All of the above, but particularly the points summarised in:

Discussion.. Evidence that the results obtained are not due to chance: strongly suggest that the alignments found are real, intentional and show a planned programme to find alignments for the lunar perturbation limits.  This would have enabled prediction of eclipses.

The implications of the evidence for precise lunar alignments is that the currently accepted nature of society in the Early Bronze Age requires revision.  The setting up and use of lunar alignments would require widespread organisation and suggests that a hierarchical society is likely to have existed with an elite class.  (MacKie 1977, 1981)

While we perhaps can never be certain that the rationale for the determination of the lunar perturbation maximum, and presumably its time, was for the purpose of the prediction of lunar eclipses, there would seem to be no other logical reason for what could only have been a significant effort.

If the lunar alignments are correct then so must be the value (± 23º 54’.2 ± 0’.7) for ε as determined by Thom from the presumed solstitial sites Kintraw, Ballochry and others (Thom, MLO pp36-38 ) as that value is central to the determination of the value used for the extreme declination of the moon, namely ± 29º 03’.  Thus the claims made by Thom that the above sites were set up for the specific purpose of the precise measurement of the solstices about 1750 BC ± 100 years would seem to be correct.