^{page 176 note 1} In a leaflet issued by the York Reman Excavations Committee an account is given of the formation of the Committee and acknowledgement is made of much valuable assistance given to the Committee and to the writer of this report.

^{page 177 note 1} The difference in level would not be more than 6 ft. and may have been rather less, for the outer face of the wall round the east corner showed a chamfered projection, which would naturally be taken to be a cornice crowning the wall proper (see below, p. 191), 6 ft. above the level of the cobbled surface inside the north-east wall; and the top of the wall may have been stepped up a foot or two at the corner.

^{page 177 note 2} About 2 ft. Cf. Köchly, and Rüstow, , Griech. Kriegsschriftst., ii, 2, p. 94Google Scholar: Τὰς δὲ έπὶ τῶν τειχῶν ὲπάλξεις ἐγγωνίους γίνεσθαι, ὥστε ὑποβλέϕαρα ἔχειν οὐκ ἔλαττον ἔχοντα το βάθος σπιθαμῶν τριῶν, etc.

^{page 178 note 1} I am here working upon a suggestion made to me by Mr. A. O. Curle.

^{page 178 note 2} Der römische Limes in Österreich, Heft xi, fig. 3; Heft xiii, fig. 37 and fig. 38, Schnitt A—B.

^{page 178 note 3} The eight pieces not illustrated are: from under undisturbed stratum of cobbles and gravel, four fragments of Samian ware, viz. a piece showing lip and deep plain margin of a heavy bowl of form 37, part of a large cup of form 27 with the flattened contour characteristic of late examples of the type, a bit of a cup of form 33, and part of the side of a platter of form 31; from dip between wall and clay bank, a fragment of a Samian platter of form 31, a fragment of a bowl that seemed to have been of the type Ludowici So, a piece of a pot of yellowish-grey ware with a recurved rim like fig. 94, no. 6, and a bit from the body of a rough-cast beaker coated with a brown slip.

^{page 180 note 1} Haverfield, , Eph. Epig. ix, p. 538Google Scholar; Haverfield and Macdonald, Roman Occupation of Britain, p. 154.

^{page 181 note 1} If the relation of the clay rampart to the wall in Section C-D was the same as its relation to the corresponding wall in Section G-H (see below), we must take 10 ft. as the minimum height of the rampart-walk (supposing that the earth bank and the clayey mass in front of it never stood higher than they do now), and we may allow the palisade to have added another 5 ft. Again, on the evidence of our Section G-H, we must suppose that the front of the clay rampart extended quite 6 ft. beyond the line of the inner face of the wall. If the earth bank behind be allowed a rearward slope of 35°, the back of the rampart would be some 35 ft. behind the line of the inner face of the wall. We should then get a total breadth for the clay rampart of quite 41 ft. That would be a massive structure, but not much more massive than the stratified ramparts of such auxiliary forts as Lyne and Camelon.

^{page 181 note 2} This is given simply as an illustration of what is said in the text. A well-founded restoration can only come, if it come at all, by more excavation.

^{page 183 note 1} I am indebted to Messrs. Brierley and Rutherford, the architects, for a plan from which fig. 98 is derived.

^{page 187 note 1} The outer edge originally lay 5½ ft. further out than is shown in Section G-H. See below, p. 190.

^{page 188 note 1} At Niederbieber the foundation of the wall was laid deeper than the bottom of the ditch (Bonn., Jahrb. cxx, p. 263Google Scholar).

^{page 189 note 1} Similar piles have been found from time to time at various points along the line of the fortress wall.

^{page 190 note 1} C.I.L. vii, 241.

^{page 190 note 2} The dimensions agree fairly closely with those of the ditches of the legionary fortresses of Novaesium and Carnuntum.

^{page 190 note 3} A spread of gravel ran along the outside edge of the ditch at Novaesium (Lehner, Novaesium, p. 213).

^{page 193 note 1} About that time there was trouble enough, both external and internal, to account for extensive damage to the wall. If the disaster did not occur until after the middle of the fourth century and if the reconstruction dates to the same period as the establishment of the signalling stations on the Yorkshire coast, the existing remains in the Museum Gardens would be impressive evidence of the thoroughness of the Theodosian restoration. Cf. Ammianus Marcellinus xxviii. 3: In integrum restituit … castra multiplicibus quidem damnis adflicta, etc.

^{page 193 note 2} Apparently with similar, but much smaller, towers between. For a projecting polygonal tower between the south-west gateway and the west corner see G. Benson, Later Mediaeval York p. 161. There is nothing to show that the tower of which ‘some indications’ were discovered at the end of Aldwark, near the east corner (Wellbeloved, Eburacum, p. 50) was of the same character.