Читать книгу Stonehenge: Neolithic Man and the Cosmos - John North - Страница 12
ОглавлениеNeolithic Chamber Tombs—an Introduction
IN all of those ancient cultures from which written records survive, worship of the dead seems to have been bound up not only with religion, but with law and custom generally. The dead are typically considered as guardians, upholding order within family and tribe. Farming in particular encourages an appeal to the power of deceased ancestors, for there is no more conspicuous object of inheritance in need of defence than land. Farming communities not only depend on the landscape but help to redesign it. In northern Europe, for example, they created large enclosures, roughly circular in form, by throwing up banks of earth and stone taken from the surrounding area. In the course of doing this they usually created well-defined ditches which are likely to have had a purpose going beyond the mere supply of material for the bank. Some of the enclosures might have had a simple farming purpose, and others must have functioned as gathering points of some sort, since they show evidence of communal feasting. Evidence will be presented later that like so many of the smaller ditched enclosures that surrounded tombs and acted as focal points for religious ritual, the large enclosures also had a ritual function.
Tomb and landscape together preserve what is known about Neolithic and Bronze Age ceremonials of death. Building materials were usually determined by what was locally available, earthen tombs predominating where stone was in short supply, but existing simultaneously with dolmens elsewhere. (‘Dolmen’ is a Breton, Welsh and old Cornish word for ‘stone table’, and was originally used with an eye to the slabs of stone that had become visible when the covering material, usually earth, had disappeared. The word is widely used now for any stone monument, covered or not, containing a chamber created from upright stones capped with a roofing slab.) Some had an entrance passage, but almost all were originally covered with a cairn of stones or a mound of earth, or a combination of the two. If the dolmen was once a tomb, then it is usually called a chamber tomb, even where the chamber was only of timber. Where the soil acidity is low, skeletons will survive, and since undisturbed dolmens in such regions seem invariably to contain skeletons, equating dolmens generally with chamber tombs is not likely to be seriously misleading. Not all mounds were used as tombs, even when resembling others that clearly did, and even tombs served other purposes than burial alone. Several tombs have been found surrounded by thousands of shards of pots that had originally been placed on the tomb, a witness to some or other ritual of offerings continuing long after the primary interment. Similar collections of shards deposited over very long periods of time are found at many Neolithic sites throughout Europe.
There are two broad categories of chamber tomb: a passage grave is one in which it is easy to distinguish between the burial chamber and a passage leading into it, while a gallery grave is one where there is no clear distinction, but where the chamber-passage, usually slab-lined, is long—a substantial fraction of the length of the tomb as a whole.
Although there are large patches of territory without them, chamber tombs abound in a region of Europe and north Africa lying to the west of a line running roughly from south-central Sweden to Poland and across Europe to Tunis, taking in Sardinia and Corsica. There are also chamber tombs to the northeast of the Black Sea (the Caucasus), to its southwest (north of Istanbul), in Malta, the heel of Italy, and Palestine. Early in the twentieth century, surveys were made across Europe of modern human skull-formations, and it is a curious fact that the remains of chamber tombs generally are concentrated in much the same regions as those where the modern populations with longest skulls were then found to be living (dolichocephalic, those with highest cephalic index). The skulls those tombs yield up are similarly dolichocephalic.
Chamber tombs are found in a multitude of forms and sizes, and some date from at least as early as the fifth millennium BC. Tombs built with dry-stone walling—that is, using relatively small stones—seem to have given the lead to the builders of megalithic tombs, tombs with very large component stones. In the Iberian Peninsula and Brittany, and in Britain, the two techniques went on being used side by side, even under a single covering.
Many of the earliest of the chamber tombs had a long, tapered, trapezoidal form, almost always with the burial section lying across the broad end, which was also the higher end. (Various Wessex examples will be illustrated later in the present chapter. The word trapezoidal here indicates a four-sided figure with two sides parallel and two not so.) In Poland the taper may come almost to a point, while in Denmark and Britain the plan is often a long and narrow rectangle. The traditional English name used loosely for such tombs, without regard to their internal structure, is long barrow. It is possible that some rectangular tombs were of constant height rather than wedge-shaped, but until firm evidence is offered for this idea it is best ignored. In Britain—for example at Wayland’s Smithy, on the Ridgeway near the well-known White Horse at Uffington—as well as in northeastern Europe, there are long asymmetrical barrows that were quite obviously made deliberately so. That the lie of the barrows’ sides was not random, or a product of incompetence, will soon be evident from aspects of their internal structure.
While regional fashions do assert themselves, several properties of the tombs clearly rule out the idea that they were developed in the various centres entirely independently. The problem of the spread of tomb design across Europe is a difficult one. The homeland of farming and cattle-breeding seems to have been in Asia Minor, from whence it worked its way into the Mediterranean and Europe. The spread of European languages, however, followed different routes. The historical Indo-European languages are thought to have radiated from the Balkan–Carpathian region, around the fifth and fourth millennia BC. The spread of the Neolithic peoples responsible for the long tombs, one that took place over millennia rather than centuries, seems to have radiated from Anatolia or the Balkans. As an archaeological measure of cultural movement, pottery styles link the later English long barrows with the tombs of the Funnel Beaker Culture of Northern Europe, but long barrows had been built by earlier Neolithic groups, who had also come from the continent. There are those who believe that the (Belgian) Michelsberg culture is a part of the Funnel Beaker Culture, and that it was a group of Michelsberg people that crossed to Britain and influenced the earlier Neolithic Windmill Hill culture there, with its characteristically simple ‘baggy’ pottery.
Many centuries after the immigration that had brought funnel-shaped beakers, a new immigrant people introduced beakers of bell shape—the ‘Bell Beaker people’. They favoured round rather than long barrows. The Bell Beaker expansion was an important agent of cultural change in western Europe during the earlier Bronze Age, say from 2500 to 1800 BC. It is impossible to do justice to it in a brief space, but its movements were evidently largely sea-borne, with influences passing from the Low Countries down the Rhine, and to Britain and Brittany, thence to the Atlantic rim of France and Spain, as well as into the western Mediterranean.
The Funnel Beaker Culture—also known as ‘TRB’ from its German name, Trichterbecherkultur—covered at one period or another most of the area from the Low Countries in the west to central and southern Poland in the east, and from southern Scandinavia down to Bohemia and Moravia in the south. No doubt the great rivers of Europe, not forgetting the Elbe, helped with the dispersion of its influences. Most sea navigation would have been chiefly coastal, with crossings to Britain kept as short as possible, and presumably made in skin-covered boats capable of carrying at most two or three tonnes.
There are other traces of contact between the distant peoples of the Danube area and those of the north and west. The styles of fortification and enclosure they had in common are hard to overlook, if only because they are on such a massive scale. Use was made of rings of concentric ditches broken at intervals by what have been described as ‘causeways’—a potentially misleading name, in view of the fact that while some ‘causeways’ were several metres wide, others were only a few centimetres across. Causewayed enclosures are typically found on promontories, with the ground falling away—in Denmark often to water or bog—on two or three sides. Links between some of these structures and the long barrows will be proposed in Chapter 3.
Quite apart from similarities of form, and cultural traces of those responsible for them, it will soon emerge that there are very specific astronomical parallels between the long barrows of Northern Europe and Britain. While this reinforces the idea of a continental source for custom and ritual in Britain, it seems quite possible that a number of island practices developed independently, and then passed back to the continent. Megalithic tomb styles as such do not seem to be traceable to the cradle of the TRB culture, but their relatively rapid spread becomes a little less mysterious if we are prepared to suppose that there were established and continuing cultural links over great distances between different groups. If only by hearsay, these groups could have known much about each other—much more than we know about them, for instance. In short, the essential movements need not have been of whole peoples, but of ideas, carried in all probability by a few individuals. It has often been said that the outward forms of northern and western graves agreed far more closely than the grave goods in them, and the rituals to which they point. The astronomical principles they embody, simple as they might appear to us, must have represented a great mystery to many of the prehistoric community, so yet again the movement of a few experts might be of greater significance in the cultural transmission than the movement of whole peoples.
The densely populated long barrows of the earliest Neolithic Windmill Hill culture in Wessex have been intensively studied. There are less numerous groups in other regions, such as Sussex and Kent, the Hampshire Uplands, the Chilterns and East Anglia, the Lincolnshire Wolds, Yorkshire, Wales, Western Scotland, and Ireland. The British long barrows, like their Northern European counterparts, are usually a few tens of metres in length, although an unusual bank barrow running across the ditched camp at Maiden Castle in Dorset, and later surrounded by an Iron Age fort, is as long as 540 m. More than half of all British earthen long barrows are between 30 and 60 m long, and outside the Sussex area few groups of small barrows are without at least one long example. The Stonehenge district is well populated with earthen long barrows, some of the properties of which are summarized in Fig. 5.
Long barrows in Northern Europe frequently occur in tight clusters, and conventional wisdom is that this was not so in Britain—although this is largely a question of definition. Colin Renfrew has claimed, on the basis of the distribution of around 120 long barrows in Wessex, that the region was divided into five territories, each with its group of long barrows and a so-called ‘causewayed enclosure’. The prime example of such a site, but perhaps not the oldest, is at Windmill Hill itself.
FIG. 5. Earth-covered long barrows, all well within an hour’s walk of Stonehenge. The crosses mark the barrows, the directions of the lines represent the approximate directions of the barrows (with the cross always the lower end) and the lengths of the lines are proportional to the estimated lengths of the barrows (but they are not to the scale of the map). The average length is about 40 m. The circle marks the main surviving circle of stones at Stonehenge, to the same scale as the barrow-lines. The naming of barrows might seem odd, if it is not realized that they are conventionally numbered within parishes. There are more than fifty long barrows on Salisbury Plain as a whole.
The deep ditches that flanked the long barrows and provided the soil for the covering-mound were in some cases three metres or more deep, and occasionally ran in a U-shape round one end of the barrow. In some instances they formed a virtually complete perimeter. They were often shallower, say the height of a man, and they play an important part in our story, for it will be argued that significant observations of the sky could have been made by people standing at suitable points in these ditches.
In Britain, as well as on the continent, the grave mound often covered a wooden mortuary house. The forms of some of these will be discussed in the following sections. The evidence has often been taken to point to structures in the shape of a ridge-tent, with very heavy end-posts supporting the ridge-pole, on which inclined and close-packed rafters rested. This ‘tent’ interpretation has been offered for structures found in Denmark, Poland, and Germany, and in all of these centres examples have been found of burning in the timber structure. The first structure was in some cases first covered with a layer of stones (in Britain often flints), over which was laid turf. Those responsible for excavating them have remarked on the excellence of the verticals in the pits that held the posts—which might have been seven or eight metres high and must often have weighed as much as two or three tonnes apiece. Excavators of shafts on English sites—Normanton Down (Grinsell’s no. 330) is a good example—have commented similarly on the evident use of a plumb line in the digging of shafts and the dressing and setting of stones. These marks of technical competence are of some importance to our later argument that the posts helped to create a network of sight lines within the structure, for the very fact that the plan of a mortuary house at the heart of a long barrow lacked symmetry must then mean that this was deliberate. That the same plan can be found in more than one barrow serves only to strengthen the conclusion. And if our explanation of it is to be rejected, then another explanation must be found for what must have been a deliberate act.
Customs seem to have varied, but in many cases in Britain it appears that the dead were first exposed to the elements, and the bones only moved to the mortuary house at a later stage. (This is not a customary reading of the evidence in northeastern Europe, although barrows there may contain several interments.) In some cases at least, the covering mound of the earthen barrow proper was added only after the house had stood for many months, or even years. The sheer weight of the soil or stone or chalk rubble covering the mortuary house could then cause its collapse, as was the case at Fussell’s Lodge (Clarendon Park, barrow 4a), 12 km southeast of Stonehenge—where the house was an oak structure, about six metres long and a little over a metre wide. In this particular case the house contained the remains of more than fifty men, women and children. It was certainly well established before the end of the fifth millennium BC.
Apart from human remains, several long barrows have been found to contain the skulls of oxen as well as their hooves. This has been thought to hint at the hanging of hides (with horns and hooves still attached) as cult objects. Evidence for such a cult from the Bronze Age has been found in the form of wooden horns at the corners of a rectangular wooden temple-like structure at Bargeroosterveld, in Drenthe (the Netherlands). These are important pieces of evidence as to religious practice, and might be connected with astronomically guided ritual, the cult of the bull linking with the celestial bull, Taurus. What seems to be an example of this, requiring a new interpretation of the Uffington ‘White Horse’, will be given in Chapter 4.
Of the many hundreds of long barrows recorded in Britain, a large proportion seem to have been used to house multiple burials—typically five or six, but occasionally twenty or thirty or more. It has been argued that such long barrows indicate an egalitarian society. This presupposes what is not at all certain, that long barrows represent the norm rather than the exceptional means of burial; that they are for the family unit, and not shared, as religious centres might have been shared among a much more extensive group; and that bodies were placed in them close to the time of death. Variations in grandeur, complexity, and grave goods, seem to indicate variations in wealth and political power. A little light is thrown on this question by two facts taken in conjunction. First, there is the sheer magnitude of the enterprise of building a long barrow: this required, say, between 5,000 and 15,000 man-hours. Second, as time progressed, the number of burials in each long barrow seems to have fallen drastically, often to one or two. The great investment of time and labour has therefore been often seen as a symptom of a steady growth of political or religious hierarchies, the barrows having been set aside for the burial of highly favoured personages. The social function of the tombs might of course have been much the same, whatever the number of burials, large or small.
It does not follow that the religious meaning remained constant. In places as far apart as Ile Carn in Brittany and Corrimony in the northeast of Scotland, individuals were selected for burial alone in a vast tomb. On what social or religious grounds this favoured treatment was meted out is a difficult ethnographical problem, but it is one that should not be addressed without reference to barrows that were used for no burial at all. Horslip, Beckhampton Road, and South Street, for example, all in Wiltshire, have left no evidence that they were ever used for burial. If barrows generally are to be understood as adjuncts of a religion of the stars, capable of religious meaning even in the absence of burials, all arguments for the distribution of social power based on the number of burials will need to be reviewed. What if, for example, in the gradual reduction in the number of burials at certain places and times, we are seeing no more than a slow transition from a religion based mainly on ancestor-worship to one at a higher level of abstraction, based, for instance, on a more sophisticated and extensive mythology of the heavens? Selection for burial might in this case have been made on spiritual grounds, and might in principle have had no implications for a decline in the breadth of the power base. Spirituality might as well have resided in an epileptic, an innocent chosen by lot, a captive, a priest, or a prince. A barrow without burials could mean simply further progression along the same spiritual road. The whole question of a possible religion of the heavens is one that cannot be easily evaded.
There are well over five hundred known long barrows in continental Northern Europe, of which more than a hundred have been investigated professionally. Relatively few have been adequately charted for our present purposes. Rough statistical analyses have occasionally been made that divide the compass into eight or sixteen sectors and take counts of general orientations of the tombs. Across Europe, orientations have seemed to show a definite tendency to cluster around certain preferred directions, and the same tendency has often been remarked in Britain. Such coincidences tell us virtually nothing of value, however, if the aim was everywhere to be astronomically precise, since the most probable astronomical orientations—whether involving Sun, Moon, or star—depend heavily on factors that are ignored in such simple accounts, notably geographical latitude, irregularities in the height of the local horizon, and even the form of the monument itself. Poor statistics make a poor guide. The only acceptable approach is through a comprehensive set of detailed analyses of individual cases—a programme far beyond the scope of this book. A few important examples will be touched on in the remaining parts of this chapter: they conform closely to a few simple principles, and it would be very surprising to find that others failed to do likewise. But first to a related theme, a favourite object of archaeological comment, namely the orientation of the body within the tomb.
There seem to have been many local variations in burial rites. In the Wessex long barrows the bones were generally placed in the tomb in a disarticulated state, sorted by type, presumably following exposure of the body elsewhere. Where a skeleton in a chamber tomb is found in a recognizable direction, however, it is often with the head to the west or to the east–occasionally both are found in the same tomb. Opinions differ as to whether these variations indicate differences of age or sex. In a study of the positions of very many skeletons in tombs in Bohemia and Moravia, W. Schlosser and J. Cierny found some interesting patterns. It seems that the culture responsible for linear pottery (in the centuries around 5000 BC) had a very strong preference for placing the head roughly to the east; that among the ‘corded ware’ peoples (around 2200 BC), west and east were favoured in roughly similar numbers, with a slight preference for west; but that the Bell Beaker peoples favoured placing the head to the north or south (in the ratio now of about two to one), with a small subgroup roughly northeast or southwest, and a few roughly southeast. Three thousand years later, many Germanic groups were still putting the feet of the corpse to the north. In an English Saxon cemetery at Fairford, for example, this habit was found to be almost universal. J. Grimm, writing on German mythology in 1854, drew attention to something that appeared to point back from the Middle Ages to an ancient pagan tradition of worshipping towards the north. Thus in some stories of Reinard the Fox, the wolf did just that, while the fox adopted the Christian convention, facing east.
It has to be said that in the Bell Beaker class of burials the bodies were no longer laid out full length, but in a contracted posture, with the knees drawn up in the manner of an embryo, and that for this reason, a precise direction of the body simply cannot be specified. Nevertheless, some rough statistics from the Wessex round barrows belonging to the Bell Beaker culture fit well with the continental evidence. In a count of 55 instances made by William Long in 1876, on the basis of previous reports, in 35 cases (in other words, more than six out of every ten) the head was said to have been to the north. Six had heads placed to the northeast, and five to the east. All eight points of the compass with the exception of south were mentioned in one report or another, but the three named were the most common. (West and southeast accounted for three each, northwest for two and southwest for one.)
Rough and ready as these data are, it does seem that customs in the orientation of the body of the deceased are almost as characteristic as pottery types, although possibly not as subtle. They were obviously ‘astronomical’ in some degree, but how are they to be interpreted? Some of the problems to be expected have already been mentioned. Death has some plain analogues in the heavens. It seems natural to represent death by the western half of the horizon, the place of the dying Sun for peoples in the northern hemisphere, and this is borne out by many later religious rituals known from direct textual evidence or report. (The eastern half of the horizon tends to represent life or birth.) In those Neolithic practices in which a near north–south line was preferred, the body might have been lain with head turned to face an appropriate easterly or westerly direction. Why east or west? Was it the general phenomenon of rising and setting of heavenly bodies that counted? (It was obviously not a precise arrangement, if the Sun was involved at any other season than the equinoxes.) It will shortly be noted in connection with tomb architecture that an approximate north–south line was associated with the behaviour of the star Deneb, in which case that more obvious explanation counts for nothing. Granting for the sake of argument that this was so, how should one interpret the intentions of those who laid out a body east–west? Might the relatives of the deceased have looked across the body, perhaps, towards Deneb? Or did they look over the length of the body towards Aldebaran, the brightest star in the constellation of Taurus, which happened to rise near east and set near west? There is evidence from some much later societies that the feet are placed towards the land of the dead, but this of course is often related to the daily passage of the Sun.
Is the argument helped out by the orientation of the tomb itself, as for example in certain passage graves that restrict a line of sight? At Gavr’inis in Brittany, for instance, and Newgrange in Ireland, where there are chamber tombs dated to within a century or two of 3500 BC, the main corridors were clearly directed towards midwinter sunrise, in the sense that they opened towards it and were illuminated by it. (There might also be lunar directions implicit in both.) And if we settle on a solar ceremony at the laying out of the body, there remains the question of the timing of events. Was the body preserved until the Sun was in the right position, say setting at one of the solstices? Or was the Sun’s direction preserved, marked out as something appropriately sacred, say at the time of one of the solstices, and used for burial even if the Sun was not rising or setting in that direction at the time of the burial? Another possibility is that in some cases there was human sacrifice, a death that was contrived rather than natural, at precisely the appropriate season.
There is some evidence from modern anthropology as to comparable burial customs, but it is not coherent. All told, it is virtually impossible to place a reliable ‘astronomical’ interpretation on even comprehensive statistics as to the directions of burials, if they are merely to the four cardinal points of the compass. Even so, there is a conclusion to be drawn that is far from trivial. Whether or not the ritual interpretation can be discovered, there is no doubt that burial was in these cases an activity governed by rules relating to celestial phenomena, for only by reference to the heavens can the cardinal points be found. To progress further it is necessary to look to the tombs in which the bodies were placed.
Long barrows were built in Britain from perhaps the middle of the fifth millennium BC to the early third, and as time progressed, several changes of style and custom took place. To hint at a change in custom one must first be quite sure that there was a consciousness of the need to act in a particular way. That there were architectural rules of some sort is strongly suggested by the fact that in a large proportion of the examples surveyed, the higher and wider end of trapezoidal structures heads (roughly) east. This might not be thought a very profound principle, but it was true of the long mounds of the Linear Pottery culture in Northern Europe, and it is reminiscent of the characteristic habit of that same culture in Bohemia and Moravia, where the body was so often placed with the head to the east. This encourages a search for more precise rules.
Those who have recorded the directions of unexcavated long barrows have usually been content with rough approximations, referring, say, to only eight or sixteen points of the compass. The original structures have almost always been largely destroyed, modified by rabbits and badgers, or ploughed out of existence—usually by the farmer, but occasionally in the course of military exercises with tanks. (Unfortunately here, the better the barrow the greater the challenge.) The line of the major axis of a barrow is often as much as we can expect, and this seldom to better than a degree or two, unless we are convinced that it was deliberately aimed at some distant point. One should not be satisfied with the axis alone, for the principal lines of sight will be shown to have often differed from that line.
A direction without reference to the location is worthless. The midwinter Sun at Stonehenge rises about 40° south of east; at Newgrange, the angle is nearly 44°. Such directions vary with geographical latitude and the altitude of the horizon, but they also vary with the century. All such data are needed. The first two items of information can usually be taken from the Ordnance Survey. For structural details, an archaeological survey is needed. It is worth bearing in mind that whatever the quality of the original, a third-hand copy might be in error by as much as 20°. One much-copied plan—by one of archaeology’s most expert draftsmen—displaces the West Kennet long barrow by 15°.
An approximate date for construction calls for independent archaeological evidence. The orientation of a monument might provide a date, but only in conjunction with an opinion as to precisely what the prehistoric practices of orientation were. The vicissitudes of star positions are at the root of one method of astronomical dating. What are often called the ‘fixed’ stars are anything but fixed. They move among themselves by very small amounts (their so-called ‘proper motions’), and these can often be ignored, although it is dangerous to do so when very long time-spans are concerned. Less subtle is another slow change in the apparent positions of all of them, making them seem to drift very slowly round what can be thought of as the celestial sphere (the sphere on which they seem to be carried round the sky with the daily motion). The movement of the stars round the sphere follows a path parallel to the Sun’s yearly path, but it takes them roughly seventy years to cover a single degree. (The Sun’s yearly path through the background of stars is called the ‘ecliptic’.) The change may be slow, but in spans of time of the order of four or five thousand years, the shift in star and constellation positions, amounting to about a sixth of the way round the sky, is highly significant.
The drift in question is actually a consequence of the instability of the Earth, which in relation to the Sun provides the frame of reference against which star positions are measured. Just as the longitudes of places on the Earth are measured from the Greenwich meridian, so star longitudes are measured from a point on the celestial equator (a great circle on the celestial sphere directly above the Earth’s equator). By convention, this is the point where the Sun, moving along the ecliptic, crosses the equator in spring, one of the two ‘equinoctial points’. What was earlier described as a drift in star positions, amounting to a slow increase in their ‘celestial longitudes’, is really a drift of the zero point from which their positions are measured. The drift is for this reason now called the ‘precession of the equinoxes’.
The increasing longitude of a star changes its apparent (angular) distance from the north pole of the sky, and in turn affects the point on the horizon over which it rises and sets. To take just one example: the Pleiades rose 10° south of east at Stonehenge in the early forty-second century BC, but twelve centuries later they rose due east.
This brings us back to a serious problem, hinted at previously. How are we to know whether an orientation indicates a concern with the Sun rather than with a star? east–west directions in long barrows, for example, are relatively uncommon but not unknown. Like alignments that seem to be to the rising and setting Sun at the solstices (midsummer and midwinter), they might well have been directed to a particular star. The change in the direction of sunrise and sunset at the solstices is very slow—say a fifth of a degree of the horizon in a thousand years. Over the same period of time, the point of rising of the Pleiades would have moved by twelve degrees. Continuity of custom is less obvious with the stars than when the Sun was involved. Prehistoric peoples might have aligned all their barrows on the rising or setting of a particular star or group of stars and have done so with great deliberation and precision at different periods of history, and yet have left us with a scattered set of compass directions, in short, with an impression of carelessness and imprecision. Fortunately there is a way of cutting through this problem, for it will soon emerge that the custom was to align long barrows on two or more stars simultaneously, and to consider mostly very bright stars. This greatly reduces the number of ways of interpreting individual cases, and from a study of well excavated examples a coherent picture gradually emerges, as clear in its way as that of the later monuments with their more stable solar orientations.
In brief, it turns out that the Wessex long barrows were mostly stellar, while the later circular monuments were solar and lunar. The later long barrows already show signs of change. There are two different criteria here—alignment and illumination. Matters were sometimes so arranged that at one of the Sun’s extremes it could illuminate the end of a long gallery in the tomb, through a suitable entrance slot. Even this will prove to have followed a similar arrangement with stars, from an earlier period.
FIG. 6. Four different examples of taper in long barrows, one from Poland, one from northern Germany, one from Lincolnshire, and one from Wessex. All have left interesting traces of mortuary houses (mh), the sides of all of which are approximately parallel to the sides of barrows erected later on the same site, or perpendicular to them. Note the letters (a, b) marking these highly significant properties. The scales and orientations of the different barrows are approximately correct, with Wayland’s Smithy, the smallest, about 55 m long. The outlines of barrow/mortuary house were formed by either timber posts (t) or stones (s). From left to right, the materials were: s/t, s/s, t/t, s/s. A mortuary house at Wayland’s Smithy, hinted at in the figure but redrawn in Fig. 19 was of timber.
First Thoughts on the Taper of Long Barrows
While some of the longer and earlier examples of earthen long barrows were parallel-sided, or nearly so, and just possibly of constant height over their length, most were tapered in height and width. (An idea of the plans of some of the different styles may be had from Fig. 6. The rationale of their three-dimensional forms is the subject of this chapter, and the differences are too subtle to be readily illustrated at this stage, but Fig. 53 should give an idea of what taper in height and breadth entails.) It is doubtful whether the property of taper originally had any astronomical purpose, since it is found in many early post-built houses throughout Europe, but the precision with which those houses were built—for instance in the alignment of their side walls—is usually greatly inferior to that of most large tombs. Precision might have signified respect, a wish to give something perfect to the dead, but there are good reasons for thinking that it served also to direct the eye towards significant risings or settings over the horizon. Even supposing that the trapezoidal form was first adopted for the housing of the dead by analogy with the housing of the living, support for this other explanation is strong. It is buttressed by the fact that time and again, on an astronomical reading, different sites seem to indicate an allegiance to a relatively small number of select bright stars.
Accepting the astronomical idea provisionally, two or three potential explanations of the taper in plan offer themselves. (Taper in height will turn out to be another question.) The first of these is easily appreciated in terms of what seems to have been a strong desire to find lines that point to the rising of an important star in one direction and the setting of another in precisely the opposite direction. This is easier said than done, but by taking two lines at a fairly small angle, it can always be done. Using those two lines as the orientations of the sides of a barrow, the assumption that stars were viewed along the barrow’s sides provides one possible explanation of taper.
It is not difficult to imagine others. As illustrated in Fig. 1 (Chapter 1), the direction of the rising midsummer Sun at an arbitrary place on the Earth’s surface is usually near to, but not precisely the same as, the reversed direction of the setting midwinter Sun. Since the (angular) height of the horizon helps to determine the directions of rising or setting over it, if there were hills of different heights in the two directions, the angle between the lines might be several degrees. This difference might have been built into a long barrow in various ways. The broad (interment) end of the barrow might have been made to point to midsummer sunrise, and another part, say a side or axis at the other end, to midwinter sunset. (There is an instance fitting this description in the Winterbourne Stoke group in Wiltshire.) On level parts of Salisbury Plain the difference between the ideal directions would have been only about 3°, so that there an edifice with only a 3° taper could have served both directions at once. It would be wrong to place much emphasis on a solar explanation at this stage, however. It is certainly not relevant to the early history of the taper of long barrows, the alignments of which will prove to have been primarily stellar.
At first sight it might seem that a single building alignment—say that of a central ridge along it—cannot cater for two opposed solar extremes—for both midsummer rising and midwinter setting, for example. This is where taper in height might have come into play. Something like the technique to be explained in broad outline here might have been applied at the Skendleby 2 (Lincolnshire) long barrow and more certainly at Stonehenge itself. The directions in which the first glint of the rising Sun and the last of the setting Sun are seen depend appreciably on the altitude of the horizon. By directing a tomb to horizons of different altitudes, or alternatively, by creating an artificially elevated horizon out of the tomb itself, the directions of rising and setting can be adjusted, and under certain circumstances brought into exact opposition. (The Skendleby case is illustrated in Fig. 40 in a later part of this chapter.)
This last possibility puts paid to a claim that has been made on many occasions, that the minor variations in the directions of chambered tombs and long barrows show that their builders were very casual about directions, and were happy enough if they could get them roughly right. Our hypothesis is capable of explaining why the chamber-tombs in an otherwise apparently coherent group are almost, but not at all precisely, parallel—say with a scatter of 4 or 5°. The differences in direction, which might in some cases have been the results of variations in the height of the distant land-horizon, might in others be the result of taper in height, that is, of differences in the slopes of monuments that were themselves used to set artificial horizons close at hand.
Apart from the rectangular and trapezoidal forms of long barrow, which are found often in an area stretching from Jutland (Denmark) to Western Pomerania, there is the near-triangular form already mentioned, not unlike the trapezoidal: it is a long isosceles triangle with very slightly concave sides, higher at the wide end. This is the form at the site of one of the best collections of well-excavated long barrows in Europe, the Sarnowo group in Kujavia (Poland). There are nine in all, ranging in length from 30 m to 83 m, and judging by a radiocarbon measurement, their dates are probably all within a few centuries of 4500 BC. They are all on spurs of land, with the wide ends always in the northeast quarter, and facing down gently sloping ground.
The Sarnowo barrows are in a parlous state, and it is at present difficult to assess more than their general orientations. (Even this is a dangerous undertaking, since two modern surveys display compass directions which may differ by as much as 14° on a single monument.) Nevertheless, there seems to be much the same pattern in the directions of the barrows as is to be found in a selection of four or five of their English counterparts. Three are conceivably aligned on the following stars: Aldebaran or the Hyades (nos. 2, 4, and 5); the Pleiades (no. 6); two perhaps simultaneously on Bellatrix and Regulus (nos. 1 and 9); one simultaneously on Deneb and Rigel (no. 7); and one on Sirius (no. 3). Only one (no. 8) seems to be aligned on the Sun—and even there it is possible that Orion’s belt is somehow indicated. All of these are possibilities within the approximate period 4500 to 4100 BC. The uncertainties are chiefly inherent in the disordered state of the remains, and archaeological plans that cannot provide exact orientations. Perhaps at some future date the Sarnowo stellar orientations will be studied more thoroughly. Perhaps other stars were observed across the mounds, as at the Wessex barrows. Whether or not the long barrows reflect back on earlier European practice, they are vital to an understanding of later stone and timber circles relating to Stonehenge. They are also intimately related to the Avenue that leads up to Stonehenge and to the strip known as the Cursus, to the north of both.
The Orientation of Long Barrows
There are more than sixty long barrows on Salisbury Plain, and many hundreds elsewhere in Britain and Ireland. They tend to be on high ground with open views of the surrounding country in two or three directions. In Britain, Ireland, and on the continent, there was a mild preference for a near east–west orientation, a fact already the subject of comment as long ago as William Stukeley’s Abury, a Temple of the Druids (1743). While crude statistical surveys of the orientations of large numbers of chambered tombs will obliterate most of their essential differences, there are a few simple rules of this sort that are easily appreciated, if not easily interpreted. As a purely descriptive measure, the key direction may be regarded as that from the narrow end to the broad. (One might have said ‘looking out from the entrance’ were it not that many barrows had blind entrances.)
Taking the chief regions of Britain and Ireland with remains of such tombs, there are definite tendencies that are more or less duplicated in places of similar geographical latitude. Cairns in Orkney are like those in Shetland, in that in both places most cairns tend to look southeast. Those who built the earlier passage graves of Brittany had also favoured this direction. Coming down to the northeast of the Scottish mainland, the tombs of Caithness resemble those of Ross and Cromarty, in that most look very roughly east. The nearby Clava cairns, on the other hand, look southwest, a quality they share with tombs of an outwardly different pattern in Ireland, the ‘wedge tombs’ that are especially numerous in the southwest of the country. (This last name derives from the fact that the burial chamber is trapezoidal in shape, although it is usually given a round cairn as cover, so hiding a certain similarity with the overall wedge shape of many English long barrows.) Surviving megalithic tombs in Ireland number about twelve hundred, and more than a quarter of these, mostly in the northern third of the country, have a small court at the entrance. Such ‘court tombs’, especially common in the Atlantic coastal regions of Mayo, Sligo and Donegal, often look northeast; and in this, if not in their shape, they resemble the long barrows of the Clyde region, across the water in Scotland. The passage graves of the Boyne valley on the eastern coast of Ireland tend to look east and southeast, as do the long barrows of the Cotswold–Severn area in England.
Such broad rules as these are subject to many exceptions, but they confirm the principle that custom changed appreciably with place and time. This is hardly surprising, bearing in mind that the time-span of the groups mentioned here covers more than two millennia. Contrary to a common belief, these tendencies tell us absolutely nothing about orientation towards the Sun. After studying several individual examples more closely it will become clear that interpreting rough proclivities of the sort outlined above is a dangerous business. Even without doing so, the dangers of summarizing motives of the architects of the long barrows too hastily can be seen from the fact that, in England as a whole, there are at least ninety long barrows along a roughly north–south line; and yet in no single region was the number great enough to have merited mention in the brief list of tendencies just given.
FIG. 7. A general plan of the Fussell’s Lodge long barrow, drawn by P. J. Ashbee.
Many chambered tombs from the fifth millennium BC still survive in eastern France. Relatively few barrows in Ireland and Britain have been assigned such an early date, but among them there is the mortuary house from the Fussell’s Lodge tomb—which radiocarbon dating puts in the calendar (corrected) range 4250–3950 BC. This barrow, roughly 12 km from Stonehenge and 5 km from Salisbury, has been well excavated (1957) and described by Paul Ashbee. When it was first erected, Salisbury Plain was heavily wooded and is unlikely to have had more than a few dozen families living on it. The tomb was an earthen long barrow, raised after a period of time over a mortuary house of still greater interest. This first structure, of wood, was at an early stage apparently covered with turf, and then with crushed chalk. Later a cairn of flints covered the burial area. This cairn stage might have followed the collapse of the timbers of the mortuary house, or might indeed have precipitated the collapse—in either case probably within a few decades of the foundation structure. The flint cairn had wings anticipating the later façade (see Fig. 7). The mortuary house was used as a place for the systematic deposition of the bones of the dead—skulls in one place and long bones in another. Many bones were missing altogether. The bodies had presumably been exposed to the elements, or buried and disinterred after decomposition. Despite the small population, building the mortuary house—barrow apart—would have required a high degree of social organization. It was bounded at east and west by two colossal oak trunks or split trunks, each—depending on length—weighing in the region of 2 to 10 tonnes, assuming a green state. When the time came to erect an earthen long barrow on the site, its mound was enclosed within a retaining wall of nearly two hundred timber posts, a palisade bedded in a deep trench. The old mortuary house was at the eastern end of this enclosure, the whole lying roughly east—northeast. Into this frame were packed many hundreds of tonnes of chalk that had come from the ditches flanking its long sides, chalk that had been dug out with antler picks. Ashbee estimated that if men worked ten hours a day at the task of infilling, it would have taken no less than 487 man-days. Whatever the precise dimensions of the mound—and here a somewhat smaller mound than his will be preferred—it was certainly no light undertaking.
FIG. 8. The letters B show where the collections of bones were found. The central pit might originally have been used for a central support to a massive beam—or double beam—spanning the two main uprights. Note that the palisade trench interferes with the socket for one of these, showing that it is a later and independent structure.
The planning of the barrow must have been no less taxing, for there can be little doubt that it was done with reference to the peculiarities of the landscape in regard to the risings and settings of certain bright stars. Even more interesting is the case of the original mortuary house. At a cursory glance the posts bounding this are badly skewed, but it will be seen to have pointed within two or three degrees of 22° north of east. Since 20° is the approximate mean direction of the splayed sides of the later barrow, this is not very surprising.
To progress any further, we must make reference to certain striking resemblances between its plan and that of the first mortuary house on the site of the barrow at Wayland’s Smithy. As will be described shortly, the Wayland’s barrow had split trunks with the flat of the D-shaped section to the inside, and the trunks were of diameter 1.2 m. The pits were of much the same dimensions as at Fussell’s Lodge. The areas on which the bones were placed were at both sites about a metre across, although the length of the area at Fussell’s Lodge was greater. Lines of sight from the posts of what has been interpreted as a porch would have skimmed the two sides of the western trunk at Fussell’s Lodge had it also been about 1.2 m across, and the other is presumed to have been of much the same size. The stacked bones were found to lie in a neat line not very different from 22° to the east–west line. This can reasonably be taken as a first approximation to the direction of the axis.
Now it so happens that another long barrow, today virtually obliterated, was on the skyline at a distance of 1.85 km and in a direction just 21.7° south of west, as seen from the Fussell’s Lodge barrow. The sighting of one barrow from another will later prove to be so common that this line must be treated as important, and doubly so when we discover that it indicates the setting of the star Aldebaran over the neighbouring barrow at about the right period of prehistory. (As for the precise date, some of the deciding factors must first be explained.) Working from this assumption, we are no longer, therefore, entirely dependent on the edifice itself in establishing at least this property of its orientation. The observed altitude of the barrow (not the star) would have been about 1.32°, depending on the height and position of the observer. This raises an important point: the quoted altitude is based on the assumption that the observer is standing at ground level. A study of the relatively extensive and important collection of human remains from the Fussell’s Lodge barrow, by D. R. Brothwell and M. L. Blake, led to the conclusion that the average adult male stood 170 cm (5 ft 7 in) and the average adult female 157.5 cm (5 ft 2 in).
At a given date, any star has a particular position on the celestial sphere, that is, can be assigned a given set of sky coordinates comparable to latitude and longitude on the Earth’s surface. There are two particularly useful reference planes comparable to that of the Earth’s equator: one is the celestial equator above it, and the other is the ecliptic (see Chapter 1 for a simple account, and Appendix 2 for more detail). Taking the equatorial system for the time being, declination is the coordinate corresponding to terrestrial latitude. If its value is known, then using methods outlined in Appendix 2 one can say exactly where on the horizon the star would have been seen rising and setting at any date in the past from any given place—and conversely, what was the date at which it rose or set in such and such a position. A star of declination in the neighbourhood of –12.67° can be said categorically to have been seen setting over the now defunct barrow, but only if the star was visible down to the horizon. Aldebaran had this declination in 4365 BC, but for reasons to be explained, the star would not have been seen as low as the true (natural) horizon.
Whenever a distant natural horizon enters into the calculation it is necessary to take possible tree cover into account, since a star descending into even distant trees leaves one unable to decide precisely when it disappears. In the present example the horizon was topped by a barrow, however, so that tree cover can there be ignored. (Trees seen against the sky do not necessarily modify the effective horizon altitude when the Sun and Moon are being observed. If the trees are not in leaf, as at winter solstice and spring equinox, and the horizon is defined by a fairly sharp ridge, the Sun and Moon are often easily visible through the trees down to ground level.)
In the reverse direction of the axis at Fussell’s Lodge, 21.7° north of east, the ground above and below the tomb slopes at about 3° to the horizontal, this being effectively also the altitude of the eastern horizon. The ground is near at hand, and was no doubt cleared of trees. It is no accident that the slope of the final earthen barrow, according to a plausible reconstruction, was also of about 3°. One may see relatively faint stars at such an altitude. Over the barrow to the west, however, where the horizon altitude is lower, few stars are bright enough to be visible at the horizon proper. This difference needs to be explained.
How low in the sky one may detect a star depends on weather conditions, of course, but even under the most favourable conditions there is a limit to what is possible that depends on the star’s brightness. Astronomers classify the luminosity of stars by their so-called magnitude. The brighter the star, the smaller the magnitude. Sirius, the brightest star in the sky, has now a magnitude of –1.46. Only one other star then visible from Britain, Arcturus, has (and had) a markedly negative visual magnitude. Our concern here will be with only the brightest stars, of which about 45 have magnitudes less than 2, although not all of those are visible from northern latitudes.
Typical minimum altitudes for viewing the very brightest stars under favourable viewing conditions from the downs of southern England, are these: Aldebaran 2.00°, alpha Crucis 2.42°, Altair 2.01°, Antares 2.32°, Arcturus 1.43°, beta Centauri 1.86°, beta Crucis 2.35°, Betelgeuse 2.06°, Capella 1.48°, Deneb 2.35°, Pollux 2.24°, Procyon 1.71°, Regulus 2.44°, Rigel 1.53°, Rigel Centauri 1.58°, Sirius 0.74°, Spica 2.12°, Vega 1.48°. The cluster of stars known as the Pleiades is a difficult case, but it was probably rarely seen below 4°, with 4.4° more typical.
These ‘extinction angles’ are for specific viewing conditions, and the accuracy with which they are quoted here is rather excessive, although useful for purposes of comparison. Actual values are as uncertain as the weather itself. Climatic conditions were not identical to those of modern times, and even local air currents may affect the issue. Fortunately it will soon be discovered that in very many cases precise values are unimportant—that is, when the height of the actual (natural or artificial) horizon is appreciably greater than the extinction angle.
To return to Fussell’s Lodge. Since Aldebaran might have been seen only down to about 2°, it is necessary to review the earlier statement of what was seen over the neighbouring barrow. The declination, recalculated for this slightly higher altitude, turns out to be –12.07° and the corresponding date 4245 BC. The shift of more than a century from such a seemingly insignificant change of altitude draws attention to the importance of working with precise data.
If we are to believe that the star Aldebaran was observed setting along the axis of the mortuary house, how would it have been sighted? It is wise to consider the possibility that a distinction has to be drawn between an initial act of design and later acts of ritual observation. At the foundation of the mortuary house, however, the alignment would have been the same on both occasions: just as the designer had done, so others would have been able to look in the direction defined by the two massive vertical posts, each presumed to have been a split trunk (by analogy with Wayland’s Smithy, where the post holes are also ovals). The fact that the large posts define a line to Aldebaran’s setting raises an interesting possibility. The skull of a domestic ox was found by the eastern post, and it seems likely that this star—which in historical times was regarded as the eye of the bull—had already acquired an association with the bull, although not necessarily quite the same one.
The alignment on the setting Aldebaran will be acceptable only if the date is confirmed by independent evidence. To decide on what might have been seen in the eastern direction, at the altitude set by the natural horizon (very close to 3.0°), one must again decide on a precise azimuth. There is of course no barrow in this case, and simply reversing the line to the first barrow yields nothing of real interest, but from the very similar structure at Wayland’s Smithy there is reason to think that the line grazing two of the posts of what has usually been regarded as an ‘entrance porch’ will be significant.
For reasons to be explained shortly in connection with that other monument, the posts will be referred to as ‘scaling posts’. Taking the four relevant post holes at Fussell’s Lodge (see Fig. 8), and making estimates of post diameters on the basis of later evidence and of the lines of bones, the azimuth appears to be about 61.8° from north (with an error unlikely to be much more than half a degree). The star Spica rose at an altitude of 3° in this direction when its declination was 19.51°, a declination it had in approximately 4250 BC. The agreement in date is remarkable, but one should not make too much of it, in view of the various uncertainties.
The scaling posts fairly certainly served a similar function, defining lines to the rising and setting of other stars, both over natural ground and over certain artificial horizons. For the time being only the first sort of alignment will be considered. Unfortunately, even here there are two possible arrangements. It will be found repeatedly hereafter that viewing was apparently at right-angles to certain lines in the barrows and their associated ditches. Briefly, the question is whether, in looking across such a line, the line is defined by a part of the barrow’s structure on the near or the far side of the barrow from the point of view of the standing observer—or in this case, on the near or far side of the quadrangle of scaling posts. This often makes a difference of a few centuries to a derived date, although occasionally it rules out a solution completely. For some barrows, making the right choice will prove to be of fundamental importance, but for the moment all plausible results will be quoted.
As shown on Fig. 8, there are two more alignments to be found yielding dates in rough agreement with the others. With estimated dates in parentheses they are: to the rising of beta Crucis (4220 BC) and to the rising of beta Centauri (4220 BC). (The posts nearer to the monuments are very close to aligning on the setting of Deneb to the north and the rising of Rigel to the south, but if this was the intention, the line does not seem to have been as accurately engineered as the others.) It is here assumed that the observer was standing on level ground.
A few brief remarks are in order here, in view of the coherence of these results. The posts were only two metres apart, and various assumptions have been made as to their sizes and positions in their sockets. The line to the rising of Rigel (grazing the sides of the two posts in a way to be discussed in connection with Wayland’s Smithy) was over a nearby natural horizon of about 4.6°, assuming no tree cover. Whether or not there was forest cover, the Deneb alignment is poor, but there is much evidence from elsewhere of the enormous importance of this star. It is assumed that nearby trees were removed in Deneb’s direction. (The hill over which Deneb set was destined to become the site of a sizeable Iron Age monument, Figsbury Ring.) Dropping Deneb from the process of averaging dates from the six stars, an astronomical date can be quoted for the mortuary house of 4235 BC (rounding to fives). The probable error here is anything up to a century either way. With Aldebaran (4250 BC) more confidence is in order, since the neighbouring barrow is a far more accurate marker than nearby posts; and yet the Aldebaran date might just possibly relate to activity on the same site before the mortuary house was built. The radiocarbon range for charcoal from one of the D-post pits, perhaps relating only to a clearing operation, was 3230–3150 bc, which when corrected produces a range of calendar years from 4250 to 3950 BC. The agreement is not perfect, but respectable.
There are one or two conclusions to be drawn from these findings. The first concerns the purpose of the various structures. When the mortuary house was covered with a cairn, that might have obliterated some of its earlier sighting-line properties, although trunks would have had a longer lifetime than smaller timber. Some of its astronomical properties, however, as yet to be described, almost certainly remained to it. These probably—but not necessarily—involved observation across the cairn from short ditches, small versions at the eastern ends of the later (much enlarged and deepened) ditches that flanked the great barrow.
FIG. 9. The Fussell’s Lodge long barrow, viewed from the north. An outline of its probable appearance at the time of its foundation. The shading is added to suggest the lie of the walls and roof.
The mortuary house was no doubt a self-sufficient edifice with a wider spiritual function than this name for it might suggest. The argument that it was a religious focus is supported by what can be said of the visibility of the various stars. There are stars that are so close to the north pole of the sky as never to rise or set. Whether or not they are visible depends only on cloud cover and whether the Sun is up or not. The visibility of stars like Aldebaran, Spica and beta Crucis that do rise and set is obviously limited twice over, restricted as it is by the horizon and the need for the Sun to be below the horizon. It emerges that there was no season at which all the phenomena of rising and setting mentioned in conjunction with the Fussell’s Lodge mortuary house could have been seen. The conclusion is that it was erected, and its site carefully chosen, on the basis of knowledge about risings and settings acquired over a reasonably long period of time, indeed with long-term meaning, and recorded in wood or stone in relation to the terrain. This is a small point, but it makes it unlikely that any astronomical ritual enacted at Fussell’s Lodge was designed around celestial events appropriate to only one season—that of a founding burial, for example. Observation of the stars at such sites must have been a fundamental rather than a purely ephemeral affair.1
It is not yet possible to offer a convincing argument for a conclusion that will be reached at a later stage, to the effect that the scaling posts had an architectural function in establishing directions that would be needed in the later building. For this the more complete evidence of Wayland’s Smithy is called for. It is conceivable that those posts also supported a platform on which the dead were exposed, before the larger bones were finally brought into the mortuary house; but on the whole this seems unlikely, for the platform would neither have been very substantial nor well fitted to the purpose.
One can only guess at many of the finer details of the mortuary house, such as the height of the massive uprights, and whether they bore another split trunk of comparable girth as a lintel between them. This might have made for a less leaky house than otherwise, but the sheer weight and rarity of such a splendid trunk makes the idea improbable, as does the fact of slightly different depths at the two ends—suggesting posts of somewhat different heights. Fig. 9 is offered here as a mere outline of the arrangement at the time of foundation, and in it no attempt is made to indicate supporting timbers, apart from a short intermediate trunk that could have been interposed to take most of the weight of the roof. (It is placed in Ashbee’s ‘pit B’, which is 60 cm deep and 50 cm across. The fill was found to contain burnt bones, and the area was covered with bones, but not from the earliest date.) Access to the interior was probably from the eastern side of the southern D-post. The form suggested for the roof will be justified in a later section.
FIG. 10. A general plan of the long barrow at Wayland’s Smithy (after R. J. C. Atkinson). This shows both phases of the tomb, and later field ditches superimposed on them.
The chambered long barrow at Wayland’s Smithy was extensively restored in the 1960s. It is trapezoidal, and despite its asymmetry shows great regularity in its geometrical design (see Fig. 10). An earlier mortuary house found during its excavation closely resembled in overall style the mortuary house at Fussell’s Lodge, and the two, taken together, provide valuable information as to astronomical practice. They offer a challenge, too, because although they are so similar in form, their orientations are radically different.
It is hard to think of a more famous long barrow than that at Wayland’s Smithy. Here lived an invisible smith, who would shoe travellers’ horses at a groat apiece: the owner would leave horse and coin at the spot, and return to find the horse shod. Francis Wise printed the traditional story in 1738, and it has been retold often since then, perhaps the best known versions being those in Walter Scott’s Kenilworth and Tom Brown’s Schooldays by Thomas Hughes, who was born nearby. The name of Wayland, always a craftsman of some sort, is not uncommon in Germanic mythology. The oldest literary reference to him is in the Anglo-Saxon poem Deor, but there is a reference to Wayland’s Smithy in a charter of king Eadred, dated 955, while the Franks Casket in the British Museum, from the early eighth century, depicts the same character. In later Berkshire stories, Wayland the Smith was not always so well-disposed, as the imp Flibbertigibbett discovered. When hit on the head by a large sarsen2 stone, thrown at him by the smith, the imp went away snivelling—as a proof of which the sarsen stone is still to be seen at Snivelling Corner, 2.5 km to the northwest. (It has to be added that Wayland’s abilities were as nothing compared with those of a British Parliament that in 1974 managed to throw the entire monument from Berkshire into Oxfordshire—an unfortunate act, bearing in mind the archaeological habit of listing monuments by parish and county.) Stones of this kind are typical of the downs to the northwest of Marlborough, and the barrow that has somehow become associated with the name of Wayland can be regarded as the easternmost of the group of stone-chambered long barrows on the Marlborough Downs, although it is relatively isolated from monuments of the same period. In style, it fits most comfortably into the Cotswold-Severn class of chambered tombs, but again only as an outlying example.
FIG. 11. The central area of Wayland’s Smithy, phase I.
At Fussell’s Lodge, in looking along the line of the barrow in either direction, one looks upwards to the distant horizons. At Wayland’s Smithy, which is almost on the highest ridge of the downs, the surrounding terrain lies below, except to the east. Along the prehistoric track known as the Ridgeway, which passes close by, lies Uffington Castle, 2 km to the east. This is a natural plateau with the remains of an Iron Age hillfort. Adjacent to it is the Uffington White Horse, a figure large enough to be visible from many miles distance, when not hidden by the downs that enfold it. Formed out of chalk exposed where the turf has been cut away, it is of indeterminate age. It was once supposed to be Saxon work, and is now said to be of the late Iron Age, but in a later chapter it will be suggested that the White Horse has Neolithic origins. Whatever the answer, the neighbourhood of Uffington has long been a very special place.
The barrow at Wayland’s Smithy was newly excavated in 1962–3 by Richard Atkinson and Stuart Piggott, who confirmed earlier suspicions that it had been constructed in at least two phases. The first structure (at the centre of Fig. 10) had contained a wooden mortuary house lying between two extremely sturdy end-posts made out of split tree trunks, each 1.2 m across the dividing diameter. The turf had been stripped from the ground before a pavement of sarsens was laid on the chalk. The skeletal remains of at least fourteen persons were found on the pavement. A bank of timbers, which Atkinson took to have rested against a mortised ridge pole, and sarsen stones to both sides of their bases, had formed the walls. The central tomb was at some stage covered by an oval mound outlined by sarsen slabs up to a metre tall, but these were not set into the ground. That particular mound was a cairn of small sarsen boulders at its base, topped with chalk rubble taken from ditches to the east and west. The oval structure, which even with its ditches fits into an area of only 15 by 20 m, was later covered completely by the final tomb, and is no longer visible.
One curious fact about the skeletal remains in the older tomb is that many of the smaller bones were missing, particularly hands, feet, kneecaps, and lower jaws. The absence of jaws is intriguing, in view of the fact that in some cultures with a literature testifying to their beliefs—the Egyptian, for example—the jaw-bone was preserved separately, and was supposed to have the spirit of the dead person attached to it. It is often missing from Irish barrows, and near the south entrance to the circle at Avebury a fragment of skull was found in the topmost layer of rubble, together with no fewer than five mandibles. A similar collection was found at the Sanctuary, not far from Avebury, and such examples could be multiplied. Perhaps the mandible was associated with the voice of the dead.
Just as at Fussell’s Lodge, there are posts at Wayland’s Smithy arranged in a quadrangle at one end of the first mortuary house—here the southern end. One pair of posts is nearly parallel to the eastern kerb of the later structure, which was about 22.3° west of north. The ditches are not straight, but the average angles of reasonably straight sections are about 19.5 and 20.5° west of north. Putting a median line through the central tomb by eye, it lies at about 21° west of north, but judging by what was found at Fussell’s Lodge, the lines grazing the D-posts are probably what mattered. The directions of their flat faces are about 22° north of east. The fact that these are almost at right angles to the line of posts mentioned earlier, and to the later east wall, makes us suspect that—even for the first structure—we should be paying special attention to angles in the neighbourhood of 22°. A very straight Iron Age ditch (ditch 400 on the plan of the whole, see Fig. 10) was in a direction 22.4° west of north. Even more significant: the virtually straight section of the Ridgeway between the tomb and Uffington Castle lies at an average of 21.5° north of east, that is, nearly at right angles to the directions under discussion. (For a map of the area see Fig. 80 in Chapter 4.)
In his report of the excavation of 1962-3, Atkinson pointed out that after the massive tree-trunks had rotted and collapsed, many small boulders from the surface of the mound fell into the void created. This suggested to him—in view of their large diameters—that the timbers may have projected far above the top of the mound, and that ‘perhaps carved or painted, they may have formed a landmark visible for miles around’. Were they, perhaps, together with some celestial object, observed from a point on the Ridgeway, say from the dip in the path, 800 m east of the tomb, or even from the shoulder of Whitehorse Hill? It is extremely probable that such sightings took place.
The visitor to Wayland’s Smithy today sees only the later barrow on the site, a trapezoidal mound nearly 55 m long and tapering from about 14 m wide at its southern end to under 6 m at its northern. The ditches that originally flanked it, as much as 4.5 m wide in places, and of graded depths (of roughly human proportions), are today completely filled and invisible. The southern façade originally had six large sarsen stones, the highest rising to a height of about 3 m on either side of the entrance to the burial chamber. (The arrangement roughly resembled that of timber façades known from earlier barrows.) Two stones are now missing. The eastern long side will here be taken as having been at 22.3° west of north, with the façade precisely at right angles to it and the western long side at 13.7° west of north.
Earlier excavators of Wayland’s Smithy reported a much disturbed interior, with the disordered bones of perhaps eight skeletons, including that of a child. The chamber was in the form of a cross, the main corridor being 1.8 m high at the crossing, and 1.35 m in the transepts. Before the second barrow was built, the ground had been cleared by fire, and a specimen of charcoal from a tree branch or small trunk yielded a radiocarbon reading corresponding to a calendar date in the probable range 3800-3440 BC. The degree of silting was thought to point to an interval of perhaps not more than fifty years between the two barrows; but this was to assume that the ditches were not regularly cleaned, and much of the evidence of this book will be that this was very probably—indeed religiously—done. Bearing in mind the strong similarity of the first mortuary house at Fussell’s Lodge and its Wayland’s Smithy equivalent, these dates imply that the overall architectural style might have been preserved for four centuries or more.
After the Iron Age ditch was dug, perhaps even before, it seems that the area between it and the barrow came under the plough. During the Romano-British period the ditch was re-cut, and a fourth ditch was cut to the southeast of the barrow. What is particularly interesting is the fact that the angle of that third ditch is almost identical to those incorporated into the older structure. Why this persistent orientation, not to say precision, and why was the line of the Ridgeway displaced by exactly a right angle from it?
The near-northern line is towards the setting of Deneb in the fourth millennium. No precise discussion of dates should ever begin with this star, however, since it has the property that its horizon positions change very slowly with the centuries. This fact, while it is annoying for anyone who wants to use a Deneb direction to establish a date, would have been an excellent reason for early peoples’ fidelity towards it. To give an idea of just how constant was this star: its declination (angular distance) from the celestial equator in 4200 BC was almost exactly 37°, and this gradually reduced to 36.22° around 2800 BC before increasing again. (It reached 37° around 1400 BC, and had not quite reached 39° by 200 BC.) Its risings and settings, which depend on that declination, changed correspondingly little over this enormous period of time, and if any religious ritual required its observation at an early date, it would not be at all surprising to find that a similar ritual continued to be used, especially at a site like Wayland’s Smithy, which lay close to centres of intense human activity for so long. It is no doubt for reasons relating to the behaviour of Deneb that the directions of the eastern edge of the second long barrow and the central line of the Iron Age ditch 400 differ by only about a tenth of a degree, while they might be two millennia apart in age. To reinforce the point: Deneb had almost exactly the same declination in 1600 BC as in 4000 BC, so that the direction of the star’s setting was the same at both times. This does not mean that it had not changed in the intervening period; but even then it had changed little.3
There are other conspicuous lines, such as that of the façade, the flats of the split trunks, and the Ridgeway itself, that are at right angles to the Deneb line. From the dip in the Ridgeway, and looking along it and up over the barrow, at 21.55° south of west, Aldebaran could have been seen setting around 4200 BC, when its declination was –11.83°. Alternatively, the setting Betelgeuse could have been seen along the same line four centuries later (3780 BC), and for the moment there is no way of deciding between these alternatives. There is a ‘disused pit’ marked on the Ordnance Survey actually on the Ridgeway, close to the optimum point for viewing either of the two stars, and it would be interesting to know whether the pit is of prehistoric origin, or whether its position is merely fortuitous. (The monument was constructed in a place cleared of woodland and sarsens, and the finds of pottery and stone implements on the site have raised suspicions that there might have been a trading centre here.) From the shoulder of Whitehorse hill, the setting of Aldebaran would have been seen around 4300 BC. The rising of Altair in the reverse direction over the shoulder of Whitehorse hill (with or without tree cover, since the extinction angle is paramount) occurred at about the same period of history, and this might be an indication of the date of the track itself. Betelgeuse, at the later date, had no very bright star as its ‘opposite’.
Turning to the posts at the southern end, which so closely resemble those at Fussell’s Lodge, it emerges that—within the limitations of accuracy set by relatively close posts—looking southwards along the direction set by the western edges of the D-posts (16.8° east of south), alpha Crucis would have been seen rising, the optimum date being around 3940 BC, or 3890 BC if we take the other edges (making a direction 16.0° east of south). Two of the scaling posts align perfectly at 23.1° with the eastern edge of the northern D-post and the direction of Deneb’s setting, around 4040 BC, but there are the usual large uncertainties in a Deneb date.
As for Aldebaran and Altair, which seem to be indicated by the Ridgeway, the southernmost pair of scaling posts is suited to the marking of Aldebaran’s setting in one direction around 4144 BC,4 and the rising of Altair in the other, around 4190 BC.5 An azimuth of 21.0° north of east and south of west is assumed here, taking the posts to have been centrally seated in their holes.
This spread of dates is not very satisfactory. The posts are of course too close together to allow very accurate dating, but it is a fact that if we assume the post in the southeastern hole displaced to make the azimuth 19.5°, the two alignments bring the named stars more or less into agreement in time, with Aldebaran at 3984, Altair at 3970, and alpha Crucis (as before) at 3940 BC. While this may seem to be a very devious move on our part, there are arguments in its favour. First, it makes for consistency with yet other dates, that are to be derived in a different way from the first barrow, but only after an analysis of its successor. And second, as an examination of the excavator’s drawing reveals, the hole in question is oval, unlike the others, and the required shift leaves the post well within the oval. In short, the first barrow can be provisionally placed at, say, within a century of 3965 BC.
In reporting on the excavation, Atkinson noted that the posts south of the main chamber must have predated it, since the northeastern post hole was covered by the sarsen slab leaning inwards at that point. This offers support for our view that what we are calling ‘scaling posts’ were destined to be discarded after they had been used to ensure a correctly aligned mortuary house, and were not for use after its completion. Fussell’s Lodge offers comparable evidence.
There is more to be said about potential sightings over the first mortuary house, but the way to discovering them is through an examination of the later barrow. The neat trapezoidal form of this will provide evidence that risings and settings were observed by people standing in the ditches flanking it, and will lead on to a realization that the first structure had very probably already been used in the same way. Parts of the ditches will prove to have been for continuing ritual use, and other parts to have been of use only at the design stage, like the scaling posts.
It will be seen from Fig. 13 that the ridge line of the old mortuary house at Wayland’s Smithy was in one direction more or less the line of the eastern wall of the new grave chamber, while in the other direction it passed through one vertex of the trapezoid. Apart from a very slight shift in orientation, there was an evident awareness at the new barrow of the traditions by which the old was planned. The precise form of the new barrow, however, is something about which we are not yet in a good position to speculate. On the face of things, the line linking the two structures could not have been the line of a ridge running all of its length, since that would have made the surface slope too steeply at its northwest corner. Reasons will be given for thinking that the ridge was displaced very slightly from the perfect line to the northwest vertex.
Judging by the higher end, the angle of view from the ditches could hardly have been less than eight degrees to the horizontal and is unlikely to have exceeded 15°. Taking this as starting point, consider then which important stars could have been seen from the two ditches, on the assumptions (1) that viewing was at right angles to the long edges of the barrow, whether the near or the far side, and (2) that viewing from opposed positions in the two ditches set the same altitude for viewing.
What seems to have been a constant concern for creating features at right angles to the ditches at many long barrows—for instance here and at Fussell’s Lodge, but in quite a different way there—offers superficial support for (1), while (2) is underwritten by the near-symmetry of the flanking ditches of most long barrows, especially of their inner edges. Accepting these principles provisionally, potential lines of sight of a typical pair of observers are shown in Fig. 12. The assumption in the case illustrated is that the perpendiculars are to the near sides of the barrow. The lines are chosen to pass over the northeast corner of the new sarsen chamber, which it will be recalled lies on the line through the old D-posts. This point would have been at the same distance from both observers. Assuming that the section across the barrow was reasonably symmetrical at its high end, we have all that we need.
FIG. 12. It is wise to keep an open mind over the question of whether viewing was at right angles to the far edge (or far ditch) of the barrow, or to the near edge (or ditch). There seems to be no evidence that a right angle was taken with an intermediate ridge.
FIG. 13. The long barrow, Wayland’s Smithy II. Observations were made by people standing in the ditches, for example at B, using the barrow as an artificial horizon. The barrow set horizons of identical altitudes to observers in opposite ditches, but the directions (azimuths) were not the reverse of one another. Another potential observer stands on level ground at A and looks up the ridge of the barrow. The dimensions given to the chamber are only approximate, and are for the outer surfaces of the stones.
Since the joint principles (1) and (2) will be invoked on many other occasions, a defence of their great naturalness may be offered here. It is easy to believe that a perfect arrangement was regarded as one in which the rising of one star exactly opposes the setting of another. There would have been times when this was found over open ground, but if the horizon was to be set by a barrow, then the simplest and most appealing arrangement would surely have been a parallel-sided barrow in ridge-tent form, with the observers square on to it and at the same distances. Granted that the observers are at the same level and equally distant from the ridge, the two stars would have had equal altitudes. (There is no suggestion that the person observing them was necessarily conscious of viewing at right angles to any particular feature, or that the altitude of view was a significant angle—although evidence will be gradually accumulated that the latter might have been the case.) The stars would almost invariably have failed to cooperate in this ideal scheme, however. While not all of these conditions could in general have been met, the barrow could nevertheless always have been built with (a) the direction of one star perpendicular to the ridge, the same ridge being used for the other star, observed obliquely now (that is, not at right angles to the ridge) from exactly the same distance. Alternatively, the two observers could still have looked along lines perpendicular to the ridge, and yet view at different altitudes, either by (b) standing at different distances from the ridge, or (c) varying the levels of the ground on which they stand. Judging by the structural features of barrows, options (b) and (c) seem to have been generally disliked. There are exceptions to this rule, but even a cursory survey of styles of ditching round long barrows shows that if viewing was indeed from them, then there was a preference for symmetry of viewing position—in distance and in level. Even when a barrow runs along a contour on steep ground, so that one side is appreciably higher than another (it will later be seen that this was so at Giant’s Hills, Skendleby, Lincolnshire), this ambition was evidently aimed at and achieved in ingenious fashion, by stepping the ditch edge.
It appears that oblique viewing was usually preferred to (b) and (c); but that having sacrificed directly opposed viewing, their architects managed to preserve viewing at right angles to a new ridge, or at least across a new edge. Our double principle ((1) and (2) together) is in fact an extremely natural way of preserving monumental symmetry; and if, on a plan, ditch symmetry seems to be absent, then that is likely to be because one or both of the ditches is parallel to an edge that is not obvious. The directions of ditches offer important clues as to the directions (azimuths) of viewing, but here there is a difficult decision to be made—one to which allusion was made in an earlier section. Is viewing to be at right angles to the local ditch or to that on the other side of the mound?
FIG. 14. Lines of sight (altitude 11.7°) from the eastern ditch over the crossing in the burial chamber at Wayland’s Smithy. The vertical lines in the ditch (terminating in points corresponding to the observer’s eye) represent two possible viewing positions that share the same line of sight. The broken lines are suggested limits to the cross-section of the mound at this place. The stones fronting the mound (not shown here) would all have fitted under the sight-lines, and in the case of three out of the four surviving stones, the fit would have been rather precise. Only one (at extreme left) would have failed to fit under the three-part broken line.
The most natural solution at first sight is one that puts the right angle near at hand. At the Beckhampton Road long barrow—which will be placed in the thirty-fourth century BC—it was seemingly there the far edge that counted. The scaling posts at Fussell’s Lodge and Wayland’s Smithy are ambiguous, but the placing of the ditches in relation to the latter will turn out to fit better with the idea of viewing at right angles to near edges. All told, we are left with little choice in the matter: every example must be worked out for both possibilities, and the solution preferred that is more consistent, internally and with radiocarbon dates.
All possible angles of view are considered until a pair of declinations is found that were simultaneously held by two bright stars at some likely period of prehistory. This is done on both assumptions concerning the right angle, whether it is with the near or far side. There are two qualifying objects that seem most likely to have been observed across the barrow at Wayland’s Smithy, namely the Pleiades (setting) and Spica (rising). (The position of the Pleiades will always be taken as that of the brightest star in the cluster, Alcyone, then of magnitude 2.86.)
First for a near-side perpendicular: for viewing at an altitude of 11.7°, a pair of simultaneously valid declinations is derived. The Pleiades (at azimuth 247.7°) would have had declination –4.14°, and Spica (azimuth 76.3°) declination 17.47°. These values are appropriate to dates in the neighbourhood of 3670 BC. The altitude of 11.7° is close to a gradient of 1 in 5. Whether the true ridge was visible from the ditch depends on the precise shape of the barrow. If viewing was from mid-ditch, the distance to the ridge (17.4 m) would imply a maximum height for the barrow of about 3.6 m, again depending on the form of the cross-section. This is the height above the observer’s eye, but judging from the only relevant ditch section available to us, the eye must have been very near to ground level (see Fig. 14). The height of the top of the stone covering of the chamber is approximately 2 m, which is well below the required maximum.
If this solution is to be accepted, with more than a metre’s depth of soil above the stone capping, then at least it fits with the greatest problem in viewing the Pleiades—their high extinction angle (about 4.4°). The mound, on this hypothesis, would have set an appreciably higher angle of view.
FIG. 15. Three potential solutions (marked by small circles) for the viewing of stars across the Wayland’s Smithy long mound. Continuous lines are for viewing at right angles to the far edge of the barrow, broken lines for the near edge.
Repeating the calculation for a set of azimuths perpendicular to the far edges, other dates are obtained, in one case with very different implications for the form of the barrow. The stars are the same (Spica and the Pleiades), the date 3710 BC, but the angle of view is only 5.35°. The lines of sight would in this case not have been high enough to clear the chamber, and so this is rejected, but another two solutions present themselves on this second hypothesis. Both are for the thirty-sixth century. One pairs the settings of Procyon and Regulus (around 3581 BC, altitude 11.70°), the other the settings of Procyon and Capella (around 3549 BC, altitude 11.85°). These various solutions are shown graphically in Fig. 15.
Which of the three viable solutions are we to favour? All the dates derived fit comfortably into the range found from the radiocarbon dating of a branch or small trunk that the excavators linked with the operation of clearing vegetation from around the previous barrow by fire (2820 ± 130 bc, or 3620 ± 180 BC). This is an ill-starred situation, and other evidence must be brought to bear on it.
Procyon and Capella are both brighter than Spica, and they and Regulus are much brighter than the Pleiades. If one inclines to the Spica–Pleiades combination, it must be for other reasons. Independent dating from other possible alignments might offer us a way.
FIG. 16. The stone mortuary chamber of Wayland’s Smithy II, showing how rays from the brightest stars of the Southern Cross, when rising, fell diagonally across the chamber, grazing carefully shaped stones. (The blocking stone at the entrance, used to seal the tomb, is omitted. The places of missing stones are not indicated, but should be obvious.) The faces of the large stones at the ends of the transepts are likewise carefully placed so as to parallel the sides of the mound, and so would have aligned on alpha Crucis and Deneb before the mound covered them. If the arrangement indicated is not an illusion, then it is plain that religious conceptions were more important than astronomical observations. The upper plan covers the entire mound, and shows how beautifully the geometry of the chamber is related to that of the mound. The details hardly need to be spelled out.
The tomb begs to be considered as an artificial horizon for a view of the southern sky over the length of the barrow. The slope of the barrow was evidently small: judging by its dimensions, known and inferred, the angle set by it to an observer standing at ground level was a little under 2.36°, depending on its form. The massive central stones at the southern end would have been hidden to an observer at the north. Now it so happens that during the period in question the only reasonably bright star in anything approaching the required direction was the rising alpha Crucis, which has an extinction angle close to this figure. The star could have been seen over the chamber by an observer standing at the northeast corner along a line exactly parallel to the western long side. This was very precisely true around 3750 BC, assuming an extinction angle of 2.42°. (It has to be remembered that estimates of extinction angles are not absolute. To give an idea of potential uncertainties: bringing the angle down to 2.1° would in the present case advance the year by a century.) Alternatively, an observer standing at the northwest corner would have seen the same star rising at the same altitude, as if it were coming out of the large sarsen stone at the southwest corner. (This stone is unfortunately now missing, but was probably higher than the corresponding stone on the other side.)
There are interesting resemblances between old and new chambers: the internal measurements of the stone chamber were such as to bring in the ray from the rising alpha Crucis to the northeast corner of the chamber at this time. (With an estimated azimuth of 167.8° and an extinction angle of 2.35°, a year of 3640 BC is obtained.) There is, however, a new and quite unexpected alignment with the other main diagonal of the chamber, this time on the rising of beta Crucis.6 As in the other case the line is uncertain—they are estimates for standing observers rather than prone skeletons—but an azimuth of 150.1° and an extinction angle of 2.42° suggests a period around 3680 BC. The west ditch is just long enough to accommodate that same alignment over the crossing point in mid-chamber (see Fig. 16).
In sum, taking these additional dates into consideration inclines us to take the Spica–Pleiades combination, with its earlier date (3670 BC) and its more intuitively acceptable style of viewing at right-angles to near edges. Doing so, moreover, brings us to the middle of the radiocarbon range. All the dating is subject to error. For the time being, all too little is known of the ditch floor, but one thing at least seems probable, albeit based on only two sections: the floor falls gradually towards the north, as the ditch is pulled in to the tail of the barrow. This is surely how the angle of view was preserved, while lowering the height of the barrow. (Note that the principle of viewing at equal altitudes was generally implemented by having opposed sections at equal levels. It does not require level ditches, although they were indeed usually level.) When the ditch is completely excavated, the barrow’s form will be better known, but even now one can say—on the basis of ditch sections at the transept and at mid-mound—that the barrow was falling over this stretch at a gradient of about 4.3°. This gradient would have brought it down almost perfectly to ground level at its very tail, but perhaps the gradient was levelled out towards that end, so that the tail retained some height.
FIG. 17. The central area of the next figure.
This analysis has an important consequence for the sequence of building. For those who reject an astronomical interpretation, it makes little difference whether the mound preceded or followed the chamber. The precise astronomical alignment of the stones of the chamber, however, makes it seem certain that they were set up first, rather as the scaling posts for the older barrow had been.
Some people have seen the massive sarsens flanking the entrance to the surviving barrow as male to the right and female to the left. Alpha Crucis has a natural partner, beta Crucis, then slightly brighter, and to be seen on the side of the ‘male’ stones flanking the entrance. (One might equally say that it illuminated the ‘female’ side of the interior, and it would take a Jungian psychologist to decide on the more probable interpretation.) The two stars might well therefore have been regarded as male and female. Was gamma Crucis (magnitude 1.63) their child? The fourth star making up our cross is relatively feeble—but a second child, perhaps?7
FIG. 18. A general view of the planning of the mortuary house and its ditches at Wayland’s Smithy. For lettering, see the previous figure, and for a detail, see the next figure.
FIG. 19. A detail of the previous figure, showing the shape of the shallow pitched roof of the earliest mortuary house at Wayland’s Smithy. No attempt is made to show constructional details. Hatching is added simply to make the pitch (and the ridge) more evident.
Frame and Form at Wayland’s Smithy I
What has now emerged about potential uses of the later barrow at Wayland’s Smithy reflects back in an important way on the earlier phase of activity on the site. All the evidence of scale rules out the idea that observation of risings and settings of stars across the earlier structure was done by people standing at ground level, which would have required a higher central edifice than the later barrow. But what if Spica and the Pleiades were observed from the ditches, as it seems likely that they were with the second barrow?
An artificial barrier of wood or chalk or stone over the mortuary house would then have provided the horizon. Only two or three of the edges defined by what is known of the monument are acceptable for the purposes of observing the two stars, on the hypothesis that viewing was at right angles to the edges. They are the lines labelled KB and JA (or parallels to them) in Fig. 18, and they set azimuths (measured from north) of 73.2°, towards the rising of Spica, and 246.9°, towards the setting of the Pleiades, at an altitude and period as yet undetermined.
Note that whereas the second direction represents a shift of less than a degree from what was set by the later monument, the other direction differs from its later equivalent by three degrees. This alone should lead one to expect an appreciable time interval between the two monuments.
Applying exactly the same argument to the first phase of the barrow as that applied earlier to the second, perfectly consistent results are obtained only around the year 3940 BC. The two stars—the only bright stars with the property of being seen as stipulated over the right spread of centuries, would both have been observed at an altitude of 10.4°, when the Pleiades’ (Alcyone’s) declination was –5.59° and Spica’s was 18.58°.
Since Procyon, Capella and Regulus offered alternatives with the later mound, it is worth mentioning here that the dates they yield, whether on the assumption of viewing at right angles to near or far edges, are either much too early or too late (4555, 3490, 3415 BC).
However implausible it may seem to claim that we can make such a precise statement (overlooking small uncertainties in the azimuths) about a monument that has left hardly more than two-dimensional traces of its existence, there is much circumstantial evidence for our conclusion, as will appear in the course of the following suggested reconstruction of the procedures adopted by its builders:
The holes were prepared for the split trunks N and S, and N was erected. Scaling posts were set up at D and A, aligned with the eastern edge of N on the star Deneb. (Post d was also used along the same line, but for mechanical reasons and not geometrical, so that here it will be passed over in silence.) The opposition of Aldebaran’s setting and Altair’s rising was well known at the site, as was the fact that their directions over a natural horizon were almost perfectly at right angles to the Deneb line. Post B was set up to provide the requisite Aldebaran—Altair line. Some fine trimming in the position of A was needed, hence the oval hole. Post C was next placed so that it yielded a right-angle DCB, needed for the ‘viewing at right angles’ rationale of the monument. Once the lines of the artificial horizons had been so defined by the scaling posts AD and BC, the experimental part of the operation began. Beams were tied across those posts (see Fig. 18), the first of them perhaps to d too, and carefully adjusted in level and height (always kept equal) so that the rising of Spica and the setting of the Pleiades was always viewed across them at right angles. This was easily arranged: they simply used the uprights, C with D and A with B, to define the direction in conjunction with the appropriate beam. But viewing now had to be precisely done from a definite height, ground level, and the ditches were dug accordingly. Viewing was later going to be by people standing more or less in the ranges UV and XW of the ditches. The UV area was dug first, and the ditch extended to a point (E) suitable for observing the Pleiades. There was to be no wasted effort: digging was to be done as far as was needed, and no further. The same was true for a point G in the western ditch. (But later the ditch was extended to F so that an overland sighting of Spica was possible.) Care was needed to ensure that E and G were at precisely the same distances from their respective beams. At last the requisite height of the beams was found (it was about 1.09 m above eye level), and the roof of the mortuary house could be erected.
This was in the first place a wooden structure, and its being pitched would have aided the run-off of water, although it was only at a shallow angle of 10° or so. The perfect pitch for observing the stars was 10.4°. It was the two edges that were really crucial, and these were defined by marks cut into the split trunk N and by the short posts flanking S. These short and puny posts did not bear any load: they were levelling posts, each hammered into the chalk until level with the edge of the corresponding beam. The nearby sarsens bore the weight of the roof ridge.
By an unhappy circumstance—heavy sarsens in the ditch-fill—the inner ditch sections were not always excavated to the bottom, and never where they are of greatest interest to us. However, the inner ditch gives every sign of having been a normal viewing ditch, flat-bottomed and within a few centimetres of 1.65 m below the thin chalk soil (rendsina) under the cairn. (The estimated heights of eight adult males in the tomb yield an average of 1.70 m, and the average eye-level of the tallest three would have been about 1.65 m.) This has important implications for the astronomical argument. Clearly the remnants of the later cairn must allow for the idea that the viewing angle did not greatly differ from 10.4°. This they do, but the two sarsens at the southern D-post—which on our reconstruction supported the roof of the earlier wooden chamber—if set upright in the (known) stone holes from which they came, set a more reliable criterion. Each would have been almost exactly 1.05 m above the floor. The viewing angle requires the ridge at this point to be a minimum of about 1.20 m, and for more comfortable viewing 1.30 m, so that there is ample leeway for a (probably solid) roof of 15 or even 20 cm thickness. Trunks of this diameter could have provided a frame for the roof, bringing it up to the right height. The need for waterproofing might be thought to argue for a covering of turf or even compacted chalk, but solid trunks, split, trimmed to shape, and calked in some way would have been better.
The shallow ridge of the roof runs from K to I. What might easily have been taken as a sign of incompetence on the part of the builders, who might have been thought simply incapable of positioning one massive trunk precisely opposite another, is now on the contrary seen as testimony to genius of a high order. The trapezium outline of the horizon-roof is very much the same as the shape of the later long barrow, and no doubt its pitches were much the same. Furthermore, there is an exact analogy between the placing of the stone mortuary house within it, to the side of the central line, and the mortuary house under the roof between the split trunks (compare Figs. 13 and 19). In other words, the great barrow at Wayland’s Smithy was very probably a replica of that original roof, in all but the fine details of the angles, which the precession of the stars had altered.
The observing ditches were dug so that the ranges XW and UV were equidistant from the roof edges, just as G and E had been from their beams. Each range had space for six or seven people standing side by side. (Could they have been three men and three women, like the sarsens fronting the barrow?) Each person probably had a place-marker, say a ridge cut in the chalk. Note that the chalk at X was dug so as just to accommodate a viewing position, but with little more expenditure of energy than was necessary. Later extensions to the ditch took it up to other viewing positions L, M, and Q—and again took it no further than was necessary—from which the stars under discussion here could have been seen in different ways. Which stars they were will be appreciated without further comment, if Fig. 18 is examined for parallels to the lines of sight already discussed.
In possession of our several new principles, it is possible to re-examine the first chamber at Fussell’s Lodge, here previously dated at around 4235 BC on the basis of individual stars—Aldebaran, Spica, beta Crucis, and beta Centauri. The broad principles embodied in the use of scaling posts, whose positions are at Fussell’s Lodge known fairly accurately, are now taken to be the same as at Wayland’s Smithy; and these provide a pair of directions, across which it will be supposed that viewing was at right angles. Examining them for stars at low altitudes, no qualifying pair emerges at all, but as the altitude is increased it becomes clear that the setting Arcturus to the north8 and the rising Betelgeuse or Bellatrix to the south9 are possible candidates.
FIG. 20. Two possible forms of roof for the Fussell’s Lodge mortuary house, seen from different viewpoints. The first form, anticipating the shape of the barrow, is favoured here.
Although they are similar, the mortuary houses at the two sites differ in ways reflected in the final forms of their associated barrows. First, it seems that one of the right angles within the quadrangle of scaling posts is exactly as before, but that the other is not. It seems to have been drawn to the axis, and perhaps relates to a post in Ashbee’s ‘Pit III’ at the centre of the four. Second, the Fussell’s Lodge example is not itself plainly associated with a pair of viewing ditches and there was certainly none for use with the scaling posts. As for the way they functioned, much the same procedure could have been followed as at Wayland’s10.
FIG. 21. Potential right angles in the (highly stylized) forms of the barrows at Fussell’s Lodge and Wayland’s Smithy, drawn to approximately the same scale.
Again it is necessary to consider viewing at right angles to near (qualified as explained) and far edges. In both cases the stars would have been the same. At an altitude of no less than 20.4° a perfect fit for the first alternative is produced, with Arcturus at declination 54.89° and Betelgeuse at –14.36°. The required year is 4240 BC.
FIG. 22. A cross-section of the ditches at Fussell’s Lodge (P. Ashbee’s section CD), roughly 15 m from the wide end and 25 m from the narrow. Note that the revetting posts have one third showing and two-thirds buried—a sound constructional principle.
Again it is necessary to consider viewing at right angles to near (qualified as explained) and far edges. In both cases the stars would have been the same. At an altitude of no less than 20.4° a perfect fit for the first alternative is produced, with Arcturus at declination 54.89° and Betelgeuse at –14.36°. The required year is 4240 BC—complementing in an unexpected but very welcome way the date (for the foundation barrow) previously obtained by averaging over individual (and different) stars, that is to say, 4235 BC.
Taking the second alternative, the altitude increases to 23.64° and the year is 4180 BC. In all cases small errors in the fundamental directions can alter appreciably the dates derived. (An idea of how sensitive the method is to error in the directions of key edges will be given briefly in a later section.) This alternative fits marginally more comfortably with the radiocarbon dates for material from the D-post pit (4250–3950 BC), but the first is to be preferred if Wayland’s style is to be followed. Far more important than those dates, however, is the fact that procedures derived from a study of Wayland’s Smithy can explain this monument equally well.
Was there a viewing ditch for the mortuary house? The palisade trench for the later barrow would not have served, for it would have been much too cramped, but short sections of the ditches known from the final barrow would have served very well. These ditches were relatively cavernous in the necessary region, reaching a depth of about 3.6 m below the base of the mortuary house. This great depth, slightly more than twice human height, might at first seem to be an insuperable objection to the idea that viewing was from them, but in fact the ground chalk is cut away, especially on the inside edge. The excavator, P. Ashbee, interpreted this as a weathering ramp, and no doubt some weathering has taken place—although presumably fairly uniformly. The angle of this ramp, however, is suspiciously close to what we require it to be, and not only in one section but in all seven places where Ashbee’s profiles allow us to measure it. To both sides of the house, an adult observer of normal height standing at the back of the true floor of the ditches—almost symmetrically now—would have been perfectly placed to look along a line barely skimming the ground at between 16° and 19°. The angles are not easy to measure accurately, but rejecting Ashbee’s section K (which lacks the flatness of the chamfer of the others, but certainly produces something in excess of 14°), the other six range from 16° to 19° and average at about 17.5°. This might not seem particularly close to 20.4°, but that is hardly surprising, for in its present form it does not belong to the mortuary house phase at all. It is a figure germane to the final barrow phase, and as we shall see shortly this figure is within a degree of what is predicted—again very welcome evidence in favour of the general principles being proposed.
The implied height of the ridge of the mortuary house is 2.45 m, almost identical to the height of the final barrow at this point, as estimated by Ashbee—although the loaf shape he favoured for the final barrow does not fit with the ideas being put forward here. Accepting the idea of two distinct phases, with the cairn an intermediary, building around the old house was just as rooted in tradition as at Wayland’s Smithy. Again, the form derived for the roof of the old mortuary house very probably resembled the flared form of the later barrow, for which there is the evidence of the palisade trench. It is conceivable, however, that the roof had a simpler form into which the gradients for the two different directions were nevertheless worked. The two alternatives are illustrated in Fig. 20. The fact that the pit between the split trunks is somewhat east of centre seems to favour the flared alternative, for if the pit was meant for another short trunk supporting the roof, it would have been appropriate to place it at the centre of gravity of the roof. It is in fact placed precisely where one would judge the centre of gravity of the flared, barrow-shaped, roof to have been.
At Wayland’s, we worked back from the later to the earlier. At Fussell’s Lodge, we now know something of the local tradition, and also the directions of the bounding lines of the barrow across which viewing at right angles would have been planned. The outline of the later structure has been corrupted to some extent through the collapse of parts of the palisade trench, but averaging across sizeable lengths, the lines of sight seem to have been at 16.2° west of north and 23.3° east of south. From those data, using the same sort of argument as before, the conclusion is drawn that the date of the barrow was around 4180 BC and that viewing was at an altitude of 18.1°. The same star, Arcturus, was observed setting, but now the rising Bellatrix rather than Betelgeuse seems to have been observed. (The stars now observed were at declinations 54.63° and –17.83° respectively.) Even if viewing was at right angles to far edges, the same stars produce the only acceptable solution, with the altitude 20.23° and the year 4082 BC.
The lower angle seems preferable, as already explained. Other options might be thought relevant, for it does seem that the Fussell’s Lodge barrow has certain right angles deliberately built into it. (Fig. 21 shows three, and the Wayland’s barrow is added, to scale, as a reminder of the arrangement there.) If, instead of viewing across lines a and d, lines b and c were taken, the (equal) altitudes of the same stars Arcturus and Bellatrix would have been about 16.0° and the year approximately 4140 BC. If planning was based on b and d, then the angle was 17.1° and the year 3960 BC. The differences are not as great as one might have imagined, and simplicity favours the original choice.
In appearance, with viewing at such relatively steep viewing angles (see Fig. 22), the barrow must have resembled an upturned boat. It is conceivable that there was chalk outside the revetting posts of the barrow, so that only the tops of them were visible. In this case, the whole thing would at first have taken on the appearance of an isolated giant white wave in the landscape. If duly trimmed as the stars changed their declinations, and regularly scoured to prevent vegetation settling on it, this form could have been kept for decades. The revetting posts would have rotted, perhaps within a century, but the downward wash of the cover might have been repaired even for several centuries. Of course it could be that alignment on stars was merely a foundation activity, regarded as done once and for all, and that the stars were not observed over the barrows thereafter, or were observed for only a few years. Perhaps, in time, closer attention to ditches will produce the much-needed evidence—the burning of lamps on the walls, the tread of feet on the floor and on sloping platforms, provision for drainage inside and on the lip, scouring tools on the ditch bottom with unexpectedly late radiocarbon dates, and so forth.
It has been shown by experiment that weathering of the surroundings of a ditch can cause it to fill in five or ten years, and it has usually been taken for granted that the ditches of a long barrow filled up quickly in this way, soon after its completion, since it has been assumed that the only purpose of the ditches was to provide raw material for the barrows. If the barrows were used as suggested here, however, and not merely in a foundation activity, then they would have been regularly cleared of rubble, baled out, and possibly even recut from time to time, to accommodate changing star positions. They might even have been given penthouse roofs to keep them dry, and so have weathered much more slowly than if they had been exposed. This last is a purely speculative remark, in the absence of any remains; but such roofs would have been as nothing by comparison with those often postulated for circles like the Sanctuary and Woodhenge, which would make them into prehistoric equivalents of St Paul’s Cathedral. And needless to say, the bigger the roof postulated, the more surprising that no traces of its fallen timbers have been found.
A radiocarbon date for the long barrow on the Lambourn Downs, 4 km north of Lambourn, provides it with a venerable status that has not had much influence on its preservation. Corrected in the usual way, the date falls in the calendar range 4200 ± 200 BC. Part of the long barrow not in woodland has been badly ploughed down, and part in woodland has been interpenetrated by tree roots. It was excavated in a peremptory way by Martin Atkins in the 1850s and rediscovered by Leslie Grinsell in 1935. It was so ruthlessly ploughed that a rescue excavation was mounted in 1964 by John Wymer, who found that very little remained of the exposed mound.
Grinsell described it in outline, and following his plan its central axis seems to have been about 74.7° east of north, which is probably true to better than a degree. Wymer’s plan is more detailed, and shows that the ditches were tolerably straight, and had mid-ditch azimuths (corrected for 4.0° compass error) of about 76.2° (south side) and 73.2° (north side). The average corresponds exactly to the figure obtained from Grinsell’s plan. It differs only slightly from that of the Fussell’s Lodge barrow, but so, of course, do the geographical latitudes and horizon profiles at the two places, and the conclusions to be drawn from the two are very different.
Scattered sarsen boulders are still to be found at the eastern end of the barrow, deriving from a chamber there, and at the exposed western end many of the boulders still lying at the edge of the field are what were removed by the farmer, through deep-ploughing. The modern partial excavation of the head of the mound, within the wood, revealed that sarsen boulders in the mortuary area followed a well-defined line (azimuth about 72°). The Lambourn barrow was perhaps a two-stage structure, like those at Wayland’s Smithy and Fussell’s Lodge, containing a primary wooden-framed tomb that was covered over by the present barrow at a later stage.
Atkins, who tells us that a farmer had already dug into it before him, and that ‘human remains and a quantity of black earth’ were found, also unearthed other skeletons. One was lacking its skull—as a result, he thought, of the farmer’s enthusiasm. More recently, Wymer found a crouched burial in a rough sarsen stone cist, together with seashells, but this was not a primary burial. It is unfortunate that the character of the first structure is unknown, and that we are left with only the decrepit remains of the outer barrow as a guide.
The slope of the badly worn mound when Grinsell charted its profile in 1936 was about 2.3° upwards to the east, and an estimate of the barrow’s maximum height to be made shortly suggests that it might originally have been about 3.0°. The critical land horizon (eastwards) is 2.02° and had there been trees on the horizon 10 m tall, the angle would have risen to 2.32°. (Corresponding western altitudes are 1.54° and 1.97°.) It looks very much as though, just as at Fussell’s Lodge, the slope of the barrow was made to mask the distant horizon, as would be explained if the viewing of a rising star to the east was along its spine by a person standing in the natural hollow at its western end. This appears to have been a ditch, judging from old aerial photographs, and Grinsell took it to be such, like those he knew from Dorset barrows, but Wymer did not detect any such ditch.
Accepting an azimuth of 72° for the line of the barrow’s sarsen core, the rising of the star Altair would have been seen at the stated altitude around the year 3960 BC—a very tentative date, of course.
The barrow was certainly tapered, like that at Fussell’s Lodge, and judging by the ditches the taper amounted to only about three degrees. Observation towards the west was also in principle possible, using the natural horizon. Another way of looking at a star on the western horizon would have been to use the gentle incline up to the tomb entrance from the place of the present road (there is no path in that position now). A person of average height would have seen the foreground (land, track) coincide with the far horizon—assuming trees of more or less 10 m. The high end of the tomb, at 2.6 m above the surrounding ground or thereabouts, would have made a third (intermediate) horizon for the ray to skim, rather as at Wayland’s Smithy, which is less than 5 km away. The artificial altitude would in this case have been 2.0°, which is the extinction angle for the most likely star in that direction, the setting star Aldebaran. There is little point in setting down what are no more than directed guesses, based on a very superficial survey, but a line of sight of azimuth 250.5° would have been needed for the fortieth century BC, not far from the reverse of the supposed direction to Altair’s rising. Both of these alignments might have been embodied in a set of scaling posts, as at the head of the barrows at Fussell’s Lodge and Wayland’s Smithy.
Another possible observation along the barrow to the west could have been in connection with the setting of the Pleiades at the same period, as at Wayland’s Smithy. Not enough detail is known of the Lambourn structure to speculate on a precise alignment. Another long barrow over which the setting of the Pleiades might have been observed is at Nutbane, in north Hampshire, 20 km or so from Stonehenge, a barrow that is perhaps five centuries younger.
The use of the Lambourn barrow as an artificial horizon for crosswise viewing admits of two distinct solutions. One of these involves the setting Arcturus to the north together with the rising Bellatrix to the south. The estimated date (4272 BC) fits well with the corrected radiocarbon date of 4200 ± 200 BC, and yet reluctantly the solution must be rejected. It requires viewing in both directions at an angle of approximately 19.0° to the horizontal, a figure that implies a mound 4.7 m high, seemingly much more massive than the ditches could have provided.
The alternative solution, which is tentatively accepted, is that the setting of Vega was observed to the north and the rising of Sirius to the south, around the year 3970 BC, both stars being viewed at an altitude of 11.0°.11 The implied maximum barrow height (2.6 m) fits very well with what is known of the ditch sizes and their potential for providing material for the mound. The derived date is probably not in error by more than a century. The dates obtained for Altair and Aldebaran along the barrow’s length are compatible with it, but why then does it not square with the radiocarbon date?
The latter was based on a patch of burnt wood from the floor of the tail end of the south ditch. It could have originated with timber from an old mortuary house, or from posts that had been removed from an old structure by burning—broadly speaking, the bigger the post the older the timber. As one of the long barrows at Skendleby will later demonstrate, a set of more than a dozen radiocarbon dates from a single site may span two millennia, and any one of them in isolation might have given an entirely false impression. In a sentence: the Lambourn barrow we (barely) know is perhaps not as old as is generally supposed.
Although the Lambourn barrow is now in a relatively isolated spot, there was considerable later prehistoric activity nearby. It is unlikely to have been by chance that, almost in line with it, and half a kilometre to the east, is a disc barrow. Mostly to the southeast of it, and within a circle of radius under a kilometre, there are remains of at least thirty round barrows of various types of the early and middle Bronze Age. A dozen of these lie together in a group with the collective name of ‘Seven Barrows’ (the English having always found barrows difficult to count), and six of them lie virtually in a line that passes directly over the long barrow, while another four lie roughly on another line, parallel to the first. The former and better alignment is about 40° north of west (or south of east), but this depends on how one selects the barrows, for the line is not perfect. It seems obvious that these much later tombs were aligned on the setting of the Sun over the long barrow at the summer solstice—a direction which changed very slightly with historical period. (The calculation depends on the horizon altitude, which in turn depends on the line chosen.)
It is hard for us now to comprehend how such a simple mound as the Lambourn long barrow could have kept its reputation as a place of great sanctity for perhaps two thousand years after it was first erected. Having accepted the idea, however, we seem to catch a glimpse of one reason for shifting allegiance away from the stars to the Sun in these matters of alignments: the directions of the Sun’s risings and settings are more or less constant over very long periods of time. For most of the stars this is not true, even though Deneb, as we have already seen, had a certain constancy in its behaviour.
One of the most valuable of modern reports on long barrow excavations is that in which Paul Ashbee, I. F. Smith, and J. G. Evans describe respectively excavations that took place between 1959 and 1967 at Horslip, Beckhampton Road, and South Street, all in or near Avebury. The Horslip barrow lies a kilometre due south of Windmill Hill, and so is often called after that enclosure—which in turn has given its name to a whole culture. Selected radiocarbon dates for the three barrows produce the ranges 4000 ± 160, 3550 ± 185, and 3300 ± 100 BC in calendar years (3240 ± 150, 2750 ± 135, 2517 ± 90 bc). There is no unambiguous evidence that any of these barrows was built in more than a single phase, although Horslip probably was; and none has produced evidence that it was used for burials at all in the earliest phase, a discovery that has caused much surprise. All three barrows have been ploughed down considerably in recent centuries, but even so, the second and third retained extremely valuable clues as to the astronomical and architectural procedures followed by those who built them.
FIG. 23. Plan of the Horslip long barrow. Details of most of the inner area had been removed by ploughing. The letters indicate Paul Ashbee’s ditch sections. The broken-line trapezium represents no more than a directed guess at the outline of the limits of the barrow. The line of pits at the north western end might represent viewing positions for an earlier mortuary house on the site.
The Horslip barrow was originally perhaps of trapezoidal form, since the ditches that flank it are slightly splayed (Fig. 23). In 1743 William Stukeley described it as being ‘of large bulk, length, and height’, but ploughing after his time virtually removed all surface features. Its orientation was roughly 45° south of east, and in this direction the ground falls away. The southeastern end is the higher in relation to the surrounding ground, but the overall slope of the ridge of the barrow when it was erected was very probably between four and five degrees upwards in the opposite direction, that is, to the northwest. Observation in this direction seems unlikely, for want of bright stars, but looking the other way, over the natural horizon, the rising of Sirius might have been seen. The line is not securely known, and trees might have been a factor, but all plausible directions indicate a date within a couple of centuries of 4000 BC.
FIG. 24. Sections of the ditches at the Horslip long barrow, with labels following those of Fig. 23. The lines of sight are arrived at entirely on astronomical grounds. The verticals represent the height of the eye of a male observer of average height. The broken horizontals show the excellent levelling of observing positions, which ditches purely for quarrying purposes would not have had. The maximum height of the barrow can be judged easily enough from the suggested lines of sight, and is not indicated. The broken lines of sight are what might have been used had the ditches changed direction here, for example with a lower barrow. Note that observers’ heights may still be equal.
The orientation of the barrow can be estimated from the average direction of the inner edges of the ditches (135.0° for the northeastern and 140.8° for the southwestern), or from an interesting pair of pits (134.5° through their centres) that might well have been related to the line of the ridge itself, or perhaps to that of an earlier mortuary house on the site, as at Wayland’s Smithy. The pit near the northwest was roughly a metre square and three quarters of a metre deep; the other, about a third of the way up the ridge from the southeastern end, was circular in section and had dimensions about a third as great as the first. Other pits in a rough line following the southwestern edge are also represented in two cross-sections of the ditches, illustrated in Fig. 24.
Even from a single ditch, there are possibilities of multiple viewing positions. It is conceivable that the pits (here at b and d) are meant to add more, but the spread of a reasonably stable mound, of a height to be suggested shortly, makes it more probable that they are vestiges of an earlier phase in the history of the barrow, perhaps relating to the two pits on the axis. The ditches are interesting for their equalization of the lower levels, the places which on our basic assumption offer the ideal positions for viewing. Section E appears to be an aberration, failing to conform to our principles, and yet in the EF section there is a ledge providing viewing from the very same level. A change of ditch and barrow directions here could well have changed the requirements for viewing heights, to fit a later date, but still, it seems, the possibility of viewing at equal angles is being retained. The height of the spine of the barrow above the old ground level was perhaps around 3 m. The evidence for this comes from the astronomical possibilities offered by the lines of the ditches.
Applying yet again the principle of viewing at equal altitudes, at right angles to barrow and ditch edges, near or far, there is only one likely star to the north, namely the rising Deneb, but to south there is a choice of two, Aldebaran and Betelgeuse. The setting of Betelgeuse seems preferable, since it produces dates fully consistent with the radiocarbon date—which came from a piece of horn of red deer from a ditch fill. Taking the option of right angles to near edges, the date is 3809 BC (altitude 16.21°), while far edges give 3940 BC (altitude 12.26°). Coupling Aldebaran and Deneb, the angles would have been little changed, but the date would have moved earlier by three and a half centuries. Out of these four options, 3940 BC seems most acceptable. (The latitude is 51° 25' 58". Azimuths are taken as 50.8° and 224.7° (near) or 44.7° and 230.8° (far). The favoured declinations are: Deneb 36.76° and Betelgeuse –12.69°.)
In the case of the Horslip barrow, there is an odd piece of evidence supporting our preference. There is no chamfered inner edge to the ditch to suggest a viewing altitude, but if one tries to superimpose lines of sight on the ditch sections (as in Fig. 24) the floors of the ditches turn out to be better suited to the lower angle of view (12.3° rather than 16.2°), in the sense that they can accommodate more usable viewing positions. The sizes of the excavated ditches therefore give a rough idea of the height of the barrow, twice over, since they also provided the materials for it. They are quite consistent with the implied height of around 3 m above eye level.
The finer points of barrow construction, as illustrated by Fussell’s Lodge and Wayland’s Smithy, are at Horslip completely beyond recall. As already suggested, there was very probably an alignment down the line of the barrow’s axis on the setting of Sirius over natural ground (the declination being around –26.2°). This in itself would have been notable enough, in combination with the alignments on Deneb and Betelgeuse, but in the thirty-ninth century BC the Horslip site had another very remarkable property: the natural horizons there are not equal, but are such that the direction of the rising of Sirius was then at right angles to that of its setting. (Without tree cover the extinction angle was operative to the southeast but the altitude to the southwest was 1.26°.) In view of the Neolithic preoccupation with right-angled viewing, this would have endowed the site with exceptional importance, even though the right-angle property was relatively short-lived, and within a century or so was to be lost. But was it matched by another, which persisted?
It will be shown later in this chapter that the planet Venus, the brightest of the planets, might just have been seen when she was at her southern limit of setting, along the same direction as Sirius. The planet would have had this property already in the fifth millennium, and would have kept it long after it ceased to hold for Sirius. While it is a property to which one is led by considering the barrow, it is really one that belongs to the site. There are later qualifications to be made, in regard to a fundamental difference between Venus’ extremes of rising and setting, but assuming that the property was discovered at the time to apply to Sirius and to one or more of the four extremes of Venus (risings, settings, north or south), there would have been ample reason for building a barrow along one of the key directions. It is just conceivable that this property influenced the choice of a site for the Windmill Hill settlement in the first place.
West Kennet and its Star Chambers
Today almost as well known as Wayland’s Smithy, the chambered long barrow on a ridge of the chalk downs to the south of the hamlet of West Kennet in Wiltshire is even grander in scale. Just over a hundred metres in length, and tapering from about 24 m at its eastern façade to 12 m at the western end, it is (with that at nearby East Kennet) the largest in the region, and among the largest in the country. Its sheer size has enabled it to survive a number of crude onslaughts. One of these was by a certain Dr R. Toope of Marlborough, who in 1685 raided it for bones, for their supposed pharmaceutical properties. John Aubrey gave a vague account of the tomb at about the same time. It is in a letter to him from Toope that we learn of the doctor’s attitude to prehistory, and of how already in 1678, alerted to the large quantities of bones in the area around the Sanctuary nearby, he had acquired ‘many bushels’ with which he had made ‘a noble medicine’ that relieved many of his distressed neighbours. (One might doubt whether those particular bones, lying near the surface and closely packed in the open field, skull to skull, were prehistoric at all, although Toope did insist that their feet were all directed to the Sanctuary.)
William Stukeley made valuable drawings of the West Kennet long barrow, and described it, in 1723–4. ‘It stands east and west’, he wrote, ‘pointing to the dragon’s head on Overton-hill.’ (This is the hill on which stood the Sanctuary, now known to have comprised a succession of concentric rings of timber posts, before they were replaced by rings of stones.) The long barrow was partially excavated in 1859 by John Thurnam, who reported that farmers had cut a wagon-road through it, and had raided it for flints and chalk rubble. The definitive excavations were done in 1955–6 by Stuart Piggott and Richard Atkinson, like those later at Wayland’s Smithy.
The West Kennet barrow has much to tell us about the evolution of a single long barrow, but it shows, too, that it is a mistake to consider that changes in ditch direction necessarily imply additions to an old barrow at a later time, as is often supposed. It will be discovered that it was probably built in two distinct phases, a century or so apart, but that there is a change in direction within the first barrow. Quite apart from its intrinsic interest, this long barrow has an importance deriving from the fact that it is in the neighbourhood of an extraordinary complex of prehistoric remains. Silbury Hill, the largest man-made mound of prehistoric Europe, is less than a kilometre away (NNW), and less than 5 km in the same direction is Windmill Hill. Roughly 1.3 km east of the barrow lies the site known as the Sanctuary, and from there the Kennet Avenue of stones stretched, initially roughly westwards, then turning closer to north and heading for the southern entrance to the stone circle at Avebury. (The Avenue will be discussed again in Chapter 5.) Avebury, by far the largest of all British stone circles, is only a little over 2 km to the north of the West Kennet barrow. The almost equally large long barrow at East Kennet is just under 1.5 km to the southeast. On most days, all of these places are to be seen from the West Kennet barrow.
The barrow is flanked by ditches, running alongside it for all of its length. Its orientation was said by Stuart Piggott, in his report of the excavations, to be 265° 20'—that is to say, a person looking at the façade would be facing 4° 40' south of west. To define an axis, however, one must be able to identify the intentions of the builders, and this figure presumably rests on the assumption that the stones of the central passage were set symmetrically about an axis that agrees with the spine of the mound. Another interpretation, to be offered shortly, substitutes a figure of 267.0°.
FIG. 25. Sections of the West Kennet long barrow, as drawn by Stuart Piggott.
The mound had a central core in the form of a long cairn of sarsen boulders, rising to about 2 m in the excavated section. This core was in turn overlaid with chalk rubble from the ditches, and at the higher end this virtually doubled the height. A kerb round the barrow, resembling that at Wayland’s Smithy, was robbed of its stones long ago, and consequently a much less accurate picture is available of the outline of the tomb. As for the difficulty of assessing the orientations of the sides, their sheer length makes this a little easier, compensating to some extent for the loss of the kerbstones. The northern ditch apparently changes direction so that it is in at least four sections, while the southern ditch is in at least two. The bounding edge of the mound seems to follow those changes closely (Fig. 26). Viewing across the barrow, it will be shown that section A was paired with B, but not with C, which is a linking section only, while D was matched with both E and F, but at times a century apart.
FIG. 26. Outline plan of the West Kennet long barrow and ditches, showing the deliberate changes of direction, resulting in four sections to the north but only two to the south. (The inset scale at upper left is 10 m long.)
The most obvious stone structures, now largely restored, are at the eastern end, built out of massive local sarsen stones and dry-stone walling of oolite slabs. The sarsens were brought up the hill perhaps 2 km or more from the southeast, and the smaller stones from a distance of 10 km or more. The chambers form a ‘cross of Lorraine’. They are arranged around a roofed passage running for more than 8 m down the spine of the barrow: a burial chamber opens off it at its end, lengthening it to 12 m, while two others open off at each side. The passage is as high in places as 2.3 m. It is entered from a crescent-shaped forecourt formed out of large but rather irregular uprights, and a false front closes off the forecourt, making the façade seem more or less flat. There are two stones flanking the entrance, immediately behind the large blocking stone in the middle of the façade. They should not be dismissed as merely helping to seal the entrance, for they have another important function. As will be seen from Fig. 28 (from which blocking stones are removed), they help to limit lines of sight from particular points within the five chambers. These lines of sight—of which more shortly—seem to have been directed to five significant stars.
FIG. 27. The five chambers at the eastern end of the West Kennet barrow, with surviving blocking stones in position.
FIG. 28. The West Kennet chambers with blocking stones removed, showing potential lines of sight, limited by the edges and faces of stones. The other lines added to the figure (see also Fig. 29), show the great care that went into the planning, and justify the position given to the axis.
As for the number of burials, the solicitous Dr Toope missed the bones of at least forty-six individuals, for he—and even Thurnam—found only the chamber at the western end. The others were closed with blocking stones. As at other Neolithic sites, the skeletons were often incomplete. Some have held that this implies a systematic removal of certain bones—especially skulls and long bones—but on this point opinion is divided. The adult men seem to have been between 1.57 and 1.73 or even 1.80 m in height, the women from 1.49 to 1.64 m. The characteristically round skulls of the Beaker peoples are not represented, although their pottery is found among what are presumed to have been later ritual offerings. They were perhaps well represented in Dr Toope’s medicine, since the western chamber was the most readily accessible.
FIG. 29. The surprisingly systematic construction lines, abstracted from the previous figure, defined by the faces of stones. Close parallels and perpendiculars are indicated by similar marks. On this basis it is easy to make conjectures concerning slight changes of plan, forced on the builders by their materials, terrestrial and celestial. Note that four stones are almost perfectly in line with the axis, whilst five stones (including the three marked with a circle) are parallel to ditch sections.
In time, the chambers were filled to the roof with chalk rubble, stone and earth. Numerous bone objects and flints were found in the modern excavation, as well as pottery that included early Windmill Hill ware, Peterborough ware, grooved ware (Rinyo–Clacton), and Beaker types. This pottery shows that the tomb was in use for perhaps more than a thousand years, and deposits of Romano-British ware show that some sort of interest in it continued for well over three millennia. (Much material from the tomb is displayed in the Devizes Museum.) Long after the tomb ceased to be used for burial, it must have continued to function as a focal point for religious practice—much as do the cathedrals of the Christian era quite regardless of the fact that they too are the sites of burials. If it is hard to envisage such loyal devotion to a low ridge of earth and stone, that is only because we are ignorant of the systems of belief preserved by those who revered it. At what stage the enormous blocking stone 3.7 m high was placed across the façade it is impossible to say, but this must have come last, if the lines of sight of which we have spoken were precisely that, and not theoretically drawn rays, falling on the sightless dead. The massive stone sealed off the tomb’s contents, but even then did not mark an end to its ritual function.
FIG. 30. Suggested profile of the original West Kennet long barrow, looking across it from the south. The barrow might have been of constant height over even more of its length. The ground rises very slowly to the west and falls to the east. The two lowest lines are conjectural levels of the observer’s eye and the ditch floor.
The barrow tapers off towards its western end and loses height in relation to ground level, but the ground rises gradually to the west. The profile of the barrow is now degraded, but in its prime was probably not very different from that shown in Fig. 30. Its present contours suggest that there was originally a well-defined ridge. This is borne out by Thurnam’s statement: ‘Dr Took, as they call him, has miserably defaced South Long Barrow by digging half the length of it. It was most neatly smoothed up to a sharp ridge.’ One assumes that there is intended irony in the substitution of ‘Took’ for ‘Toope’.
The rays capable of entering the five chambers do so in much the same way as the pair of rays (from the stars in the Southern Cross) at Wayland’s Smithy, being limited to a greater or lesser degree by the uprights. Those passing to the southern and northern parts of the central chamber (W and X, Fig. 28) have slightly more latitude than the others, but if it is assumed that at least two stones must always be involved in fixing the direction, even they are tightly governed. Most of the key uprights are carefully worked, and in this respect alone can be distinguished from the rest. In all cases a normal adult male could just have stood upright to see the rising star appropriate to the chamber, and there is no question of this being an imagined possibility only for the deceased lying on the ground. The horizon over which all these stars rose is therefore here assumed to be the natural horizon, so that extinction angles hold good.
It seems that the axis was not directed to any star, but that at least three and possibly four of the chambers were, and no doubt consciously. There is some uncertainty in the precise original positions of the stones, which are critical, and the following brief statement is made solely with a view to a date found in the following section. The alignments proposed are all valid for dates in the neighbourhood of 3625 BC, for rising stars (with azimuth, assumed extinction altitude, declination and year in parentheses): Spica, in chambers U and V (63.7°, 2.1°, 17.53°, 3630); the Pleiades in chamber Y (101.8°, 4.4°, –3.99°, 3640); and Betelgeuse in chamber Z (109.9°, 2.0°, –10.85°, 3610). It is conceivable that beta Tauri12 was to be seen at W and Antares13 at X in the western chamber.
FIG. 31. Section of the northern ditch of the West Kennet Long Barrow near the end of the section marked B in Fig. 26. An adult male (eye at M) of average stature would have stood in the ditch bottom to see the brief appearance of Sirius over the tomb. A woman or shorter person (eye at W) or even a child (eye at C) could in principle have seen the same phenomenon. Of the two rays marked, the upper is at 13.5° to the horizontal, and the lower at 12.5°.
FIG. 32. Years at which the two stars Sirius and Arcturus could have been seen at right angles to the various sections (A to F) of the West Kennet barrow, plotted against altitude. The years corresponding to equal altitude viewing in opposite directions are marked with circles. Azimuths from north are marked on the graphs. The broken lines in the neighbourhood of points 1 and 3 represent shifts in azimuth of half a degree.
It is assumed here that the star Spica entered chambers U and V, although once the chambers were surrounded by the rubble of the mound, the ray could not pass into V at all. At the time of foundation it might have been allowed free passage, and in entertaining this possibility one is reminded of Wayland’s Smithy, where there was a strong presumption that the vertical stones of the chamber were set up astronomically before the building of the mound was begun around it, and without any hope of astronomical use thereafter. (The capping stones were presumably—but not necessarily—dragged up some sort of inclined plane, but even this need not have been the mound itself.) Surely the same is true at West Kennet. Several of the stones there are notable for their relatively flat faces, and the very fact that there are so many good parallel lines in the plan of the chamber might be thought to hint at alignment on common stars. In fact three stones (chambers U and V) seem to align on the setting alpha Centauri, three stones (chamber W and the western chamber) on the setting Procyon, and three (chambers Z and Y) on the setting Deneb. All this holds good for dates within two centuries of 3600 BC; and there are potential alignments too on Spica, Bellatrix, and Antares. The greatest problem here is that it is impossible to quote azimuths without an inordinate number of qualifications: the stones are at most a couple of metres across, they have been moved slightly in their long history, and in some cases their hidden sides were possibly those used in aligning them. Uncertainties of two or three degrees (even of one degree, taking averages) leave too much room for doubt over specifics. That the plan of the chamber has a definite geometrical rationale can hardly be denied, however; and we are certainly not short of independent evidence that Neolithic geometry and stellar astronomy were closely allied.
There is good reason for assuming that viewing would have taken place across the West Kennet barrow, in accordance with the double principle of viewing at right angles to the barrow—here closely parallel to the line of the nearby ditch—at equal altitudes from opposite sides. Any lingering doubts as to the broad correctness of these ideas should fade as the pairing of the ditch-sections stipulated earlier is investigated. As before, in all cases an initial search is made for possible stars, within the limits set by the barrow’s form and age. There can be little doubt that the star observed across sections A and D looking north was Arcturus, while looking south across any of the four northern sections, Sirius was most probably the main object of attention. (Rigel is a candidate that initially falls by the wayside, since it seems to provide dates between 3222 and 3057 BC. These dates are much too late, if the radiocarbon dates available are to be believed.) Here are the two brightest stars in the sky.
In discussing similar cases at an earlier stage, little was said about the errors involved in the method. At West Kennet the situation is precarious, for the only part of the ditch so far excavated is close to being in the worst of all possible places, that is, near the change of direction from B to C (Fig. 26). Putting everything aside for the time being except the directions of the ditch (and matching barrow edge), all of them involving averaging over sizeable lengths, our principles lead to the following results—quoted with pseudo-precision, without any regard for their qualities. Viewing at right angles to edges near at hand (this will prove to be the likelier option):
FIG. 33. A supplement to Fig. 32, with the graphs for Rigel (across C) and Vega (across A), both observed by observers at ground level, but at the same period as Sirius and Arcturus from the ditches.
(1) across sections A and B, viewing at 13.5° in 3638 BC
(2) across sections A and C, viewing at 12.5° in 3447 BC
(3) across sections D and E, viewing at 13.3° in 3624 BC
(4) across sections D and F, viewing at 12.7° in 3512 BC.
Viewing at right angles to far edges:
(5) across sections A and B, viewing at 13.4° in 3677 BC
(6) across sections A and C, viewing at 14.5° in 3395 BC
(7) across sections D and E, viewing at 13.8° in 3677 BC
(8) across sections D and F, viewing at 14.5° in 3514 BC.
In the case of West Kennet it appears that the two different starting points make only about forty years difference to the final results at the head of the barrow, and almost none at the tail. There is a slightly more material difference in the sizes of the derived viewing angles, however. Combining the very limited evidence of the one known ditch section with the height of the mound opposite it, an angle of about 11.9° is obtained for that point. Since some slight settlement and loss of the surface layer can be assumed, any angle between 12° and 14° must be considered acceptable. The height of the observer’s eye is taken as 1.6 m, following the male skeleton sizes at the barrow itself. Even those of smaller stature could be accommodated, for as will be seen from Fig. 31, there was a viewing platform which could have been used by people of all heights above about 1.10 m—so perhaps even children were admitted to the ritual of observing the stars over the barrows. There is some later support for this intimation of democracy, in the ditches at Stonehenge, Woodhenge, Mount Pleasant, and elsewhere.
The later date is here preferred to the earlier. This preference is partly because the later dates fit better with the radiocarbon date (from three specimens of human bone from the barrow, giving a calendar range of 3575 ± 215 BC), and partly it implies marginally smaller viewing angles and a mound of reducing height in relation to the observer.
The head of the barrow seems to be a few years later than the third section. This need not be real, for it must be accepted that our azimuths are subject to error. The results spelled out above are shown graphically in Fig. 32, and broken lines in the neighbourhood of the points representing (1) and (3) are added to show the errors to be expected corresponding to errors in azimuth of half a degree. Roughly speaking, a single error of this amount moves the date by four years. A less superficial analysis might one day therefore easily reverse the order of building these sections.
There are two small residual problems. The year under (2), is obviously entirely spurious, as therefore is the altitude too. It would be interesting to find from future excavation how the altitude was handled in this region. Perhaps levels were changed to preserve the visibility of one star, or to substitute another. One possibility is that section C of the ditch was simply being made to change direction to link up with a section (E) that had already been—or was planned to be—cut with a different orientation from B’s. But why change direction at all? The answer had perhaps something to do with the wish to reduce the barrow’s height, and with the lie of the natural ground making up the tail of the barrow. The last date quoted, however, would seem to indicate that an addition was made to the mound, over a century after the first mound. The only other likely explanation is that viewing angles were not being kept equal, or equal enough.
There is an interesting alternative explanation, that might be extended to preserve the integrity of the entire barrow. Before explaining it, one must consider the way in which the ditches would have been used to observe Arcturus and Sirius. These stars would both have been seen behaving in a very striking and memorable way. At Wessex latitudes Arcturus did not rise and set at all over a level horizon; but here, looking from the southern ditch over the barrow, the star would have dipped down to touch the tomb, seeming to rest there, and then have risen again about half an hour afterwards. Some four and a half hours after that, Sirius would have risen over the tomb, as seen from the northern ditch, and again in much the same time would have sunk back into it at a point not far away.
The two brightest stars were not alone in these comings and goings. Vega would have copied the behaviour of Arcturus very closely, descending into the tomb and rising again, although as seen from the ditch it would have lacked the finesse of Arcturus. In fact the optimum altitude for observing Vega during the period in question (around 3620 BC) would have been a low angle of 7.8°, if the star was to be viewed at right angles to section A, as Arcturus had been viewed. Vega’s visitation would then have begun almost simultaneously with Sirius’ re-emergence.
On what grounds can it be claimed that this additional item of very striking behaviour was in fact observed? Quite simply on the grounds that the angle 7.8° is close to what should be expected—if the mound height conforms to the altitudes found from the ditch—for an observer standing at ground level near the back edge of ditch A.
Having gone so far, did the people of West Kennet leave matters in such an unsymmetrical state? Could they not find a fourth star, in the southern sky, to complete this remarkable pattern? They undoubtedly did so, and the star was Rigel. Rigel, like Sirius, rises only a little way out of the tomb and then quickly falls back in again. On this occasion an altitude of about 7.5° is needed, and this is so similar to the angle from the ground looking north that we know exactly what we must do. We must consider the possibility that equal altitude normal viewing took place from ground level across section A and the aberrant section C. The azimuths taken are 162.4° and 356.5°, and in the usual way an altitude and year are found—7.6° and 3610 BC—at which the arrangement would have been perfect (Fig. 33). It seems probable, therefore, that the ditches will eventually prove to have platforms at the back, at the same level, capable of yielding this common angle.
The anomalous date having been removed, we are now left with a series of dates. Rounding to tens they are (1) 3640, (2) 3610, (3) 3620, and (4) 3510 BC. There is no point in trying to eliminate the spread in the first three dates, in the absence of a more complete excavation. Those dates are so close that for want of better evidence we may take the date of the mound as their mean, 3620 BC. The fourth still seems to indicate a later phase.
The pattern of observing the stellar phenomena was therefore as follows: Arcturus from ditch A, Rigel14 from behind ditch C, Sirius from ditch B, Vega15 from behind ditch A. The seasonal limits during which these phenomena could have been seen vary from star to star, but all three could in principle have been seen on every clear night between roughly a fortnight before the autumnal equinox and a month before the winter solstice. The interval between the mid period of Arcturus’ visit, and the mid period of Vega’s visit the same night, was about six hours.
At this six-hour period, and in this way, the West Kennet people had evidently contrived a series of apparitions, involving four spectacularly well-chosen stars, of a character that might have been interpreted as somehow relating to the spirits of the dead. Only one star visible in Wessex was the equal of any of these four. (Capella was then Vega’s equal—it is now less bright—and both were brighter than Rigel. Canopus, second only to Sirius, never rose at these latitudes, and was unknown.) It would be hard to imagine a simpler monumental design showing more intellectual brilliance than this.
There is, in conclusion, a remote possibility that Sirius might have played a part in the selection of the West Kennet site many centuries before the present long barrow was built. It is a strange property of the entrance to the barrow that, from it, two low hills to the south are more or less symmetrically arranged, one 37.3° west of south, the other 38.8° east of south. Although these angles are not equal, neither are the altitudes. (With five-metre trees, for instance, the altitudes would have been 1.8° and 1.2° respectively.) This fact compensates almost perfectly in the case of Sirius, so that if it rose precisely over one hill it set precisely over the other. Without trees, this would have been so around the year 4460 BC. Trees would have brought the date a decade or two later. The hills are not prominent, and the early date must seem very improbable, but the possibility is a very curious one. Even stranger is the fact that at exactly the same period it fits the star Rigel Centauri16 (not to be confused with Rigel), which has the same declination as Sirius in the year 4450 BC. Whether or not this is a pure coincidence, it seems worth putting on record.
The same property fits the star beta Centauri in the forty-second century and Rigel in the thirty-fifth. The Rigel case could not explain the choice of site, although the other might.
The West Kennet Neighbourhood. Silbury Hill
The East Kennet barrow, the later of the two neighbours, is now tree-covered, and has not been excavated in recent times. Except in orientation, it seems to resemble its neighbour: it is only a metre or so longer, there are signs of a sarsen burial chamber, and there are traces of flanking ditches. The barrow as a whole looks as though it might have been aligned on the rising of Rigel. On the basis of its direction, about 145°, it might be tentatively placed not far in time from the West Kennet barrow. In view of the sophistication of the latter it would be foolish to draw any conclusion about a cult of Rigel. Such might have lasted here until the building of the artificial mound at nearby Silbury Hill, however, for the star would have been visible from the top of that mound, rising over the East Kennet barrow in the late third millennium. Against this idea is the fact that the first stages of the construction of the hill had begun centuries before—but in the mid-fourth millennium, Sirius would have had the same property. There are too many imponderables about the height reached at various times for any specific claim to be made.
FIG. 34. The character of the internal structure of Silbury Hill.
It seems likely that when two monuments were involved in a stellar alignment, the star was seen from the newer and over the older of the two. (The converse arrangement would have been perfectly possible, of course.) The principle seems to be illustrated at West Kennet. The Sanctuary, a succession of concentric circles, first of timber and later of stone, was to the east of the long barrow, across the valley of the river Kennet, and on its own small hill. (Its name was one used by local people, according to Stukeley.) Robbed of its stone in the eighteenth century, the site was rediscovered and excavated in 1930 by Maud Cunnington. As seen from the centre of that succession of circles, the West Kennet barrow would have been in a direction about 8.5° south of west, and at about 0.5° altitude, slightly less than that of the spur of the hill behind the barrow. Aldebaran could have been seen setting over the barrow when the star’s declination was about –3.96° (taking its extinction angle of 2.0° as the deciding altitude). This was its value in approximately 2800 BC, a date compatible with other archaeological evidence relating to the early phases of the Sanctuary. There is a pair of post holes defining a diameter of the oldest structure (Sanctuary IA) in exactly the required line.
The most impressive of the monuments immediately surrounding the West Kennet barrow is Silbury Hill. Built on a spur of chalkland, its base is nearly 30 m below the level of the barrow, and yet the Hill rises 10 m or so higher. (Its top is currently about 3 m lower than the East Kennet barrow.) It is a cone, with steepest slope close to 30°—a gradient that was perhaps chosen deliberately for its ‘one in two’ property, one unit rise to two up the slope. Several exploratory tunnels have been dug into the hill, the latest of them by Richard Atkinson (1968–70), who suggested four phases in its construction. After the first two phases had been completed, the hill was still not particularly remarkable for its height. It is probable that there was never any primary burial there. Certainly none has ever been found, although some have suspected that the shaft from an eighteenth-century excavation destroyed signs of one. The third phase required a change of plan, resulting in a stepped cone. The stepping can still be easily seen by the casual observer. The last phase was an extension of the main ditch to the west, presumed to be for chalk to fill in the steps. The overall construction involved moving well over 300,000 cubic metres of chalk—less than would have been needed had part of its base not been the end of the spur of the natural hill—and those responsible for the third phase especially knew much about soil mechanics, for they gave it great stability through a series of internal walls, built out of chalk blocks, the resulting cells being then filled with rubble (see Fig. 34).
What was the final purpose of this extraordinary structure? Since it is thought to have been four or five centuries in the making, one need not suppose that its purpose remained unchanged. Attempts have been made to find an astronomical explanation for it. In 1902 the American writer Moses B. Cotsworth, in his Rational Almanack, suggested that it carried a colossal gnomon to cast shadows on the world’s largest sundial, marked out on the ground below. He was evidently influenced by the Rev. Edward Duke, who in 1846 had made it out to be the centre of a vast planetarium on which the planets were represented by ‘temples’ in the neighbourhood: Stonehenge was Saturn, Avebury was the Sun and Moon, and so on. One must be charitable to all enthusiasts, in the hope of being treated likewise. Sundials and planetaria are really no more exotic than the vision of one respected contemporary archaeologist, Marija Gimbutas, who sees the hill as ‘a gigantic representation of the Pregnant [Earth Mother] Goddess in a seated position’.17
There is a range of Silbury radiocarbon dates running from 2725 ± 110 down to 2145 ± 95 bc. The extremes of this range correspond to 3630 and 2500 BC in calendar years. Building probably began a century or so after the West Kennet barrow, an undertaking by the same group of people.
From Silbury Hill, the West Kennet barrow is nearer and more prominent than that at East Kennet. The place where it stands is such that looking over the head of the future West Kennet barrow from a point at the centre of Silbury Hill, Rigel could have been seen rising in the thirty-ninth century. The place of rising would gradually have moved to the forecourt area, and eventually off the scene completely, but at the date accepted here for the foundation of the present barrow, namely 3625 BC, beta Centauri rose precisely over the chamber as seen from the same point. An earlier long barrow near Windmill Hill (Horslip) is more or less on the same line, and the East Kennet barrow is near it too, so that many further possibilities are added to the list—for example, Rigel rose over the place of the future West Kennet chamber as seen from the Horslip barrow in 3710 BC. It might be thought that when barrows are not intimately related—for instance directed one towards another, as at Fussell’s Lodge—such alignments are best ignored. That they are part of a conscious strategy covering a much wider territory, however, will be demonstrated in the following chapter. (Maps will also be found there and in Chapter 6 showing the principal long barrows and other prehistoric remains in the West Kennet region. See Figs. 69 and 101.)
Beckhampton Road. Stars and the Sun
The long barrow off Beckhampton Road lies less than 4 km west of that at West Kennet. As mentioned earlier, it is one of an important trio excavated between 1959 and 1967—in this case in 1964 under the direction of I. F. Smith. It was described by William Stukeley in 1743 as ‘pyriform’ (pear-shaped), and ‘longish, but broad at one end’. It had been over 50 m long, and was originally flanked by splayed ditches, with its broad end roughly northeast. That end had been transformed when a round barrow was superimposed on it, millennia before nineteenth-century farmers levelled and ploughed the whole thing. In that same century it was dug into by archaeologists too. One of them, John Thurnam, recruited his labour from among his patients at the Wiltshire County Asylum in Devizes. Despite all of this activity, the ground beneath the barrow retained a few secrets of immense interest, to be recovered in the modern excavations.
Especially important were signs of the care with which directional properties had been assigned to the barrow. A row of stake holes was found to run along the axis, changing direction abruptly in a middle section, before returning to the original direction. Offset from this bent axis were traces of further straight rows of stake holes, making for at least twenty bays at the wider end. The staking out of some of the bays on the northern side, where the round barrow had been superimposed, and probably that of others at the narrower end, have been lost. Other rows of stake holes mark the edges of the mound, and those to the southeast side make an extremely well-defined line. Some rows were pushed out of true by the weight of soil, although much care had obviously gone into establishing them in the first place. Generally speaking, stake holes had a depth of from 15 to 60 cm, and the stakes, of diameters up to about 8 cm, had been pointed, so they must have been hammered into the ground. Much the same procedures were followed at the South Street barrow.
In Fig. 35, the better-defined rows of stake holes have been indicated, that is, when they are of reasonable length and have clearly not suffered distortion through soil pressure. (To see the individual holes, the original report will have to be consulted.) The excavation brought to light even more valuable structural evidence than this, in the form of hollow casts representing rods and poles that had been used in various ways and had later rotted. Traces of poles were found to the side of each row. They had clearly been tied in some way to the uprights, certainly not to hold back the material of the mound, but surely to fix a straight line. The line was important because it was what settled the precise form of the barrow, and on the assumptions made here that form was intimately related to a line of sight. From signs of accidental spillage of material during the construction it was clear to the excavators that the bays, or opposite pairs of bays, were built one at a time.
Bays such as those at Beckhampton Road are reminiscent of others in barrows at Sarnowo (Poland), Skendleby (Lincolnshire), Ascott-under-Wychwood (Oxfordshire), and the neighbouring South Street barrow. Paul Ashbee has noted, furthermore, that various Cotswold long barrows dug into in the nineteenth century were found to be split axially and into bays with stones, and that stake holes might there have been missed. The Beckhampton Road barrow is an important guide to the meaning of these bays, and it is no exaggeration to describe them as an astronomical plan, analogous to the scaling posts of the barrows at Fussell’s Lodge and Wayland’s Smithy.
The stake holes constituting the bent axis of the barrow begin and end (rr) at two almost perfectly identical azimuths, paralleled at the northeastern end by the bounding rows p and t. The accuracy with which these are laid out is noteworthy: averaged over their lengths, the mean azimuth is 49.8°. It gives great satisfaction to find that the transverse lines of stake holes turn out to be perpendicular to one side or the other, for this is exactly what our principle of perpendicular viewing at equal altitudes leads us to expect. As explained earlier in this chapter, however, it is chastening to find that the b-rows are at right angles to features on the opposite edge of the mound, judging by row q to the distant side. There can be no doubt that viewing was along the lines of the b-rows from the southeastern ditch (sector EG at the very least), and that viewing was along the lines of rows a looking southeast from ditch sector AB and perhaps BC. The mound edges seem to be perversely arranged, but they matter less than the ridges. The a-rows are well suited to row r, but not to row q, and the b-rows to row q (where they are not to be found) but not to row r.
FIG. 35. The overall structure of the Beckhampton Road long barrow, as indicated by the ditches, the approximate edges of the mound (coarse broken lines) and selected rows of stake holes (continuous lines within the mound). A few individual stake holes are marked at the northeast. These are joined by broken lines (x and y). Note the precision with which they cross on the axis, and note also the merest hint of a scaling-post construction near the letter x. Where transverse stake hole rows are close to parallel, they are labelled with the same letter: there are only three clear groups. Judging by consistency alone, the estimated edges of the ditch and the mound are plainly less reliable as astronomical indicators than the stake hole rows. The ditches are divided into sectors (see letters A to K) according to supposed astronomical function. Ox skulls were found at the points marked.
Does this not fit uncomfortably with the principle of perpendicular viewing? Not at all. The axial rows cannot be taken as a simple ridge line. The supposed lines of sight would have been tangential to the mound below the ridge, unless the spine came to a sharp edge. Any axial ridge would not have been strictly seen, and would have been slightly flattened, if only by weathering, so that the effective ridge would have been double. The angle of view on this occasion was almost certainly about 15.3°. The reasoning follows the principles laid down earlier. Starting from the azimuths of a and b (both derived with the help of longer rows t and s but not the less reliable edge of a turves line e), stars at equal altitudes are sought. Ditch depths are compatible with this idea, although further excavation of them would not come amiss. From azimuths 135.8° and 309.8° and a latitude of 51° 24’ 28", it seems that the setting of Deneb was observed to the northwest and the rising of Bellatrix to the southeast. A strict application of previous principles, overlooking imperfections in their implementation, provides a common altitude 15.32° and a year of 3320 BC. There are many imponderables here, but the date is probably accurate to better than a century. (The altitude is finely tuned, since Deneb’s declination changes so slowly, and no other bright star offers itself.) As for the date, it is perhaps not without significance that the mid-section of the barrow (row q) aligned accurately on the setting of Rigel around 3320 BC, over the natural horizon.
Stukeley’s ‘pyriform’ barrow was so only because it had been deformed by the later addition of a round barrow to the wider end. It is clear that it originally had a trapezoidal form, and since the ditches extend so far to the southwest, we must suppose that the barrow ended in the neighbourhood of Z (Fig. 35). The directions of lines x and y were clearly those of a chevron-shaped façade. How this was used is less clear, but over natural ground, Pollux set in the direction of x (around 3360 BC) and beta Centauri rose in the direction of y in 3380 BC—or say a century earlier with a cover of seven-metre trees.18
The lines of x and y meet almost precisely on the axis, at a point well fitted, therefore, for a triple observation, if only there were something to be seen along this part of the main axis. There was indeed something to be seen: the last glint of the setting midwinter Sun (that is, at the solstice) was along this line. Precise dating is impossible, but taking the azimuth as 226.2° (it differs from that of the westernmost part of the axis by half a degree) and the altitude of the horizon set by a treeless valley floor as 2.3°, the direction is virtually perfect for the period in question. Viewing could have been along the ridge itself, given a platform on which to stand, but it would have been easier to look along the edges p and t. Long barrows may bend, but never, as far as can be seen, in such a way as to block views along their edges at the head of the barrow.
This is a highly significant finding—the first Wessex long barrow considered here that clearly incorporated an alignment on the Sun, although there must surely be many more, since six or seven centuries earlier a Lincolnshire long barrow (Skendleby 2) had done the same, and a square, overtly humble, but cleverly designed barrow at Grendon in Northamptonshire had done so too, say a century before South Street. (They will be discussed later in this chapter.) Sirius and important stars in Orion were within a degree of aligning, but not to the standards of accuracy found previously. The alignments of later circular monuments were almost invariably on the Sun and Moon, and it seems probable that the Beckhampton Road long barrow was colonized by the builders of the round barrow at its head precisely because it aligned on midwinter sunset.
Three ox skulls were found more or less in the line of the ridge, which again gave rise to speculation that the ox heads, perhaps with skins and hooves attached, had been set up on posts penetrating the back of the barrow. There is no sign of Aldebaran here, but as it happens there is an independent case to be made out for a switch of bull symbolism to midwinter sunset.
The main disadvantage of the scaling posts of the old structures at Wayland’s Smithy and Fussell’s Lodge was that they required a transference of levels, not very difficult when structures were of wood, but more so with earthen mounds. In some way or another, a horizon had to be built up until it produced the desired effect—but strictly in accordance with the requirements of right-angled viewing. There would have been only minor difficulties with the progressive shaping of the artificial horizon, as the bays were filled in, one by one, until at last rising and setting were seen in the required directions. The main intellectual problem was quite different: it was that of settling the directions. To reconstruct the method it is necessary to make a conjecture as to the starting point. There is a strong sense in which Beckhampton Road is a Rigel barrow. Rigel is the brightest of the stars concerned, and the alignment on it belongs to the site in the pre-barrow period. Assuming, then, that this was the star from which the builders began, here is a potential procedure:
Some method or other is needed to establish the positions of a pair of ridge poles at equal heights and at right angles to the lines of view to opposed stars, northern and southern in this case. Once they were established, ditches could be marked out, if not extensively dug at this stage, with appropriate sections parallel to the ridge-poles. There was still a degree of freedom left to the architects, who would have tried to accommodate other stars in their length-wise barrow lines. The central portion of the line of stakes will be assumed to align on the setting of Rigel, and the other pair on Deneb and Bellatrix:
(1) The stake holes XX were arranged to be in line with Rigel’s setting.
(2) Fences of stakes, to which poles were tied as cross-pieces, were set at right angles to that first line, and a viewing ditch cut parallel to it, for sighting on Bellatrix. Horizontal poles were raised along the ridge line until the star rises at that level and in the direction set by the transverse stakes. Hides could have been draped over them to give an obscuring horizon. The lengths of such poles would have been between 1.5 m and 3 m, and their weight offers no problem. Infilling of ditch material to make a solid horizon might have begun at this stage—in slices almost the reverse of an archaeological dig.
(3) A provisional ditch was started for viewing Deneb. (Approximate positions were of course known all along.) Another ridge-pole, at precisely the same level, was varied in its orientation until Deneb was seen to set precisely at right angles to it. Fences of stakes at right angles to the Deneb ridge-pole could then have been set up to confirm the arrangement, with viewing directions again made more easily visible by poles tied along them. This established perhaps all remaining directions.
(4) Ditches could next have been completed and the fences largely filled in. An arch of spaced parallel poles, like stringers in the body of an aircraft, could have been tied to the rows of stakes, to define the barrow’s form, although the removal of the stakes from one section might have been thought necessary before an adjacent section was added. The curve of the arch might have been fixed by templates, curved perhaps like longbows, with tow. The builders’ job was to fill up the arch with the materials of the mound, and to dress it smooth in a suitable way—with marl, chalk gravel, and turves. Section by section the barrow was so assembled, its wedge-like shape guaranteed by the lines of the ditches.
FIG. 36. The overall plan of the South Street barrow. The three most important sections are combined into one, for convenience, at the right. They correspond to the broken lines A, B, and C on the plan. Lines of sight to the eyes of observers in possible viewing positions, at approximately the correct calculated angle, are shown on those sections—which are, however, not perfectly chosen for our present purposes. The diagonal line of stakes (D) crosses the entire barrow. The larger sarsens are marked on the plan (notably S), in their present positions. On section B, however, sarsen S is placed upright, as it might at one time have stood, even before the barrow.
Assuming angles derived from the same sort of astronomical analysis as used for earlier barrows, now for the setting Deneb in one direction and the rising Bellatrix in the other, the resulting profile of the barrow conforms closely with I. F. Smith’s estimates of the shape the barrow originally had. Pollen analysis showed that the barrow had been built in an area of grassland, but with arable and wooded areas nearby. This barrow was not on high ground, and the best of all possible motives for the relatively steep angles of view of those who looked across the barrow from the ditches was to observe the stars well clear of nearby trees—in this case 15 m trees further away than 75 m would have presented no problem except in the line of the barrow—that lay along an arable valley floor.
The last of the important trio of long barrows in this district is that off South Street in the parish of Avebury. It was erected in an area of arable land that had been cleared in the early fourth millennium BC. The surrounding country, predominantly woodland, contained hawthorn, oak, birch, elm, alder, pine, and other species, with all-pervasive hazel scrub invading the open pasture from time to time. It will be shown that this barrow was always extremely humble in outward appearance and material structure, and as an artificial horizon set rather low altitudes, but that it was high enough for sight-lines to clear nearby scrub and woodland. William Stukeley, in his book Abury (1743), described it as ‘broad and flat, as if sunk into the ground with age’; but broad and flat it had always been. Humble though it appeared, it embodied a most beautiful astronomical symmetry, and one that leaves us in no doubt that the people of Wessex were still actively developing their art.
The barrow was excavated under the direction of J. G. Evans in 1966–7. The road—South Street—runs across a corner of it, and the site has been much ravaged by time, not to mention proximity to the village. There are two large standing stones 120 m to the west of it, known as the Longstones, or Adam and Eve, but these belong to a much later period, as will be shown in connection with the Avebury avenues. Although there are no great quantities of stone in any of the three Avebury barrows, here at South Street there are several large and small sarsens at its principal end, and the larger stones were evidently set in place before the main construction, since all lines of stake holes curve round them where necessary. This barrow repeats in several respects the structure at Beckhampton Road, revealing a similarly straight row of stake holes defining an axis, and indeed more complete transverse rows than at that other barrow. The South Street barrow was divided by them into about 20 bays to each side of the axis. Again they were occasionally deformed in places by pressure of the mound, and again there were numerous traces of collapsed rods that had originally been fixed to them. There is a pair of ditches, uninterrupted in the South Street case, and splayed but not very markedly so. A simplified plan of the site as a whole is shown in Fig. 36, where rows of stake holes are represented by continuous lines (marking the axis and bays). Once again, those who dislike the element of approximation in all this may consult the original publication for the many hundreds of holes.
Applying the principle of viewing at equal altitudes, it is found that looking across the ditches, along the directions set by the bay-fences, Vega was observed rising in the northeast, and that Sirius was seen setting at an altitude of about 9° to the southwest. The date was in the neighbourhood of 3260 BC. A date seven centuries earlier, with Arcturus and Bellatrix, is rejected because it makes for an improbably high barrow, in fact over 5 m along the spine. The radiocarbon dates also speak against that solution.
This very general conclusion has to be qualified in several respects, since the large number of paired lines of stake holes means a large number of solutions, and not all are superficially in agreement. As Fig. 36 shows, these lines are by no means constant in direction, but it would of course be mistaken to take averages of azimuths, north and south. While—on our basic assumption—opposed altitudes are ideally equal, it need not be supposed that they were constant along the whole length of the ditches. Until the ditches are fully excavated we shall remain ignorant of the pattern of altitudes, but even now it can be said categorically that (1) the known ditches were almost perfectly equal in depth (in relation to the Ordnance Datum rather than surface levels) at the eastern end; and that (2) the ditch floor was somewhat lower there than in the neighbourhood of the point at which the compass rose is placed on Fig. 36. This second fact, that might at first sight seem to refute our basic principles, will on the contrary eventually serve to vindicate them.
From a naïve application of the principle of equal-altitude viewing, ten of the twenty pairs of stake hole rows are found to be usable to yield ten corresponding dates (and altitudes). The graphs giving a result for the first bay (actually based on the second pair of fences) are shown on the right-hand side of Fig. 37. The date derived is 3260 BC and the altitude 9.42°. Until such time as further information on the ditches becomes available, this result will be the most reliable of the ten, since the corresponding observers are known to have stood at equal depths. Performing the same calculation for all ten of the straightest pairs of rows of stake holes, one finds a mean result of 3220 BC, but this is certainly less reliable than the first result, since it is known that not all corresponding points in the ditches further west are at exactly equal levels.
FIG. 37. The dating of the South Street barrow. On the right, the graphs for the setting of Sirius as seen in line with the first bay from the north ditch and the rising of Vega as seen similarly from the south, meet opposite the year 3260 and altitude 9.42°. On the left is the evidence for the sighting of Regulus and Bellatrix along the line of the diagonal row of stake posts at about the same time, but now at altitudes differing by about 0.92°. On the case of Capella, see the text.
The date of 3260 BC, probably good to within half a century, fits well with two out of four radiocarbon dates obtained from the site. The four, briefly, are these: from oak under the mound (2810 ± 130 bc, equivalent to 3535 ± 165 BC); from an ox vertebra (2750 ± 135 bc, 3495 ± 150 BC); from a red deer antler, north ditch (2670 ± 140 bc, 3360 ± 260 BC); from another, second bay of the mound (2580 ± 110 bc, 3190 ± 150 BC).
Without a knowledge of precisely where the observer stood, it is impossible to give a precise height for the barrow at the eastern end, but it was very close indeed to the height of a man’s eye above the carefully levelled ground on which the mound was erected (see sections at the right hand side of Fig. 36). This means that it was just possible to look over the barrow from its façade, for instance, in ways that might have been indulged in before the barrow was built.
The calculated height in question is so close to the length of the sarsen marked S on the figure that it is tempting to suppose that it originally stood upright. It needed no significant stone hole, since it would have been held up by the chalk of the mound. This was not the opinion of the excavator, J. G. Evans, however, who noted that the mound material gave no evidence of collapse. He noted too that the three largest had been prepared for the positions in which they were found by having a slice of stone removed. We might have been inclined to see in S not a sighting stone but a stone that was meant to remain only just covered by the mound. Oddly enough, although it is too low for the mound in its prone position, it could have served this very purpose at an earlier stage, when a still lower angle had to be set by the mound. (This lower angle will be discussed later.) Attrition of such carefully prepared mounds by rain, wind, and invading plants must have been a serious problem, and regular shaping must surely have taken place.
Other sarsens shown on our figure seem to have had a part to play in the planning of the diagonal D, and the spine as far as the third and fourth bays. The idiosyncratic line of stakes marked E in the diagram was very probably a simple alignment on the rising Deneb—but this explanation must be abandoned if the southern ditch eventually turns out not to be unusually shallow.
Had there been only the transverse rows and the ditches, South Street would have added nothing to what has already been discovered about long-barrow astronomical practice, but a most unusual aspect of this barrow lies precisely in that very straight row (D) of stake holes that ran diagonally across it. As emphasized by the placing of the compass rose in the figure, this line has a most beautiful symmetry with the axial line of stake holes, for the angle between them is almost perfectly bisected by a north–south line.
There are two possible explanations for this diagonal, one of them almost certainly correct, the other perhaps true of an earlier period of history. If, before the barrow was built, observations had been made at the same horizon altitudes along the two lines (axis and diagonal), both to the south or both to the north, then the symmetry implies that the same star would have been seen rising along one line and setting along the other. In the case of the South Street barrow, horizon altitudes were far from being equal in any pair of the four key directions, but this offers no problem, for the stars to be seen rising and setting in this way would have been seen at their extinction angles—Pollux to the north around 3600 BC or Bellatrix to the south around 3920 BC. In either case, therefore, one could have expected perfect symmetry with respect to the north–south line. As it happens, the Pollux date falls comfortably within the range of radiocarbon dates obtained from the piece of oak found under the mound. This is surely not the date of the mound as it is now known, or of the diagonal row of stakes, but it is quite possible that the direction was embodied in some nearby material structure—say a pair of stones or another barrow. In fact at a distance of only 300 m the line runs along the southern edge of another long barrow (one of three parallel barrows northwest of the Beckhampton roundabout).
Whatever the truth of this explanation, why the diagonal line was incorporated into the barrow itself is of more immediate interest, and can be explained in a different way, not necessarily ruling out the first. It is difficult to give a precise argument for want of data from the ditches, but the rising of Regulus could have been observed along it to the northeast and the setting of Bellatrix to the southwest. Assuming that the line was regarded as very important, the entire mound could have been designed around it, with these new stars in mind. Perhaps somewhere else in the neighbourhood, or perhaps on this very site, these two stars were regularly observed along this line at equal altitudes (just over 6°) in the thirty-fifth century (point T on Fig. 37). Had this been done at South Street—and the radiocarbon dates tell us that this was not impossible—the mound would have been lower then, or the ditches shallower, or both; and then as the years passed, the direction no longer functioned as it had done, and it became necessary to look south at a higher angle than north in order to keep the old direction.
By 3260 BC the difference was about 0.92°. Had Capella and Bellatrix been taken as the pair of stars, it would have been found that there was never a time when they could be seen at equal altitudes along the diagonal. As the figure shows, they were much closer, however. The reason for opting for the larger separation is to be found in the ditches. It is not difficult to calculate that no matter what the height of the mound, within reason, the required difference in viewing angles would be provided if the observer looking north were standing about 33 cm higher than the observer looking south. The precise value depends on where in the ditches the observers stood, but on the evidence at present available it seems that there was probably a difference of levels somewhere between 25 and 40 cm. The required viewing altitudes for the year 3260 BC (see the figure) imply a barrow that would have been about 33 cm lower at the crossing point of the diagonal than at the barrow’s eastern end. This is precisely a quarter of the way along its length of 36.8 m as judged by the bays—note that it comes after five out of twenty bays. If it sloped regularly, therefore, the barrow would have fallen from 1.66 m at its head to 33 cm in the last bay, measuring above the base so carefully prepared for it.
These proportions are unlikely to have happened by accident, and one cannot help wondering whether the derived viewing altitude across the first bay, estimated at 9.42°, was not likewise deliberately chosen as a gradient of 1 in 6 (or 9.46°). There is much evidence from a later date for the use of such simple ratios. One can only make this assumption freely, on the other hand, if there was no other constraint, and there can be little doubt that there was one very important constraint accepted by the designers. The first four rows of stake holes on the northern side were almost certainly meant to be exactly at right angles to the axis—the average angle was around 89.8° even at the time of the excavation. This was surely the builders’ second precondition—the first being acceptance of the diagonal. It is unlikely that they waited until the old Bellatrix line (the axis) was at a perfect right angle to a one-in-six Sirius line, before setting up the barrow. Had they done so, however, they would only have been playing yet another round in a science of barrow-symmetry and precise alignment that one suspects was becoming an end in itself.
Skendleby: Giants’ Hills 1 and 2. Stars and Sun
While the focus of attention is Wessex practice, there are other important long barrows that are worth mentioning since they reveal so much in common with it. Two of these are at Giants’ Hills, Skendleby, Lincolnshire. Both Skendleby barrows were constructed using the technique of fences of stake posts, now emphatically holding multiple horizontals (hurdling), and not only single fence-top poles. Each barrow was at some stage fronted by a façade of heavy timber posts, which in the earlier case were used to mark the extreme directions of the rising of the midwinter Sun and the setting of the midsummer Sun. Skendleby 2 long predated Beckhampton Road, and seems to have set its perpendiculars towards far edges, as did that barrow. At a later stage two other early instances will be added to our list of long barrows with solar alignments.
The two Skendleby barrows are near neighbours on the same hillside—they are only 250 m apart—and are among fifteen or so long barrows on the Lincolnshire Wolds. Skendleby 1 (latitude 53° 13' 05") was excavated under the direction of C. W. Phillips in 1934, and Skendleby 2 (latitude 53° 12' 40"), in what amounted to a rescue operation, under the direction of J. G. Evans and D. D. A. Simpson in 1975–6. It is likely that each went through at least two distinctive phases, which it will be convenient to distinguish as Skendleby 1A and 1B, and Skendleby 2A and 2B, and so forth.
Both were almost wholly surrounded by ditches, and those flanking the barrows seem at first sight to have been by no means at equal levels. Had this been wholly true it would have meant that our fundamental assumption was inapplicable, and that our method would have needed to be applied, if at all, in a modified form. (To every observing altitude from one ditch there corresponds a unique ridge height and therefore a unique observing altitude from the other ditch. Given a complete ditch survey it would in principle be possible to extend the method to such cases, matching disparate observing levels.) In fact on closer examination it appears that a ledge was deliberately placed in the lower ditch to be at the level of the floor of the upper (see the left lower part of Fig. 38). This ledge was vital to the planning of the barrow: it was clearly cut to satisfy the general precept that observation must be at equal distances from the ridge, and from places at equal levels below it. As can be seen from the same figure, however, when the ditch was cut of necessity to a lower level, for materials, the possibility was retained of observing at precisely the same angle albeit from that lower level. At Skendleby 1 the same principle is to be seen in reverse (see the upper left of the same figure). Here the key positions for the planning of the ditch were in mid-ditch, but an additional place was provided on a wide inner ledge. Something similar was seen at the Horslip long barrow.
The excavation of Skendleby 1 was a model for its time—perhaps one day a return will be made to Phillips’ attempt to draft ditch contours—and the monument yielded good geometrical detail. Fig. 39 abstracts what from our point of view is the most important information. The façade, in what was described by Phillips as a revetment trench, is at R. The posts in it had been split, and the flats of the semicircles set against the outer curve, to the northwest, suggestive of a wall intended to be a solid curtain. H was a large hole, B a platform for the burial, and S a line of stones, perpendicular to the axis of the barrow.
FIG. 38. Important sections of the ditches at barrows 1 and 2 (in their final forms) at Giants’ Hills, Skendleby. Both barrows run close to contours on hillsides, and the fall of the ground across the barrows is easily appreciated here. A rough idea of likely mound forms is given by the lines of sight, astronomically derived. The upper section corresponds to the line XY in Fig. 39 and the lower to a roughly similar place on the similar barrow Skendleby 2. Key viewing positions at the planning stage were x, y, p, and q. Note the optional viewing positions, z and r, to the left (south west) of the sections.
Skendleby 2 had been severely damaged by long ploughing, but the mapping of the barrow was done meticulously, and radiocarbon dating produced no fewer than fourteen dates in all, covering a perplexing spread of time—almost two millennia. Its end-ditches were clearly meant for observation. There were actually two end-ditches at the northwest: the first had been filled in before a new one was cut about 10 m further from the façade. This might mean that the mound was at some stage lengthened—say from 60 m or less to 70 m. The former, as it happens, is the length of Skendleby 1. The deeper ditch to the front of the façade was well fitted to observation over the mound. The rough equality of the distances of observers, from side ditches to ridge, and from front ditch to façade, will eventually be used to confirm that there was at Skendleby 2 a set of three equal viewing angles.
FIG. 39. Outline of Skendleby 1 (upper figure). The contours are at 2 ft (about 60 cm) intervals, and although unnumbered, their change can be estimated with the help of Fig. 38, showing the section XY. The mound base (outer contour of hole H) is 236 ft above the Ordnance Datum. The positions of many of the revetting posts round the edge follow the conjectures of C. W. Phillips. The stake holes for the fencing of the bays, and collapsed poles, once tied to them, are at the northwest. The studied proportions of the southeastern end are intimated in the lower figure (drawn to a larger scale), where construction lines are conjectured. The width of one of our hypothetical bays is double the interval between revetting posts, that is, about 3 m.
These twin barrows provide us with an object lesson in the importance of precision, something that can be best illustrated by the fact that, while the barrows differ in direction by only a few degrees, and are of much the same form and dimensions, the star Deneb was much implicated in the design of one but not the other (Skendleby 2, but not Skendleby 1, which was begun several centuries later). There are some difficult problems here, not only internal, but of the relations between the two barrows. Even after two exemplary excavations, there is much simple but desirable surveying information that is lacking, and that might one day allow a more complete analysis.
The bowed form of the façade of Skendleby 1 seems to hint at a clever architectural device for holding back the soil of a mound. On these grounds it might be made the later of the two—and the astronomical arguments fall in with this idea. In both cases, a mere glance at the plans is enough to suggest that there are components of the barrows that are skewed with respect to the axes of the mounds. This should be enough to alert us to the presence of multiple celestial alignments.
Consider first Skendleby 2. The southern edge of the mound has been irretrievably lost, but long stretches of the northern edge are known in minute detail—for which the excavator’s report must of course be consulted—and a perpendicular to the first and best sets an azimuth of 225.7°. An average for five transverse fences of stakes on the south side gives azimuth 45.8°. Together, these suggest that the principle of setting right angles to far edges is being observed here, as at South Street. The figure of 45.8° is doubly significant, as will be seen, for the observation of Deneb. This is the first crucial azimuth.
The façade sets a good line at right angles to a row of thin stakes (see Fig. 40, azimuth 133.8°) at its centre. It was likewise almost exactly perpendicular to four posts of intermediate size at the northeast side, which were no doubt an aid to setting out the façade, with properties to be explained. The direction of the façade is comparable with that of the front edge of the mound and that of a transverse stake fence on the northern side, but again these will be kept in reserve. It will be assumed that viewing across an unspecified mound (not the last erected there) from the northern ditch was done at an azimuth of about 223.8°. This is the second crucial azimuth.
FIG. 40. Outline of the Skendleby 2 long barrow. The coarse broken lines pick out some of the symmetries of the plan. Lines numbered 1 and 2, passing through the gap in the façade, are sight lines for solar observation in two directions. For further detail see Fig. 41.
Combining it with the first, in the now standard way, one finds that the rising of Deneb could have been seen over the mound of Skendleby 2A to the northeast from the southwest ditch at altitude 13.96° around 4020 BC, together with the setting of Betelgeuse at the same date and altitude, looking from the other ditch to the southwest.
In this first phase of the monument, the ditches might well have differed somewhat from the later ditches, as is known from the excavation, but it must be said that the change cannot have been great, for what is known of them fits very well with the derived viewing angle. The ditch to the front, moreover, is everything that we could wish for. By a fortunate chance, the excavators took a section almost exactly where it is needed. The short row of stake holes in front of the façade (not to be confused with the long row along the axis) meets the front ditch at a very special point indeed (Fig. 41). Lines from this point to the extremes of the observation trench run due west and due north. (This is our oldest example of a more than tacit occurrence of precise cardinal points.) Not only this, but there are two steps cut into the chalk at this point (see the right-hand side of Fig. 42), and from the upper step, observation could have been made over the top of the façade at the very angle (nearly 14°) to be derived from the assumption of cross-viewing from the side ditches. This angle has two remarkable properties. First, it allows the setting of Deneb to be seen in a direction virtually identical to that of the line of stakes in front of the façade, at the same date as the crosswise observation of Deneb’s rising. (The discrepancy in azimuth is about a quarter of a degree, which is of the same order as the accuracy of our present measurements.) Second, taken in conjunction with the levels of the ground and distance of the ditch, it carries the important implication that the height of the eye of a man of normal stature making the observation would have been exactly the height of the operative posts in the façade itself. In other words, the same man could just have seen over the relevant part of the façade, had he stood against it at ground level. The assumption here is that the height of the eye (E, in the figure) is made equal to that of the average West Kennet male (nominally 1.66 m).
It is hard to avoid the feeling that the near-perpendicularity of the two Deneb directions was the result of an attempt to make them precisely so. Had they been so then each would have been precisely 45° from north, that is, yielding azimuths of 315° and 45°, rather than 314.2° and 45.8°. The discrepancy might have been adjusted by a slight difference in viewing altitudes. It is not likely to be entirely due to our own error in the azimuths, since the need to bring Betelgeuse to the same altitude was surely enough to thwart any ambition to produce a Deneb right angle. What is certainly of great interest is the similarity between the Skendleby 2A’s use of perpendicular sighting lines to Deneb and the presence of perpendicular sighting lines to Sirius over the natural horizon at the Horslip long barrow. The date assigned above to that barrow was only three quarters of a century later than the date obtained here for Skendleby 2A.
The façade held thirty-one posts, and showed signs that at some stage posts had been removed by burning. This is of course one of the easiest ways of removing an old heavy post, but it should not be assumed that removal was once and for all. The façade functioned, if the present analysis is accepted, for many centuries, and might have been replaced many times. Traces of the posts of the façade show that they were by no means of constant girth, and some posts—especially towards the middle—were no doubt appreciably taller than others. How they were arranged it is impossible to say, but the excavation produced fragments of charcoal from small branches that the excavator interpreted as possibly indicating a wattle and daub arrangement. This fits very well with the idea of a curtain of timber through a gap in which observation of the first glint of the rising midwinter Sun was seen, in a way that will be explained.
FIG. 41. Plan of the area around the façade of the Skendleby 2 long barrow. An observer in the front ditch, looking over the façade in a direction originally marked with four stakes, saw the setting of Deneb. Note the north–south and east–west directions to the corners of the façade from the observer. An observer at the gap could observe the setting Sun at its midsummer extreme, along line 1. Line 2 is the sight-line towards the midwinter rising Sun from the ditch at the northwest (see Fig. 40).
The line of four posts meets the backfilled ditch at its very centre, and there are two equally spaced parallels to it that graze the holes for the large D-posts in the burial area and bring us to the corner of the (original) end-ditch. Over a natural treeless horizon, midwinter sunrise was at altitude 0.57° and azimuth 132.7°, and midsummer sunset at an altitude 1.44° and azimuth 311.6°. From a point near the northwest corner of the original back ditch (see Fig. 40), the Sun would have been observable at its midwinter extreme along the line marked 2 in the figure, grazing one of the burial split trunks and one side of the gap in the façade. The burial post would have covered the solar image as it rose, being about 10 per cent larger in angular diameter. The first glint of the Sun’s image would thus have been trapped between two posts, almost (but not quite) in a manner that was to become classic over the following two millennia. Midsummer sunset was similarly observable by an observer looking along the line of sight marked 1. The posts flanking the gap in the façade were unusual, as it happens, in that they had flattened faces to the inside. It was easy to follow the rising Sun, by shifting one’s own position slightly, opening and closing the slit at will. The angular width of the D-post was well matched to that of the Sun, as seen from the end-ditch, being about 10 per cent larger. From the façade, of course, the post was nearer and the margin much greater.
While we do not have a ditch section at the observer’s position—a high priority for any future excavation—we do have one not far from the centre that offers support for the claim that the end ditch was used for solar observation. As shown in Fig. 42, a ledge there—had it been used for viewing—would have brought the observer’s eye to the same level as the top of the posts over which the setting of Deneb was observed. Given no other obstacle, in other words, the observer can see perfectly well down to the natural horizon. This suggests solar viewing, since no star, not even Sirius, is visible down to that level (0.57° without trees). Observations would have been made by someone standing at the point described, but level with the step and at the inside edge of the ditch—in fact more or less at the old ground level.
FIG. 42. The front ditch (right-hand side) and (original) rear ditch of the long barrow Skendleby 2A. The observer stands in the front ditch on the upper step in the chalk, with eye at E, and views the setting Deneb at altitude of nearly 14°. The same observer at ground level would just have been able to see over the posts at F. The lower step was cut for the later barrow (Skendleby 2B). At the left of the figure a section is reproduced probably resembling others from which midwinter observations of the Sun could have been made with Skendleby 2A (and possibly later). The observer’s eye (0) is at this end evidently also on a level with the top of F. The places actually used no doubt offered a more comfortable place to stand than the one shown. At the optimal point for the natural horizon the observer stood virtually at the old ground level, but with eye at the same level as shown here.
This all leads to the suspicion that here are signs of the very first activity of this sort on the site. The old ditch at this end could have been cut only after the ditches for crosswise viewing of Deneb and Betelgeuse were added. If the unsymmetrical solar arrangement ever made it seem desirable to view the Sun from a more central place, this would have demanded a higher altitude, easily arranged by building up the relevant façade posts. (For viewing along any of the set of three parallels, for instance, they would have been 2.9 m above ground level at mid-façade.)
One cannot be certain that there was no tree cover on the horizons needed for solar observations of the sort described, but both horizons are set by land fairly close by. It was shown from an analysis of molluscs sampled during the excavation that the immediate neighbourhood of the barrow was grassland, and that the mixed deciduous woodland that had covered the area had been cleared not long before. There are good reasons for thinking that the site was chosen for its solar properties: to bring the sighting lines to the solstices as near to parallels as possible, a higher altitude was needed towards midsummer setting than towards midwinter rising. The idea of perfectly reversible solstice alignments will be met with repeatedly elsewhere, but only at a much later date. Skendleby 2A is remarkable chiefly because it is such an early example of a carefully designed solar structure.
What of the implications of the mound, our Skendleby 2B? It should be plain enough that it brings alternative azimuths into the reckoning. The mound edges mentioned earlier provide much the same azimuth as the second crucial figure, and it seems that an effort was made to keep it, presumably not only for continuity’s sake in some abstract sense but because of its implications for observation of the Sun (the change in which is so slow that we can here ignore it). This, however, does not explain the direction of the very well defined line of stake posts and larger rubble running down the spine of the barrow.
Viewing across this would have provided azimuth 228.9°, which replaces the previous 233.9°, in our conjectural reconstruction of observing practice. Each of the options may be combined with a direction (48.9°) perpendicular to the spine of the barrow and the other related lines. In view of the barrow’s imperfections, it is unlikely that much more reliable data than this will ever emerge. Applying the usual method then yields a rather surprising result. The first option suggests viewing of the rising of Deneb and the setting of Aldebaran at an altitude of 15.76° around 3750 BC. The second option suggests viewing of the rising of Deneb and the setting of Betelgeuse at an altitude of 15.75° around 3730 BC. What is surprising is that the same mound could have been used to observe the setting of Aldebaran and Betelgeuse at more or less the same dates, as long as the directions of viewing were not too stringently set. (Aldebaran and Betelgeuse could have been seen setting across Skendleby 2A at points 5° apart in azimuth.) We have always assumed—for want of any sign of permanent direction markers—that observers after the initial architectural act had no precise means of knowing that the direction of view was strictly at right angles to any particular edge. It is for this very reason that a long barrow, used as explained, could have avoided becoming plainly outdated, in the minds of those who used it, for decades, even centuries. From the moment of foundation of a barrow, the stars were slowly shifting their places of rising and setting, but it would have been long before any but the expert realized the extent of the drift.
This is to assume that the barrow was used for ritual observation after its foundation, and that it was not just a question of ‘building the stars into the monument’ and then paying little attention to its properties. That the front ditch was still being used at Skendleby 2B to observe the setting Deneb, in exactly the same way as before, seems very likely, and the evidence is beautifully simple. The upper step on the inside wall of the ditch (Fig. 42) is just 32 cm above the lower, and this is exactly what would have been needed to change the viewing angle from 14.00° to 15.75°, at the distances in question, assuming that the property was to be retained whereby the relevant façade post came to the height of the observer’s eye. As for the character of what are here being referred to as steps, it is not to be supposed that they resemble those at St Paul’s. They were rough-hewn. They do seem to have been worn with some use, although very frequent use would surely have worn them more. These things have not been properly studied, since ditches have usually been regarded only as quarries for mound building. Either observation was done relatively infrequently, say by a single person at selected seasons, or the step was protected, perhaps by a wooden platform that it was meant to support. The second seems in any case not unlikely, just as ladders allowing entry to ditches in general would have been almost essential.
Skendleby 2B, with its long mound, perhaps built in more stages than suggested here, was thus brought into existence to redress the changes of the centuries. Some centuries later, and perhaps because of dissatisfaction with yet further changes nearby, a completely new barrow was built along similar lines, namely Skendleby 1. Very briefly, starting from azimuths 36.3° and 218.6°, one may conclude that the rising of Vega was seen from the southern side and the setting of Bellatrix from the other, both at altitude 17.4° around 3120 BC. It has to be said that there is another possibility, and that around the year 3550 BC the rising of Vega could have been combined with the setting of Betelgeuse, at altitude 19.0°. Other things being equal, the (later) option with lower altitude seems more fitting to the known scale and architecture of the ditches. Perhaps the two phases were both known; and it might even be prudent to assume that there was a still earlier ‘façade’ phase, as at the barrow nearby—in which case the labels Skendleby 1A, 1B, and 1C ought to be reserved. The chief reason for favouring the year 3120 BC (with the usual uncertainties) is that Skendleby 1 has provided two radiocarbon dates, equivalent to 3125 ± 225 and 3000 ± 300 BC (2460 and 2370, each ± 150 bc). The first of these could hardly be closer to our figure, although it would be unwise to claim better than an accuracy of a century or so in our own dating.
It seems that there were three centuries between Skendleby 2A and 2B. In a third phase, or even in a tail appended to 2B, it might have been used with Deneb still, but with Bellatrix in place of Betelgeuse—at the mound head around 3420 BC, say. Deneb would eventually have become unusable there. It is just conceivable that in due course Skendleby 2 passed into yet other phases. It would in this case have been built up slightly higher, making the viewing angles around 20°, and it could then have been used successively with Vega and Aldebaran (around 3310 and 3070 BC, in different ways) and Vega and Betelgeuse (around 3030 BC). On the whole it seems safer to put aside all these options, and to associate a highly specific structure only with a first probable use, that is, the use for which it seems to have been deliberately designed.
Did Skendleby 1 have any solar use? Its axis is less than 2° away from the required direction. There is an almost perfect alignment set by the revetting posts running from the northwest corner of the mound towards the right-hand (southwest) edge of the entire façade, but this part of the façade has been lost. The burial area might have been implicated in some alignment using the middle of the façade, but it is difficult to see how, and much less is on record concerning the fine structure of the façade here than at the other barrow.
Most of the fourteen radiocarbon dates associated with Skendleby 2 provide support for the gamut of astronomical dates offered here. The oldest, for a piece of façade charcoal, was equivalent to about 4250 ± 100 BC (1965 ± 80 bc). This was presumably large and very old timber, and is recognized as anachronistic. There followed three highly consistent specimens, of oak charcoal and antler, the oldest being 3910 ± 120 BC (3155 bc), and so belonging to phase 2A. One item of charcoal from the north burial post pit (3790 ± 150 BC), a solar observing post, shows this phase continuing in use. Another specimen from the same post was well over a century earlier—supporting the idea of post renewal. The mound and burials produced more, hovering around 3500 BC, all with similar ranges of uncertainty; and there were others still later. These dates, however, are less interesting by far than the simple fact that the Sun’s turning points were being accurately observed at a Lincolnshire long barrow in all probability just before the beginning of the fourth millennium.
As a postscript to this section, it is worth putting on record a curious fact about the nearest stretch of the A1028 road, marked on the Ordnance Survey as a ‘Roman Road’. A consistently straight tract of this is still in use, and covers nearly 4 km. The now disused part of the road continued in exactly the same direction for 7km more, taking it south to the village of Burgh le Marsh. The section still in use passes within 400 m of both Skendleby barrows, and its azimuth is 133.7°—differing therefore by only a fifth of a degree from the line of stake holes in front of the façade of Skendleby 2. This does not imply any immediate connection with the barrows, of course, but it does seem highly probable that the road embodies an alignment on the Sun, and that this was pre-Roman. The gradient is very slight (averaging 0.23° upwards in the direction 313.7°). The direction of the road never matched that of the extreme of midwinter sunrise, but it did once perfectly match that of sunset (last glint) at the summer solstice in the Neolithic period. Using the figures quoted here implies a date in the neighbourhood of the twenty-second century BC, but there are slight uncertainties in the direction and altitude, and these are far more critical than in the case of dating by the stars. (The direction of midsummer sunrise changed by only two-fifths of a degree over the previous two millennia, for the stated altitude.) This is our first example of a trackway aligned on the Sun, but as will be shown in the following chapter, it was by no means an isolated instance.
A barrow nearer to those of Wessex that seems to have shared some of the contrivances of Skendleby 2 has a most unusual shape, for its mound was in the form of a parallelogram, and not especially elongated. It was excavated at Barrow Hills, Radley, Oxfordshire, in a rescue operation led by Richard Bradley in 1984—as one of many prehistoric monuments on a site destined to be submerged under a housing estate on the northeastern edge of the town of Abingdon. Lying on a gravel terrace a little more than a kilometre north of the River Thames there, this site had seen Neolithic, Bronze Age, Romano-British and Anglo-Saxon settlement. (The tradition of refusing to treat the land as a museum is clearly an ancient one in Abingdon.) The nearby Abingdon ditched enclosure was separated from the barrow by a very shallow valley, and it is assumed that the two were set up at much the same time.
The long barrow is difficult to analyse not only because most of its surface detail has been lost but because it had ditches that were superimposed on it, in some places cutting through an older trench, assumed to have been a post trench. Fig. 43 shows the ditches in rough outline. It was suggested by Bradley that there were four phases:
(1) A fenced enclosure, the posts of which were later removed. Our outline figure hints at the existence of the narrow trench in which these were taken to have been placed. He thought that the grave might have belonged to this phase, or more probably to the next. It was very shallow and held two crouched burials of adult males laid along the axis of the site with heads at opposite ends of the grave and legs laid across one another.
(2) A three-sided ditch (surrounding the first trench to the north of points x and y on our figure) and large post holes E and F, the former seeming to have held a split trunk. (For the notation, see Fig. 46.)
FIG. 43. Outlines of the ditches surrounding the long barrow at Barrows Hills, Radley.
(3) The open end (below x and y, where a causeway was left) was closed with another length of ditch avoiding one of the large posts.
(4) The outer ditch was cut in a series of segments between 2m and 5 m long, with a causeway at the southernmost corner. There were deposits of pottery and flint scrapers near both causeways.
Stretches of the sides of the outermost ditch were clearly deliberately curved, in contrast to the inner ditches, which contain some quite passable straight lengths, for example on the ditch floor. Averaging over three of these in the short direction and four in the long, the plan can be provisionally taken to have been meant as a parallelogram with sides at azimuths 136.0° and 216.0°.
It is very unfortunate that all detail of the structure of the mound has been lost, but one thing is clear: the trench and inner ditches were intimately related in shape. The material thrown up from the trench would not have been enough to build a mound worthy of the name. Even the material from the inner ditches would not have created an artificial horizon to a person standing anywhere but in the ditch. This presents a difficult problem. Why did they cut into the trench at all?
There are many possible answers, but it is not easy to find one that carries conviction. The first phase might have begun with the outer ditch, perhaps starting from the south side, and looking north to Deneb’s rising. By the usual procedures—possibly using E and F as scaling posts—the required angle of view would have been known. The mound area might then have been marked out and surrounded with short revetting posts to retain the soil. But the outer ditch alone (about 90 cu m) would not have provided enough material to create a revetted mound setting an angle of 12.3°. Perhaps labour was in short supply, or for some other reason it was decided to build a less ambitious mound. For the same viewing angle they would have needed to move closer. Since the mound was so much lower, the revetting posts were no longer needed, so that they and the trench would have been dispensed with.
A first impression of the ditches at Radley is that they are a rather chaotic patchwork, but this is not so. There are surprisingly many points of agreement between the excavation outline plan and a geometrical scheme shown in Fig. 44, tending to support—rather than prove—the idea that the outer ditch was planned in relation to the mound limits, the positions of the southern posts, and the grave area (ill-defined to the excavators). It should be noted that all circles in the scheme are centred on the parallelogram, and that the centre in all probability was originally at the head of the grave. The scheme otherwise speaks for itself. Only those with the experience of making geometrical constructions on the open landscape, with antler pick and ox-bone spade, have a perfect right to say that agreement could have been closer.
Another explanation of the discarded trench might be that an attempt was made to build a mound with high edge—say half a metre—from a shallow inner quarry ditch, but that quantities were misjudged, and the revetting, growing increasingly unstable, was abandoned in favour of a low mound. The inner ditch (about 75 cu m) would have provided just enough material for a 12° mound tapered to the ground, without posts (Fig. 45).
No matter what the explanation, it can make little difference to the astronomical geometry of the situation that has come down to us, which has interesting echoes of Skendlebury 2. There, matters were so arranged that Deneb could be seen rising across the barrow and setting down its axis. Here, at Radley, conversely, Deneb was to be seen rising along the length of the barrow (by the observer at C), and setting across the mound (by observers at points marked D, or between them). More details will be given shortly. At Radley, however, there was exactly the same doubling of function with respect to Bellatrix too: it would have been seen rising across the mound by observers between points marked A, and setting along the barrow’s length by an observer at C.
There is some latitude in the placing of observers in the inner ditches, since all that we require of them is that the lines of sight pass over points of maximum altitude of the mound. In fact each of the viewing positions marked on the figure has something more to recommend it, however insignificant it may seem (B was at mid-ditch to E, B was off centre at a point where the ditch was cut slightly deeper, and so on).
Ignoring for the time being the slight discrepancies between those ditch lines and azimuths that are exactly multiples of 45°, the following solution illustrates principles found from earlier barrows. Ensuring the same viewing altitude in the two directions would have guaranteed that Deneb—observed at right angles to a short edge and parallel to a short edge—was to be seen rising and setting. Neither here nor at Skendleby is viewing strictly along an axis. The same goes for Bellatrix with respect to the long sides. All four possibilities, however, would have held simultaneously only at one particular time in history. To find that time and the shared viewing altitude for the latitude of Barrows Hills (51;40, 45°), our usual procedure is followed. Given the ideal as explained, the year turns out to be 3700 BC and the viewing altitude 12.3°.
FIG. 44. Potential geometrical construction lines for the entire original system of ditch and mound at Radley. Only the surviving ditch edges are shown. The short sides of the parallelogram in particular are based largely on the ditch structure and not only on its outline as drawn here.
FIG. 45. The probable overall shape of the Radley mound, in idealized form.
FIG. 46. The mound area and inner ditches of the Radley long barrow, in the form of a parallelogram. The mound was limited by a narrow trench, probably before these ditches were cut. The broken lines show what is here taken to have been the ideal aimed at. The short sides are drawn here at exactly 45° to the cardinal points, differing only very slightly from what would be suggested by internal features of the ditches. The lines of sight are drawn parallel or perpendicular to the sides. Another ditch, omitted, but more or less surrounding what is shown here, was about 2 metres distant from it. Points A, B, C, D, and R are potential viewing positions for the stars Bellatrix, Deneb, and Rigel.
