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