•20

D. G. HOGA11TH

slabs were found on tho rloor north-east of the anta. Sherds over all tliis
space were scarce and late Mycenaean.

As soon as the absence of constructions near the surface had beeu estab-
lished this area was sounded more deeply, and the rock appeared at moderate
depth with early painted wäre lying on it. A very well preserved early
ground plan was there found to uuderlie later constructions just south-
east of the Palace (sce Fig. 11). In H 3 an orifice was revealed, lined
with earthenware cylindcrs, which proved to be the mouth of the well
(Fig. 10), described below (p. 58). The earth was cleared out of this to 'a
depth of nearly six metres and found to be mixed with sherds exclusively
Mycenaean. Below this point infiltration of water from the sea stopped
progress.

To east of the Palace block, however, a large late structure, whose
walls showed above the original surface, was now opened. Its cement floor-

Fio. 10.—Tue Well Mouth.

ing appeared at a depth of about -50, and, in the passage to söuth of
it, remains of a marble paving at the same level. Further excavation herc
resulted in the finding of the interesting earlier house described below (Fig.
32) but the whole space was singularly barren of pottery. In the north of
the next block on the south (J 2: 9) were found two complete pithoi with
incised decoration, at a depth of 1'45 metres below the latest floor level (Fig. 11).

In the eastern part of J 2 and 3 similar deeper sinkings revealed the
best preserved pre-Mycenaean house on the site (described below, p. 44),
whose south wall almost touches the line of the drain, explored in 1898
(Fig. 12). The remarkable preservation of early buildings in this region is due
perhaps to the artificial Alling in of a space before the later Palace. This house,
based on rock at 3'50, yielded an enormous amount of pottery, including the
curious vessel described below (p. 207), found in tho doorway betwreen rooms 2
and 3. The inside ofthe walls showecl mach rough cement still adheribg, but