io8 A History of Art in Sardinia and Judaea. nized as a palace. The superstructures have entirely disappeared, leaving nothing but the foundations, which are perfect throughout, and still preserve, in places, the second course, with a salience of about 60 centimetres (Fig. 294), quite enough to enable the explorer to make out the main dispositions of the building. The plans of T6xier and Barth betray signs of hurry ; that of the former is invested with an appearance of regularity and a state of preservation which are apt to mislead. To look at his walls, we should never suspect that the blocks of the lower course were anything but consistent in size. Again, he does not seem to have observed that the sinister passage of the central hall was closed at the back, a detail which calls to mind Assyrian palaces (11 in plan) ; ^ whilst he left out the north-west angle of the building, perhaps on account of its worn condition. Nevertheless, even on this side, a patient observer is able to trace the walls. On the other hand, subsidiary buildings, a little in front of the main entrance, which are omitted in Earth's plan, are carefully indicated in Texier's albeit somewhat too near the palace. They were built of small stones, and in Barth s estimation would scarcely have been noticed, when juxtaposed with the huge blocks of the larger edifice.^ Further diggings would doubtless bring to light the foundation- stones hidden underground, as well as all the inner dispositions, perhaps remains of pavement, decoration, etc., and enable the architect to make a perfect plan. Our time was too short to attempt doing more than compare on the spot the drawings of our predecessors. On the eve of our departure, we were still at work at 9 p.m., by a clear moonlight, trying to finish our verification. Barth made a complete plan of the building under notice, which we reproduce as being on the whole the more accurate of the two (Fig. 295). To adopt his measurement, the edifice, properly so called, was about 57 m. long by 42 in width ; it formed a rectangle, except at the north-west angle, now disappeared. Some of the foundation- stones are 5, 6, and even 7 m. long, and 2 m. high. The supporting walls are sometimes as thick as the outer wall, sometimes thinner. These walls were obtained by bevelled masonry, i.e. the stones fitted one into the other like carpentry, recalling a similar arrange- ment at Passargae. The vertical faces are rough and rudely cut, ^ Texier, loc. cit., torn. i. Plate LXXX. ^ This is the reason he adduces for not having drawn them.
