lived, and no other satisfactory explanation is possible than that in these lanes grew the sallow or willow tree. Many of these roads are, moreover, on the Yorkshire side of Manchester, and throughout the rural districts of England where this name occurs (Salford or Saltersgate) salt mines are non-existent.

In the History of Newton Chapelry, Mr. Crofton mentions numerous Saltersgates in the lanes between Manchester and Stockport, a district with which I happen to be intimately acquainted, and where thirty years ago the willow flourished.

ROBIN HOOD'S WELL.-The words "Robin Hood's Well" occur on the map on Whitworth Moor, two miles north-west from Whitworth, at a height of about one thousand one hundred feet above the sea-level.

The words "Robin Hood's Bed" occur on the map on Blackstone Edge Moor, one thousand five hundred and fifty-three feet above the sea-level, close to the Yorkshire border.

BURY.

The accompanying illustrations show at a glance the strength of the position of the old town of Bury; the castle, the market place, and the church all close together on a hilly plateau with a steep descent to the river, which formerly followed a more easterly and sinuous course than now.

* Canon Raines's chief instance is the road from Rochdale to Burnley. The first portion of this road, however, runs through a deep valley, about nine miles long, through Whitworth and Facit to Bacup, with a stream, the Spodden, at the bottom, the very place to favour the growth of sallows or willows. As already mentioned, Salterford is the ford at the sallow tree, from sallow, A.S. sealh, sallow or willow, the common name of various species of the genus salix or willow kind. Saltersgate, consequently, is the road lined with sallows.

W

Castle Steads

Bury is not mentioned as a Roman station either in Watkin's Roman Lancashire or in Harrison's archæological survey of the county, although the great Roman road from Manchester to Ribchester passed within one and a half miles of it on the westerly side. A plan of the castle is, however, to be found in vol. xlvii. of the Royal. Society's Transactions (published in 1753), and in an accompanying article the author describes it as a Roman camp.* This old plan of the castle is valuable, for it shows the former bed of the Irwell, as described above, immediately to the front or north of the castle and church with the words "a steep descent to the river." The castle is shown with a frontage to the north of two hundred yards. The plan is called "Coccium, now Bury Castle Croft." This steep descent is still a feature of the town, although all visible traces of the castle have disappeared in the lapse of time.

To the north of the site the ground was (and still is) marshy, and was thus useful, as in so many other similar instances, for defensive purposes. It is, however, now being gradually filled up for the prevention of floods, and in a few years this important feature of the site will be obliterated.

The former course of the river is not completely shown on any plan which I have seen, but an inspection of the site, looking north from the edge of the castle plateau, leads irresistibly to the conclusion that at a point threequarters of a mile from the castle the river at some unknown date swept round at the base of the cliff gradually from the plain in a nearly semi-circular line, rising almost imperceptibly until it reached to the full

* Philosophical Transactions, vol. xlvii., for the year 1751 and 1752. London, 1753.

height of the castle plateau at a point nearly opposite to the parish church. The river at its widest point thus flowed about one-third of a mile more to the east than

it does at present. This marshy ground, measuring about three-quarters of a mile from north to south, and one-third of a mile from east to west, rises slightly from west to east, and I surmise that at one time, perhaps in the time of flood, some impediment caused the river to take the present easier line to the west. The castle would thus be bereft of one of its principal defences.

The Roman camp was probably six hundred feet from north to south, as well as from east to west, and some of its foundations might, therefore, be found by digging on the southerly side of Bolton Street.

From discoveries made during excavations in the year 1865, and recorded in Baines's Lancashire, and again in the summer of 1905, it seems clear that a mediæval castle succeeded the Roman camp on this site. The building whose foundations wer were thus discovered was well buttressed, and appears to have been a keep or peel tower, measuring externally about one hundred and thirteen feet by one hundred and ten feet. It stood near the easterly side of the Roman camp.

Some of the stones from this tower have been used in recent years in building the volunteer barracks, and it may be, indeed, that some of them have been of Roman workmanship. With the hope of arriving at some definite conclusion as to their age, Mr. Frederick Howarth, of Bury, has very kindly made for me a carefully figured drawing of one of these walls with the masons' marks. Unfortunately, a good many of these stones have been discharged into a tip. On this site coins from the reigns of the Edwards onwards have been discovered from time to time.