We can imagine the men would get impatient {37} at having to climb a mound, descend into a ditch, and then climb another mound at each journey north or south. And if rough carts were used, a causeway would be a necessity. By degrees continuous traffic would of itself sink a path across the mounds; and what would be easier than to make deliberate gaps and fill up the ditch with the soil? The Romans were a methodical people, and in undertaking an enormous work like the Wall they would certainly have points, at regular distances apart, to which building materials had to be brought. The only difficulty in the theory seems to be that a gap every 45 yards could hardly be necessary. Supposing this theory to be correct, then the Vallum _is older_ than the Wall and its contemporary buildings, but very little older. It has already been proved that the Vallum is not older than the original forts, because it always curves round to the south when it approaches a fort, in order to avoid it. The Stone Wall appears to have been an after-thought, found necessary for the final solution of the defensive problem. It has long been a question whether there was not originally a turf-wall right across the island, thrown up hastily to mark out the course of the Stone Wall, and as a temporary defence, and then gradually replaced by the Stone Wall. The presence of a piece of turf-wall, more than a mile long, between Birdoswald and Wallbowers, running north of the {38} Vallum and south of the Stone Wall, has to be accounted for; and for two reasons it seems probable that this was only a temporary structure: first, because there is no carefully-laid stone foundation under the turves, as is the case with the Antonine Wall; secondly, because no military way accompanies it, and the Romans always had a road accompanying their fortifications. Mr. Simpson thinks this particular stretch represents a mistake in the laying out of the works, because it is so close to the Vallum that a mile-castle could not have been built at the usual position: which mistake was corrected by the Stone Wall, but the fruits of which it was not thought necessary to remove. Or--another theory--this turf-wall may have been a temporary local barrier, thrown up during some interruption of the building of the Stone Wall. It is certain that the turf-wall is the earlier, for where it ends, at Wallbowers, its ditch goes under the Stone Wall, and had to be filled up when the Stone Wall was built. {39} CHAPTER V THE WALK: WALLSEND TO WALBOTTLE I decided to begin my walk at the Newcastle end because I thought it would be the least interesting part, and I wanted to get it over. However, it proved far more interesting than I expected. I reached Newcastle from King's Cross at five o'clock on a May morning, and, booking my luggage, I started off at once, knapsack on back, for Wallsend.