Two of the roads of Roman Britain, popularly called Fosse Way and Icknield Way, came through our area, crossing just south of Axminster.

The Fosse Way ran diagonally from Lincoln in the northeast to Seaton in the southwest. Fosse means "ditch" and is from Latin fossa, "dug", yet the English word traces back only to about 1400; if the Romans gave this road a name in the style of the Via Appia and Via Flaminia and Via Portuensis, those great roads of theirs in Italy, we don't know it.

Perhaps the Fosse Way was the successor to more ancient ridgeway tracks that followed the line of the limestone Jurassic scarp across England; but the Roman road did so only approximately, since it connected a series of towns in or either side of this line—Lindum (Lincoln), Ratae (Leicester), Corinium (Cirencester), Aquae Sulis (Bath), Lindinis (Ilchester). It reached the sea by availing itself of another smooth northeast-southwest feature, the Axe valley.

The Fosse Way corresponds also to the temporary frontier of southeastern Britain, occupied by the Romans in the first phase of their invasion of 43 A.D. So perhaps it started as a military road, defending against the tribes as yet unsubdued and enabling deployment of troops along this line before the further advance under the general Ostorius Scapula in 47. Indeed, a fosse was often a defensive ditch, so the name could be a memory of this origin.

Roman roads were straight, and then the English twisted and obscured them—that's almost proverbial, and is largely true. (As said by, I think, G.K. Chesterton, "The rolling English drunkard made the rolling English road.") The roads were a network of links between towns, and typically, but not always, each link was laid out in several straight sections between sighting-points. You can see this clearly in the fine Ordnance Survey map of Roman Britain, published in 2001. The Fosse Way includes, in its central section through the vale of Warwickshire, a particularly long almost-straight stretch of more than sixty miles.

These roads, created for imperial purpose by surveyors and engineers and troops giving orders to forced labourers from the native population, had stone foundations and surfaces of rammed gravel, cambered to drain into side ditches. They may or may not have followed earlier routes, and multitudes of lesser and rougher tracks will have coexisted with them. Almost the whole course of the Fosse Way is known—that is, physical traces of it have been found—except at our Axe valley end. From near Ilminster in Somerset, through Axminster to the probable end at Seaton, the solid red line on the O.S. map becomes a dashed line, meaning that the route marked is only "possible".

Yet a lot of it is probable, because still preserved as roads or other features. North out of Axminster, Fosse Way lives on as the Chard Road—but has lost the straightness. For one thing, it kinks at the bridges by which it jumps from side to side of river and railway. As you drive toward Chard, you pass (just at the Devon-Somerset border) a turning to a secondary road that heads to Crewkerne—but look again at the map: it's been shaped to look like a turning, but the B road is the one that continues the straight Roman line up the Axe valley.

In the other direction, as you come into Axminster, there is a roundabout where Stony Lane turns off—but again, it is the present main road that bends; Stony Lane is the one that continues the straight line. I'm not sure, and I don't know whether archaeologists are sure, of where the Fosse then runs under the built-up areas of the town, since it is apparently picked up next by South Street, which doesn't lie in the same line as Stony Lane. South of the town and after Abbey Gate, the present road's curves are for a while not large enough to take it far from what will have been the straight Roman line, but later they are, and the Fosse Way presumably crossed the Axe at some point (if it kept in a straight line, this would have been near Colyford), to reach Seaton.

The other major Roman road, the east-west one, came from Durnovaria (Dorchester), tribal town of the Durotriges, and went on to Isca (Exeter) of the Dumnonii. No physical traces have been found between Eggardon Hill (east of Bridport) and Axminster, so the exact route through our area is not known; yet it seems likely to correspond quite closely to the modern A 35 highway. It's the only almost easy way to go long-distance from east to west, dodging the steepest of cross-valleys and making use of the only roughly east-west ridgetop (between the Char and Lim valleys). It even seems likely that the newest parts of the highway—the cutting along the north face of Lyme Hill, and the Axminster bypass—bring the route back to the straighter Roman one. They revive the Roman road but destroy it—I imagine that the cutting has gouged out, and the bypass buried, any traces there might be. The Romans, I think, would not have come along the level top of Lyme Hill only to have to descend the Corkscrew; later people must have shifted the route a little southward for easier access to Lyme.

Some call this road part of the Icknield (or "Ickneld") Way. Well, the Icknield Way is another ancient route that followed a sinuous northwest-southeast scarp: the chalky Cretaceous scarp, from Norfolk through the Chilterns to the Berkshire downs and the uplands of the Salisbury Plain.

Icknield, however, is a pre-Roman Way. Much of it still exists as a braided set of tracks—a ridgeway along the top of the escarpment, a main track above the spring line, and a summer track below. But none of this route was followed by a Roman road. From Sorviodunum (Old Sarum, by Salisbury) a Roman road pushed on in the same general direction to Vindocladia (Badbury Rings, a now townless site in eastern Dorset); but this is known as Ackling Dyke, and was really part of one of the Roman routes radiating from London. From Badbury it could be said to be continued to the sea by a road that turned south to Poole; or to be continued with a more westerly deflection by the road that led to Dorchester, which only in this stretched sense could be said to be part of the Icknield Way.

As for the name Icknield, its earliest known application was to the Berkshire section; yet it is now mostly applied to the Chiltern section; and it is presumed to derive from the Iceni, the British tribe that occupied what is now Norfolk. Some fancy that there was a lady called Icenhild. But hild ("battle") is an element in Germanic names (Hilda, Hildegard, Mechthild or Matilda, Kriemhild and Brunhild the rival queens in the epic Nibelungenlied; this is curiously like the Navajo habit of using words for "war" in female names). There was certainly Boudicca or Boadicea, who in 60 A.D., having been atrociously provoked by the Romans, led the Iceni in a spectacular rebellion, which almost succeeded and was ruthlessly crushed. So just possibly a later race remembered this Way as having led to the realm of the Icenian battle queen.

I think, though, there is something deeper in the name. I was born at a village north of Birmingham called Streetly, showing that it had been on a Roman road (street is Anglo-Saxon straet, from Latin via strata, "paved way", from the root of sternere "to strew" or "stretch" or "spread"). And the sandy path that went past us through a piece of wild land was pointed out as the line of the Roman road—and people said that it was Icknield Street. But on the map you can see that, running toward Worcester in one direction and York in the other, it is yet another line roughly parallel to the Fosse Way and the real Icknield. They had mis-heard: its name is Ryknild Street. And that seems to suggest that something other than "Iceni" underlies not only "Icknield" but "Ryknild" and "Ackling".

From our western part of the "Icknield Way", a branch road may (or may not) have run down to Charmouth and then along the coast toward Lyme. Along the inland end of Lyme town lies "Roman Road". If the tradition that it is such is true, then it dipped straight down into the valley and had to climb as steeply out the other side, and presumably would have continued to climb up Clappentail Lane, and then followed a line like the present coast road. Possibly. But no settlement is known at Lyme earlier than 774; and the Ordnance Survey maps show no Roman road here.