The French took a different course from all these. They built circular churches like other nations, apparently in early times at least, which were intended to stand alone; but in no instance do they appear to have applied them as naves, nor to have added choirs to them. On the contrary, the clergy always retained the circular building as the sacred depository of the tomb or relic, the Holy of Holies, and added a straight-lined nave for the people. Of this class was evidently the church which Perpetuus built in the fifth century over the grave at St. Martin at Tours. There the shrine was surrounded by seventy-nine pillars arranged in a circular form: the nave was lined by forty-one—twenty on each side, with one in the centre of the west end as in Germany. When the church required rebuilding in the 11th century (1014?), the architect was evidently hampered by finding himself obliged to follow the outline of the old basilica of Perpetuus, and having to labour on the same foundation so as not to disturb either the shrine of the saint or any other place which had become sacred in this, which was the most celebrated and revered of the churches of Gaul. All this is made clear in the plan of the new church (Woodcut No. 575). The arrangement of the circular part and the nave exactly accord with the description of the old church, only that the latter has been considerably enlarged according to the fashion of the day. But the juxtaposition of the two shows how nearly the chevet arrangement was completed at that time. [Illustration: 575. Plan of St. Martin at Tours. Scale 100 ft. to 1 in.] Another church, that of Charroux, on the Loire, looks as though it had been built in direct imitation of the church of Perpetuus. The round church here retains its pre-eminence over the nave, as was the case in the older examples, and thus forms an intermediate link between the old church of St. Martin, which we know only by description, and the more modern one, of which a plan is given (Woodcut No. 575). St. Bénigne, Dijon, is another transitional example which may serve to render this arrangement still more clear. It was erected in the first year of the 11th century, and was pulled down only at the Revolution; but before that catastrophe it had been carefully measured and described in Dom Plancher’s ‘History of Burgundy.’ As seen by him, the foundations only of the nave were of the original structure, for in the year 1271 one of its towers fell, and so damaged it that the whole of that part of the church was then rebuilt in the perfect pointed style of the day. Without entering too much into detail, it will suffice to state that the part shaded lightly in the woodcut (No. 577) is taken literally from Dom Plancher’s plan, regarding which there can be no doubt, and the contemporary descriptions are so full that very little uncertainty can exist regarding the dimensions and general disposition of the nave.