from which these face masks were made, was highly resistant to damage. It provided the horseman with effective protection against attacks to the face (Meijers and Willer 2007).

These early first century face masks – like the helmets of the Kops Plateau type, dating around the mid-first century AD – seem to have been primarily made for use in battle and not just in tournaments or hippika gymnasia, as was often assumed (Künzl 2008, 112–4). In the second and third centuries the construction of these cavalry helmets changed, which made them unsuitable for the battle. Most likely by then they were only ceremonial in function, perhaps used as parade helmets.

A second part of a cavalry helmet was found in 2015, in a heap of clay and sediment at the earth removal and dredging operation in Millingerwaard (northwest of Millingen aan de Rijn). It is a bronze brow plate of a cavalry helmet, relief embossed from sheet bronze, which was subsequently tinned to give a silvery look (on loan in the Museum Het Valkhof, inv. no. 2016.15). Originally the brow plate was crescent-shaped, but was found in a folded and slightly bent condition. These deformations though may well have been the result of the dredging operations.

The brow plate is divided in three horizontal zones, separated from each other by pearl and ovolo borders. The middle zone, between the two ovolo borders, is decorated with a wreath of oak leaves, in three leaf fascicles, set against a background of punches (Fig. 1.4). This zone ends with a rosette on both sides of the brow plate. But part of the rosette on the left side is missing however. Holes for fixing the plate to the helmet are visible on both ends, situated behind the rosettes, and also on the front, in the center of the brow plate (Fig. 1.5).

The decoration on the upper zone consists of a waves motif, rolling from the centre of the plate to the left and to the right. Again the background of this decoration is punched. The center of the brow plate is decorated with a crescent-shaped, floral motif, with large curly petals and a rosette in the center (Fig. 1.6). The background of this floral motif is also punched. The lower zone of the plate, between the ovolo and pearl borders, is undecorated.

These bronze, mostly tinned, brow plates were applied to iron cavalry helmets of the Weiler/Koblenz-Bubenheim type, dating to the second quarter of the first century (Born and Junkelmann 1997, 17ev?). A cavalry helmet of this type from the river Waal near Nijmegen has a comparable bronze brow plate, decorated with a wreath of oak leaves, which are embossed in high relief (Braat 1939, 40–2; Robinson 1975, 98–9; Klumbach 1974, 46–7, no. 33; Heijden and Koster 2017, 23). Another brow plate with a wreath of oak leaves also came from the river Waal near Nijmegen and is decorated in the same way (Braat 1939, 39–40; Klumbach 1974, 47–8, no. 34, taf. 34; Robinson 1975, 99, no. 272).

We should assume that the soldiers wearing helmets decorated like this, did not do so because of their beauty. The wreath must have had a special meaning for them (cf. for instance the helmet of the Weiler/Koblenz-Bubenheim type from Xanten with the gilded wreath of olive leaves and olives: Prittwitz und Gaffron 1991, 237–9; Prittwitz und Gaffron 1993). The oak wreath, the corona civica, was reserved for Roman citizens who had saved the life of a fellow citizen. Most likely the corona civica, was chosen on purpose as the decorative motif for these helmets. The horsemen wearing the helmets may actually have been awarded these insignia. In this way they could show their wreath permanently on their helmets, allowing them to distinguish themselves from their colleagues. It means also that these cavalry helmets, with corona civica, must have been used by Roman citizens, serving in a cavalry unit. The horseman could therefore have served in the cavalry unit of a legion, or as the leading officer, the praefectus alae, of an auxiliary cavalry unit (Prittwitz and Gaffron 1991, 237–9; Prittwitz und Gaffron 1993).

Figure 1.5 Right side of the brow plate with rosette and fixing hole. Photo: Museum Het Valkhof, Ronny Meijers, Nijmegen

Figure 1.6 Centre of the brow plate. Photo: Museum Het Valkhof, Ronny Meijers, Nijmegen

Bibliography

Born, H. and Junkelmann, M. (1997) Römische Kampf- und Turnierrüstungen. Sammlung Axel Guttmann, bd. VI. Mainz am Rhein, H. Born.

Braat, W. C. (1939) Romeinsche helmen in het Rijksmuseum van Oudheden. Oudheidkundige Mededelingen uit het Rijksmuseum van Oudheden te Leiden 20, 29–46.

Enckevort, H. van (2007) Romeinse gezichtshelmen uit bijzondere bergplaatsen in Nijmegen. In F. Willer and R. Meijers (eds) Achter het zilveren masker. Nieuw onderzoek naar de productietechnieken van Romeinse ruiterhelmen/Hinter der silbernen Maske. Neue Untersuchungen zur Herstellungstechnik römischer Reiterhelme, 9–16. Nijmegen, Museum Het Valkhof.

Enckevort, H. van and Willems, W. J. H. (1994) Roman cavalry helmets in ritual hoards from the Kops Plateau at Nijmegen, The Netherlands. Journal of Roman Military Equipment Studies 5, 125–37.

Franzius, G. (1993) Die römischen Funde aus Kalkriese. In W. Schlüter (ed.) Kalkriese – Römer im Osnabrücker Land, 107–92. Bramsche, Rasch.

Hanel, N. and Willer, F. (2009) Nachweis einer Metallklebemasse bei einem römischen Reiterhelm aus Xanten-Wardt. In A. W. Busch and H.-J. Schalles (eds) Waffen in Aktion, Akten der 16. Internationalen Roman Military Equipment Conference (RoMEC), Xanten, 13.-16. Juni 2007, 209–15. Mainz am Rhein, Philipp von Zabern.

Heijden, P. van der and Koster, A. (2017). Romeinse helmen in Nederland. Alle helmen en viziermaskers van de limes en de rest van Nederland. Leiden, Hazenberg Archeologie.

Humer, F. (2006), (ed.) Legionsadler und Druidenstab. Vom Legionslager zur Donaumetropole. St. Pölten, Amt der NÖ Landesregierung.

Junkelmann, M. (1996) Reiter wie Statuen aus Erz. Mainz am Rhein, Philipp von Zabern.

Junkelmann, M. (2000) Römische Helme. Bd. VIII, Sammlung Axel Guttmann. Mainz am Rhein, H. Bonn.

Klumbach, H. (1974) Römische Helme aus Niedergermanien. Bonn, Rheinisches Landesmuseum Bonn.

Künzl, E. (2008) Unter den goldenen Adlern. Der Waffenschmuck des römischen Imperiums. Regensburg/Mainz, Schnell und Steiner.

Meijers, R. and Willer, F. (2007), eds, Achter het zilveren masker. Nieuw onderzoek naar de productietechnieken van Romeinse ruiterhelmen/Hinter der silbernen Maske. Neue Untersuchungen zur Herstellungstechnik römischer Reiterhelme. Nijmegen, Museum Het Valkhof.

Prittwitz und Gaffron, H. H. von (1991) Der Reiterhelm des Tortikollis. Bonner Jahrbücher 191, 225–46.

Prittwitz und Gaffron, H. H. von (1993) Der schiefe Prunkhelm. In H.-J. Schalles and C. Schreiter (eds) Geschichte aus dem Kies: neue Funde aus dem Alten Rhein bei Xanten, 59–63. Köln/Bonn, Rheinland.

Robinson, H. R. (1975) The Armour of Imperial Rome. London, Arms and Armour.

Willems, W. J. H. (1991) Een Romeins viziermasker van het Kops Plateau te Nijmegen. Jaarboek Numaga 38, 9–18.

Willer, F. and Meijers, R. (2007) Onderzoek naar de productietechnieken van de ijzeren gezichtshelmen uit Nijmegen. In R. Meijers and F. Willer (eds) Achter het zilveren masker. Nieuw onderzoek naar de productietechnieken van Romeinse ruiterhelmen/Hinter der silbernen Maske. Neue Untersuchungen zur Herstellungstechnik römischer Reiterhelme, 31–50. Nijmegen, Museum Het Valkhof.

Chapter 2

If you go down to the woods today… A rare item of Roman horse gear from the Dutch–German border

Clive Bridger with a contribution by Frank Willer

There can be few friends and colleagues in our quirky field of research so obviously non-military as the subject of our celebratory offerings. And yet, in some far-off time in a former millennium, even the genteel Carol van Driel-Murray delved into the inner echelons of Roman military affairs – nay, she was even so carefree as to organise a conference in Nijmegen on military equipment and edit its resultant proceedings (van Driel-Murray 1989a). In that publication she herself published a paper on horse equipment from Vindolanda, albeit mainly concerned with her beloved leather (van Driel-Murray 1989b). In much the same way, this aberrant author has hitherto managed to survive a life in archaeology without too much direct contact with Roman militaria, but all good things must come to an end. In the following paper I wish to present an almost unique piece of Roman horse gear, the existence of which came about from a cross-border joint-venture in the collegial manner so much enjoyed by the author over the years with friends and colleagues in Leiden and Nijmegen.

Until recently the author was responsible for all new archaeological finds that happened to be found in the Lower Rhineland, i.e. all chance finds accruing from non-excavations, including those from (registered) metal detectorists. In September 2012 a Dutch metal detectorist contacted the Museum Het Valkhof in Nijmegen with the news that he had discovered an unusual bronze object. When asked by Louis Swinkels for some details, he received a few photographs and the information that it had been found in the Reichswald, i.e. on German territory (Fig. 2.1). The finder, who lived in Heythuysen (Limburg), had no authorisation to search in the forest, which is legally forbidden in wooded areas in Northrhine-Westphalia, and this was not the first time he had been looking for – albeit modern – militaria in this area.¹ Indeed, the new object had already been discussed in an internet forum (http://www.romanarmytalk.com/rat/forum-8.html from 19 August to 6 September 2012). Via Swinkels and his colleague Stephan Weiß-König the author received more detailed information on the find. It took almost a year to arrange a personal meeting with the finder, who was well aware of the illegal way he had acquired the object. This finally took place at the end of 2013 in the former Museum Kam in Nijmegen, where I could provisionally document the object; it was registered under the activity NI 2012/0105 of the Amt für Bodendenkmalpflege im Rheinland.²

The finder reported that he was looking for Second World War militaria, when northeast of a forest clearing in Goch-Nergena, Kreis Kleve he uncovered the find at a depth of 0.3 m (or 0.6 m according to the internet forum) in a sandy layer without any further finds or features. After its discovery the artefact had been neither cleaned nor restored, the edges displayed some signs of recent dehydration, but the metal was generally in a good condition, having lain for centuries in forest sand bereft of any agricultural fertilisers (Fig. 2.2).

The copper-alloy artefact comprised two separate pieces joined together by a hook-and-loop connection. Possible modern-day manipulations could not be observed, i.e. the object was such as it had been deposited in the ground. The whole assemblage weighed 301 g. Since March 2015 the find has gone on loan to the LandesMuseum Bonn (no. F2015/5), where it was restored, allowing further details to be observed, which I documented in April 2016. My thanks go to Frank Willer for his expertise in the restoration and for his following note.

Figure 2.1 Map with find-spot. C. Bridger using the Software MapInfo 8.0

F. Willer (LVR-LandesMuseum Bonn) writes:

During the restoration in the LandesMuseum Bonn soil and corrosion adhering to the original surface were removed. Parts of the smaller part had been lost or were broken. These were joined and missing parts added according to comparative finds, in order to make a suggested reconstruction. Despite its age non-intrusive conductivity measurements of the alloy using eddy current methods resulted in values of 12 megasiemens per metre (MS/m). Based upon reference measurements of ancient copper alloys, this indicates without doubt an alloy of brass (Willer 2014). For a more exact analysis a small bore probe was extracted from the middle of a break. An investigation at the Curt-Engelhorn Centre for Archaeometry in Mannheim using X-ray fluorescence analysis (RFA) resulted in the clear definition of a special brass or complex alloy with 77 per cent copper, 21.8 per cent zinc and 1 per cent tin. As experiments have shown, even minimal additions of tin within 1 per cent can positively influence the characteristics of the alloy (Schmauder and Willer 2004). Accordingly, similar alloys are still produced today. The lack of lead within the 1 per cent range indicates a fresh product, by which no additions of old metal (bronze) occurred, as was often the case. Both parts were cast in the lost wax technique and then worked into its present shapes. Contrary to the Roman technique of smelting bronze, the production of brass casts was very complicated. However, the particular characteristics of the brass alloy, such as the extremely good workability and the ability for polishing the metal to a golden colour, seem to have justified the effort.

The smaller part is multi-angled and is linked by a loop around a hook attached to the larger piece. The loop is 5–6 mm thick, 2.1 cm wide and terminates in two flatly hammered lugs held by two flatly hammered rivets on the underside of a 3.8 cm wide and 10.2 cm long, rhomboid plate (Fig. 2.3). Both rivets appear on the upper, visible face as two round heads with diameters of 0.8 cm and 1 cm. Behind them is a third rivet-head also with a diameter of 1 cm, but today without any visible object attached to it on the underside; one can assume, therefore, that an organic material is missing from here, very probably leather. The rhomboid plate is convex and tapers on both sides to 1 cm, before both turning at right-angles and transmuting into two shanks. One shank is broken off, but the other is almost entirely preserved to a length of 13.6 cm. The latter has a width of 0.9–1.0 cm and a thickness of 2–2.2 mm, is slightly convex and at the end of the upper angle bends again at right-angles to the front, but is then only preserved for 1.7 cm. On each of the upper sides of both shanks is a cast roughly D-shaped loop, whose outer diameters measure 2.7–2.8 cm, the inner ones 1.8 cm, the thickness is 4.2–5 mm. The intermediate space between the two is 10.5 cm, which is the maximum width, the shanks narrowing to 8.9 cm. At the transition to the loop the thickness of the shank increases to 2.8–3.0 mm. The conserved piece has now been reconstructed with the length of the lower shanks measuring 14.2 cm before terminating in a semicircular band of 6.6 cm in height (Plate 2).

Figure 2.2 Find in its unrestored state. Photo: R. Meijers, Museum Het Valkhof, Nijmegen

Figure 2.3 Reverse of the restored metal bridle. Photo: J. Vogel, LVR-LandesMuseum Bonn

Figure 2.4 Connecting hook of the restored pendant. Photo: J. Vogel, LVR-LandesMuseum Bonn

Since the 1970s this type of artefact has generally been known in Roman archaeological circles as a hackamore, although, since the hackamore proper operates without a bit, the term psalion (Gr. ψάλιον; headpiece) is preferred by equine experts (Junkelmann 1992, 24–33; Hyland 1992, 75–8; 1993, 63–4; Găzdac 1999, 743; Hyland 2013, sections curb and psalion and psalion; Fahr 2015, 131, n. 1; Bishop 2015, 380). It comprises the rhomboid noseband with lateral shanks and loops, to which reins or a longe could be attached in the way of a cavesson and was combined with a bit. Since the Roman examples were also provided with a band running under the chin, they do resemble modern-day hackamores, which, though, are overwhelmingly made of leather or rope and are more akin to leather drop nosebands. The metal plate applied pressure to the animal’s sensitive nasal bone, i.e. it was among others things an effective method for training young, work-shy horses. In connection with a lower curb band, to which knots could be fastened, which could have had a painful effect on the sensitive parts around the snout, this could have had eine recht bösartige Hebelwirkung zur Folge (Junkelmann 1992, 29). Hackamores and similar bridles allow riders to control their mounts with just a couple of fingers, giving them the freedom to deal with other activities while mounted. However, with a metal noseband this necessitated a very close co-operation between rider and horse, as the metal bridle could more easily inflict pain upon the animal, when compared with conventional ones (Junkelmann 1992, 21–3; cf. also Schwinden 1987). The further the horse tried to escape from the constraints of the bit, for instance by opening its mouth, the more pressure was applied by the metal plate on the nasal flesh.

Numerous examples of psalia are known from the northwestern provinces.³ In our own case only a part of the object has been preserved, for the lower portion of the cheek-shank, the lower angles and the chin-bar or chin-strap are missing.⁴ This specimen seems to be petite, since the maximum distance between the two bridle loops measures just 10.5 cm, but this actually corresponds to the average of such nosebands (Simon Ortisi 2005, 20–1). Such bridles date from the Augustan period to the middle of the third century AD. Our item with the rhomboid noseband can be attributed to the type 3 or 5 of Annabel Taylor (Lawson 1978) or group III.A with undecorated lozenge according to Christina Simon Ortisi (2005, 25–6; also Simon 2014, 21). However, the latter includes a special form III.B.4, which the authoress described as a Raute mit Zierknöpfen (Simon Ortisi 2005, 40, 160, 210, fig. 62; Simon 2014, 31 with 32 fig. 30.) This form is not discussed in her analysis and only one example from the north-western provinces with a terminus ante quem of AD 69 is mentioned in her catalogue (Augsburg, Kornhausgasse 4, Römisches Museum Augsburg, inv. no. 1985,722). Photographs of this find from Augsburg show, however, that a strap with loop was once present on the back,⁵ i.e. one can assume that an appliqué backed with an organic material had been present here, too. The new find has now made it apparent that the Augsburg piece is not a special case, but an example of her group III. A. The rivet-heads do not represent a decoration, but mainly served the fastening of a larger part.

The larger part of the new find is made of brass consisting of a 1 mm thin, slightly convex, hammered plate, whose preserved measurements are 16.2 × 10.1 cm, whereby the original edges are hardly preserved and only present in the central portion. On the underside there are remains of a tin coating, the upper side has no decoration. At the mid-point of the outer rim is a round terminal knob with a diameter of 1.6–2 cm, which merges into a 0.45 cm wide rib on the underside and thence terminates in a hook in the form of a wolf’s or dog’s head (Fig. 2.4). The 2.0 cm high and 3.6 cm long head faces the terminal knob and resembles a wolf or a greyhound with ears pinned tightly back; on top longitudinal lines could indicate lines of the animal’s coat. A connection between a horse and a wolf might be that both are attributes of Mars, as well, of course, as Lupa representing the allegory of the founding of Rome.

The plate was originally in the shape of a figure of eight and corresponds to the well-known type of the double lobed or so-called winged pendants, the analysis of which has mostly been done by Eckhard Deschler-Erb (1998; previously Bishop 1988, 150–1 with table 6). He described the finds, which occur between Central Spain, England, Potsdam (!) and Greece, as winged pendants with hooks in the shape of heads of waterfowl or canines, which are mostly adorned with grooved or punched decoration. Our example corresponds to his type B in Augst (Deschler-Erb 1999, 50–1, 161–2 with pls 22–4 nos 438–500. Bishop type 7a: Bishop 1988, 149 fig. 46). With former measurements of at least 16.8 × 10.1 cm it belongs to the largest category, of which only eight examples are known (Deschler-Erb 1999, 116–7 and another new example dealt with below). The largest specimen ever found comes from an extraction pit in Kerkdriel-De Zandmeren (NL), i.e. again without an archaeological context, and measures 20 × 15.2 cm (Nicolay 2005, 59, 440 pl. 87 no. 158.1; 2007, 57, 398 pl. 87 no. 158.1). The second largest example comes from the early military phases in Cirencester (UK), measures 18 × 11.2 cm and closely resembles our specimen; in addition, it has a lug with two flatly hammered rivet-holes connected to it in the same way as the new find (Webster 1982, 109 with 110 fig. 36, no. 100; Bishop and Coulston 2006, 120 fig. 70, no. 9). Based upon Deschler-Erb’s considerable amount of material, he dates this type mainly to the second third of the first century AD, after which it fades out during the Flavian period (Deschler-Erb 1999, 51). For instance, at Vindonissa (CH), which was mainly garrisoned between AD 14 and 101, 36 examples have been found (Unz and Deschler-Erb 1999, 39–40, pls 49–50, nos 1329–64, where, however, only three psalia were listed: 50 with pl. 68 nos 1941–3). In Germania inferior the new find corresponds to the type A3 of horse-pendants in the Batavian region, which Johann Nicolay (2007, 55) dates to the first third of the first century AD. However, the one example mentioned above is the only one of this type within his 101 horse-pendants from the Batavian area, a fact which makes his dating somewhat tenuous. Deschler-Erb (1998, 119) also lists only a few examples from the Lower Rhineland: two from Moers-Asciburgium (DE), one from Xanten-Vetera I (DE), four from Xanten-Colonia Ulpia Traiana (DE) and one from Nijmegen (NL).

In summary, therefore, we are dealing with an artefact from around the mid-first century AD which is identifiable as a metal noseband bridle (Gr. ψάλιον) with an attached pendant which covered the horse’s nostrils (Plate 2).⁶ Hitherto, such pendants were associated with horses’ chests (Gr. προστερνíδια), as depicted on numerous sculptures, in particular the tombstones of cavalrymen, where they are assumed at least partly to have functioned as apotropaic decorations. Before the discovery of this object, these pendants had never been brought into connection with bridles and, as far as I am aware, this was the first example of such a pendant found connected to a bridle to be published (Lat. frontal, Gr. προμετωπíδιον). Consequently, we now seem to have a new function for this type of pendant, which had not appeared previously in academic literature. For instance, Deschler-Erb (1999, 49) was of the opinion that horse-pendants possessed no practical function. A check through the volumes of the Corpus Signorum Imperii Romani showed that as a result of the poor preservation of the materials many equine heads show insufficient details, whereby some could well have originally displayed painted ornaments (cf. Garcés Estallo and Graells Fabregat 2011, 23). A higher degree of detail is provided by bronze horse heads, which, however, often show decorative elements above, but rarely below the bridles.⁷

The first illustrated examples from Annabel Taylor (1975, pl. 53 no. 2 from Heddernheim (DE) pl. 57 no. 1 from Weißenburg (DE) pl. 59 no. 1 from Ruvo di Puglia (I) and pl. 63 no. 4 from Valkenburg (NL)) show bridles which were capable of attaching such pendants through loops, eyelets or holes. All pendant plates in Augst (CH) and Vindonissa (CH) are shown without loops or other parts, so that no direct link to bridles exists there. Several metal nosebands from the Bay of Naples display perforations or pieces missing, but these could also have served the attachment of strap-loops on the underside (cf. Simon Ortisi 2005, pls 13–5, 17–9, 22, 36; Simon 2014, pls. 2, 10). One from Pompeii, though, does still bear a small pendant (Simon 2014, 173 with pl. 25 (PH 17a, b)). Two roughly similar comparisons to the new find, at least with a corresponding loop, are the above-mentioned ones from Cirencester and Augsburg. With the former it was mooted that it might reflect the find of a Thracian cavalry unit (Bishop and Coulston 2006, 261).⁸ The object from Augsburg is too fragmentarily preserved to say definitely whether a pendant was attached or not, although a corresponding connecting strap-loop might suggest this.

Figure 2.5 Psalion from the Collection C. Giannelli. Photo: M. Ostini, CH-Lugano

In the spring of 2015 the missing link appeared, when an exhibition of historical equestrian finds Cavallo: storia, arte e artigianato was held in Travagliato near Milan (I). It comprised objects from four millennia from the private collection of Claudio Giannelli of Lugano in Switzerland, and consisted almost entirely of horses’ bits. The superbly produced book accompanying the collection was published that autumn.⁹ In it there is a picture of an almost complete psalion of about 12.5 × 20 cm above which, i.e. towards the animal’s eyes, there is a series of three, overlapping double-lobed pendants with their flattened terminal knobs pointing down towards the snout (Fig. 2.5; Cascarino 2015, 142 fig. 29)!¹⁰ The uppermost pendant is attached to a c. 20 cm broad cross-band, the ends of which are broken off. The attachment is performed by a rectangular hole in the cross-band, through which a loop runs ending in a canine-headed terminal facing downwards. The three pendants measure approximately 17 × 9 cm (terminal knob missing), 14.5 × 13 cm and 12 × 9 cm, so that the first also belongs to the largest group of such pendants. This whole assemblage, which is roughly 29 cm in length, would correspond to a prometopidion. The front of the psalion’s rhomboid plate is missing, which allows one to speculate, whether here a further pendant akin to our own find originally lay over the nose. Unfortunately, in the publication neither a provenance nor a description of the piece is given and the writer of the relevant chapter places it in the milieu of the later cataphracti, which is clearly incorrect (Cascarino 2015, 141). Personal contact with Sgn. Giannelli revealed that he acquired the piece at an auction of ancient armour from the Axel Guttmann collection, whose famous objects mainly comprised pieces bought on the antiquities market or from illegal sources, especially in the Balkans. Consequently, this stunning and unique find is bereft of a provenance.¹¹

At present, we seem to have only two examples of Roman frontalia/prometopidia comprising winged pendants attached to psalia. We can assume, therefore, that providing a horse’s head with such an additional nose-plate seems to have been very unusual, or at least this must have occurred in a situation which was deemed not worthy of visualisation, since there seem to be no representations of such nasal pendants on any equine statuary. When would such a provision be pertinent? For example, we could imagine a festive decoration of the animal for a parade or its protection in training or for the cavalry games, the hippika gymnasia (Arrian Ars Tactica in Roos and Wirth 1968, 34–44). It is known from both archaeological and literary sources that horses’ eyes were protected at such games.¹² The forehead could also be covered with a chamfrain of leather and metal, which sometimes also extended down to the nose, as on the well-known example from Newstead (UK) (see also Junkelmann 1992, 202–10; Schuckelt 2014, 13–6, 42–103 and most recently Schamper 2015, 53 f, 227–34 with pls 48–59 (58 ex.)). Hitherto, however, such a protection or an adornment for the nostrils was unknown, providing we ignore muzzles, which mainly served the protection of humans and other animals from aggressive horses (Garcés Estallo and Graells Fabregat 2011).

It remains to discuss whether the given find-spot in Goch-Nergena seems feasible. The finder was well aware that he had contravened several German laws and regulations regarding trespass, theft and heritage law. Consequently, if he had been interested in developing an alternative crime scene, it would have been more logical for him to have chosen another find-spot over the nearby border on Dutch territory. Nevertheless, the find-spot does remain problematic, since a proper archaeological context is lacking. Owing to the use of the terrain as a royal, then a state forest for several centuries, the archaeological documentation of this area has been difficult.¹³ Despite this, several Roman sites are known here and the new find-spot lies just within the 8.5 × 7.5 km large area investigated by Erika Riedmeier-Fischer (1998, 160–1, 280–1), who was able to establish 35 Roman sites here, but hardly any of the first century.¹⁴ In the south of this area in Goch-Hassum, some 7 km from the find-spot, there is a new site discovered by a registered metal detectorist, which contains numerous metal finds of the first century AD, among which are many military finds (unpublished; ABR NI 2011/0099, 2011/0232, 2011/0235, 2011/0349, 2012/0214). In 1982, some 5.5 km northwest of the find-spot, amongst other features a well with a dendrochronological date of AD 22±5 was uncovered, which indicates a further early Roman settlement here (ABR NI 1982/0011).¹⁵

We cannot be certain that the new object presented here is attributable to the Roman military, but this is probable. Perhaps we could imagine that the horse and its rider may have been attached to one of the early Roman units stationed along the nearby limes section between Xanten and Nijmegen. How the piece of harness came to where it was found, be it as the result of a riding accident, an aggressive act, an unhappy loss or possibly for the recycling of old metal, must remain hypothetical, as long as the site remains uninvestigated. However, together with the second unprovenanced find, we hope that it may provide impetus to look at such pendants in future in a different light.

Notes

1Cf. The Calgary Herald from 4 May 2007 on him finding and handing back a bracelet lost by a Canadian soldier in World War II.

2My thanks go to both colleagues for arranging the meeting, as well as to Ronny Meijers for initially photographing the find. For a provisional publication of this object in German see Bridger (2015).

3Cf. Taylor (1975); Lawson (1978); esp. Simon Ortisi (2005, 155–61, 275–86), who lists 63 examples from the towns on the Bay of Naples, as well as 111 specimens from the rest of the Roman Empire. Since then: ten from the Batavian area (Nicolay 2007, 45 with pls 49–50); Bürghöfe (Franke 2009, 38 with pl. 31 no. 681); Nijmegen, Kops Plateau, find 270/008 (unpublished); Krefeld-Gellep (Fahr 2015). There are no examples in the immediate area of the Rhine delta (van Roemburg 2011, 40). Now Simon (2014, 11–55).

4In June 2014 I was able to discuss the find personally in Salisbury with Dr Annabel Lawson, née Taylor, whom I would like to thank very much indeed. Using her considerable unpublished archive she could prove that the correct way of wearing the Roman bridle corresponds to that described by Littauer (1969, 291–2), whereby the loops face towards the eyes and the lower angle towards the nose, in order to hold a bit; cf. also Palágyi (1989, 132 with fig. 6a, 135 with fig. 9.3); Junkelmann (1992, 27 fig. 21, 28 figs 25–6, 31 fig. 28b–c), as well as method B of Lawson (1978) and Simon Ortisi (2005, 215 with fig. 81 right, 217 figs 96–7); Simon (2014, 37–52).

5My thanks are due to Frau S. Schneider M.A. from the Museum in Augsburg for providing photographs.

6A possible opposite direction upwards towards the forehead is illogical and is also contradicted by the method of wearing the lower chin-band towards the mouth.

7E.g. Junkelmann (1992, figs 183–4); Rasbach (2014, 40–3). The head from Augsburg, which belonged to a harnessed team of horses, displays a leather strap-end on which a lunulate amulet hangs between the nostrils: Hahn (2014, 49 with 51 fig. 1).

8Cf. the tombstone of the Thracian (Dansalian) cavalryman Flavius Bassus in Cologne, whose horse is also shown with a metal bridle (e.g . Junkelmann 1992, 77 fig. 84).

9The book C. Giannelli (ed.), Equus frenatus. Morsi dalla Collezione Giannelli is not on sale commercially, but can be bought directly from the editor by sending an email to him: [e-mail].

10 My sincere thanks go to Sgn. Giannelli for providing the measurements and further information, as well as a copy of his sumptuous book.

11 Christie’s auction London, 28 April 2004, "The Axel