The spaces below the aisles in the third bay from the west form chapels--that on the right dedicated to S. Martin, and that on the left to S. Gilles. Before the last extension of the building these chapels were at the extreme west end. They have western doorways, which still retain the wooden doors. Each of these doors was of four divisions in height, covered with subjects carved in low relief. They are executed either in cedar or oak (I am uncertain which, for they are covered with paint), and the subjects, inscriptions, and borders are all obtained simply by sinking the ground three-sixteenths of an inch. The figures are, of course, only in outline, but it is still evident that they were carefully painted with draperies, etc., so as to be thoroughly distinct. There is some appearance of the ground having been painted with broad horizontal bands of colour, but the traces are so indistinct that it is difficult to speak positively.[44] The doors are hung folding, and those to the chapel of S. Gilles contain subjects from the early life of our Lord, whilst those in the chapel of S. Martin contain subjects from His Passion. The meeting-rail in the former fortunately contains an inscription of extreme value: “Gaulfredus: me: fct: Petrus: epi”; after which some letters are lost. If my reading of the last letter but one as “p” is correct, I think it leads to a most important inference. No one who looks at the design of these gates can doubt that they are thoroughly Eastern in their character; and, upon searching for the lists of bishops of Le Puy since my return, I was delighted to find that the first bishop of the name of Peter[45] was consecrated at Ravenna by Leo IX. in A.D. 1043, and died at Genoa A.D. 1053, as he returned from the Holy Land. Gates of the same description are said to exist in the churches of Chamaillères and of Lavoulte-Chilhac in the same district, whilst other evidence of intercourse with the East is afforded by fragments of _tissus_ preserved at Monestier, at Pébrac, and at Lavoulte-Chilhac. These _tissus_ are all extremely Eastern in their character, and very similar to the famous cope at Chinon described by M. de Caumont in the _Abécédaire_, and to the Le Mans _tissu_ described by M. Hucher in the _Bulletin Monumental_ (1846, p. 24). The date ordinarily attributed to them is the middle of the eleventh century, which exactly tallies with the return of Bishop Peter from the Holy Land. I dwell on this the more because, if the inference I have drawn from the inscription be true, it gives the date also to the second portion of the construction of the cathedral, to which the chapels in the porch undoubtedly belong; and the result would be that whilst I should date the earliest portion of the church at about the end of the tenth century, or quite the commencement of the eleventh, the second portion would be dated at about A.D. 1050; and, finally, there is little doubt as to the whole having been completed in the