JUST a mile down the road from where I live in Wales is a Welsh place-name, Rhydypennau, “Ford of the Heads”. According to local tradition, a battle was fought here against the Romans in the first century, and the heads of those slain were thrown into the ford by the Britons. The ford is situated on the Nant Ceiro, “Stag Stream”, which runs through our grounds. The ford is situated at the meeting of three parishes Tirymynach, “Monks’ Land”, for the terrain belonge to Strata Florida abbey; Ceulanamaesmawr; and Geneu’r Glyn. Nant Ceiro is a boundary stream.

Boundaries and thresholds were of great significance in early Celtic life and religion. For example, one of the important duties of the Druids would seem to have been the fixing and maintaining of boundaries. Caesar says as much in his De Bello Gallico (“Gallic Wars”):

For they (the Druids) have the right to decide nearly all public and private disputes…and disputes concerning legacies and boundaries (Tierney 1960, 271, 13).

Fords, in early Celtic tradition both continental and insular, were places of single combat and of sinister portents. The motif of The Washer at The Ford is well known in Celtic lore. In Welsh tradition, for example, Urien Rheged goes to Rhyd y Gyfartha, “the Ford of Barking”, in Denbighshire, to find out why hounds always barked there. No one else had been bold enough to investigate. Reaching the edge of the ford, Urien can see nothing but a woman washing. The hounds stop their barking, and Urien seizes the washer-woman and has intercourse with her. She tells him to return at the end of a year, when he will receive his son. When he goes back at the appointed time the woman presents him with a son and a daughter, namely, Owein son of Urien and Morfudd daughter of Urien (Bromwich 1961, 459). There are Irish examples of this motif, while the name of the River Clyde in Scotland – Clota, the name of a goddess meaning “the washer” – mentioned by Tacit-us, suggests an early date for the motif.

Fords were places of single combat, rivers clearly forming boundaries between tribal terrains, and thus they acted as neutral territory. In medieval Irish texts single combat between heroes regularly took place at fords. The boy-hero Cu Chulainn, in the epic tale Tain Bo Cuailnge, who protects Ulster from Connacht in the first century B.C. according to tradition, comes to a ford. There he proceeds to cut down the forked branch of a tree with a single blow of his sword. He then drives it into the middle of the stream so that no chariot can pass over the stream. While he is engaged in this work two warriors from the enemy and their two charioteers approach the boy. Single-handed, he cuts off their four heads and impales them on the four prongs of the forked branch. He then writes an inscription in Ogam down the side of the great branch.

The ensuing remarks of the hero, Fergus, make it clear that this is indeed a boundary stream. When someone asks him who could have performed this amazing deed he replies: “The man who could have performed this deed is Cu Chulainn…it is he who would have come to the boundary (criche) accompanied only by his charioteer” (O’Rahilly, 1976, I, 341 ff.). This warlike episode would have applied equally to the Gauls, as attested by the classical commentators. The heads were left at the ford.

This epic tale of the deeds of Cu Chulainn and of how he left the decapitated heads of his enemies at the ford which marked a boundary would seem to cast some light upon the discoveries of human skulls in watery places where they had been deposited in antiquity. I shall return to this theme of severed heads and sacred waters in due course, but first I want to look at some prime pieces of evidence for the cult of the human head itself in antiquity.

We have very good grounds for believing that the severed head was regarded by the Celts as a sacred the classical symbol: writers testify to as much. I have discussed this topic in some detail in my earlier publications (see bibliography); for the present purpose it will serve to look at some new or comparatively new discoveries in Europe and in the British Isles.

An important group of sanctuaries is situated in Celto-Ligurian territory in south-east France in the vicinity of the Rhone delta, and dating to the pre-Roman period. All are within easy reach of Massalia, the Greek trading post established circa 600 B.C., which opened up the Celtic world of the hinterland by means of trading up and down the Rhone. The architecture of the Celto-Ligurian temples is sophisticated, and the cultural influence of Greece is evident. Thus it is remarkable to find that here, where access to classical civilisation was easy, the cult of the human head flourished. One is reminded of the presence of the archaic-looking heads which abound in early Christian churches and Norman cathedrals, the baleful albeit protective gaze of which strongly belies the benign quality of the Christian faith. Some of the best-known of these edifices are at Roquepertuse, situated some twenty miles inland from Massalia. The temple was cut out of limestone, and a great janus-head, painted in red and black, was surmounted by a goose-like raptor of which only the beak has survived.

This is echoed by the large sculpted goose which surmounted the lintel supported by pillars in which head, painted in red and black, was surmounted by a goose-like raptor of which only the beak has survived. This is echoed by the large sculpted goose which surmounted the lintel supported by pillars in which were cut niches for heads or skulls. Forensic analysis has shown that the heads had all belonged to young men in the prime of life (Piggott 1968, 56).

Another temple, associated with a spring and a cave- sacred places for the Celts – was situated at Glanum, near Saint-Remy-en-Provence. Here a lintel with skull- niches was found. This temple, re-used in the second century B.C., is likely to have had an early origin, as may well have been the case with some or all of the other temples. Here too, in the first century B.C. Celtic deities such as Sucellos and the horse-goddess Epona were venerated; the association between horses and sacred waters is well-attested for the Celtic world. The god Glanis was the god of the shrine, “the pure or clean one”. Glan means pure, holy, clean in both Welsh and Irish. (Salviot 1979).

At Mouries, Bouches-du-Rhone, remains of an earlier stone-built sanctuary had been incorporated into Gallo-Greek structures of the fourth century B.C. Here pillars were decorated by figures of horses and riders, and the head cult is likely to have flourished here too.

Another sanctuary at which an earlier temple had been incorporated into a fourth-century edifice is at Saint-Blaise. Here the jamb of a doorway had niches for severed heads or skulls cut into it. Perhaps the most on the northern outskirts of Aix-en-Provence. Here the threshold incorporated a re-used pillar on which are carved in outline twelve mouthless human heads, all but one of which are upright, the other inverted. Here also is the carving on a stone slab of a schematic human head flanked by niches for the display of the real thing. Some fifteen skulls of adult males were recovered here; some of them still retained the large iron nails by which they must have been fixed or suspended from some wooden structure. There was also a series of stone sculptures of squatting warriors, some holding a severed head in one hand and a lance or spear in the other. The temple must have existed as a sacred place well before its destruction in the year 123 B.C.

Hillforts, too, boasted their own quota of skull- trophies, set up, no doubt, as guardians, as in later times. A skull was set meaningfully in the wall of L’lmpernal, the oppidum of the Cadurci, situated near Cahors, Lot. At Bredon hillfort, Gloucestershire, skulls had decorated the lintel of the gateway. When this was set on fire during an attack early in the first century A.D. the heads came crashing down. The head cult is illustrated, too, by a sculpture of a sinister nature from Noves, situated – significantly perhaps – close to the River Durance, the Druentia, a tempting name. The sculpture is of a wolf-like creature of fierce countenance, with a human limb hanging from its lower jaw and an object now broken at either end in its mouth; it is powerfully ithyphallic. Known locally as “le Tarasque de Noves”, the sculpture represents a monster which, according to medieval legend, used to emerge from the river and devour people – leaving, it would seem, their doleful heads intact. Two lugubrious male heads are firmly held down, one by each fearsome talon. The monster may well date from the third century B.C. The story is remembered and marked by an annual fete in July known as the fete du tarasque. Paper replicas of the monster are carried through the streets of Avignon and, from what I was able to gather from an informant who had witnessed the fete, the actual sculpture was sometimes displayed.

The legend would serve to link the tradition with the river, and the whole may be based on the discovery of actual skulls in the water (see below). (Megaw and Megaw 1986, 170).

The classical writers amplify the evidence of archaeology in their descriptions of the custom of head-hunting by the early Celts. There can be little doubt that the taking of heads was an essential way of life and an integral part of Celtic religious belief and practice up to the Roman conquests, first of Gaul and then of Britain. Thereafter the taking of heads was banned, as were human sacrifices; but later literary and archaeological evidence demonstrates how deeply- rooted in the tradition this practice must have been. On the Arc d’Orange, Vaucluse, one of a series of triumphal arches set up by the Romans to mark their conquests of the troublesome Celts, the latter are portrayed with human heads hanging from their saddles – just as Diodorus Siculus describes them. Scalps also are depicted.

Heads were preserved with herbs and oils, and kept in chests, by the Gaulish nobles who would not part with them for any sum of money. Their apotropaic and protective powers were clearly appreciated from a very early date: I shall return to this point later. That the skull was used as a drinking cup in solemn circumstances, e.g. in temple rites, is likewise made clear by the classics. Livy, for example (born 59 B.C.), describes how the consul Postumius was lost in Gaul together with his army in a huge forest called Litana, “The Broad One” (early Irish lethan, modern Gaelic leathan). The Gauls ambushed and surrounded Postumius’ two legions and virtually destroyed them. The consul died fighting, spoils were taken from his body and, predictably, his head was severed, and it and the booty were taken by the Boii, the Celtic tribe in question, to their “holiest temple”. “Then, after they removed the flesh from the head (excarnation of skulls and of bodies generally is well attested in archaeological work and must have been widely practised) they adorned the skull with gold according to their custom. They used it as a sacred vessel to give libations on holy days, and their priests and the custodians of their temple used it as a goblet” (Koch & Carey 1994, 31). The libations were presumably sacred water from the springs which served the temples; blood, perhaps, from sacrificial victims; and on occasion the mead or ale which were the favoured drinks of the Celts, although wine-drinking early became popular when supplies became available.

Many rites connected with heads are attested from sacred sites in Gaul and the British Isles in the pre- Roman period. One or two of these can be mentioned in this context. The sanctuary of Gournay-sur-Aronde, a late Iron-Age Gaulish sanctuary situated in Belgic territory, is one of an important series of temples in this region of northern Gaul. The importance of the entrance – the ritual significance of the entrance to such structures is well attested – was here stressed by the erection of a huge portico above the gate. This seems to have been closely similar to the portico at the Rhone temple at Roquepertuse. Here too, at Gournay, human skulls were displayed (Brunaux 1988, 27). The evidence lies in the cervical vertebrae and the incisors that were rcovered; these had become detached and had fallen as the heads rotted. A skull found in the ditch probably originally came from the portico. On either side of the entrance were two huge piles of the skulls of cattle and of weapons (cf. Thames and otherdeposits, infra). The animal heads would likewise have exerted some apotropaic power. Some type of funerary excarnation clearly took place at Gournay, as in Ireland (infra); for example, a skull was cut into the shape of a mask; one thinks of the tin mask found in a drain in the waters of the springs of Sulis at Bath (Aquae Sulis), perhaps intended to fit over an actual skull, or the face of a priest, or that of an idol. Many heads were placed in significant positions in shrines and temples, pits, shafts and wells down the Celtic ages, testifying to the deeply rooted nature of this most Celtic of cults. A few of these will be noted here.

At Odell, Buckinghamshire, the decapitated head of a woman had been placed behind the wicker lining of a Romano-British well, perhaps to protect the waters and ensure a good supply.

At the water-shrine of Springhead, which stands at the head of the Ebbsfleet Valley in northwest Kent, and one and a half miles from the River Thames, traces ofhead-ritual at two periods are evident. The temple ison the site of Vagniacis, which is from a British word meaning a marshy or boggy place – a good description of the site. Four babies were found to have been buried- two on the west side of an early floor and two on the east side of a later floor. This happened in the original room of the Antonine period, and again some ten to fifteen years later in a new floor. In the first floor the baby on the southwest corner had been decapitated: in the later construction the infant in the northeast corner had been decapitated. The excavators supposed that one pair had been offered as a foundation sacrifice, and the second pair at a ceremony of rededication. As is well attested from the dramatic Gournay shrine, constant rebuilding, and thus rededication, must have been a regular feature of Celtic sanctuaries. What happened to the heads is not recorded; excavation did not recover them. It is likely that they were placed in the sacred waters of this impressive Springhead shrine which must have been accorded worship some time before the second century, according to the evidence revealed by excavation.

This is reminiscent of the remarkable Celtic sanctuary at Libernice, near Kolin in Slovakia which dates to the third century B.C. (see Rybova and Soudsky 1962). Consisting of an elongated ditched enclosure, it revealed many traces of quite dramatic