‘Celtic’ Wheel Symbolism: The Archaeological & Iconographical

Evidence For The Links Between Time, Agriculture, & Religious

Ideas In The Celtic World From Later Prehistory To The Roman

Period.

Kevin Jones 2003

I certify that the work submitted herewith is my own, and that I have duly acknowledged any quotation from the published or unpublished work of other persons

Signature Date CONTENTS

Introduction 1

Methods of structuring time in ancient societies 11

Archaeological evidence for ideas about time in the Celtic world 22

• The Coligny Calendar 23

• The Jupiter-Giant Columns 34

• Summary of analysis 41

Wheel symbolism 45

• A statistical analysis of wheel symbolism 45

Reinterpreting wheel symbolism 59

Conclusions 68

Concordance table for coins 69

Tables 1-5: 75

Key to wheels 81

Plates 1-2: Wheel symbols 82

Key to coins 84

Plates 3-35: Wheel symbolism in coins 88

Plate 36: Coin-using tribes in Late pre-Roman Iron Age Britain 121

Plate 37: Distribution of Jupiter-Giant columns 122

Plate 38: Distribution of Epona finds 123

Plate 39: Map of Celtic regions and people 124

Bibliography 125

13,886 words, excluding references, captions, tables and titles 1,511 words references, captions, tables and titles

INTRODUCTION

The wheel is a ubiquitous religious symbol found in both Celtic and Romano-Celtic contexts. It presumably represented a widely understood religious concept that related to the central core of beliefs. However, with no indigenous literate statements of those beliefs, archaeologists are forced to rely upon Classical authors, inscriptions, and iconographical analysis for insights. Although these can provide valuable information, they also present problems. Iconographical interpretation can be enormously subjective, and interpretative errors may result from unconscious and potentially anachronistic cultural preconceptions. Similarly, the works of authors such as Tacitus can be misleading, since they interpreted Celtic beliefs according to their own cultural notions (Germ. 43). The reverse is also true; indigenous religion adopted Roman concepts, but reinterpreted them according to its existing framework (Green 1976, 6-

7, 8-10, 119). To compound matters further, this also altered indigenous religion to some undetermined degree (Derks 1998, 105, 107).

There are problems even when Celtic and Roman deities are directly equated. The limited range of Roman deities used might suggest that a limited number of indigenous, multifunctional deities were described by many regional epithets

(Webster 1986b, 54; Green 1986a, 32; Derks 1998, 95). However, how do we understand the equation? Some representations are alien to Roman conceptions, 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. The typical function of the

2 Roman god, as we understand it1, may not, therefore, accurately indicate the function of the Celtic deity. On the other hand those responsible for name-dualling presumably saw a relationship, although it may be difficult to define without further information.

Of course, none of this helps where indigenous deities are described as genius loci, or where they are unconnected to Roman concepts.

Matters might be simpler if we understood the relationship between the iconography and religious ideas about time and cosmology. A possible cosmology is suggested by the Old Irish source Cath Maighe Tuired, and this may have a connection with the iconography of the Jupiter-Giant columns. However, making a direct equation between the Irish sources and the iconography is complex, and beyond the scope of the present paper. It is in any case wiser to first establish a robust model derived from the archaeological data before including material from the Celtic sources.

It is therefore proposed to analyse the role of time in Celtic and Romano-Celtic religion and its relationship to agricultural fertility. Such a link is suggested by the iconography of the Jupiter-Giant columns. The summit group typically includes the wheel-god, who was frequently identified with the celestial deity Jupiter; celestial phenomena were the basis for structuring time in antiquity. Further, the bases of

Jupiter-Giant columns often feature the deities of the days of the week or the four seasons, suggesting a further connection with time. Additionally the wheel-god occasionally has the attributes of Mars; the Romano-Celtic Mars is associated with indigenous religious calendars. The old fertility connections of the Classical Mars are

1 Assuming that we understand Roman deities as the Romans understood them.

3 even more marked in his Romano-Celtic equivalent2, and the religious calendars can be shown to have an agricultural orientation (Green 1976, 30; 1986, 36, 113). There therefore appears to be, at some level, a relationship between the wheel-god, wheel symbolism, agricultural fertility, religion, and concepts of time.

Nevertheless, this approach is likely to raise objections. Collis, for example, has argued that there was no such thing as Celtic society, and that there were never any

Celts in Britain or Ireland (Collis 1985; 1994). Chapman argued that the whole notion of being 'Celtic' is and was a construct imposed by the centre on diverse peripheral cultures that had no sense of a common identity, no political unity and possibly even ethnically diverse origins (Chapman 1992). In short, the Celts were entirely a creation of Roman propaganda and 18th century romantic visions of European ‘noble savages’.

One might conclude from these arguments that there is very little chance of finding common religious ground between such cultures, and that such a line of inquiry is therefore futile. This, however, is a counsel of despair. To address Collis’s rather

Thatcherite argument first, there were and are Celtic-speaking populations.

Populations form societies, and Celtic-speaking societies may, or may not, demonstrate some common features that are not shared with non-Celtic societies. It is not however necessary to assume that either a common ethonym or an overarching

Pan-Celtic awareness of a Celtic society are necessary prerequisites for such similarities.

2 I.e cornucopias, ithyphallic statues.

4 The objection of possible ethnic diversity runs the risk of confusing ethnicity and culture. The ethnic origins of individuals may have no relation to their cultural affiliations, their native language or even their own perception of their ethnicity. A substantial number of white Americans are descendants of black Africans (Stuckert

1976, 139; Parra et al 1998); presumably they would consider themselves neither ethnically nor culturally Afro-American. Ethnicity is actually a complex and fluid cultural construct defined by perceived existing differences and similarities, rather than a biological reality (James 1999, 76-77).

In any case it would be rash to make dogmatic assertions about ethnicity from, for example, Caesar’s observations. Language, laws and institutions may, in Caesar’s opinion, have separated the Belgae, Celtae and Aquitani (BG 1.1); however, this does not necessarily indicate three ethnic groups. The reported differences in language may simply reflect the differences between P- and Q-Celtic, which at that time may have been little more than differences in dialect (Olmsted 1979, 107-110). Dialectal differences may sound like language differences to an outsider. Similarly the reported differences in laws and institutions could be accounted for by Hellenising and

Romanising influences; Caesar in fact makes the observation that the Belgae were furthest from such influences (BG 1.1).

The emergence of shared cultural features, including language3, is in any case not precluded by different ethnic origins (Wadell & Conroy 1999, 129-130). Celtosceptic arguments nevertheless seem to assume that Celtic-speaking cultures were discreet entities that did not acculturate, an assumption that is at odds with the evidence. As

3 Language is a large part of culture.

5 the preconquest Romanisation of Britain demonstrates, the shifting patterns of Celtic clientship resulted in culture contact and consequent acculturation over considerable distances (Haselgrove 1982; 1984: Millett 1994, 34). Similarly early Bronze Age

Ireland was in contact with southern Europe, the Mediterranean area, and to a lesser extent, with what later became Gaul. In the later Bronze Age, these contacts became largely restricted to northern Europe (Raftery J 1972, 2). La Tène objects reached

Ireland as early as the fourth century BC, and there are strong suggestions of on-going contacts with the general area of eastern Gaul and the Rhineland (Raftery J 1972, 5-

6). Analysis of gold objects in the National Museum of Ireland demonstrates that the vast majority of Irish gold objects, from both the pre-Roman and Roman periods, were either manufactured from imported gold, or were imported as manufactured items (Kelly 1976, 46). There is also sufficient evidence to propose some form of contact between parts of Ireland and Iberia during the first millennium BC, such as features of hillfort construction, and the skull of a North African Barbary ape found at

Navan Fort (Emain Macha), (Raftery B 1972, 49-50; Raftery B 1989, 141). One can also find evidence that religious ideas also travelled by these routes. Navan Fort was a major ritual site that was in use from the seventh to the first centuries BC; its closest parallels are the sixth century BC enclosures at the Goloring and the Goldberg in

Germany (Powell 1958, 172-173; Green 1986a, 20; 1989, 151-152; Raftery B 1989,

141). Irish stone idols suggest continental and British artistic influences, which in turn imply the emergence of a common artistic language for expressing religious ideas appropriately (Rynne 1972, 79-99; Ross 1967, 153, 337-338; Green 1984, 137;

Mallory 1989, 24).

6 These were therefore societies that participated, from the Bronze Age onwards, in a network of relationships with other societies (Wells 1995). Bronze Age trade was manipulated by elites and may, like Iron Age trade, have involved kinship links

(Pydyn 1999, 17); it also operated over very long distances (Shennan 1982). It is probable that trade and diplomatic relationships were linked to religion, as in antiquity. It would therefore be reasonable to propose that trade was accompanied by the emergence of a widespread cultural koiné that included religious ideas (Koch

1986; Wadell & Conroy 1999, 129-130).

Such mechanisms are seen in the emergence of the Aegean and Mesoamerican cultures (Renfrew & Shennan 1982, 90). They typically result in the emergence of common cultural patterns, a common language, and shared religious ideas. They also characteristically involve the intensification of production to support both elites and specialists. What they do not require however is the development of either political unification or of an overarching common identity. The various centres may still remain politically autonomous while participating in the wider structure that supports them (Renfrew & Shennan 1982, 90). The requirement that societies must be uniform and politically unified in order to exhibit a common culture is therefore an unnecessary restriction. It also runs counter to a known historical example: the

Greeks.

Prior to the Persian wars Greeks primarily defined their identity by their polis, and expressed it through highly individual and very regional polis religions (Renfrew

1987, 217; Sourvinou-Inwood 1993, 11). Poleis cultures were not uniform; there were substantial differences even in funerary ritual. Poleis were not even the universal form

7 of Greek social organisation, merely the most widespread (Thuc. I.5; Snodgrass 1980,

44). Furthermore the ancient Greek world was never politically unified, and had a multitude of political forms. Had the Greeks been a preliterate culture we might now be debating a Hellenosceptic argument.

Common forms of cult and sanctuary building emerged in the Greek world from the eighth century BC onwards. These were however the consequence of shared trading and diplomatic needs rather than the expression of a common identity (Polignac 1995,

38-39: Morgan 1993, 34: Sourvinou-Inwood 1993, 11-12). Indeed, these common forms may predate the emergence of ‘Greekness’. It is, for example, difficult to see any specifically Greek consciousness in the Odyssey, although the epics can be reasonably taken to represent the political and social realities of the eighth and early seventh centuries BC (Od. 9.175-6; van Wees 1992, 55; Hall J 1997, 46). It is therefore probable that contact with other cultures crystallised ‘Greekness’ from those pre-existing criteria that defined it: a shared language, cultural similarities, and shared religious ideas (Collis 1984b, 62; Long 1986, 131; Renfrew & Bahn 1996, 181). In short, cultural identity preceded ethnic identity, rather than vice versa.

Nevertheless the Hellenic identity remained, in contrast to polis identity, relatively unimportant until the fifth century Persian invasion defined it by its antithesis, the barbarian non-Greek (Persae 434; Long 1986, 131-132; Renfrew 1987, 217; Hall

1989, 7, 16, 60; Hall 1989, 3-6, 9; Morris 1998, 10). Even then, it was not universally accepted; indeed some poleis joined the Persians, while others were swift to abandon

Panhellenism after the threat had passed (Diodorus Siculus XI 2.5-314). Others subordinated it to the older polis identity, thus allowing them to claim that their

8 opponents were less Greek than they were (Thuc. I. 68-71; Rawson 1969, 43-44;

Long 1986, 113, 127, 151-153; Morris 1998, 50). The polis identity was so entrenched that it continued to be a major factor in the Greek world throughout the

Roman period (Magie 1950, 504, 599: Mitchell 1993, 81; Hall 1989, 7).

Culture contact and communal threat were thus important factors in developing the

Hellenic identity. It is possible that some Celtic-speaking populations were developing a common identity for similar reasons. Although the possibility of topai cannot be discounted, one might note Brennus’ disparaging remarks about Greek religion, and Gaulish concerns for ‘the general welfare’ in Bello Gallica (Diodorus

Siculus XXII 9.4; BG VII.2, 4). Nevertheless, this late possibility does not mean that it is possible to speak of a Celtic nation or of a unified people.

It is however possible to suggest that long-term interactions, such as trade and clientship, led to convergent evolution and consequent development of sufficient cultural similarity for archaeologists to be able to use the term ‘Celtic’ as a shorthand.

Linguists are quite content to refer to Common Celtic, a language from which both the

P-Celtic and Q-Celtic language families originated, initially as dialectal variants,

(Olmsted 1979, 107-110; Koch 1981, 21). This suggests mutual comprehension between near neighbours at the very least. There can also be agreement, despite

Celtosceptic arguments, that an object is 'Celtic' rather than Greek or Roman.

Additionally common ideas and treatments of themes appear on Celtic coins, despite widely separated origins, while other cultural elements, such as the La Tène artistic style, have a very wide distribution (Allen 1980, 150). There are therefore shared elements that constituted ‘a community of custom or culture’ (Renfrew & Bahn 1996,

9 181). These shared elements included a deity associated with a wheel and the use of the wheel as a religious symbol (Green 1984, 13, 132). This in turn allows the subject of ‘Celtic’ religion to be addressed in reasonably meaningful terms.

Nevertheless Celtosceptic arguments need to be accommodated. There are, after all, potential problems in comparing archaeological data drawn from different periods and geographical locations. It is necessary to first establish whether common features exist in the Celtic data. This is best done by using the Viennese method of developing independent models from each example, and comparing the results (Karl a, b). If the models match, their common features can then be compared with models derived from 'non-Celtic' sources to see if there are any characteristic differences. If such differences are found, then new lines of inquiry and new interpretations are possible.

It is then up to Celtosceptics and the antinativists of Celtic literature studies to explain the anomalies in their arguments.

10 METHODS OF STRUCTURING TIME IN ANCIENT SOCIETIES

Before considering specifically ‘Celtic’ ideas about time, it is first necessary to consider the evidence for, and general principles of, primitive time measurement. The practicalities of time measurement depend upon culturally-independent physical and astronomical principles, although their actual application may be determined by a culture’s needs, resources and available technology. These may however be similar, given similar environments, similar types of agriculture, and similar levels of technology.

The principle developments in structuring time may be interwoven with the development of agriculture. Simple methods of marking the passage of time, such as the passing of the seasons and the phases of the moon, undoubtedly existed at a much earlier date, and possible simple Palaeolithic lunar counts exist (Renfrew & Bahn

1996, 378). However, the introduction of agriculture made communities dependent upon knowing ‘when’ they were in the year. Agricultural communities need, for example, to be able to accurately predict when the rains will come, when the time for sowing arrives, and when to gather fodder. Farmers also need to estimate how much fodder is needed for feeding animals through the winter, and whether the harvest has produced a surplus that can be used for trade or for making beer (Amarosi et al 1998,

4). This requires a reasonably accurate knowledge of the duration involved, and the anticipated consumption of resources during that duration. This marks a fundamental shift in perception; it is significant that there are no Mesolithic monuments.

11 Measuring and structuring time thus involves an element of prediction for purely practical agricultural purposes, and may therefore be considered an agricultural technology. However, in many societies the agricultural year is also the sacred year; agriculture itself often has a divine origin (CMT 149-161; Macneill 1962, 4-5; Erdoes

& Ortiz 1984, 11-13). Measuring and structuring time is therefore, on both a practical and a religious level, an attempt to gain insights into the workings of nature, and to use those insights in order to discover order and predictability within nature (Gell

1992, 304); it is a form of occult knowledge (Gell 1992, 303). Such ideas may also affect the social order; the divinely established order may be thought to be dependent upon maintaining certain patterns of social behaviour or observing particular taboos.

Conversely, any deviation from the perceived ‘natural’ order, such as poor crop yields, may be given a religious explanation (Shiel 2001, 243-244). It may, for example, be explained as the consequence of religious fault, requiring appropriate remedial steps (NH XVIII, 41-43; AM 17-21; CMT 46-48; Cross & Slover 1936, 99-

100; Kelly 1976, xvii; Erdoes & Ortiz 1984, 173-176; Kelly 1988, 20-21; Gordon

1990, 253). This is implied in Livy’s account of the ‘Bacchic Conspiracy’ (Livy

XXXIX 8.6, 13.10-11, 14.4. 15-16, 18.4-5; Philips 1991, 264).

Structured ideas of time can also affect the social order in other ways. Agriculture makes possible higher population densities (Hassan 1999, 694, 698-699). Larger societies have more resources, particularly in the form of people. Structured ideas about time make it possible to organise those resources more efficiently, thus making more complex societies possible. In turn, the evolution of more complex societies makes the development of more refined systems necessary, simply in order to co-

12 ordinate the available resources (Bogucki 1999, 856-857). There may be an analogy with the role of writing in early societies (Snodgrass 1980, 83; Moreland 1999, 647).

The earliest methods of structuring time appear to have been both lunar and seasonal

(Baity 1973; Thorpe 1982; Chamberlain 1988; Gingerich 1988, 40). There is, however, a discrepancy of approximately eleven days between the tropical4 and lunar years, and a correction mechanism is required to synchronise the two. Various cultures have achieved this using simple non-mathematical rules-based methods and seasonal markers (Gell 1992, 300). For example, the Trobriand Islanders kept their lunar calendar in step with the sun by observing the first appearance of the marine paolo worm; if it did not appear in the lunar month of Milamal, they simply added another month5 (Gingerich 1988, 41). The old Hebrew calendar was similarly synchronised by the Sanhedrin’s judgement on the appearance of the first spring vegetation (Gingerich 1988, 41). Such terrestrial indications are, however, affected by climatic variations, and are therefore extremely inaccurate. Greater consistency can be achieved by marking the beginning of the year with a celestial event instead, such as a solstice or the heliacal rising of a star6.

In practice, there are inherent problems with using heliacal risings in temperate Europe.

The most serious is that, in temperate zones, stars rise up from the horizon at an

4 The tropical year is 365.2425 days in length. The name comes from the Greek term tropoi, or ‘turning point’. It is defined as two successive passages of the Mean Sun at the Mean Vernal Equinox (Neugebauer 1975, 60 n.2, 1082). However, in antiquity it was defined as the period of time between two successive solstices (Neugebauer 1975, 293).

5 That is, the lunar calendar had drifted by approximately one month in respect to the marker for the tropical year.

6 A star is said to be heliacally rising when, after a period of invisibility, it appears just before dawn (Renfrew & Bahn 1996, 381).

13 insufficiently steep angle for accurate observation (Barlai 1988, 436). Constellations can be used to indicate approximate seasons, but this is even less accurate since it involves not one but several stars (Hesiod, Work & Days 379-408, 595-624). One might therefore note the number of megalithic structures orientated to either the summer or winter solstice7. Further, unless these societies were content with an undivided year they presumably used a solstice to mark the start of each year, subdivided the year into lunar months, and reconciled the two using direct observation. This would appear to be suggested by the orientation of some Scottish megalithic structures for low precision observations of the full moon nearest to the summer solstice (Ruggles 1988,

22); high precision observations are not really necessary for such purposes. A 5th millennium model furnace from Slatino, west Bulgaria, may be such a method of reconciling lunar and solar time (Radoslavova 1993, 109-110).

Nevertheless, such methods do not imply the existence of calendars. A calendar is, by definition, a sophisticated mathematical and artificial structuring of time. These are pre-calendrical rules-based systems; they apply a simple system of non-mathematical rules to the direct observations of celestial and terrestrial events.

7 TheT sun’s rising and setting points cannot be used to determine the timing of the equinoxes; the east- west line, which lies midway between the solstices, must be determined first (Sinclair & Sofaer 1993, 182).

14

Figure 1: Chavin de Huantar (Renfrew & Bahn 1996, 393)

There remains the matter of the technology required to apply these basic principles.

One widespread method of marking the limits of the year is to use an aperture to direct a beam of sunlight onto a particular spot on one day of the year. The Zuni of the

American south-west use this technique (Zeilik 1988, 146). The South American site of Chavin de Huantar (Figure 1) is a particularly spectacular example of this principle, which is also found at Newgrange and Maes Howe (Figure 2; Ruggles 1988, 14-15;

Renfrew & Bahn 1996, 393). Alternatively the sun itself can be framed by a structure as it rises on a particular day, as at Stonehenge. There is an element of religious theatre in such monuments, which also illustrates the ability of such societies to plan and direct large amounts of resources to communal ends.

Figure 2: Newgrange (Renfrew & Bahn 1996, 385)

15 Nevertheless, such methods can only mark one particular annual solar point.

Structures such as Newgrange and Chavin de Huantar are therefore relatively inflexible in use, despite their impressive appearance. Although some may synchronise solar and lunar time, they cannot, of themselves, be used to subdivide the tropical year further, and thus mark major agricultural phases such as harvest-time. It could therefore be argued that such monuments are exceptional precisely because they were both costly of resources and relatively inflexible in use. They were effectively an evolutionary dead-end in time measurement.

The most primitive method for marking the agricultural phases of the tropical year is the use of local horizon marks, which is also found amongst the Zuni (Zeilik 1988,

146, 149). An activity or festival is indicated by the sun rising above a particular horizon feature, as observed from a standardised watching point. This requires a suitably open landscape with a horizon possessing singular features (Zeilik 1988,

146). There are several probable megalithic examples of structures using local horizon marks, such as the observing platform at Brainport Bay, Argyll (Ruggles 1988, 15;

Hicks 1988, 478). Nevertheless, horizon systems are unsuited to many locations; it is for example impossible to use them if the horizon is either featureless or occluded by trees. These problems can, however, be overcome by marking celestial alignments with structures of stone or wood, either directly or by the use of shadow.

The use of shadow is a particularly flexible technique. The structure required, a simple pillar (a gnomon) and markers, is simpler, far less labour intensive than a

Stonehenge, and thus more suited for small, decentralised societies, or societies that are mobile to some degree. Additionally it can be used to indicate a variety of solar

16 positions throughout the year. Single standing stones are widespread throughout

Europe, while pillars are a marked feature of Celtic iconography. The method was used in the Graeco-Roman world, the most sophisticated example being Augustus’ sundial (Mor. 410-411; Vitruvius, De Arch. IX, i.1; Neugebauer 1975, 737-748, 615-

624; Claridge 1998, 190-192).

Figure 3: Arrangement of solstitial sunrise and sunset points around a gnomon

In its simplest form, the desired occasion is indicated by a shadow cast by a pillar at sunrise on a particular day falling on a given marker (Figure 3). However, this method immediately presents problems. When exactly does sunrise occur? Is it to be taken as the point when the first flash of the sun appears above the horizon, or when the sun's orb first clears the horizon? Errors in construction and observation can cause considerable inaccuracy, although this may neither matter nor be observable to a simple and fairly local agricultural community; in any case inaccuracies are easily corrected in a rules-based system (Danaher 1981, 221-222).

17 Equally importantly, the sunrise method requires either a high point such as a hilltop, or a landscape that is open to the horizon, such as moorland. Even a nearby hill will throw out the first appearance of the sun by an appreciable amount, although the error will be consistent (Barlai 1988, 438). However, it is impossible to use sunrise observations in wooded or urban landscapes, where the appearance of dawn will be hidden.

An alternative, more technically advanced method is to mark the length of the shadow at noon, when the sun is at its highest; the length of the shadow varies with the seasons. Noon observations have other advantages, in addition to overcoming the limitations of a local horizon. From the observer’s point of view it means that they do not have to get up quite so early, which would no doubt appeal to an almost universal human desire. More importantly the method is capable of considerable accuracy, even by modern standards. All that is required is the technological refinement of the shape and dimensions of the gnomon. This method was used by the later Greek calendar makers for their observations. Of all the contemporary methods available, it was the only one capable of sufficient accuracy for calendar construction.

Calendars enable societies to create finer divisions of time, thus enhancing the organisation of resources. However there are problems inherent in all calendars, despite their sophistication. Rules-based methods are automatically self-correcting; calendars however are not. A solar calendar must contain 365 days; however the tropical year is 365.2425 days long. A mathematical mechanism is therefore needed to correct for this discrepancy. The modern calendar achieves this by using a 400-year cycle and a leap year rule to accommodate the difference. Leap year and leap day rules

18 are thus characteristic of calendars; they are unnecessary in rules-based systems.

Furthermore, the use of such approximations means that calendars contain inherent minor errors that are cumulative with the passage of time, leading to a growing discrepancy with the seasons. This calendar drift cannot occur in rules-based systems.

There are further practical problems. The tropical year is best subdivided by the use of lunar months. Although it is relatively easy to reconcile solar and lunar time with rules- based systems, it is far more difficult to do so with calendars. Calendars need an accurate long-term rule to equate X lunar synodic months8 with Y tropical years, while simultaneously correcting for the approximation used for the lunar component

(Neugebauer 1975, 614). This is more complex than it seems. The synodic month is approximately 29.5 days in length. Since the month needs to be a whole number of days, most calendars compromised by using alternating months of twenty-nine and thirty days

(Neugebauer 1975, 353). However, a year composed of six months of each type is 354 days long, or approximately eleven days shorter than the tropical year. A calendar based on lunar divisions will therefore very rapidly get out of step with the seasons unless some mechanism is found to periodically add an extra month of thirty days whenever the accumulated error amounts to approximately a month. Furthermore, a lunar year of 354 days is shorter than the true lunar year by 0.36706 days; it is therefore necessary have to insert one leap day every third year to keep in step with the phases of the moon. This may be done by adding an extra day to a month of twenty-nine days.

8 The mean interval between consecutive conjunctions of the moon and sun. It corresponds to the cycle of lunar phases. See Weisstein 2003, for a more advanced definition.

19 However, it is not simply a matter of mathematics. Before a rule can even be derived, the exact lengths of both the tropical year and the synodic month must be measured as accurately as possible. The length of the tropical year was, in antiquity, established by observing the number of days between successive summer solstices, as indicated by the shortest shadow cast by a gnomon at noon (Neugebauer 1975, 293). Since this period varies slightly, it requires many years of repeated observations and accurate equipment to obtain an accurate mean value (Neugebauer 1975, 61-62). The length of the synodic month also had to be established by similarly repeated measurements. The development of soli-lunar calendars was thus a significant achievement in both applied mathematics and accurate astronomical measurement. It was, for the ancient world, cutting edge technology requiring the most skilled individuals available.

One of the earliest attempts to reconcile solar and lunar time was the octaetris, attributed variously by Censorinus to either Cleostratus of Tenedos (c. 500 BC) or Eudoxus of

Cnidus (390-c.340 BC) (Neugebauer 1975, 620). However, a similar cycle was used in

Babylon between 529 and 504 BC.

The octaetris used an eight year cycle, with five ordinary years of twelve months each and three leap years of thirteen months each. Since, in the sixth century BC, the length of the year was accepted to be 365 days, the octaetris contained 8 x 365, or 2,920 days. This was very close to the total of 99 lunations (99x29.5 = 2,920.5 days). However, the shortcomings of the octaetris were revealed in the fourth century BC by more accurate measurements of the length of the year.

20 The most well-known and successful example of a soli-lunar calendar is the 19-year

Metonic cycle, described by Meton of Athens in 432 BC (Neugebauer 1975, 622-623). It had previously been discovered by the Babylonians (Neugebauer 1975, 354-357). A long-term rule equated nineteen tropical years, totalling 6939.61 days, with 235 synodic months, totalling 6939.69 days. Within the cycle, 110 of the months are 'hollow' months of twenty-nine days and 125 are 'full' months of thirty days. The cycle is divided into twelve years of twelve months, and seven years of thirteen months; extra months are added in years 3, 5, 8, 11, 13, 16, and 19. It is a convenient and accurate cycle; the total number of days in the cycle comes to (110 x 29) + (125 x 30), or 6,940 days, which is very close to the required values. Furthermore it gives, using the mean value of the synodic month (29.530588 days), a mean tropical year of 365.2504 days, which is again very close to the actual value.

21 ARCHAEOLOGICAL EVIDENCE FOR IDEAS ABOUT TIME IN THE

CELTIC WORLD

The east-west grave alignments of the east and central European Neolithic and Copper

Age suggest the early importance of the solar path (Barlai 1988, 436). Since it has long been assumed that Celtic wheel symbolism represents a solar cult, it would be uncontroversial to suggest that pre-Celtic and Celtic people used the tropical year, and that solar symbolism might appear in various ritual practices. Nevertheless, such practices do not, in themselves tell us anything about the structuring of time; they only illustrate a particular cultic application of associations based on those ideas and their religious interpretation.

The available evidence for Celtic ideas about structuring time is late, but nevertheless valuable. The related Coligny and Antres calendars reveal interesting insights, given an understanding of the practicalities of time measurement. The Irish sources additionally mention the oenach or festivals. The antinativist school in Celtic literature would, however, argue that the literature has been so contaminated by

Christian influence that any deductions are meaningless (McCone 1990). It is beyond the scope of this paper to counter antinativist arguments. It is nevertheless sufficient to note that a system of festivals similar to the Irish oenach can be discerned both in the Coligny Calendar and in Roman Britain, at a period when the argument of

Christian influence does not apply (Olmsted 1979, 114-116; 1992, 90; Davies 1984,

113, 119; Millett 1994, 40).

22 The Coligny Calendar

The Coligny calendar was discovered in 1897, at Coligny near Bourgs (Ain) (Olmsted

1992, 16). This was originally the territory of the Sequani, a centre of Celtic culture that lay on the principle route to Masallia and the Mediterranean (Plate 39; McCluskie

1990, 105).

The calendar, which was accompanied by a statue of Mars dated to AD 50-150, is a religious soli-lunar calendar using alternating months of twenty-nine and thirty days; the mathematical mechanism is firmly based on Greek work (Mac Neill 1928, 28, 62;

Olmsted 1992, 16, 70). It is probably Hellenistic in origin, despite the first century

AD date of the bronze plates. The Hellenistic origin is suggested by the Greek-derived orthography, and the use of archaic forms of the language (Mac Neill 1928, 6;

Olmsted 1992, 72-73). Additionally there is, despite the use of the Latin alphabet and enumeration, a complete lack of Latin, and a lack of any reference to the Julian system or the imperial calendar (Olmsted 1992, 71). The first two features might be anticipated in any calendar created after Caesar’s conquest, and all would be expected in any system created during or after the Augustan period.

Since the underlying mathematical system is Greek-derived, references can be made to Greek manuscripts on calendrical systems. It can, for example, be assumed that the length of the tropical year was measured using noon observations of a pillar or gnomon, since this was the only available accurate contemporary method for doing so

(Neugebauer 1975, 293; q.v. pages 14 n. 1, 19). Nevertheless, the Coligny calendar differs in some fundamentals from any recorded Greek calendar, despite the use of

23 Greek technology. Olmsted’s conclusion, that it initially used a thirty-year cycle that was later refined to twenty-five years, appears to be convincing (Mac Neill 1928, 27;

Olmsted 1992, 17, 103-106). There are, however, no parallels for either cycle in the

Greek world, suggesting that the Sequani applied Greek methods independently. This conclusion is supported by the various correction systems, such as the TII marks, which again have no Greek parallels (Olmsted 1992, 70, 117-119). Similarly the value for the mean year9 is, in the twenty-five year cycle, more accurate than those given by

Greek systems; 365.2422 days as opposed to 365.2504 days for the Metonic calendar

(Olmsted 1992, 3). This again suggests independent work, which of necessity involved many years of repeated observation and accurate measurement. Since the astronomical and mathematical methods were, by contemporary standards, complex, the development of the Coligny calendar was a major intellectual accomplishment

(Heath 1931, 14-15, 20-24, 29-32).

However there is more to the Coligny Calendar than an indigenous application of

Greek calendar theory and mathematics, no matter how advanced. The principal festivals are separated by a quarter year (McCluskey 1990, 105); this pattern corresponds to the Irish oenach (Olmsted 1979, 114-116; 1992, 90). Although

Olmsted remarks upon this, he is almost certainly incorrect in suggesting that these festivals are just the solstices displaced by calendar drift (q.v. page 19; Olmsted 1992,

89, 90, 130-134). It is an understandable error; the festivals are similarly fixed with respect to the solar year. However, calendar drift required for such displacement would mean that the Sequani developed a mathematically advanced soli-lunar

9 The mean year is taken as being the number of days in a complete cycle divided by the number of tropical years in the cycle.

24 calendar in 850 BC-1100 BC, which may be as much as six centuries before even the earliest Babylonian example (Olmsted 1992, 90). Further, the solstices are fairly insignificant in the Irish sources, and the midsummer fire festival may even be a

Scandinavian introduction.

Given the lack of evidence, Olmsted’s assumption would appear to be untenable.

Unfortunately it means that he has equated Samonios 1 with the midwinter solstice, despite observing that Samonios ends the summer and begins the winter half of the year (Olmsted 1992, 89, 91). His reconstruction makes far more sense if Samonios 1, like Samhain10, corresponds to November 1 in the modern calendar.

Olmsted’s only supporting argument for equating Samonios 1 with the midwinter solstice was that he found the pattern of the Irish oenach otherwise inexplicable

(Olmsted 1992, 90). In fact they are a practical, rules-based method of crudely dividing the solar year into quarters for agricultural and religious purposes (Danaher

1981, 220, 222-223; Mackie 1988, 228). Since the oenach fall midway between the solstices and the equinoxes, the system is easily maintained by the direct observation of solar events (Danaher 1981, 221-222).

There is really only one way of achieving this pattern with the technology available; stick a pillar in the ground and place markers at points equidistant between the equinoctial line and solstices (Figure 4; c.f. Mackie 1988, 226). The oenach are of necessity paired; the appropriate pair of festivals is indicated when the tip of the shadow cast by the rising sun falls on a particular marker, while the specific festival is

10 Samhain=end of summer, i.e. start of winter.

25 indicated by the shadow travelling either towards or away from the relevant solstice point. The existence of an identical pattern of festivals in the Coligny calendar would therefore similarly suggest an earlier pre-calendrical, rules-based system, also based on the use of a gnomon. The arrangement of the festivals around the gnomon is reminiscent of a wheel, suggesting the hypothesis that wheel symbolism represents the solar path.

Figure 4: Use of a gnomon, markers and shadow to determine the festivals from their sunrise and sunset points

Further observations can also be made about the Coligny calendar. Firstly, it splits the year into a waning and waxing half (Olmsted 1992, 76). Although there is a Greek parallel, an identical division of the year was, interestingly, used until quite late in

Ireland and Scotland (Olmsted 1992, 57, 193). Secondly, the names of the months in the Coligny calendar are entirely indigenous and owe nothing to Greek influence

(Olmsted 1992, 190-201). This suggests that the rules-based system divided the

26 tropical year into lunar months and, of necessity, reconciled the two by direct observations of celestial events. If the solstice, as indicated by the shadow cast by a pillar, did not occur in the expected lunar month, another month was added; a 30-year cycle emerges automatically from this procedure, although other cycles are possible

(Figure 5). The association of solar and lunar imagery in Celtic coins should therefore be noted (Figure 7).

Figure 5: Intervals in a 30-year rules-based cycle

This system, in its earliest form, is almost certainly pre-Celtic. Neolithic and Bronze

Age structures aligned to the solstices are far too numerous to list, and single standing stones are common in north-west Europe. The division of the tropical year into eighths is a logical extension of such technology. Similarly the lunar subdivision of the tropical year is both widespread and well attested in both the archaeological and anthropological literature, while the mechanics of equating the two are well

27 understood. The technology also potentially existed in the Celtic world; pillars of wood and stone are common in religious contexts, and archaeological examples are well known. The ‘royal’ site of Emain Macha, for example had a massive free- standing central oak pillar at the centre of a wheel-like pattern (Raftery B 1989, 141;

Lynn 1992, 43; Green 1997, 56). One might also note Caesar’s remarks about the druids’ interest in, and discussion of, the movements of the heavens (BG VI.14).

Figure 6: Aristotle, Meteor. II.6 (Kidd 1999, 197)

Now although this argument is attractive, the pattern of festivals might equally be contemporary with the creation of the Coligny calendar. After all, ‘shadow clocks’

28 were well known in the Graeco-Roman world, and diagrams similar to Figures 3 and

4 are described by both Aristotle and Posidonius (Figure 6; Aristotle Meteor. II.6;

Posidonius 137a; Mor. 410-411; Vitruvius De Arch. IX, 1.1; Neugebauer 1975, 737-

748, 615-624; Claridge 1998, 190-192). This argument would, however, presuppose that such knowledge was exclusively Greek. In fact it depends only upon celestial mechanics, and would have been observed by anyone noting the annual movement of the shadow cast by a pillar. It is a fairly simple step to then divide the tropical year into agricultural quarters by marking positions midway between the solstices and the equinoctial line. I have seen one Oklahoman woman, using Native American knowledge, use this same method for determining planting times; in her case she applied it to gardening rather than agriculture, but the same principle applies. This method does not, however, appear in Greek texts.

If it were, however, to be accepted that festivals in the Coligny calendar were contemporary with the application of Greek technology, one is left trying to explain their arrangement. They have no Greek equivalents; Aristotle’s diagram, for example, is concerned only with the solstices and winds (Figure 6). Furthermore, the Irish festivals are generally accepted, on several grounds, to be the equivalent of the festivals in the Coligny calendar. There are no indications of any calendrical system in

Ireland prior to the introduction of Christianity, let alone a Greek-derived soli-lunar calendar (Danaher 1981, 222). Even the antinativists have not gone so far as to suggest a Christian or Classical origin for the oenach. An indigenous pre-calendrical system is therefore most likely.

29 It might be objected that there is no direct archaeological evidence for this method.

However, the archaeological footprint would be small. Although stone pillars can be substantial and difficult to remove, the markers need be no more than a billet of wood, a suitably sized stone, or even a cut in the sod filled with chalk. Their removal would be unlikely to leave any record, and without the markers there is no evidence as to the structure’s use. Furthermore, posts do not need to be stone; a wooden post would leave no evidence except for a posthole of unknown function, as at Navan Fort.

Nevertheless, corroborative evidence can found by examining the structure of the festivals. In a post-and-marker system duration is marked by a shadow moving across a surface from one marker to another. The eccentricity in the earth’s orbit means that the period from the winter to the summer solstice is shorter than that between the summer and winter solstices (Neugebauer 1975, 57-58; Vilhjámsson 1993, 73).

Consequently, while the physical distance between the markers is equal, the duration it expresses is not; the spacing of any festivals derived from this technology will

Figure 7: Tolna or Kapostal type coins showing analemma (Allen 1987, vol. 1, Plate XXII, S92-S96); with horse-wheel symbolism for comparison.

30 directly reflect this orbital eccentricity by slight inequalities in the two halves of the measured year. This is seen in both the Coligny calendar and the Irish oenach; the

Samhain-Beltane period is three days shorter than the Beltane-Samhain period

(Danaher 1981, 219). The spacing of the festivals therefore conforms to what would be expected from a post-and-marker system.

Figure 8: (A) Analemma traced over a year by the shadow cast by noon sun; the summer loop is nearest the gnomon (B) Modern analemma curve showing degree by which sun deviates from its mean position throughout the year (Urschel 2003)

There is also the evidence of coins from the Danube basin, which substitute an analemma for the wheel symbol in horse-and-wheel symbolism (Figure 7); the design also appears in degenerate form (Allen 1987, vol. 1, Plate V, 78, 79; Plate VI 98). The analemma is well known to astronomers, and its use strongly confirms the identification of the horse as a solar symbol; it was used exclusively in Greek astronomy to set up vertical pillar sundials (Vitruvius, De Arch. IX 1.1; Sawyer 1995;

University of Denver 1999; Holtz 2002; Urschel 2003).

31 The annual movement of the sun, as traced by a shadow cast by the rising sun, produces the cross or wheel patterns seen in Figures 3 and 4. However, if daily observations are made at noon, variations in the movement of the sun cause the shadow to annually trace a figure-of-eight pattern, the analemma (Figure 8). The symbolism of the Danubian coins is therefore entirely consistent with the observational methodology necessary to create the Coligny calendar. Further, the coins associate lunar symbols with the horse and the analemma, strongly indicating soli-lunar time division. This does not necessarily indicate a calendar, although it would be interesting if one were to be found in this region.

Finally, the Bronze Age Etzelsdorf cone suggests that the use of a

gnomon predates the emergence of the Celts as a distinct group

(Figure 9; Harding 2000, 348-349). Similar cones are known

from Avanton (Poitou), and Schifferstadt (Rheinland-Pfalz), and

are reported to have been found in 17th-18th century Ireland

(Harding 2000, 348-349; Paterson 17 March 2002).

The Etzelsdorf cone is almost certainly a religious artefact; it is

highly decorated, made of gold, and apparently lacking in any

practical function. The shape suggests a model gnomon11; the

shape is ideal for time measurement. Further, Menghin’s analysis

Figure 9: suggests that the iconography represents the 19-year soli-lunar Etzelsdorf cone (Green 1997, 57)

11 The suggestion that these artefacts are priests’ hats is very dubious (Harding 2000, 348-349). One may wear a traffic cone, but that does not make it a hat. Has anyone tried the experiment of wearing them?

32 cycle (Paterson 17 March 2002); this cycle is, however, unhelpfully described as the

Metonic cycle. It should however be noted that an approximate 19-year cycle may emerge from rules-based systems. Since there are no suggestions of the leap year or leap day systems required by mathematical calendars, the cones do not describe the

Metonic cycle, but a 19-year rules-based soli-lunar cycle, similar in principal to the underlying system proposed for the Coligny calendar.

There is therefore strong evidence that the Coligny calendar incorporated an earlier rules-based system. Such a system would be suitable for a fairly unstratified, possibly segmentary society, with small-scale settlement patterns, such as those from which the early Bronze Age ranked societies of central, northern and north-western Europe developed (Renfrew & Shennan 1982, 10; Burl 1989, 145). The pillars are easily erected by a single agricultural community. Furthermore, the principles are extremely simple, and could have been either communicated by group interactions or independently discovered at different times by several groups.

In contrast to the rules-based system, the refinements seen in the Coligny calendar indicate specialists and specialist knowledge. Such knowledge takes time to learn and time to put into practice, during which time specialists need to be supported. This suggests the availability of surplus resources. Further, one might expect such a major religious project to be performed on behalf of a communal sponsor, such as a ruler or an elite group. This suggests a society that is both stratified, and developing aspects of centralisation, which in turn makes such innovations necessary. Increasing stratification, centralisation and specialisation in Celtic societies is seen in the

33 archaeological record, particularly as consequence of contact with Hellenistic culture

(Filip 1978, 429; Collis 1984a, 87, 188; Millett 1990, 25-29).

To recap, a gnomon is necessary for the construction of the Coligny calendar; it could not have been devised without one. The Danubian coins indicate that the necessary techniques were known by other contemporary groups. However, the Etzelsdorf cone suggests that the use of pillars in time measurement may date back to at least the early

Bronze Age. The evidence further suggests that the Coligny calendar was a methodological and mathematical refinement of an earlier and archaic rules-based system. These used sunrise, rather than noon, observations; consequently the festivals and solstices form a wheel-like pattern around a gnomon, suggesting the hypothesis that wheel symbolism represented the solar path. The Danubian coins, by replacing the wheel with an analemma in horse-and-wheel symbolism, suggest that the two symbols are equivalent. Since the analemma represents the annual solar path, it again suggests the hypothesis that the wheel may represent the solar path.

The Jupiter-Giant Columns

In the post-and-marker system the solar path is represented by a wheel-like pattern of shadow cast by a pillar. In the Roman period wheel symbolism is associated with pillars, most notably with the Jupiter-Giant columns.

The Jupiter-Giant columns are predominantly a phenomenon of eastern Gaul and

Germany, although at least three probable British examples are known (Haverfield

1917, 191; Phillips 1976; Henig 1986, 161; Green 1986b, 66; Henig 1993, 8-9, no. 18;

White & Barker 1998, 18, 94). Concentrations are found in Upper and Lower

34 Germany, Raetia and Belgic Gaul; they are especially common in Middle Rhineland, both east and west of the river (Plate 37; Green 1984, 133). The centre of the main scatter is the territory of the Gaulish tribe of Lingones, neighbours of the Aedui (Plate

39; Green 1984, 133). The majority of columns appear to span the period between the first and fourth centuries AD; the earliest may have been erected in the Flavian period at Mainz (Green 1984, 132). Many were publicly erected in cities, although others were located near villas or at remote sanctuaries such as Le Donan (Green 1984, 179).

The dedicants were not normally soldiers, but members of the local population (Green

1984, 179).

Some columns are decorated to imitate trees, as at Hausen, and the columns are generally considered to represent a tree (Green 1984, 111; 1986b, 67); Powell suggested that they may have had native forerunners in wood (Powell 1958, 164).

However, this does not mean that the forerunners were trees; they may only have been thought of in such terms.

The summit group often consists of a rider-god, frequently carrying a wheel, riding down a giant (Green 1976, 10; 1986b, 67). The wheel is undoubtedly the symbol of the Celtic Jupiter and is so completely identified with the deity that it sometimes appears alone, as a substitute for the god (Green 1981, 111; 1984, 168). The Mouhet giant, for example, kneels in the attitude of a captive, a wheel balanced on his back

(Green 1986a, 59).

35 The giant is portrayed as being either snake-limbed or fish-tailed (Green 1986a, 194).

Since snakes appear to be chthonic12 in Romano-Celtic iconography, the imagery may either convey the imposition of celestial upon chthonic forces, or their marriage

(Green 1981, 111; 1986, 185-186; Green M J A 1996, 136); the two are not necessarily mutually exclusive. It has therefore been suggested that elements of the tale of the Gigantomachy are involved; however, this suggestion may unduly overstress the Classical elements (Green 1984, 175). Nevertheless, it is necessary to consider this possibility, considering the rider god is dualled with Jupiter. It is probable that, for example, the representation of a cuirassed rider may owe something to both Roman military tombstones and Greek influences (Mackintosh 1995, 2, 65).

Figure 10: Horse and wheel, Kalleby, Bohuslän, Sweden (Glob 1974, 151)

There are some very non-Classical features in the iconography of the Jupiter-Giant columns, despite the use of Roman art forms and technology (Green 1976, 8-10, 119).

12 One might note in passing that the conflation of the subterranean and the aquatic is suggested by Romano-Celtic contexts; pits and wells, for example, have similar functions (Green 1984, 95, 211; 1986, 101-102, 145, 156).

36 The snake element, for example, is seen in other very non-Classical contexts (Green

1984, 175, 186-187). Similarly, the Classical Jupiter was never equestrian, and is never shown as equestrian when the god is presented in a Classical guise (Green 1984,

175-176). The rider-god also occasionally has the attributes of Mars, rather than

Jupiter, which suggests some very non-Classical ideas about divinity (Green 1976, 10,

30); presumably the indigenous deity had properties of both (c.f. Wightman 1986,

559). Furthermore, the combination of wheel and horse is a widespread motif of the pre-Roman Celtic coins; since the wheel can substitute for the god the imagery is effectively identical. A Bronze Age petroglyph from Bohuslän additionally suggests that the imagery may predate the emergence of the Celts as an identifiable group

(Figure 10). There are similarities between Scandinavia and Central Europe during the

Urnfield period, probably due to contact through the amber trade (Shennan 1982;

Green 1984, 20; Pydyn 1999, 17). The iconography of the Jupiter-Giant columns may therefore be reasonably interpreted as both indigenous and ancient, despite the use of

Roman technology and artistic styles.

The wheel-god’s horse has been identified as a solar animal (Green 1984, 192). It has been therefore suggested that since the wheel is a solar symbol and the horse a solar animal, the deity is therefore a solar god. This is, at first sight, an attractive proposition. Since the sun was used to structure the year, it would suggest a direct iconographical connection with time. However, the identification rests upon two assumptions; that the wheel is a solar symbol, and that the horse is a substitute for the wheel, and therefore identical with the god (Green 1986a, 59). The first has not been proven; the second is almost certainly incorrect. As any rider would observe, horse

37 and rider are different entities; the rider masters the horse. If the rider is not in control, they are in deep trouble, as any Roman or Celtic cavalryman would have known.

The horse is the theriomorphic form of the goddess Epona, whose distribution in eastern Gaul coincides almost exactly with that of the equestrian Romano-Celtic

Jupiter (Green 1984, 196, 301; 1994, 54; Plates 37, 38). There is a further connection in that Epona and the Jupiter-Giant columns are both associated with water, as at

Allerey (Green 1984, 196). It cannot be argued that the dedicants were unaware of

Epona’s symbolism, since these monuments were largely erected by and for local people; they would have automatically equated the horse with the goddess.

Furthermore, there are no reasons to suggest that the eastern Gaulish worshippers of

Epona were culturally unrelated to those of the rider-god. One is therefore left with the deduction that, in eastern Gaul at least, the rider-god’s horse was seen as Epona.

The equation may have been more widespread; Epona was adopted by the Roman cavalry and she was sufficiently well-known to appear in Roman literature (Metam.

III.27).

The summit group is therefore best interpreted as a composite of the Romano-Celtic sky god and Epona. It is probably a hieros gamos; paired deities are common in

Romano-Celtic contexts (Green 1984, 116, 174). Indeed, large numbers of the columns are dedicated to Juno, spouse of Jupiter, which strongly suggests that such pillars were perceived to be associated with a goddess and consort, despite the summit image of the god (Green 1984, 174). It may therefore be incorrect to emphasise the

Jupiter aspect, simply because it is the only human image present.

38 Celtic horse symbolism is generally accepted, on several grounds, to be solar. The

Jupiter-Giant columns’ imagery of paired deities can therefore be interpreted as the wheel-god controlling the sun in the same way that a rider controls a horse. This is consistent with the identification with Jupiter, and again suggests that the god’s attribute, the wheel, represents that control by representing the solar path that defines the sun’s limits. However this runs counter to the generally accepted view that the wheel-god is a sun god. Further, since the horse is identified with Epona, at least in eastern Gaul, the idea of an originally solar goddess must be considered. This however presents a few problems. While Epona’s association with the period of the midwinter solstice suggests some sort of solar connection, she has, in the Roman period, no obvious solar connections beyond this (Wightman 1986, 559; Green 1989,

23). On the contrary, she has chthonic/aquatic/fertility associations. There are therefore several possibilities:

• The Celtic horse was not solar.

• Epona, although identified with the horse, was never solar, unlike other Celtic

horse symbols.

• Epona lost her solar functions early in the Roman period.

It is unlikely that Celtic horse symbolism was non-solar. The horse is widely associated with naturalistic representations of the sun and celestial imagery; the

Danubian analemmas mentioned earlier further confirm a solar identification. It is equally unlikely that Epona was an exception to Celtic horse symbolism. She was

39 prominent in Aeduan territory; the horse symbolism on Aeduan coins shows no significant deviation from similar imagery on other Celtic coins (Figure 11; Plate 8,

78-89; Plate 11, 128-133, 135; Oaks 1986, 77). It is therefore most probable that

Epona was, despite her chthonic associations, an originally solar goddess who lost her solar attributes early in the Roman period.

Figure 11: Aeduan silver coin, 80-50 BC (Compagnie Generale de Bourse 2003)

This leads to an interesting deduction. The parallels between Epona and an Irish solar goddess, Macha, have observed on several occasions (Olmsted 2001, 158-159; 1979,

139; Oaks 1986, 78; Sessle 1994, 10-11). Macha additionally has chthonic/submarine/fertility associations, and the two are almost certainly linked by the Rom inscription (Olmsted 1988, 317, 354-355; 1994, 158-159). Macha’s solar attributes are linked to female ‘tribal’ ruling qualities; these were anathema to both

Roman mos maiorum and Roman rule. The loss of such attributes would leave chthonic/fertility goddesses, such as the Matres that Epona resembles (c.f. Oaks 1986,

82). The cult form of the Matres, although linked to local communities, do not appear to predate the Claudio-Neronian dynasty (Derks 1998, 123, 124, 128 n. 245), thus

40 suggesting Romanised deities. Epona’s horse probably survived however, minus its solar meaning, due to her adoption by the cavalry.

It is thus most probable that the meaning of the summit group was Romanised by changes in the function of the goddess, and that the Romano-Celtic imagery is not therefore identical in meaning to the Iron Age imagery. Nevertheless, the iconography of the Jupiter-Giant columns retains the original connection with time suggested by the Iron Age imagery. The bases of some columns feature carvings of the deities of the days of the week, while others carry carvings of Sol and Luna, or of the four seasons (Bauchhenss 1981, Tables 18, 19; Huskinson 1994, 3). These are new

Romanised forms; the four seasons were, for example, a popular Roman motif.

Similarly Sol and Luna are obvious Latin deities, although their association does suggest soli-lunar calendars such as the Coligny calendar. The concept of the seven- day week is a Middle Eastern idea that was universally adopted by the Roman world during the first century BC (Sarton 1959, 328). Since the imagery of time could no longer be incorporated in the Romanised interpretation of the summit group, it has been given a new Roman form and displaced to the base, a presumably symbolically appropriate position. Nevertheless, it should be noted that the summit group made the columns useless for actually measuring time. Consequently they present time as an abstract concept, as part of a wider religious message involving cosmology and fertility.

Summary of Analysis

The results of the analysis can be summarised as follows:

41 • The Coligny calendar and the Antres calendar are associated with a deity who

is identified with the Classical Mars. The rider god, a celestial deity, is dualled

with Mars as well as Jupiter. Celestial events were used to structure time.

• The Romano-Celtic Mars has marked associations with fertility. This

association is strengthened in the Jupiter-Giant columns by the representation

of a probable hieros gamos involving Epona, who was also associated with

fertility. Both the Antres and the Coligny calendars, and the earlier rules-based

system underlying the Coligny calendar, structure the agricultural year.

• The Coligny calendar, and the earlier rules-based system, used a pillar to

structure time. The iconography of the Jupiter-Giant columns associates pillars

with abstract concepts of time.

• The summit group of the Jupiter-Giant columns features wheel symbolism,

and probably represents celestial control over the originally solar horse. In the

rules-based system underlying the Coligny Calendar the shadow cast by the

rising sun indicates the solstices and festivals, which form a wheel-like pattern

around a pillar.

The Coligny calendar and the Jupiter-Giant columns therefore share, on a cognitive level, identical features. The only difference lies in the Jupiter-Giant columns conveying in abstract what the Coligny calendar does practically. However, in the imperial period indigenous methods of structuring time would have been superseded by the imperial calendar.

42 Figure 12: Etruscan cista featuring Classical wheel symbolism, Museum of Volterra (Cook 1925, 261)

The test however lies in comparing this data to non-Celtic material. While the

Classical Mars is associated with agricultural fertility, he is not associated with wheels, pillars or calendars. Although wheels appear in Greek graves of the Archaic period, and as votives at Delos and Kameiros, they are not the attribute of any Greek deity (Green 1984, 75). Similarly, while wheels appear in Greek religious art, again they are not the attribute of a deity (Figure 12); instead they are associated, for example, with the punishment of Ixion (Figure 13; Cook 1925, 253, 198-205).

The association of wheels with pillars is similarly not seen in non-Celtic contexts.

Gnomon, although widely known in antiquity and associated with calendars, are similarly not overtly connected with agricultural fertility outside Celtic contexts.

Again, while Roman pillars may have a summit statue of an emperor, the pillar is purely functional, to gain height for the image. Although such pillars may have

43

Figure 13: Ixion bound to the wheel (Cook 1925, 205) influenced the development of the Jupiter-Giant columns, they do not convey ideas about time or agricultural fertility, nor are they decorated with wheels (Wightman

1986, 555). Finally, rider gods in other parts of the Empire are not associated with wheels or pillars (Delemen 1999, 80, 83). This is therefore a specific, non-Classical complex that can reasonably be described as ‘Celtic’, even if its origins may lie in a pre-Celtic period.

44 WHEEL SYMBOLISM

The evidence suggests that wheel symbolism may represent the solar path rather than the sun. It is therefore necessary to analyse wheel symbolism more deeply in order to either prove or disprove this hypothesis. If the wheel is a sun symbol, the choice of spoke number is unlikely to be purposeful, although high spoke numbers would be expected. It is the appearance of the symbol, rather than the number of spokes, that suggests the sun. Similarly, if the sun was identified with the wheel, one might expect wheel symbolism to follow the pattern of real wheels.

A Statistical Analysis of Wheel Symbolism

Wheel symbolism appears early in the archaeological record; four-spoked examples appear in central Europe during the second millennium BC (Green 1984, 17). They are plentiful in both the middle and late Bronze Age, the petroglyphs of Scandinavia and North Italy being notable and apparently closely linked examples (Green 1984,

19, 20, 23, 24). They continue to be widespread in Iron Age and Roman contexts

(Green 1984, 33, 47-48).

The table and pie chart at Tables 1 and 2 summarises data provided by Green in her seminal study of wheel symbolism (Green 1984). Over two hundred lead four-spoked wheel symbols from Lavoye were mentioned in the original survey, but not included

(Chenet 1919, 246; Green 1984, 288). Since there appears to be no good argument for ignoring them, they have been included in this analysis, and given a nominal quantity

45 of 200. This nominal value may result in some slight underestimation of four-spoked wheel symbols.

The statistical pattern is interesting; four-, six- and eight-spoked wheel symbols make up 94.30% of the sample (Table 2). However, the raw data is worth investigating further. Firstly the sample apparently assumes that everything round is, of necessity a wheel symbol. The nine-spoked example from Vienne (Isere) is very dubious; it is one of five barrel-shaped objects that radiate from another barrel (Plate 1, fig 1; Green

1984, 129). Barrels and wheels are not necessarily equivalent, and this example should have been excluded. Similarly the example from Bois-de-la-Neuve-Grange

(Plate 1, fig. 2) could only be called a wheel by a very wild stretch of the imagination, and again should not have been included in the data. The 22-spoked example from

Farley Heath (Plate 1, fig 3) is also dubious; as Green remarks it is only just a wheel, and may in fact be intended as a naturalistic representation of the sun, rather than a wheel symbol (Green 1984, 279). The sixteen-spoked example quoted in the original data turns out to be the eight-spoked half-wheel on the Gundestrup Cauldron (Figure

14; Plate 1, fig 4). The original data assumes that this must represent a complete sixteen-spoked wheel13 (Green 1984, 139); however, the artist may have intended to represent the perfectly formed, unbroken, eight-spoked half-wheel portrayed14. The example from Niederwürzbach (Plate 1, fig 6) should also have been excluded, since the spoke-number is uncertain. It may be a four-spoked example, but there is no way to be absolutely sure.

13 Actually it would represent a complete 15-spoked wheel.

14 The artist had no problem representing perspective in other areas, so it is not the result of a lack of technical ability. Similarly the central figure’s thumb is over the end of the felloe. It would therefore appear to be deliberate.

46 Figure 14: Detail of Gundestrup Cauldron (Allen 1980)

The corrected data is at Table 2. The exclusions do not have an enormous impact on the data; four-, six- and eight-spoked wheels make up 94.88% of the data. However,

Green’s sample covers the whole region west of Italy, and a period from the Iron Age to the Late Roman Empire. It is therefore highly probable that the data is distorted by the inclusion of both non-Celtic and Romanised symbolism. For example, the seven- spoked wheel from Oehringen (Plate 1, fig 7) may be from a representation of

Fortuna, and thus represent Classical rather than Celtic symbolism (Green 1984, 338-

339). In fact Green remarks that wheel-bearing images of Fortuna have been included simply because they are uncommon in the study area (Green 1984, 116). Similarly the five-spoked wheel from Petronell (Plate 2, fig 10) is from the grave of a soldier of

Legio XV Apollinaris (Green 1984, 345; Krüger 1972, 47-48, Fig. 542). To judge from his trinomina and his regiment he was, at the very least, highly Romanised

(Lendering 2003); Roman symbolism was almost certainly intended, despite the accompanying six- and eight-spoked wheel symbols (Plate 1, figs 8-9). Roman

47 symbolism may also be intended by the five-spoked example from Susa, near Turin

(Plate 2, fig 11; Green 1984, 342).

Unfortunately it is difficult to judge the effects of acculturation from this sample alone. A few examples are definitely Iron Age (Plate 1, fig. 5; Plate 2, figs. 12-14), while others come from the late Roman period. Others are completely undateable

(Green 1984, 76). It is therefore necessary to compare this sample with the known

Iron Age wheel symbolism of Celtic coins (Allen 1980, 149). With the exception of

Britain, these went out of circulation with the conquest of Gaul. Since this predates any marked Romanisation of Celtic religion, it should be possible to distinguish the effects of Romanisation.

The data for the following analysis were culled from several thousand coin reports.

Coins were excluded if the coin was too worn for the spoke numbers to be counted; this is likely to affect all coins equally. There has been only one assumption in compiling the data; namely that monnaie-á-la-croix represents wheel symbolism. This assumption is supported by several arguments:

Figure 15: Monnaie-á-la-croix

48 • The rim of a wheel is suggested by both the rim of the coin and by the frequent

presence of a circular design on such coins, plus explicit examples of wheel

symbolism (Figure 15).

• Coins associating wheel symbolism with

the horse are ubiquitous, and foreshadow

the later Jupiter-Giant columns (e.g.

Plate 5, figure 40). However some

Danubian coins substitute a cross for the

wheel (Figure 16). Conversely,

Danubian coins seem to have less wheel

symbolism. The cross is directly Figure 17: Moldavian cross-marked series, Dumbrăveni hoard type (Allen associated with the later Romano-Celtic 1987, Plate IX, figs. 145, 146)

Jupiter, and has been interpreted as having solar associations (Green 1991, 49). It

is therefore possible that regional variation in symbolism is involved, and that a

cross was, at least in some areas, a rimless wheel symbol. This variation in

symbolism would be consistent with multiple convergent traditions.

Figure 16: Monnaie-á-la-croix showing eye, axe, lunar and solar symbolism

49 • There are many examples of monnaie-á-la-croix using celestial symbolism such

as the eye or the axe (Figure 17). The eye is directly associated with celestial

imagery in Treviran coins, and with Mars Ocellus at Carlisle (Figure 18; Henig

1984, 73). Model axes are also associated with celestial deities; they are

frequently marked with crosses and are, on some coins, directly associated with

celestial imagery such as crescent moons (Henig 1984, 149; Green 1984, 99-

100).

Figure 18: Treviran coins showing eye symbolism (Allen 1980, Plate 18, figs. 246-249)

Monnaie-á-la-croix therefore uses the celestial symbolism associated with the celestial deity later dualled with Mars and Jupiter. The indigenous attribute of this deity is the wheel; it is therefore probably safe to assume that, since the shape suggests a wheel, wheel symbolism is intended.

The only other comment to make is that representations of chariots may imply realism. However, these are copies of Greek coins; the artist may have been aware of

Apollo’s chariot and regarded the imagery as appropriately solar. It should be noted that the use of religious imagery on coins was common in antiquity.

50 Approximately 90% of the continental coins show four-spoked wheel symbolism; the remainder are exclusively six or eight–spoked examples (Table 3). There are no other spoke numbers observed. This is not consistent with the hypothesis of a solar symbol, but it is consistent with a representation of the solar path (q.v. Figures 3 & 4). By contrast, British coins favour the six-spoked version, although the preferences vary from region to region (Tables 3-5). Importantly, deviations from the 4-6-8 pattern are exclusively British, and are primarily from the core contact zone, particularly the territory of the Atrebates; the incidence of deviations decline in the peripheral zone

(Tables 3-5; Plate 36).

The British pattern is undoubtedly the

result of pre-conquest Romanising

influences. The principal influence on

continental coins was, until quite late,

Greek; Britain, however, lay outside

Figure 19: Coin of Verica, with vine leaf design the region of direct Greek influence. (Hooker & Perron 1999-2002, VA 525-1) Conversely, the conquest of Gaul meant that the south-east of Britain was exposed to marked pre-conquest Romanising influences (Rodwell 1978; Stead 1967; Stead 1976;

Millett 1994, 31, 34; Cunliffe 1984, 15). The coins of the Atrebates, for example, are particularly well known for drawing on Roman motifs and symbolism (Figure 19;

Jersey 2003). The variations may therefore reflect the political proximity of the elites to Rome.

51 Figure 20: A & B: Coins of the Cantii, showing pentagram; C: coin of the Atrebates with 5- spoked wheel (Hooker & Perron 1999-2000, VA 163-1, VA 164-1, 660050)

Nevertheless these Romanised forms may still convey ideas about time and celestial movements. The five-spoked examples from the Atrebates may be related to the inclusion of the pentagram in coins of the neighbouring Cantii (Figure 20). The pentagram is a Middle Eastern symbol that has planetary symbolism, most probably related to Venus (Wikipedia 2003; Alabe Inc 2003); the pseudo-Venus is prominent in later Roman Kent (Jenkins 1958). The substitution of the pentagram for the wheel in wheel-and-horse imagery suggests equivalence with wheel symbolism. Seven-spoked examples may similarly refer to planetary symbolism, while ten-spoked wheels may refer to the Republican year of ten months. Spoke numbers of nine and higher are due entirely to three coins of the Atrebates (Figure 21; Plate 15, 169; Plate 17, 215, 216).

The symbols in question may not be intended to be wheel symbols at all, but naturalistic representations of the sun. The variability of spoke-numbers in the three symbols on each coin suggests that appearance was the dominant factor.

Figure 21: Coins of the Atrebates showing probable naturalistic sun symbols (Hooker & Perron 1999-2000, 680103, 680104)

52 Figure 22: Petroglyph, Stora Backa, Brastad, Bohuslän, Sweden (Gelling & Davidson 1969)

A brief survey of Bronze Age wheel petroglyphs shows that four-spoked wheels make up 90% of the sample, although six and eight-spoked wheels are known, as demonstrated by the petroglyphs at Bohuslän15 (Figure 22; Glob 1969, figs. 53-74).

The frequency of spoke numbers in continental Celtic coins therefore suggests possible continuity with Bronze Age ideas. A nine-spoked Iron Age wheel symbol from Fully is known, but this may be an error; the example from Bachos Binos shows an eight-spoked wheel that nearly became a nine-spoked version (Plate 2, figs. 14, 15;

Primas 1974, 98; Green 1984, 330).

15 In Figure 22 the wheels symbols have an approximately aligned axis.

53 In contrast, the frequencies seen in Green’s sample deviate markedly from those of the continental coins; the pattern is more akin to that of the pre-Conquest British coins. This is almost certainly due to Roman influence. Some examples may represent purely Roman symbolism; others may be factory wasters from mass production sites such as Allier, or the result of private experimentation with symbolism. Some may not be wheel symbols at all, but true wheel models deposited in an analogous manner to the model ballista washer from Bath (Cunliffe & Davenport 1985, 182). However the majority are almost certainly due to the influence of Graeco-Roman thought on

‘Celtic’ ideas. Twelve-spoked wheels, for example, are not seen in Iron Age symbolism; the second or third century AD examples from Housesteads, Felmingham

Hall and Wavendon Gate may therefore indicate the influence of the zodiac or even true solar imagery (Plate 2, figs. 16, 17; Green 1984, 169-170; Maynard 1992, 51).

Jupiter gained a solar interpretation in the later Roman period, and it should be expected that this would be reflected in his Romano-Celtic counterpart. It should not therefore be assumed that wheel symbols of the Late Roman Empire are identical with the symbolism of the Iron Age, nor with that of the immediate post-Roman period.

Nevertheless, several potential objections could be raised to the observed statistical pattern. The pattern itself may be an artefact created by the limitations of the material used by the craftsman, and thus have no significance. It could also be argued that it is easier to make wheel symbols with an even number of spokes. These objections are easily dismissed. Speaking as a former professional weaponsmith, it is possible, using simple metal-working techniques, to make metal wheel symbols with more than twenty spokes. Similarly it is, from personal experience, not particularly difficult to do a simple casting with odd numbers of spokes, even with the technology available

54 to Bronze Age craftsmen, while the presence of very small eight-spoked wheels on

Celtic coins suggests that the size of the depiction was not a limiting factor for stamped work.

This is borne out by the archaeological evidence. Realistic secular representations of chariot wheels are found, for example, in sheet metal work (Frey & Lucke 1962; Frey

1968; Metzler 1986). The examples studied have variously five, six or seven spokes, which immediately discounts the argument that it was difficult to portray odd numbers of spokes. Similarly a relief at Buzenol shows a Gaulish reaping machine with ten-spoked wheels, while a depiction at Dijon shows an eight-spoked wagon wheel (Green 1984, 272). Roman masons similarly had no problems representing twelve-spoked wheels, as Trajan’s column demonstrates (Green 1984, 273). It should be noted however that eight-spoked wheels appear to be the norm in purely Roman secular and religious contexts (Green 1984, 273). A coin from Praeneste, for example, shows Fortuna wearing an eight–spoked wheel earring (Champeux 1982, Plate VI, 4).

This therefore raises the question as to how closely wheel symbols conform to real wheels. The Llyn Cerrig Bach chariots had between seven and twelve spokes, while the two wheels from Glastonbury Lake Village had ten and twelve spokes respectively (Vouga 1923; Fox 1946, 12; Green 1984, 269; Egg & Pare 1993, 217).

The wheels from the Garston Slack cart burials also had twelve spokes (Stead 1989,

2). Both the Newstead wheel and a La Tène wagon wheel from Bell in Hunsrück were ten-spoked, while the Dejbjerg cult wagons have twelve and fourteen spokes (Fox

1946, 12; Megaw 1970, 129; Green 1984, 269). As a general rule, real Hallstatt and

La Tène wheels have high spoke numbers (Jope 1956, 551; Green 1984, 269).

55 The pattern seen in realistic, secular Romano-Celtic representations of wheels therefore concurs, within the limits of the material used and the size of the image, with known examples of real wheels. A similar realism is seen in depictions from purely Roman contexts, where eight spokes are favoured. There is therefore a probable relationship between Roman representations and Romano-Celtic secular representations.

By contrast, the statistical pattern seen in religious wheel symbolism runs counter to

Romano-Celtic secular representations of wheels, to representations from purely

Roman contexts and, in the pre-Roman period, to examples of real Hallstatt and La

Tène wheels. Indeed the oldest and commonest form of Celtic wheel symbol has four spokes, and these are only found in religious contexts (Green 1984, 276). While four- spoked cast bronze wheels are known from Urnfield hoards, it is questionable whether such expensive items were ever intended to be practical (Green 1984, 266). Practical four-spoked wooden wheels are known from very early Aegean examples, but they required characteristic bronze spoke terminals, which are not reflected in four-spoked wheel symbolism, even in the Hallstatt period (Pare 1987). It would in any case be rather problematic to argue that Urnfield and early Aegean wheel design influenced wheel symbolism in the Late Iron Age and Romano-Celtic period.

One is led to conclude that Celtic and Romano-Celtic artists used one set of rules for portraying realistic representations of wheels, and a quite different set of rules for creating wheel symbols for religious contexts. Since there are no visible, practical sources for the observed statistical pattern, and since wheel symbolism was of a religious nature, the inspiration for the statistical pattern may lie in religious ideas,

56 and in perceptions of what constituted an appropriate, meaningful symbolism. In short, the number of spokes was significant for both artisans and their customers, and represented part of a common symbolic language. This might be unsurprising since there does seem to have been a pattern of number significance in Celtic society

(Boucher 1976, 212).

Figure 23: True solar symbolism in Celtic coins (Hooker & Perron 1999-2002, VA 226-1, VA 286-1, VA 290-1)

These religious ideas do not however involve sun symbolism. The difference between real spoked wheels and wheel symbolism suggests that the former had no part in the origins of the latter. The whole notion that it developed from likening the sun to a wheel is therefore in error, although the symbolism itself was probably assimilated to the idea of a wheel at a later date. Further, indigenous wheel symbolism does not conform to the pattern expected for sun symbolism, which suggests that the wheel is not even a sun symbol. The expected pattern is however seen in the true solar symbolism found in Celtic coins; none have less than nine spokes, although there is a

57 slight preference for eleven spokes (Figure 23; Plate 13, 149; Plate 16, 195; Plate 17,

213; Plate 18, 240; Plate 20, 276, 289; Plate 32, 530). In short, Celtic wheel symbolism and true Celtic solar symbolism are apparently mutually exclusive.

It might however be objected that the number four had solar connections, and was therefore appropriate to sun symbolism. This however moves the argument away from that of a true solar symbol, as well as begging the question of why the numbers six and eight should also have significance. Any explanation should encompass the variation seen in indigenous wheel symbolism.

58 REINTERPRETING WHEEL SYMBOLISM

The Coligny calendar and the Jupiter-Giant columns appear to share equivalent cognitive features that are not found collectively in non-Celtic contexts. They differ only in that one directly structures time, while the other represents time in abstract cosmological terms. In both instances the evidence suggests that the wheel represents the solar path rather than the sun. The choice of spoke-number in wheel symbolism appears to be determined by solely by religious ideas; it does not however conform to the pattern expected of a true sun symbol, nor does it show any connection to real wheels until the Roman period. It is therefore worth considering the generally accepted hypothesis that the wheel is a solar symbol.

The Classical Mars and Jupiter were never solar. Furthermore, where an epigraphic dedication accompanies wheel symbolism, it is to the Roman sky god – not to a sun god (Green 1986b, 68). However the wheel-god, who is dualled with Jupiter and

Mars, is frequently described as a sun god, a description for which Müller’s discredited nineteenth century Germanic ‘solar school’ of comparative mythology is largely responsible (Ross 1967, 7; Olmsted 1979, 129). This interpretation rests entirely on the wheel being the god’s attribute, since the wheel has been identified as a solar symbol for over one hundred years (Gaidoz 1884, 14; Frazer 1922, 843). The nave of the wheel is the sun, the spokes are the sun’s rays, and the rim is the outer nimbus. It is further argued that the sun rolls in a wheel-like manner through the heavens.

59 This is a very subjective analysis of a presumed subjective perception. Further, the sun does not appear in any way appear to rotate as it travels through the sky.

However, in some cases no further supporting argument is given for this reasoning, beyond the subjective observation that it is ‘obvious’ (Gaidoz 1884, 14). Other authors have accepted the identification unquestioningly, as if it had been proven beyond doubt (Altheim 1938, 76; Allen 1980, 149; Green 1984, 156; Mynard 1992,

51).

There are, however, some supporting arguments for this hypothesis. It has been remarked that Greek poets likened the sun to a wheel (Cook 1925, 57-93). It should not be necessary to remark that the poetic similes of Greek poets have no bearing on the symbolism of a completely different culture from a completely different area, particularly when the earliest central European form of this symbol predates them by approximately nine hundred years. To insist on this point is, to paraphrase Tacitus, interpretatio Graeca. One may as well cite the interpretations of Polynesians.

The solar hypothesis has been supported by arguing that the sun is male in almost all cultures (Green 1991, 22). This is a circular argument; the identification of the Celtic wheel-god as a sun god rests only upon the hypothesis that the wheel represents the sun. There are in fact some notable European exceptions to the generalisation of the male sun. The sun is female in the Poetic Eddas, and the word for ‘sun’ takes the feminine case in Gaelic at all stages of its development16 (Poetic Eddas, Voluspo 5).

Antinativist arguments apart, several deities in the Irish sources can be firmly

16 It is meaningless in Gaelic to represent a feminine object with a masculine deity; Gaelic is as gender- orientated as French.

60 identified with the sun; they are all female, and the identification is not in doubt (Met.

Dinds. IV, Ard Macha, 127, lines 45-48; O’Rahilly 1946, 293; Mac Cana 1955).

Macha, for example, has the epithet grian (fem. noun), or ‘sun’ and is also associated with horses. There is also the Romano-British goddess Sulis (RIB 149); the element sul is cognate with the Irish feminine noun, súil, which meant both ‘sun’ and ‘eye’17

(Rivet & Smith 1979; Green 1991, 19-20; 1995, 96). Interestingly later traditional

Gaelic poetry described the feminine queenly sun as the eye of the great (masculine) god18 (Carmichael 1900-1941, III, 306-307, 310-311). One might note in this context the association of an eye with celestial imagery, including wheels, in Treviran coins, the reverse of which show a horse19 (Figure 18; Green 1991, 19). It should also be recalled that although the modern Welsh word for ‘sun’ is masculine, its ancestor,

Brythonic, was spoken in an area where Latin, Roman divinities, imperial associations with the sun, and Roman power were every day realities. Changes in perceptions of deity may well be reflected by changes in language, particularly since the Latin, masculine word for sun, sol, was close to the Brythonic sul. Of course, this again implies that Romanisation wrought profound changes on indigenous religion.

The occasional explicit identification of the wheel-god with Apollo has also been taken as proof that the wheel-god is solar. Apollo Moritasgus at Mount Auxois,

Alesia, for example, physically resembles Jupiter (Green 1984, 203). However the

Classical Apollo was not identified with the sun itself; he was instead a lawgiver who was the master of the sun, and who directed the fate of the world (Altheim 1938, 401;

17 In modern Irish Gaelic it has become limited to ‘eye’ (Mac Mathúna & Ó Corráin 1997, 602). A century ago it meant both.

18 C.f. Epona’s epithet, Regina Sancta.

19 Jupiter-Giant columns were abundant amongst the Treviri (Green 1984, 204).

61 Gelling & Davidson 1969, 121). This may have relevance for the iconography of the

Jupiter-Giant columns; the wheel-god rides a horse, an animal that has been identified with the sun, and which probably represents the theriomorphic Epona. As any horse- rider will know, the rider of a horse is not the horse, but its master; at Mauvieres

(Indre) Apollo Grannus is referred to as Atepomarus (Great Horseman) (Green 1984,

194, 203). Apollo additionally presided over crops and herbs, and protected flocks

(Laing 1982, 23). An indigenous deity who was a lawgiver, who directed Fate, who was master of the sun, and who was additionally connected with the agricultural cycle could readily be equated with Jupiter, Mars or Apollo. It does not however mean that the deity was a sun god.

The argument that the wheel is a sun

symbol is thus very flawed, although it

should be admitted that perhaps the right

argument has not yet been framed.

However, it should be noted that an

identical cross-and-circle symbol arose in

Mesoamerica (Figure 24; Aveni et al

1978; Iwaniszewski 1993, 290-291). It is

impossible to identify this symbol as a

‘sunwheel’, since the wheel was not used

Figure 24: Mesoamerican cross-circle (Aveni in pre-Columbian America. The 1988, 447) Mesoamerican design apparently relates to points in the tropical and agricultural year. Similar alignments have been observed

62 for Danish wheel petroglyphs (Christensen N I & Christensen E I 2001); this would be anticipated by this interpretation. Further, the earliest Danish wheel petroglyphs are

Late Neolithic and Early Bronze Age in date, and thus predate the earliest known spoked wheel north of the Alps by 500-1000 years (Green 1984, 30; EuroPreart

2003); in fact the symbol appears as early as the Danish Middle Neolithic (EuroPreart

2003). This supports the earlier deduction that spoked wheels were not the inspiration for wheel symbolism (page 57). An alternative explanation of Celtic wheel symbolism is therefore called for.

It has already been seen that it was

necessary to use a pillar to establish the

framework of the agricultural and religious

year. From a geocentric point of view, the

heavens move in a circular, wheel-like

manner around a static point, the celestial

pole (Figure 25). This was a familiar

Figure 25: Timelapse photograph of stars rotating around the celestial pole (Krupp observation in antiquity, although it is 1983, 104) readily obvious to anyone who watches the night sky for long enough (Aristotle, On the Universe 391b16-392a4; De Caelo II.4

287a11-22; Ovid, Fasti I.119-120; IV. 179-180; Virgil, Georgics I, 242-251; Lucan,

Bell. Civ. I.552-3, VI.464-465; Vitruvius, De Arch. IX, i.2-3). The wheel-god is also associated with night and day, both of which revolve, equally obviously, around the celestial axis (Aristotle, On the Universe, 6.399a, 1-5; Green 1981, 112). 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.

63 A line from the celestial pole to a local observer can be considered as a local celestial axis. The disposition of the solar rising and setting points for the festivals and the solstices are identical, whether one is considering a local celestial axis or a gnomon. A gnomon and the celestial axis are therefore functionally equivalent, and the former is probably a representation of the latter. Similarly the Jupiter-Giant columns 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). Lucan’s remark that the Pole Star looks down upon the Celts is therefore particularly pertinent, and may indicate a characteristic Celtic belief (Lucan Bell. Civ., I.458-460). It should also be noted that the corollary of a local celestial axis is a central place, such as that mentioned in the land of the Carnutes (BG VI.13).

In geocentric astronomy the rising and setting points of the solstices define the limits of the sun’s annual movement; this astronomical fact was well understood in antiquity

(Posidonius 137a). These points form a diagonal cross if considered in relation to the celestial axis. Since this cross, which is associated with the Celtic Jupiter, represents the limits of the sun’s annual travel, it could be considered as representing the divine law that established those solar limits, thus framing the agricultural year and establishing agricultural fertility (Green 1978, 20; 1989, 84-85; 1991, 49; Lynn 1992,

48). The Danubian substitution of the analemma therefore makes sense. It also represents the law that governs the sun’s movement (Ballew 2003; Urschel 2003); it is not a representation of the sun itself. All that is required for this substitution of symbols is a change in methodology, not technology.

64 The complete four-spoked wheel symbol would therefore be a symbol of the law that, by establishing the turning of the heavens and the limits of the sun, established the agricultural year. The six-spoked examples would be obtained if north-south line, the line from the pole, is also considered, although the symbolism is identical. Attempts to represent both the festivals and the solstices, would produce an eight-spoked wheel

(Hicks 1988, 1988, 471; Lynn 1992, 43). Interestingly, the eight-spoked half-wheel on the Gundestrup cauldron may suggest the division of the sky into day and night

(Figure 14); solar observations are applicable only to daytime, which is one half of the daily rotation around the celestial axis.

This interpretation ultimately derives from the use of a pillar to

structure time for agricultural purposes. The Etzelsdorf cone

and related artefacts are probable sacred representations of such

pillars. Further, it is generally accepted that the Jupiter-Giant

columns represent trees, and that there is therefore an

iconographical equivalence between the two. Deities associated

with trees, such as Mars Olludios or Mars Rigonemetis could

therefore be represented by an aniconic pillar of the type used

to structure time20 (Henig 1984, 73). This would further

account for the association of calendars and the Romano-Celtic

Mars. Furthermore, the rider god sometimes carries a spear, an

Figure 26: Pole-tip attribute of Mars; a pole tip from a shrine at Brigstock from Brigstock (Green 1976, XXV, b) apparently fuses spear and phallus (Figure 26; Green 1986a,

20 Which makes it even less likely that the cones were priests’ hats; people are less likely to wear gods on their heads.

65 98). The Romano-Celtic Mars may therefore have represented the generative aspect of the sky god that mediated between the celestial and chthonic (Eliade 1957, 35; Green

1976, 51; Green M J A 1996 120-121). There are significant indications of confusion between Mars and Mercury, the Classical messenger of the gods (Green 1976, 29-30;

1986, 37). It is therefore significant that the imagery of time is found at the base of the

Jupiter-Giant columns, where the chthonic and celestial could be considered to fuse, thus generating time and agriculture.

This brings us to the Le Châtelet wheel-god

(Figure 27). According to this interpretation of

wheel symbolism he holds, in one hand, a

wheel that symbolises the law that regulates the

heavens and the agricultural year. In the other

he holds a model gnomon, symbolising the

celestial axis that transmits that law to earth,

and by which the agricultural year is

established. Over his shoulder he carries a ring

containing S-shaped objects, which may

represent stylised snakes, symbols of the

Figure 27: Le Châtelet wheel-god chthonic forces subjected to the celestial law (Green 1984) that orders the movements of the heavens.

If this interpretation is correct, and wheel symbolism is a representation of divine law rather than a solar symbol, one would expect it to be used apotropaically. Unexplained illness, sudden death and plagues were, in the Roman world, consequences of

66 witchcraft. Witchcraft was opposed to divine law (Bell. Civ. VI.510-513, VI.529-532;

Dio 52.36.3; Gordon 1990, 253-255; Luck 1999, 137). A symbol of divine law, and the means of transmitting it, would consequently avert such forces and maintain health. Direct and indirect evidence suggests that Romano-Celtic wheel symbols were used in such a manner, while the Romano-Celtic Mars, unlike his Classical counterpart, is associated with healing (Green 1976, 29; 1984, 64, 80, 85; Webster

1986a, 58). Wheel symbolism is also seen in Iron Age military equipment, and it is probable that all Iron Age models were also amulets (Green 1984, 35, 36-37, 42). The observed use is therefore consistent with the hypothesis. Further, such symbolism could be later assimilated to Roman deities such Fortuna who, since they symbolised

‘the will of the gods’, could be considered in similar terms, as at Agey (Green 1984,

116-117).

67 CONCLUSIONS

The evidence therefore suggests that wheel symbolism represents the solar path, and that the symbolism was derived from a post-and-marker method of structuring time for agricultural and religious purposes. The structure of both the Coligny Calendar and the oenach is consistent with such technology. Furthermore, as expected, some

European examples of wheel symbolism have solar orientations. The technology however predates the emergence of the Celts as a recognisable group (c.f. Green

1984, 302), and was probably both independently discovered at different times by different groups and communicated by group interactions. Later acculturative processes, such as trade, led some of these groups to develop a koiné later described as ‘Celtic’. This koiné included the religious use of a wheel symbol as a representation of divine law.

Nevertheless, if the wheel represents the solar path then it is not a sun symbol, nor is the wheel god a sun god, although he may have gained solar attributes in the Roman period. On the contrary, the solar deity was probably female, although possessing chthonic/aquatic/fertility aspects. However, Romanisation resulted in the goddess losing her solar attributes, changing the originally solar deity to the equivalent of a

Matre. This suggests that Celtic ‘Earth Mother’ deities may be as much a product of

Romanisation as Celtic ‘sun gods’. Overall, there therefore appears to be a need to extensively re-evaluate Celtic and Romano-Celtic iconography, although space forbids expansion on this theme.

Kevin Jones

68

CONCORDANCE TABLE & DATA Figure Original reference Spokes Figure Original reference Spokes 1 Allen 1980, 102 4 56 Allen 1987, Vol 2, 53 4 2 Allen 1980, 103 4 57 Allen 1987, Vol 2, 54 4 3 Allen 1980, 4 4 58 Allen 1987, Vol 2, 55 4 4 Allen 1980, 168 4 59 Allen 1987, Vol 2, 56 4 5 Allen 1980, 48 4 60 Allen 1987, Vol 2, 57 4 6 Allen 1980, 91 4 61 Allen 1987, Vol 2, 60 4 7 Allen 1980, 92 4 62 Allen 1987, Vol 2, 63 4 8 Allen 1980, 93 4 63 Allen 1987, Vol 2, 66 4 9 Allen 1980, 94 4 64 Allen 1987, Vol 2, 67 4 10 Allen 1980, 95 4 65 Allen 1987, Vol 2, 68 4 11 Allen 1980, 96 4 66 Allen 1987, Vol 2, 69 4 12 Allen 1980, 97 4 67 Allen 1987, Vol 2, 71 4 13 Allen 1980, 98 4 68 Allen 1987, Vol 2, 79 4 14 Allen 1980, 99 4 69 Allen 1987, Vol 2, 81 4 15 Allen 1980, 100 4 70 Allen 1987, Vol 2, 82 4 16 Allen 1980, 101 4 71 Allen 1987, Vol 2, 107 4 17 Allen 1980, 169 4 72 Allen 1987, Vol 2, 108 4 18 Allen 1980, 178 4 73 Allen 1987, Vol 2, 140 2x6 19 Allen 1980, 216 6 74 Allen 1987, Vol 1, 158 4 20 Allen 1980, 223 4 75 Allen 1987, Vol 1, 185 4 21 Allen 1980, 241 4 76 Allen 1987, Vol 2, 247 4 22 Allen 1980, 242 4 77 Allen 1987, Vol 2, 248 4 23 Allen 1980, 436 4 78 Allen 1987, Vol 2, 319 4 24 Allen 1980, 281 4 79 Allen 1987, Vol 2, 320 4 25 Allen 1980, 283 4 80 Allen 1987, Vol 2, 321 4 26 Allen 1980, 293 4 81 Allen 1987, Vol 2, 322 4 27 Allen 1980, 379 6 82 Allen 1987, Vol 2, 323 4 28 Allen 1980, 429 4 83 Allen 1987, Vol 2, 324 4 29 Allen 1980, 249 8 84 Allen 1987, Vol 2, 326 4 30 Allen 1980, 246 - 85 Allen 1987, Vol 2, 328 4 31 Allen 1980, 247 - 86 Allen 1987, Vol 2, 329 4 32 Allen 1980, 248 - 87 Allen 1987, Vol 2, 330 4 33 Allen 1980, 453 4 88 Allen 1987, Vol 2, 331 4 34 Allen 1980, 454 4 89 Allen 1987, Vol 2, 335 4 35 Allen 1980, 462 6 90 Allen 1987, Vol 2, 594 4 36 Allen 1980, 467 4 91 Allen 1987, Vol 2, 595 4 37 Allen 1980, 472 4 92 Allen 1987, Vol 2, 596 4 38 Allen 1980, 490 8 93 Allen 1987, Vol 1, S5 4 39 Allen 1980, 510 8 94 Allen 1987, Vol 2, S97 4 40 Allen 1980, 575 6 95 Allen 1987, Vol 2, S98 4 41 Allen 1980, 63 6 96 Allen 1987, Vol 2, S103 4 42 Allen 1980, 57 6 97 Allen 1987, Vol 2, S105 4 43 Allen 1980, 116 2x4 98 Allen 1987, Vol 2, S108 4 44 Allen 1980, 67 6 99 Allen 1987, Vol 2, S109 4 45 Allen 1987, Vol 2, 39 4 100 Allen 1987, Vol 2, S111 4 46 Allen 1987, Vol 2, 41 4 101 Allen 1987, Vol 2, S112 4 47 Allen 1987, Vol 2, 42 4 102 Allen 1987, Vol 2, S114 4 48 Allen 1987, Vol 2, 43 4 103 Allen 1987, Vol 2, S115 4 49 Allen 1987, Vol 2, 44 4 104 Allen 1987, Vol 2, S116 4 50 Allen 1987, Vol 2, 45 4 105 Allen 1987, Vol 2, S118 4 51 Allen 1987, Vol 2, 46 4 106 Allen 1987, Vol 2, S119 4 52 Allen 1987, Vol 2, 48 4 107 Allen 1987, Vol 2, S120 4 53 Allen 1987, Vol 2, 49 4 108 Allen 1987, Vol 2, S121 4 54 Allen 1987, Vol 2, 50 4 109 Allen 1987, Vol 2, S122 4 55 Allen 1987, Vol 2, 51 4 110 Allen 1987, Vol 2, S123 4

69

CONCORDANCE TABLE & DATA (contd.) Figure Original reference Spokes Figure Original reference Spokes 111 Allen 1987, Vol 2, S124 4 169 Hooker & Perron 2002, 001495 10 112 Allen 1987, Vol 2, S125 4 170 Hooker & Perron 2002, 001496 8 113 Allen 1987, Vol 2, S126 4 171 Hooker & Perron 2002, 001501 6 114 Allen 1987, Vol 2, S127 4 172 Hooker & Perron 2002, 001858 8 115 Allen 1987, Vol 2, S128 4 173 Hooker & Perron 2002, 010104 6 116 Allen 1987, Vol 2, S129 4 174 Hooker & Perron 2002, 010333 6 117 Allen 1987, Vol 2, S130 4 175 Hooker & Perron 2002, 010549 8 118 Allen 1987, Vol 2, S131 4 176 Hooker & Perron 2002, 010554 8 119 Allen 1987, Vol 2, S132 4 177 Hooker & Perron 2002, 610133 8 120 Allen 1987, Vol 2, S133 4 178 Hooker & Perron 2002, 610137 8 121 Allen 1987, Vol 2, S134 4 179 Hooker & Perron 2002, 610139 8 122 Allen 1987, Vol 2, S135 4 180 Hooker & Perron 2002, 610140 8 123 Allen 1987, Vol 2, S136 4 181 Hooker & Perron 2002, 620006 8 124 Allen 1987, Vol 2, S137 4 182 Hooker & Perron 2002, 630178 6 125 Allen 1987, Vol 2, S138 4 183 Hooker & Perron 2002, 630180 8 126 Allen 1987, Vol 2, S139 4 184 Hooker & Perron 2002, 630184 6 127 Allen 1987, Vol 2, S141 4 185 Hooker & Perron 2002, 630189 5 128 Allen 1987, Vol 2, S286 4 186 Hooker & Perron 2002, 640026 8 129 Allen 1987, Vol 2, S287 4 187 Hooker & Perron 2002, 640027 6 130 Allen 1987, Vol 2, S288 4 188 Hooker & Perron 2002, 640028 8 131 Allen 1987, Vol 2, S289 4 189 Hooker & Perron 2002, 640049 8 132 Allen 1987, Vol 2, S294 4 190 Hooker & Perron 2002, 660050 5 133 Allen 1987, Vol 2, S297 4 191 Hooker & Perron 2002, 660054 8 134 Allen 1987, Vol 2, S457 4 192 Hooker & Perron 2002, 660057 8 135 Allen 1987, Vol 2, 325 4 193 Hooker & Perron 2002, 670045 8 136 Allen 1987, Vol 1, S168 8 194 Hooker & Perron 2002, 670046 6 137 Allen 1987, Vol 1, 20 4 195 Hooker & Perron 2002, 670063 6 138 Allen 1987, Vol 1, 24 4 196 Hooker & Perron 2002, 680050 8 139 Allen 1978, 24 4 197 Hooker & Perron 2002, 680051 8 140 Allen 1978, 25 4 198 Hooker & Perron 2002, 680052 8 141 Allen 1978, 26 4 199 Hooker & Perron 2002, 680054 6 142 Allen 1978, 50 4 200 Hooker & Perron 2002, 680055 6 143 Allen 1978, 95 3X4 201 Hooker & Perron 2002, 680058 6 144 Allen 1978, 98 4 202 Hooker & Perron 2002, 680061 8 145 Allen 1978, 57 8 203 Hooker & Perron 2002, 680064 6 146 Allen 1978, 71 6 204 Hooker & Perron 2002, 680068 8 147 Allen 1978, 86 8 205 Hooker & Perron 2002, 680069 6 148 Allen 1978, 88 8 206 Hooker & Perron 2002, 680070 6 149 Allen 1978, 89 6 207 Hooker & Perron 2002, 680073 8 150 Allen 1978, 92 8 208 Hooker & Perron 2002, 680074 6 151 Allen 1978, 102 5 209 Hooker & Perron 2002, 680075 6 152 Castelin 1985, 11 8 210 Hooker & Perron 2002, 680076 6 153 Castelin 1985, 12 8 211 Hooker & Perron 2002, 680089 6+6 154 Castelin 1985, 14 4+6 212 Hooker & Perron 2002, 680090 8 155 Castelin 1985, 16 4 213 Hooker & Perron 2002, 680093 8 156 Castelin 1985, 20 4 214 Hooker & Perron 2002, 680099 6 157 Castelin 1985, 21 4 215 Hooker & Perron 2002, 680103 11+11+9 158 Castelin 1985, 76 8 216 Hooker & Perron 2002, 680104 10+11+12 159 Castelin 1985, 90 4 217 Hooker & Perron 2002, 690142 6 160 Castelin 1985, 93 4 218 Hooker & Perron 2002, 690143 5 161 Castelin 1985, 104 4 219 Hooker & Perron 2002, 690144 5 162 Castelin 1985, 105 4 220 Hooker & Perron 2002, 690145 6 163 Castelin 1985, 109 4 221 Hooker & Perron 2002, 690166 4 164 Castelin 1985, 139 8 222 Hooker & Perron 2002, 690320 6 165 Hooker & Perron 2002, 000298 6 223 Hooker & Perron 2002, 720029 8 166 Hooker & Perron 2002, 000299 8 224 Hooker & Perron 2002, 720033 8 167 Hooker & Perron 2002, 000677 6 225 Hooker & Perron 2002, 720035 8 168 Hooker & Perron 2002, 000849 4 226 Hooker & Perron 2002, 720039 6+8+6

70 CONCORDANCE TABLE & DATA (contd.) Figure Original reference Spokes Figure Original reference Spokes 227 Hooker & Perron 2002, 720041 6 279 Hooker & Perron 2002, 954022 6 228 Hooker & Perron 2002, 720043 8 280 Hooker & Perron 2002, 960098 4 229 Hooker & Perron 2002, 730133 6 281 Hooker & Perron 2002, 960104 4 230 Hooker & Perron 2002, 730136 6 282 Hooker & Perron 2002, 960114 6 231 Hooker & Perron 2002, 740028 6 283 Hooker & Perron 2002, 960115 6 232 Hooker & Perron 2002, 740030 8 284 Hooker & Perron 2002, 960116 6 233 Hooker & Perron 2002, 740034 8+6 285 Hooker & Perron 2002, 960117 6 234 Hooker & Perron 2002, 820155 6 286 Hooker & Perron 2002, 960980 8 235 Hooker & Perron 2002, 820161 8 287 Hooker & Perron 2002, 961178 8 236 Hooker & Perron 2002, 840802 5 288 Hooker & Perron 2002, 961180 6 237 Hooker & Perron 2002, 850093 6 289 Hooker & Perron 2002, 961184 6 238 Hooker & Perron 2002, 860140 4 290 Hooker & Perron 2002, 961363 4 239 Hooker & Perron 2002, 860150 6 291 Hooker & Perron 2002, 961479 8 240 Hooker & Perron 2002, 860155 6 292 Hooker & Perron 2002, 961481 8 241 Hooker & Perron 2002, 880113 4 293 Hooker & Perron 2002, 961789 6 242 Hooker & Perron 2002, 900090 4 294 Hooker & Perron 2002, 961790 6 243 Hooker & Perron 2002, 900116 6 295 Hooker & Perron 2002, 962424 8 244 Hooker & Perron 2002, 910346 6 296 Hooker & Perron 2002, 962426 8 245 Hooker & Perron 2002, 920530 8 297 Hooker & Perron 2002, 962427 8 246 Hooker & Perron 2002, 920666 6 298 Hooker & Perron 2002, 962443 6 247 Hooker & Perron 2002, 930181 6 299 Hooker & Perron 2002, 962464 8 248 Hooker & Perron 2002, 930249 8 300 Hooker & Perron 2002, 962466 8 249 Hooker & Perron 2002, 930269 6 301 Hooker & Perron 2002, 962478 8 250 Hooker & Perron 2002, 930648 6 302 Hooker & Perron 2002, 962556 8 251 Hooker & Perron 2002, 940034 4 303 Hooker & Perron 2002, 962656 4 252 Hooker & Perron 2002, 940588 4 304 Hooker & Perron 2002, 962698 5 253 Hooker & Perron 2002, 940639 4 305 Hooker & Perron 2002, 962785 6 254 Hooker & Perron 2002, 940815 8 306 Hooker & Perron 2002, 962791 8 255 Hooker & Perron 2002, 940848 6 307 Hooker & Perron 2002, 962880 7 256 Hooker & Perron 2002, 940937 6 308 Hooker & Perron 2002, 962881 4 257 Hooker & Perron 2002, 940961 6 309 Hooker & Perron 2002, 963008 8 258 Hooker & Perron 2002, 941010 4 310 Hooker & Perron 2002, 963009 8 259 Hooker & Perron 2002, 950342 6 311 Hooker & Perron 2002, 963014 6 260 Hooker & Perron 2002, 950346 4 312 Hooker & Perron 2002, 963214 4 261 Hooker & Perron 2002, 950541 6 313 Hooker & Perron 2002, 963324 6 262 Hooker & Perron 2002, 950624 4 314 Hooker & Perron 2002, 963389 8 263 Hooker & Perron 2002, 950719 8 315 Hooker & Perron 2002, 970016 8 264 Hooker & Perron 2002, 950735 4 316 Hooker & Perron 2002, 970017 8 265 Hooker & Perron 2002, 950805 6 317 Hooker & Perron 2002, 970026 6 266 Hooker & Perron 2002, 950827 8 318 Hooker & Perron 2002, 970796 8 267 Hooker & Perron 2002, 950828 8 319 Hooker & Perron 2002, 970798 8 268 Hooker & Perron 2002, 951019 8 320 Hooker & Perron 2002, 970799 6 269 Hooker & Perron 2002, 951260 6 321 Hooker & Perron 2002, 970803 6 270 Hooker & Perron 2002, 952905 8 322 Hooker & Perron 2002, 970996 6 271 Hooker & Perron 2002, 953069 6 323 Hooker & Perron 2002, 970997 6 272 Hooker & Perron 2002, 953071 8 324 Hooker & Perron 2002, 971033 5 273 Hooker & Perron 2002, 953470 7 325 Hooker & Perron 2002, 971158 5 274 Hooker & Perron 2002, 953474 6 326 Hooker & Perron 2002, 971206 4 275 Hooker & Perron 2002, 953729 6 327 Hooker & Perron 2002, 971468 4 276 Hooker & Perron 2002, 953733 8 328 Hooker & Perron 2002, 971711 8 277 Hooker & Perron 2002, 954020 8 329 Hooker & Perron 2002, 980193 6 278 Hooker & Perron 2002, 954021 8 330 Hooker & Perron 2002, 980242 8

71 CONCORDANCE TABLE & DATA (contd.) Figure Original reference Spokes Figure Original reference Spokes 331 Hooker & Perron 2002, 980441 6 383 Hooker & Perron 2002, 951173 4 332 Hooker & Perron 2002, 980442 6 384 Hooker & Perron 2002, 951387 4 333 Hooker & Perron 2002, 980444 6 385 Hooker & Perron 2002, 961474 4 334 Hooker & Perron 2002, 980447 6 386 Hooker & Perron 2002, 963103 8 335 Hooker & Perron 2002, 980451 6 387 Hooker & Perron 2002, 991060 6 336 Hooker & Perron 2002, 980453 6 388 Hooker & Perron 2002, 991704 6 337 Hooker & Perron 2002, 980455 8 389 Hooker & Perron 2002, 991061 6 338 Hooker & Perron 2002, 991334 8 390 Hooker & Perron 2002, 000385 6 339 Hooker & Perron 2002, 980462 6 391 Hooker & Perron 2002, 000433 6 340 Hooker & Perron 2002, 980462 8 392 Hooker & Perron 2002, 610038 6 341 Hooker & Perron 2002, 980469 8 393 Hooker & Perron 2002, 610039 6 342 Hooker & Perron 2002, 980583 8 394 Hooker & Perron 2002, 610040 6 343 Hooker & Perron 2002, 980586 6 395 Hooker & Perron 2002, 610042 6 344 Hooker & Perron 2002, 980593 6 396 Hooker & Perron 2002, 610055 6 345 Hooker & Perron 2002, 980748 5 397 Hooker & Perron 2002, 610057 6 346 Hooker & Perron 2002, 980749 5 398 Hooker & Perron 2002, 610059 6 347 Hooker & Perron 2002, 980751 5 399 Hooker & Perron 2002, 640096 6 348 Hooker & Perron 2002, 980752 5 400 Hooker & Perron 2002, 650008 6 349 Hooker & Perron 2002, 980753 5 401 Hooker & Perron 2002, 650009 6 350 Hooker & Perron 2002, 980756 5 402 Hooker & Perron 2002, 660015 6 351 Hooker & Perron 2002, 981018 4 403 Hooker & Perron 2002, 660016 6 352 Hooker & Perron 2002, 981033 7 404 Hooker & Perron 2002, 660017 6 353 Hooker & Perron 2002, 981306 6 405 Hooker & Perron 2002, 660022 6 354 Hooker & Perron 2002, 981307 6 406 Hooker & Perron 2002, 660024 6 355 Hooker & Perron 2002, 981308 7 407 Hooker & Perron 2002, 660026 6 356 Hooker & Perron 2002, 981424 6 408 Hooker & Perron 2002, 660027 8 357 Hooker & Perron 2002, 981811 7 409 Hooker & Perron 2002, 660266 6 358 Hooker & Perron 2002, 981892 8 410 Hooker & Perron 2002, 670013 6 359 Hooker & Perron 2002, 981961 5 411 Hooker & Perron 2002, 670019 6 360 Hooker & Perron 2002, 981979 4 412 Hooker & Perron 2002, 670020 5 361 Hooker & Perron 2002, 982072 5 413 Hooker & Perron 2002, 670022 8 362 Hooker & Perron 2002, 982142 7 414 Hooker & Perron 2002, 670024 8 363 Hooker & Perron 2002, 982143 6 415 Hooker & Perron 2002, 680004 6 364 Hooker & Perron 2002, 982360 6 416 Hooker & Perron 2002, 690058 6 365 Hooker & Perron 2002, 990414 6 417 Hooker & Perron 2002, 690059 6 366 Hooker & Perron 2002, 990445 4 418 Hooker & Perron 2002, 690067 6 367 Hooker & Perron 2002, 991224 6 419 Hooker & Perron 2002, 690068 6 368 Hooker & Perron 2002, 971335 4+4 420 Hooker & Perron 2002, 690069 6 369 Hooker & Perron 2002, 981919 4+4 421 Hooker & Perron 2002, 690070 6 370 Hooker & Perron 2002, 982108 8 422 Hooker & Perron 2002, 690089 6 371 Hooker & Perron 2002, 630093 8 423 Hooker & Perron 2002, 690093 6 372 Hooker & Perron 2002, 680026 4+4 424 Hooker & Perron 2002, 690094 6 373 Hooker & Perron 2002, 720020 4+4 425 Hooker & Perron 2002, 690095 6 374 Hooker & Perron 2002, 830241 4 426 Hooker & Perron 2002, 690098 6 375 Hooker & Perron 2002, 840090 4+4 427 Hooker & Perron 2002, 690099 6 376 Hooker & Perron 2002, 850096 4+4 428 Hooker & Perron 2002, 690101 8 377 Hooker & Perron 2002, 900980 4+4 429 Hooker & Perron 2002, 690102 8 378 Hooker & Perron 2002, 930796 8 430 Hooker & Perron 2002, 690103 8 379 Hooker & Perron 2002, 930797 4+4 431 Hooker & Perron 2002, 690104 8 380 Hooker & Perron 2002, 940118 4 432 Hooker & Perron 2002, 690106 8 381 Hooker & Perron 2002, 941035 4 433 Hooker & Perron 2002, 690108 8 382 Hooker & Perron 2002, 941267 4 434 Hooker & Perron 2002, 720009 6

72 CONCORDANCE TABLE & DATA (contd.) Figure Original reference Spokes Figure Original reference Spokes 435 Hooker & Perron 2002, 730044 6 487 Hooker & Perron 2002, 972224 6 436 Hooker & Perron 2002, 730047 6 488 Hooker & Perron 2002, 981284 6 437 Hooker & Perron 2002, 730048 6 489 Hooker & Perron 2002, 982187 6 438 Hooker & Perron 2002, 730049 6 490 Hooker & Perron 2002, 990573 6 439 Hooker & Perron 2002, 730050 6 491 Hooker & Perron 2002, 982244 8 440 Hooker & Perron 2002, 730052 6 492 Hooker & Perron 2002, 982298 8 441 Hooker & Perron 2002, 730053 6 493 Hooker & Perron 2002, 990307 6 442 Hooker & Perron 2002, 730054 8 494 Hooker & Perron 2002, 000344 6 443 Hooker & Perron 2002, 730055 8 495 Hooker & Perron 2002, 000536 6 444 Hooker & Perron 2002, 730056 8 496 Hooker & Perron 2002, 000537 6 445 Hooker & Perron 2002, 810011 8 497 Hooker & Perron 2002, 001924 8 446 Hooker & Perron 2002, 820053 6 498 Hooker & Perron 2002, 610445 8 447 Hooker & Perron 2002, 820068 6 499 Hooker & Perron 2002, 610879 8 448 Hooker & Perron 2002, 820071 6 500 Hooker & Perron 2002, 610880 8 449 Hooker & Perron 2002, 820072 6 501 Hooker & Perron 2002, 610924 6 450 Hooker & Perron 2002, 820074 6 502 Hooker & Perron 2002, 610928 6 451 Hooker & Perron 2002, 820075 6 503 Hooker & Perron 2002, 680979 4 452 Hooker & Perron 2002, 820077 6 504 Hooker & Perron 2002, 681000 8 453 Hooker & Perron 2002, 820081 6 505 Hooker & Perron 2002, 681210 6 454 Hooker & Perron 2002, 820083 6 506 Hooker & Perron 2002, 681212 6 455 Hooker & Perron 2002, 820090 6 507 Hooker & Perron 2002, 820538 6 456 Hooker & Perron 2002, 820093 6 508 Hooker & Perron 2002, 870540 8 457 Hooker & Perron 2002, 820094 6 509 Hooker & Perron 2002, 880133 6 458 Hooker & Perron 2002, 820098 6 510 Hooker & Perron 2002, 910033 8 459 Hooker & Perron 2002, 820104 6 511 Hooker & Perron 2002, 921035 6 460 Hooker & Perron 2002, 820107 6 512 Hooker & Perron 2002, 930321 6 461 Hooker & Perron 2002, 820109 6 513 Hooker & Perron 2002, 930620 8 462 Hooker & Perron 2002, 820110 6 514 Hooker & Perron 2002, 930688 6 463 Hooker & Perron 2002, 820111 6 515 Hooker & Perron 2002, 930751 8 464 Hooker & Perron 2002, 820112 6 516 Hooker & Perron 2002, 940032 6 465 Hooker & Perron 2002, 820113 6 517 Hooker & Perron 2002, 940820 8 466 Hooker & Perron 2002, 820119 6 518 Hooker & Perron 2002, 941577 6 467 Hooker & Perron 2002, 820121 6 519 Hooker & Perron 2002, 951114 6 468 Hooker & Perron 2002, 820122 6 520 Hooker & Perron 2002, 952445 6 469 Hooker & Perron 2002, 820123 8 521 Hooker & Perron 2002, 952452 6 470 Hooker & Perron 2002, 830409 6 522 Hooker & Perron 2002, 961986 6 471 Hooker & Perron 2002, 840729 6 523 Hooker & Perron 2002, 963064 6 472 Hooker & Perron 2002, 930289 8 524 Hooker & Perron 2002, 970844 6 473 Hooker & Perron 2002, 941349 6 525 Hooker & Perron 2002, 980485 8 474 Hooker & Perron 2002, 950785 6 526 Hooker & Perron 2002, 980663 6 475 Hooker & Perron 2002, 950883 6 527 Hooker & Perron 2002, 982180 6 476 Hooker & Perron 2002, 951026 6 528 Hooker & Perron 2002, 990474 6 477 Hooker & Perron 2002, 951246 8 529 Hooker & Perron 2002, 990521 4 478 Hooker & Perron 2002, 951266 6 530 Hooker & Perron 2002, 990861 6 479 Hooker & Perron 2002, 953345 6 531 Hooker & Perron 2002, 000297 6 480 Hooker & Perron 2002, 961703 6 532 Hooker & Perron 2002, 000877 6 481 Hooker & Perron 2002, 963258 8 533 Hooker & Perron 2002, 001111 6 482 Hooker & Perron 2002, 970667 6 534 Hooker & Perron 2002, 001112 4 483 Hooker & Perron 2002, 971291 6 535 Hooker & Perron 2002, 001531 4 484 Hooker & Perron 2002, 971292 6 536 Hooker & Perron 2002, 010048 4 485 Hooker & Perron 2002, 971296 6 537 Hooker & Perron 2002, 010224 2x4 486 Hooker & Perron 2002, 971722 6 538 Hooker & Perron 2002, 610427 6

73 CONCORDANCE TABLE & DATA (contd.) Figure Original reference Spokes Figure Original reference Spokes 539 Hooker & Perron 2002, 610429 8 568 Hooker & Perron 2002, 820363 6 540 Hooker & Perron 2002, 660293 6 569 Hooker & Perron 2002, 820411 8 541 Hooker & Perron 2002, 670158 4 570 Hooker & Perron 2002, 850077 6 542 Hooker & Perron 2002, 670159 4 571 Hooker & Perron 2002, 900668 8 543 Hooker & Perron 2002, 670367 8 572 Hooker & Perron 2002, 910493 6 544 Hooker & Perron 2002, 680187 2x4 573 Hooker & Perron 2002, 930917 4 545 Hooker & Perron 2002, 680188 2x4+4 574 Hooker & Perron 2002, 930997 4 546 Hooker & Perron 2002, 680189 2x4+4 575 Hooker & Perron 2002, 940315 2x4+4 547 Hooker & Perron 2002, 680199 2x4 576 Hooker & Perron 2002, 940869 6 548 Hooker & Perron 2002, 991710 6 577 Hooker & Perron 2002, 941064 6 549 Hooker & Perron 2002, 680352 4 578 Hooker & Perron 2002, 941254 8 550 Hooker & Perron 2002, 680353 4 579 Hooker & Perron 2002, 941355 6 551 Hooker & Perron 2002, 680939 6 580 Hooker & Perron 2002, 951018 6 552 Hooker & Perron 2002, 680940 6 581 Hooker & Perron 2002, 951144 8 553 Hooker & Perron 2002, 680942 6 582 Hooker & Perron 2002, 990647 6 554 Hooker & Perron 2002, 680956 6 583 Hooker & Perron 2002, 953535 6 555 Hooker & Perron 2002, 680957 5 584 Hooker & Perron 2002, 960990 8 556 Hooker & Perron 2002, 680960 8 585 Hooker & Perron 2002, 961101 4 557 Hooker & Perron 2002, 680961 8 586 Hooker & Perron 2002, 961489 4 558 Hooker & Perron 2002, 681082 8 587 Hooker & Perron 2002, 962388 8 559 Hooker & Perron 2002, 681130 8 588 Hooker & Perron 2002, 971990 8 560 Hooker & Perron 2002, 720524 8 589 Hooker & Perron 2002, 980426 6 561 Hooker & Perron 2002, 730225 2x4 590 Hooker & Perron 2002, 980427 6 562 Hooker & Perron 2002, 730332 4 591 Hooker & Perron 2002, 980428 6 563 Hooker & Perron 2002, 730333 4 592 Hooker & Perron 2002, 980432 6 564 Hooker & Perron 2002, 730671 8 593 Hooker & Perron 2002, 980433 6 565 Hooker & Perron 2002, 740071 4 594 Hooker & Perron 2002, 980434 6 566 Hooker & Perron 2002, 740173 6 595 Hooker & Perron 2002, 980437 6 567 Hooker & Perron 2002, 820242 4 596 Hooker & Perron 2002, 981871 6

74 Table 1: Summary of Green’s data (Green 1984, 283-287)

75 Table 2: Wheel symbolism (after Green 1984)

76 Table 3: Analysis of coins; All coins & continental coins

77 Table 4: Analysis of coins; All British coins & Atrebates

78 Table 5: Analysis of coins; Corieltauvi & Dobunni

79 Table 6: Analysis of coins: Iceni & Trinovante-Catuvellauni

80 Key to wheels

1 Vienne (Isere) 2 Bois-de-la-Neuve-Grange 3 Farley Heath 4 Gundestrup Cauldron 5 Gundestrup Cauldron 6 Niederwürzbach 7 Oehringen 8 Petronell 9 Petronell 10 Petronell 11 Susa 12 Starĕ Hradisko 13 Wederath 14 Fully 15 Bachos Bino 16 Housesteads 17 Felmingham Hall

81

Plate 1: Wheel symbols (Green 1984)

82

Plate 2: Wheel symbols (Green 1984)

83 Key to Coins

1 Cadurci, LVXTERIOS, Rhone Valley & Garonne Basin 2 Northern monnaie-á-la-croix, Rhone Valley & Garonne Basin 3 Rhineland; countermarked 4 Central western lyre type, Limoges 5 Danubian, Sasthieni 6 Rhoda type, south-western Gaul 7 Rhoda type, south-western Gaul 8 Rhoda type, south-western Gaul 9 Monnaie-á-la-croix, Rhone Valley & Garonne Basin 10 Monnaie-á-la-croix, Rhone Valley & Garonne Basin 11 Monnaie-a-la-croix, Rhone Valley & Garonne Basin 12 Monnaie-á-la-croix, Rhone Valley & Garonne Basin 13 Monnaie-á-la-croix, Rhone Valley & Garonne Basin 14 Monnaie-á-la-croix, Rhone Valley & Garonne Basin 15 Monnaie-á-la-croix, UNTIKIA, Rhone Valley & Garonne Basin 16 Monnaie-á-la-croix, Rhone Valley & Garonne Basin 17 Central western lyre type 18 Central Gaul lyre type 19 Osismi 20 Redones 21 Suessiones 22 Nervi, Epsilon type 23 British, with CAMUL (Cunobelin, Camulodunum) 24 IAZUS, Rhone Valley 25 Allobroges, bouquetin type 26 KAΛETEΔOY variant 27 Western Gaul, VIREDIOS 28 South or western Gaul, CMEP 29 Treviri 30 Treviri 31 Treviri 32 Treviri 33 Lexovii, SIMMISSOS PVBLICOS LIXOVIO. CATTOS VERGOBRETO 34 North-western Gaul – as 33 35 Dobunni, Britain 36 Cunobelin, Britain 37 Central Gaul 38 Redones, Amorica 39 Northern Gaul 40 Dobunni, Britain 41 Danube, Tótfalú type - dump 42 East Noricum - dump 43 Rhone Valley & Garonne Basin, Bridiers Emporiae type 44 Danube 45 Massalia type 46 Monnaie-á-la-croix, Rhoda type

84 Key to Coins (contd.)

47 Monnaie-á-la-croix, Rhoda type 48 Monnaie-á-la-croix, Rhoda type 49 Monnaie-á-la-croix 50 Monnaie-á-la-croix 51 Monnaie-á-la-croix 52 Monnaie-á-la-croix 53 Monnaie-á-la-croix 54 Monnaie-á-la-croix 55 Monnaie-á-la-croix 56 Monnaie-á-la-croix 57 Monnaie-á-la-croix, attributed to the Tolosates 58 Monnaie-á-la-croix, attributed to the Tolosates 59 Monnaie-á-la-croix, attributed to the Tolosates 60 Monnaie-á-la-croix, attributed to the Tolosates 61 Monnaie-á-la-croix, attributed to the Tolosates 62 Monnaie-á-la-croix, attributed to the Tolosates 63 Monnaie-á-la-croix, attributable to Volcae Arecomici, oppidum Nemausus 64 Monnaie-á-la-croix, attributable to Volcae Arecomici, oppidum Nemausus 65 Monnaie-á-la-croix, attributable to Volcae Arecomici, oppidum Nemausus 66 Monnaie-á-la-croix, attributable to Volcae Arecomici, oppidum Nemausus 67 Monnaie-á-la-croix, attributable to Volcae Arecomici, oppidum Nemausus 68 Monnaie-á-la-croix, attributable to Volcae Arecomici, oppidum Nemausus 69 Monnaie-á-la-croix, attributable to Nitiobroges or Petrucorii 70 Monnaie-á-la-croix, attributable to Nitiobroges or Petrucorii 71 Monnaie-á-la-croix, Bavaria and Switzerland 72 Monnaie-á-la-croix, Bavaria and Switzerland 73 Emporion type 74 Danubian Tótfalú hoard type 75 Danubian. Imitation Philip III tetradrachm 76 Allobroges 77 Allobroges 78 Aedui or Lingones Kaletedou series 79 Aedui or Lingones Kaletedou series 80 Aedui or Lingones Kaletedou series 81 Aedui or Lingones Kaletedou series 82 Aedui or Lingones Kaletedou series 83 Aedui or Lingones Kaletedou series 84 Aedui or Lingones Kaletedou series 85 Aedui or Lingones Kaletedou series 86 Aedui or Lingones Kaletedou series 87 Aedui or Lingones Kaletedou series 88 Aedui or Lingones Kaletedou series 89 Aedui or Lingones Kaletedou series 90 Volcae Arecomici, Massalia type 91 Volcae Arecomici, Massalia type 92 Volcae Arecomici, Massalia type

85 Key to Coins (contd.)

93 Huş-Vovrieşti type 94 Monnaie-á-la-croix 95 Monnaie-á-la-croix 96 Monnaie-á-la-croix 97 Monnaie-á-la-croix 98 Monnaie-á-la-croix 99 Monnaie-á-la-croix 100 Monnaie-á-la-croix 101 Monnaie-á-la-croix 102 Monnaie-á-la-croix, attributed to Tolosates 103 Monnaie-á-la-croix, attributed to Tolosates 104 Monnaie-á-la-croix, attributed to Tolosates 105 Monnaie-á-la-croix, attributed to Tolosates 106 Monnaie-á-la-croix, attributed to Tolosates 107 Monnaie-á-la-croix, attributed to Tolosates 108 Monnaie-á-la-croix, attributed to Volcae Arecomici, oppida Nemausus 109 Monnaie-á-la-croix, attributed to Volcae Arecomici, oppida Nemausus 110 Monnaie-á-la-croix, attributed to Volcae Arecomici, oppida Nemausus 112 Monnaie-á-la-croix, attributed to Volcae Arecomici, oppida Nemausus 113 Monnaie-á-la-croix, attributed to Volcae Arecomici, oppida Nemausus 115 Monnaie-á-la-croix, attributed to Volcae Arecomici, oppida Nemausus 115 Monnaie-á-la-croix, attributed to Volcae Arecomici, oppida Nemausus 116 Monnaie-á-la-croix, attributed to Volcae Arecomici, oppida Nemausus 117 Monnaie-á-la-croix, attributed to Volcae Arecomici, oppida Nemausus 118 Monnaie-á-la-croix, attributed to Volcae Arecomici, oppida Nemausus 119 Monnaie-á-la-croix, attributed to Volcae Arecomici, oppida Nemausus 120 Monnaie-á-la-croix, attributed to Nitiobroges or Petrucorii 121 Monnaie-á-la-croix, attributed to Nitiobroges or Petrucorii 122 Monnaie-á-la-croix, attributed to Nitiobroges or Petrucorii 123 Monnaie-á-la-croix, attributed to Nitiobroges or Petrucorii 124 Monnaie-á-la-croix, attributed to Nitiobroges or Petrucorii 125 Monnaie-á-la-croix 126 Monnaie-á-la-croix, attributed to Ruteni 127 Monnaie-á-la-croix, attributed to Ruteni 128 Aedui or Lingones Kaletedou series 129 Aedui or Lingones Kaletedou series 130 Aedui or Lingones Kaletedou series 131 Aedui or Lingones Kaletedou series 132 Aedui or Lingones Kaletedou series 133 Aedui or Lingones Kaletedou series 134 Massalian type minim 135 Aedui or Lingones Kaletedou series 136 East Norican inscribed tetradrachm 137 Imitation tetradrachm of Philip II, with wheel-countermark 138 Branch-arm type tetradrachm 139 Garonne basin, Rhoda type

86 Key to Coins (contd.)

140 Monnaie-á-la-croix, Tolosates, Toulouse, mid 2nd century BC 141 Monnaie-á-la-croix, Cadurci 142 Rheims, copy of Philip II stater 143 Britain, Tasciovanus, Catuvellauni 144 Britain, Cunobelin 145 Osismii 146 Suessiones 147 Britain, copy of Suessione stater 148 Britain, Selsey 149 Wonersh, Surrey 150 Britain, Selsey 151 Britain, Commius 152 Dobunni stater, CATTI 153 British I stater, Coritani 154 British stater, Freckenham type, Iceni 155 Allobroges (Rhonetal) 156 Monnaie-á-la-croix 157 Monnaie-á-la-croix 158 Vichy type 159 KAΛETEΔOY 160 Senones quarter stater 161 East Gaulish 162 East Gaul wheel stater 163 North Swiss Büschel coin 164 East Norican Wuschelkopfgruppe tetradrachm 165-367 Atrebates 368-386 Corieltauvi 387-490 Dobunni 491-527 Iceni 528-596 Trinovante-Catuvellauni

87 Plate 3: Celtic coins (Allen 1980)

88 Plate 4: Celtic coins (Allen 1980)

89 Plate 5: Celtic coins (Allen 1980)

90 Plate 6: Celtic coins (Allen 1987)

91 Plate 7: Celtic coins (Allen 1987)

92 Plate 8: Celtic coins (Allen 1987)

93 Plate 9: Celtic coins (Allen 1987)

94 Plate 10: Celtic coins (Allen 1987)

95 Plate 11: Celtic coins (Allen 1987)

96 Plate 12: Celtic coins (Allen 1978)

97 Plate 13: Celtic coins (Allen 1978)

98 Plate 14: Celtic coins (Castelin 1985)

99 Plate 15: Celtic coins (Hooker & Perron (1999-2002)

100 Plate 16: Celtic coins (Hooker & Perron (1999-2002)

101 Plate 17: Celtic coins (Hooker & Perron (1999-2002)

102 Plate 18: Celtic coins (Hooker & Perron (1999-2002)

103 Plate 19: Celtic coins (Hooker & Perron (1999-2002)

104 Plate 20: Celtic coins (Hooker & Perron (1999-2002)

105 Plate 21: Celtic coins (Hooker & Perron (1999-2002)

106 Plate 22: Celtic coins (Hooker & Perron (1999-2002)

107 Plate 23: Celtic coins (Hooker & Perron (1999-2002)

108 Plate 24: Celtic coins (Hooker & Perron (1999-2002)

109 Plate 25: Celtic coins (Hooker & Perron (1999-2002)

110

Plate 26: Celtic coins (Hooker & Perron (1999-2002)

111 Plate 27: Celtic coins (Hooker & Perron (1999-2002)

112 Plate 28: Celtic coins (Hooker & Perron (1999-2002)

113 Plate 29: Celtic coins (Hooker & Perron (1999-2002)

114 Plate 30: Celtic coins (Hooker & Perron (1999-2002)

115 Plate 31: Celtic coins (Hooker & Perron (1999-2002)

116 Plate 32: Celtic coins (Hooker & Perron (1999-2002)

117 Plate 33: Celtic coins (Hooker & Perron (1999-2002)

118 Plate 34: Celtic coins (Hooker & Perron (1999-2002)

119 Plate 35: Celtic coins (Hooker & Perron (1999-2002)

120

Plate 36: Coins using tribes in Late Iron Age Britain (Rudd 2001)

121

Plate 37: Distribution of Jupiter-Giant columns along the Rhine (after Bauchhenss 1981).

122 Plate 38: Epona finds in eastern Gaul and the Rhineland (after Magnen & Thevenot 1955). Red boundary = Plate 37.

123 Plate 39: Map of Celtic peoples (after RK 1999). Red boundary = Plate 38.

124 BIBLIOGRAPHY

AM Audacht Morainn, trans. Kelly Bell. Civ. Lucan, The Civil Wars BG Caesar, The Gallic War CMT Cath Maighe Tuired, trans. Gray De Arch. Vitruvius, The Ten Books on Architecture Met. Dinds. Metrical Dindsenchas, trans. Gwynne Metam. Apulieus, Metamorphoses (The Golden Ass) Mor. Plutarch, Moralia NH Pliny, Natural History Persae Aeschylus, Persians Thuc. Thucydides

Primary Sources

Aeschylus, Aeschylus I: Suppliant Maidens, Persians, Prometheus, Seven Against Thebes, trans. Smyth H W (1973), London & Cambridge, Mass., 108-207. Apulieus, The Golden Ass, trans. Kenney E J (1998), London. Aristotle, The Complete Works of Aristotle, 2 vols., ed. Barnes J (1984), Princeton, New Jersey. Caesar, The Gallic War, trans. Edwards H J (1922), London & New York. Cassius Dio, Dio’s Roman History trans. Cary E (1914-27), London. Cath Maige Tuired, trans. Gray E A (1982), Kildare. Cross T P & Slover C H (1936), Ancient Irish Tales, Dublin. Diodorus of Sicily, vol.11, trans. Walton F R (1957), London & Cambridge, Mass. 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. Kelly F ed. (1976), Audacht Morainn, Dublin. Livy, vol. 11. trans. Sage E T (1936), Cambridge Mass. Lucan, The Civil Wars, trans. Duff J D (1928), London & New York. The Metrical Dindshenchas: RIA Todd Lecture Series Vols. VIII-XII, trans. Gwynn E (1900-1935), Dublin. 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. Poetic Eddas, trans. Bellows H A (1936), http://www.sacred-texts.com/neu/poe/ Posidonius, The Fragments: Volume III, The Translation of the Fragments, trans. Kidd I G (1999), Cambridge. Thucydides, History of the Peloponnesian War, trans. Crawley R (1910), London & New York. 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.

125 Secondary Sources

Alabe Inc (2003), Venus cycle 1900-2050, http://www.alabe.com/VeCycle-1.pdf Allen D F (1978), An Introduction to Celtic Coins, London. Allen D F (1980), The Coins of the Ancient Celts, Edinburgh. Allen D F (1987), Catalogue of the Celtic Coins in the British Museum with Supplementary Material, 3 vols., London. Altheim F (1938), A History of Roman Religion, trans. Mattingly H, London. Armbruster C W & R A Williamson (1993), Sun and sun serpents: continuing observations in south-eastern Utah, in Ruggles C (ed.), Archaeoastronomy in the 1990s, Loughborough, 219-226. Audin A (1945), Les Fêtes Solaires: Essai sur la Religion Primitive, Paris. Aveni A F (1988), The Thom paradigm in the Americas: the case of cross-circle designs, in Ruggles C L N (ed.), Records in Stone, Cambridge, 442-472. Aveni A F, H Hartung & B Buckingham (1978), The pecked cross symbol in ancient Mesoamerica, Science 202, 267-279. Baity E C (1973), Archaeoastronomy and ethnoastronomy so far, Current Anthropology, 14, 389-449. Ballew P (2003), Maths Words Pg 13, http://www.pballew.net/arithm13.html Barlai K (1988), Some archaeological problems in East-Central Europe, in Aveni A F (ed.), World Archaeoastronomy, Cambridge, 436-440. Bauchhenss G & P Noelke (1981), Die Iupitersäulen in den Germanischen Provinzen, Koln & Bonn. Beard M, J North & S Price (1998), Religions of Rome: Volume 1. A History, Cambridge. Bertrand A (1897), La Religion des Gaules, Paris. Blank W M (1999), Imitations of the coins of Massalia; Silver coins, http://www.kernunnos.com/dlt/massaliaImit_jpgs.html Blench R & M Spriggs (1999), Archaeology and Language III: Artefacts, Languages and Texts, London & New York. Bogucki P (1999), Early agricultural societies, in Barker G (ed.), Companion Encyclopedia to Archaeology, 2 vols., London & New York. Bóna I (1960), Clay models of Bronze Age wagons and wheels in the Middle Danube basin, Acta Archaeologica Hungarica 12, 83-111. Boucher S (1976) Recherches sur les Bronzes Figurés de Gaule Pré-Romaine et Romaine, Paris & Rome. Broda J (1993), Astronomical knowledge, calendrics and sacred geography in ancient Mesoamerica, in Ruggles C L N & N J Saunders (eds.), Astronomies and Cultures, Niwot, Colorado, 253-295. Brøndsted J (1939), Danmarks Oldtid II: Bronzealderen, Copenhagen. Burl H A W (1982), Pi in the sky, in Heggie D C (ed.), Archaeoastronomy in the Old World, Cambridge, 141-169. Carmichael A (1900-1941), Carmina Gadelica, 5 vols., Edinburgh. Carney J (1955), Studies in Irish Literature and History, Dublin. Castelin K (1985), Keltische Münzen: Katalog der Sammlung des Schweizerischen Landesmuseums Zurich, Vol. 2, Bern. Chamberlain von Del (ed.) (1988), Proceedings of the First International Ethnoastronomy Conference, Washington DC. Champeux J (1982), Fortuna, Rome. Chapman M (1992), The Celts: The Construction of a Myth, London & New York.

126 Chenet G (1919), Rouelles de plomb et persistence d’emploi des rouelles Gauloises, Bulletin Archéologiques: Comité des Travaux Historiques et Scientifiques Fasc. A, Antiquités Nationales, 243-251. Childe V G (1951), The first wagons and carts – from the Tigris to the Severn, Proc. Prehistoric Soc. 17, 177-194. Christensen N I (September 2001), Guds borg ved Vårby Å, Arkæologi På Nettet 34, http://www.ancient-astronomy.dk/sepmag01.htm Christensen N I & E I Christensen (November 2000), Hjulkors i Tanum - en arkæoastronomisk undersøgelse , Arkæologi På Nettet 24, http://www.ancient-astronomy.dk/novmag00.htm Claridge A (1988), Rome: An Oxford Archaeological Guide, Oxford. Collis J R (1984), Adieu Hallstatt! Adieu La Tène!, Revue Aquitania, Supplement 1, 327-330. Collis J (1984a), Oppida: Earliest Towns North of the Alps, Sheffield. Collis J (1984b), The European Iron Age, Oxford. Collis J R (1994), Celtic fantasy, British Archaeological News, new series 11, 5. Compagnie Generale de Bourse (2003), Ædui - Éduens (Région du Mont-Beuvray) (IIe - Ier siècles avant J.-C.); Denier à la tête casquée, fourré - c. 80-50 AC, http://www.cgb.fr/monnaies/vso/v15/gb/monnaiesgb99d4.html?depart=503&nbfic=15 15 Cook A B (1914), Zeus: A Study in Ancient Religion, Vol. I, Cambridge. Cook A B (1925), Zeus: A Study in Ancient Religion, Vol. II, Cambridge. Cunliffe B & P Davenport (1985), The Temple of Sulis Minerva at Bath: Volume 2, Oxford. Cunliffe B W (1984), Relations between Britain & Gaul in the First Century BC and the First Century AD, in Macready S & F H Thompson (eds.), Cross Channel Trade between Gaul & Britain in the pre-Roman Iron Age, London, 3-23. Cunliffe B (1997), The Ancient Celts, Oxford. Cunliffe B & T Rowley (1976), Oppida: The Beginnings of Urbanisation in Barbarian Europe, Oxford. Danaher K (1981), Irish folk tradition and the Celtic calendar, in O’Driscoll R (ed.), 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 (eds.) Military and Civilian in Roman Britain, BAR Series 136, Oxford, 93-128. Dechelette J (1910), Manuel d’Archéologie Préhistorique: Celtique et Gallo-Romaine: II Archéologie Celtique ou Protohistorique, Paris. Dechelette J (1927), Manuel d’Archéologie Préhistorique: Celtique et Gallo-Romaine: IV Second Age du Fer ou Époque de la Tène, Paris. Delemen I (1999), Anatolian Rider Gods, Bonn. Derks T (1998), Gods, Temples and Ritual Practices, Amsterdam. Egg M & Ch. Pare (1993), Keltische Wagen und ihre Vorlaufer, in Dannheimer H & R Gebbard (eds.), Das keltische Jahrtausend, Austellungskatalogue der prähistorischen Staatssammlung München Band 23, 209-18. Eliade M (1957), The Sacred and the Profane, trans. Trask W R (1959) London & New York. Erdoes R & A Ortiz (1984), American Indian Myths and Legends, New York. EuroPreArt (2003), Denmark Vester, Lilla Strandbygård I, http://www.europreart.net/cgi- bin/baserun.exe?_cfg=record.cfg&_fil=code%3D%22denma080%22

127 Février P-A (1973), Origin and growth of the cities of southern Gaul to the third century AD: An assessment of the most recent archaeological discoveries, J. Roman Studies LXIII, 1-28. Filip J (1978), Celtic strongholds: an indicator and reflection of the evolution and structure of Celtic society, Archeologické Rozhledy XXX, 420-432. Fox C (1946), A Find of the Early Iron Age from Llyn Cerrig Bach, Anglesey, Cardiff. Frazer J G (1922), The Golden Bough: A Study in Magic and Religion, London. Frey O H (1968), Die Situla von Kuffarn, ein figgurlich verzierter Bronzeblecheimer aus Zeit um 400 v. Chr., Veröffentlichungen aus dem Naturhistorischen Museum NF 4. Frey O H & W Lucke (1962), Die Situla in Providence; Ein Beitrag zur Situlkunst des Osthallstattkreises, Römische-Germanische Forschungen Band 23, Berlin. Gaidoz H (1884), Le dieu Gauloise du soleil et le symbolisme de la roué, Revue Archéologique 2, 7-37. Gamble C (1982), Leadership and surplus production, in Renfrew C & S Shennan (eds.), Ranking, Resource and Exchange: Aspects of the Archaeology of Early European Society, Cambridge, 100-105. Gell A (1992), The Anthropology of Time: Cultural Constructions of Temporal Maps and Images, Oxford & Providence, Rhode Island. Gelling P & H E Davidson (1969), The Chariot of the Sun and Other Rites and Symbols of the Northern Bronze Age, London. Gingerich O (1988), Reflections on the role of archaeoastronomy in the history of astronomy, in Aveni A (ed.), World Archaeoastronomy, Cambridge, 38-44. Glob P V (1969), Helleristninger i Danmark, Copenhagen. Glob P V (1974), The Mound People, trans. Bulman J, London. Gordon R (1990), Religion in the Roman Empire: the civic compromise and its limits, in Beard M & J North (eds.), Pagan Priests, London, 233-257. Green H J M (1986), Religious cults at Roman Godmanchester, in Henig M & A King (eds.), Pagan Gods and Shrines of the Roman Empire, Oxford, 29-56. Green M J (1976), The Religions of Civilian Roman Britain: A Corpus of Religious Material from the Civilian Areas of Roman Britain , BAR 24, Oxford. Green M J (1978), A Corpus of Small Cult Objects from the Military Areas of Roman Britain, Oxford. Green M J (1979), The worship of the Romano-Celtic Wheel-god in Britain seen in relation to the Gaulish evidence, in Latomus 38, fasc. 2, 345-367. Green M J (1981), Model objects from military areas of Roman Britain, Britannia XII, 253-269. Green M J (1984), The Wheel as a Cult Symbol in the Romano-Celtic World, Brussels. Green M J (1986a), The Gods of the Celts, Stroud, Gloucestershire. Green M J (1986b), Jupiter, Taranis and the solar wheel, in Henig M & A King (eds.), Pagan Gods and Shrines of the Roman Empire, Oxford, 65-76. 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 (1994), The Gods of Roman Britain, Princes Risborough. Green M J (1995), Celtic Goddesses; Warriors, Virgins and Mothers, London. Green M J (1997), Exploring the World of the Druids, London. Green M J A (1996), Celtic Art, London.

128 Green M J A (2001), Seeing the Wood for the Trees: The Symbolism of Trees and Wood in Ancient Gaul and Britain, Aberystwyth. Green M J A & P Webster (2002), Artefacts and Archaeology, Cardiff. Hall E (1989), Inventing the Barbarian, Oxford. Hall J M (1997), Ethnic Identity in Greek Antiquity, Cambridge. Hardie P (1992), Augustan poets and the mutability of Rome, in Powell A, Roman Poetry and Propaganda in the Age of Augustus, London, 59-82. Harding R (2000), European Societies in the Bronze Age, Cambridge. Härke H (1989), Transformation or collapse? Bronze Age to Iron Age settlement in West Central Europe, in Sørensen M L S & R Thomas (eds.), The Bronze Age- Iron Age Transition in Europe, Oxford, 184-203. Haselgrove C C (1982), Wealth, prestige and power: the dynamics of late Iron Age political centralisation in south-east England, in Renfrew C & S Shennan (eds.), 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.), Oxford, Military and Civilian in Roman Britain, 5-64. Hassan F A (1999), Population dynamics, in Barker G (ed.), Companion Encyclopedia to Archaeology, 2 vols., London & New York, 672-713. Hatt J-J (1965) Essai sur l’évolution de la religion gauloise, Rev. des Etudes Anc. vol. 57, 80-125. Hatt J-J (1970), Celts and Gallo-Romans, Geneva. Hatt J-J (1989), Mythes et Dieux de la Gaule, Paris. Haverfield F (1917), Roman Cirencester, Archaeologia 69, 161-209. Heath Sir T L (1913), A History of Greek Astronomy to Aristarchus, Oxford. Heath Sir T L (1921), A History of Greek Mathematics, 2 vols., Oxford. Heath Sir T L (1931), Manual of Greek Mathematics, Oxford. Heath Sir T L (1932), Greek Astronomy, London. Heath Sir T L (1932), Mathematics in Aristotle, Oxford. 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 (eds.), Pagan Gods and Shrines of the Roman Empire, Oxford, 159-170. Henig M (1993), Corpus Signorum Imperii Romani Vol. 1 Fasc. 7; Roman Sculpture from the Cotswold Region, Oxford. Hicks R (1984), Stones and henges: megalithic astronomy reviewed, in Krupp E C (ed.), Archaeoastronomy and the Roots of Science, Boulder, Colorado, 169- 210. Hicks R (1988), The year at Dromberg, in Aveni A F (ed.), World Archaeoastronomy, Cambridge, 470-482. Hooker J & C Perron (1999-2002), Celtic Coin Index, http://www.writer2001.com/cciwriter2001/index.htm Holtz J (2002), Why the earliest sunset, latest sunrise, and shortest day of the year occur on different dates, http://members.aol.com/jwholtz/analemma/analemma.htm Huskinson J (1994), Corpus Signorum Imperii Romani Vol. 1 Fasc. 8; Roman Sculpture from Eastern England, Oxford. Iwaniszewski S (1993), Mesoamerican cross circle designs revisited, in Ruggles C (ed.), Archaeoastronomy in the 1990s, Loughborough, 288-297.

129 James S (1999), The Atlantic Celts: Ancient People or Modern Invention, London. Jenkins F (1958), The cult of the pseudo-Venus in Kent, Archaeologia Cantiana, vol. 72, 60ff. Jones G D B (1984), Becoming different without knowing it; the role and development of vici, in Blagg T F C and A C King (eds.) Military and Civilian in Roman Britain, BAR Series 136, Oxford, 75-92. Jope E M (1956), Vehicles and harness, in Singer C, E J Holyard, A R Hall & T I Williams (eds.), A History of Technology, Oxford, 537-562. Karl R (forthcoming a), Erwachen aus dem langen Schlaf der Theorie? Ansätze zu einer keltologischen Wissenschaftstheorie, in Poppe E (ed.), Keltologie heute; themen und Fragestellen, Akten des 3, Deutschen Keltologensymposiums – Marburg, März 2001, Studien und Texte zur Keltologie 5, Münster. Karl R (forthcoming b), Celtic studies as cultural studies, Celtic Cultural Studies Online, http://www.cyberstudia.com/celtic-cultural-studies/ Karl R & Stifter (forthcoming), Carpat Carpentum; Die keltischen Grundlagen des ,Streit’wagens der irischen Sagentradition, in Eibner A, R Karl, J Leskovar, K Löcker & C Zingerle (eds.), Pferd und Wagen in der Eisenzeit; Arbeitstagung des AK Eisenzeit der ÖGUF gemeinsam mit der AG Reiten und Fahren von 23 25.2.2000 in Wein, Wiener keltologische Schriften 2, Wein. Koch J T (1981), The loss of final syllables and the loss of declension in Britonnic, Proc. Harvard Celtic Colloquium 1, 21-51. Koch J T (1986), New thoughts on Albion, Ierné and the Pretanic Isles, Proc. Harvard Celtic Colloquium 6, 1-28. Krüger M-L (1970), Corpus Signorum Imperii Romani: Corpus der Skulpturen der Römischen Welt: Österreich, Band 1, Faszikel 3 - Die Reliefs des Stadtgebietes von Carnuntum: I Teil; Die Figürlichen Reliefs, Wein, Köln & Graz. Krüger M-L (1972), Corpus Signorum Imperii Romani: Corpus der Skulpturen der Römischen Welt: Österreich: Band 1, Faszikel 4 - Die Reliefs des Stadtgebietes von Carnuntum: II Teil; Die Dekorativen Reliefs, Wein, Köln & Graz. Krupp E C (1994), Echoes of Ancient Skies, Oxford. Laing J (1982), The Greek and Roman Gods, North Pomfret, Vermont. Lendering J (2003), Legio XV Apollinaris, http://www.livius.org/le-lh/legio/xv_apollinaris.html Lewis M J T (1966), Temples in Roman Britain, Cambridge. Lewis H & H Pedersen (1937, reprinted 1989), A Concise Comparative Celtic Grammar, 3rd ed. 2nd impression, Göttingen. Littauer MA & J H Crouwel (1974), Terracotta models as evidence for vehicles with tilts on the ancient Near East, Proc. Prehistoric Soc. 40, 20-36. Long T (1986), Barbarians in Greek Comedy, Carbondale & Edwardsville, Illinois. Luck G (1999) Witches and sorcerers in Classical literature, in Flint V, R Gordon, G C Luck & D Ogden, Witchcraft and Magic in Europe: Volume 2, Ancient Greece and Rome, London, 91-158. Lynn C J (1992), The Iron Age mound in Navan Fort: A physical realization of Celtic religious beliefs? Emania 10, 33-57. 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.

130 Mac Mathúna S & A Ó Corráin (1997), Collins Pocket Irish Dictionary, London & New York. Mac Neill E (1928), On the notation and chronography of the Calendar of Coligny, Eriu X, 1-67. MacGregor A (1989), Bone, antler and horn industries in the urban context, in Serjeantson D & T Waldron (eds.), Diet and Craft in Towns, Oxford, 107-128. MacGregor M (1962), The early Iron Age metalwork hoard from Stanwick, N R Yorks, Proc. Prehistoric Soc. 28, 17-57, Mackie E (1988), Investigating the prehistoric solar calendar, in Ruggles C L N (ed.), Records in Stone, Cambridge, 206-231. Mackintosh M (1995), The Divine Rider in the Art of the Western Roman Empire, Oxford. Magie 1950, D (1950), Roman Rule in Asia Minor to the End of the Third Century after Christ, 2 vols., Princeton. Magnen R & E Thevenot (1953), Épona: Déese Gauloise des Chevaux Proctectrice des Cavaliers, Bordeaux. Mallory J P (1989), The career of Conall Cernach, Emania 6, 22-28. McCluskey S C (1990), Astronomies and rituals at the dawn of the Middle Ages, in Ruggles C L N & N J Saunders, Astronomies and Cultures, Niwot, Colorado, 100-123. McCone K (1990), Pagan Past and Christian Present in Early Irish Literature, Maynooth. Megaw J V S (1970), Art of the European Iron Age, Bath. Megaw J V S & M R Megaw (1996), Ancient Celts and modern ethnicity, Antiquity 70, 175-81. Metzler J (1986), Ein frühlatènezeitliches Gräberfeld mit Wagenbestattung bei Grosbous-Vichtne, Arch. Korrbl. 16, 161-77. Millett M M (1990), The Romanization of Britain, Cambridge. Mitchell S (1993), Anatolia - Land of Men and Gods in Asia Minor, vol. 1, Oxford. Mollerup A (2003), Sundial construction, http://www.geocities.com/macsida/sundial_construction.html Moreland J (1999), Production and exchange in historical archaeology, in Barker G (ed.), Companion Encyclopedia to Archaeology, 2 vols., London & New York, 637-671. Morgan C (1993), The origins of pan-Hellenism, in Marinatos N & R Hägg (eds.), Greek sanctuaries: New Approaches, London & New York, 18-44. Morris I (1998), Archaeology and archaic Greek history, in Fisher N & H van Wees (eds.), Archaic Greece: New Approaches and New Evidence, London, 1-92. Mynard D G (1992), Windows on the Past: Archaeology in Milton Keynes, Milton Keynes. Neugebauer O (1975), A History of Ancient Mathematical Astronomy, Berlin. Neverov O (1986), Nero-Helios, in Henig M & A King (eds.), Pagan Gods & Shrines of the Roman Empire, Oxford, 189-195. Oaks L S (1986), The goddess Epona, in Henig M & A King (eds.), Pagan Gods & Shrines of the Roman Empire, Oxford, 77-84. Olmstead G S (1979), The Gundestrup Cauldron, Collection Latomus 162, Brussels Olmstead G S (1988), Gaulish and Celti-Iberian poetic inscriptions, Mankind Quarterly XXVIII 4, 339-387. Olmsted G S (1992), The Gaulish Calendar, Bonn. Olmsted G S (1994), The Gods of the Celts and the Indo-Europeans, Innsbruck.

131 Olmstead G S (2001), A Definitive Reconstructed Text of the Coligny Calendar, Washington. O’Rahilly T F (1946), Early Irish History and Mythology, Dublin. Parra E J, A Marcini, J Akey, J Martinson, M A Batzer, R Cooper, T Forrester, D B Allison, R Deka, R E Ferrell & M D Shriver (1998), Estimating African American Admixture Proportions by Use of Population-Specific Alleles, Am. J. Hum. Genet., 63:1839-1851, http://www.journals.uchicago.edu/AJHG/journal/issues/v63n6/980409/980409.html Pare C (1987), Wheels with thickened spokes and the problems of cultural contact between the Aegean world and Europe in the Late Bronze Age, Oxford Journal of Archaeology 6(1), 43-68. Paterson T (17 March 2002), Mysterious gold cones ‘hats of ancient wizards’ http://portal.telegraph.co.uk/news/main.jhtml?xml=/news/2002/03/17/wwiz17.xml&s Sheet=/news/2002/03/17/ixworld.html Phillips C R III (1991), Nullum crimen sine lege: Socioreligious sanctions on magic, in Faraone C A & D Obbink, Magika Hiera; Ancient Greek Magic and Religion, Oxford, 260-278. Phillips E J (1976), A Roman figured capital in Cirencester, J. Brit. Archaeol. Assoc. 129, 25-41, with pls. 10-11. Piggott S (1968), The earliest wheeled vehicles and the Caucasian evidence, Proc. Prehistoric Soc. 34, 266-318. Polignac F de (1995), Cults, Territory, and the Origins of the Greek City-State, trans. Lloyd J, Chicago & London. Powell T G E (1958), The Celts, London. Primas M (1974), Die Latènezeit im Alpinen Raum, in Vogt E (ed.), Ur- und Frühgeshichtliche Archäologie der Schweiz: Band IV, Die Eisenzeit, Basel, 89-104 Pydyn A (1999), Exchange and cultural Interactions: A Study of Long Distance Trade and Cross-Cultural Contacts in the Later Bronze Age and Early Iron Age in Central and Eastern Europe, Oxford. Radhakrishnan S (1923), Indian Philosophy, 2 vols. London. Radoslavova T (1993), Astronomical knowledge in Bulgarian lands during Neolithic and early Bronze Age, in Ruggles C (ed.), Archaeoastronomy in the 1990s, Loughborough, 107-116. 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 (1983), A Catalogue of Irish Iron Age Antiquities, Marbury. Raftery B (1989) Barbarians to the West, in Barrett J C, A P Fitzpatrick & L Macinnes (eds.), Barbarians and Romans in North-West Europe from the Later Republic to Late Antiquity, BAR International Series 471, Oxford, 117- 152. Raftery B (1994), Pagan Celtic Ireland, London. 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. Rawson E (1969), The Spartan Tradition in European Thought, Oxford. Renfrew C (1987), Archaeology and Language, Harmondsworth. Renfrew C & P Bahn (1996), Archaeology: Theories, Method and Practice, 2nd ed., London. Renfrew C & S Shennan (1982), Ranking, Resource and Exchange: Aspects of the Archaeology of Early European Society, Cambridge.

132 Rhys J (1909-10), The Coligny Calendar, Proc. Brit. Acad. 1, 207-289. Richmond I (1938), The Roman fort at Bewcastle, Cumberland & Westmorland Soc. Trans. 38, 195-237. Ritchie J N G (1989), Archaeology and astronomy: an archaeological view, in Heggie D C (ed.), Archaeoastronomy in the Old World, Cambridge, 25-44. Rivet A L F & C Smith (1979), The Place Names of Roman Britain, London. RK (1999), Map of Celtic kingdoms, http://worldcoincatalog.com/AC/C1/CelticKingdoms/CelticKingdoms.jpg Rodwell W (1978), Rivenhall and the emergence of first century villas in northern Essex, in Todd M (ed.), Studies in the Romano-British Villa M Todd, Leicester, 11-32. Ross A (1967), Pagan Celtic Britain, London. Rudd C (2001), Coin using tribes in Late Iron Age Britain, http://www.kernunnos.com/celticcoins/index.html 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. Sarton G (1959, reprinted 1987), Hellenistic Science and Culture in the Last Three Centuries BC, New York. Sawyer III F W (1995), Of analemmas, mean time, and the analemmatic sundial, http://www.longwoodgardens.org/Sundial/Analemma.html Sessle E J (1994), Exploring the limitations of the sovereignty goddess through the role of Rhiannon, Proc. Harvard Celtic Colloquium 16, 9-13. Shennan S J (1982), Exchange and ranking: the role of amber in the earlier Bronze Age of Europe, in Renfrew C & S Shennan, Ranking, Resource and Exchange: Aspects of the Archaeology of Early European Society, Cambridge, 33-45. Sherratt A G (1986), Two new finds of wooden wheels from late Neolithic and early Bronze Age Europe, Oxford J. Archaeology 5(2), 243-248. Shiel R (2001), The potential for using religious belief to derive environmental information on past societies, with a case study on the environment of Attica, in Albarell U (ed.), Environmental Archaeology: Meaning and Purpose, Dordrecht, 220-248. Sims-Williams P (1998), Celtomania and Celtoscepticism, Cambrian Medieval Celtic Studies 36, 1-35. Sinclair R M & A Sofaer (1993), A method for determining limits of the accuracy of naked-eye locations of astronomical events, in Ruggles C (ed.), Archaeoastronomy in the 1990s, Loughborough, 178-184. Smith R A (1925), British Museum Guide to Early Iron Age Antiquities 1925, Ipswich. Snodgrass A M (1980), Archaic Greece: The Age of Experiment, London. Sourvinou-Inwood C (1993), Early sanctuaries, the eighth century and ritual space, in Marinatos N and R Hägg (eds.), Greek Sanctuaries: New Approaches, London & New York, 1-17. Sprockhoff W (1955), Central European Urnfield culture and Celtic Le Tène: An outline, Proc. Prehistoric Soc. 25, 257-281. Stead I M (1967), A La Tène III burial at Welwyn Garden City, Archaeologia 101, 1- 62. Stead I M (1976), The earliest burials of the Aylesford Culture, in Sieveking G de G, I H Longworth & K E Wilson (eds.) Problems in Economic & Social Archaeology, London, 401-416.

133 Stead I M (1989), Cart burials at Garton Station and Kirkburn, in Halkon P (ed.), New Light on the Parisi; Recent Discoveries in Iron Age and Roman East Yorkshire, Hull, 1-6. Stuckert R P (1976), The black ancestry of white Americans, in Hammond P B (ed.), Physical Anthropology and Archaeology: Introductory Readings, London & New York, 135-139. Surber D (2003), Browsing Celtic coins in the Wildwinds databank, http://www.wildwinds.com/coins/celtic/i.html Thomas J (1991), Understanding the Neolithic, London & New York. Thorpe I J (1981), Ethnoastronomy: its patterns and archaeological implications, in Ruggles C L N & A W R Whittle (eds.), Astronomy and Society in Britain During the Period 4000-1500 BC, Oxford, 275-288. Toutain J (1920), Les Cultes Païens Dans L’Empire Romain. Première Partie: Les Provinces Latines. Tome III: Les Cultes Indigènes Nationaux et Locaux, Paris. University of Denver (1999), M Vitruvius Pollio and the analemma, http://www.du.edu/~jcalvert/classics/analemma.htm Urschel B (2003), The Analemma, http://www.analemma.com/Pages/framesPage.html US Naval Observatory (2003), Phases of the moon, http://aa.usno.navy.mil/data/docs/MoonPhase.html#y2005 Van Wees H (1992), Status Warriors: War, Violence and Society in Homer and History, Amsterdam. Villefosse A H de (1881), Note sur un bronze découvert a Landouzy-la-Ville (Aisne), Revue Archéologique 41, 1-13. Vilhjámsson T (1993), Time reckoning in Iceland before literacy, in Ruggles C (ed.), Archaeoastronomy in the 1990s, Loughborough, 69-76. Vermeule C (1987), The Cult Images of Imperial Rome, Rome. Vouga P (1923), La Téne; Monographie de la Station, Leipzig. Waddell J & J Conroy (1999), Celts and others: maritime contacts and linguistic change, in Blench R & M Spriggs (eds.), Archaeology and Language IV: Language Change and Cultural Transformation, London & New York, 125-137. Warde-Fowler W (1933), The Religious Experience of the Roman People, London. Webster G (1986a), What the Britons required from the gods, in Henig M & A King (eds.), Pagan Gods and Shrines of the Roman Empire, Oxford, 57-64. Webster G (1986b), The British Celts and Their Gods Under Rome, London. Webster J (1999), At the end of the world: Druidic and other revitalization movements in post-conquest Gaul and Britain, Britannia XXX, 1-20. Wedlake W J (1982), The Excavation of the Shrine of Apollo at Nettleton, Wiltshire 1956-1971, London. Wells P S (1995), Trade and exchange, in Green M (ed.), The Celtic World, London & New York, 230-243. Weisstein E W (2003), Synodic month, http://scienceworld.wolfram.com/astronomy/SynodicMonth.html White R & P Barker (1998), Wroxeter: Life and Death of a Roman City, Stroud, Gloucestershire. Wightman E M (1986), Pagan cults in the province of Belgica, ANRW II, 18(1), 542- 589. Wikipedia (2003), Pentagram, http://www.wikipedia.org/wiki/Pentagram Zeepvat B (1991), Roman Milton Keynes, Milton Keynes.

134 Zeilik M (1988), Keeping the sacred and planting calendar: archaeoastronomy in the Pueblo Southwest, in Aveni A F (ed.), World Archaeoastronomy, Cambridge, 143-166.

135