ROMANO-BRITISH SUFFOLK no doubt from 1 6 ft. to 20 ft., they form one mass with it. It seems probable that the builders decided on the additional defences when they had partly carried up their wall, and that the bastions were then added, their upper portion becoming one with the wall behind them. The wall and bastions were of the usua.1 construction for such work, a core of rubble with a facing of faced flints (in this case) with, at short intervals, lacing courses of brick. Occasionally the mortar appears coloured with brick-dust, which is used capriciously, and it is doubtful if there was any particular reason for its use. The greater portion of the area of the fortress was on high level ground, but west of the gates in the north and south walls it sloped rather steeply, as has been said, to the river. The walls named were carried down this slope to join the west wall, which adjoined the river whose waters perhaps then ran nearer this wall than at present. Of this wall nothing is visible. The remains of its foundations were however uncovered in 1859 by that able antiquary the late Henry Harrod, and the existence of a wall similar to the others sufficiently proved,* It was pushed as far forward towards the river as the marsh would allow. How it disappeared is clear enough, for it was used as a quarry, being easy to get at, during the Middle Ages, Tiles from it may be found in the fabric of the church at Reedham on the Yare, not more than two miles from Burgh, and no doubt in other places. The south wall of the fortress will be seen by a reference to the plan (plan II) to beat an obtuse angle with the east one. This looks like an intention to get the south-west angle of the inclosure as near the river bank as possible. The distance between the banks at this south-west corner is roughly 330 ft., at the corresponding north-west angle it may be 600 ft. The distances here named are by no means beyond those attainable by the darts of the larger ballistae used by the Romans in the field, and there is nothing improbable in supposing that such ballistae might have been planted on platforms backing the river wall, of such power as effectually to command the passage of the river opposite the fortress.^ The gates are three in number, as has been said, but of the south gate only a trace remains. Those in the north and south walls were only 5 ft. wide and offer no peculiarities of construction. The eastern gate, the prin- cipal one, in the centre of the eastern wall, is commanded by bastions on either side. In a line with the inside face of the wall across the gateway a trench was found in the soil 15 in. wide, a clear indication of the former existence of a timber threshold. Inside the gate on either hand were the remains of a low wall with a return at a distance of 10 ft. 7 in. from the threshold. It is possible these were vestiges of the walls of the guard-houses. The gateway was no doubt a single arch, the rampart walk passing over it, and it was c'.osed by a door of two valves in the usual fashion. As in other fortresses of the same date there seems little trace of any ditch. Along the east front there is no sign of such a feature. Some signs of this ° See 'Norf. Arch. 1859, v, 146 et seq. ' See for such a platform an inscription in Corpus Inscr. Brit. Latinae, vii, no. 104.5. T'** French have made experiments as to these engines. In one of the halls of the Musee des Antiquites Nationales de St. Ger- main near Paris may be seen full-size working models of balRstae of several sizes, restorations made from the plans of General de Reffye. All are presumably for use in the field, and the largest can throw a dart, according to the weight, for a distance from 150 to 310 metres. Larger and more powerful engines of the sort might have carried bolts with effect to still greater distances than those named. 285
