with the remains of the vaulting shaft, the space between the gable and the north-east respond of the tower, and at the centre between the two deeply recessed windows there are traces of an altar and the remains of a pedestal with a small fragment of a figure little less than life-size. This has generally been assumed to be part of a seated figure, but that is evidently a mistake, as the size and position of the feet prove, and a careful examination of the face of the pier shows blocks of freestone now cut flush with the wall, on portions of which the head and body of the figure had no doubt been sculptured. Whose statue was this which so evidently formed part of the original design of this arcade? We have no clue.

A few feet above this in the thickness of the wall there is a narrow passage leading from the dormitory to the choir, where its opening is seventeen feet above the choir, probably giving access to a rood loft or to the church, though another access to the church seems to have been provided from a door in this same passage close to the outside of the choir wall, which may have led to the sacristy by a covered passage. The arrangement seems to have been made chiefly with the view of giving access to the choir during the night. The east wall of the sacristy is old, but the west wall is not, and no trace of a door is now to be found in it. Above the passage the thick wall has been carried up to the flat ceiling already referred to, and then the wall is thinned from the inside to the same thickness as the west wall. When this old ceiling existed the transept would have quite a symmetrical appearance, but it has a somewhat lopsided appearance now, as the axis of the roof, as of the gable, is midway between the external faces of the walls-as it has always been.

In the west wall of the transept there is a very beautiful late door, opening into the cloister, having a semicircular head, but it is peculiar in this, that the door check and rear arch are towards the outside, so that the door must open outwards. This is a very unusual arrangement, and there is nothing now to indicate that this door had given access to any porch or other apartment. The cloister has, apparently without any such obstruction, extended continuously round the four sides of the garth, and

the fragments of arches and piers which remain prove that it was an exceptionally beautiful structure, exhibiting many peculiar and interesting features.

Here we find a variety of architectural puzzles hard to be understood. The distance between the church and the refectory is 73 feet; and from the west wall to the west wall of the chapter-house vestibule 66 feet; and the remaining foundations of the cloisters show that they extended round the square, and that the width between the walls was 10 feet 10 inches with slight variations. The remains of capitals, bases, shafts, and archstones preserved in the chapter-house are sufficient to enable us to reproduce the design. The walls-if I may so speakconsisted of a succession of arches resting on coupled shafts on a low base-course or sill, part of which still exists apparently in situ. The total height from the floor to the eaves appears to have been 8 feet 6 inches. The arrangement just mentioned seems to have been followed on three sides of the quadrangle, but on the fourth, namely, the south side, we find instead of the low base or sill a wall about 3 feet 6 inches high of uncertain age. I hope to have an opportunity of examining this wall more carefully by excavations at different points. In the meantime I think there is good evidence of its antiquity. It was probably part of an older cloister, similar to that which still exists at Oransay, where the cloister on the south side differs considerably from the other.

Although I do not feel quite satisfied about this, I may mention, as favouring its probability, that this south cloister adjoins the remains of the wall I have already referred to as probably the oldest portion of the existing remains, and also that it occupies the same relation to the church and the built-up door as the corresponding cloister at Oransay does to the church and the still open door there.

There is abundant evidence that the three cloisters were of much more refined and beautiful design, and that they were not erected before the thirteenth century. They were of a type most unusual here, but not uncommon among the earlier Italian churches, as, for example, at San Zeno at Verona, and at Amalfi. The wall-what little there is of it-is carried on a continuous arcade of small arches (having a span of 2 feet