of about 1.20 m. Their present height varies between 0.70 m. and 0.90 m., but they must have been at least 0.40 m. higher, as the level of the uppermost house-floor here and there is as much as 0.30 m. above the top of the foundations; this being due to the fact that after the desertion of the settlement the foundations served as a supply of building-stones for other house-builders; just as the builders of these foundations took stone from earlier walls and thus demolished them to some extent. The method of joining the walls is very faulty. They are not tied together by means of binding-stones but simply butted against each other, leaving a "sabre-cut" at the joints, and thus reducing the strength of the walls. The upper walls were built of sun-dried bricks. All that is left of them is a layer of hard clay above the house-floors. The inside walls have been covered with a mud-plaster mixed with fine-grained sand, of which some. traces remain here and there on the foundations.

As far as can be seen the roof has been covered in the same way as that of the Alambra house. Some plates of baked clay, of the type used to catch the drip from the eaves, were found. They are of exactly the same shape as those in Alambra.

No threshold constructions were found.

The house floors consist of a substructure of small rubble-stones or rough gravel, and sea-shells, covered with a layer of trodden earth or clay, mixed with lime-gravel. Only in room 6 was the covering layer of lime, 3-4 cm. thick of the same kind as the house-floor in the Alambra house.

The walls of rooms 1 and 2 are very much demolished and as a rule reach only 0.20 m. above the second house-floor. Room 1, measuring 2.10 m. in length and 1.78 m. in width, is a vestibule leading to the inner room 2. On account of a northern divergency of the north wall the length of the vestibule is not constant, measuring 2,15 m. at its eastern and 2.05 m. at its western end. Its front is open to the east. To the west an entrance, 0.68 m. wide, opens into room 2 at the right end of the wall. The southern wall of room 1 does not join the eastern wall of room 3, but the end of the latter is just in flight with that of the former. The small interstice of 0.10 m. between the walls is doubtless due to a sinking of the

1 The house-floors are counted from the surface, as they appeared during the digging.

ground, which moved the east wall of room 3 eastwards 0,10 m. out of its original direction. Taking this sinking of the ground, also conspicious in the sinking of the house-floor of room 6,1 into account, the east wall of room 3 originally should have touched the southern end-point of the south wall of room 1. Between the end of the two walls a rather flat stone was found measuring about 0.45 m. in length and width. At the corresponding place at the end of the northern wall of room 1 no such stone was found, but that does not testify against its original existence there, as the wall at this point is demolished much below the level of the second housefloor. These stones cannot have been anything else than bases for wooden pillars, supporting a wooden beam. In this way the walls were joined, and at the same time these pillars served as parastades, protecting the wall-end of the open vestibule from rain, and reducing the pressing weight of the roof-beams on the walls at their weakest point.

Considerable amounts of charcoal, found between the north and south walls of room 1, bear evidence to the former presence of such pillars.

Room 2 measures 2.65 m. in length, averaging 1.95 m. in width. According to the northern divergency of the north wall the width varies betwen 1.90 m. at the western end of the room and 2.00 m. at its eastern end. The western corners of the room curved, as already mentioned; but only the corners, not the wall between them. The rooms are a closed unit, communicating with each other but with no other part of the building, and form together the well-known megaron type, in its simplest form consisting of a inner main room and an open vestibule, or ante-room. Usually the main room is quadrangular, but there are some specimens of a horse-shoe type from Greece. As a rule the entrance, leading from the vestibule to the main room, opens in the middle of the cross-wall, though by no means always exactly in the middle; and sometimes, as here, it opens at the end of the cross-wall. Nothing in the room indicated its use. Only fragments of domestic pottery were found.

2

1 See section fig. 4, p. 32.

2 References at Fimmen: Die kret.-myk. Kultur p. 41 sqq; Blegen: Korakou, Boston and Newyork 1921 p. 76 sqq.

3 E. g. the door-way between the megaron and opisthodomos in a house of Troy II: Dörpfeld: Troja und Ilion, Athens 1902, the door-way I on fig. 23, p. 81.

Leaving room 1 and turning to the south, we approach the main entrance of the house, measuring 1.58 m. in width. This considerable width makes it most probable that the door was a double door, with two wings meeting in the middle.

Through the door we enter room 5, measuring 3.85 m. in length and 3.15 m. in width. That it had no roof is proved by the fact that fire-making and cooking were practised along the walls, especially the north and east walls, where a 0,20 m. deep layer of ashes and carbonized matter was found. The smoke has blackened the wallsides and the heat has baked the underlying clay-floor. Several pieces of cooking vessels, blackened by the smoke, give further evidence of the cooking. Under such circumstances roofing is impossible, as such cooking along the walls could not, on account of the smoke, be practised elsewhere than under the open sky. If the room' had been roofed it would have been necessary to concentrate the cooking to one place, from which the smoke was led out into the open air through a hole in the roof. Besides the cooking-pots a great number of other domestic vessels, such as jugs and jars, were found, together with two saddle-querns of the usual type, and a round hard mass (about 0.15 m. in diameter) of red colouring material, found outside one of the broken pots.

We thus see that the room was an open inner court, where the food was prepared and cooked. It could also serve as a source of light for the surrounding rooms.

West of the inner court is room 6, measuring 3.35 m. in length and 2.80 m. in width. The entire front side of this room opens on the court, having no dividing wall. The border between the two rooms is marked by a row of small flat stones, west of which the lime-floor of room 5 is laid out, while the floor of the inner court is made of trodden earth.

The walls do not meet each other at right angles but run approximately parallel, the northern parallel to the southern, and the western parallel to the stone setting, the room thus forming a rhomboid. The southern wall of the room does not join the southern wall of room 9 but ends blindly, only touching its western point, as shown by the map; the same peculiarity as observed at the joint of the south wall of rooms 1-2 and the east wall of room 3. The explanation is the same here: a wooden beam, resting on two wooden parastades at the ends of the southern and northern