Sir Norman Lockyer, writing early in the twentieth century, was one of the first to suggest that there was a widespread calendrical practice in Neolithic and Bronze Age Britain and Ireland that involved dividing the year into eight equal parts. This calendrical practice fed through into Celtic times two millennia later, and thence to medieval pagan and Christianized calendrical festivals. It has nothing to do with lunar months but instead involves counting the days from one of the solstices. The eight division points are the two solstices themselves, the two equinoxes, and the four mid-quarter days.
When Alexander Thom started to survey scores of British megalithic monuments several decades later, and to examine their astronomical potential, he noticed accumulations of alignments upon the position of sunrise or sunset at each of these eight dates. There was even some hint of an interest in the dates halfway in between them. On the basis of this evidence Thom proposed that there existed in Neolithic Britain a “megalithic calendar” dividing the year into sixteen equal parts, each twenty-two or twenty-three days long, demarcated by “epoch dates,” which in the modern (Gregorian) calendar would correspond, to within a day or so, to June 21 (solstice), July 14, August 6 (mid-quarter day), August 28, September 20 (equinox), October 13, November 5 (mid-quarter day), November 28, December 21 (solstice), January 12, February 4 (mid-quarter day), February 27, March 22 (equinox), April 14, May 6 (mid-quarter-day), and May 29. While the six-teen-division calendar never gained wide acceptance, there was a great attraction in the idea that the eight-fold division of the year represented a thread of continuity stretching through the Celtic calendar and into the even more remote past. Indeed, mid-quarter day alignments have been found among stone circles, such as Beltany in County Donegal, Ireland; among Neolithic passage tombs dating back to the late fourth millennium b.c.e., such as Dowth (one of the large tombs in the Boyne valley) and Bryn Celli Ddu on the Isle of Anglesey (Ynys Mon) in Wales; and even in one Earlier Neolithic site (a U-shaped setting of timber posts) at Godmanchester in Cambridgeshire, England.
Yet this argument may be circular. There is a distinct danger of our picking out “calendrical” alignments because they seem interesting—be-cause they form part of our list of astronomical targets deemed worthy of attention—while ignoring many other alignments that do not point at calen-drical targets. Lockyer was certainly aware of existing ideas about the Celtic calendar, as was Thom. Reassessments of Thom’s data, paying stricter attention to the fair selection of data, have failed to duplicate his results regarding the calendar, as have systematic studies of local groups of monuments. Added to this, the whole notion of a precise “Celtic” calendar, confidently assumed by Lockyer and Thom and their contemporaries, is now known to be highly problematic and questionable in itself.
Finally, one of the other criticisms of Thom’s megalithic calendar was that he did not find calendrical sites as such (although the site at Brainport Bay in Argyll, Scotland, was subsequently claimed as one), but rather a scatter of individual monuments with isolated alignments, each of which indicated sunrise or sunset on just one of the various calendrical epoch dates.