^{page 2 note 1} For the plan of the Roman city see R. E. M., and Wheeler, T. V., Verulamium, A Belgic and two Roman Cities, pl. CXIXGoogle Scholar; or Corder, P., Antiquity, xv, June 1941.Google Scholar

^{page 2 note 2} St. Albans & Herts. Archit. & Archaeol. Soc. Trans. 1953.

^{page 4 note 1} There were no surface indications in this field; but in the field north-west of Bluehouse Hill slight superficial traces exist.

^{page 5 note 1} A Samian sherd from layer 18 fitted one from layer 11.

^{page 5 note 2} R. E. M., and Wheeler, T. V., op. cit., pl. cxix.Google Scholar

^{page 5 note 3} Ibid., p. 75.

^{page 6 note 1} Op. cit., pl. xx, Section A-B, where the bank seals an offset in the wall.

^{page 6 note 2} I have had the benefit of Mr. B. R. Hartley's preliminary examination of this material. He has picked out nine pieces as having been manufactured after c. A.D. 160.

^{page 6 note 3} However, the gap is not large. That the wall had been in existence some years by A.D. 235 seems certain from the coin hoard found buried in the floor of a wall-tower, op. cit., 62.

^{page 7 note 1} Gratitude is due to Mr. J. Lunn for salvaging the inscription fragments and realizing the importance of the site: through some mischance it was not scheduled.

^{page 8 note 1} On the comparable inscription from Wroxeter (Atkinson, Wroxeter, pp. 177 ff.) the heights (as measured by the present writer) are in l. 1 9⅛ in., 1. 2 8½ in., 1. 3 7½ in., 1. 4 6¾ in., 1. 5 6 in. The die of the inscription is 11 ft. 10 in. wide by 3 ft. 9¾ in. high.

^{page 9 note 1} Part of a superscript bar above a numeral has survived before DESI[GN.

^{page 9 note 2} Mr. Birley, E. kindly suggested this title from the analogy of Vollmer, Inscr. Bavar. Rom., no. 257 a, b (Koesching)Google Scholar. See also ILS, 267.

^{page 10 note 1} Professor R. Syme kindly suggested this. The shorter form, PR PR, is used on the contemporary inscription moulded on the water-pipe found at Chester (EE ix, 1039), but there may have been more reason for brevity on a pipe than on a monu mental inscription.

^{page 10 note 2} The left part of the serif is only in. narrower than that on two R's in 11. 5 and 6, and is equal to the same serif on E, which is the nearest extant letter comparable to L.

^{page 10 note 3} The restoration conventus is hypothetically possible for an assize-area, but it is hard to see why such a group, otherwise unattested in Britain, should have built an important public building in a town which, because of its proximity to London and Colchester, was hardly its headquarters. The letter N seems less likely as the comparable N in ]NATA[ has a less pronounced backward serif.

^{page 10 note 4} The present writer thanks Mr. J. Lunn, Director of the Verulamium Museum, for kindly granting access to the fragments. He also makes grateful acknowledgement to Mr. E. Birley for discussing some of the problems in the restoration.