G. 3.--The Atrebates, who play a not unimportant part (as will be seen in the next chapter) in Caesar's connection with Britain, were apparently in possession of the whole southern bank of the Thames, from its source right down to London--the river then, as in Anglo-Saxon times, being a tribal boundary throughout its entire length. This would make the Bibroci a sub-tribe of the Atrebatian Name, and also the Segontiaci, if Henry of Huntingdon (writing in the 12th century with access to various sources of information now lost) is right in identifying Silchester, the Roman _Calleva_, with their local stronghold Caer Segent. G. 4.--But the whole attempt to locate accurately any but the chiefest tribes found by the Romans in Britain is too conjectural to be worth the infinite labour that has been expended upon the subject by antiquaries. All we can say with certainty is that forest and fen must have cut up the land into a limited number of fairly recognizable districts, each so far naturally separated from the rest as to have been probably a separate or quasi-separate political entity also. Thus, not only was the Thames a line of demarcation, only passable at a few points, from its estuary nearly to the Severn Sea, but the southern regions cut off by it were parted by Nature into five main districts. Sussex was hemmed in by the great forest of Anderida, and that of Selwood continued the line from Southampton to Bristol. Kent was isolated by the Romney marshes and the wild country about Tunbridge, while the western peninsula was a peninsula indeed when the sea ran up to beyond Glastonbury. In this region, then, the later Wessex, we find five main tribes; the men of Kent, the Regni south of the Weald, the Atrebates along the Thames, the Belgae on the Wiltshire Avon, and the Damnonii of Devon and Cornwall, with (perhaps) a sub-tribe of their Name, the Durotriges, in Dorsetshire.