founded.

Such passages over the Thames were of two sorts: there are first the original fords, numerous and primeval; next the crossing places of the great roads.

Of the original fords we have already drawn up a list. Few have, merely as fords, proved to be of strategic or commercial value. Oxford may have been an early exception; and the difficult passage at Abingdon founded a great monastery but no military post: the rise of each was connected, as was Reading (which had no ford), with the junction of a tributary. Wallingford alone, in its character of the last easy and practicable ford down the river, had for centuries an importance certainly due to geographical causes alone. Two principal events of English history—the crossing of the Thames by the Conqueror and the successful challenge of Henry II. to Stephen—depend upon the site of this crossing. Long before their time it had been of capital importance to the Saxon kings, so early as Offa and so late as Alfred. If the bridges built at Abingdon in the fifteenth century had not gradually deflected the western road, Wallingford might still count the fourteen churches and the large population which it possessed for so many centuries.

Apart from Wallingford, however, the fords, as fords, did little to build up towns or to determine the topography of English history. Of more importance were the crossings of the great roads.

When one remembers that the south of England was originally by far the wealthiest part of the country, and when one considers the shape of Ireland, it is evident that certain main tracks would lead from north to south, and that most or all of these would be compelled to cross the Thames Valley. We find four such primeval ways.

One from the Straits of Dover in the south-east to the north-western centres of the Welsh Marches and of Chester, the Port for Ireland, and so up west of the Pennines. This came in Saxon times to be called the Watling Street, a name common to other lesser lanes.

Another, the converse to this, proceeded from the metal mines of the south-west to the north-east until it struck and merged into other roads running north and east of the Pennines. This came to be called (as did other lesser roads) the Fosse Way.

A third went more sharply west from the southern districts, and connected them not with the Dee, but with the lower Severn. This track ran from the open highlands of Hampshire through Newbury and the Berkshire Hills to Gloucester, and was called (like other lesser tracks) the Ermine Street.

Finally, a fourth went in a great bend from these same highlands up eastward to the coast of the North Sea in East Anglia. This was called in Saxon times the Icknield Way.

All these can be traced in their general direction throughout and for most of their length minutely. All were forced to cross the Thames Valley, which so nearly divided the whole of South England from east to west.

Of these four crossings the first in point of interest is that which the Ermine Street makes over the upper Thames at Cricklade.

These old roads are of capital importance in the story of England, and though historians have always recognised this there are a number of features about them which have not been sufficiently noted—as, for instance, that armies until perhaps the twelfth century perpetually used them; for the great English roads, though their general track was laid out in pre-historic times, were generally hardened, straightened, and embanked by the Romans in a manner which permitted them to survive right on into the early Middle Ages; and of these four all were so hardened and strengthened, except the Icknield Way. Not one of them is quite complete to-day, but the Ermine Street is perhaps the best preserved. It is a good modern road all the way from Bayden to Gloucester, with the exception of a very slight gap at this village of Cricklade.

It originally crossed the river half-a-mile below Cricklade Bridge, so that the priory which stood on the left bank lay just to the south of the old road. How and when the old bridge at Cricklade fell we have no record, but one of the most important records of the Thames in Anglo-Saxon history is connected with this passage of the river.

The importance of Cricklade as a station upon the upper Thames does not only proceed from its being the crossing place of a great road, it is also the point when the first important tributary stream, the Churn, joins the Thames. Above this junction the Thames nowadays is hardly a stream; and even in the eighteenth century and earlier, before the digging of the Severn and Thames Canal, it must have depended on the weather whether there were any appreciable amount of water in the upper part or not. It would probably be found, if records could be examined, that the mills at Somerford Keynes were not continually worked throughout the year, even when the supply of water had been left undiminished by modern engineering. But when once the Churn (which, as we have seen, has a larger volume of water than the Thames) had fallen in at Cricklade the two formed a true river, with depth in it always sufficient to support a boat, and with a fairly strong stream, as also with a width sufficient for minor traffic; and it is after Cricklade that you get a succession of villages and churches dependent upon the river and standing close to its banks.

But though this piece of hydrography has its importance the chief meaning of Cricklade in history lay in the fact that it was the spot where this Ermine Street on its way from the south country to the Severn Valley got over the Thames, and the village connected with it was entrenched certainly in Roman and probably in pre-Roman times. This entrenchment may still be traced.

The crossing of the Thames by the Icknield Way, unlike the crossing of the Ermine Street at Cricklade, presents a problem.

Cricklade, as we have seen, is a perfectly well-established site, and we owe our certitude upon the matter to the fact that the Romans had hardened and straightened what was probably an old British track. But with the crossing of the Icknield Way no such complete certitude exists, for the Icknield Way was but a vague barbarian track, often tortuous in outline, confused by branching ways, and presenting all the features of a savage trail. Doubtless that trail was used during the four hundred years of the high Roman civilisation as a country road, just as the similar trail, known as the "Pilgrims' Way" from Winchester to Canterbury, was used in the same epoch. There are plenty of Roman remains to be found along the track, and there is no doubt that all such roads, even when the State was not at the expense of hardening or straightening them, were in continual use before, as they were in continual use after, the presence of Roman government in this island; but the Icknield Way does not approach the river in a clear and unmistakable manner as would a Roman or a Romanised road. It is on this account that the exact point of its crossing has been debated.

The problem is roughly this: the high and treeless chalk downs have been used from the beginning of human habitation in these islands as the principal highways, and any single traveller or tribe that desired in early times to get from the Hampshire highlands to the east and north of England must have begun by following the ridge of the Berkshire Hills, and by continuing along the dry upland of the Chiltern Hills, which continue this reach beyond the Thames. But the spot at which the pre-historic crossing of the Thames was effected cannot be determined by a simple survey of the place where the Thames cuts through the chalk range. Wallingford up above this gorge has certain claims, both because it was the lowest of the continually