A HISTORY OF NORTHAMPTONSHIRE on or off these roads has been treated as comparatively immaterial and unworthy of serious discussion. This is to invert the true relation of the two subjects. In this as in other volumes of the Victoria County History we have preferred to describe the sites first and proceed from them to the roads. Our sources for determining the roads are of two kinds — written and archjEological. The archaeological evidence is supplied by actual remains, as when we dig up ancient metalling along a line where a Roman road might be expected, or when we find a still existing track which runs with persistent straightness from one Roman site to another. The written evidence is more elaborate. Charters tell us of ' streets ' bounding estates in early days. Placenames like Stratford, if of estab- lished antiquity, suggest ancient and usually Roman roads.* Parish and county boundaries sometimes preserve curious information. But our chief written evidence is the Itincrarium Antonini, a Roman roadbook which gives the distances and ' stations ' along various routes in the empire. Its exact age and its object are uncertain and do not now con- cern us ; its accuracy, which matters more, is by no means unfailing, and it is sometimes more useful in testifying that a road ran in a particu- lar direction, as for instance from Colchester to Lincoln, than in telling us the precise course of the road and the precise sites of the ' stations ' along it. For our present purpose two of the Itinerary routes are important. We give the distances as given in the original in Roman miles, thirteen of which may be reckoned as equivalent to twelve English miles. (i) Part of route from Carlisle through Wroxeter and London to the Kentish ports : Venonae (High Cross, Leicestershire) to Bannaventa, 17 miles; B. to Lactodorum, 12 miles; L. to Magiovinium, 17 miles; M. to Durocobrivae, 12 miles [Itin. Ant. 470, 471). This route recurs in a route from London to Lincoln : Durocobrivae to Magiovinium, 1 2 miles; M. to Lactodorum, 16 miles; L. to Bannaventa (misspelt Isanna- vantia), 12 miles; B. to Tripontium, 12 miles ; T. to Venonae, 8 miles (///>/. Ant. 476, 477); and again in a route from York to London: Venonae to Bannaventa, 18 miles; B. to Magiovinium, 28 miles; M. to Duroco- brivae, 12 miles [Itin. Ant. 479). The three versions agree substantially. (2) Part of route from London by Colchester to Lincoln and the north: Camulodunum (Colchester) to Villa Faustini, 35 miles; to Icini, 18 miles; to Camboritum, 35 miles; to Durolipons, 25 miles; to Duro- brivae, 35 miles; to Causennae, 30 miles; to Lindum (Lincoln), 26 miles, or according to a less well attested reading, 16 miles {Itin. Ant. 475). With this we may compare a list of names given by the Ravenna Geographer (429-30) : Manulodulo Colonia, Durcinate, Duroviguto, Durobrisin, Venta Cenomum, Lindum Colonia. Corrupt in spelling as these names are, we may regard them as somewhat the same as the Itinerary names. ' Portvv.iy must not be included among these placenames ; it does not necessarily or usually denote a Roman road. 202
