Hutchinson was struck with the linear character which the ruined streets of this fort had in his time, and was reminded, by their appearance, of the arrangements of the Polybian camp. This will be observed in a greater or less degree in all the stations, and there cannot be a doubt but that the dwellings were arranged in rows parallel to the four sides of the stations, and hence, intersecting each other at right angles. It was necessary that the Roman camp, whether of a temporary or permanent character, should be nearly uniform in its plan. If the troops rested but for a night, each man knew the part he had to fill in preparing the fortification, and could set about it at once; in the event of a sudden attack in the darkness of the night, each knew his position, though he may never have rested upon the spot before. Suburban buildings have occupied the space between the station and the river, and ruins more extensive than usual are spread over the ground to the south. There is no appearance of any habitations having been erected to the north of the Wall. Whenever the surface of the contiguous ground is broken, fragments of Samian ware and other marks of Roman occupation appear. [Sidenote: CHESTERS.] Two remains of great interest are found within the station. One of these is an underground vault near the middle. Its masonry is rough, and somewhat peculiar; the sides incline slightly inwards, but the roof, instead of being uniformly vaulted, is formed of three ribs arched in the usual manner, and the intervals between them are in technical language—‘stepped over,’ that is, the stones of each course are made to project inwards a little, until, at length, one laid on the top completes the junction. The woodcut, which is here introduced, together with the following extract from Hodgson’s description of it, will give a tolerably correct idea of this curious structure. [Illustration: Vault at Cilurnum] This vault, when it was first found, was supposed to have been the Ærarium of the station. Between the joinings of the floor, which were of thin free-stone flags, were found several counterfeit denarii, both of copper and iron plated with silver. The approach to it was by four steps downwards, the lowest of which was a large centurial stone, which had borne an inscription, but nearly all of it had been purposely erased. On the outside of the threshold was found, in a sadly decayed state, its original door of wood, strongly sheathed with plates of iron, and the whole firmly rivetted together with large square nails. Within the door, which had opened inwards, the end wall was two feet thick, plastered and painted. Its internal area is ten feet by nine, and its height to the crown of the arch six feet four inches.[85] [Illustration: Ground-plan, Hypocaust, CILURNUM.] [Illustration: Hypocausts at Cilurnum]