A reference to my ground-plan[325] will show that the church remains very much in the state which the documentary evidence describes. The choir has nine chapels round its chevet, as described, and has lofty arches, a series of very small openings in lieu of triforium, and a clerestory of two-light windows, of decidedly late but still good Middle-pointed character. The columns, in the usual Catalan fashion of this age, are clusters of rather reedy mouldings, with no proper division or subordination of parts, and consequently of poor effect, and there is no division by way of stringcourses above or below the triforium. On the exterior the east end is not seen to much advantage, as it is built into and against a steep hill, so that at a distance of a few feet only the eye is on a level with the top of the walls of the chapels round the apse. The roofs, too, have all been modernized and lowered. The only peculiarities here are a series of trefoiled openings, just under the eaves of the roof, into the space over the vaulting, and perhaps devised for the purpose of ventilation: and the gurgoyles projecting from the buttresses, which are carved and moulded stones finished at the end with an octagonal capital, through the bottom of which the water falls, and which almost looks as if it were meant for the stone head of a metal down-pipe. When the choir was built, some considerable portions of the church consecrated in A.D. 1038 were left standing. The nave was probably entirely of this age; and a portion of what was no doubt one of the original towers still remains on the north side, between the cloister and the nave. This tower has pilasters at the angles and in the centre, and is divided into equal stages in height by horizontal corbel-tables. An apse of the same age remains on the east side of what seems to have been the south transept of the early church: and from its position we may, I think, assume with safety that the church was then finished with three or five apses at the east, very much as in the church of San Pedro, close by, which I shall have presently to describe. In addition to these early remains there is also a magnificent and all but unaltered cloister. I cannot find any certain evidence of its exact date, though it seems to have existed in A.D. 1117, when an act of the Bishop Raymond Berenger was issued in the “cloister of the cathedral.”[326] The character of the work confirms, I think, this date. The plan is very peculiar, forming a very irregular trapezium, no two of the sides being equal in length. It has on all four sides severely simple round arches carried on coupled shafts: these are of marble, and set as much as 20 inches apart, so as to enable them to carry a wall 3 feet 1½ inches thick. This thickness of wall was quite necessary, as the cloister is all roofed with stone, the section of the vaults on the east, west, and south sides being half of a barrel, and on the north a complete barrel vault.