## Horizon Astronomy

The basic astronomical formulae are (2.1) to (2.4). Solving these equations for the case in which the altitude, h = 0, when an object of declination 8 is on the horizon, we have approximate expressions for azimuths and hour angles of rise and set:

 cos f ' (6.1) sin A = - cos S x sin H, (6.2) sin H = - sin A cos S (6.3) cos H = - tan f x tan S. (6.4)

Recall that observations are made through Earth's atmosphere, which both refracts and scatters light. Therefore, for accurate results, corrections to the altitude for elevation, dip, and refraction must be made to determine the true azimuth and hour angle of the object. Consult §3.1.3 for details.

Twice the hour angle of rise or set is the time spent by the object above the horizon; as far as we know, there is no direct evidence that this type of information was recorded in any way in megalithic times. The azimuth of rise or set, on the other hand, can be and probably was marked by sightlines directed to points on the horizon.

It is the reality and the purpose of the layouts of these sightlines that constitute the main subjects of debate in megalithic astronomy. In Megalithic Sites in Britain, Thom (1967) described 145 stone circles and other monuments, and the solar and lunar alignments. With subsequent works (Thom 1971, 1974, A.S. Thom 1984, among others), the claims were sharpened and better substantiated. But criticisms abound (many are summarized by Hicks 1984a). To see why, we first describe how alignments are measured, and then we discuss the types and results of such measurements.

Sightlines are best measured with a theodolite, a survey device capable of high-precision measurements of altitude and azimuth. A measurement of the altitude and azimuth of the Sun in the daytime, or of a star at night, made at a precisely recorded instant of time (necessitating a precision chronometer ), permits both longitude and latitude to be obtained (see §3.2.1). The determination of azimuth for surveying purposes requires that the direction of north (or some other direction) be known precisely, but measurements relative to some given direction can be made and later corrected. North can be established in several ways: by magnetic compass and application of the correction to true north; by measurements of the apparent azimuth and altitude of Polaris (or other appropriate stellar marker for previous epochs); by solar gnomon; or by bisection of the angle relating sunrise to sunset azimuths. Tables giving the differences in A and h between Polaris and the NCP at a given instant on a given date are provided in astronomical almanacs. To save time, the positions of landmarks that are included on accurate survey maps can be measured, and the true azimuth and altitude readings corresponding to all relative measures can be found later. The reduction requires tables of the right ascension and declination of the Sun for the dates of observation, and the star positions, corrected for precession. Corrections must be made for refraction and dip. Recall from §3.1.3 that refraction lifts the image of the rising or setting Sun (or Moon) above the horizon by an amount that is approximately equal to half a degree, but which varies with atmospheric temperature and air pressure, especially within the last few kilometers of the observer. The effect of refraction is to cause objects to rise sooner at smaller azimuths and set later at larger azimuths4 than they would without refraction. Similar corrections must be applied for measurements of the objects on the horizon (distant fore-

4 Recall that we have defined azimuth to be measured from the North Point (in the Northern Hemisphere) positive eastward. With a similar convention for treating azimuths at Southern Hemisphere sites, measured from the South Point eastward, the same comment holds for the Southern Hemisphere.

Figure 6.5. An example of an astronomical alignment involving distant foresights and backsights, and observations of a setting Moon near maximum standstill. Drawing by E.F. Milone.

sights) or closer objects. Dip causes the horizon to be depressed with effects similar to those of refraction, in the sense that lower altitudes are visible than with a flat horizon. Finally, there may be instrument corrections (scale and zero-point adjustments).

The kinds of alignment-checking measurements that can be made depend on the type and layout of structure being studied. In the relatively straightforward case of a shaft through a building (or tomb), its bearing can be measured by sighting along its length. In the case of two or more standing stones, the measurement of the bearing can be made by aligning the instrument along the stones. On the other hand, a ring of stones presents many more possible alignments and an accurate survey of the site may take considerable time. In some cases, mountains on a distant horizon may align a rising or setting object with a backsight (an object closer to the observer). In principle, such an alignment is purely arbitrary unless features of the local terrain, such as the side of a hill, greatly restrict the field of view, or unless the backsight is a suitable distance away from the observer. Both types of conditions serve to improve the probability that the alignment is intentional. Such an arrangement seems to have been found at Kintraw (Thom 1971/1973/1978, pp. 36-40; MacKie 1974), in the British Isles, for example, but even this site is not without controversy. The measurements, although sometimes inconvenient to make, are straightforward. Surveying is only the beginning, however. The interpretation is more difficult. Figure 6.5 illustrates an astronomical alignment involving distant foresights, backsights, and an observation of a setting Moon at maximum standstill. The more distant the foresights, the more potentially precise the observation, to the limits set by visual acuity and atmospheric conditions. Note that the "artificial foresight" could function as a backsight.

Once the measurement of azimuth of a distant foresight has been made and the corrections determined, the declination of an astronomical object conforming to the direction of the alignment can be obtained by the equations of §2.2.4.

This result can be compared with stellar positions that have been precessed back to the expected epoch, to the Sun at specified times of the year, or to the Moon sometime within the nodal regression cycle.

As we described previously, Ruggles (1981/1982a,c/1983) reviewed the evidence provided so carefully by Thom and associates and strongly criticized Thom's conclusions concerning the high precision of megalithic astronomical alignments. Nevertheless, Ruggles (1988a) concluded that alignments exist to three levels of precision:

(1) At the highest precision, there is some evidence, although marginal, indicating a preference for six specific values of declination, 8, viz., -30°, -25°, -22.5°, +18°, +27+, and +33°.

(2) At an intermediate level, there is a strong preference for alignments indicating -31° < 8 < -19°.

(3) At the lowest precision, alignments indicating the range -15° < 8 < +15° are present but rare, perhaps avoided.

We now turn from general propositions to particular cases.

### 6.2.5. Brittany

In Brittany, we find early passage graves made from gigantic rocks and covered with earth. Most of these show a clear preference for an orientation of the passage to the southeast. A few show a precise alignment to a position that we can regard as astronomically significant. Burl (1985, pp. 23-24) suggests that only the general orientation to the southeast was important and that within that range, the precision of an alignment, for example, to the winter solstice, was purely accidental. This is a solution that has found substantial favor with many archaeologists, not least because it relieves them both of the necessity to understand the astronomical evidence, and of the tedious labor of checking alignments, horizons, and the possible astronomical-geometrical relationships of the site being studied. But there would seem to be at least two other possibilities. One is that the bulk of the population being buried in these great tombs was only interested in a general orientation, but that a small minority was deeply interested in great precision and that these are the people responsible for such alignments as that of Dissignac, Brittany (to be discussed later in this section), to the winter solstice rising Sun. The second possibility is that most or all of the population was interested in a variety of different precise alignments that were important to them but lack significance for us. At the moment, we know of no objective external criteria that would allow us to choose between these alternatives.

The earliest case that involves a precise alignment is found in one of two passage graves at Dissignac in Brittany. This burial monument has a corridor so oriented that the rising winter solstice Sun illuminates the burial chamber. Both graves are in a common mound that was built in several stages, and the graves may be of slightly different dates. Mohen (1990, p. 304) assigns the initial construction to about 4500 b.c. The tomb having the alignment lies to the south-

Figure 6.6. The large menhir, broken pieces of which have been incorporated into the lintel of the passage grave at Gavr'inis and in the tomb of the Table des Marchands at Loc Mariaquer: Note the enormous scale of the complete menhir and its elaborate decoration. Drawing by Sharon Hanna.

Figure 6.6. The large menhir, broken pieces of which have been incorporated into the lintel of the passage grave at Gavr'inis and in the tomb of the Table des Marchands at Loc Mariaquer: Note the enormous scale of the complete menhir and its elaborate decoration. Drawing by Sharon Hanna.

west, with a 7-m long passage leading to a rectangular chamber, which contained many pieces of broken, extensively decorated pottery, beads, and a range of stone artifacts (Burl 1985, pp. 98-99). The entrance capstone has representations of shepherds' crooks and axes, appropriate for a small farming community. The northeast tomb has a similar passage, but with different orientation and with a bend that prevents light from entering the inner chamber. At this writing, early in the 21st century, this site appears to mark the beginning of the archaeoastronomy record.

One of the more interesting Breton sites investigated thus far is the passage grave at Gavr'inis. One of the lintels of this monument is a broken piece of a giant menhir, which once stood 14m high (see Figure 6.6). Another piece of the same menhir is built into the Table des Marchands tomb at Loc Mariaquer.

The latter site is also the location of another menhir, the tallest one known, Le Grand Menhir Brisé, which would have stood 20.3 m high when erected. It has been argued that this monument was so large that the people who tried to erect it were unable to do so and that it fell and broke. The discovery of another broken menhir of the same class and similar material at the same site suggests that both were broken in the same way, which is much less likely to be due to technical incompetence in both cases. Both Burl and Mohen have separately suggested that both monuments were deliberately torn down. However, the differing positions of the broken parts of Le Grand Menhir Brisé constitute a good argument for the view that the monument was toppled by earthquake (Thom and Thom 1978a, p. 98ff). Interestingly, Burl and Mohen draw opposite conclusions about the incorporation of part of a broken menhir into a passage grave. Burl (1985, p. 109) suggests that this "is evidence of the indifference prehistoric people could have for the handiwork of earlier societies," whereas Mohen (1990, p. 172) writes, "Using blocks again in other monuments symbolizes the endurance of a cult whose rites would suffer complete destruction of some of its sites." In any case, the presence of two broken giant menhirs in the same area is strongly suggestive that the same event was responsible for breaking both, whether due to deliberate human action, as Burl and Mohen suppose, or by natural events, as others have thought.

Thom and Thom (1978a, pp. 100-102) had presented the hypothesis that Le Grand Menhir Brisé acted as a foresight for megalithic astronomers gathering information on the movements of the Moon and suggested the identification of a number of sites that could have functioned as observing points for making these observations. From an archaeological viewpoint, the best support for such a view would have been some identification that the sites involved formed a complex. If they were all the same kind of monuments and all derived from about the same period, the hypothesis would have been supported. However, Burl (1985, p. 136) argues that three of the postulated backsights were probably not Neolithic monuments at all. He further argues that an alignment to a mound at Tumiac was inaccurately determined by Thom (if true this would be a rarity), and that the remaining monuments, menhirs, mounds, and passage grave are dissimilar in both architecture and date. Thom's postulated astronomical date of 1700 b.c. is later than some of the monuments. The most damaging case against the universal foresight hypothesis is that Le Grand Menhir Brisé was probably toppled before some of the other monuments were constructed and long before 1700 b.c. On the basis of 14C dating (see §4.3), Burl (1985, p. 108) suggests that Gavr'inis was built "in the centuries around 3500 b.c." The hypothesis demonstrates that it is easy to be misled when there is a plethora of monuments. A straight line, after all, requires only two points, and there is a high probability that any line drawn through the monument laden region will pass over some monument. A line from a particular monument to a specifed point on the horizon can almost certainly be extended backward to pass over some monument from which the foresight could have been seen.

The Gavr'inis site, however, has considerable importance beyond its role in the story of the great menhirs. As at Dissignac, the passage grave at Gavr'inis is illuminated by the winter solstice Sun. Burl (1985, pp. 110-111) notes an added interesting difference between this site and Newgrange (described in §6.2.6):

Looking from Stone 19, at the left-hand entrance of the chamber, toward Stone 1, the bearing is 128°, almost perfectly in line with the midwinter sunrise. The main axis of the passage is 134° towards the low-lying Arzon peninsula and the orientation is close to that of the major southern moonrise. It has been calculated that the two alignments, one solar, the other lunar, intersect halfway down the passage level with Stone 7, the white quartz slab whose undeco-rated surface may have been illuminated by the light of the rising Sun and Moon.

Gavr'inis is among the most lavishly decorated of the Breton tombs. Representations of long-horned cattle abound on the earliest Breton monuments, and the menhir that became the Gavr'inis capstone shows representations of such cattle. These have often been called oxen or bulls, but if the Breton cattle were similar to the contemporaneous Egyptian breed, the exceptional length of the horns would signify cows rather than bulls.

There is also extensive decoration on the companion passage grave at Le Table des Merchands. Here, however, it may be notational. Müller (1970, pp. 107-108) has drawn attention to the 56 shepherd's crooks, divided into 29 on the left and 27 on the right, and 19 accompanying curved elements, suggestive of the nearest whole number of days of synodic and sidereal months and of the Metonic cycle. At present, however, there is otherwise little evidence to connect the symbols with the cycles. Moreover, scholars disagree on what is being depicted. Burl (1985, p. 136) sees 53 crooks and an "anthropomorphic figurine," and Twohig (1981, p. 97) sketches the decoration differently. Whatever the outcome of this dispute, the motifs and style of the Breton passage grave decoration are closely paralleled at Newgrange and Knowth in the Boyne Valley of Ireland, and it is generally accepted that there is a close historical connection. The orientation to a winter solstice sunrise is common to Gavr'inis and Newgrange and suggests that both alignments are indeed intentional and arise from convergent or parallel traditions5 that stem from an earlier passage grave tradition in both areas.

0 0