I have been having a discussion on a Facebook site on the differing characteristics of ring cairns, circular banked enclosures, Iron Age round house remains, and medieval cattle pounds.
We have one of the latter in Newport, characterised by massive stone walls about 2m high and almost 2m wide. It''s well preserved, as you might expect for something less than a thousand years old and used up to maybe 100 years ago. There is a single entrance with an iron gate still in good repair. Internally, the diameter is about 10m -- but there is effectively no limit to the size of pounds like this, since there is no roof to support -- a pound could be 30m or more across, the the builders had enough energy and enough building stones for the completion of the task.
But when you want to built something with a heavy conical thatched roof on top, real engineering problems appear. What we were discussing on the Facebook page was this -- how big can a hut be before the weight of the roof becomes unsupportable?
There are lots of "circular features" in the prehistoric record for Preseli, and the Dyfed Archaeological Trust field workers are often quite circumspect about their origins:
It seems to me that features with a diameter of between 5m and 20m are sometimes considered to be the remnants of huts -- and referred to as " hut circles" -- while larger features (with a diameter of over 30m) are often deemed too large to have been roofed, and are therefore referred to as animal enclosures. So is there an effective limit of about 20m as the maximum diameter for a genuine Iron Age roundhouse?
I know nothing at all about structural engineering, but here are a few images culled from the web:
This is an interesting animation: http://www.bbc.co.uk/history/interactive/animations/ironage_roundhouse/index_embed.shtml
One of the old round houses at Castell Henllys was taken down a few years ago -- not sure whether that was because the thatch had rotted to the point where is was leaking at an unacceptable level, or whether the structure was collapsing. Somebody up at the site told me the other day that the wall posts embedded into the ground were rotting -- that probably meant that they were no longer able to withstand the outward pressure exerted by the weight of the roof.
Now the building is being erected again, no doubt with lessons having been learned! Note that the wall pillars are quite sturdy and short, with a maximum exposed length of about a metre (the building is on a slight slope, and the wall plate is carefully levelled). So when all of the rafters are in place the wall plate will take the weight of the roof. The rafters will be locked at the top of the cone by the centre post, which is currently supported on a scaffolding platform (that will be removed once the structure is complete). There may be no need for "ring beams" like the ones shown at the top of this post) since the preferred method here for stopping the collapse of the cone is a series of crossing purlins like these: