12.') 6 VALLUM E0JLANU5L Tlie wall was everywhere accompuuied on its iiiirtbern side by a broad and deep fosse, which may still be traced, with trifling interruptions, from sea to sea, even where the wall has quite disappeared. It traverses indiflerently alluvial soil and rocks of sandstone, limestone, and basalt. Thus, on Tapper Moor, enormous blocks of whinstoue lie just as they were lit ted out of the fosse. East of llcddou on the Wall, the fosse is 34 feet wide at the top, 14 at tiie bottom, and about 9 deep. In some places it is 40 feet wide at the top, and in others 20 feet deep. Hodgson, in his History of Northumberland (iii. p. 276), states a fact curious if true: " A little W. oi Port(/ate, the earth taken out of the fosse lies spread abroad to the N. in lines, just as the workmen wheeled it out and left it. The tracks of their barrows, with a slight mound on each side, remain unaltered in form." It is scarcely credible, how- ever, that slight elevations of earth, and superficial traces in it, should, for more than a thousand years, have successfully resisted the constant operation of the natural agencies which are sufiieient to disinte- grate the hardest rocks. The Valluji, or earth wall, is uuiforndy S. of the stoue wall. It consists of tln'ee ramparts and a fosse. One rampart is close to the S. edge of the ditch. Of the other two, which are considerably larger, one is situated N., the other >S. of the ditch, at the distance of about 24 feet from it. These larger ramparts are even now, in some places, 6 or 7 feet liigh. They are composed of earth, in which masses of stone are often imbedded, for the sake of which they are sometimes i[uarried. The fosse of the vallum was probably snialler than that of the murus. Iso outlets through the S. lines of fortification have been discovered ; so that the gateways of the stations appear to have originally been the only means of communication with the countiy. At distances averaging nearly 4 miles, stationary camps were erected along the line. Some of these, though connected with the wall, were evidently built before it. The stations are four-sided and nearly square, but .somewhat rounded at the corners, and contain an su'ea averaging from 3 to 6 acres, though some of them are considerably larger. A stone wall, about 5 feet thick, encloses them, and was probably in every instance strengthened by a fbsse and one or more earthen ramparts. The stations usually stand upon ground with a southern inclination. The great wall either falls in with the N. wall of the stations, or else usually conies up to the N. cheek of their E. and W. gateways. The vallum in like manner generally approaches close to the S. wall of the stations, or comes up to the S. side of the E. and W. portals. At least three of the stations, however, are quite detached from both Hnes of fortification, being to the S. of them. These may have been erected by Agricola. Narrow streets intersecting one another at right angles traverse the interior of the stations ; and abundant ruins outside the walls indicate that ex- tensive suburbs were required for the accommodation iif those connected with the soldiers stationed in the camps. The .stations were evidently constructed with exclusive reference to defence ; and hence no traces of tesselated pavements or other indications of luxury and refinement have been discovered in the mural region. According to Ilorsley, there were IS stations on the line of the wall, besides some in its innnediate vicinity; VALLUM ROMANUM. but Hodgson reduces the number to 17, believing that in one instance Horsley mistook a mere tempo- rary encampment for a station. In ascertaining the number and names of the stations, our principal literary authority is the No- tltla Imperii, supposed to have been compiled about the end of the reign of the emperor Tlieodo.sius the younger. The 69th section of this document con- tains a list of the prefects and tribunes under the Duke of Britain : the portion relating to our subject is headed, " lievi per Uneani Valli," and contains the names of 23 stations, evidently arranged in their order from E. to W. The heading, however, mani- festly implies, not, as it seems sometimes to have been interpreted, that all the stations were actually on the line of the wall, but that they were along it, that is, parallel to, or at no great distance from it. It is clear, therefore, that as remains of stations exist both to the N. and to the S. of the wall, as well as ac- tually on its line, nothing but the remains themselves can enable us to name the stations with certainty. Now the first 12 stations mentioned in theNotitia have been accurately identified by means of inscrip- tions found in the ruins of the stations. Of these we subjoin a list, with the ancient and modern names, taken chiefly from the plan prefixed to Mr. Bruce's work : — Segedunum - - Wallsend Pons Aelii - - Neiccastle. Condercum - - Bemcell. Vindobala - - Rutchester. Hunnum - . - Halton Cliesters. Cilurnuni- - - WalwicJc Chestcrs. I'rocolitia - - - Carraicburgh. Borcovieus - - Housesteads. Vindolana - - Little Chesters, or Chesterholm. Aesica ■ _ - Gi'eat Cheslers. JNlagna (Magnar) - Carvoran. Amboglanna - - Birdoswald. All these are on the actual line of the wall, except Vindolana and Magna, which are a little to the S. of it. West of Amboglanna no evidence has yet been discovered to identify any of the stations ; and it is to be feared that many antiquities which might have enabled us to do so have been destroyed ; for it ap- pears that the country people, even quite recently, regarded stones bearing inscriptions as " unlucky," calling them "witch-stones," the evil influence of which was to be extirpated by pounding them to powder. Besides this, stone is scarce in that jjart of the country ; and hence the materials of tl;e wall and stations have been extensively employed in the construction of dikes and other erections in the neighbourhood. It appears from the plan already referred to that there were stations at the places now called Cambeck Fort, Stamoix, Burgh, Drumbnrgh, and Bowness; the first a little to the S., all the rest on the line of the wall. Of the remaining eleven stations mentioned in the Notitia, the plan identifies Alionis with Whitky Castle, some miles S. of the wall. ^Ir. Bruce places Bremetenracum a little W. of the village of Bramp- ton ; I'etriana, he thinks, is probably the same as Cambeck Fort. It is possible that something may yet be done to elucidate what is still obscure in connection with these most interesting monuments of Eoman Britain ; and the Duke of Northumberland had, in 1S53, given
