A HISTORY OF DERBYSHIRE of more moment. It would represent their first definite enrolment in the Empire, as men liable to taxation and to conscription. 1 It would imply the systematic organization of this district in the Roman system. It remains to determine the Romano-British name of Brough. Fortunately the task is easy. A milestone found at Buxton in 1862 states that that place was 10 (or perhaps 11) miles ANAVIONE (p. 226). This might be interpreted either a Navione, 'from Navio,' or Anavione ' (from) Anavio.' But the Foligno incription just cited shows that the latter is preferable. If we proceed to inquire where Anavio was, we shall find only one Roman site distant i o or 1 1 miles from Buxton and connected with it by a road. Brough, therefore, is Anavio, and the Anavionensian Britons are the hillmen of High Peak. With this result we may connect two items in the Anonymus Ravennas. He mentions a river Anava next to a river Dorvantium, and a place Navione * or Nanione next to Aquis. If we take Dorvantium to be the Derwent, Aquae to be Buxton, and Navione to be a mis-spelling of Anavione, 8 everything falls into its place. We may even go on to think that the river at Anavio was called Anava, and that the name still survives in Noe. 4. MELANDRA Melandra is a Roman fort in the parish of Glossop and the town- ship of Gamesley, near to the Cheshire border and the Dinting station of the Great Central Railway. Its position is no less significant than that of Brough. Near Glossop, the gorge of Longdendale, meeting lesser valleys, begins to widen towards the Cheshire and Lancashire lowlands. Here is the easiest entrance to the north Derbyshire hills. Here, too, is the nearest point to those hills that is readily accessible from the western plains. The ancient soldier, wishing to plant a fort within striking distance of High Peak and yet within safe reach of western communica- tions, would find his fittest site near Glossop. The actual position selected for the fort agrees well with Roman custom (fig. 1 6) . Not far from Glossop a low spur of hill runs out from the south side of the valley into Longdendale, and on the end of this spur stands the fort. The ground falls away from it on all sides except for a narrow neck to the southward, which quickly changes to a rise. But on the other three sides a slope of varying steepness descends a hundred feet to the river Etherow and the Glossop Brook, which meet a little way north of the fort, and to a little ravine which descends to the Glossop Brook. It is a well-protected site, and it commands a wide outlook over Longdendale and the great hills around it. And at the same time 1 Corf. Insc. Lat. xi. 52i3 = Dessau 1338. Watkin, who first saw that this inscription referred to our district, misinterpreted it both by taking censitor as the officer of a cohort, and by eliciting a cohort with the impossible title of Brittonet a Navione (Archteologjcal Journal, xli. 255 : Derb. Arch.Journ. vii. 83). 8 Parthey and Finder gave Nanione as the reading of the best MSS. But Prof. J. S. Phillimore, who examined the chief MS. for me, the Vatican Urbinas 961, tells me that it reads Nauione. In either case the divergence from Anavione is slight. 8 Omissions of initial letters are not the least common among the many errors of the Ravenna lists. 2IO
