additional objection of crossing the Rother at a more unfavourable spot.
The entire length of road between Pevensey and Limme being thus forty miles, and, allowing for inevitable digressions, more probably forty-five, that would be beyond a single day's journey for any body of soldiers under ordinary circumstances, in the regular routine of change from one station to another. Some halting-place therefore between the post in Kent and that in Sussex must have been requisite in order to complete the chain of communications, essential to maintaining the military occupation of the country; and for such an outpost the supposed site of Anderida at Newenden, previously described, is not unlikely to have been selected; with the manifest advantage of finding there defensive (British) works already provided, together with roads, such as they were, diverging thence on every side, counterbalanced only by the inconvenience of adding a few miles to the march. Possibly Newenden might have been reached from Staple Cross by Northiam, or if, as conjectured above, the Roman intercourse from Pevensey with Limme passed through Sandhurst, it would thence (most directly) proceed toward Tenterden either partially or entirely along still-existing bye-lanes from Ringle Crouch Green at the eastern extremity of the village to the western side of that of Rolvenden, or else by some other line long since obliterated. Again, from Sandhurst the distance is about three miles, very nearly straight, and as nearly level, to Newenden Castle Toll, whence however the road must have retrograded in some degree through Rolvenden in order to reach Tenterden, unless it should have been practicable, which is doubtful, to cross the adjoining marshy valley north-eastward, and pursue the left bank of the old channel of the Rother by Small Hithe and Reading Street to Apledore, or rather to Apledore Heath, the remainder of the route being then free from serious obstructions by Ham Street, Aldington, and Court at Street to Limme.
And here the question for consideration occurs, whether the names Reading Street, Ham Street, Court at Street in Kent, and Boreham Street in Sussex justify the inference, that the Roman road passed through those places? That this is actually the fact with regard to Court at Street is affirmed by
