Thanks Cian! I've hiked my way down through the messages and seen that Fourwinds has got there just ahead of me.
I'll just quote a few lines from Burl (1988):
"There are more quadrilaterals and trapezoids amongst Four-Posters than precise rectangles. It was the circle that was the basis of the design."
As ever it's the Southern Ireland sites that can be guaranteed to be different. Doesn't it often seem that when you're reading about ASC's, four-posters, or whatever that you can just predict the use of a phrase such as 'degenerate form of ...'. Maybe I'm just being parochial. I'd like to think that the people around here were bravely striking off in a new improvisational direction.
Robinstown Great, as you mentioned, can be fitted to a circle. As can Mullaghmore.
Lettergorman, Gortnacowly and Maughanaclea SW could be, but may not have been.
Slightly further west, Cappaboy Beg retains its four stones, as does the more distant Reenkilla and neither can be put on a circumference.
I used a circle to show a possible range of positions in my drawings, however as Fourwinds states, even if the stones were placed on a circle, the choice of point to draw from would affect the outcome. Particularly in the case of Gortnacowly where the is such a large discrepancy in the size of one monolith. Thom (1967) sates that the centre points of the stones should be used and this is what I initially did with this site. However it looked better to me to keep a roughly even amount of each stone inside the perimeter, so ultimately I used the inner edges as my markers.
My gut feeling, for what it's worth, is that Lettergorman was on a circle but that the other two 'Three-Pointers' possibly weren't. They're just that bit further west and they share the one dominant stone template with their geographical neighbours.
So askew it is!
Take it handy,
g.