- Chapter Eleven -
replaced. This was achieved by raising the lintel of the porch free of the tenons,
replacing the uprights and lowering the lintel back into position.
The third type of Celtic house, still a round-house, is one with a solid wall
of stone. Remains of these houses can still be seen in the classic stone country of
Cornwall, Wales and Scotland. With the exception of Chysauster in Cornwall where
the walls survive to a considerable height, the remainder survive as barely discernible
circles of stone rubble. This type of house hardly challenges the great double-
ring houses for size and splendour. Rather they perhaps reflect the poverty of their
landscape in so far as they have to use stone simply because the traditional timber
building materials are in extremely short supply. Nevertheless, these houses still hold
a deep fascination, surviving as they do in the final romantic Celtic landscapes. One
house, the evidence for which was excavated on Conderton Hill in Gloucestershire,
an outlier of the Cotswolds, has been built as a structure on two separate occasions
by the writer. The evidence comprised just the foundation layers of the stone wall,
its width being just under 1.0 m. The external diameter of the house was just over
7 m. Experiments during the excavation led to a conjectured wall height of about
1.0 m. The doorway was a mere 600 mm wide. The actual construction of this
type of house wall is relatively straightforward in that it is a standard dry-stone wall
with an inner and outer face. The only observation to be made is that in the original
and in normal practice the inner part of the wall is not rubble-filled. Each and every
stone is carefully positioned. What was remarkable was the sheer quantity of stone
needed to build a relatively modest structure. On both occasions the wall absorbed
in excess of 80 tonnes of stone.
The real challenge, however, lay in erecting a cone-shaped roof on top of a
dry-stone wall. The lateral thr~t of each rafter butt during building was enough to
dislodge the upper courses of stonework. Again one has to use the argument of
'without which, then nothing'; the only obvious way was to use a wall-plate around
the inner edge of the wall to spread the thrust, each rafter being simply notched into
place onto the wall-plate. The roof construction follows exactly the same sequence
as the single-wall houses described above. For each of the two structures a different
method of thatching the eaves was employed. The first made the thatch protrude
over the edge of the wall by some 200 mm, the other started the thatch over the
centre of the wall but with an under layer of sloping flat stones to shoot the rain-
water to the edge of the wall. This latter type of eave thatching is traditional in
both west Wales and the highlands of Scotland. The final aspect makes the houses
look dramatically different: the former with its thatched eaves looks wide and
comfortable while the latter appears narrow and quite prim. Which is the more
accurate is a matter of debate; perhaps they both are and reflect regional differences
which survive even into the present. The longevity of such a house is incalculable
and, provided the roof is kept in good order, it makes a snug and comfortable
dwelling against whatever extreme the climate chooses to provide.
Finally, with regard to solid-walled Celtic houses there is every likelihood, though
conclusive evidence has yet to be found, that the houses could have been built of turf
walls. Traditionally turf houses or soddies are known from Wales and Ireland and
were even taken to the prairies of America. But what traces would structures like
these leave? A scatter of pottery, enhanced phosphate and magnetic susceptibility