following remarkable hypothesis: 'Their tops are shaped in an oval,
conical form, meant to make a joint in a socket to erect the
superstructure on.' These words are quoted from a 'Report of a
Conjoint Visit of the Geological and Philosophical Societies to the
Dumbuck Crannog, 8th April, 1899.' {26}

The result of the excavations, so far as I can gather from
observations made during my second visit to the 'crannog,' and the
descriptions and plans published by various societies, may be briefly
stated as follows.

The log-pavement within the circle of piles was the upper of three
similar layers of timbers placed one above the other, the middle layer
having its beams lying transversely to that immediately above and
below it. One of the piles (about 4 feet long) when freshly drawn up,
clearly showed that it had been pointed by a sharp metal implement,
the cutting marks being like those produced by an ordinary axe. The
central portion (about 6 feet in diameter) had no woodwork, and the
circular cavity thus formed, when cleared of fallen stones, showed
indications of having been walled with stones and clay. Surrounding
this walled cavity--the so-called 'well' of the explorers, there was a
kind of coping, in the form of five or six 'raised mounds,' arranged
'rosette fashion,' in regard to which Mr. Donnelly thus writes:

'One feature that strikes me very much in the configuration of the
structure in the centre is those places marked X, fig. 20, around
which I have discovered the presence of soft wood piles 5 inches in
diameter driven into the ground, and bounding the raised stone
arrangement; the stones in these rude circular pavements or cairns are
laid slightly slanting inwards.' {27}

From this description, and especially the 'slanting inwards' of these
'circular pavements' or 'cairns,' it would appear that they formed the
bases for wooden stays to support a great central pole, a suggestion
which, on different grounds, has already been made by Dr. David
Murray.

The surrounding piles were also attached to the horizontal logs by
various ingenious contrivances, such as a fork, a natural bend, an
artificial check, or a mortised hole; and some of the beams were
pinned together by tree-nails, the perforations of which were
unmistakable. This binding together of the wooden structures is a
well-known feature i