Agr. Tacitus, Agricola

AM Kelly trans., Audacht Morainn

Ann. Tacitus, Annales

Bell. Civ Lucan, De Bello Civili

BG Caesar, Bellum Gallico

CA Cóir Anmann, in Stokes & Windisch 1897, 288-444

CMT Gray trans., Cath Maige Tuired

De Arch. Vitruvius, The Ten Books on Architecture.

De Excidio De Excidio Britanniae, in Giles 1885, 293-380

LM Buhler trans., The Laws of Manu

Met. Dinds. Gwynn trans., Metrical Dindsenchas

Mor. Plutarch, Moralia

NH Pliny, Natural History

RIB Roman Inscriptions in Britain

San. Corm. O'Donovan trans., Cormac's Glosssary

Tec. Corm. Meyer trans. The Instructions of King Cormac mac Airt (Tecosc Cormaic)

The interpretation of Romano-Celtic iconography presents major problems. With no indigenous literate statement of belief, archaeologists are forced to rely upon contemporary Latin authors, inscriptions, and analysis of the symbolism. Although these methods can provide valuable information, they also present problems. Classical authors, such as Tacitus, interpreted Celtic beliefs according to their own cultural framework (Germ. 43). The reverse is also true; the Jupiter-Giant columns, for example, appear to use Roman art forms and technology to express non-Roman ideas (Green 1976, 8-10, 119). Interpretation distorted those ideas, while the adoption of Roman concepts by indigenous religion altered the latter to some degree (Derks 1998, 105, 107).

The same pattern of underlying non-Roman thought, and the problems of understanding it, can be seen in name dualling. The limited range of Roman deities used might suggest that a limited number of indigenous, multifunctional deities were described by many regional epithets (Webster 1986, 54; Green 1986, 32; Derks 1998, 95). Some representations are alien, despite name-dualling, suggesting that Classical ideas were not uppermost in the artist’s mind. Presumably the artist’s efforts found favour with the dedicator, suggesting a common symbolic language that had little in common with Roman norms. Consequently the typical function of the Roman god, as we understand it, may not necessarily be an accurate indication of the function of the Celtic deity, and it is of no help with those deities unassociated with a Roman divinity.

This dissertation explores the possibility of interpreting the iconography using the inter-related Irish and Welsh sources (O'Curry 1861, 472; Gruffydd 1953, 8-9, Mac Cana 1958, 1, 16-101). Early Irish societies had very strong direct links with the former La Tène area from the fourth century BC onwards, with probable additional Iberian contacts (Raftery J 1972, 5-6; Raftery B 1972, 49-50; 1989, 141; MacMullen 1990, 46-48). The name of the religious site of Emain Macha is almost certainly cognate with the Gaulish deity Imona mentioned in the Rom inscription (Olmsted 1988, 317). Emain Macha was in use from the seventh to the first centuries BC; its closest parallels are the sixth century BC enclosures at Goloring and the Goldberg in Germany (Raftery B 1989, 141; Green 1989, 151-152; 1986, 20). Irish sources also suggest connections with continental iconography, while Irish stone idols suggest continental and British artistic influences (Rynne 1972, 79-99; Ross 1967, 153, 337-338; Green 1984, 137; Mallory 1984, 24).

Although they are late, the Irish sources preserve much that has been convincingly demonstrated as archaic. For example, Audacht Morainn was compiled in approximately AD 700, but it probably belongs to a very much earlier date (Kelly 1976, xxxiii). The Botorrita inscription echoes some of its passages, and this is no later than the first century BC (Olmsted 1988, 365). Exact parallels for Audacht Morainn are also found in the Vedas (Dillon 1947, 6; 1975, 128). It may therefore be possible to use these archaic elements to detect common patterns underlying the regional variation of Celtic societies.

Audacht Morainn is concerned with rulership, which in Ireland was expressed as kingship. This was apparently central to Celtic pre-Christian religion (Dillon 1973a, 1; Oosten 1985, 21). Irish kings married the goddess of the province, and maintained this relationship by demonstrating Truth (Tec. Corm. 1.7, 10, 45, 46; Met. Dinds. IV, Carn Máil 134-143; Dillon & Chadwick 1967, 87-88; Dillon 1972, 2; 1973a, 2; Oaks 1986, 78; Kelly 1988, 18). Truth is a supernatural, masculine, life-giving, ordering force that is equated with heat, and which maintains the fertility and prosperity of the kingdom (AM 17-21; Dillon 1947, 6; 1973, 16; Kelly 1976, xvii; 1988, 18). The behaviour of the king, described by one source as a god, was therefore dictated by the necessity of mediating the divine by demonstrating Truth (Stokes & Windisch 1891, 209-211, 214-215; Cross & Slover 1936, 505; Dillon 1972, 7). Collective social behaviours were dictated by his perceived success in doing so (Dillon 1975, 113). The presence of this concept in other Celtic societies may leave archaeological traces of similar behaviours.

An Irish king, by his Truth, was expected to increase communal prosperity (Tec. Corm. 1.19, 23-24, 46; AM 13, 17-21). He also demonstrated his possession of the source of fertility by his largesse, and was eulogised for his generosity by the poets (AM 59; Posidonius 67). As in pre-conquest Britain, resources were apparently controlled by the elites, and imports were probably used to enhance social ties (Tec. Corm. 1.25-26; O’Donovan 1868, 111; Kinsella 1969, 28; Haselgrove 1982; 1984; Millett 1994, 34; Cunliffe 1995, 23).

Irish kings also demonstrated Truth through warfare (AM 15, 30). Warfare, or rather raiding, was an essential basis of society in both Ireland and Gaul, since military dominance increased communal status and prosperity (Dillon & Chadwick 1967, 89; Derks 1998, 45-46). Inferior communities supplied tribute and hostages to the dominant king, although theoretically he had no authority over the subordinate territory (Tec. Corm. 2.34; Hogan 1931-32, 204; Dillon & Chadwick 1967, 89). Conversely, a king who lacked hostages was not a king. He was unable to enforce any law or alliance, since hostages were an integral part of the legal system. He therefore lacked Truth (Kelly 1988, 18-19).

The issue of tribute suggests potential areas of conflict between Rome and Celtic communities. Both Tacitus and Dio infer that complaints about injustice and excessive demands were behind the Boudiccan revolt (Agr. 15; Ann. XIV, 31-32; Dio LXII.3.1-5). In Cath Maige Tuiredh communal prosperity is reduced by excessive demands for tribute (CMT 25, 36, 39, 45). The failure of fertility was a failure of natural law that indicated the ruler’s lack of Truth (falsehood), which was identified with injustice and illegitimate power (AM 17-21; CMT 46-48; Cross & Slover 1936, 99-100; Kelly 1976, xvii; Kelly 1988, 20-21). Warfare was therefore an inevitable, highly ritualised process of divine judgement, because a ruler guilty of falsehood could not win battles (AM 30; CMT 129; Dillon 1975, 131; Gray 1982, 106-107; Kelly 1988, 18; Derks 1998, 45-46). Rulers were equally at fault, and equally vulnerable, if there was unseasonal weather, plague, or failure of agricultural yields (AM 12; Dillon 1975, 130-131; Kelly 1988, 18). In the mythology such a king would either be deposed or killed (Dillon & Chadwick 1967, 88; Dalton 1971, 2; Dillon 1973, 17). A similar mechanism may be behind populations surrendering Gaulish leaders to Caesar (BG VII, 89; VIII, 38).

The king’s perceived falsehood, whatever its origin, also meant that he could no longer command the respect of allies or subservient peoples. Alliances could abruptly switch or collapse under these circumstances. Defections during the conquest, such as that of the Bodounni, appear to fit this pattern (Dio LX, 21.4-5). Similar patterns are seen in the Roman world (BG IV, 5; 30; V, 39; 283; Birley 1988, 23; Millett 1994, 66, 68).

There are therefore common patterns of behaviour in similar contexts, suggesting common underlying motivations. Interestingly, Posidonius saw the druids as representatives of the Logos, the Greek first principle of the universe (Tierney 1960, 222). The Irish Truth shows marked analogies to both the Logos and the Vedic rta (Radhakrishnan 1923, 78-79; Dillon 1973, 16-18; 1975, 127, 133; Zaidman & Schmitt-Pantel 1992, 143). Given its archaic nature, it would be remarkable if it were not to be found in other Celtic cultures. It might therefore be useful to study the mythology of kingship and Truth, and to compare it with the iconography.

According to myth, the source of Truth is the divine lawgiver, judge and king of Ireland, the Dagda or Excellent God (CMT 81; Met. Dinds. IV, Ailech II, 105, lines 43-44; Ailech III, 111, lines 41-46). His name is paralleled in Iberia by Endovellicus or ‘Good God’ (Green 1987, 790). The Dagda is additionally a warrior, a poet, a harper and a smith; he also possesses an inexhaustible cauldron of plenty (CMT 6, 119, 163-164; Brown 1913, 238; Kelly 1988, 63). Cath Maige Tuiredh supplies an almost iconographic account of the Dagda, his exaggerated genitals exposed, carrying a wheeled club, with his feet shod in horsehide (CMT 93). The association of warrior, wheel and club is interesting; a mould at Cambridge shows an armoured deity, generally identified as Taranis, also carrying a club and a wheel (H J M Green 1986, 39).

The Dagda is one of the chief protagonists of CMT, a mythical battle between the gods (Tuatha De Danann) and the not-gods (Fomor) that shows pronounced similarities with the battle of the Asuras and the Devas in the Vedas (Gray 1982, 1). The Fomor are monstrous submarine giants (Macalister 1938-42, IV, 13; Gray 1982, 132). They lack Truth to the same degree that the Dagda possesses it, and their rule of Ireland was distinguished by barrenness. An affliction of barrenness is also ascribed to the uncontrolled ravages of giant snakes (Met. Dinds. II, Berba, 62-63; Ross 1967, 149-151, 152-154; Green 1993, 58). The submarine Fomor and snakes are therefore functionally equivalent.

Snakes appear to be chthonic in Romano-Celtic iconography (Green 1981, 111; 1986, 185-186; Green M J A 1996, 136). However, Celtic myths make little distinction between the chthonic and the submarine. The conflation of the subterranean and the aquatic is also suggested in Romano-Celtic contexts; pits and wells, for example, have similar functions (Green 1984, 95; 1986, 101-102, 145, 156). Similarly, the giant associated with the rider-god may be either fish-tailed or snake-legged (H J M Green 1986, 43; Green 1986, 194).

The rider-god of the Jupiter-Giant columns often carries a wheel, which is undoubtedly the symbol of the Celtic Jupiter (Green 1976, 10). In some contexts it may appear alone (Green 1976, 18; 1981, 111; 1984, 168). For example, the Mouhet giant kneels in the attitude of a captive, a wheel balanced on his back (Green 1986, 59). However, although the Jupiter-Giant columns are predominantly a phenomenon of eastern Gaul and Germany, wheel symbolism is not exclusive to them (Green 1976, 18). It also appears in other contexts, as for example at Bisley-with-Lippiatt (Green 1976, 18; H J M Green 1986, 42-43).

The rider-god occasionally has the attributes of Mars, rather than Jupiter (Green 1976, 10, 30). Mars had old connections with fertility (Green 1976, 29; 1986, 36). Some representations of the Celtic Mars have very emphasised genitals and often possess a cornucopia, both reminiscent of the Dagda’s attributes (Webster 1986, 58; Green 1986, 113; 1986, 36). Other representations are even ambiguous; Mars Lenus at Chedworth has the attributes of Sucellus, a divine smith usually connected with Jupiter (Green 1976, 24, 29; 1984, 96-97). Similarly the Classical attributes of an armoured sky god from Earith could identify it with either divinity (H J M Green 1986, 51). The sky god was also identified with Apollo, whose lyre could be equated with the Dagda’s harp (Green 1984, 203; 1986, 172). Additionally, two of the Dagda’s thirty or forty epithets are Roth (Wheel) and Eoghaid or Horseman (CMT 93; Macalister 1938-42, VII, 181; Gray 1982, 100). It would therefore appear that the collective attributes of the Romano-Celtic sky god are identical to those of the Irish Dagda (see page 5).

The sky god’s wheel has been interpreted as a sun symbol for at least eighty years (Frazer 1922, 843; Green 1986b, 68). However the Dagda does not have a solar function. On the contrary, the sun is explicitly identified as a goddess, whose attribute it never is, and this identification is not in doubt (Met. Dinds. IV, Ard Macha, 127, lines 45-48; O’Rahilly 1946, 293; Mac Cana 1955). The wheel also lacks any specifically solar symbolism in the sources, despite an apparent familiarity with Gaulish dentate wheels (Met. Dinds. IV, Tlachtga, 189, lines 29-40). Given the preceding arguments, it is unlikely that Irish beliefs originally differed much from those of other Celtic peoples. The wheel must therefore represent some other aspect of celestial imagery.

It was a familiar observation in antiquity that the sky moves in a wheel-like manner around the stationary celestial axis (Aristotle, On the Universe 2.391b, 10-25; 2.392a 1-10; Ovid, Fasti I.119-120; IV. 179-180; Virgil, Georgics I, 242-251; Lucan, Bell. Civ. VI. 464-465; Vitruvius, De Arch. IX, i.2-3). The wheeled nature of the Dagda’s club, and representations such as Le Châtelet and the Etzelsdorf cone, suggest similar imagery (Green 1984, 98; 1997, 57). The wheel god is also associated with night and day, both of which revolve around the celestial axis (Green 1981, 112; Aristotle, On the Universe, 6.399a, 1-5). A similar idea may be intended by the Boa Island Janus (Figure 2). The rim of the sky god’s wheel may therefore illustrate the primary characteristic of the sky from a geocentric point of view; that it moves in circular manner around a stationary axis.

The celestial axis may be considered in local terms as an imaginary line from the celestial pole, marked by the Pole star, to a local observer. The corollary is a local centre that is the counterpart of the celestial pole, which in turn implies central religious places (Eliade 1957, 36-37). These are certainly evident in Ireland, and probably existed in Gaul (BG VI, 13; Lynn 1992, 41; Mac Cana 1996, 14). Lucan’s remark that the Pole Star looks down upon the Celts is therefore pertinent, and probably indicates a characteristic Celtic belief (Bell. Civ. 458-460). Gildas remarks that Britain is ‘poised in the divine balance . . . which sustains the whole earth’ (De Excidio II, 3). This is based neither on Classical ideas, nor on biblical authority. A continuing thread concerning the celestial axis runs though later Celtic religious poetry, such as The Measurements of the Universe (Hull 1913, 7).

Trees and columns form an important part of the iconography of Celtic and Romano-Celtic religion. They are probably equivalent; Jupiter-Giant columns are sometimes decorated to imitate trees, as at Hausen (Green 1986b, 67). Both apparently mediate between the celestial and the chthonic, as does the celestial axis (Eliade 1957, 35; Green 1976, 51; Green M J A 1996 120-121). Sacred trees are associated with kings in the Irish sources, who also mediate the divine, and the ‘royal’ site of Emain Macha had a free-standing central oak pillar (Met. Dinds. III, Temair Luachra, 239, 33-36; IV, Bile Tortan, 65-69, lines 57-60; Raftery B 1989, 141; Lynn 1992, 43; Green 1997, 56).

Wheel symbols are restricted in the number of spokes they possess (Table 1); four is both the oldest form and the commonest number (Green 1984, 18). In geocentric astronomy the rising and setting points of the solstices define the limits of the sun’s annual movement (Posidonius 137a). These points form a diagonal cross if considered in relation to the celestial axis (Figure 3), and there is evidence to suggest that crosses have some solar significance (Green 1991, 49). However, the idea that the sky itself was shaped in the form of a cross appears, in an abstract form, in Plato (Timaeus, 36 b-c). The association of crosses with the Celtic Jupiter suggests that they represent the law that established the limits of the sun (Green 1978, 20; 1989, 84-85; Lynn 1992, 48). Crosses are also found on model axes, one of which (Figure 2) possesses the double triangle shown in Figure 3 (Green 1978, 32-33; 1984, 69, 144, 163-165).

In antiquity, the movement of celestial bodies was thought to generate the time by which mortal life was measured (Plato, Timaeus 38 b-c; Hardie 1992, 72-73). A four-spoked wheel is therefore an appropriate symbol for celestial divine law (Aristotle, On the Universe, 6.399a, 15-399b, 25; 6.400b 5-15, 25-30; Radhakrishnan 1923, 79). Six spokes are obtained if the equinoctial line is also considered, although the symbolism is identical (Figure 3). This symbol appears on the Boa Island Janus (Figure 2). An east-west orientation was typical of Celtic and Romano-Celtic shrines, such as Heathrow (Lewis 1966, 32; Derks 1998, 180; Cunliffe 1997, 204).

This raises the question of the potential use of pillars as shadow clocks for organising communities. The Irish festivals, which form an agricultural calendar, fall at regular intervals between the equinoxes and the solstices, and were probably based on direct solar observations (Danaher 1981, 221-222; Green 1986, 74). The pairing of the festivals would produce an eight-spoked wheel, as in Figure 4 (Hicks 1988, 471; Lynn 1992, 43). Again, a ten-spoked wheel results from considering the equinoctial line. Interestingly, the eight-spoked half-wheel on the Gundestrup cauldron may suggest the division of the sky into day and night (Olmsted 1969, 66, Plate 2c); solar observations are applicable only to daytime, which is one half of the daily rotation around the celestial axis.

Shadow clocks can only be inferred in Ireland, although indigenous terminology was used for the solstices and equinoxes (Joyce 1903, I, 464-471). On the other hand, they were well known in the Graeco-Roman world (Mor. 410-411; Vitruvius De Arch. IX, i.1; Neugebauer 1975, 737-748, 615-624; Claridge 1998, 190-192). Furthermore, the Coligny calendar used the solstices to construct an identical pattern of festivals, which supports Dio’s suggestion of a pre-Roman festival on August 1 at Lugudunum (Dio 54, 32, 1; Suetonius Claudius 2,1; Olmsted 1969, 114-116; 1992, 1, 90, 175; Ross 1995, 433-434).

The twelve-spoked wheel from Felmingham Hall is probably derived from the zodiac (Green 1986b, 74). However, the zodiac is not primarily solar. The signs merely mark the divisions of the celestial equator, which is at right angles to the celestial axis (Aristotle, On the Universe, 2.392a 10-15; Posidonius 126; Vitruvius De Arch. IX, i.3-4). In geocentric terms, a year is marked by the complete rotation of the zodiac. The existence of a twelve-spoked wheel, and the possible planetary aspects of seven-spoked wheels, suggests the introduction of Mediterranean ideas to Romano-Celtic religion (Green 1984, 169-170). However, it may also imply that these ideas were reinterpreted to fit a pre-existing indigenous framework.

This interpretation of wheel symbolism suggests consistent imagery expressed through purposeful choices in spoke-numbers. The uncommon forms are probably aberrations, either due to individuals experimenting with symbolism, or to manufacturing errors, such as overstamping or slips in mould-making.

The interpretation is also consistent with the mythology. In CMT the destruction of Fomor power is the direct consequence of the sky god sexually uniting at Samhain with the Morrigan (Mór Rígan: Great Queen), a female representative of the chthonic/aquatic powers (CMT 84; Webster 1986, 32). The duplication of this sexual adventure in CMT 93 would seem to be due to localised versions of the same tale (Gray 1982, 98, n.84:360). As Lévi-Strauss remarked, ‘repetition has as its function to make the structure of myth apparent’ (Lévi-Strauss 1968, 229). Divine union may also be suggested by deep, narrow, circular shafts, common in Gaul and Germany, containing posts or other objects suggesting a ritual function (Ross 1976; Green 1976, 4, 51; 1986, 20, 132-135; Webster 1986, 109-110).

Like the Dagda, the Celtic Jupiter is associated with water (Green 1981, 113; 1984, 201; Webster 1986, 62). In northern Gaul, water sources are principally associated with fertility goddesses (Derks 1998, 141 n. 58). Romano-Celtic iconography furnishes several instances of celestial forces being conflated with a fertility goddess or chthonic symbols such as snakes, suggesting their equivalence (Green 1976, 18; 1981, 112-113; 1984, 201). Both the Gundestrup goddess and snakes are associated with serpentine S-symbols, and snakes are associated with goddesses at Ilkley and Sommerécourt (Olmsted 1969, 65-66; Green 1989, 25-26). This suggests that the chthonic/aquatic forces were seen as inherently female in function (Green 1976, 18; 1986, 50).

In Romano-Celtic iconography the feminine principle primarily denotes fertility (Wightman 1986, 557). The Morrigan, mentioned above, is one of several epithets for a divinity called Dé Ána, or ‘goddess of prosperity’; in later literature she is called Sovereignty. The epithet Dé Ána is paralleled in Romano-Celtic contexts by several solitary ‘mother goddesses’ depicted as Abundance (Wightman 1986, 557). In Romano-Celtic iconography ‘mother goddesses’ are often found with regional names (Green 1976, 22: Derks 1998, 123). The same pattern is seen in Ireland; Mór Muman (Great Mother), who is identified as the sun, gave her name to the province of Munster (CA 1; Mac Cana 1958, 77).

However, the Irish sources present a more complex picture. Falsehood, the source of sterility and disruption, is considered an essentially feminine characteristic (Tec. Corm. 16; Kelly 1988, 207). This is both conditioned by and reflected in the language; Fír (true) is masculine, while gó (falsehood) is feminine (Thurneysen 1946, 44; 196). Eithne Uathach is seen destroying fertility before her marriage, by eating children (CA 170); Eithne is a variation of Étaíne, a common name for Dé Ána. Similarly the Fomor, by the sterility of their rule, are collectively an essentially feminine power (see pages 6 & 14). They are only overthrown, and agricultural fertility assured, as a consequence of the Dagda’s sexual liaison with a Fomorian woman. The inference is that the feminine is only productive when constrained by marriage, and that it is after marriage that the goddess becomes Dé Ána; there are marked Greek parallels (Versnel 1993, 270-280; Gould 1985, 25). In addition, since the Irish Truth is the masculine life-giving force, the implication is that life is imposed by the male upon the female, who otherwise takes no active part, much like planting a seed. The same idea is found in both Graeco-Roman and Vedic thought (Eumenides 658-661; Mor. V 374-374; LM IX, 33; O’Flaherty 1980, 29).

Sovereignty is a divine queen, often associated with trees or columns; she is frequently identified with the horse, which is a solar and military symbol (Ross 1967, 223; Green 1984, 192; 1986, 59, 117, 164; 1993, 46; Oaks 1986, 78-80; Lynn 1992, 50). Despite explicit or implicit identification with the sun, these deities have aquatic features (Met. Dinds. IV, Ard Macha, 127, lines 45-48; Mac Cana 1958, 77, 89; 1996, 51; Oaks 1986, 78; Green 1989, 17); for example, Macha (called Grian or sun), is the daughter of Sainrith mac Imbaith (Ocean) (Kinsella 1969, 7). Later Gaelic folk poetry continued to associate the queenly sun with the sea, and the sun has marine associations in the Vedas (Carmina Gadelica III, 310-311; Wallis 1887, 48).

The Irish sovereignties are interchangeable and equivalent (Gray 1982, 118; O’Hehir 1983, 172; Green 1989, 188-189; Ross 1967, 219). Similarly, the Rom inscription shows an archaic profusion of epithets for a goddess identified as both Macha and Epona, both of whom are hippomorphic and have marine associations (Green 1984, 195; 1986, 164; Olmsted 1988, 317, 354-355). Morrigan’s name is synonymous with Dea Rigiina, Rigantona (Rhiannon), and with Regina Sancta (Epona) (Gruffydd 1953, 67, 103, 98; Oaks 1986, 78; 81). Both Morrigan and Epona have midwinter associations, the former through her intimate links with Newgrange (Met. Dinds. II, Brug na Bóinde, 19, lines 5-16; Stokes 1894, 292; Olmsted 1969, 139; Green 1989, 23). Midwinter is the point when the sun’s heat gets stronger.

The sky god mounted on the hippomorphic goddess has archaic overtones of union, control and sexuality (Dillon & Chadwick 1967, 88; Dillon 1973, 3; O’Flaherty 1980, 152-154). The Irish sources represent Sovereignty as the sky god’s wife; Macha, for example, was the wife of Nemed (Heaven) (O’Donovan 1868, 121; Puhvel 1970, 165-166). The Romano-Celtic sky god and his consort are probably represented by a mounted god and goddess at Kingscote (Green 1976, 19). The sky god’s consort also appears to be represented by Minerva from the Willingham Fen hoard, and there is a strong case for generally regarding the Romano-Celtic Minerva in this light (H J M Green 1986, 45). Minervan imagery is also suggested by Dea Rigiina’s spear; her cauldron suggests the marine, feminine, fertility symbol of Celtic myth (Brown 1913, 245-46, 249; Green 1976, 18; 1997, 132; Green M J A 1996, 68-69; Ford 1977, 69).

Minerva’s attributes parallel Sovereignty’s intimate association with the corollaries of Truth; justice, success in war, wisdom and fertility (Dillon 1947, 6; AM 17-21, 55, 56). Like Minerva, Macha is described as the goddess of women (Met. Dinds. IV, Ard Macha, 127, 49-52). Like Sovereignty, some Romano-Celtic Minervas possess marked aquatic symbolism, as at Chichester (RIB 627; Green 1976, 22; Cunliffe & Davenport 1985). At Stonea a gold votive leaf dedicated to Minerva, and a bust of the goddess, suggests a major shrine dedicated to a tutelary goddess of the Iceni; in Irish terms, Sovereignty (H J M Green 1986, 45). Warfare was also a significant part of Sovereignty’s role, and in earlier centuries Polybius equated Minerva’s Greek equivalent, Athena, with a Celtic war goddess (Pol. II, 32-33).

Minerva was twinned with Brigantia, who was a focus for loyalty in northern Britain (Henig 1986, 161). Brigantia was described as a territorial goddess, and it is possible that she became the divinity of Britannia Inferior (Henig 1984, 83; 1986, 161). A relief from Birrens depicts her with Minerva’s gorgon aegis, spear and shield (Toynbee 1964, 174-175; Henig 1986, 161). Her crown suggests that she was a provincial Fate (Henig 1984, 211-213). Brigantia is also winged as a Victory, and possesses an omphalos, the latter suggesting central places (Henig 1984, 213). Victory was originally a symbol of the Roman people but became assimilated to the imperial cult through the Augustan theology of victory (Fishwick 1987, 113, 115, 117). The association with rulership is again seen in a somewhat earlier altar to Dea Victoria Brigantia and the Imperial Numina at Greetland, Yorkshire (RIB 627), while a dedication to Caelestis Brigantia suggests that she is the sky god’s consort (Richmond 1967, 194). Brigantia also has aquatic associations, as illustrated by an altar inscription from Brampton to Dea Nympha Brigantia (RIB 2066; Henig 1986, 161; Green 1989, 17).

Brigantia is an example of Severan syncretism. However, her collective symbolism is also that of a Sovereignty; she was probably expected to communicate a dual message. The same principal is seen in dedications to genius loci, and may be present even in purely Classical representations (Webster 1986, 61; Wightman 1986, 545). Ambiguity and double meanings are a feature of Celtic art (Green 1986, 222-223).

Minerva was also twinned with the indigenous goddess Sulis (RIB 149). The element sul is cognate with the Irish feminine noun, súil, which meant both ‘sun’ and ‘eye’ (Rivet & Smith 1979; Green 1995, 96). This usage continued as late as the nineteenth century, when the feminine sun was described as the eye of the great god (Carmina Gadelica, III, 306-307). This description has both Romano-Celtic and Vedic parallels (Wightman 1986, 551; Radhakrishnan 1923, 77).

The iconography of the Bath pediment (Figure 6) is interesting (Cunliffe 1986, 6-7). The Minervan imagery of Gorgon’s head and shield is conflated with the state imagery of a shield surrounded by an oak wreath (Res Gestae 34). The imperial imagery is further strengthened by the twin Victories (Fishwick 1987, 108-109; 112-113). However, the iconography may have another double meaning; after CMT the severed head of Balor of the Evil Eye was placed in a tree that was later used to make a shield (Macneill 1908, 135-136). The association of gorgon’s heads with aquatic beings is documented (Toynbee 1964, 135-137). The iconography further suggests Sovereignty's myth by associating probable solar imagery with overt marine symbolism, in the form of sea creatures (Cunliffe & Davenport 1985, 26; 182-183). The solar imagery is probably indigenous, despite the possible late Neronic date, since Nero’s self-identification as the sun god meant that his imperial solar imagery generally suffered damnatio (Pliny NH XXXIV, 45; Cassius Dio LXIII, 6; Bell. Civ. I, 46-51; Cunliffe & Davenport 1985, 33).

It is unlikely that a purely Roman interpretation was intended. The site appears to have been used in the LPRIA, although the Roman construction may have obliterated much evidence (Selwood 1985). The area was relatively unromanised when the temple was built, and the population would have expected to see Sulis' traditional attributes represented in some form. Both the local elite and the Roman authorities would have been interested in identifying tribal authority with imperial power. The sculptors from north-east Gaul would have been familiar with similar mythology, and able to execute it in a form that could be read according to the cultural prejudices of the observer (Blagg 1979, 103; Cunliffe & Davenport 1985, 179; Lynn 1992, 42).

There are other indications of Sovereignty's attributes. For example, Sulis had marked connections with war (Cunliffe & Davenport 1985, 182; RIB 143, 144, 147). The curse tablets can be considered as appeals to divine justice or Truth (Cunliffe & Davenport 1985, 181-182). Possible fertility elements are also present, such as model breasts and a schist plaque depicting the Dea Matres (Cunliffe & Davenport 1985, 182, 183; Henig, Brown, Baatz, Sunter & Allason-Jones 1985, 8).

Similarly, the second century alterations at Bath created similarities to the temple of Mars Lenus at Trier (Cunliffe & Davenport 1985, 180). RIB 140 suggests that one dedicant regarded Sulis as identical to the Treveran Nemetona, consort of Mars Loucetius. The name Nemetona is the feminine equivalent of the Irish Nemed (Heaven), husband of Macha, suggesting that Nemetona was the sky god’s consort. Three other inscriptions, plus a masculine mask, suggest a perceived association of masculine divinity with the temple (Tomlin 1985, 152, 164, 180; Cunliffe 1985, 37).

The collective evidence from the site is therefore consistent with a typical Sovereignty. The identification is further supported by the later inscription of Gaius Severius Emeritus restoring the temple to the Virtue and Deity of the Emperor, thus directly associating Sulis with rulership (RIB 152; Cunliffe & Davenport 1985, 131; Tomlin 1985, 63). This association was also implied by state imagery on the pediment.

It is probable, however, that the political facts of the empire led to the majority of such goddesses changing function (Oaks 1986, 82). The incorporation of Gaul into the Roman empire apparently resulted in the loss of Epona’s Aeduan Sovereignty role, and her evolution into a protectrix of cavalry (Oaks 1986 79-81). Similarly, Minerva also appears on some early Gaulish public monuments, and at Willingham Fen, in a manner that suggests a connection with the Deae Matres (Green 1976, 22). These are a homogenous group that appear to have the same regional, chthonic/aquatic, fertility role as Sovereignty, but lack the ruling solar aspects (Gruffydd 1953, 104; Toynbee 1964, 172; Phillips 1977, 86; Green 1984, 196-197; 1986, 78-79, 99; Derks 1998, 123; Puhvel 1970, 167). Although the cults are indigenous, they do not appear to be pre-Roman. In Gaul they flourished from Gaius onwards, and dominated private cult (Green 1986, 102; Derks 1998, 119, 124, 128 note 245).

Sulis’ name, and her probable first century solar imagery, contrasts with the equally probable representation of Sol on the late third century Façade of the Four Seasons (Cunliffe & Davenport 1985, 120, 183; Cunliffe 1986, 9). By contrast, the third century Birrens Brigantia shows no obvious solar symbolism, despite being otherwise identifiable with Sovereignty. However, a third century Celtic Jupiter from Felmingham Hall is assimilated with the sun, and the Celtic Apollo developed similar associations (Green 1984, 64-65; H J M Green 1986, 45; Webster 1986, 62 note 20). This suggests a change in divine roles under the empire.

Perceptions of native goddesses had changed by the second or third century AD; only the name remained the same (Derks 1998, 119). Roman rule led to the disappearance of martial ideology from Celtic societies, and weakened the tribal structure (Roymans 1996, 37-40, 99). Emperor worship was also promoted, and emperors had been increasingly associated with Sol since Augustus (Neverov 1986; Beard, North & Price 1998, 259; Webster 1999, 13). It is therefore probable that there was a progressive transfer of the originally regional, solar, ruling attributes to appropriate male divinities, although the regular appropriation of Matres by soldiers suggests some perceived residual military symbolism (Henig 1984, 49). The remaining fertility aspects could be interpreted as chthonic or aquatic, leading to a diversification of forms. Aquatic associations would have facilitated assimilation with either nymphs or Venus, and the symbolism of pseudo-Venus’s appears to be related to the Deae Matres (Jenkins 1956, 61, 62, 65; 1957, 38; Toynbee 1964, 83; Green 1976; 20-21, 22, 119; 1984, 196, 200; 1986, 95, 164). Popularity and mass production probably no doubt divorced some from their original context and made them entirely anonymous except to their worshippers (Green 1978, 16-17).

A small relief from Cirencester shows a Mother holding a swaddled baby (Toynbee 1964, 172). This is a widespread image throughout the Romano-Celtic world (Green 1976, 23). Epona is similarly shown with a horse and foal, while a bronze plaque of Minerva from Maiden Castle suggests that she bears a male child (Green 1976, Plate VIII, h; Derks 1998, 81). The Classical goddess was never pregnant; however Sovereignty’s pregnancy is central to CMT.

In CMT the daughter of Indech mac Dé Domnann, who is almost certainly a double of the Morrigan, promised the Dagda that lying with her would destroy Fomor power (CMT 93). Lugh’s birth to Sovereignty is the predicted cause of the death of his grandfather, Balor, and the consequent destruction of Fomor power (Delaney 1991, 7; Green 1992, 38). Balor’s death is the cause of the simultaneous death of the Fomor king, Indech mac De Domnann, suggesting that the two are doubles (CMT 135). The defeat of the Fomor is directly related to the establishment of agricultural fertility (CMT 149-161; Macneill 1962, 4-5).

Lugh’s Welsh counterpart is Lleu (Ford 1977, 89-109). Both are separated from their mothers at birth. Both could be easily identified with Apollo; Lugh is a harper, and both names mean ‘light’ (CMT 73; Oosten 1985, 49). The name of Apollo Borvo (boiling) similarly suggests heat (Wightman 1986, 554; Green 1986, 162). However, the Irish Lugh cannot be a sun god, since Irish solar deities are female. It is however probable that the sun was seen as giving birth to heat as a separate entity. This conforms to climatic realities; in northern latitudes the sun is always visible, but it does not always provide warmth. Further, both the Irish sources and the Peñalba inscription link Lugh with the harvest, suggesting probable conflation of the sun’s heat and the ripening grain (Macneill 1962, 1, 3, 43-44, 343-344; Olmsted 1988, 357-358); the agricultural cycle waxes and wanes with the year. The identification of the Romano-Celtic Apollo with the sun is therefore probably due to the acculturative processes mentioned earlier (Green 1984, 203; Wightman 1986, 554).

However, the Romano-Celtic Apollo presents some problems, since he also represented the sky god (see page 7). In the Irish sources, the sky god is the source of Truth, which is equated with heat, but Lugh is not his equivalent. The Romano-Celtic Apollo, by conflating the two, suggests ideas of divine function that differ from Classical norms, and there is some evidence that the Celts saw nature as personifying the divine (Lynn 1992, 47; Green 1993, 50; 1995, 9; 1995b 465). It also suggests that it was possible for Romano-Celtic deities to implode into further multifunctionality, as well as diversifying, and that interpretation of this deity may have rested entirely with the worshipper.

Lleu is raised by Gwydion, and the latter has many of the properties of Mercury. Gwydion is a trickster who poses as a tradesman and is responsible for the theft of pigs (Ford 1977, 92-94, 99-103). The root of his name, gwydd, means ‘tree’. Interestingly, the Sanskrit equivalent of Mercury, Agni, is described as pillar that supports the heavens (Wallis 1887, 22, 29); in Romano-Celtic contexts Mercury often has a geographical name, suggesting central places (Green 1986, 98). In Irish mythology Manannan, another trickster associated with trade and pigs, takes Gwydion's role (CA 156; San. Corm. 114; O’Donovan 1868, 114; Macneill 1908, 118; Gray 1982, 91). Although Manannan is commonly seen as a sea god, his presence on the sea converts it to dry land (Meyer 1895, 16-21). Manannan also has another Welsh alter ego, Manawyddan, who appears to have mastered all mercantile trades (Ford 1977, 78-79).

In both instances the paired deities are inseparable. Mercury and Apollo are similarly linked in Romano-Celtic iconography, and even conflated (Wightman 1986, 553); at Reims Mercury possesses Apollo’s lyre (Wightman 1986, 552). Since Lugh was master of all arts, the roles in Roman terms (but not Celtic terms) overlapped (Dillon 1973, 14). There are also indications of confusion between Mercury and Mars (Green 1976, 30; 1986, 37). Mars’ traditional attribute was a spear, a weapon also associated with Lugh. Spears were evidently very important in Romano-Celtic iconography. A number of model votive spears and spear-shaped pole-tips have been recovered from religious contexts, such as Uley, Owmby, Woodeaton and Felmingham Hall (Green 1976, 42-43; Plate IX, i; Henig 1984, 22).

The Romano-Celtic Mars’ exaggerated genitals suggest that this deity was seen as a procreative aspect of the sky god, with whom he overlaps (CMT 93; Green 1976, 29). His ability to impart the life-giving force of Truth would give this deity healing abilities, as with Mars Lenus (Green 1986, 36). In CMT Lugh wields the spear of Gorias (gor: heat, hatching); the name implies that it also imparts Truth and life (CMT 4; Macalister 1938-42, IV, 143; Gray 1982, 74-75). Spears have phallic symbolism, and the equivalence of semen and fire is seen in both Vedic and Roman myth (Gjerstad 1973, 53; Green 1995, 40). At Tongres Mercury is triple-phallused, triplism being a common method of multiplying power; many representations associate this deity with fertility (Green 1976, 11; 1987, 79; 1986, 84-85). The Romano-Celtic function of Mars and Mercury would therefore seem to be identical; they represent the particularly phallic aspect of the sky god, which the name Mars Olludius (Great Tree) suggests is associated with the celestial axis (Green 1983, 43). A poletip from a shrine at Brigstock, which apparently fuses spear and phallus (Figure 7), supports this suggestion (Green 1976, Plate XXV, b; 1986, 98). Minerva’s spear may therefore, in indigenous contexts, suggest sexual union, which is consistent with her role as consort, and its function as a priapic symbol.

This would suggest that the divine union consisted of the sky imparting the masculine fiery essence of life to the aquatic female. This appears in the Vedic material, and in Anaxagorus (Aristotle, Meteorology, I.3.339b, 20-25; Eliade 1978, 209). It is consistent with the earlier implications regarding conception (page 15), and is suggested by the Aquitanian ritual of rolling a blazing wheel into water (Green 1986, 164). It would also explain the virtual absence of medical ex votos at Bath; the hot waters, charged with life, would restore health in a non-specific way.

The combination of male and female elements in sexual union is inherently ambivalent (O’Flaherty 1980, 292-293, 295-296). The Triad of Bolards sandwiches a hermaphrodite deity between Cernunnos and a Mother (Buckley 1981, 307). Gwydion is similarly sexually ambivalent since, like Agni, he gives birth (Ford 1977, 96-98; O’Flaherty 1980, 49, 299). However, the same idea can be represented by a divine pair who are seen as one unit (O’Flaherty 1980, 296). This would account for the Romano-Celtic Mercury acquiring a consort who, in some cases, has no separate identity (Green 1989, 56-57). In CMT Manannan’s role is taken by Tailltiu (CMT 55; Gray 1982, 91); in other accounts both foster Lugh, one after the other, which suggests an element of ambiguity. Like Rosmerta, the consort of the Romano-Celtic Mercury, Tailltiu has definite fertility aspects through her connection with Lughnasad (Macalister 1941, IV 117; Green 1976, 31; 1986, 97; Gray 1982, 127). Rosemerta’s patera is possibly linked with the marine Cauldron of Celtic myth (Green 1986, 97).

Mars and Hercules appear to be different aspects of the same god in northern Gaul; the former has a more southerly distribution (Derks 1998, 95-96, 242). The two were also linked in Greek thought (Aristotle, On the Universe, 2.392a 25-30). Bronze representations of Hercules are occasionally ithyphallic, an attribute of the Romano-Celtic Mars, while Hercules’ principal attribute, a club, is a probable representation of the celestial axis (Green 1976, 25). Again, an apparent association between the Wheel God and Hercules supports the idea of a purely procreative aspect (Green 1984, 95).

As noted earlier, the heat of the sun is probably conflated with the corn. Lugh is killed during Samhain as a consequence of his wife’s adultery (Met. Dinds. IV, Loch Lugborta, 278). Both the Celtic Apollo and seeds are connected with chthonic forces (Green 1986, 160-161, 194). Apollo Cunomaglos (Hound Lord) is associated with dogs, which are both chthonic and connected with agriculture (Jenkins 1957, 68-69, 70, 76; Webster 1986, 59; Henig 1986, 166; Green 1986, 194). In the LPRIA, the harvested grain was stored over winter in underground pits (Reynolds 1974; 1981, 88-89). Bodies might be buried in these pits, which at Danebury are associated with ritual pits (Green 1986, 133; Wait 1995, 492). Apollo Cunomaglos is linked to Apollo Maponus whose myth, like Lugh’s, involves being separated from his mother at birth (RIB 1121; Ross 1967, 208-209; Richmond 1967, 206-210; Green 1995, 64-66). Perhaps they are different aspects of the same deity.

The agricultural cycle is suggested in Noinden Ulad by the debility of the Ulstermen from Samhain to spring (Kinsella, 257, note 7). The same cycle may also explain the association of Bacchus with the rider-god at Wroxeter (Hutchinson 1986, 141-142); The equal prominence given to Lycurgus on a probable Jupiter-Giant column at Cirencester makes little sense in Classical terms (Phillips 1976; Henig 1986, 161; Green 1986, 64-65). However, Gwydion’s fostering of the infant Lleu is analogous to Mercury succouring Bacchus, and Lleu’s killer, Gronw, is central to Lleu’s story (Wightman 1986, 548; Ford 1977, 103-106). The iconography, which is probably of a Julian date, therefore makes sense in terms of indigenous myth (Green 1986, 65). It may also provide some insight into the surviving corners from Bath (Cunliffe & Davenport 1985, 118, 183; Cunliffe 1986, 8). The only element missing is the slaying of an aquatic being.

Two conclusions arise from this analysis; firstly the commonly accepted theory of the solar wheel appears to be fallacious. Its reinterpretation as a symbol of celestial law reveals archaic features observed in other mythologies. Secondly, it suggests that Romano-Celtic cults were not merely a continuation of pre-conquest belief, any more than the society in which it functioned was a pre-conquest society. Romano-Celtic beliefs cannot be projected backwards to draw conclusions about LPRIA beliefs since they are a product of a different environment. They evolved in a provincial, acculturated society, and reflected the nature of that society. The progressive acculturation of the elite, who supported local cults, and the pervasive presence of Roman ideas, made this evolution inevitable. Roman ideas and forms were adapted to express indigenous concepts, which changed as a consequence, and in consequence of the changed political circumstances. The multifaceted, unspecialised and polyvalent Celtic deities consequently evolved, either diversifying or conflating in the process, to fill needs and religious niches created by, and peculiar to, evolving Romano-Celtic societies.

Augustus, Res Gestae Divi Augustus, Brunt P A & J M Moore (1967), Oxford.

Aristotle, The Complete Works of Aristotle: Vol. 1, ed. Barnes J (1984), Princeton, New Jersey.

Kelly F trans. (1976), Audacht Morainn, Dublin.

Best R I, O Bergin, M A O’Brien & A O’Sullivan eds. (1954-83), The Book of Leinster, Dublin.

Buhler G trans. (1886), The Laws of Manu, Oxford, reprinted Delhi 1964.

Caesar, The Gallic War, trans. Edwards H J (1922), London.

Carmichael A (1900-1941), Carmina Gadelica, 5 vols., Edinburgh.

Cross T P and Slover C H (1936), Ancient Irish Tales, Dublin.

Dio Cassius, Roman History, trans. Cory E (1914-27), London.

Dunn J (1914), The Ancient Irish Epic Tale of Táin Bo Cúalgne, London.

Euripedes, The Plays of Euripedes, 2 vols., trans. Milman S D (1908), London & New York.

Ford P K trans., (1977), The Mabinogi and Other Welsh Tales, Berkely.

Giles J A (1885), Six English Chronicles, London.

Goodall D ed. (1996), Hindu Scriptures, London.

Gray E A (1982), Cath Maige Tuired, Kildare.

Gwynn E trans. (1900-1935), The Metrical Dindshenchas: RIA Todd Lecture Series Vols. VIII-XII, Dublin.

Hancock W N ed. & trans. (1865), Senchus Mor: The Ancient Laws of Ireland, London.

Hesiod, The Works and Days: Theogony. The Shield of Heracles, trans. Lattimore R (1959), Ann Arbor.

Homer, The Odyssey, trans. Murray A T (1919), London.

Hull E (1913), The Poem Book of the Gael, Edinburgh.

Kinsella T trans. (1970), The Tain, Dublin.

Lucan, The Civil Wars, trans. Duff J D (1928), London & New York.

Macalister R A S ed. and trans. (1938), Lebor Gabála Érenn, Dublin.

Macneill E trans. (1908), Duanaire Finn: The Book of the Lays of Fionn: Part 1, London.

Meyer K trans. (1895), The Voyage of Bran Son of Febal, London.

Meyer K trans. (1909), The Instructions of King Cormac mac Airt: RIA Todd Lecture Series vol. XV, Dublin.

Murphy G (1953), Duanaire Finn: The Book of the Lays of Fionn: Part 3, London.

O’Donovan J trans. (1868), Cormac’s Glossary, Calcutta.

Ovid, Fasti, trans. Frazer Sir J G (1931), London & New York.

Plato, Timaeus and Critias, trans. Taylor A E, (1929), London.

Pliny, Natural History, trans. Rackham R (1942), Cambridge Mass.

Plutarch, Moralia, vol. V, trans. Babbitt F C (1969), Cambridge, Mass.

Polybius, The Histories, trans. Schuckburgh E S (1889), 2 vols., London.

Posidonius, The Fragments: Volume III, The Translation of the Fragments, trans. Kidd I G (1999), Cambridge.

Stokes W H (1894), The Prose Tales in the Rennes Dindsenchas, Revue

Celtique, 272-336.

Stokes W H & Windisch E (1880), Irische Texte, Leipzig.

Stokes W H & Windisch E (1887), Irische Texte, Zweite Serie. 2 Heft., Liepzig.

Stokes W H & Windisch E (1891), Irische Texte, Dritte Serie. 1 Heft., Liepzig.

Stokes W H & Windisch E (1897), Irische Texte, Dritte Serie. 2 Heft., Liepzig.

Suetonius, The Twelve Caesars, trans. Graves R (1979), London.

Tacitus, Annals trans. Church A J & W J Brodribb (1921), London.

Tacitus, Agricola & Germania, trans. Handford S A (1970), Harmondsworth.

Virgil, The Georgics & Eclogues of Virgil, trans. Williams T C (1915), Cambridge, Mass.

Vitruvius, The Ten Books on Architecture, trans. Morgan M H (1914), London & Cambridge, Mass.

Alcock J (1986), The concept of genius in Roman Britain, in Henig M & A King, Pagan Gods and Shrines of the Roman Empire, Oxford, 113-134.

Allason-Jones L (1989), Women in Roman Britain, London.

Allen D F (1980), The Coins of the Ancient Celts, Edinburgh.

Beard M, J North & S Price (1998), Religions of Rome: Volume 1. A History, Cambridge.

Birley A (1988), The People of Roman Britain, London.

Breatnach R A (1953), The Lady and the King; a theme of Irish

literature, Studies XLII, 321-336.

Blagg T F C (1979), The date of the temple at Bath, Britannia X, 101-

108.

Breeze D J & B Dobson (1976), Hadrian’s Wall, London

Brown A C L (1913), Notes on Cauldrons of Plenty and the Land-Beneath- the-Waves, in F N Robinson et al eds., Anniversary Papers by Colleagues and Pupils of George Lyman Kittredge, Boston, 235-

250.

Buckley (1981), The Aeduan area in the third century, in King A & M Henig, The Roman West in the Third Century, Oxford, 287-315.

Bunham B C (1995), Celts and Romans: towards a Romano-Celtic society,

in Green M J ed., The Celtic World, London & New York, 121-144.

Burkert W (1985), Greek Religion, Oxford.

Charles-Edward T M (1970), The date of the four branches of the Mabinogion, Trans. Hon. Soc. Cymmrodorion, 263-298.

Christison R (1880), On an ancient wooden image found at Ballachulish Peat Moss, Proc. Soc. Antiq. Scotland, vol. 15, 158 ff.

Claridge A (1988), Rome: An Oxford Archaeological Guide, Oxford.

Collingwood R G & Wright R P (1965), The Roman Inscriptions of Britain: Volume 1, Inscriptions on Stone, Addenda and Corrigenda by Tomlin R S O (1995), Oxford.

Collis J (1987), Celtic Europe before the Romans, in Wacher J ed., The Roman World, vol. 1, London & New York, 15-37.

Crumley C L (1974), Celtic Social Structure: The Generation of Archaeologically Testable Hypotheses from Literary Evidence, Ann Arbor, Michigan.

Cunliffe B (1979), The Celtic World, London, Sydney & Toronto.

Cunliffe B (1985), The Temple of Sulis Minerva at Bath: Volume 2, The Finds from the Sacred Spring, Oxford.

Cunliffe B (1986), The sanctuary of Sulis Minerva at Bath: A brief review, in Henig M & A King, Pagan Gods and Shrines of the Roman Empire, Oxford, 1-14.

Cunliffe B (1988), Greeks, Romans and Barbarians, London.

Cunliffe B (1995), Iron Age Britain, London.

Cunliffe B (1997), The Ancient Celts, Oxford.

Cunliffe B & P Davenport (1985), The Temple of Sulis Minerva at Bath, Oxford.

Dalton G F (1971), The Ritual Killing of the Irish Kings, Folklore 81, 1-22.

Danaher K (1981), Irish folk tradition & the Celtic calendar, in O’Driscoll R, The Celtic Consciousness, Portlaoise, 217-242.

Davies J L (1984), Soldiers, peasants, and markets in Wales and the Marches, in Blagg T F C & A C King, Military & Civilian in Roman Britain, Oxford, 93-128.

Delaney F (1989), Legends of the Celts, London.

Derks T (1998), Gods, Temples and Ritual Practices, Amsterdam.

Dillon M (1947), The Archaism of Irish Tradition, London.

Dillon M (1972), The consecration of the Irish Kings, Celtica vol. 10, 1-8.

Dillon M (1973), Celt and Hindu, Dublin.

Dillon M (1975), Celts and Aryans, Simla.

Dillon M & N Chadwick (1967), The Celtic Realms, London.

Dorson R M (1955), The eclipse of solar mythology, J. American Folklore 68, 393-416.

Durkheim E (1912), The Elementary Forms of the Religious Life, trans. Swain J W (1965), New York.

Eliade M (1978), A History of Religious Ideas, vol.1, Chicago.

Eliade M (1957), The Sacred and the Profane, trans. Trask W R (1959) London & New York.

Evans J (1987), Graffiti and the evidence of literacy and pottery use in Roman Britain, Archaeol. J 144, 191-204.

Ferguson J (1987), Ruler worship, in Wacher J, The Roman World; Volume 2, London & New York, 766-784.

Finley M I (1954), The World of Odysseus, reprinted 1999, London.

Fishwick D (1987), The Imperial Cult in the Latin West, 2 vols., Leiden.

Fitzpatrick A P (1989), The uses of Roman imperialism by the Celtic barbarians in the later Republic, in Barrett J C, A P Fitzpatrick & L Macinnes, Barbarians and Romans in North-West Europe from the Later Republic to Late Antiquity, Oxford, 27- 54.

Flannery K V & J Marcus (1998), Cognitive archaeology, in Whitley D S ed., Reader in Archaeological Theory: Post Processual and Cognitive Approaches, 31-48.

Frazer J G (1922), The Golden Bough: A Study in Magic and Religion, London.

Gerriets M (1987), Kingship and exchange in pre-Viking Ireland, Camb. Medieval Celtic Studies 13, 39-72.

Gjerstad E (1973), Early Rome VI: Historical Survey, Lund.

Goodchild R & J R Kirk (1954), The Romano-Celtic Temple at Woodeaton, Oxon. 19, 15-38.

Green H J M (1986), Religious cults at Roman Godmanchester, in Henig M & A King, Pagan Gods & Shrines of the Roman Empire, Oxford 29- 56.

Green M J (1975), A Romano-British cult symbol from Icklingham, Suffolk, Proc. Cambridge Antiq. Soc. 66, 55-61.

Green M J (1976), The Religions of Civilian Roman Britain, BAR 24, Oxford.

Green M J (1976b), A Corpus of Religious Material from the Civilian Areas of Roman Britain, Oxford.

Green M J (1978), A Corpus of Small Cult Objects from the Military Areas of Roman Britain, Oxford.

Green M J (1981), Model objects from military areas of Roman Britain, Britannia XII, 253-269.

Green M J (1983), A Celtic god from Neatherby, Cumbria, Trans. Cumberland Westmorland Antiq. Archaeol. Soc. 83, 41-47.

Green M J (1984), The Wheel as a Cult Symbol in the Romano-Celtic World, Brussels.

Green M J (1986), The Gods of the Celts, Stroud, Gloucestershire.

Green M J (1986b), Jupiter, Taranis & the solar wheel, in Henig M & A King, Pagan Gods & Shrines of the Roman Empire, Oxford, 65-76.

Green M J (1987), Provincial cults, in Wacher J, The Roman World; Volume 2, London & New York, 785-795.

Green M J (1989), Symbol and Image in Celtic Religious Art, London & New York.

Green M J (1991), The Sun Gods of Ancient Europe, London.

Green M J (1992), Animals in Celtic Life and Myth, London & New York.

Green M J (1992), Dictionary of Celtic Myth and Legend, London.

Green M J (1993), Celtic Myths, London.

Green M J (1995), Celtic Goddesses; Warriors, Virgins and Mothers, London.

Green M J (1995b), The gods and the supernatural, in Green M J ed., The Celtic World, London & New York, 465-488.

Green M J (1997), Exploring the World of the Druids, London.

Green M J A (1996), Celtic Art, London.

Gruffydd W J (1953), Rhiannon: An Inquiry into the Origins of the First and Third Branches of the Mabinogi, Cardiff.

Hamp E P (1972), On dating and archaism in the Pedeir Keinc, Trans. Hon. Soc. Cymmrodorion, 95-103.

Hamp E P (1974), Mabinogi, Trans. Hon. Soc. of Cymmrodorion, 243-249.

Hardie P (1992), Augustan poets and the mutability of Rome, in Powell A, Roman Poetry & Propaganda in the Age of Augustus, London, 59-82.

Haselgrove C C (1982), Wealth, prestige and power: the dynamics of late Iron Age political centralisation, in Renfrew C & S Shennan eds., South East England, Ranking, Resource and Exchange: Aspects of the Archaeology of Early European Society, Cambridge, 79-88.

Haselgrove C C (1984), Romanisation before the conquest: Gaulish precedents and British consequences, in Blagg T F C & A C King eds., Military and Civilian in Roman Britain, BAR Series 136, Oxford, 1-64.

Haselgrove C (1990), The Romanization of Belgic Gaul: some archaeological perspectives, in Blagg T & M Millett eds., The Early Roman Empire in the West, Oxford, 45-71.

Hatt J J (1970), Celts and Gallo-Romans, Geneva.

Haverfield F J (1917-1918), Roman Cirencester, Archaeologia 69, 161- 209.

Haynes I (1993), The Romanisation of religion in the auxilia of the Roman Imperial army, Britannia XXIV, 141-157.

Heichelheim F M (1946), Bacchus or a young satyrs, not Mercury, L’Antiqite Classique 15, 227.

Henig M (1983), The Maiden Castle ‘Diana’: a case of mistaken identity?, Proc. Dorset Nat. Hist. Archaeol. Soc. 105, 160-162.

Henig M (1984), Religion in Roman Britain, London.

Henig M (1986), Ita intellexit numine inductus tuo: Some personal interpretations of deity, in Henig M & A King, Pagan gods and shrines of the Roman Empire, Oxford, 159-170.

Henig M, D Brown, D Baatz, N Sunter & L Allason-Jones (1985), Objects from the sacred spring, in Cunliffe B, The Temple of Sulis Minerva at Bath: Volume 2 The Finds from the Sacred Spring, Oxford, 59-227.

Hicks R (1981), Archaeoastronomy and related problems: Old world approaches vs. new, in Williamson R A, Archaeoastronomy in the Americas, Los Altos, 43-50.

Hicks R (1988), The year at Dromberg, in Aveni A F, World Archaeoastronomy, Cambridge, 470-482.

Hogan J (1931-32), The Irish law of kingship with special reference to Ailech and Cenél Eoghain, Proc. RIA XL, Sect. C, 186-254.

Hutchinson V J (1986), The cult of Bacchus in Roman Britain, in Henig M & A King (1986), Pagan gods and shrines of the Roman Empire, Oxford, 135-146.

Jacobsthal P (1944), Early Celtic Art, 2 vols., Oxford.

Jenkins F (1957), The cult of Dea Nutrix in Kent, Archaeologia Cantiana, vol. 71, 30 ff.

Jenkins F (1957), The role of the dog in Romano-Gaulish religion, Latomus vol. 16, 60-75.

Jenkins F (1958), The cult of the pseudo-Venus in Kent, Archaeologia Cantiana, vol. 72, 60ff.

Jenkins F (1962), The horse deity of Canterbury, Archaeologia Cantiana, vol. 77, 142ff.

Johns C M & Potter T (1983), The Thetford Treasure, London.

Johns C M (1981), A gold votive leaf and bust of Minerva from Stonea, Britannia XXII, 101-104.

Joyce P W (1903), A Social History of Ireland, 2 vols., Dublin.

Kelly F (1988), A Guide to Early Irish Law, Dublin.

Krappe A H (1927), Balor With The Evil Eye, Lancaster.

King A (1990), Changes in Romano-Celtic religion, in Blagg T & M Millett eds., The Early Roman Empire in the West, Oxford, 220- 241.

Lewis M J T (1966) Temples in Roman Britain, Cambridge.

Linduff K (1979), Epona: A Celt among the Romans, Latomus 38(3), 817-837.

Lynn C J (1992), The Iron Age mound in Navan Fort: A physical realization of Celtic religious beliefs? Emania 10, 33-57.

Macalister R A S (1917-19), Temair Breg: A study of the remains and traditions of Tara, Proc. RIA XXXIV, Sect C, 231-399.

Macalister R A S & R Praeger (1928-29), Report on the excavation of Uisneach, Proc. RIA XXXVIII, Sect C, 69-127.

Mac Cana P (1955), The theme of the king and the goddess in Irish literature, Etudes Celtiques VII, 76-114; 356-413; VII (1958),

59-65.

Mac Cana P (1958), Branwen Daughter of Llyr, Cardiff.

Mac Cana P (1996), Celtic Mythology, London.

Mackie N (1992), Ovid and the birth of Maiestas, in Powell A ed., Roman Poetry & Propaganda in the Age of Augustus, London, 83-97.

McLeod N (1992), Early Irish Contract Law, Sydney.

MacMullen R (1981), Paganism in the Roman Empire, Yale.

MacMullen R (1990), Changes in the Roman Empire, Princeton, New Jersey.

MacNeill M (1962), The Festival of Lughnasa, Oxford.

Mallory J P (1989), The career of Conall Cernach, Emania 6, 22-28.

Manning C (1988), A note on sacred trees, Emania 5, 34-35.

Megaw J V S (1970), The Art of the European Iron Age, Bath.

Millett M (1984), Forts and the origins of towns: cause or effect, in Blagg T F C & A C King eds., Military and Civilian in Roman Britain, BAR Series 136, Oxford, 65-74.

Millett M (1994), The Romanization of Britain, Oxford.

Minahane J (1993), The Christian Druids, Dublin.

Neugebauer O (1975), A History of Ancient Mathematical Astronomy, Berlin.

Neugebauer O (1957), The Exact Sciences in Antiquity, 2nd ed., Providence RI.

Neverov O (1986), Nero-Helios, in Henig M & A King, Pagan Gods & Shrines of the Roman Empire, Oxford, 189-195.

Oaks L S (1986), The goddess Epona, in Henig M & A King, Pagan Gods & Shrines of the Roman Empire, Oxford, 77-84.

Ó’Cuív B (1954), Lugh Lámhfhada and the death of Balar ua Neid, Celtica 2, 64-66.

O'Curry (1861), The Manuscript Materials of Ancient Irish History, Dublin.

O'Flaherty W D (1980), Women, Androgynes and Other Mythical Beasts, Chicago.

O’Hehir B (1983), Eachtra Airt meic Cuind, in Ford P K ed., Celtic Folklore & Christianity, Santa Barbara.

Olmstead G S (1969), The Gundestrup Cauldron, Collection Latomus 162, Brussels.

Olmstead G S (1988), Gaulish and Celti-Iberian poetic inscriptions, Mankind Quarterly XXVIII 4, 339-387.

Olmstead G S (1992), The Gaulish Calendar, Bonn.

Oosten J G (1985), The War of the Gods: The Social Code in Indo- European Mythology, London.

O’Rahilly T F (1946), Early Irish History and Mythology, Dublin.

Phillips E J (1976), A Roman figured capital in Cirencester, J. Brit. Archaeol. Assoc. 129, 25-41, with pls. 10-11.

Phillips E J (1977), Corpus Signorum Imperii Romani. Great Britain 1 Fasc. 1, Corbridge, Hadrian’s Wall East of the North Tyne, Oxford.

Puhvel J (1970), Aspects of equine functionality, in Puhvel J ed., Myth and Law Amongst the Indo-Europeans: Studies of Indo- European Comparative Mythology, Berkeley & Los Angeles, 159-

172.

Radhakrishnan S (1923), Indian Philosophy, 2 vols. London.

Raftery B (1972), Irish hill forts, in Thomas C ed., The Iron Age in the Irish Sea Province, CBA Research Report 9, 37-58.

Raftery B (1989) Barbarians to the West, in Barrett J C, Fitzpatrick A P & Macinnes L, Barbarians and Romans in North- West Europe from the later Republic to Late Antiquity, BAR International Series 471, Oxford, 117-152.

Raftery J (1972), Iron Age and Irish Sea: problems for research, in Thomas C ed., The Iron Age in the Irish Sea Province, CBA Research Report 9, 1-10.

Reece R (1988), My Roman Britain, Oxford.

Reece R (1981), The third century: Crisis or change, in King A & M Henig, The Roman West in the Third Century, BAR International Series 109 (i), Oxford, 27-38.

Reinach S (1895), Epona, Revue Archaeologique 1895, part 1, 113, 309. Addenda ibid 1898, part 2, 187; 1899, part 2, 61; 1902, part 1, 227; 1903, part 2, 348.

Renfrew C (1987), Towards and Archaeology of the Mind: The Puzzle of Indo-European Origins, Cambridge.

Renfrew C & E B W Zubrow eds. (1994), The Ancient Mind: Elements of Cognitive Archaeology, Cambridge.

Reynolds P J (1974), Experimental Iron Age storage pits – an interim report, Proc. Prehistoric Soc. 40, 118-131.

Reynolds P J (1981), The material culture of the pagan Celtic period, in O’Driscoll R, The Celtic Consciousness, Portlaoise, 79-94.

Richmond I A (1967) Roman legionaries at Corbridge, their supply- base, temples and religious cults, Archaeol. Aelina 4^th series, 21, 127-224.

Richmond I A & J M C Toynbee (1955), The Temple of Sulis Minerva at Bath, JRS 45, 97-105.

Rivet A L F & C Smith (1979), The Place Names of Roman Britain, London.

Romm J S (1992), The Edges of the Earth in Ancient Thought, Princeton, New Jersey.

Ross A (1967), Pagan Celtic Britain, London.

Ross A (1995), Ritual and druids, in Green M J ed., The Celtic World, New York & London, 423-444.

Ross A & Feachum R (1976), Ritual rubbish: The Newstead pits, in Megaw JVS ed., To Illustrate the Monuments: Essays on Archaeology Presented to Stuart Piggott, London, 230-237.

Ross A (1995), Ritual and the druids, in Green M J ed., The Celtic World, London & New York, 423-444.

Roymans N (1996), From the Sword to the Plough: Three Studies on the Earliest Romanisation of Gaul, Amsterdam.

Ruggles C L N (1988), Recent developments in megalithic astronomy, in Aveni A F, World Archaeoastronomy, Cambridge, 13-26.

Rynne E (1972), Celtic stone idols in Ireland, in Thomas C ed., The Iron Age in the Irish Sea Province, CBA Research Report 9, 79- 98.

Selwood L (1985), The Celtic coins, in Cunliffe B & P Davenport, The

Temple of Sulis Minerva at Bath, Oxford.

Sheldon H, I Haynes & L Hannigan (1999), The Archaeology of London, Oxford.

Simpson C J (1993), Once again, Claudius and the temple at Colchester, Britannia XXIV, 1-7.

Smith R ed. (1942), The Tree of Life, New York.

Tarrant H (1990), The distribution of early Greek thinkers and the question of ‘alien influences’, in Descoeudres J-P ed., Greek Colonists and Native Populations, Oxford.

Thurneysen R (1946), A Grammar of Old Irish, trans. Binchy D A & O Bergin (1975), Dublin.

Tierney J J (1960), The Celtic Ethnography of Posidonius, Dublin.

Tomlin R S O (1985), The curse tablets, in Cunliffe B, The Temple of Sulis Minerva at Bath: Volume 2 The Finds from the Sacred Spring, Oxford, 59-227.

Toynbee J M C (1964), Art in Britain under the Romans, Oxford.

Versnel H S (1981), Faith, Hope and Worship: Aspects of Religious Mentality in the Ancient World, Leiden.

Versnel H S 1993, Inconsistencies in Greek and Roman Religion II; Transition and Reversal in Myth and Ritual, Leiden.

Wait G A (1995), Burial and the Otherworld, in Green M J ed., The Celtic World, London & New York, 489-514.

Wallis H W (1887), Cosmology of the Rig Veda, London.

Webster G (1986), What the Britons required from the gods, in Henig M & A King, Pagan Gods & Shrines of the Roman Empire, Oxford, 57- 64.

Webster G (1986), The British Celts and Their Gods Under Rome, London.

Webster J (1995), Sanctuaries and sacred places, in Green M J ed.,

The Celtic World, London & New York, 445-464.

Webster J (1999), At the end of the world: Druidic and other

revitalization movements in post-conquest Gaul and Britain,

Britannia XXX, 1-20.

Wightman E M (1986), Pagan cults in the province of Belgica, ANRW II, 18(1), 542-589.

Wolf G (1988), Becoming Roman: The Origins of Provincial Civilisation

in Gaul, Cambridge.

Wright R P (1943), The Whitley Castle Altar to Apollo, JRS 33, 36-8.

Zaidman L B & P Schmitt-Pantel, Religion in the Ancient Greek City, trans. Cartledge P (1992), Cambridge.