Events on Site - The Stones and the Stars: Building Scotland's Newest Megalith - page 244

Sun moved several diameters left and right with the differences in
air temperature.
Archie Thom pointed out that the Sighthill megalith is par-
ticularly vulnerable to such effects because its skyline is in a sense
too good. The effects of refraction are not only much greater but
also much more susceptible to variation between 0° and 1° alti-
tude. The Neolithic builders must have been aware of the effect,
because they typically used sight-lines with elevations of 2-4°,
where variations in refraction were much less, and if they used a
true horizon it tended to be a sea horizon, with a cooling, calming
effect, and ideally looking down from an elevation of 1° or 2°, as at
the Scalisaig stone on Colonsay (Fig. 4.23 ).
If the observations had agreed exactly with prediction, rela-
tively little would have been learned. The original idea was that
if the features of the Sighthill circle were all based upon surviving
Neolithic sites, and yet it functioned as an observatory and the
project members could testify that it had been intended from the
outset to be one, then anyone maintaining that the ancient sites
were not observatories would have to point to some significant dif-
ferences. Now, however, it was obvious that if the Sighthill align-
ments had been determined by observation, rather than calculation,
the sunset alignments would have been a great deal more accurate
than they are, even within the limitations of a circle 40 ft across. If
horizon features were used as 'foresights,' as Thom suggested, then
much greater accuracy would indeed have been attainable. For
example, I had calculated the most northerly setting of the Moon
(at lunar standstill) to be where the natural line of the hillside met
the stepped roof of the right-hand distillery building (Fig. 8.17 ) . In
Neolithic times, no doubt the priority would have been to find a
site further north, from which the significant moonset would be
seen in the notch formed by Ben Lomond's rim on the skyline - a
shift of less than 2° in azimuth is needed. Then, observing the rim
of the Moon rather than its center, accuracy to within a minute or
two of arc, could indeed have been achieved as Thom maintained.
The 'errors' actually enable us to make an important point,
which otherwise wouldn't be obvious. As explained above, at the
ancient sites nothing happens exactly where it did, because the
tilt of Earth's axis has changed by half a degree meantime. It's
enough to let the skeptics say that the calculated alignments are