foundations were exposed lying beneath the hedge on the north side, and here the presence of the three ditches underneath the rapidly sloping surface was ascertained. On the other sides nine sections in all were made. In these the black deposit clearly marked the outline of the ditches off from the subsoil in which they had been cut. For the most part they appeared to be V-shaped, scarp and counterscarp lying at the same angle. But in that portion of the inner ditch which was cleared out on the west front, as well as in that which masked the west gate, the sides became perpendicular about eighteen inches or two feet from the bottom, which, again, was about a foot in breadth and quite flat. The same type of ditch was noted at Bar Hill.

In other respects the different sections displayed a considerable want of uniformity. Thus, along the east front the breadth of the berm, or space between the wall and the edge of the innermost of the triple line of ditches, varied from three and a half feet to six feet. To the north of the gate on this side, owing to the deflection of the outermost ditch, the space between the outer face of the wall and the edge of the counterscarp of the outermost ditch measured ninety-three feet, whereas to the south of the gate the corresponding dimension was only sixty-six feet. The normal width of the innermost ditch was here probably about twelve and a half feet, and its depth twelve to fourteen feet, the two outer ditches being smaller and shallower. In the section taken diagonally across the south-east corner, the middle ditch proved to be eleven feet wide by eight feet deep, and the outer ditch thirteen feet wide by nine feet deep. Generally, it may be said that the middle ditch had been less important than either of the other two. Again, the outline of the sections cut upon the south presented one marked point of contrast with those cut upon the east. While the berm remained about the same, the sectional line was much more zigzag in character, a feature which was also noted in some of the ditch sections at Novaesium.[1] It is probable that this was the result of alterations on the original plan.

Lastly, in both of the southern sections, the outer ditch assumed the appearance of two ditches running parallel and separated only by a slight midrib. Whether this is due simply to changes brought about in the course of periodical cleaning, or whether it was part of the original design, is uncertain. Professor Ritterling has explained a somewhat similar arrangement at Hofheim as being intended to facilitate the fixing, in the bottom

1 Koenen, 'Beschreibung von Novaesium,' Bonner Jahrbücher, Heft 111/112, p. 211, Fig. 12.

30