earliest period, Londinium, now London. Goods appear to have been shipped from the Continent direct to London, and the wrecked cargo of Samian pottery found on the Pan Rock in the mouth of the Thames estuary (p.163) is doubtless one of the relics of such traffic. Passengers probably used the short crossing from Boulogne to Kent, where Richborough was a frequented port, and Lympne and Dover were used as landing—places. Other routes were not unknown. Troops seem to have been not seldom sent by a long sea passage from Fechten or other German port on the Rhine estuary direct to northern Britain. The discovery of a pig of Mendip lead near the mouth of the Somme suggests passages across the Channel at points west of its narrowest portion, and in this, as in other early ages, there was occasional intercourse between the south-west coast of Britain and the opposite shore of the Continent.
   Lastly, the roads. In considering these we must put out of our minds the Four Great Roads which are named in one or two documents and chronicles of the twelfth and succeeding centuries: Watling Street, Icknield Street, Ermine Street, and Fosse. This category of Four Roads appears to be an invention of lawyers and antiquaries utilizing early English road—names which they knew from charters or otherwise. Certainly Icknield Street, which runs along the Berkshire downs and the Chilterns, is, for most of its course, neither Roman in origin nor Roman in use, and the notion of Four Great Roads is alien to all that we know of the Roman road system in Britain. In the south and midlands of the province, with which alone we are here concerned, we can distinguish five roads or groups of roads. Like the modern railways, which indeed they much resemble, they radiate principally from London. One road ran south-east, through Rochester and Canterbury, to the Kentish ports. A second ran west to Silchester, and thence by various branches to Winchester and Exeter, to Bath and Gloucester, to Herefordshire and south Wales. A third, known since Saxon days as Watling Street, crossed the midlands north-westwards to Wroxeter, near Shrewsbury, and gave access to the fortress at Chester and to the military districts of north Wales and northwestern England : by a branch it also led to Leicester and the north-east. A fourth road ran to Colchester and the eastern counties, to Lincoln, York, and the military districts of the north-east. To these four roads, which start from London, we must add two which do not touch that town, but which connect the north-east of the province with the south-west: the Fosse, which joins Lincoln and Leicester with Bath and Exeter,³ and the Rycknield or Icknield Street, a road of somewhat uncertain course and of very puzzling early English name, which connects south Yorkshire and Derby with Gloucester-shire. These must be understood as being the main roads, divested of branches and intricacies for the sake of clearness, and placed in a category by themselves. It will be obvious that the province possessed an adequate supply of internal communications.
   Such, in the main, was that large part of Roman Britain in which ordinary civilized non-military life prevailed: the lowlands of the south, the east, and the central plain. It was permeated by the simpler forms of Roman civilization, but it lacked its higher developments. It was not devoid of
   ³ For a discussion of the suggestion that the Fosse may have originated as a Roman frontier-line in the early years of the conquest, see R. G. Collingwood, Journal of Roman Studies, xiv, 252.