Roman Manchester

by Charles Roeder

ROMAN MANCHESTER.

BY

CHARLES ROEDER.

MANCHESTER:
RICHARD GILL, TIB LANE, CROSS STREET.
1900.

ILLUSTRATIONS.

---

RECENT ROMAN DISCOVERIES IN DEANSGATE AND ON HUNT'S BANK, AND ROMAN MANCHESTER RE-STUDIED (1897-1900).

BY C. ROEDER.

THE extensive excavations in lower Deansgate during the last four years have naturally been of particular interest to local antiquaries, and therefore invited research. I undertook, at the request of the Society, to watch operations all over the area for the purpose of extending our knowledge of Roman remains, and unstintedly gave all my spare time, so as not to miss the slightest chance of any discovery that might add to our previous information. It proved a very arduous task, but the ground having been familiar to me in all its geological and topographical bearings for many years, no feature of its physiography had difficulties left to unriddle for a thorough comprehension of the localities where results would most likely spring up, and no time was wasted in fruitless search. I had, of course, to condense considerably my materials, which otherwise would have extended my paper to unwelcome size; and have consequently refrained from giving more than a general summary of the leading results. "Roman Manchester" has fared very badly compared with Roman Chester, or even little places like Wilderspool, Ribchester, and Melandra. The earlier antiquaries were satisfied to look at the "square tower," or to view one or two centurial stones left in the wall. Dr. Dee, who might have done much, devoted himself to other studies rather than give us a description of the ruins at Castlefield or leave a plan of how it appeared in his time. Whitaker, with all his eagerness and enthusiasm, practically wrote for men of the town in his own time, who knew all its features well; if he had confined himself to locating and drawing all the sites of discoveries and speculating less, he would have done a great service; as it is, his two large volumes have carefully to be studied through, and the few golden grains have to be extracted from the heap of dross that covers them. But we owe to him the registration of some valuable discoveries which but for him would leave us incapable of understanding some important points which have since extended our knowledge. Thompson Watkin has collected all the literary evidence up to his time, and incorporated his labours in his great work, Roman Lancashire, but not being a resident and thus missing most of the important excavations effected during his time, it has scarcely, if at all, increased our information; besides an author, who had to treat the whole County Palatine, could not be expected to concentrate his attention on Manchester, which held out so little encouragement to him for work or practical research.

Unfortunately, interest in Roman matters lay in a deep slumber, so far as local efforts were in question, and all the great opportunities from 1849 to even recent times were allowed to pass without profiting by the exceptional and transitory chances that lay open to practical investigation. Although records of discoveries and finds are wanting, I know from navvies who have worked on the spot for many years past that the soil was exceptionally rich in Roman "spoils," which my own experience during the last thirteen years confirms. I have lost no opportunity, so far as has been in my power, to rescue and record whatever could be found, and with a view to preserve from loss or indifference any evidence left of Roman Manchester I have spared no time or labour to rescue them for the town. I may say that I have collected numberless objects and fragments, too large and varied for classification; part I have already given to the Grammar School, the Geographical Society, Peel Park, and Oldham, but the main bulk of my collection is awaiting still a proper home,^[1] for which, above all, the Town Hall should be the most suitable centre.

Finally, I have to acknowledge the great help received from many quarters during these arduous labours; without the intelligent help of contractors, foremen, and navvies I should have made little progress or result; their assistance and interest, at all times, has been unsparing and valuable. I have also to thank Professor F. Haverfield for much help and information; to Mr. Thomas Rogers, Mr. Charles Bailey, and others for the naming of botanical and other specimens; and particularly to our esteemed member Mr. J. J. Phelps, who later on joined me in work, and who not only has prepared all the splendid drawings and sketches, but has given a great amount of labour and devotion to make this paper more attractive and worthy of acceptance. To present a complete clear account of Roman Manchester I have put all the previous strands together and incorporated Whitaker's and Watkin's data. Whitaker's account had carefully to be restudied, and all his sites to be identified, to make them intelligible and available for a reconstruction of a plan. I shall now immediately proceed to the description of the station.

Walls.

The northern and southern walls^[2] measured respectively 175 yards, and the western and eastern walls 140 yards, covering thus an area of 5 acres and 10 perches or 24,500 square yards.

The eastern wall, in 1765, stood still 6 feet to 7 feet in height, and was lowered to 3 yards to 4 feet 80 yards away from its northern termination, where the Porta prætoria was placed. The latter was still visible in 1810. It was here that, according to Dr. Holme^[3] the centurial walling-stone (Coh. I. Fris. Quintiani Pedes XXIIIII.), measuring 15 inches by 11 inches, of millstone grit, was found. It is of interest to note that the centurial stone of the Frisian Cohort at Melandra^[4] was discovered very near the east angle. Probably it was, therefore, inserted in Manchester at or over the gateway. Dr. Holme mentioned that along with it, at the same point, coins of Trajan (98-117) and Hadrian (117-138) were found. Unfortunately, he does not state the exact

situation in which they occurred. He suggests they might have reference to a time when the wall may have undergone a repair. We shall see later on, when speaking of the hypocaust, that such a hypothesis is not improbable. The wall at the south-east angle in Whitaker's time still stood 10 feet high.

The southern wall. The gateway was exhibited 65 yards away from its south-west angle and afforded a ready passage from the station to the Medlock. The walls were still 7 feet high.

The western wall was still rising 7 feet from the ground and 40 yards from its north-west angle the Porta Decumana was placed.

The northern wall had its gate 48 yards from the east side, it still stood 4 feet to 6 feet high. Whitaker took occasion to examine its internal structure. It consisted of rough undressed stones (new red sandstone), angular pieces of rock, smooth round boulders, all bedded in strong white-brown mortar. He found the walls raised from the breadth of 7 feet to 8 feet at the base. Sir Henry Dryden, who also measured a remnant of the wall in 1843-4 on the east side, states its thickness to be 7 feet, the old base was almost entirely gone, the old joints at the base were about ½ inch wide, the joint inside 2 inches or 3 inches wide, the stones 12 inches and 15 inches long, and 5 inches to 8 inches thick. The probable height of the wall was 12 feet.

We see then from the position of the gates that they were not uniformly placed. The northern gate stood at a distance of 48 yards from the east angle, the southern 65 yards from the west angle, the eastern 80 yards from the north angle, and the western gate 40 yards from the north angle, contrary to the usual rule, and the principal streets, as Whitaker remarks, must therefore have crossed the area obliquely from east to west, and likewise obliquely from north to south. The cause of this deviation is probably to be sought in the physical exigencies of the ground with which the builders had to deal.

Foundations.—I shall first speak of the substructure of the walls. Whitaker fortunately examined three sides. He found the foundation of the western side laid on two beds of blue, well-worked clay, the lower being nearly a foot in depth, and remarkably stiff and solid (pages 32 and 33, edition 1770). The southern side was laid in two courses, not a foot in depth, not of actual clay but of claymortar, clay and sand incorporated together, and both lying upon a deep bed of river sand. The eastern side rose from two courses of boulder stones cemented with clay. The northern side has been trenched at various points during the extensive railway operations from May, 1897, to September, 1898; at the east and west side of Collier Street. These I have carefully watched. The trenches were carried down into the underlying new red sandstone rock, and were made 40 feet long and 28 feet apart from each other. At trench No. 3^[5] they penetrated a little into the castral area, cutting the Roman wall obliquely.

I give a cross section.^[6] Starting from the surface we have on its eastern side:—

36 inches modern rubbish;

5.6 ⁠„⁠rough, irregular blocks of sandstone;

2¼⁠„⁠a course of brownish-black Roman mortar, mixed with mould;

5½⁠„⁠oblong, dressed blocks of red sandstone;

1¼⁠„⁠a course of mortar;

6⁠„⁠yellow-brown, stiff boulder clay;

12 inches rough, undressed sandstone;

19 ⁠"⁠ consisting of six consecutive layers of well-selected, oval-shaped, small boulder stones, carefully laid in rows, each 2 inches to 3 inches high, all of 5 inches long, and imbedded between regular bands of stiff brown clay, below the natural deposit of

70 ⁠"⁠ valley gravel, which again rest immediately on the scooped surface of the red rock.^[7]

On the western side of the same trench (about 10 feet wide), the layer of 6 inches clay is absent.

In trench No. 2, starting from the bottom, we find:—

10-12 inches boulder stones, of medium to large size, bedded on clay;

10 ⁠"⁠ stiff brown clay;

7 ⁠"⁠ large, dressed red sandstone;

8 ⁠"⁠ mortar.

Rest obliterated by some fireplace, built into the wall at this point, which cuts off the remaining 3 feet.

At duplicate-trench No. I, we have a cross and longitudinal section of the wall 10 feet long, which shows:—

23 inches rubbish;

29 ⁠"⁠ greenish soil;

4-5 ⁠"⁠ boulder-stones, bedded in stiff boulder clay. underlaid by

45 ⁠"⁠ valley gravel.

And about a yard to the east, instead of boulder stones, we meet with rubble or red sandstone, resting on the clay, and below a light yellow soil, which I may call the pre-Roman or original surface, which spreads over the valley gravel.

I had another and still better opportunity of examining the nature of the foundation wall in March, 1898, at the western side of Collier Street, where, for the erection of a small coalyard, the northern wall was longitudinally laid open on the inner (intra-castral) side for a length of 38 feet, and 4 feet deep (see section). On its more eastern side, we have a layer of 22 inches Roman mortar, and 20 inches big, roundish sandstone, bound together by brown cement (specimen kept), which seems to have been poured between and over the various layers of stones. On the western side the order changes, and we have four layers of soft, decomposed rubble, separated by layers of middle-sized boulder stones; below, 6 inches to 7 inches, stiff, greenish, sometimes drab coloured boulder clay with boulder stones imbedded, the foundation resting here on greenish marshy clay. At the base, little flat pieces, or chips, of sandstone flags have been laid down. In the central part, at the bottom, for the space of 72 inches, are placed regular oblong and dressed blocks of red sandstone and millstone grit, of various length, varying from 19 inches by 10 inches to 24 inches by 12 inches. In all probability it formed the northern gate. Under the wall, at a spot marked x on the section, I discovered a great number of promiscuous land shells and bones of the newt, which must have been carried hither by some rain floods or the current of an ancient rill. These objects will be described in the appendix. In the various sections along the northern wall no pounded or broken tiles were discoverable in the mortar.

At trench No. 3 the thickness of the foundation walls varied from 5 feet to 6 feet, east of Collier Street it measured 7 feet.

It is evident from the above description that there exists

a constant and considerable variation in the substructure of the wall, and it appears that the builders were mainly guided by the nature of the ground they had to deal with, adapting their material to the varying ground and levels. At one point they might have strong gravel, at another loose sand or again marshy, yielding ground and swamp, or the outcropping red rock. For the same reason the rough outline wall varied and oscillated between 5 feet to 7 feet in thickness.

Defences.—The fort on the eastern side was bounded by a fosse, of which Whitaker has left a description.^[8] On the south-east angle, upon the lower margin of the Medlock, a deep and narrow gully, cut through the solid rock, existed in 1765. From this point, he says, the ditch seems pretty plainly to have mounted directly up the little garden that now lies on the eastern side of Castlefield,^[9] the rocks of the right side of the garden appearing evidently cut away and sloping towards the west.

On the south it was naturally protected by the curve of the Medlock. The southern part of the Roman stronghold stood on the rocky ledge of the ridge or plateau, which commenced at Knot Mill, and extended along the northern banks of the Medlock. It may still be viewed in Castlefield Street, at the coal wharf, and along the towing path of the Rochdale Canal. The southern side rose on the crest of the rock and stood on higher ground than the northern part of the castrum, towards which the rock gently dips away. Green's Map of Manchester (1789) gives a good representation of the appearance and direction of the banks and slopes which embrace the west, south, and east side of the fort. "For the greater safety the northern bank of the river was carefully scarped, and the long stroke of the pickaxes appeared in 1764 for the whole compass of the bank, upon the face of the rocks which are below the present edge of the water, and descended nearly to the original surface of it within 1½ yards from the stony bed of the river. Deep in the artificial soil, with which the face of the bank was covered, were found in 1765–6 a fibula, an urn, a coin [ . . . Reduci ... an Aug. Cos], and an unguent bottle of black glass in a little hollow upon the rock. Along one particular part of the margin, from the eastern boundary of the field beyond the mouth of the subterranean Bridgewater tunnel, the rock was cut down either into a very sharp descent or into an absolute perpendicular, and it extended along the whole semicircular margin of the river. About 20 yards to the east appeared in 1766 a large flight of rude stairs leading to the river, viz., seven steps about three yards long, from ¾ yard to 1 foot in breadth and from 10 inches to 4 inches in depth, and very visibly worn away in the centre. The lower part of them had been actually cut down into an absolute perpendicular."^[10]

On the west was the natural barrier of a lofty bank (see Green's Map), forming a sharp slope of 50 yards to the swampy ground below it. It extended in one continuous line up to the present Water Street, and envelops the whole southern wall from its western to its eastern angle. Beneath this bank and rampart spread out a morass, or marshy ground, about 100 yards in breadth and 300 yards in length, beginning at the margin of the Medlock^[11] on the south, extending betwixt the bank and the channel of a rill to the north-west of the Roman suburb. Whitaker adds (page 27): "This lately continued so bad a morass that even in the dry summer of 1765 horses sank up to the belly in it."

I may now proceed to the rather intricate question of the set of ditches which were placed along the line of the northern wall.

In order to fully understand the character of their structure I have prepared a plan which shows at a glance the work done during the railway excavations in 1897–8 in that direction.^[12] Ten double trenches were made along the wall, and in addition to this the whole area now covered by the new police station, in Bridgewater Street, has been more or less excavated for foundations. I have followed operations almost daily while the work has been in progress; personally examining, measuring, and collecting all the time, and noting the physical features of the successive strata penetrated until the rock was reached. It proved a tough piece of work, and the labour has been great to unriddle satisfactorily the real nature of the northern defences. It must be understood that there was not a single trench that exhibited in unbroken succession the original arrangement. Since the beginning of last century the ground has been interfered with incessantly for the erection of streets and yards, part has been excavated for gravel or for tipping rubbish, and the subsoil has been redisturbed for sinking drains and culverts, and it is only by piecing up shreds of untouched patches here and there that I have been enabled, in a manner, to reconstruct a running section from the point of the northern wall across Bridgewater Street to Worsley Street. As the trenches were made at an angle of 45 degrees to the course of the Roman wall, the section had likewise to be prepared again at right angles. It will thus be seen that the diagram produced presents merely an approximate view of the probable course and appearance of the fosses and ramparts situate on the northern side.

The Romans were obliged to carry their wall on the northern part through a stratum of greenish or blueish and sometimes dark-coloured clay, which reappears inside the castrum. The rocky neck or ridge on which Mancunium stands does not cover the whole internal area; in fact, it falls away gradually at a point about 200 feet away from the northern wall, apparently in a slope, the sides of the ridge being covered by a deposit of ancient valley gravel. This gravel rests practically on the under-lying and undulating red rock, which rises and falls at certain points.

The greenish-blue clay, which occurs at trench No. 3, immediately rests in the hollow of this valley or river gravel, and we can follow it right along our section for a distance of 194 feet. It clearly appears to a geological eye to mark the existence of an old transverse valley trending towards the Irwell, which at one time collected the surface-water of the higher grounds to the east. In all probability this dingle or valley stretched from Camp Street on one side, where we are on a constant rise, right across to the ridge on which the Roman station stands. At Campfield and Alport Town we stand again on the outcropping edge of the red rock, as proved by the recent railway excavations, and the rock after plunging away reappears on the southern side of castrum. This miniature hill-and-valley system, which is so characteristic a feature of the new red sandstone in Deansgate and the cause of the abundance of so many wells and springs in that particular locality, is in full evidence from Ivy Street to Great Bridgewater Street, that is, along the northern side, and throws the underground and surface water into the depression or valley formed between the castrum and Camp Street.

Now at the time the Romans perched their castrum on the northern banks of the Medlock Bridgewater Street formed part of an open valley mingling with stretches of marshy swamp, moorland, patches of heather, sprinkled with some clear pools here and there, and traversed by little rills speeding down from the higher adjacent levels. They had to dig their northern wall along the edges or margins of this marshy ground, and we, therefore, find the foundation walls on that side by necessity sunk down into the swamp. We have seen already that under the foundation of the northern wall on the eastern side of Collier Street a large accumulation of heterogeneous land shells were discovered by me in the green marshy clay beneath, which clearly affords another proof of the existence of an old marshy ground.

At the northern or outer side of the wall this soft marshy bed is cut down to a slope of 45 degrees. At trench No. 2, 2 feet to 3 feet away from the wall, and at the very bottom of the first fosse, a solid causeway was struck in May, 1897, running apparently parallel to the wall. It was formed of large dressed blocks^[13] of compact, heavy, white millstone grit, measuring 11 inches by 7 inches by 6 inches each. The upper face of the stones was convex, and they had the hard Roman mortar still adhering to its two sides, the bottom had no mortar, and was rough. The width of this causeway was 2 feet 7 inches. It was laid on the marshy ground, and the navvies had great difficulty in breaking it up, in consquence of its solidity and the stones being so tightly cemented together. Unfortunately, I was not present at the moment when taken up in situ to see whether the stones bore any rut-traces, and how the two sides terminated. Later on in 1898 this causeway was found again in the same position in the duplicate—trench No. 1 on the east side of Collier Street, close to the wall.

Being at the bottom, and so close to the wall, the natural conclusion seems to be that it was made for securing a firm grip, and giving greater stability, and for procuring a foothold to carry and facilitate the transport of the building-stones and material when building at the foundations. In trenches Nos. 2 and 3 we see the first hollow or fosse, 16 feet wide and 4 feet to 5 feet deep, carried into the river gravel, then follows a gravel-ridge, feet wide, which formed apparently a small rampart. Then comes a little hollow, 5 feet wide and 1 1/2 feet deep. From its banks, sliding down in an angle, we observe a promiscuous stratum of Roman mortar and broken pieces of rubble stone, and having the appearance of surplus building rubbish. It looks as if the material had been discharged from the top after the work was finished. The natural ground is disturbed at this part. Then follows another hollow, 16 feet wide; a gravel ridge, 6 feet wide; a hollow, 22 feet wide; ridge, 4 feet; then the visible part (for 14 feet) of another hollow, the slope of which is flanked with big angular and rounded blocks of white millstone grit, measuring from 11 inches by 8 inches to 21 inches by 24 inches—perhaps to strengthen the sides of the fosse. Reaching now Bridgewater Street (30 feet wide), which, of course, is inaccessible for underground examination, we are left to conjecture. The only partially-seen hollow probably runs on for another 14 feet. Close to the new police station begins another

hollow, down whose descent a flight of stone steps seems to have been carried, consisting of a series of large, roughly-shaped, white millstone grit blocks^[14] in situ, which measured 27 inches by 13 inches, 16 inches by 11 inches, 12 inches by 9 inches. They were placed in the blue marshy soil, at a depth of 62 inches. There I met with some small fossilised bones, burnt bones, decayed wood and rootlets, and a piece of Roman brick at the bottom. After this we have another gravel ridge, 13 feet wide, which is succeeded by another long stretch of a marshy depression, 46 feet wide, extending to the back of the new police station at Worsley Street.

We have thus a longitudinal section of 198 feet, drawn at right angles, from the northern wall.

Summing up again, to make the matter clearer, and substituting fosse for hollow, we have then—

Fosse I. ... 16 feet wide, 4 feet to 5 feet deep.

5 feet, gravel ridge or rampart.

5 feet, rubbish-ditch.

2 feet, gravel ridge.

Fosse II. ... 16 feet.

6 feet, gravel ridge.

Fosse III. ... 22 feet.

4 feet, gravel ridge.

Fosse IV. ... 14 feet, visible part 3 feet to 4 feet deep.

(?) 14 feet, probably carried under Bridgewater Street.

(?) 14 feet, gravel ridge (?), probably carried under Bridgewater Street.

Fosse V. ... 27 feet, flanked with stone steps.

6 feet, gravel ridge.

47 feet, marshy depression.

198 feet.

In the first fosse, close to the wall, scarcely any pottery has been found; when we recede from the wall it becomes abundant, and more so near the edges of the fosses; at 90 feet away, the almost complete fragments of a large amphora (4 feet high) was found by me.

Whitaker, of course, misses the true character of the northern defences, and, being guided by his eyes alone, was naturally led away by mere appearances. What he saw was certainly a large hollow, 35 yards to 40 yards, or 105 feet to 120 feet wide, as he says, and this hollow is also well shown on Laurent's old map. The slope on the wall-side is nothing else but the large talus of soil and rubbish accumulated in the course of ages, and the northern termination of his fosse would coincide with the gravel ridge, past our Fosse IV. The intermediate four fosses were of course levelled up long ago, and would appear to him level ground, and formed thus the bottom of his so-called great fosse.

We learn from Whitaker that the northern defences, or his great fosse, runs 30 yards beyond the eastern end of the wall, and terminated in a high bank, which was raised upon the course of his ditch, and sloped away into the former part of it; the north-west part sloped north and south, and the north-east part sloped south and west, and was sunk 5 feet deep and extended 75 yards in length beyond the wall, even to the great road (Alport Lane).^[15]

This latter incidental information is of particular interest; it simply shows that the northern fosses turned at right angles, that is, they swept away in a curve far beyond the north-east side of the castrum independently, without inter-branching with the eastern ditch on the side of Knot Mill.

The construction of the northern defences, I need scarcely remark, differs entirely from what we see at Wilderspool (Veratinum) and Ribchester (Bremetonacum). We must again remember that the ground on the northern side consisted of marsh, low swamp, and undulating gravel ridges, which would offer natural advantages in the adoption of this parallel work.

A similar arrangement is observed in the Roman station at Birrens, in Annandale (4 acres in extent), where on the north-east side we have also a series of parallel trenches, "thus establishing a resemblance between this work and the stations of the German limes not yet proved in any other Roman or native work in Britain." At Birrens we find six ditches bending round the angles without coalescing.^[16]

At the edge of the marshy depression, inside the new police station, a little black pit^[17] was discovered. A few yards away from it in situ on the river gravel I found a corroded brass coin, which Mr. Robert Blair refers to Antoninus Pius (138-161). Now above this gravel we have in succession a stratum of Roman soil, about 22 inches thick, abounding in ornamental Samian ware, typical, according to Professor Haverfield, who examined these finds, of the second century; then upwards a gravel road, 3 inches deep, and on the top of it again 19 inches of Roman soil with quantities of Roman pottery in it.

This part of the section then, as I read it, seems to mark four distinct phases: (1) The undisturbed Roman period, when the ground was still mere marsh and bog; (2) conversion into fosses; (3) the filling up of the fosses when no longer required to make room for the suburb; (4) the formation of a road over it in the second century; (5) the final accumulation indicating when the station fell into decay.

Gates.—We have once more to return to the northern wall. The situation of the Porta Principalis dextra evidently was a little west of Collier Street, now occupied by the little coalyard. It was 6 feet wide, but unfortunately its structure could not be traced, the upper part having been taken out for a culvert. At its base we find a course of rectangular stones, varying in length and height, 11 inches and 12 inches long, 19 inches by 10 inches and 24 inches by 12 inches; these, with the exception of the latter, which is of white millstone grit, were all of red sandstone. Only the internal part of the wall being exposed to view, it was, of course, impossible to follow the passage outside the wall and the access to it from the fosses. The stones were placed close together without the use of mortar. At Whitaker's time (see pp. 36-7, vol. i.) the gateway was still rising 4 feet to 6 feet high.

He also tells us (Principal Corrections, p. 12) that in the northern wall itself small circular arches were left, at certain distances, five or six in number, the facing of one side still remaining over two of them. They were found below the natural surface of the ground. He noted them carefully in 1772. One that was accidentally laid open from end to end disclosed the design of all. He says, "As the Romans carried their walls upwards they took off the pressure from the parts below and gave a greater strength to the whole, by turning little arches^[18] in their work and fixing the rest of the wall upon them. We must again bear in mind that practically the northern wall was directly laid in the marshy ground and stood on the first fosse which reached up to it. Whitaker omitted to give the distances between the set of arches and their dimensions.

Towers.—The recent excavations at Ribchester and Melandra have shown the existence of four corner turrets built within the rounded angles of the walls. Their structure is particularly well shown at Melandra. They are built square and stand detached, 4 feet away from the angles, measuring 10 feet by 11 feet, the walls being 3 feet thick. Indications and traces of a similar turret were found at our station by Whitaker in 1768–9. He says: "A man hunting for treasures made trenches through the foundations of the stationary walls, near the south-east angle, and carrying his operations along the interior line came to a new and distinct wall, lying 3 feet to 4 feet from the wall, which proved to be 3 feet thick and 12 feet long, fairly curving at the angle." Whitaker followed up the work where the treasure seeker had left off, digging downwards for 1½ yards, and found this wall ranging parallel with the line of the station. It consisted of large and squarish flakes of red rock, and was cemented with a new species of clay mortar, a brown, compacted mass of sand and clay, tempered with some sprinkling of lime. Each irregular layer seemed to rest, he says, upon a course of fine sand, and the lowest was framed of the massiest stones and lay on a deep bed of sand that had been previously beaten and compacted together, each spade-depth of it appearing successively smooth and hard upon the surface and the third lying on the natural gravel. (See Whitaker, page 13.)

The stations at Ribchester, Melandra, and Manchester were no doubt all built on the same general plan, and turrets must have existed at our station on the remaining three angles. Traces of the turret situated in the north-west angle in Manor Street, which part is still unexplored and little interfered with, may yet be found at a future time, when the property on the north-west side of the station will be pulled down.

Interior Buildings.—Inside the present timber-yard, about 48 yards to the west of the eastern gate, in a central position, we have still preserved the remains of the wall of a building which, judging from its situation, indicates doubtlessly the site of the prætorium; the wall still rises about 26 inches from the ground, its thickness is yet 2 feet wide, and its length 20 feet.^[19] Unfortunately, this old monument of Roman times is fast crumbling away for insufficient shelter, and the rainwater is slowly but surely disintegrating and dissolving the piece away, as I have convinced myself from recent examination.

The prætorium at Melandra is similarly placed and measures 25 yards square, the inner walls 2 feet and the outer walls 3 feet thick. In its rear it has three additional buildings, analogy leads us to think that the same arrangement prevailed here.

The construction of the Altrincham railway in 1849, which diagonally cut through the whole area of the castrum from east to west, destroyed and effaced whatever foundations were in existence, and narrow-minded officialism prevented antiquaries from entering the precincts to note the many discoveries which are sure to have been made. To formulate any idea of the internal buildings, and how they were planned and assigned, we are thrown upon the researches now made at Melandra, Ribchester, and Wilderspool for completing the lacunae. All these places were erected at periods not much removed from each other, and we know already that in the main features they were much the same in plan and structure. When these excavations are complete we shall be able to realise much better than at present what Mancunium probably looked like.

Internal Area.—I have already pointed out that the southern, and, probably, also the central part of the castrum, stood on a higher level than the more northern part, and that the ground was therefore levelled up. In May, 1898, a long trench was made in Collier Street, leading to Southern's timber-yard, from the large stone post at the passage of the small coalyard to inside the railway arch, where we are on intra-castral ground. After penetrating the layer of rubbish and modern paving, we come upon a course of paving stones, underlaid by stiff, earthy clay, yellow soil, and stones. This shows that the area, just as at Castleshaw and Ribchester, was levelled with soil and stiff clay, and that it was paved with stones. Mr. Howarth, who tried to watch the excavations in 1849, also mentions the existence of pavements inside the castrum.

Drain.—On the 6th May, 1897, on driving the foundation for the new railway pier on the west side of Collier Street, an arched drain or culvert was found, 12 feet below the surface, and raised on the rock. It was 4 feet high, the arch spanning 3 feet in the centre, the bricks were 6½ inches by 2½ inches by 2½ inches, of "bull nose" shape. At the bottom it was laid on flags or blocks. I only arrived on the scene a few days after it was destroyed and used up for concrete, but I obtained a sketch from the foreman, an intelligent man, to whom I am indebted for other help. The drain seems to have been constructed like that laid open at Birrens, which is engraved in the work cited above.

Ovens.—The excavations at Wilderspool and Melandra afford a good idea of their construction in these little stations; two have been found and described by Mr. May at the former place, and at the latter locality an oven has recently been discovered by Messrs. Hamnett and Garstang, near the southern turret. At Manchester several ovens were laid bare in 1789 by Mr. Perry, cut in the rock, which, to conclude from the situation, appears manifestly to have been placed, as at Melandra and elsewhere, near the south-east angle. Close to, as in Wilderspool, he found millstones. Of these I have also recently discovered three large segments, two of coarse millstone grit and one of trachyte. Another such one is noticed by Mr. May at Wilderspool.

Wells.—A square well was discovered in 1830 a few yards to the west of the wall. It had four upright posts at the angle, closed in with other logs and floored in the same manner, the logs 5 inches to 6 inches diameter, and a number of boulders or ballistæ were found at the bottom. Whitaker (see p. 360) speaks of another well outside the station which was in evidence in 1764–5. It was sunk for several yards into the rock. "When the little alehouse,^[20] which now stands opposite the gate of the Castlefield, was erected, upon opening of the ground to form the cellar of the house a hole appeared in the rock about six feet in the square and filled with rubbish; three coins of brass were found in it and a piece of thick, short gold wire.

Cemetery.—In the rectangular piece of ground, bounded on the north by the Roman road to Buxton (which struck off from the Ribchester road a little north of the northern gate) on one side, and the east gate and the stem of the roads to Slack and Chester on the other side (see diagram), a number of sepulchral urns have been unearthed at different periods.

1762. Whitaker records the discovery of an urn with bones on the eastern boundary of the field (vol. i., p. 59).

1765. Another sepulchral urn, same field, lower down the slope, 7 feet deep, and resting on the rock.

1849. Numerous urns were discovered again here, also a grave of a vertical form, cylindrical, cut in the rock, charred bones were found in it (Corbett).

1832. A tile tomb was found on the opposite side (south side) of the Medlock. This is evidently near Great Jackson Street, close by the Roman road to Chester, where many other Roman sepulchral stones have been secured. The coffin was of oak, and enclosed in a casing of flanged tiles, 20 inches long, 16 inches broad, and 2½ inches thick.

We see that all the sepulchral urns and tombs are situated at the sides and along the Roman roads which issue from the north-eastern side of the station, following in this respect the usual practice of the Romans.

Exterior Buildings.—While the more northern part was thus appropriated for sepulchral purposes the remaining space, described by the eastern ditch, the Medlock, and the south-east and south wall, was covered with a number of large and extensive buildings. Enclosed as they were by the northern and eastern ditches and the steep banks of the Medlock they were well protected and defended on all sides. Whitaker marks the position of these buildings on his "Ground Plot of Manceinion," taken August 8th, 1765, with the letter A (in his fanciful way, he calls them British foundation). The groundwork was found ½ yard below the surface. The more northerly building measured 16 yards by 12 yards, the walls were 6 feet thick and the remaining height 3 feet, the door "which almost covered the whole side" looking north. The foundation walls consisted of three layers of common boulder stones, bedded in clay, a little lower in the field, and running for 30 yards to 40 yards together the same foundations were found, and a single layer of small boulder stones, bedded in clay, resting on the plane of the rock (pages 24–5, vol. i.). The purpose or nature of these large official buildings cannot be given now.

Crossing Deansgate, and where the Crown Inn is situated (see vol. ii., p. 84, 1775), just behind the Roman well (already described) and apparently by the Buxton road, on a plot of land then called Alport Field, was discovered in 1776, "a long line of disjointed foundations, commencing nearly from the well, extending 1 yard in breadth, 14 yards long, and 2 yards in depth, consisting of unhewn stone, broken bricks, and adhering mortar."

Water Mill.—"A little above the ancient ford of the Medlock, the sluice of a mill was accidentally discovered about twenty-four years ago (1747). There, on the margin of Dyer's Croft and opposite to Mr. Markland's^[21] constructions, the current of the Medlock accidentally swelled with the rain, and, obstructed in its course by a dam, broke down the north bank and disclosed a long tunnel in the rock beneath. This tunnel I have since laid open in part. It appeared entirely uncovered at the top, was about 1 yard in width, and another in depth, but gradually narrowing towards the bottom. The sides showed everywhere the marks of the tool upon the rock and the course was parallel with the channel. It was bared by the torrent only for 25 yards in length, but must have been evidently continued for several yards further, having originally begun, as the nature of the ground evinces, just above the large curvature in the channel of the Medlock." (See Whitaker, p. 316, vol. i.)

Northern Roman Suburb.—Whitaker writes in 1771 (see p. 205): "Betwixt the Castlefield and Alportfold is an area of 16 acres to 17 acres, now chiefly converted into gardens. It lies immediately without the northern barriers of the station, it extends up to the new houses in Campfield, or Tickle-pitcher's Lane [see Tinker's map], and the new church [St. John's Church, founded 1769]. The soil on the southern part [towards the station] of this area is absolutely one great body of adventitious earth, fragments of bricks, pieces of hewn stone, and remains of urns. Huge blocks of millstone grit have recently been dug up with the mortar adhering to them within the circuit of the area, and the whole level of the ground appears to have been traced with streets of regular pavement in a variety of directions across it."

In his "Plan of the original town of Manchester, about 300 A.D.'" he has laid down the outlines of the various streets which were constructed on the two sides of the Ribchester road in four rectangular blocks. He says: "To the one original street (the Ribchester road) others were gradually annexed. Indeed, in 1765–6, a pavement near the south-west extremity of the area, extending more than 2 yards in breadth, and seeming to tend nearly parallel with the original street of the road, was found, and this must undoubtedly have commenced with it by a cross street. This actually commenced with it by no less than three cross streets: (1) One was laid along the margin of the northern fosse, and the remains of it, a narrow causeway about 1 yard in breadth, have been recently dug up for several yards; (2) another was lately found along the northern hedge of the first great garden; (3) a third, about an equal distance from both, remaining only about ½ yard in width."

And, in 1765–6, another pavement was discovered, situated almost as much to the east as the former was to the west of the principal street, lying 3 yards in breadth and ¾ yard in depth. This commenced at the northern hedge of the Castlefield, and in the middle of the private garden, pointed less obliquely across the neighbouring lane, and evidently carried a direction towards Alport Fold. Another street was found by me, in August, 1897, in the centre of the block between old Trafford Street and Great Bridgewater Street, which pointed to Tickle Street (crossing Deansgate). It was carried here over an old pit, filled up with clay, bits of charcoal, and Roman pottery, and could be traced for 12 feet, and was paved with flags of red sandstone, some measuring 12 inches by 9 inches by 5 inches, others 8 inches by 4 inches by 3 inches, of which I numbered twenty, all in situ, on the right side. (See Plan.)

In 1839, in laying the foundations of the "Hall of Science," in Tonman Street, which is situated immediately on the west side of Ribchester Road, a bronze statuette of Jupiter Stator was found in the ground, with a silver coin of Trajan (98–117). The ornamental bowls of Samian ware which I have collected on the northern area (in Bridgewater Street, Worsley Street, Liverpool Street, &c.), mostly refer to the second century. This would give us an approximate indication of the date when this suburban part was founded. In the later empire the whole force lived outside the forts, when not on duty. Mancunium was only a small station with a garrison of perhaps not more than five hundred soldiers, besides women, traders, artisans, craftsmen, natives, &c., and these probably dwelled mostly on this side of the station.

As we have no records of the discovery of regular foundations between Camp Street and the station, either by Whitaker, who would surely have specially alluded to it, or by subsequent local testimony, we may conclude that no buildings of important character were erected over this area, and that the houses were only of the ordinary or inferior order. The "huge blocks of millstone grit with their mortar adhering to them" found "within the circuit of the area," as Whitaker says, may more likely come from the station itself when dismantled and turned into a "quarry in the middle ages."

But the Romans did not confine themselves, as Whitaker assumes, to the west side of Deansgate only, which he fancies was the proper "Brito-Roman town," but they also occupied the whole of the land on the east side comprised between Great Bridgewater Street to Hewitt Street, bounded on the east by Watson Street, on the north side, and by Rowe Street in Gaythorn on the south side, as proved by the recent excavations in connection with the Great Northern Railway. We have seen already that foundations were discovered in Whitaker's time at the corner of Trafford Street (Crown Inn), running 14 yards in length; also, close by, a Roman well, and in 1840 at the same spot a large hoard of Roman coins, mostly of gold and silver. Mr. Esdaile also speaks of some stone structure, not far away, on the banks of the adjacent Tib, and we have further attestation by the marvellous number of Roman objects I have obtained from the area under discussion at almost every footstep, particularly in Gaythorn and Knot Mill Station. Here again the ornamental Samian pottery of the second century type prevails; at Gaythorn Row Barritt reports the discovery in 1788 of a lump of Sal ammoniac, together with a Roman coin of Tetricus (267–272), and some bronze rings, one with a blueish fluted bead on it. The number of finds, from Knot Mill Station up to Rowe Street and across the Rochdale Canal tunnel to Trafford Street, are almost unmatched for variety and intrinsic value. The Roman soil at Gaythorn, which overspreads the original pre-Roman yellow soil as a stratum of 3 feet thick, is packed with pottery of every description, iron nails, charcoal, lead, broken tiles, fragments of glass. Along the railway arches of the Altrincham line, between Gilbert Street and Mount Street, the excavations in August, 1897, which were carried down to a depth of 17 feet in the gravel, have incidentally demonstrated a change of the bed of the Tib since Roman times.^[22]

In the ancient river silt along these arches the excavators handed up to me a number of complete urns and Samian pateras and patellas, and another fine large bowl embossed with ornamental work and a circular inscription, which unluckily was stolen by some boys from the cabin and smashed.

Almost in the centre, between Gilbert Street and Mount Street, I laid open and excavated an old botontinus (see plan), which measured 7 feet across; its upper part was covered with a thick layer of charcoal, and the rest to the bottom was filled with a grey, sandy, soft loam, containing Samian ware, burnt bones, glass, lead, amphoræ, black-and-white ware, iron nails. It was dug into the river gravel, which forms the original subsoil at Gaythorn. Its depth was about 3 feet to 4 feet. Another Roman botontinus appeared at the corner of Great Bridgewater Street in August, 1897, facing No. 330, Deansgate, which I have photographed (see plan). It spreads in basin-shape; its greatest depth (of the first layer) is 20 inches, and consists of Roman soil, pottery, iron nails, bricks, calcined bones, blackish-brown sandy loam, and at the bottom contained the upper part of a big amphora. Below this we have a 4-inch layer of charcoal, then clay, and at the base the pre-Roman yellow soil, followed by the usual river gravel. Later on the vicinal road, already described, was laid over it.

These botontini, which demarcated the boundary of some territorial property, refer, of course, to an earlier time, when the life was more contracted; later on, when the place was more opened out and attracted settlers and traders, it outran its original bounds with an increasing population.

But we have not quite done. So far we have shown a Roman occupation of land on the northern side of the Medlock, but it did not terminate here. We discover evidence of Roman activity on the southern side of the rim across Knot Mill Ford in Hulme sufficient to show their presence. First, we have Whitaker's statement of the water mill, opposite Markland's check manufactory at Mount Pleasant (occupied in 1772 by John Wallford), at the top of Jackson's Lane, immediately on the left side of the Chester road. The same authority relates the discovery of a Samian bowl, stamped Advocisi; of an amphora, stamped Vabeo; of an unglazed pitcher, stamped Nonovi, and another Samian bowl, stamped Of. A. Ascui. We also have the fragmentary centurial stone, inscribed Coh. I. Fris., found south of the Medlock. Then there is the tile tomb found (1832) on the same side, and no doubt, according to Roman custom, located on the side of the Roman road, and the three sculptured stones dug up (1821) in Hulme, 6 feet deep, out of the river gravel (sub-soil),^[23] which I consider to have been here in their original situ. I have only recorded the few instances specially noted and known, which mere accident has preserved; but it is clear that there must have been numberless finds of whose existence no notice has been taken. Roman antiquaries by some unfortunate infatuation, partly due to a

Sepulchral Monumental Stone, from Hulme

(red sandstone). Carved in relief on a block 22½ inches high by 24 inches wide. Compare also figure No. 169. p. 79. Catalogue of the Roman Inscribed and Sculptured stones in the Grosvenor Museum, Chester, by F. Haverfield, Chester 1900.

Dr. Colley March writes with respect to it: "It is undoubtedly Mithraic. The Phrygian cap, the inverted torch, the weary posture, point to a setting sun, to the western solar phase. The caviform niche is worth notice. The sculpture is perhaps, only part of a larger scheme."

Such figures also occur on sepulchral monuments in Germany. A Frisian centurial stone was also found in Hulme, which is suggestive.

Now in Peel Park Museum, Salford.

Monumental Effigy, from Hulme.

Carved in red sandstone. In relief; height 28 inches; base 19½ inches. Represents the head of a female with a coronal of braided hair; a similar mode of dressing is seen in the first illustration of Earwaker's Chester.

Now in Peel Park Museum, Salford.

Monumental Figure, from Hulme.

Of red sandstone. In relief, measuring 22 inches high by 18 inches wide at the base. The head has been broken off. Dr Colley March observes: "It may be a German Matrona, and so might well be 'standing, fully draped, and holding fruit.'"

Now in Peel Park Museum, Salford.

constant reference to Whitaker and directing their eyes on Castlefield alone, have failed to grasp the situation. They should systematically have watched all excavations in Gaythorn, Hulme; the area between Chetham College and the Hanging Ditch; the elbow in which ancient Salford was situated, and the area enclosed between the confluence of the Medlock and Irwell at the foot of Castlefield and at Ordsall. Our ideas of Roman Manchester would then have been clearer and more satisfactory and creditable.

From our general survey of the distribution of the more substantial buildings we come to the conclusion that the principal and public buildings were erected to the east and south of the station and enclosed by the curve of the Medlock (the northern bank). The northern side of the ground, more or less moist and swampy, was parcelled out to the garrison and the mixed population; and to Gaythorn and Hulme, on the two banks of the Medlock, where we are on rising ground, with better drainage, and a ready access to the river, the higher ranks may have resorted. Here also the Romans, when building this solid fort on the eminence at Castlefield, may have erected their temporary workshops until their walled fort was raised and finished, for it offered the best convenience for that purpose.

Roads.

Road to Condate (and Chester).—It constitutes the second Iter and is 24 Roman miles long. According to Whitaker, it issued at the eastern gate on a common stem^[24] at the extremity of the field, then winding along at the less precipitous bank higher up at the old ford of the Medlock, having Great Jackson Street on the left, leaving Hulme on the right, it passed the Cornbrook, near Pomona Gardens, where, as I am informed by Mr. Healy, a ganger, an old structure, consisting of a stone arch of red sandstone, topped by another one, was found 11 feet below some twenty years ago. Then it passed along to Stretford and Broadheath. On issuing from the eastern gate it was 14 yards wide and 1½ yards deep. Even in Whitaker's time all traces in the immediate neighbourhood of Hulme were already obliterated. Mr. George Esdaile exhibited in 1885 a drawing of a section of the Chester road, ¾ mile long, according to which the Roman road in that direction was formed of a layer of gravel boulders, a layer of red gravel, a thick bed of gorse, ling, and brushwood, in thickness from 3 feet to 4 feet. At Edisbury, in Cheshire, the road was found 30 feet to 36 feet wide, and thus seems to have pretty well kept up its uniform width throughout its length.

Road to Coccium (Wigan), 13 yards wide, issued from the western gate. It was 17 Roman miles long and formed the tenth Iter. It probably avoided the morass, crossing at a very acute angle the high bank which fringed the western wall, descending the crest near the south-west angle of the station, and following the southern bank of the Medlock to the Woden's Ford^[25] (= Hulme Bridge). According to Barritt, it was a fine paved causeway, and pointed out yet (in 1832) as a broad ridge of gravel and stones on the south side of Regent Road, in the first field on the west side of Ordsall Farm.^[26]

Road to Cambodunum (Slack, and York), 23 Roman miles long. It was at its issue from the fork at the east gate 14 yards wide and 1½ yards deep, made of melted bricks and broken millstones, constructed with strong gravel, mingled with large boulders and fragments of sandstone. It was seen by me again in August, 1897, in the block between Trafford Street and Great Bridgewater Street, slanting diagonally through the central area in a direction pointing to the site of the Wesleyan Chapel in Great Bridgewater Street. It could be traced for 19 feet. Starting from below, it rested on the yellow pre-Roman soil, in which I found a hazelnut imbedded, then follows: 3 inches stiff, yellow-reddish clay; 5 inches burnt brick, iron nails, small rubble stones, charcoal, and pieces of pottery; 8 inches boulder stones, 8 by 6 by 4 inches; 8 inches middle-sized gravel; 2 inches rubble stone and clay; total, 26 inches; topped by a further deposit—probably of a later period—of 19 inches loamy soil and gravel dispersed through it pieces of coal and Roman pottery; 26 inches rubbish (modern).—See section.

At the Wesleyan Chapel the ground must have sloped very considerably, for Roman pottery was found here at a depth of 14 feet to 15 feet.

At Ancoats Lane the road is 16 yards to 17 yards wide, with a ridge of ¾ yard of gravel and ¼ yard of marl. In Butler's Lane it is reduced to 5 yards to 6 yards.

In the miry expanse of Newton Heath, where it pointed directly upon the chapel, it was 8 yards to 10 yards wide, and consisted of a ridge of ½ yard of gravel.

At Street Lane it reached the margin of the moss. The construction of the road here is of great interest. Whitaker says: "The road is carried for no less than 400 yards across the hollow of the moss. I made a large incision into the bed of the turf beneath the Roman gravel; they appeared plainly to have originally trenched the line of the moss, the larger and more solid plates of turf they appeared to have laid upon the original face of the moss and raised the level of the line with them more than a yard in height. Upon sinking a pit along the sides of the gravel and for 1½ yards into the black soil, no ling or heather was found upon the surface of the soil and immediately below the gravel. It was first found about a yard below the surface, and then found in considerable quantities."

Close to either margin of the "Street"^[27] and 3 yards under the crown of the Roman road Whitaker found frequently the fir, birch, and oak (p. 311). On this road a hoard of Roman brass coins (a.d. 135–235) has been found at Hollinwood; at Oldham Park a silver denarius of Domitian (a.d. 81–96) and a patera at Oldham. This road was also well described by Thomas Percival in 1751–2.

Road to Aquæ (Buxton), striking off, according to Whitaker, from the street on the northern side of the ditches that extended from the northern wall^[28] and intersecting the Slack and Kinderton road at Alport Lane, and proceeding along the left of Gaythorn Fields, and crossing the Tib beyond its mounted Calley Banks and along its edge into Garrett Lane to Garrett Hall, and passing along the margin of the Medlock into Longsight. Traces of this road were seen by me in Owen's Yard, at the corner of Bridgewater Street, during the excavations. It seems to have been a boulder-paved road at the beginning.

Road to Ribchester.—The description of its course by Whitaker is so minute that it is rather surprising how little Thompson Watkin and others have troubled to work it out. On the ordnance map, 5 feet to the mile, it is laid down along the present lower Deansgate, which, of course, is quite wrong. Watkin disposes of Whitaker in a few laconic words and says he prefers Just's account, who, however, is silent, and only touches upon the road from Hunt's Bank as his starting point. If Watkin had availed himself of Berry's (1750), Tinker's (1772), and Green's (1787) maps Whitaker's account would have been intelligible enough.

The road issued at the northern gate in Collier Street (48 yards away from its eastern angle), passing the parallel ditches. Whitaker says: "It was found 1765 in the adjoining garden still visible from its ridge, 5 yards wide. In 1751 it was found in the second garden,^[29] proceeding in the line of a hedge [this garden is bounded by Priestner Street and Tickle Street], bordered with large squarish stones at the sides, raised into a convexity of ½ yard above the ground. Crossing the narrow lane beyond both some trace of the convexity lately appeared and pointed across the level of Campfield to Mr. Philips's two houses in Quay Street [which is the house marked 45 on Berry's map and would be near Reiss Brothers' warehouse]. There the road was discovered in 1760 near the doorway of the more easternly house, more than ½ yard below the surface of the ground, from 4 to 5 yards in width, and more than 1 yard deep in strong gravel. In the gardens and in Campfield it appeared to be continued in a slanting line considerably to the east of it. This obliquity of the road was necessarily occasioned by the great curve of the Irwell."^[30]

Whitaker adds (see p. 121): "In the second garden, near Castlefield, and on the site of Mr. Philips's house, the pavement was dug up, consisting of the largest boulders, and having two layers of stones upon a bed of gravel."

On the 14th January, 1898, when the old deanery, in Deansgate, was pulled down for the new arcade, next to Mr. Armstrong's shop, the road was discovered by me again. It was found 5 inches below the foundation of the old deanery. The top consisted of a layer of sandstone flags, then clay and rubble or red sandstone, 5 inches; gravel, 6 inches; blackish soil, bricks, charcoal, iron nails, scoria, 4 inches; yellow pre-Roman soil, below the river gravel.

Whitaker has a remark (vol. ii., 407): "The parsonage plainly proves to have been near the course of the old and then forgotten road to Ribchester." I have taken photographs of this section of the road.

I picked up traces of the road again in Wood Street, which enables us to produce the slanting road from Quay Street to St. Mary's Street. The road proceeded then along upper Deansgate, falling into Hanging Ditch.

To sum up, the Ribchester road on leaving the northern gate in Collier Street passed across Bridgewater Street, Worsley Street, Liverpool Road, Tonman Street, Dumville Street, Tickle Street, upper Quay Street, Hardman Street, Cumberland Street, Spinningfield, Wood Street, upper Bridge Street, Lower King Street, St. Mary's Street, and then along the present upper Deansgate.

It was bedded on a stratum of strong gravel, ½ yard to 1 yard deep, bordered with large squarish stones at the sides, at its lower end, paved with boulder stones; and, at the deanery, it was flagged with white sandstone, and resting in a layer of clay and gravel, &c.

The road to Condate (the Second Iter) was 14 yards wide and 1½ yards deep; to Coccium (the Tenth Iter), 13 yards wide; to Cambodunum (the Second Iter), 14 yards wide and 1½ yards deep at the castrum, but constantly varied in width, and contracted to 8 yards and 10 yards at Newton Heath and onwards; the one to Ribchester was only 5 yards wide,^[31] but more elaborated and flagged as we have seen.

The construction of the roads is so intimately connected with the history of the Roman conquest of Lancashire and Yorkshire^[32] that it is necessary at this point to devote a few remarks to the subject for a clear understanding of the probable date of their formation. We have, therefore, to concentrate our attention for a few minutes on the position of Chester, as the military focus, at the estuary of the river Dee. It was probably in existence already, as a fortress, about the latter part of Claudius's reign (50–54), and held by Suetonius in his campaign against Anglesey (61). The chief work (in the words of Professor Haverfield) was a double defence against Wales and the Irish pirates, for which its position was admirably fitted, by enabling the Roman generals to operate by sea and land. With the northern belt of forts stretching from Segontium, along the Eleventh Iter to Deva, and connected down the south by Uriconium with Caerleon, on the Severn, Wales, in course of time, had little chance of serious resistance. Ten or fifteen years later, it was occupied by the second legion, and after that by the twentieth. Chester (and Caerleon) had all the characteristic features of a distant borderland, the province was purely of a military order and, in reality, a military frontier.

On the eastern flanks they had to reckon with the troublesome and powerful people included under the generic name of the Brigantes, a confederation of minor clans that dominated practically the backbone of England, comprising practically Lancashire, Yorkshire, Westmorland, and the Midlands, from sea to sea, and from the Mersey to the Solway Firth on the west side, to the Humber and the Tees on the east side. To master them it was necessary to strike from the two sides, with Deva and Lindum as a base line.

The first serious attempt on the west was probably made by Cerealis in 71–74, to effect which he had to traverse the territory of the Cornavii, who dwelt in Wirrall and along the southern banks of the Mersey, and camps and fortifications were placed through the forests and heathery plains of North Cheshire. Condate is on this line; from thence they tracked to Mancunium to enter South Yorkshire for an attack on their capital. Our station was of no little strategical importance in this large scheme, and thus became one of the original links of a great base, which required a firm and powerful hand. The double road from Hunt's Bank and the eastern gate led direct to Cambodunum and Isurium, and must have been amongst the first constructed. Cerealis reduced a great part of the district without, however, having settled the country. The final reduction was reserved for Agricola (78–85), who then completely cowed the Brigantes. It was probably in his time that Mancunium was made into the impregnable final stone-built stronghold. To guard himself securely against the piratical coast tribes of West Lancashire the Segantii, who nestled in and invested the densely wooded creeks and estuaries of the Ribble, Wyre, and Lune, harassing by land and sea, he threw another close line of forts up from Condate^[33] and Manchester to Wigan and Walton for the protection of the Ribble; continuing it viâ Kirkham and Poulton to Fleetwood—the Portus Segantiorum—and another one from Walton to Lancaster to the Lune.^[34] This is in all probability the next road in priority; the direct communication to Ribchester and Lancaster (disusing Walton) falls into another scheme, and was meant for a different purpose, when the conquest was carried further north to Scotland for the domination of the Picts and Scots. The other secondary road, such as the one to Buxton, is of later growth and lesser significance, and need not be considered in this place. To sum up, in order of time, we have: (1) The road from Chester, viâ Condate, to Manchester; (2) the double roads to Yorkshire, of which more hereafter; (3) the road to Wigan and the estuaries, probably built by Agricola; (4) the later direct road to Ribchester and the north.

Alport.

The area of Alport Town proper, as far as Roman discoveries are concerned, has proved barren. The red rock comes pretty close to the surface, little of the original soil is left, and on the Deansgate side it is covered with about 4 feet of rubbish. At Gratrix's warehouse, on the site of which many great Roman or Anglo-Saxon discoveries were expected, no finds or foundations were discovered.

So much has been fabulated about Alport that a few words must be devoted to the discussion of its etymology. Professor W. W. Skeat informs me that: "Ald is the Northumbrian and old Mercian spelling of what we now spell old. 'Port' is very common in early English. The sense 'town' is as old as Edward the elder and Athelstan; it is not, however, a true English word, being borrowed from the Latin. Port also occurs in the sense of gate, but this is unusual and only found in quotations from the Gospel or the Psalms, where the Latin version has porta. Practically it need not be considered. For port in the sense of road I find no authority whatever. It follows that Aldport almost certainly means old town or old village, and such a form may easily be as old as a.d. 900." Our earliest reference is: Alde Parc, 1282; Alde Porte, Al-Ald-Aldeport, 1320–2.

In 1322 it covered an approximate area of 95 acres, as described in Mamecestre, and comprised, according to John Palmer,^[35] the whole area between the river Irwell and the Tib, and between the Medlock and the present Quay Street, which practically coincides with the boundary lines of Roman Manchester, excluding, of course, the Hulme fee.

In the earlier maps of Berry and Casson (1750) we find a small block of buildings, then a farmstead, called All Port Town, which extends across the present Deansgate into Camp Street.

The almost total absence of Roman pottery, &c., during the recent excavations seems to show that it must have been left to mere cultivation,^[36] being on the outskirts; at its eastern side it was crossed by the road to Slack.

Any occupation of the site as a town by the Mercians or Northumbrians is most improbable, as it would have been the worst place one could choose for ready defence or protection. The old Roman fort or castle must by that time have been already dismantled, a mere wreck of skeleton walls, and a previous prey to Picts and Scots. It is more in harmony with practical sense to assume that the rocky ridge at Hunt's Bank, which forms a natural defence, became their burgh, and that the old disused Roman town in Deansgate was in contra-distinction called the Aldport, or old village or town; Aldport was separated into Over Aldport, which in 1473 was a close of pasture (towards Castlefield), and Nether Aldport, which in 1473 was a park (towards the deanery). Aldport Lodge or New Park is mentioned 1567–88.

The popular idea of the antiquity of Aldport Town must be discountenanced; it is a more modern creation, dating from the early part of the last century, when Deansgate proper outgrew its old boundaries at the deanery, and crept on slowly down the two sides to Knot Mill.

Roman Manchester (Aldport) had its centre in Knot Mill, the Northumbrian, Mercian, and Norman Mamecestre kept round Hunt's Bank and the old Baron's Hall. The discovery of Anglo-Saxon sceattas in Campfield (eight coins, dating from 450–600, and now in the possession of Mr. Hardy at the Portico Library) proves nothing. They were found, we only can say, on a Roman highway that led both to Chester and York, and may refer to the time of the struggles between Eadwine and Æthelfrith.

Recent Excavations between Dyer's Lane and Fleet Street.

The recent excavations, at the end of April, 1900, at the east side of Deansgate, along the Great Northern Railway, for the erection of shops between Dyer's Lane and Fleet Street have offered an excellent opportunity for examining a large space of ground, of which the lower strata of original soil have been left undisturbed, thus preserving and allowing a study of the Roman deposit.

We have on the top: 55 inches of modern soil and old brick foundation, (a) 33 inches of cultivation soil (Roman), (b) 17 inches of trodden ground (Roman), 14 inches coarse sand (river gravel), and at the base the red rock.

(a) Consists of a light, sandy clay, free from any other admixture or accumulation, and showing the presence of little roots (casts and the outer bark), with scarcely any or only sporadic Roman crockery; it has all the appearance of cultivated ground. Below, in (b), the upper surface of the 17 inches is rather stony, and made of tough clay, which apparently had been traversed and trodden and tightly pressed. In it we have a good many fragments of black, grey, and red ware of large size, and parts of Samian bowls and dishes and some Roman mortar, it is much mixed with charcoal and also contains bits of iron in it, pieces of wood, and a seed of Atriplex patula was found at the bottom of the underlying sand, which, of course, formed the original surface. This section shows again in a striking manner the constant change of level. Practically, in Fleet Street the original level was 7 feet below the present flags. At the Wesleyan Chapel in Old Trafford Street Roman pottery in large profusion was discovered by me on the gravel at a depth of 14 feet to 15 feet, at Eltoft Street it occurs at a depth of 6 feet, which demonstrates again the undulating nature of Deansgate at the time of the Romans. The posterior levelling up of the whole district has wiped these features off and makes us forget what it really once was.

The river Tib (now culvered over) formed a conspicuous Roman boundary, and carried a great volume of water in its pristine state, running, as it did, in many small curves, it often overleapt its boundaries and over-flooded its banks. It is found now in Chepstow Street, on the premises of Messrs. Wadkin and King, as recent excavations have shown (August, 1897), 9 feet 4 inches below the surface. Its banks are there covered with 7 feet 6 inches of rubbish; its activity is shown by a layer of 2 feet of black and greenish clay, teeming with decayed leaves, roots, and branches, followed by 5 feet green and drab clay, which rests on the usual river gravel and boulder clay.

The Series of Buildings on the North Bank of the Medlock.

One of the most interesting discoveries in connection with the station was accidentally made in 1771 on levelling the bank of the Medlock for the construction of a new coal wharf. The foundations, which were then disclosed, were situated within the irregular semi-circular projection below the level of the station, 25 yards distant from the edge of the water and directly opposite the small bridge on the opposite side of the river corresponding to the site of the present coal wharf in Castlefield Street. The bridge is marked on Green and Laurent's old maps, and we have no difficulty in locating these Roman structures.

Thompson Watkin has alluded to them in a very superficial and unsatisfactory manner, notwithstanding Whitaker's minute description, which will be found in his Appendix. It is due to the co-operation and ability of our esteemed member, Mr. J. J. Phelps, that I can add a plan of reconstruction of these buildings which he has worked out in scale from the measurements and details put on record by the former writer. In referring to the plan we notice that in front of the first building a pedestal was placed on a flooring 7 to 8 inches thick, made with pieces of soft red rock and bedded in clay; 8 feet immediately to the east was

Building I., 20 feet long and 10 feet wide, floored with Roman cement (mortar and pounded brick), the floor, 9 inches thick, "resting on a body of marl about as many in depth; at the further end from the pedestal were found human and animal bones lying under a decayed arch of Roman bricks, and fragments of urns." Nine feet away to the east followed

Building II., 30 feet long and 10 feet wide, made of the same cement, and lying 2 or 3 feet lower in the ground, and covered with flags. Below this came a second floor, composed of pilæ of millstone grit blocks and square tubular tiles, 16 inches high, resting on a cake of cement, 2 feet thick; this, again, lay on a third flooring of pure mortar, 3 inches thick, on which was erected a series of pilæ of flat square tiles and flags, and 3 feet high, the whole built on a subsoil of red sand. The exterior wall of Building I. had a thickness of 2 feet 3 inches, Building II. of 4 feet, formed of regularly dressed stones. The space between the pilæ of floor iii. was closely filled on every side with loose earth. Three feet to the east occurred

Building III., but all a mere mass of confusion. In the rubbish were found a couple of Roman coins, and three round tile tubes, 16 inches in height, inserted into each other. It is clear from the description that the three buildings formed the hypocaust of the station,

representing probably the frigidarium, the sudatorium, and the præfurnium. The pedestal, which stood about 2 yards to the west of the first building, was evidently in its original site, being fixed on a regular base of red rock and clay, and Mr. Phelps suggests that on it the altar, dedicated to Fortuna Conservatrix by the centurion Lucius Senecianius Martius, of the Sixth Legion (which was stationed at York and came over to Britain about 121 a.d. with Hadrian), was placed. Such a conjecture is not very improbable when we recollect that the altar was found in 1612 in the bed of the Medlock, not far away from the position of these buildings, which were close to the edge of the river (only 25 yards away). It is easy to conceive that the altar, originally occupying the base, into whose measurements it marvellously fits, was thrown down from thence into the water at a later time. Mr. Phelps, on paying a special visit to the Ashmolean Museum, Oxford, where it is now preserved, examined the altar very carefully, and found in the centre at the top a hole, sunk 2 inches deep, probably for the fixing a separate focus, which is gone, and may still be embedded in the Medlock not far away from the spot where the altar was found.

Another point has to be taken into consideration. We must infer from the plan that the sudatorium was built apparently on an anterior and older hypocaust, which either had been destroyed or rebuilt for some reason subsequently, for otherwise we cannot understand the meaning of the third floor with its duplicate series of pilæ, filled on all sides with loose earth, and rebuilt over with a thick flooring of cement. We know that Antoninus Pius (138–161) had some troubles with the Brigantes, and also that the Caledonians in the reign of Commodus (180–192) ravaged Lancashire, and we are inclined to think that the station may then have been closely beset and suffered structural damage;^[37] that for its relief and repair a body of troops may have been hurried down from York or the Great Wall, and that possibly the centurion Lucius Senecianius Martius,^[38] in commemoration, erected this altar to Fortuna Conservatrix. Another altar, dedicated to the same genius, has been found at Cilurnum, on the Roman wall; another at Netherby in Cumberland.^[39]

But to return to the hypocaust. I mentioned that between the eastern wall and the ditch, which defended the east side, two other large buildings were found. These seem to have freely communicated with those placed within the semi-circular area on the south side of the station by a paved causeway, and access to the bed of the river and the river boats was gained by the flight of steps midway between them.

The presence of blocks of millstone grit in the foundation walls of the station manifestly proves a previous intimate acquaintance with the borderland of the south-western Brigantes of Yorkshire, for the stone could not have been carried a day's march from the station unless it were already tolerably well held by Cerealis and Frontinus. We may, therefore, assume that its final erection, built with strong walls, is due to Agricola. Agricola's military schemes were on a more elaborate scale, and his ambitions included the whole reduction of the north, with Lancashire and Yorkshire as a base of operations. To this circumstance the erection of Mancunium as a strong ROMAN ALTAR FOUND AT CASTLEFIELD, 1612.
⁠NOW IN THE ASHMOLEAN MUSEUM, OXFORD.

border-fortress must be attributed. From the material which has been collected and discussed in the previous pages we shall be able to offer a slight picture, so far as the results of these investigations permit us legitimately to do.

General View of Mancunium.

The northern and southern walls measured respectively 175 yards, and the western and eastern ones 140 yards; the station consequently covered a little over 5 square acres and was of rectangular shape, the corners rounded. It was provided with four detached turrets in its angles, and the walls had an average thickness of 7 feet. The foundations were made of clay, boulder stones, rough sandstone, bound together with copious mortar. It had four gates, all irregularly placed, and the northern wall was apparently strengthened by the insertion of six small circular arches at the base. The defences consisted on the northern side of five or, perhaps, six parallel fosses and small ramparts; on the east side a fosse was drawn round to the Medlock, cut into the rock;^[40] the southern side was naturally defended by the curve of the river, whose bank was additionally scarped for greater security; and on the west side rose a lofty bank, skirted by a swampy slope that fell away to the Irwell. The northern wall was built immediately into a swampy hollow, and stood in the first fosse, without the appearance of rampart or berme. No trace exists of the gates, which at Melandra were double-arched and had guard-rooms attached to them. At the south-east angle were placed the ovens, cut into the rock, and a well was made near the western wall,^[41] where probably the altar of the Rætians and Noricians was also placed. The principal buildings, comprising the prætorium,^[42] and probably the granary, were about the^[43] centre of the station, still indicated by the remnant of a piece of wall, 20 feet long, which has been permanently preserved. The cemeteries followed the margin of the road to Buxton and the stem of the road to Slack and Chester immediately on their issue from the station. Sepulchral monuments on the south side of the Medlock line also the side of the road to Chester. The hypocaust was built on the southern side of the station close to the banks of the river, into which steps were made for access to the water.

The area of the castrum was paved and drained; the south-eastern part, which stood on the rock and was much higher than the north-western part, was levelled with clay and earth. A number of roads issued from the gates—on the eastern side we find the road to Slack (York) and Buxton and to Chester; on the western gate, leading to Woden's Ford or Hulme Bridge, we have the road to Wigan and the estuaries.

From the northern gate, near Collier Street, we can yet trace in part the road to Ribchester, which at Hunt's Bank sends off a branch to Blackstone Edge. Outside the station we have on the east side, between the wall and the ditch, some important and large buildings that were connected by a paved road with the hypocaust. On the north side of the castrum, rectangular to the Ribchester road, a number of parallel streets were built, reaching up to Camp Street and probably even to Quay Street, the occupied area being defined to the west by Lower Byrom Street. This part of the suburbs was most likely used by the soldiers and traders.

Stepping across Knot Mill, we have evidence of a large building and a Roman well at the Crown Inn (Trafford Street). Gaythorn has yielded a very rich field of finds; so have Trafford Street and Great Bridgewater Street, which also must have been largely occupied. We must now cross the Medlock; here on its south banks we obtain traces of a water mill and other evidence which makes it clear that it was also populated. We have seen already that the erection at Trafford Street and Gaythorn of botontini, and the later construction of a road over the former one, indicates an expansion of the boundaries of the settlement in that direction. Towards Fleet Street and Alport Town, on the east side of Deansgate, the occurrence of finds is remarkably thinning out. Recent excavations have shown that here the original soil is only covered to a slight extent by a layer of Roman "trodden" soil; it seems to have been left in later times to free cultivation, for it is covered with a layer of good light sandy clay almost free from any fragments of pottery. Taking, now, a general view, the "larger" Mancunium was enclosed and bounded on the west by the Irwell, on the north by Quay Street, on the east by the banks of the Tib, and on the south by the Medlock and certain parts in Hulme alongside of the southern banks, perhaps reaching to a little beyond Great Jackson Street.

The station was built by some auxiliary cohorts of the Frisians, we have the names of the first centuria of Masavo, and Quintinianus and Candidus. Those of Candidus and Masavo were seen in situ by Camden and Dr. Dee in the walls, while the centurial stone of Quintinianus was found under the rubbish at the eastern gate. Another one of the first centuria of the Frisians (officer's name incomplete) was found at Great Jackson Street, across the Medlock. These three cohorts alone built together seventy-one passus of the compass of the walls, and the general work was probably directed by an officer of the Twentieth Legion, of which a tile was discovered. These troops were all detailed from the headquarters of the chief station at Chester.

Later on we meet with tiles of the Third Cohort of the Bracaræ and an altar of the Præp. Vex. Raet. et Noric. and another erected by an officer of the sixth legion, all of which point to York and the northern Roman Wall, and tending to the conclusion that in the following century it was garrisoned not from Chester, but by troops stationed in the north. It is, of course, impossible to define the strength of men by which the station was held. It was never anything but a small third-rate castrum and similar in size (5 acres) to the forts along Hadrian's wall, which may have held each a complete cohort or ala of nominally one thousand or five hundred men (in some cases certainly one thousand).

The formation of the suburbs probably began in the second century, for most of the earliest ornamental Samian pottery scattered over its area, as between Bridge Street and Camp Street, and Gaythorn, Knot Mill, and Hulme, consists of specimens characteristic of types of the second and third century. Of the pottery inside the station nothing can be said, as a systematic search there has never been made. In order to realise the appearance of Castlefield Station and its immediate vicinity at that epoch we must keep before us the fact that the whole district in subsequent times has been more or less levelled and artificially covered by 4 feet to 5 feet of soil and rubbish. In its original state the surface was very uneven and its undulating character gave rise to a succession of little hillocks, valleys, and hollows, abounding with pools, springs, and rills. The white valley-gravel skirted the whole length of Deansgate, and was everywhere exposed to view, and on the light and yellow soil which over-spread it at Knot Mill, Gaythorn, and in the central parts the birch and hazel grew. The plants and mosses, the oak, hazel, alder, birch, bracken, which have been found, tell their own tale. Their habitats speak of turfy bogs, of woods, dry heath, moorland, woods, and shady banks; the land shells conduct us to hedge banks, garden walls, and grassy banks. Even in 1322^[44] there still existed in Alport 30 acres of heathland, 20 acres of pasturage, 3 acres of meadow, and the unenclosed wood that stretched 1 mile in circuit. The Court Leet Records remind us of Sowce Hills (1564) and Pygott Moor (1642) in Deansgate. The Irwell, Medlock, and Irk, indented by knolls, doughs, and little gullies of picturesque red rock, were overhung with spreading trees, and their crests marked with broom, heather, and gorse; the heights of Cheetwood were occupied by moorland and oakwoods, and opposite the station extended the Shaw (forest) of Salford. The river beds and banks of the Medlock, Tib, and Irwell threw their curves closer to Deansgate, often flooding the lower grounds; the wild boar and the wild cattle roamed still in the neighbourhood.

Later on the more military character of the settlement passed away. It remained in the stage of its earlier days, its importance as an actual fortress had ceased. Attention was then more seriously diverted to the north, where greater concentration was needed. After the complete pacification of Lancashire and Yorkshire, its intrinsic military utility was lessened, the inhabitants went on probably in their quiet way, cultivating the soil, clearing and draining the land, and laying out many vicinal roads or occupied with barter and trade. But there was always sufficient stir and bustle to keep the place alive. The traffic and transport of troops from London and Chester to the north must have been unceasing, for the station formed a link in the great artery of the Second and Tenth Iter, and it was a little cosmopolitan beehive, crowded with soldiers, officials, traders, and natives.

We have evidence of the navigation of the Irwell and Mersey by the discovery of the keel of an oaken canoe found in 1897 in excavating Nos. 7 and 8 docks at the Ship Canal, Old Trafford, at the depth of 15 feet to 20 feet in the gravel, and canoes at Barton (Sticking Islands), and at Veratinum (Wilderspool); by the golden bulla dredged up in the river near Eccles, and in the finding of coins in the channel of the Irwell at Victoria Bridge (306–340), Blackfriars Bridge (249–255), Quay Street Bridge (98–141), and the lid of a jar in May, 1899, 12 feet deep in the gravel of the Irwell at the Parsonage.

The great number of iron nails of all sizes, found outside the station in the suburbs, shows activity in the erection of houses, the iron scoria, cinders, and charcoal and mineral coal demonstrate the presence of smiths and smithies. We have sheet lead and leaden nails and stamps, the nails for fixing the antefix; plain green-coloured window glass, golden wire and rings, pewter dishes; glass and stone beads and blue fluted beads; bronze pins and rings, bracelets, bullæ, scales, ladles, statuettes and figures, fibulæ of silver and bronze, brooches, buckles, glass bottles and vases, plain and embossed, of greenish and blue colour; iron axes, knives, and styli, bronze celts and spearheads, terra cotta lamps, Samian ornamental fictile bowls and vases, and the thousand and one articles used in kitchen and scullery, such as dishes, jars, stands, tubs, mortariæ, amphoræ, ampullæ, pots, platters, urns, &c., too innumerable to specify. The fair amount, amongst them, of objects de luxe shows that the garrison and population enjoyed their full share of opulence and comfort, and, to judge from the enormous amount of pottery and articles found by me alone over a comparatively small area of the suburbs, the accumulation of the articles of daily life must, during a busy occupation, extending over three to four centuries, have become very great. Much has been lost during the last fifty years by indifference and ignorance, but what is left is sufficient to form a slight picture of the interesting life in Mancunium in its best days. When the place was more fully established, roads and by-roads were made which facilitated intercourse and traffic to Buxton (Aquæ) for taking the hot springs; pigs of lead were carried both by land and river from the Deceangi in Flintshire; salt was obtained from Northwich (Salinæ); slate from Wales and the Lakes; millstone blocks, for building purposes and querns, were quarried at Castleshaw and Blackstone Edge, and coal from the outcrops at the banks of the higher Irwell; the masons fetched their sandstone blocks from Collyhurst to carve and inscribe their altars and commemorative stones; and the traders of Samian and other ware, journeying from the Rhine and Gaul across the channel to London and by the Second Iter up to the north, dropped their loads at Wilderspool, Melandra, and all the many Lancashire stations. Castor ware is rare both here and at Wilderspool, for Durobrivæ, one of the chief places of manufacture, was on an iter (the fifth) which lay outside the direct connection with the Second Iter; white-ware is not so common; while coarse black or Upchurch ware is found at every step. Some of the commoner pottery appears to have been made in the locality.^[45] We have no palpable evidence, as at Ribchester, of the fate of the station on the "departure" of the Romans about 411; I particularly searched for charred wood work and other proofs of destruction or conflagration, but found nothing pointing to such a catastrophe. The latest coins discovered refer to Anastasius I., 491–518; close to the Ribchester road, at Higher Broughton, which clearly shows that the stations cannot have been entirely deserted or abandoned, but must have existed long after Ribchester, which, being more exposed and within easier reach of Picts and Scots and other marauders, was apparently already sacked or even destroyed. I have also obtained from the northern suburb at Bridgewater Street some large bronze buttons, one of which was gilt and another tin-plated, of peculiar make, which, according to Mr. C. H. Reed, are in make very like a series of studs on a Merovingian shield boss in the British Museum. He adds: "They may be Roman and, if so, of very late date, perhaps even fifth century." Mancunium, after this, disappears from history; some traditionary legends, however, lingered on through many centuries—even reaching into Whitaker's time—of a famous giant Tarquinius, said to have lived in Mancastle, at the Medlock. We have a few Anglo-Saxon sceattas (ranging from 450–600), as already mentioned, found on the Roman highroad at Campfield, which possibly may fall within the time of the struggle between Eadwin and Æthelfrith, but we are left in darkness until the tenth century, when the Mercians rebuilt Manchester, which had been destroyed by the Danes. No milestones or additional altars or inscribed stones have been saved from the station, and thus our picture remains blurred, consisting of mere outlines, and with our paucity of inscriptional records we are unable to impart fuller life to our sketch.

Comparative Table of some Features of

Ribchester (Bremetonacum). Excavated by Mr. John Garstang.	Melandra (Zedrotalia?). Excavated by Messrs. R. Hamnett and J. Garstang.	Wilderspool (Veratinum?). Excavated by Mr. Thomas May.
Berme, 25 feet; fosse, 20 feet wide, 6 feet to 7 feet deep.	Two sides ditched, the other sides naturally defended, standing on sharp slopes. Still awaiting excavation.	Ditches, 7 feet wide, 3½ feet deep; berme, 7 to 11 feet, with a rude stone pavement 18 inches wide, over one portion near the south-west angle.
Walls, 6 feet thick, height probably 12 feet—the lowest 3 feet of boulders, the upper 3 feet of four courses of regular formed stones. Wall with inner and outer facings, with stones and rubble between.	Walls, 4 feet thick, composed of large boulder embedded in clay, with three courses of flagstones on the top, the outer face is of dressed stones 12 inches thick, 12 to 21 inches in length, the inner portion filled up with large boulders, clay and gravel; no inner wall has been discovered.	No walls, and defended by a rampart 9 feet wide, the lowest about 1 foot of made ground, 6 inches of loamy sand and gravel, and one or two courses of hammer-dressed sandstone blocks for facing stones, with rubble between.
Streets paved with cobble stones, floor on clay, oak spars, over it earth.	Road: the principal road, 13 feet wide, of gravel, with boulder curbstone outside; from near the south-west tower, a gravel road, 9 feet wide, runs to a small plateau in the adjoining field.	Road, 24 feet wide, just inside the rampart on two sides, formed by a bedding of made ground, a layer of broken sandstone blocks about a foot deep, and 2 or 3 inches of gravel.
Gates, in the centre for shorter sides, at one-third of the length of the wall for the longer sides.	Gates placed as at Ribchester; gateway double-arched; at the P. P. Dextra was a portcullis.	Gates: evidently at north-east and south-west angles, foundations of the latter only traced.
	Conduit, from P. P. Sinistra to the north-west tower, flagged.
Tower on west and south corners discovered so far; their walls, 3 feet thick; chambers, 10 feet by 11 feet.	Towers: four detached corner towers; near south angle, an oven.	Towers: only slight remains of foundations outside south-west angles; two ovens.
Granary and four other buildings in the centre.	Prætorium and complex of buildings adjoining: prætorium (inner wall, 2 feet; outer wall, 3 feet), 25 feet square, three rooms excavated partially, the middle one has its floor of broken bricks and the other two of flagstones; the tile floor of granary, of tiles 8, 9, and 10 inches square, 2 inches thick; been repaired with roofing tile.	Site: not excavated yet.
Shape, rectangular, rounded; 6¼ acres, walls, 600 feet by 427 feet.	Same; walls, 122 yards by 112 yards=3⅛ acres; midway between the P. P. Dextra and the Decumana entrances is the Corn Mill; the floor of a workshop to the right of the road from the Decumana to the Prætorium.	424 feet by 420 feet=4 acres; trapezoid
Pottery, chiefly second century.	Same; also fragments of first century.	Local, intermixed with Samian of the first and second centuries.
Coins: earliest, under north-west wall, Nerva (96–98); first inscription, 161 (Sixth Legion). Sacked probably 367 by the Picts. Latest coin, 364.	Coins from Domitian (81–96), Hadrian (117–138).	Coins: earliest, consular; latest, Constantine the Great (324–336).

The Segantii and Brigantes.

We have no accurate knowledge of the old inhabitants of Lancashire at this period. In a general way we know that the west coast, from about the Mersey to the Lune, was occupied by the Segantii, while towards the slopes of the Pennine chain, and between the hills of Manchester and Leeds and Sheffield, we had the south-western Brigantes.

The former clustered, and were settled, in the estuary regions of the Ribble, Wyre, and Lune, which gave them great security, and made it difficult to dislodge or attack them. Agricola had to grip them from the two sides, from the land and sea concurrently. Casting our eyes on a map of the Roman roads^[46] to the shore lines we at once perceive that the part between Liverpool and Walton is singularly devoid of any known roads or stations. It has been argued before, most convincingly, by Jos. Boult, that the Mersey^[47] was formerly a pure freshwater lake, and that the marshes of Bidston and Wallasey were connected with those of Bootle and Crosby; that in Roman times the Mersey had no distinct existence, its estuary being of much later date, and that it was included most likely by Ptolemy with the Ribble, Alt, and the minor streams in the Æstuarium Belisama; that the Seteja has to be sought between the mouth of the Dee and the shore line of the Wirrall. However that may be, we know that at Dove Point, on the Cheshire coast, the Romans were in full evidence in the present neighbourhood of Great Meols, as shown by the multitude of objects discovered in the old forest surface, consisting of pottery, great quantities of precious fibulæ,^[48] and coins ranging from 41–388;^[49] while nothing, in spite of close research, has been found at or about Liverpool until we reach Formby. To judge from some Roman objects found at Tarbock, Downholland, and Formby, and in a direction from Wigan to Upholland and Scarisbrook a few tracks to the sea may have been in existence, but otherwise the stretch between Great Meols and Walton-le-Dale, on the Ribble, forms a blank as far as Roman manifestations are concerned. If the Mersey had been an estuary Agricola would certainly have built a stronghold and constructed a road to it, as he did at the Ribble and Lune. These collateral facts appear to establish an additional proof that the Belisama must be considered to represent the estuary of the Ribble, and that probably the Segantii spread thickest between the Ribble and Fleetwood, which latter is accepted as the Portus Segantiorum; that, consequently, more strictly their centre comprises the Fylde district. At Fleetwood (formerly called Quaggy Meols) a considerable number of coins was found in 1840, from Vespasian to Caracalla and from Constantine to Honorius, and a large paved platform, 8 feet from the surface.

It is well known to geologists and antiquaries that the whole coast line has undergone a complete change since Roman times; the estuaries were more extensive and stretched much further out to the sea; the land surface has contracted considerably, and was continuous from the Wirrall to the Lune, forming a large (now submerged)^[50] forest belt, densely wooded with birch, ash, oak, fir, pine, and alder; large herds of deer, the wolf, the boar, &c., roamed in the woodlands along the course of the Ribble, evidenced by the subsequent discovery of a surprising number of skulls and antlers during the excavations of the Preston Ship Canal, and we may conclude that the Segantii followed the chase and the fishing of the estuaries and the sea, and were also more or less piratical and in close intercourse with North Wales, if not the Isle of Man. They lived in the forests, creeks, and crannies in the many large meres with which their district is studded.^[51]

We know that they belonged to the Brythonic branch of the Celts; at Ribchester we have their gods Maponus (Apollo) and Cocidius (Mars). We have the many rivers ending in "der" (=dwr, water), such as Docker, Conder, Coker, Calder, Hodder, the Wyre, and Don, Tand, Dardow, Darwen, Alt, Lune, Douglas.^[52] The Segantii probably obtained their name from their habitats; they were a riparian set of people. It is held that the root is derived from sæ, sæwe, seo (A.-S.), saiws (Gothic), segeven (Faroese) =pontus, mare, thus making them true "water-dwellers,"^[53] and at Lea Hall, a little above Walton, we have the river Savok also, variously written Sausk and Saugk. Their centre of gravity seems to point towards Walton and Ribchester, which Agricola and his successors made into a terminal or intercepting stronghold, and which formed not unlikely their chief places of defence, and their importance for a hold on the entire adjacent coast line cannot be overrated.

I have to say now a few words about the Brigantes with whom our own district and Mancunium in particular is more intimately identified. As the name already expresses, these peoples, who formed a powerful and widely extended confederation, were a proper race of Highlanders, who occupied the entire length of the Pennine chain and its slopes and ramifications from the Solway Firth to South Yorkshire. They were composed of many clans, in Annandale (Birrens) they may have merged into the Goidelic branch, but the central and southern portion belonged doubtlessly to the Brythonic division. Brigant, plural brigantiad, Latinised Brigantes in modern Welsh means highlanders, mountaineers,^[54] from the root brig,^[55] top, hill summit, Irish bree, Gaelic braigh. They were a fierce and dauntless race of fighting men and depredators, the terror of the Lowlanders, who made frequent descents to the plains, and gave, as we know, no little trouble to the Romans, who found them a strong, rebellious, and restless nation. The southern limit of the Brigantes was about south of Leeds and Huddersfield. Of their territory the southern portion was the more important one, the northern regions were uncivilised, and, perhaps, half uninhabited. Cartismandua's kingdom probably centred in South Yorkshire.

The hoard of coins found at Honley, near Huddersfield, consists of five British silver coins, with the obverse of Volisios Dumnove and Volisios Carti(o)ve, and eighteen Roman coins, finishing with Vesp. Cos. III. According to Mr. G. F. Hill,^[56] the date of the deposit is fixed at a few years subsequent to 73 by the fact that the four coins of Vespasian are only slightly worn and the British coins in fresh condition. He classes these South Yorkshire coins with the Brigantes, and refers the Volisios Carti(o)ve to Queen Cartismandua. The hoard was probably hidden during the British wars against Frontinus or Agricola.

We possess also some altars which refer to the god, the goddess, and the nymphs of the Brigantes, as: Deo. S. Berganti, at Longwood, near Huddersfield; Dvi. Ci. Brig, at Greetland; Deæ Victoriæ Brigant, found in the Calder, near Castleford, South Yorkshire; Deæ Brigan, Adel, near Leeds,^[57] which shows their strength in South Yorkshire.^[58]

The Second Iter, viz., the Roman road from the eastern gate of the station, viâ Newton Heath, Hollinwood, Castleshaw, to Slack (Cambodunum), and the supplementary road from Hunt's Bank, viâ Long Millgate, Castleton, Rochdale, Littleborough, Blackstone Edge, to Sowerby Bridge, went straight into their country, and these hill people, who held the Yorkshire highlands so close to Mancunium, had a natural command of the western slopes which run out and spread to the foot of Manchester. We may conclude, therefore, that Mancunium was inhabited by some little local clans or offshoots of the south-west Brigantes scattered along the heights of the river courses of the Irk, Medlock, Mersey, Irwell, Roche, and the Tame.^[59]

The Britons who dwelled in our locality were Brythonic Brigantes; the Segantii, also probably of the same blood, lived far away at the west coast of Lancashire, in the Fylde, &c., separated from our centre by dense stretches of scrub, forest, and swampy, miry heaths and mosses, with difficult tracks to traverse, while the Cornavii were confined to the southern banks of the Mersey.

Description of the Pit at the present New Police Station, Bridgewater Street.

I come now to the finds in the black pit or pool. It measured 4½ feet across and 5 inches to 6 inches in depth, and was discovered near the margin of the fosse or swamp. In consequence of aqueous fermentation and subsequent pressure, the vegetable matter of which it is composed had assumed the appearance of compressed and lamellated peat. I had the contents of the pit conveyed home for closer examination, and found it mixed with white sand grains and fairly-sized rounded stones, and little square pieces of sandstone, the latter mostly blackened by the action of fire. This little pit seems to have served for the deposit of rubbish.

It contained matted layers of roots, stems, bark, and wood. Some of the branches showed the sharp edge of the knife or hatchet, some of the wood-chippings were charred on one end. The bits of sandstone flags and boulders must have come from some improvised fireplace, just as we see the navvies putting them up now. The botanical specimens extricated from the pit are described separately. In addition it contained bones of domestic animals and cinders. Among the miscellaneous objects I have to name a broken tile, some black-ware with a peculiar metallic glaze, and rather heavy in weight; iron nails, lumps of burnt clay, and a large piece of tanned and dressed goat skin.

Unfortunately no other rubbish-hole has been discovered at Castlefield, although I have carefully searched every place; their rich and varied contents generally adds a large amount of collateral evidence.

List of Mosses

(Named by Messrs. Holt and Thomas Rogers).

	HABITAT.
Thuidium tamariscinum	woods and banks.
⁠,,⁠delicatulum	limestone rock in woods.
Hypnum cuspidatum	margins of pits, marshes (common).
⁠,,⁠purum	shady banks, amongst grass.
⁠,,⁠exanulatum	swamps and wet places.
Hylocomium splendens	on grassy banks, in woods (common).
⁠,,⁠squarrosum	banks and woods (common).
Brachythecium rutabulum	on banks, walls, trees, and wet places.
Mnium affine	marshy places, shady woods.
Neckera complanata	on trunks of trees, walls, &c.
Polytrichum commune	on dry heaths and turfy grounds.
Aulacomnion palustre	turfy bogs and marshes'.
Sphagnum cymbifolium	bogs and turbaries.

Ferns.

	HABITAT.
Blechnum spicant, branch	shady woods.
Pteris acquilina, branch	in sandy, turfy soil, on hill sides and open	woods.
Polypodium dryopteris, scales	woods.

Trees.

	HABITAT.
Fir under the Roman road in Moston, Failsworth	gravelly loam soil.
Oak, leaves, also under Roman road in Moston, Failsworth	top loam soil, gravelly subsoil.
Birch, bark and branches, also under Roman road in Moston, Failsworth	heaths, woods.
Alder	moist situations.
Willow, leaves, perhaps two or three different species	such as grow near pits and banks of brooks.
Hazel, bark and nuts	good, light, moist loam soil.
Juniper, shell of a berry	sandy hills, woods.
Sloe, half of a stone	in hedges and sunny slopes.

Plants.

Heather under Roman road at Moston
Rumex obtusifolius, seed	waste land.
Polygonum lapathifolium, seeds	waste land.
Carex sp., seeds and rhizome	wet places.
(?) Comarum palustre, rhizome	pits.
Ranunculus acris, seeds	waste land, pastures.

Animals.

Wild boar, two tusks.
Ox, skull and bones.
Sheep, teeth.
Dog, impression of the paw on a brick.
British cattle,^[60] one fine complete skull and three detached long horns, found October, 1896, in Water Street, 20 yards from Bridge Street, 12 feet deep in gravel, it was not acquired by the Manchester Museum although I informed it of the find, and is now lost to Manchester.

Land Shells
(Named by Mr. Thomas Rogers).

	HABITAT.
Helix aspersa	common on walls and about gardens.
⁠,,⁠nemoralis	on walls and grassy banks.
⁠,,⁠hispida	hedge banks, under stones, in woods.
⁠,,⁠arbustorum	hedge banks, limestone, woods, and found near walls where nettles grow.
Hyalinia (Zonites) cellaria	under stones in shady places.

These were found on the marshy soil, near Collier Street, under the northern foundation walls of the station. Apparently they were already dead and empty shells when finally deposited in this situation. They must have been swept down from a higher level by rain floods or the current of a rill or brooklet that had its course along the valley or clough existent here at the time. This would account for the heterogenous and promiscuous occurrence of these shells, all found mixed up in the same nest.

The brownish-black Roman mortar which I obtained in trench No. 3, when dissolved in water, went into a soft pulp, due to the large admixture of pure vegetable mould. On examining it with the lens, I found in it small bones and a few fragments of land shells. It not only shows that the builders mixed their lime with mould or humus dug out from the close proximity, but it gives also an interesting glimpse into the natural history of the locality at the time the station was erected.

Batrachia.

	HABITAT.
Newt, skeleton bones	in pits and ponds, from under the Northern foundation wall, named by Professor Arthur Thomson, of Oxford.

Coleoptera, &c.

Aphodius sp., a dung feeder	$\scriptstyle {\left.{\begin{matrix}\ \\\\\ \\\ \ \end{matrix}}\right\}\,}$	from the black pit.   named by Mr. J. R. Hardy, of the Manchester Museum.
Phratora vulgatissima, feeds on plants, willow leaves
Niphus hololeucus, feeds on seeds and grain

Glass, &c.

(1) A fragment of green glass, ¼ inch thick, ornamented with projecting pillar ("pillar moulding"). I have another piece of this class, of blue colour, ⅛ inch thick. See also description in The Antiquities of Richborough, Reculver, and Lymne, by C. Roach Smith, 1850, p. 76 (and the drawing). It has been found previously in London and Richborough. Pillar moulding, Mr. Apsley Pellatt remarks, is the greatest modern improvement in glass making, and was supposed to be a modern invention; he cites these specimens to show that it is really a revival of a lost ancient art.

(2) A stamped devise on the bottom of a large vessel, green coloured, thickness ¼ inch.

(3) Fragment of a rim of a large urn-like vessel of circular form, also used as a cinerary urn, ¼ inch thick, green.

(4) Various pieces of thin white and greenish window glass, ¹⁄₁₆ inch to ⅛ inch thick.

(5) Broken neck of a bottle, twisted and melted by the action of fire, and covered with an oxydised white coat in and outside.

(6) A green oval flat-bottomed bead, ⅝ inch long, 3½ inches in diameter.

(7) Round green bead, ³⁄₁₆ inch in diameter, probably set in a fibula or other ornament as a decoration, the oxydised particles of the bronze (turned blue green like malachite) still adhering to it when found.

(8) A circular white stone bead, top flattened in the centre, ⅝ inch in diameter.

Lead.

(1) Large pieces of sheet lead from the Botontinus, Gaythorn.

(2) Leaden nail, head ⅞ inch long and squarish, stem quadrangular, still ⅞ inch long (broken). The antefix of terra cotta, used in architecture to cover the frieze, was in many cases fastened to the same with leaden nails; from Bridgewater Street.

(3) Large leaden nail, 4 inches long, without head, Bridgewater Street, and various lumps of lead.

(4) Leaden seal (impression oxydised away).

Iron Nails, &c.

These are very numerous in Gaythorn and Trafford Street, and Bridgewater Street; they all have very large heads and the stem is quadrangular. Sizes of lengths in inches:—

⅞, 1⅜, 1½, 1⅞, 2⅜, 3⅛, 3¾, 4, 4⅛.

The head is from ¼ inch to ⅞ inch in diameter; also a lid or hinge, 1²⁄₈ inches long and ⅝ inch in centre, with three perforations for the nails.

Bronze.

One spear head from Bridgewater Street, 2⅞ inch long.

Marble.

Half of marble ball, diameter 1¾ inches.

Fine stopper of a marble vase, crusted and the surface eaten away like sugar by the action of carbonic acid. Size: 3 inches long, 2 inches across. From Gaythorn.

Samian-ware.

The so-called Samian-ware was to the Romans what porcelain is to us moderns. It was the finest production of the potter. Its high value is shown by the discovery of broken vessels united by leaden rivets, of which three instances occur at our station. Some eighteen or nineteen distinct forms are recognised, all designed for table service, and consisting of wine jugs, bowls and cups, vinegar vessels, bowls or basins for soup, vegetables, and other viands, dishes of different shapes and sizes.^[61]

The finds in and about Castlefield are extremely numerous. I have collected myself at least five hundred

ORNAMENTAL SAMIAN-WARE.
From the Roman Station, Castlefield, Manchester.

ORNAMENTAL SAMIAN-WARE.

From the Roman Station, Castlefield, Manchester.

ORNAMENTAL SAMIAN-WARE.

From the Roman Station, Castlefield, Manchester.

ORNAMENTAL SAMIAN-WARE.

From the Roman Station, Castlefield, Manchester.

ORNAMENTAL SAMIAN-WARE.

From the Roman Station, Castlefield, Manchester.

pieces, plain and ornamental, and amongst them some almost entire. They were first introduced in the time of Trajan (98–117 a.d.). They are of a bright red colour, with fine glaze up to the third century, and of duller red colour in the fourth century.^[62]

Using Syer Cuming's classification, we have types of the following makes:—

Calathus (see his figure 1).—A drinking cup, 2¾ inches to 3 inches high, and 3¾ inches to 5½ inches across the top, of which I have found three specimens. They are unadorned, plain, wheel-made, and dark and light red, they are also stamped.

Acetabulum (figure 6).—Eleven specimens, some halves, 1½ inches to 2½ inches high, 4¼ inches wide, for dipping the bread before eating. One still bears the potter's stamp O.AT..., some are quite unstamped, and when stamped it is in a large circle. They are not decorated.

Galeola (figure 5).—From the river silt of the old bed of the Tib at Knot Mill Station I rescued a complete bowl, 4 inches high and 7½ inches across. In it the pure unmixed wine was placed upon the table. They are decorated on the sides with a variety of objects, combats of gladiators, mythological personages, fanciful devices, scenes of the chase, animals, birds, foliage, &c. Of this ornamental work we have a great number at our station, and fragments are in abundance at Gaythorn, Trafford Street, Bridgewater Street.

Paropsis (figure 7).—Four 10 inches in diameter. The sides are embellished with graceful scrolls and tendrils; bearing vine and ivy leaves. Occasionally hares pursued by dogs are introduced.

Patina (figure 9).—I obtained some halves, 2¼ inches high and 5 inches across, without decoration, and stamped m e t.

Patella (figure 10).—One example, fragmentary.

Patera (figure 13).—With ivy leaves, 2 inches high and 7½ inches across, various halves.

Lancula (figure 18).—A complete one, 1¾ inches high and 6 inches across, stamped o....a f;^[63] another, almost complete, 2½ inches high and 9 inches across, stamped r v f i t i . m ., devoid of decoration. They generally have a wire edge, and resemble the scale of a balance. They seem to have been made in "nests" of various progressive sizes. I have numerous "sides" of the following height in inches: ⅞, 1, 1⅛, 1+³⁄₁₆, 1⅜, 1½, 1⅝, 1¾, 1⅞. They seem the most numerous here.

Tympanum (figure 20).—A shallow salver or pan with upright sides, 1 inch high and 4 inches diameter; very few instances in Britain. The type I have is mended, and perforated for the insertion of rivets.

In addition we have imitations of the real Samian-ware. These are easily recognised by their palish-yellow paste, the dull false glaze, and dead ring and colour. Of these I have fragments of some ivy-decorated Pateræ, &c. Another specimen of Samian-ware (a large flat dish), from Trafford Street, is of bright red colour, without exhibiting any glaze. From Gaythorn I obtained two fragments of Samian-ware, the paste of pale salmon colour, and covered, inside and outside, with a dull black "slip." Most of the bottoms had potters' stamps or "labels" in the centre, which latter sometimes are conically raised, sometimes concave or flat. The stamp is either unenclosed or put in a large or small disc or rosetted circle.

The paste varies; some of the vessels have a fine, pure,

PLAIN SAMIAN-WARE.

From the Roman Station, Castlefield, Manchester.

homogeneous paste; others have it mixed with yellowish particles and fine sand. The colour is either greyish red (the worst class) or deep red, light red, and yellowish red and salmon. Some are very strong and well baked, while others are more soft and fragile. The thickness of the sides varies from ¹⁄₁₆, ⅛, ¼, to ⅜ inch. The glaze is either shining bright red or dull red, and runs in many delicate shades. The majority belongs, according to Professor Haverfield, to the second, and perhaps the third, century. Potters' marks abound, but unfortunately are often either too fragmentary or illegible, in consequence of the subsequent detachment of the fine glaze which covers it to identify them. A comparison of the potter's name from the various stations of Lancashire alone, when lists are more complete and the names properly authenticated, is of great interest, and will give us, when it is done, much information of the chronology and routes the traders followed and the stations they visited in their consecutive turns. Strange to say, we have little here of stamps found frequently in Chester and even Wilderspool. Of those of Melandra we have to await the report.

Black-ware.—Their paste is of ash grey, or greyblack colour, very coarse and brittle, and easily crumbles; it contains much large-grained sand, the inside very often rough. It has either a black glaze or a friction glaze; the thickness varies from ⅛, ³⁄₁₆, ¼, ⅜, to ⁷⁄₁₆ inch. We have such a great variety that it is almost impossible to classify them. It is soft burnt, and some of the larger parts, often 6 inches in diameter, bear a thick crust of black soot, from the action of fire. Their rim indicates that they could be slung over the fire. We have a great number of platters, dishes, the latter ornamented with diagonal lines. The variety of rims in shape and form is endless. Besides, we have many urns, jars, and pans. They are all of a coarse make, and for common use in the kitchen and table, the blackware has no ornamentation. Another kind of blackware has a yellow inside and outside slip. Blackware exists in enormous quantities.

We have another kind, very soft, with a black or grey slip, the paste red or brown, and a fragment of a diagonally-lined ware, hard-burnt, of which the diagonals are produced by friction; the inner coating consists of a metallic slip of the colour of blacklead.

Grey-ware.—Its texture is homogenous, light grey, soft-burnt; its thickness varies from ⅛ inch to ³⁄₁₆ inch. A Catinus from the inside of the Castrum has the potter's name, the only instance of such an occurrence at our station. The greyware has a variety of ornamental markings, and is often turned inside.

Blue-ware has a grey slip, the paste blueish-grey and burnt like stoneware with a little sand in it, its thickness ⅛, ³⁄₁₆, and ¼ inch. It is decorated with polished lines.

Brown-ware, covered with a brown or grey slip, and a red or red-brown or ashy-red paste. It is brittle and coarse; we find little of this ware.

Red-ware.—A great variety, paste red, without slip. Another variety has a red slip on the inner and outer side, the paste grey, or a grey or black slip outside only, and the paste red. One large fragment, which is unique, has a rudimentary ornament or scrollwork, hand-made, of crude execution. Another class has an outer thick, pale yellow slip. Many bottoms and rims, they are not so

common as the blackware. We have four specimen, bowls and dishes of a redware, the inner and outer side sprinkled with mica.

White-ware, of pipe clay, the paste gritty, unglazed, some large jars.

Yellow-ware.—Of the smaller fictile class, both the white and yellow ware is conspicuous for its great quantity. It is smoothly turned on the two sides, without glaze.

Castor-ware is strikingly poorly represented here, and only two pieces have been found by me.

Frilled Red-ware.—Of this class we have seven good specimens.

Rough-Cast Ornamentation, types of bar, lozenge, hatching, pellets, fir-cone zone, wave-line, zig-zag, herring-bone, diagonal and cross bands, indented bands, rows of dots, arranged in squares, are well represented. They are not peculiar to Britain, and recently I found many good examples in Bonn, Germany, of superior execution and glaze.

Lamps.—One of yellow-brown terra cotta, plain, 3 inches in diameter, and stamped s a m (?), was found in Bridgewater Street, now in possession of Mr. Cunningham, Corporation inspector, 110, Duke Street, Old Trafford. Another one from the former collection of C. Bradley (see Harland's MS. notes), from Castlefield, was acquired by the Salford Museum; part of another recently at Bridgewater Street. These are the only finds made.

Large Circular Lids.—These have not often been found. I have a large flat circular one with a knob in centre, convex, 8 inches diameter, grey; another of yellow colour, stoneware, 8 inches diameter; black colour, 6 inches diameter; red-ware, 6 inches diameter; all from Bridgewater Street; and another smaller one, 4¾ inches diameter, with a hole in the centre, yellow, made to fit in the vessel recovered from the Irwell, 9 feet in the gravel, in building the foundations for Orme & Son's Billiard Works. A stopper, 1⅛ inches diameter, ½ inch at the lower end, from Gaythorn.

Amphoræ.—We have a good many, found all over the area; one in fragments, almost complete, from the northern ditch. I possess the whole top and the apex of the bottom of one, handles and the annular tops are common. They are red, grey, or brown. An upper part of one was found in the botontinus at Trafford Street. I have also a stamped handle, and an other one with a graffiti . . . i v v (beginning incomplete).

Tubs or casks, resembling our modern washing tubs, of circular shape, the inside hooped and running in parallel rings. At Bridgewater Street large pieces of the bottom and top were obtained, 16 inches in diameter and probably 18 inches high. Another large jar of light yellow colour, also turned, but without rings, occurred at Trafford Street; it has still remnants of a blackish crust adhering to the inside. Chemical analysis, however, failed to discover whether it was formed by the lees of wine, vinegar, honey, or some other organic substance.

Ampullæ, two-handled, large, light yellow vessels, standing about 18 inches high, of stoneware, the large neck moulded separately and jointed with the body; the circular mouths 4, 5, and 6 inches in diameter. One I got almost complete and restored from the fragments.

Necks of light red and light yellow bottles, some in annular steps, are copious. I have collected some twenty different types; diameter of mouth, 1½, 1¾, 2, and 2½ inches.

Stoneware Basins of yellow light colour, 2 inches high, rather shallow, with broad rim smoothly turned, 7½ inches across, are in fair numbers.

Fragment of a fine Bowl, in shape resembling an acratophorum of light grey, exhibiting an unctious black slip, more pronounced on the inside, its sides ornamented with bands of undulating lines, divided by horizontal zones of straight lines, only one example from Trafford Street.

Urn.—A complete red type from the ancient course of the Tib at Gaythorn, from the old river sill.

Mortaria.—I have collected at least fifty of these. Their diameter varies. They measure at the top across from 9, 11¼, 12, to 13 inches; their depth is from 3½ to 4 inches and their thickness generally ½ inch. We have them in red, white, straw colour, light yellow, and whitish grey, and deep black grey. The red ones have a paste like the bricks, either pure or with sand, and moderately burnt. The white ones are of a fine pure pipeclay, the grey kind are practically a stoneware and give a good ring. The red mortaria have a yellow inside and outside slip, but the stoneware feels rough and is unglazed. Their inside is coated with small pebbles, of which the surface by use has become well abrated.

The deep black-grey stoneware has also a thick grey slip on the two sides. Instead of a coating with white and black or brown pebbles we find some of the grey ware studded with brown angular pieces of kidney-iron; some of the stoneware mortaria have no pebbles at all, but a very rough granular interior surface instead; in one case the inside is like a rough crater, and the layer of rough knobs of coarse sand grains acts like a sharp rasp.

They are in shape like modern milk-pans, flat and circular, with overlapping rims, and having lips or spouts of various form and width, some very graceful. On the two sides of the lips often occurs the maker's name. Of such I have six. One of my specimens, measuring only 9 inches across, is extremely shallow, with a downward sloping rim of 1 inch deep, the depth of the basin is scarcely more than ½ inch, and it can only have been used for some particular purpose. When pouring out the pulpy or mashed substance the mortaria were held with the two hands and fingers under the inverted rim, as is easily seen from its construction. I have a mortarium complete, minus its bottom.

Varia.—Red hair (probably human), from the black pit; boar's grease, from the Roman soil in Trafford Street, analysed by Mr. Robert Pettigrew; dressed goat-skin, a large piece 8 inches long and 4 inches wide, dressed and dyed black, found in the black pit.

Querns^[64] of millstone grit, 18 inches in diameter, 3 inches in height, and scooped out through and in the centre. Another one, 14 inches in diameter and 2 inches thick, probably from the Blackstone Edge or Castleshaw quarries.

One of basaltic lava, 11 inches in diameter and 2 inches high—also occurring at Wilderspool, Silchester, and many other places—from the millstone-lava quarries near Niedermendig and Mayen, on the Rhine probably, according to Mr. Bernard Hobson, of Owens College.

Slate Tiles, 6 inches long and 4 inches wide or so, often found at Bridgewater Street.

An Ovular Stone, 4⅞ inches long diameter, 3 inches transverse diameter, found in the Roman soil by me at Bridgewater Street, 6 feet below the surface, well smoothed, of fine grained sandstone, probably a grain or corn crusher, and not one of Whitaker's sling-stones or ballistæ.

Mortar and Building Material.

For the supply of their lime Whitaker suggested the limestone quarries at Ancoats (Ardwick), which appears quite correct. In examining trench No. 3 I paid particular attention to the composition of the material employed in raising their foundation walls, and by good chance extracted from the lowest layer, consisting of tenacious clay and boulder stones, a large lump, about 4 inches square, of half-burnt limestone, which must accidentally have rolled down during the building operations. On examining the piece with a lens I discovered the presence in it of the little shell Spirorbis, which characterises the Ardwick limestone. Later on I found a few more pieces in another trench. Even long after the departure of the Romans these valuable beds were remembered and quarried, for we find in 1322 mention made of the "kiln at the Ancoates" and of a "stany gate" leading to it (see Mamecestre, Gazetteer). It supposes that the vicinity must have been already well explored before the erection of the station in Deansgate. The late Mr. Mellor, a former manager of the Ardwick Limestone Works, states in a paper read at the Manchester Geological Society, some thirty years ago or so, that in one of the older workings old Roman mining tools were found. Probably the Romans conveyed the quarried limestone to the station, where, perhaps, on the south bank of the Medlock, at Knot Mill, they burnt it in open air stacks, mixing it with mineral coal, of which Whitaker found a large buried store close to the issue of the Roman road from the eastern gate. For greater convenience, and for ready supply, it was preferable to have the stack close by, instead of burning it in Ardwick. In consequence of the process employed the outside layers of the stack would naturally be left half burnt, hence the half burnt piece found at the bottom of the foundation wall.

A lime kiln existed even in the last century at Knot Mill, and a Kilncroft Meadow is mentioned amongst the lands belonging to the demesne of the lordship of Hulme in 1616. When we recollect the dimensions of the castral walls, 175 yards by 140 yards, 6 feet to 7 feet thick, and probably 12 feet high, to which we have to add the interior and exterior buildings on the southern and eastern sides, including the large and thick cement floorings of the Roman baths, we can form an idea of the large quantities of limestone which were carried to, and burnt and slaked at, Castlefield. We find various mixtures and preparations of mortar used in the station:—

(1) A brownish sort, rather coarse and more friable and earthy and inferior, used in the lower courses of the foundation walls, mixed with mould and humus and gravel and sand.

(2) A whitish cement, mixed with small-sized clean river gravel, poured between the blocks of red sandstone and the facings, as seen in the walls of the prætorial building.

(3) A fine cement, made of broken bricks and lime, used for the floorings of the hypocaust.

(4) A fine pinkish cement, mixed with washed sand, perhaps used for mosaic work.

The presence of limestone quarries so close to the station offered a great advantage for the solid construction of their walls and buildings, it made them practically indestructible. No lime was used at Veratinum (Wilderspool); at Ribchester a little was used in one of the foundation walls for grouting; at Melandra I found some mortar, made of lime and gravel, adhering yet to some blocks of millstone grit, none in the foundations or walls; while, on the other hand, I picked up pieces of mortar, composed of pounded brick and lime, in Castle Hill, Northwich, Cheshire.

The white, compact, coarse millstone grit was employed in the station, in dressed blocks for the causeway, at the bottom of the first ditch in the northern wall, at the base of the northern gateway, for the rude steps to the ditch at the new Police Station (unhewn); and Whitaker speaks especially of the large square blocks of dressed millstone grit, with the mortar still adhering, found scattered over the area of the northern suburb. These, of course, must have come from the upper portions and the larger official buildings. Millstone grit was also used for the pilae in the hypocaust, and for the manufacture of the millstones and querns. The stone was most likely obtained from the western slopes of the Pennine chain, at Blackstone Edge,^[65] or Castleshaw. The employment of the millstone grit in the foundation walls argues, as already mentioned, an earlier intercourse, and a closer penetration into West Yorkshire; and it is clear that it must already have been well subjugated at that time, perhaps by Petilius Cerealis (70–71), and quite secure, or it would have been impossible for the builders to venture into the distant territory—a long day's march—to quarry and bring the heavy dressed blocks to the station.

The red sandstone cropped out all along the Medlock, at Castlefield, Hulme Bridge, at the Cathedral, and along the banks of the Irwell and Irk, and another and better kind was obtained from the once famous quarries in Collyhurst.^[66] The former was used, rough and unhewn, for the foundations and fillings; the latter, dressed and squared, for the upper walls and buildings.

The sandstone flags were readily found in the coal formation of the neighbourhood.

The blue and green slate tiles, which I found in various parts in Roman soil, are partly from Wales and the Lake district.

The gravel for the foundations, of course, was quite at hand everywhere, Deansgate abounding in river gravel and sand.

The clay for making their bricks occurs nearest at Peter Street, Mosley Street, Market Street, Shudehill, &c. The bricks were likewise made and burnt near the station, for I found many lumps of burnt clay, with the cast of the incorporated straw stalks impressed in it. One brick has the impressions of the paws of a dog, while one circular flat brick tile has three deep human finger impressions on it, similar to specimens in Shrewsbury Museum. I give the analysis of the partly-burnt lime from trench No. 3, and of another specimen from Ardwick itself, which has kindly been made for me by my friend Robert Pettigrew, Esq., research chemist to Sir Henry Roscoe:—

Analyses of Limestones.

	I.—Spirorbis Limestone from Ardwick.	II.—Limestone from the Roman Foundations, Castlefield.
Calcium carbonate	95.50	96.00	per cent.
Calcium sulphate	0.08	0.09	,,
Oxide of iron and alumina	0.21	0.44	,,
Insoluble in acid	0.84	0.60	,,
Magnesia	none	trace
Organic matter	0.50	0.20	,,
Phosphates	traces	traces
Combined water, moisture, alkalies, &c.	2.87	2.67	,,

Bricks.—Various sizes of bricks have been discovered by me at Castlefield. They are very rarely complete. I obtained fragments, about 5 inches thick, of bright red colour, and another much thicker yet and perforated. A complete brick, 14 by 7 by 4 inches, weighing about 40 lb., from the same place, is now in my possession.

Flanged Tiles.—I have parts, 7 inches to 10 inches wide, 1 inch thick. There are various types at the station with flanges 1½, 2, 2⅛, 2¼, and 2½ inches high, differently bevelled and channelled, the width of the flanges varying from ⅞ inch to 1⅜ inches.

Whitaker found some entire tiles, measuring 16¾ by 11¼ by 1½ inches.

Howarth mentions some tubes 12 or 14 inches square and 1½ inches thick.

Flanged tiles generally measure 15¾ by 12 by 2¼ inches, and plain tiles 16 by 10½ by 1½ inches.

At Melandra tiles have recently been found 8 inches square and 2 inches thick.

No imbrices or semi-cylindrical tiles for covering the roofing tiles have been seen in Castlefield. These were generally 36 in. long, 1¼ inches thick, and 3 inches in diameter.

Both bricks and tiles and drain pipes were made on the spot, and dried in the sun, and laid flat on the ground; the bottom of the tiles is always quite rough and sandy. They were probably dried, at the most handy and convenient place in Hulme, close to the southern banks of the Medlock. Some of the clay lumps are black, of a bituminous appearance, and with white sand grains imbedded in the clay. The tiles are heavy and of a uniform and well-tempered clay, free from lime, and with a certain amount of sand in it. Their colour varies—bright red, flesh-coloured, and yellow-red; at Castleshaw the tiles are very light and of a pale straw colour. The size of the flanged tiles from the tile tomb at Great Jackson Street is 20 by 16 by 2½ inches.

The flanged tiles are of seven or eight distinct forms, as far as the bevelled flanges are concerned, and of different width and height. They may belong to different periods.

At Gaythorn I found two types which differ from the rest. They are of a blueish colour, very heavy, and more like stoneware, and possess a metallic ring. The tiles are also found scored at right angles or diagonally. Checkered floor tiles are rare here; only one was found by me. At Melandra they are rather common.

Antefix.—I have a fragment of light brown colour, rather light in weight, and moulded. It comes from the upper part and is too small to make out details of its particular form.

Drain Pipes.—Of these various have been found ¾ inch thick and probably 8 inches in diameter.

Flue Tiles were used in the hypocaust and have been fully described by Whitaker. The tile of the 20th Legion, and the two referring to the 3rd Cohort of the Bracaræ, are irrecoverably lost. Another inscribed tile, found inside the castrum, was most unfortunately snatched up last year by a man shortly before my arrival, and I have in vain tried to find the appropriator. Of the great number of tiles, in consequence of their fragmentary state, not a single one was lettered to give us some additional information about the garrison.

Mineral Coal.—Often found by me, also by Whitaker, in large deposits at the eastern side of the station. I have also the burnt shale from the Roman soil.

Charcoal (probably prepared from the oak) all over the area.

Iron Scoriæ, sometimes the charcoal imbedded in it, quite common.

other finds.

British blackware, at Cathedral Street.

Flint scrapers, below Roman surface on the north side of the station, Bridgewater Street.

Broken flint flake, Cathedral churchyard.

Large monolith, 6 feet high, 30 inches wide, of brown sandstone, below the Roman surface on the north side of the station at Bridgewater Street. Unfortunately, I omitted to examine it for any scorings or markings. The stone is not derived from the locality and is a "stranger." Its source I cannot trace, as it differs from local sandstone of the coal measures. It lay on the river gravel, and had to be lifted with pulleys for its heaviness and great size.

Potters' Marks.

• vabeo. Jackson's Lane, 1771, graffitti avittii scratched on bottom, also London and Gaul (Amphora).
• nonovi. Jackson's Lane, 1771, unglazed pitcher.
• of . a ascvi. Jackson's Lane, 1771, Samian.
• advocisi. Jackson's Lane, 1765, Samian bowl.
• titivs. Castlefield, 1848, also London, tittius.
• lvppa. Castlefield, 1849, Howarth, also London and Wilderspool.
• jvcvn(di). Castlefield, in Mayer's collection, Liverpool.

FOUND BY MYSELF.

No.
648.	met(ti). Also London and Wilderspool, second century bowl.
	jm (anni). Also Lancaster and London (jmanni).
646.	of. rvfi, of rvfin, London, second century.
647.	orif ai<.
46.	rvffini . M. Also London.
166.	rufiti m.
645.	iulio.
48.	ivlivmvi. (jvlii. ma, at Wilderspool.)
54.	fa . . i. (falendi, London.)
649.	of . . . .
87.	oc . . ot.
47.	. ritni.
49.	. mai.
644.	. . andi.
51.	. . . . . ai.
52.	o . . .

Imitation Label.

There are various of this class, which are, however, not real names, but only imitation goods.

For some further remarks on above Graffitti see Appendix.

General List of Former Finds Lost, or Yet in Existence.

Centurial Stones.—Coh. I. Fris. Quintiani, found 1796 at the east gate, size 15 inches by 11 inches, formerly in the possession of Dr. Chs. White (+ 1813), lost. Cast in Peel Park. See Barritt's original drawing, vol. v., Manch. Lit. and Phil. Soc., 1802, plate vii.

Centuria Candidi, seen by Camden, 1607, in situ, lost. "I saw this inscription on a long stone there."

Coh. I. Fris. Mas. Seen by Dr. Dee, in situ, lost. "John Dee had a sight of the same here."

Coh. I. Fris..... found 1760 in Hulme, lost.

Stamped Tiles.—c . iii . b r ., found 1832, at Castlefield, by Howarth. Lost. In the tabula^[67] of the reign of Antoninus Pius found at Cilurnum the name of the third cohort of the Bracaræ occurs as Coh. III. Brac. Another was found 1840, also lost.

l xx, found 1829, was taken to Worsley Old Hall, but lost on removal to the New Hall. Was in raised letters, according to Harland, who saw it.

Fragment of Inscription.—(G)eta. Worsley Old Hall. Lost.

Altars.—Præp . . Vex. Ræt . et Noric . at Worsley New Hall. Still extant.

The Vex. of the Raetii and Noric was stationed probably at Æsica on the Roman Wall in the reign of Marcus Aurelius (161–180), a shattered stone having been found there (Thompson Watkin), naming the Rætii. The Rætians also served in the second Tungian Cohort at Birrens. Rætian spearsmen also served at Habitancium, the modern Risingham, and their name also occurs on another stone, now as a lintel over a turret stair at Jedburgh Abbey.

Fortunæ Conservatrici L. Senecianius Martius 3 Leg. vi. Vict., from the site of the hypocaust, now at the Ashmolean Museum, Oxford.

objects of gold.

Golden ring, found by Stukely, 1723, at Castlefield. Lost.

Golden wire, found in the well at the Crown Inn, by Whitaker. Lost.

Massive gold signet ring, having a bloodstone and a figure of Mercury in intaglio, in possession of Chr. Bradbury, Crescent, Salford. Found 1839. Collection sold 1867.

Golden Bulla, found 1772 in the Irwell, near Eccles, was in the Leverian Museum. Lost. Another found at Overborough. These were the only two in Britain.

pewter. Four pewter dishes, found 1808, 12 feet below in clay and sand, when cutting the Rochdale Canal, near the junction with the Bridgewater Canal. In the British Museum. See engraving Croston's Baines's History of Lancashire, page 10. 17¾ inches in diameter.

iron.

Head of an iron axe, found 1829. See Croston's Baines, plate x., page 13.

Two spear heads. See Croston's Baines, plate ix., page 13.

Stylus of iron, 7½ inches long.

beads.

Two glass beads, in possession of Mr. Farr. See Trans. Lanc. and Chesh. Antiq. Soc., vol. ii. 1884.

Blue fluted beads, at Worsley Hall. See Croston's engravings, xxv., page 15.

bronze, &c.

Two bronze scales, found 1771, near hypocaust. Lost.

Bronze kettle, 5½ inches diameter and 2½ inches deep, with a flat handle, "found in one of our Mancunian mosses, now in the Mancunian Library" (Whitaker, vol. i., p. 299. 1771). Lost.

A female head and a pegasus, found 1789, 6 feet below the ground, between two large horizontal stones, with several flat brass dishes; also urn with silver coin of Trajan (98–117) at Castlefield (see Thompson Watkin's Roman Lancashire, p. 108), engraved. Lost.

Pins. See Croston's Baines, plate xiv., p. 13.

Bullæ in bronze. See Croston's Baines, plates xv. and xvi., p. 14.

Head and neck of animal. See Croston's Baines, plate xxi., p. 14.

Small bronze cross and a hinge. See Croston's Baines, plate xviii., p. 14.

Large number of small flattened rings. See Croston's Baines, plate xxvi., p. 15.

Two rings with a blue fluted bead, and found 1796 (Barritt). Lost. See engraving, vol. v., p. 534, plate vii., Manch. Lit. and Philosoph. Soc., 1802.

Bronze celt, found 1796, now at Peel Park.

Twisted bracelet, now at Worsley Hall.

Bronze boss with a Gorgon's head, now at Worsley Hall, not engraved.

Bronze stand, Croston's Baines, see engraving xvii., p. 14.

Copper ladle, Croston's Baines, see engraving xix., p. 14.

Bronze statuette of Jupiter Stator, found 1839 in Tonman Street, still in possession of the late Dr. John Leigh's family.

Small leaden female bust, 4 inches high, at Worsley Hall. See Croston's engraving, page 11.

A Hercules, "a draped male and two household gods," in possession of Mr. Esdaile, extant.

fibulæ.

Two fibulæ (bronze), one cruciform and one circular, figured in Watkin's Roman Lancashire, page 112.

Silver fibula, found 1849 in Castlefield, figured in Roman Lancashire, page 115 (now in Jos. Mayer's collection, Liverpool).

Roman clasps, found on the east side, 1765–6, according to Whitaker. Lost.

A fibula found in Castlefield	$\scriptstyle {\left\{{\begin{matrix}\ \\\\\ \ \end{matrix}}\right.}$	In possession of Mr. Farr. See Trans. of the Lanc. and Chesh. Antiq. Soc., vol. ii., 1884.
An enamelled buckle

A bronze fibula. See Croston, plate xii.

A circular bronze brooch (see Croston, plate xxiii.), inlaid with several stones, 1 inch diameter.

glass.

A "lachrymatory" of black glass on the south side of the station, found 1765–6. Lost.

A similar one, found 1782. Lost.

sundries.

A lump of sal ammoniac, at the entrance of Gaythorn Row, with a coin of Tetricus, 1788. See vol. v., p. 534, 1802, Manch. Lit. and Phil. Soc.

The various finds of pottery need not be regarded.

General List of Roman Coins found at various time.

From Castlefield (Mayer Collection, Liverpool, found 1849, inside the station, when making the railway line):—

Vespasianus	a.d. 67–79
Domitianus	81-96
Nerva	96–98
Trajanus	98–117
Antoninus Pius	138–161
Claudius Gothicus	268-270
Valentinianus	364-375

From Birchfields, Rusholme (in the brook, found 1896, now in Owens College), a hoard:—

Gallienus	a.d. 253–268
Saloninus	253–259
Valerianus Junior	254–260
Claudius Gothicus	268–270
Aurelianus	270–275
Postumus	258–267
Victorinus	265–268
Tetricus Pater	267–283
Tetricus Filius	267–273

From Higher Broughton (found about 1850 in a hedge near to the Ribchester Road; in possession of Mr. Geo. C. Yates):—

Honorious	a.d. 395–423	$\scriptstyle {\left.{\begin{matrix}\ \\\ \end{matrix}}\right\}\,}$	Third brass.
Gratianus	375–383
[68]Anastasius	491–518		First brass. Latest date.

From Ordsall Lane East, at a point near Islington Street, found 1880, while making the viaduct for the Lancashire and Yorkshire Railway (in possession of Mr. John Rogers, Davyhulme): a Gallienus (third brass), 253–268, and a Constantinus II. (third brass), 337–340.

From Castlefield, in possession of Mr. W. Broughton, Eccles:—

Alexander the Great	b.c. 336–324
Philip the Elder	a.d. 244-249
Valerianus	254–260
Postumus	258–267
Gallienus	253-268
Tetricus Pater	267-283
Florianus	276
Numerianus	283–284
Licinius, senior	307–324
Crispus	317–326
Constantine the Great	306–337
Constantius II.	337–361
Valentinianus I.	364–375

In Peel Park^[69] (presented by the late Mr. Batty):—

Nero	a.d. 54-68
Domitianus	81–96
Trajanus	98–117
Antoninus Pius	138–161
Faustina, senior	105–141
Maximianus	286–305
Constantius II.	337–361
Magnentius	350–353

At Worsley New Hall^[69] (see Croston's Baines, plate xvii.):—

Vespasianus	a.d. 69–79
Vitellius	69
Domitianus	81–96
Nerva	96–98
Trajanus	98–117
Hadrianus	17–138
Antoninus Pius	138–161
Lucius Verus	161–169
Caracalla	211–217
Magnentius	350–353

In possession of the late John Plant^[69] (where now?):—

Trajanus	98–117
Hadrianus	117–38
Faustina
Antoninus Pius	138–161
Marcus Aurelius	161–186
Maximianus	286–305
Gallienus	253–268

Unfortunately, it is never stated if the "Castlefield" finds were inside or outside the station, and at what level.

In recent times, during the last ten to twenty years, large numbers were found during various excavations, and dispersed by the navvies, as no one ever attended these operations.

Castlefield.—Found 1789, by G. Perry (lost), silver Trajanus, 98–117. Found 1723, by Dr. Stukeley (lost), golden Otho, +69.

Ivy Street (inside castrum).—Found recently, in possession of Mr. Fox-Lee, Germanicus, 4–19; Tetricus, father, 267–283.

Tonman Street, Hall of Science (on Ribchester Road).—Silver Trajan, 98–117.

Victoria Bridge (during excavation).—Found 1828, now in Owens College, Constantine the Great, 306–337; Crispus, 317–326; Constantinus II., 337–340.

Blackfriars Bridge. (See Palatine Note-Book, 1883.) Where? Six coins from 249–255. Gallienus or Valerianus (probably).

Quay Street (end of bridge).—Found 1876 (see Palatine Note-Book, 1883), where? Trajanus, 98–117; Hadrianus, 117–138; Antoninus Pius, 138–161; Marcus Aurelius, 161–180; Faustina the younger; Faustina the elder.

Gaythorn Row.—Found 1788, by Barritt (lost), Tetricus, 267–273.

At Crown Inn, Trafford Street, 1840, a hoard of two hundred coins found (all lost), mostly silver and ten of gold; only a brass of Domitianus saved, on the road to Slack.^[70]

New Gasworks, Cambridge Street, 24 feet below the level, in the silt of the ancient channel of the Medlock (lost). Once in possession of Alderman Clay. Gallienus, 253–268; Claudius Gothicus, 268–270; Aurelianus, 270–275; three third brass of Alexandria Imperial Greek.

Roman and Pre-Roman Hunt's Bank, &c.

We have still to consider the area comprised between the Hanging Ditch, Toad Lane (now Todd Street),^[71] and the Irk and Irwell, a site which has been the perennial battlefield, ever since Whitaker's times, of our local antiquaries and historians, for it is here he established his famous summer station of Mancunium, but without being able to give actual proof that the Romans really occupied this area.

In its original appearance, before the Hunt's Bank was deprived of its picturesque, steep, and rocky western face; its quondam, almost peninsular character, as it towered from the river, was more conspicuous. The older views and maps still give us a good idea of its imposing appearance. Where we have now solid Cateaton Street and Hanging Ditch, a deep ravine or gully divided Deansgate from the Cathedral, and its waters, still further increased by the small rills and channels that carved their way down Toad Lane^[71] and Hanging Ditch from the constantly rising ground to the east, which formed Snow Hill and the Shude Hills, poured their whole united volume into the Irwell near Salford or Victoria Bridge.

Unexpected and entirely new light has been thrown on the site in question quite recently during the progress of the excavations which have been executed at various points. I shall take them in proper order, beginning with—

Hanging Bridge.—The original gully at this point, according to Whitaker (see also his "Plan of the Summer Station," taken in 1765), was 8 yards wide, and traced down to 21 feet without reaching the bottom. At its western banks rose the steep knolls of Cateaton Street, at 116 feet above the level of the sea; here Whitaker placed his hanging bridge and erected a small Roman turret, and from that point, crossing through the present western tower of the Cathedral, he makes the Romans follow the precipitous rocky crest that faces the Irwell, to hew their later road to Ribchester, to the margin of the Irk; but, before we reach its rocky brink, he figures another imaginary turret, 80 yards to the south of it, and by dint of another bridge carries the road over to Strangeways, where we are at last on firm ground and where hypothesis ceases. The road runs straight along thence, and it was noticed by him at Francis Reynolds's park in forming the canal at the end of the park (see Green's map).

The Hanging Bridge, in consequence of the extensive demolitions effected in the course of the last few months, has again risen into prominence, and is a fine stone-arched bridge, probably built in the fourteenth century; if it has taken the place of a previous Roman structure at this point, we don't know yet; perhaps traces of a possible sub-structure may be found abutting against the Cathedral side. To the west of Hanging Bridge, close to one of the arches, in Cateaton Street, a shaft was sunk in August, 1900, by the Corporation, and from the decayed vegetable black sediment which was brought up from below, 20 feet deep, I obtained two pieces of Roman glass, the rim canaliculated, of green colour, probably belonging to the bottom of a goblet, of the same class as discovered at Melandra Castle by Mr. R. Hamnett. Messrs. Rogers and Holt examined the vegetable layer, which formed a matted and compressed mass composed of mosses, ferns, seeds, and twigs, &c., of which I give a list:—

	HABITAT.
Hypnum brevirostre	woods, rare in the district.
⁠,,⁠striatum	woods and shady banks, rare.
⁠,,⁠tamariscinum	woods and banks.
⁠,,⁠splendens	woods and grassy banks.
⁠,,⁠triquetrum	woods.
⁠,,⁠purum	shady banks.
Polytrichum gracile	swampy ground.
Pteris acquilina (bracken)	woods and heaths.
Oak (wood)	woods, clayey soils.
Elder (wood)	thickets.
Hazel (nuts)	moist, light soil.
Birch (twigs and bark)	heaths, woods.
Polygonum lapathifolium	seeds, waste land.

shells.

Anodonta cygnea (swan mussel)

water courses, pools, canals.

A Roman brass (second) of Hadrianus (117–138 A.D.), rev. spes aug. leaning on a column, now in the possession of Mr. W. F. Kiernan, was found in the excavations at Hanging Bridge in 1880, which was obtained from the workman.

In 1828, during the excavations at the Manchester side of Salford Bridge, a number of Roman coins, ranging from 306–340, were found.

Proceeding now to Cathedral Street. During the excavations in May, 1899, on the site at which the Corn and Provision Exchange is being erected, a patch of Roman surface was found by me at 4 feet below the ground. The ground was occupied by old brickwork and the electrical main, and consequently greatly disturbed; 12 inches of the Roman soil was left intact, 9 inches in thickness. Only the eastern face was visible, the exact trend could, therefore, not be ascertained. On closer inspection, I found it to be composed of tough and firmly beaten clay, mixed with gravel, and containing bits of charcoal, red sandstone, and, in addition, various fragments of pottery, such as unglazed red-ware, a piece of bright coloured Samian-ware—all evidence of Roman presence; and further, a small bit of a black urn, recognised at once as British or native make. On examination, Mr. Arthur J. Evans, of Oxford, confirmed this. This Roman stratum rested on fine, clear, bedded sand^[72] and layers of river gravel, together 10 feet thick; then 6 feet "bastard rock," underlaid by the red, native rock.

Returning now to Hanging Bridge the men, while excavating on the western side of the Cathedral Churchyard, two years ago, found an old cobble road, 5 feet wide, at a depth of about 10 feet, called by them a "Roman road," which went from the bridge in a slope to the western tower.

The area comprised by the Irk and Hanging Ditch and its continuation into Todd Street is practically divided into two parts, intersected on its east side by Fennel Street—the ancient Vennel,^[73] which formed a narrow road stretching east to west from the northern side of the cathedral. Mr. John Owen has an entry in his MS.^[74] (vol. xxiii., p. 129, September, 1859), saying: "In excavating on the northern side of the cathedral yard at a depth of about 7 feet the workmen thought they had discovered a pavement of rough boulder stones; it was about 9 feet wide, and appeared to point in a line from Fennel Street or rather from Long Millgate to the north door of the church."

The northern part of the area described is occupied by the Chetham College, which slopes away to the rocky banks of the Irk; while the southern part contains the cathedral, which rises up and stood on a high sandy boss, with 10 to 15 feet of running sand and gravel, as proved by the excavations made on the south side two or three years ago. Our esteemed member Mr. John Owen, with great devotion and carefulness, has followed the various consecutive alterations which have taken place between 1859 and 1872 in the cathedral, when part of the old foundations were disturbed and laid open, and his discoveries are of the utmost importance. I shall let him speak himself. He says in vol. xxiii., October 14th, 1863, page 130: "The first stone of the new tower was laid to-day on the north-east angle. The foundations of the old tower were of rubble and clay—the clay (boulder clay) apparently tempered, the rubble (of new red sandstone) and boulder stones of all sizes and shapes were imbedded in the clay up to the level of the ground surface, no mortar whatever being used." He tells me, these foundations occurred at a depth of 7 feet. The construction and nature of these old foundations correspond in every respect with the Roman foundation walls, as seen on the northern wall of the castrum in Castlefield, of which I have given descriptions and sections in the earlier part of my paper. Unfortunately, he omitted to take measurements and sections and to follow the direction and extent the walls took, as we may correctly assume that the Roman foundation did not necessarily coincide with the superstructure of the cathedral tower. But he has more to say. He wrote me on the 23rd May, 1899, on my enquiry: "The old tower was pulled down and the Roman mortar [which he also had found in addition at the same place] was found here. I noticed the mortar being broken up by one of the men. I went to look. It was made of broken brick or tile and lime. [I showed him since, on 8th September, 1900, a piece of Roman mortar of pounded brick and lime from Chester, which he at once recognised as the same material.] I had previously found some of it at the east side of the tower—this came from the west side, which I also showed to Mr. Bowers, the dean, who likewise said it was Roman. The lump, I should think, would be a couple of hundredweights." At Deansgate we find only mortar, made of pounded brick or tile and lime, used for the flooring of the second (super-constructed) hypocaust, erected much later than the original and older hypocaust. The former probably built after 121 a.d. No mortar of brick or tile and lime was found by me then at any part of the castrum, either at the walls or in the remaining remnant of the wall existing in the timber yard, and its employment must have been introduced at a later stage. It is clear that here at the north tower we have to deal with an older Roman^[75] foundation, pointing to a substantial structure which had a cement flooring.

Continuing, Mr. John Owen records:—

Vol xxiii., September 30th, 1867, p. 131: "The workmen have been engaged for several days making an excavation under the east side of the chapter house. An old rubble wall is partially laid bare and going down to the gravel, whereas the foundations of the chapter house and the adjoining wall of the church do not go down so low by about 4 feet. It extends under the chapter house from east to west, and projects 4 feet southwards beyond the line of the church foundation. It is formed of large rubble stones and filled with (gravel) sand and small (boulder) stones."

January, 1872, p. 132: "Some excavations have been going on between the piers or stair turrets for laying a foundation to support the new organ; we found a rubble foundation extending about 6 feet from the base of the south turret on its northern side, directly under the chancel arch. Immediately below the plinth of the turret, in advance, abutting against it, was the remains of another plinth, seeming to have belonged to an earlier pier of the chancel, arch, although it may not have been so, as it would not in its position account for the extension of the rubble foundation. On excavating on the south side of the northern turret we found the rubble only extending at 3 to 4 inches beyond the plinth. This was also extended about 3 feet southward to form a solid foundation for the old rood screen and organ."

February 17th, 1866: "In the north aisle, close to the timber partition of the choir, and in the first bay of the staircase turret, we found the chamfered plinth of the ancient chancel, running east and west. It consisted of three courses of good ashlar masonry, the foundation below consisted entirely of rubble, containing no remains of old work."

They are now excavating a tunnel right under the west end of the choir, and have removed a portion of the rubble.

I may also quote a remark of the Rev. H. A. Hudson:^[76] "Mr. Crowther [the architect who was in charge of the alterations of the Cathedral] admits the evidence of a Saxon church constructed with stone may 'appear too scanty to be conclusive,' and he remarks further on upon the singular absence of any Norman remains. The explanation of this singular absence of any remains, says Mr. Hudson, seems to be "that no such Saxon or Norman stone church ever existed, and that the church on this site, if any, previous to the Early English period, was not of stone, but of wood, according to Hollinworth's tradition, and the explanation becomes the more probable when it is remembered that great numbers of fragments of all sorts, from the Early English (1204–1265) and Decorated English (1265–1360) were found at various times," and no traces whatever before that, which practically means that there is no evidence of any stone structure in existence before the thirteenth century, neither of Saxon or Norman times; nor have we any record of rubble walls filled with gravel sand and boulder stones for foundations referable to these periods either here or elsewhere.

I think we have to correlate these old substructural rubble walls between the chapter house and the north turret (measuring on the west side 75 feet long and on the south side 56 feet long) with the foundation walls at the north side of the western tower (traced for 25 feet), and their composition makes them distinctly Roman; they are real repetitions of those met with at Castlefield, Deansgate. I have prepared a plan to show the situation and extent of these remnants.

We come now to the excavations made between the 26th and 31st July, 1900, at Chetham College, on the plot between the north side of the school and the south side of the library. The distance from the shafts which were made measures 154 feet to the brink of the Irk at Hunt's Bank Bridge. The total length of the excavated area measures from east to west 27½ feet, and terminates against the old boundary wall which divides the college from the Palatine Hotel. I have taken a careful section, and, beginning from the level, we have in succession and in undisturbed condition:

(1)	30	inches	recent building rubbish;
(2)	12	,,	yellowish soil;
(3)	11	,,	red sand;
(4)	8	,,	clayey sand, with streaks of stringy patches of clay, the latter apparently accumulated and formed while the ground was actually traversed, and at the western corner it shows a layer of blistered charcoal and burnt bones (2 to 3 inches long);
(5)	23	,,	clayey red sand, in which we have (at a depth of 63 inches) a large compact block of Roman mortar, 12 inches square by 6 inches thick, now in the possession of the Chetham Hospital, for permanent preservation, found against the eastern corner; another large piece was found in the middle; further, a fragment of Roman roof tile. Of botanical specimens I met in it: Ulex europeaus, the furze, a large stem; Rubus fruticosus, the common bramble, a seed; Sambucus nigra, the elder, many seeds;
			A fang of the wild boar. Below this we come upon:
(6)	15	inches	the Roman boulder road, flagged with white-grey sandstone chippings, 6 to 7 inches long and 1 to 2 inches thick, the whole depth of 15 inches constructed and consisting of boulder stones, smaller towards the top (4 to 5 inches long, 3 to 4 inches thick) and very large at the bottom (7 to 8 inches long and 6 inches thick). It stretches for 15 feet from west to east, and does not continue on the other side of the central shaft, which was run to a further depth of about 6 feet, consequently it limits its real known width (except it should extend on the west side at the Palatine Hotel) to 15 feet, as it suddenly stops short on the east side of the central shaft. Below we have:
(7)	30	,,	sand, on which it immediately rests, and, in addition,
	12	,,	consisting of a black sediment, highly decomposed, speckled with vivianite. From it I obtained: Juniperus communis, the juniper, (a berry); habitat, on open sandy plains and hills; Polygonum lapathifolium, the persicaria, (the black shining seeds); waste land; Corylus avillana, hazelnut (fragments of the shells); woods, slopes, river banks; and a fragmentary specimen of the minute shell Helix pulchella.
			Bits of burnt bones and minute pieces of reddish pottery, with grains of sand in the matrix, pieces of leather, and a femural bone of a bird.

We see from the other section, taken at right angles, that the boulder road rested on an ancient, much earlier excavated fosse, which was cut at a very steep angle into the native rock, showing towards its southern margin the black sediment which was quietly accumulating at the bottom (the black mass contains a large admixture of fine sand grains), 12 inches thick, while the 30 inch sand, which is sandwiched between the base of the boulder road and the black deposit, points to a later stage, when it was filled up for the posterior construction of the boulder road which traverses it at this point. Whitaker mentions that in 1765, at the grammar school, a channel was found through the rock in the cellar, 3 yards wide and 2 yards deep, "filled with an unctuous mass and then sand." This channel, or fosse, which he calls "the Prætorium foss," in his "Plan of the Summer Station," is, of course, as we can see, the same fosse rediscovered now at its western side. Let me recapitulate now.

At Hanging Bridge (spanning the original gully or clough) we have at a depth of 20 feet Roman glass and a Roman brass (date 137–138 a.d.); a little further down, towards the river, or river ferry, at the margin of the Irwell, we have Roman coins (306–340 a.d.).

At Cathedral Street a patch of Roman surface with contemporaneous British and Roman pottery.

At the Cathedral, under the foundation of the western tower, old Roman clay and boulder foundations, at the depth of 6 feet, and a cement flooring; on the eastern part of the cathedral we find rectangular remains of an old substructure of rubble walls, 4 feet below, and projecting 4 feet southward beyond the church foundations.

At Chetham College we observe, 5 feet below the undisturbed ground large blocks of undoubted Roman mortar, indicating the existence of compact foundation walls, similar to those of the north wall of the castrum in Deansgate; fragment of a roof tile, and below, at a depth of 7 feet, the existence of a Roman boulder-paved road, which again rests on a much older artificially, rock-cut fosse. The Roman road, in its turn, is slowly sanding up, and lying apparently waste and disused; in the sand we have the gorse bush, which could then take root and grow upon it; we have the elder and bramble, which could drop its berries, and allow its sun-dried, puckered seeds to insinuate themselves in the accumulating layers of sand drift. We notice again above the roots and seeds pieces of Roman mortar that had been flung there from a destroyed building. Twenty-three inches above the deposit of sand we find blistered charcoal and burnt bones, indicating perhaps a recurrent incursion of Hunt's Bank.

We see what a promising field is awaiting future research in these quarters, and how within the last few years evidence has been steadily accumulating to prove the territorial occupation by the Romans of Hunt's Bank. I think we are now on sure ground, and the spade, which is now busy at all points round this ancient area, may be the means to throw further light upon a number of other points which have to be fully worked out yet. This refers specially to the direction of the Ribchester Road. On my "Map of Roman Manchester," I provisionally made this road to slant from the deanery, where it is last tracked to, to a point at Cathedral Street, as I took the Roman patch discovered here as an indication of the main road, which, splitting up, sent one branch across Chetham College ground to Hunt's Bank, while the other—the Blackstone Edge road—fell, cutting Fennel Street, into Old Millgate. But since the discovery at Hanging Bridge of Roman glass and a coin, brought up from the bottom of the gully, it is possible that the Ribchester Road passed, as Whitaker believed, viâ Hanging Bridge, through the west side of the Cathedral, straight along to Hunt's Bank Bridge. I have already shown that at the Chetham Library we find the Roman road, 15 feet wide, but the excavation here made was too small and incomplete for our purposes, to ascertain and settle the exact trend it took. Fresh excavations are in sight, however, for the erection of a lodge at the entrance of the college, and perhaps we may be lucky enough to hit upon traces of the road here, if it passed on a line as suggested by Whitaker. Meanwhile, this must be left an open question. Whether the small-cobble road, which was found on the north side of Hanging Bridge, leading to the western tower of the cathedral, represents the road is doubtful, it being only 5 feet wide.

We have to speak now of the road to Blackstone Edge, a secondary road leading to the borders of Yorkshire, which has a certain interest and importance, and of which but little has actually been known. Thompson Watkin says: "The most singular fact connected with this road is that no one has ever seen or heard of the portion of it between Manchester and Blackstone Edge from Whitaker's time to the present day." Our information has increased since that statement was made.

As mentioned before, we do not know yet at what particular point in the Hunt's Bank area the junction really was effected. On the north door of the cathedral we have Mr. John Owen saying that a pavement of rough boulder stones about 9 feet wide was found, at a depth of 7 feet, pointing in a line from Fennel Street, or rather from Long Millgate, of which the ancient Vennel, "a narrow road stretching east-west from the northern side of the cathedral" (Hibbert-Ware), seems to be a continuation. Casting a glance at the "Map of Hunt's Bank," which I have prepared, its particular position seems to urge us almost to think that the Blackstone Edge road had its rise here, and thence passed into Long Millgate. Let us hear what Stukeley^[77] says: "I find the Roman road went across the churchyard originally and so by the common street (Long Millgate and Smithy Lane) to the bridge over the Irk, called Scotland Bridge, then it ascends the hill and proceeds with its original direction north-east to Rochdale."

It probably passed along Red Bank, left the Peel on the left side, and proceeded along the Brows to Smedley, on its way to Blackley. The Rev. John Watson^[78] tells us: "From the end of Long Millgate at Manchester this road went through Blackley, near the foot road to Middleton, leaving Alkrington House to the west; crossing over the meadows to Middleton Hall, and over the Barrow fields, as the foot road goes over Hopwood Demesne to Trub Smithy, and leaving Castleton Moor and Castleton Hall to the left, over the enclosures to the ground lying a little to the right of the guide-post, near Rochdale Town End; then crosses the enclosures over a ridge of land to Belfield Lane, and leaving Newbold Hall about two fields breadth to the east up the valley to Stubley Demesne, leaving Littleborough on the east, and near Pikehouse up the Light-owlers, on the north side of the hollow up Blackstone Edge."

We notice that he leaves Littleborough on the east and makes the road pass near Pikehouse; while Thompson Watkin, to make it chime in with his imaginary "trough" road, lets it proceed viâ Lydgate (half a mile south of Pikehouse) up to Captain Rosworm's entrenchment at Blackstone Edge.

We see that in reality, according to Watson, the road took a quite different direction from Littleborough; when we reach Blackstone Edge we have close to it on the left side Cold Laughton and on the right Black Castle and Black Castle Clough, on the highroad to Ripponden. Neither Watson,^[79] Whitaker, or writers previous to Thompson Watkin, or the local farmers and shepherds have ever so much as mentioned it; nor have we a road so constructed—fancied by him to be Roman—either in Britain or abroad. Why, therefore, of all places—and the Romans had elsewhere to deal with similar gradients and ground—Blackstone Edge alone should have been singled out for the construction of such a road has to be explained yet by his followers; meanwhile, it is safer to follow Watson and discard Watkin's altogether questionable road.

As to the width of the Roman road viâ Blackstone Edge, in the "New Map of the County of York, laid down by an actual survey, published for and sold by Rob. Sayer and Tho. Bowles, mapseller, London, 1728," where it is very distinctly indicated. It says: "The Roman way extends from Manchester in Lancashire unto Aldborough near Boroughbridge, is all paved with stone, and near 8 yards broad."

Passing now from Fennel Street and Hyde's Cross to Withy Grove, I was informed by one of the gangers, Jim Kennedy, that in excavating for the foundations for the Evening Chronicle buildings (between Mark Lane and Huntsman's Court) a road, 12 feet long and 2 feet 6 inches deep, made of boulder stones (medium size) and "like cement-mixed, running apparently in a line with the old Corn Exchange, was discovered, passing right through the bottom of the building" was found at a great depth. A little higher up Withy Grove on the same side (the north side) the Manchester Magazine, August 8th, 1788, mentions a large plot of land, called Cold Ar-e Meadow,^[80] a word which it has been suggested corresponds with the Latin Arx. However that may be, I wish to point out that these two sites are just a little beyond Hanging Ditch, towards the Shude Hills, the weakest point of the defence of the Hunt's Bank promontory to the east.

Mancociunium, Manucium, Mamucium, &c.

I think we are now better fitted out to define the original locality occupied by the local Brythons prior to, and at the time of the first arrival of the Roman invaders here. One was naturally inclined to look to Castlefield as their original seat, the more so as here alone we have the visible demonstration of the castrum walls; but still its topographical position offered a serious obstacle, often dwelled upon by enquirers, to carry full conviction for accepting such a view. In studying the contours of Castlefield we find that the station itself is partly built at a level of about 105 feet towards the south-west on a little rocky ridge, within the elbow of two low-banked confluent rivers, a site, in fact, well adapted and generally preferred by the Romans for the erection of their camps. No trace whatever of any anterior British occupation has come to light in the past, or during the latter extensive excavations at and around Castlefield in 1897–1900. The only discoveries consist in some sporadic finds of a flint scraper, a barbed arrow-head on the north side of the castrum, a stone hammer of gritstone at Throstle Nest (in Whitaker's time), and a large flint core at the river level in Ordsal Lane, all of which more directly point to the presence here of some earlier aboriginal people, but indications of British ditches or stockades or pottery are wanting.

On its western and northern side, as I have shown before, the ground was of a swampy and marshy nature and little suited to British defence, and Hulme, at its south, was a low and mossy holm, exposed to frequent river floods. When we come, however, to the rocky cliff of Hunt's Bank, towards which the levels continually increase, we are on different ground, and at Cateaton Street the level rises to 118 feet, a site which offers certain natural advantages of primitive defence.

We know that the Britons were in the habit of preferably selecting hills and more inaccessible eminences for their duns; round these they drew their ditches and stockades. We have seen that the northern rocky ledge, now covered by Chetham College, is traversed on the south side by a long-drawn rock-cut fosse, which swept in a curve to the Grammar School, and further strengthened by the natural double-defence of Hanging Ditch gully and its artificial prolongation terminating in the latter Toad Lane or Todd Street. This additional cordon formed a complete outer circle.

This fosse was filled up with sand and levelled, at a later period, to allow the super-construction of the posterior Roman boulder road. It is not likely that the Romans originally made this fosse to pitch their improvised camp on this comparatively narrow strip of ground enclosed between the bank of the Irk and the termination of the Chetham College Library; it is more feasible that the Britons, when encountered by the Romans, were found in fortified possession here.

The persistence of trying to fix upon Castlefield as the British nucleus has added to the difficulty of explaining the various names by which the station, in their Latinised forms, is known, and consequently the etymologies proposed have so far signally failed to answer the physical conditions involved. Mr. Henry Bradley has quite recently viewed the question of its etymology from a purely grammatical and philological point, which is only one of many other sides to look at it. He is apparently unacquainted with the topographical and strategical features of the place he discusses, and, considering that he works on incomplete data and knowledge to help him to form a definite conclusion that would relieve us and add light, his destructive attempt lands us in a position worse than ever. For his contentions I refer readers to his paper in the English Historical Review (July, 1900), from which it is evident that his premises are built upon the old assumption which postulates the Deansgate castrum as the site. He distinctly says the name hardly fitted the low-lying settlement at Castlefield. As to the other points he raises, I completely differ and join issue with him. I uphold that the rocky mass of Hunt's Bank gave rise to the name and that it furnishes all the elements for an unforced explanation. I may, therefore, be permitted to go a little more minutely into the matter for establishing my position. To understand the physiognomy of the town we have to transport ourselves back a few centuries, say, to manorial times. It has been pourtrayed to us, then, by various distinguished visitors. It was soon perambulated at that period and gravitated towards Hunt's Bank, situated between the extended angles of the Irk and the Irwell, and reaching on the south a little to St. Mary's Gate and on the northeast to Scotland Bridge. It picturesquely rose on a red, rocky promontory, which struck all alike as one of its main features. Hear

Leland (1538): The town of Manchester stands on a hard rock of stone.

Camden (1582): Where the Irk runs into the Irwell, rising in a kind of reddish stone, flourishes that ancient town being now called Manchester.

Cecilia Fiennes (1697): There is a very large church, all stone, and standeth high, so that walking round the churchyard you see the whole town.

William Stukeley (1725): Manchester, placed between two rivers, having rocky and precipitous banks, with a good prospect.

Daniel Defoe (1753): The town standeth chiefly on a rock, &c.

And, I think, we have an early indication of this characteristic point retained in one of the Brythonic names given to the locality.

The itinerary of Antoninus, which gives us the chief military roads, was to the Romans what Ogilby and Morgan (1685), John Owen (1764), and the more recent Paterson and Carey have been to modern England. There is some considerable divergence of opinion as to the exact period the itinerary has to be referred to, which differs as wide as from Antoninus Pius (138–161) to Caracalla (218–222); although Thompson Watkin argues (see Roman Lancashire, p. 79) it must, from internal evidence, be attributed to the former. The British stronghold and its name was, of course, in existence long before its compilation and probably even before the Christian era. The spelling of the names—the work of many successive scribes, Romans and maybe even non-Romans, and men more or less ignorant of the genius and phonology of the British tongue, who obtained them from soldiers, traders, and travellers, not to mention errors and omissions of copying—must even by that time have undergone a certain fluctuation or vitiation. An attempt to treat this analysis on strict and rigid philological lines seems to be futile. To ask, or hope to find in such place-names the retention of all the inflectional subtleties from the pen, or in the mouth of Roman writers, not natives, is the same as to expect modern travellers in Central Africa or Asia to preserve or bring home native place-names in their true and correct form. We need only take modern parallels and compare the spellings of the old English road guides when we come to Welsh place-names and the northern counties! Take also such purely local instances as Kersall (twenty-four spellings), Crumpsall (eleven), Ordsall (seven), Ancoats (eight), Ardwick (seven), even within the space of one or two centuries. Amongst which we find, almost to irrecognition, forms such as these:—

Kyrsaw, Kirksagh, Kerstaw, Hereshall, Karsey;
Curmisale, Cormshale, Cromshall;
Oardsall, Urdsal, Hordeshale;
Anekotes, Annecote, Antecotes;
Aderwyk, Erdwyke, Herwic.

To return to our subject. Of the many variations, found in the twenty-one different MSS., reduced to their ultimate kernel, we seem to obtain the following chief groups:—

(1) Ma(n)cocun-ium (Mam-cun-ium, contraction);

(2) Coacciun-ium (Coac(o)cun-ium);

(3) Mam-uc-ium, Mam-ut-ium, and its contractions;

(4) Man-uc-ium, &c.

Taking the root man first, compare the Welsh place-names: Maen y Campiau, Maen sigl; Maen Twrog, Maen Arthur; Tre vaen, Y maen du, Maen melyn, Pen maen mawr and bach. Coaccium, coac(o)cun, cocun, c.p. coch, pl. cochion=red, as in the Welsh place-names: Moel goch, craig coch, clawdd coch, bryn coch, &c., where the Celtic harsh gutteral ch is softened into the Latinised form with c (k) and ti, as in Mamucium, Manutium (pronounced probably as in much).

The original Brythonic form seems to have been Meini cochion=Redstones, just as we have the Welsh place-name, meini llwydion, meini birion, and the cerig cochion, in Bangor. Hollingworth's (+ 1656) spelling of Main cester appears to point to an old plural spelling (maini, main). We meet the name again in the local name of Red Bank and Red Bank Head, both of which form the continuation of the high, steep, red, rocky banks of the Irk, on the north-east part of Hunt's Bank. We have, in addition, the Great Redstone^[81] (1624), and the Great and Little Redstones, "opposite an outcrop of the New Red sandstone towards Kersall Cells and Mill Riding going up the river Irwell."^[82]

Man-uc-ium appears to be composed, in its second root, of the word awch (also contained in och-r)=edge; compare also Welsh uchel=high, as in Uwch mynydd; in Bardsey, Uwch Conwy; and the Celtic place-names, Uxelo-dunum; and, on the continent, Ocello-durum, Ocellum Durii (Fermosella, Spain), Ocellum (=Exiles, France).

In its Celtic garb it would probably be Maen-awch=Stone edge, as we have it in Stanedge, Blackstone Edge, and the German Stein-ach and Hohen-eck. Now, I consider that both forms. Red stones or rocks, and Stone edge describe exactly the physical and characteristic features of Hunt's Bank in general and particular. The first refers to the whole face of the split and cleft red promontory, and the latter more minutely describes the stone ledge itself on which the British dun must have been founded.

Roman campaigning against native tribes did not materially differ in some aspects with practices observed in modern times. To make matters homely, we have only to follow warfare in South Africa and Matabeleland to form a mental picture of what was done by the legions under Cerealis, or Agricola against the Brigantian hill and mountain clans. The natives here, after offering a certain resistance in their stronghold and from their lurking places at and about Hunt's Bank, were probably driven off, and the place was at once occupied by the conquerors for fixing their own camp. To what degree they then further strengthened the place we have no means yet to ascertain, in want of extensive excavation at the College or Cathedral. They, no doubt, remained there until, as a more suitable situation, after studying the terrain, they proceeded to the erection of the large and more expansive station at Castlefield, more suited to their purposes. Later on, the natives were probably allowed to return, while the Roman military sphere was transferred to the new site at the Medlock. In consequence of the change, Deansgate must have become in course of time the centre of Roman life, with a specific population of soldiers, officials, and traders, while the tribe people would cluster more intimately around their ancient quarters at the banks of the Irk. Thus in course of time we probably had the Brito-Roman upper (or hill) town, and the Roman lower station at Castlefield. Such parallels are familiar to us from South Africa, where we have the old native and the British quarters.

The distance between the Hunt's Bank settlement and Castlefield measures about 1,620 yards, or nearly one mile. To distinguish the two a slight change in the place-names seems to have been made at some time. This would explain the otherwise quite enigmatical and tenacious dual form of Man and Mam. We have the mam contained in such Welsh words as mamdref=chief town, mamddinas=the metropolis, mam-eglwys=the mother church, and the place-names Mam Tor and Little Mam Tor, Moel Fammau, Mam Head, Mamhilad, Mamhole, and perhaps in Mamaceæ (France), Mamertium (in Calabria).

Mam-uc-ium would, therefore, mean the mother stone edge, the original site of the Britons and the older Roman site; while the newer and their principal castrum at Castlefield, in contradistinction, went by the name Man-uc-ium. This seems to be made plausible when we recollect that the Mercians, in rebuilding the place destroyed by the Danes, speak of Mamcestre, which can only have been the site of Hunt's Bank, which in succession was again occupied by the Norman barons, who had their Barons' Hall on Hunt's Hull or Hill. We may argue retrospectively to an occupation of the same site in early Anglo-Saxon times.

The old Roman station in Castlefield at their appearance was probably, with its suburbs, a mass of ruins and in decay. We know that when the Romans in the second and third century erected their northern suburb, the ditches were filled up and levelled, and the road to Buxton, and a street close to, were built over it; nor was the ground particularly suited for their method of defence; no remains or traces of their existence have been found either at Castlefield or Alport, nor have we any local traditions or folk-lore, except the Giant Tarquin, who is more connected with Arthurian hero legend to make it probable that they had their seat here. To them, more likely, Hunt's Bank, with its natural and easy defences, afforded the same vantage ground. If my argument of the contemporary dual existence of the upper and lower town can be strengthened at all, let me point out that the Castlefield Station was traditionally known only as Man Castle in the mouth of the people, and so called by such early visitors as Leland, Camden, and others.

Mamecestre—the seat of the Norman lords—retained its name into post-Norman times, and it is only towards the end of the sixteenth century that the scribe of the Court Leet Records drops into the spelling of Man-cestre.

The subject, of course, lends itself to controversy, but I hope I have succeeded, from evidence which is circumstantial and at least carries the weight of some grains of probability to lift it in a degree from the utter submersion Mr. H. Bradley has tried, in his way, to lead us into, and which requires tools for handling, not necessarily entirely of the philological kind; and for the peace of our souls let us hope a local Roman milestone may be found soon by some lucky discoverer to cut all doubts and polemics.

Hunt's Bank received renewed attention when the Romans began to consider the construction of a new and more direct road to reach Ribchester. Thompson Watkin holds—and I think truly so—that the new road was not made until a considerably later period than that from Manchester viâ Wigan and Walton. It is one of the roads not mentioned in Antonine's Itinerary, and may fall within the middle or end of the second century. When that new road was laid out, the stone-walled castrum at Ribchester, a complement of the station at Walton, was already built, and thus connected it direct, viâ Ribchester, with the northern wall. Agricola had no share in the construction of this road.

Not much of this road is known in our locality. It has been visible at Spen Moor, near Ainsworth, composed of hard gravel, 7 to 8 yards broad and ½ a yard in thickness, and beyond that, near Harwood, as a pavement of large stones, and portions at Cunliffe Moss, 12 yards wide (see Watkin, pp. 54 and 55). On this road the latest Roman coins have been found, 491–518, near Higher Broughton.

We have seen that at Chetham College a large mass of mortar, consisting of large pieces of rubble stone and gravel and lime, has been found, which in all probability comes from the walls of a massive tower which must have been in existence close to the junction of the Irwell and Irk to guard the river passage at Hunt's Bank Bridge, to further protect the Castlefield station and hold the newly-made road.

The circular area of Hunt's Bank is curiously divided into two parts and bisected in the centre (see map of Hunt's Bank). The series of Roman buildings on the southern half, forming the site of the Cathedral, seems to point to a closer occupation of this more elevated area.

I think I have succeeded in showing that Hunt's Bank was the Roman starting-point; as a corollary it follows also that from here probably issued forth the road on which they first advanced into Yorkshire, namely, the Blackstone Edge route, which, as the map^[83] shows, is a less direct and more circuitous though easier and therefore an older track road than the one viâ Castleshaw. The latter road is a more elaborate road (see Whitaker's description) leading over the difficult and extensive mosses of Newton Heath, Failsworth, and Hollinwood that proceeded from the eastern gate at Castlefield station.

The two roads proceeded between Halifax and Huddersfield, and describe the district where the inscriptional stones alluding to the gods and goddesses of the Brigantes have been found, as at Longwood, Greetland, Honley, Adel.

To judge from the direction of the various Roman objects found between Littleborough and Castleshaw, as at Tunshill, Slences, these two lines seem to have interocmmunicated by a short cross road.

I also believe that before the construction of that part of the Ribchester road, which issues from the north wall at Castlefield to Hunt's Bank, laid as it was through the marshy and swampy ground, an anterior and older track road was used, which led up to Hunt's Bank, more on a line with the present Deansgate, which proceeds on higher levels.

It is also very probable that the Romans spread across to old Salford^[84] (at the ancient ford on the Manchester side Roman coins, 306–340 a.d., were found), and that a vicinal road went across Broughton ferry (see Green's map) to Broughton,^[85] where at Grecian Street and Albert Park a small Roman road was found by me in 1886,^[86] pointing to the ferry, and close by a piece of Roman fresco, indicative of a Roman villa, a flint flake, and the place-name of Barrow Cliff.^[87]

British Occupation.—In addition to what has been said already about their presence on Hunt's Bank, we have also from Red Bank a Neo-Celtic urn, figured and described in our vol. v., p. 295, and another one from Broughton Old Hall (now in Peel Park), and in connection with this subject I must also shortly allude to the various caves laid open during recent years on the banks of the Irwell and the Irk.

Caves.—In excavating at the river edge at the Parsonage, in May, 1899, four caves were met with, cut into the red rock, which exhibited distinct marks of the pickaxe. They were completely filled up with river ooze and mud, and measured 4 feet long, 3 feet 3 inches across, and 5 to 6 feet high (a sketch has been taken by Mr. Phelps for me). Another large cave was revealed at Old Millgate in July, 1899, facing the premises of Messrs. A. and C. Horsley, and at a place until then occupied by Jacob Libstein. Its apparent height was 12 feet, and it was about 8 feet wide and of great length. It was found during the extensive excavation for the extension of the Lancashire and Yorkshire Railway, when pulling down the front row of houses which then lined the elevated banks of the Irk. The original slopes were visible for a a week or two, and at the margin and foot of the south bank this big cave stood out. The opposite north bank, at the terminus, could also be noticed, and it showed what a fine, broad, and lovely stream this confluent must have been in its ancient and unobstructed condition, with its clear channel and tree-covered, rocky slopes. Unfortunately I had no opportunity of getting a photograph taken; nor did I venture to explore the inside.

Still higher up, at Red Bank, in Hargreaves Street, at the place now occupied by Messrs. W. Clemson & Co., a few years ago, two large separate artificial dome-shaped caves or chambers (showing the work of the pickaxe) were discovered, of which no record exists in our times. They are again cut into the red rock, and their height is 9 to 10 feet. On Green's map (1787–1794) the locality appears as quite unoccupied virgin ground, without buildings. They were professionally examined by an acquaintance of mine, who, as he tells me, explored them for a distance of thirty yards, when he risked no further. They are now used for the disposal of sewerage, and it is impossible to penetrate them thus for excavating the old floor. There is every probability that these caves have been occupied in British, and Pre- and Post-British times as an abode, shelter, and refuge, and it is only a pity that the chance for exploitation is irrevocably gone. They are sure to have yielded inestimable finds of pottery and other remains of prehistoric times.

The antiquity of the occupation of Hunt's Bank, and the banks of the upper courses of the Irk is further accentuated by the discovery of a stone hammer in 1870,^[88] at the depth of 20 feet (apparently from the bottom of the old ditch at the junction of Todd and Corporation Streets, and now in Owens College), of a flint flake by me in the Cathedral churchyard, by a perforated stone hammer near Turkey Lane, Harpurhey, of clay-iron, and another one from Harpurhey (see our vol. v., 1887); and there is every hope, now that I have pointed out these matters, that on search our catalogue of finds will increase in view of so many building operations on our old sites.^[89]

Before I close my account of Hunt's Bank—partly to refresh our minds from the multitude of details I have been obliged to force on your attention—I may wind up with a picture of the area, as it probably appeared to the eye in British and Roman times. From the botanical specimens which were brought up from the depth at Chetham College, and the bottom of Hanging Ditch, and viewing the geological features revealed during the excavations at the Cathedral, forgetting for a few minutes flagged streets, pavements, and the bustle of railways and trams, which, with College and Cathedral, I have to waive for a moment from the neighbourhood, we see ancient Hunt's Hill rising up from the precipitous red cliffs above the sparkling waters of the Irwell and the Irk in its primeval freshness, overspread with a thick stratum of gravel, and grey or bleached fine sand, on which grows the juniper, the heath, the golden shimmery gorse bush; while the brows and slopes of the gorge of Hanging Ditch, and the rivers are overhung with the graceful spray of the silvern birch, the hazel, the elder, and the soft tendrils and arched shoots of the bramble, intermixed with thick clusters of bracken; the mosses carpet the ground of shady banks, while the Polytrichum warns us of some swampy patches. The oak, fond of the clayey undersoil, covers the higher ground of the Shude Hills, and the water mussel burrows in the muddy bottom of Hanging Ditch, while eel^[90] and fish swarm up the oozy pools of the Irk. Against the blue sky, to the south-west, we descry the crenulated, white-glimmering walls of the massive castrum at the Medlock, and bits of the Roman roads glinting here and there through the fretwork of the soft foliage.

---

Conclusion.

The Society has gone to great expense and sacrifice for illustrating the paper with copious views and sections, but some of the larger detailed plans, not to overwhelm their efforts, had to be omitted from reproduction: these and a large mass of additional sections and photographs have therefore been handed over to the Reference Library, as a suitable place for reference to them, which may be desired by the members. The large collection of my Roman finds is awaiting a suitable home yet, or some public-spirited Mæcenas to provide for it, and I renew my appeal for an archæological museum for treasuring up all local finds and monuments of interest to Manchester and the district, to illustrate its ancient and mediæval history from all points and sides, and to foster and invigorate the historical instincts of the town. Until this is done, these relics will share the fate of the past^[91] and vanish—entailing a clear loss to posterity and historical continuity.

Much remains to be done yet for the reconstruction, on modern lines, of Roman Lancashire, and we require excavations at Walton and many other places yet, before filling in the blanks, to form correct notions of Roman life and administration in these parts. The position of Coccium, whether Wigan or Blackrod, has also to be settled yet in a more satisfactory manner, and, above all, every find should minutely be registered and its position defined and described. The recent work at Melandra, Ribchester, and Wilderspool has added a great amount to our knowledge since Thompson Watkin published his great work. He has laid the groundwork, and with better resources and the pecuniary assistance now rendered more freely for practical research in excavating any old important sites by the Society, and let us trust in future also by the county councils, it is to be hoped that in course of time we shall be able to gain a better knowledge and comprehension of the history of Roman times in the Palatinate.

---

APPENDICES.

---

ACCOUNT OF THE DISCOVERIES OF THE ROMAN BATHS, FROM THE MANCHESTER MERCURY AND WHITAKER'S MANCHESTER.

As the labourers at the Duke of Bridgewater's were removing some earth in the Castlefield, to make a new coal bank on the margin of the Medlock, they discovered a large wrought stone of about 2 feet square, plainly indicating to be the pediment of a pillar. Before the front, to the east of it, lay:

A floor of about 7 yards long, 4 yards broad, and 6 inches thick, composed of a strong body of lime and bruised bricks (broken tiles and Roman mortar), with a covering of flags. On the side next to the river, and at the farther end from the pedestal, were found fragments of urns, three skulls (one plainly a human one), human and animal bones lying under a decayed arch of Roman bricks, composed of finer wrought matter than stock bricks, exceedingly heavy and bright coloured. Next the pedestal were found two brass balances and a large knife.^[92] (Whitaker's Building 1.)

That large projection of the bank of the Medlock, which commences near the south-eastern and south-western points of the station, lying within the two angles of the camp, was naturally the site of all the offices. In 1771 was here found, a little to the west of the south-eastern angle and directly opposite to the small bridge on the other side of the river, when levelling the bank for a wharf, and proceeding to the east:

A large stone, like the pedestal of a pillar, but all plain on the surface, 2 feet 9 inches across at the base, gradually decreasing upwards by four stages, 8 inches, 3½ inches, 1¾ inches, and 1½ inches in height, and 2 feet 3 inches, 2 feet, and 1 foot 9 inches long, placed on a flooring 7 inches to 8 inches thick, made with pieces of soft red rock and bedded in clay, and nearly 25 yards distant from the present edge of the Medlock.

Eight feet immediately to the east of this was a:

I. Building, equally with the stone, about 2 feet below the surface of the ground, and floored with a Roman cement of mortar and pounded brick. The floor, 9 inches thick, rested on a body of marl about as many in depth. The whole building was about 20 feet long and 10 feet broad. Nine feet to the east of this was:

II. Another flooring, 2 feet or 3 feet lower in the ground, and a cake of the same cement and thickness. It lay upon loose earth, but was covered with flags. The whole about 10 feet broad and 30 feet long. The exterior wall of both buildings was discovered on the northern side, running with the river, the former about 2 feet 3 inches in thickness, the latter about 4 feet. This rose about 3 feet high, formed of regularly dressed stones, the upper shallow and the lower deep, and having extended nearly in a right line about 30 feet, it then turned in a fair angle and pointed towards the river. In the former building was dug up only one flooring, in the latter three.

Below the pavement described, and in the loose earth on which it lay, were found as the pillars of it large blocks of a millstone grit and square tubes of strong tile. And the first flooring lay on all these, the interval between the tubes and blocks was entirely filled up with earth. The former were about 16 inches high and 5 inches in diameter, filled up with mortar, once fluid. Three of these were found together, standing erect, and two of them so formed with projections as to make a third by their union.

And these and the earth all rested upon a second flooring, another cake of the same cement, near 2 feet in thickness, and lying upon a second bed of rubbish about 3 feet in depth. In the body of this earth, which was covered with the second flooring, were discovered three or four regular pillars of flag and tile, all unbroken and entire. The first was placed about 6 feet to the south of the northern wall, and the second about 17 inches to the south of that. Six feet eastward was another, and about 17 inches north of this were some remains of a fourth. They were composed of a square flag, then two layers of tile, each tile about 2 inches thick and 8 inches square, and afterwards of flag and tile in four layers alternately, all laid in mortar and pounded brick. They rose from 22 inches to 32 inches in height, closely surrounded on every side with the loose earth and clay, as it lay upon a third flooring, made of pure and unmixed mortar, 3 inches thick, and having a layer of red sand below on the natural ground. About a yard to the east was discovered:

III. A third building, but all a mere mass of confusion. In the broken ruins of it were dug up a couple of Roman coins and three round tubes of tile. These were found in the ground with their mortar adhering to the outside, each 16 inches high. They were plainly formed in moulds, and hooped with circles on the outside, and narrowed from 4 inches at one end to 2 inches at the other in diameter, and they were inserted into each other, forming a long pipe.

And these buildings were situated within the irregularly semicircular projection of the river bank, below the level of the station, and on the edge of the water.

Near the south-western angle, and along the north-eastern side of it, were found great quantities of bones heaped together, and chiefly of oxen, sheep, and cows.

Many tiles also were found in the ruins, with round holes in them, some larger, some smaller; others were made with a bend for channels. One sough was observed in the building.

In the second building several fragments of coarse tiles were found formed into hollows; also a large iron knife, with a handle of stag's horn, and an iron chisel.

In the more westerly building (I.) the beam of a balance, fitted with a hook at one end. The second and third building consisted only of one large room each, and no partitions were found in either. But they were in the first. The first building had only one flooring, and was placed so much higher than the second building. The pedestal, standing about 2 yards to the west of the first building, was evidently found in its original site, being fixed on a regular basis of red rock and clay.

On the south-western part of the projection, which has never yet been dug into, at one point was found a few years ago (1769) a great quantity of saxa rotunda de fluviis, lying in a large heap immediately under the wall, and fairly turfed over by time.

---

HYPOCAUST.

As a further support and demonstration that the altar erected to Fortuna Conservatrix on the borders of the Medlock refers alone to the hypocaust, and was erected on the platform in front of it with whose

Draught of an antient Roman Sweating-Stove.

Mr Urban,

I Send you inclosed the draught of an Hypocaustum, or Bath, discover'd at Netherby, 10 miles north from Carlisle. You may compare it with another at Cast-steeds communicated to you (see Vol. xi. p. 650, Vol. xii. p. 30, 31, and Vol. xvii. p. 60.)

The rooms mark'd c were the sudatories, or sweating places, where the people retired after bathing. a a were for exercise, &c. That mark'd f, a bath for ablution; a necessary part of the Heathen theology in the worship of Fortune, to whom the altar is consecrated, which is also in another of your Magazines (see Vol. X. p. 171.) The communicating funnels (b) supply'd the fire with fresh pabulum of air, and at the same time the pipes (d) heated the sudatories G. Smith.

A more particular Explanation.

a a a a Thirty-six pillars of square tyle, one laid above another, and a little cement between; they were about two inches thick, and each pillar was about a yard high.

b b Two funnels, or air-pipes.

c c Fifty-four pillars of solid stone 36 of which were cover'd with flags, and cemented above.

d d d Three hollow tyles, or pipes thro' the wall.

e The sacrifice room, where the altar was found, with the inscription.

f The bath, discover'd in the year 1732.

N. B. Most of the rooms were floor'd with a small thick flag, laid in cement with three course of pavement below.

[Page from The Gentleman's Mag., vol. xx., 1750, p. 27.]

measurements it fits, we have only to look at the ground plan of the hypocaust found at Netherby and alluded to already in the text. We see from it that there the altar dedicated to the Dea Sancta Fortuna Conservatrici is exactly in the same position as found in Deansgate. A better proof could not be produced. It will also help us to form an idea of what our hypocaust probably appeared to the spectator in its original condition, and for this purpose the drawing and description as found in The Gentleman's Magazine, vol. xx., p. 27, 1750, are reproduced here.

---

GRAFFITTI.

The graffitti may be arranged into three distinct series, first those ending in V, secondly those in X, and those scratched in in full name; we have thus:—

· V[93]	· X	C O I I[94]= Coh. II.
D V	ᐸ X
V V	T X	and:
W (lingulate)	X X	J V N V I (lingulate)
I V	· X	Ʌ V I T T I I
N (lingulate)

I may observe that all these graffitti have been obtained from the northern Roman suburb, viz., Bridgewater Street and Tonman Street; and the pottery collected here, as far as the Samian ware is concerned, on which they are scratched, refers to types of ornamentation of the second and third centuries. We also know that the soldiers at that time were quartered outside the castrum, and these graffitti occur a little outside beyond the northern wall of the station. I have only found thirteen of these initialled and inscribed pieces, but there is no doubt that there must have been many of that sort promiscuously scattered over this area, and probably pointing to different periods. Those written in full seem to have belonged to private individuals, while those issuing in V and X appear to have been the property of the mess-table of the soldiers. The constant recurrence of these signs is not accidental. I throw it out as a suggestion worth consideration that they may indicate, as in VV and XX, the possession of the Twentieth Legion and the Valeria Valens; and, in the case of single V, the abbreviation of the Sixth Legion, called the Victrix, where in one case we seem to recognise the mark of the centuria. It would be of interest to compare these graffitti in the same light with such found at other stations to see if this conjecture is borne out.

PERSPECTIVE VIEWS OF MANCUNIUM.

I have to add a few words of explanation with regard to Mr. Phelps's drawing. It has been prepared from a careful geological survey of the ground and a due consideration of the physical features of the locality, and is supposed to be viewed from the Gaythorn side. The structural details were obtained from the ground plan that accompanies it, and must be considered correct and properly authenticated. Only those points have been introduced of which remnants and indications were still visible, and of which we have accurate and incontestable records, such as are found in Whitaker's and Thompson Watkin's accounts, and in addition to this, the further discoveries made by myself in the last four years, so that imagination has been excluded from the the reconstruction which has been attempted. The only regret is that the former destruction and a want of personal watching within the last twenty to thirty years has prevented us from filling in more detail of buildings and structures; as it is, the view offers only a mere outline sketch. With regard to the fosses that originally surrounded the northern and eastern wall, I wish to remark that inadvertently the northern fosses which are just shown should not have been drawn curving round the western wall. These fosses as shown on Whitaker's "View of Mancunium," extended in a straight line considerably beyond the northern side and then turned, but did not merge into the western rock-excavated fosse, which slanted off at a fair distance and at an angle towards the Medlock, as already adverted to in the footnotes of my paper. The probable course the latter took is indicated on Whitaker's View by the series of gardens on the Knott Mill side, which occupied their original position. The prospective view does not show the fosses, and the northern Roman suburb has been indicated instead in faint outlines; this has been done to save the preparation of two separate views: either we have to omit the town and show only the fosses, or vice versâ. When the town was erected the fosses were probably already in toto or partially filled up, for we have the road to Buxton built over them, as proved during the excavations. The rocky height is also a little exaggerated for greater prominence. It is necessary to mention this for fear of misinterpretation.

1. For want of a proper municipal historical museum most of the rare and priceless treasures and relics of Manchester have become dispersed or lost, as will be seen in the appendix from the list of the older finds from Castlefield and elsewhere. Chester, Liverpool, Warrington, and Preston have set a better example, and greater public appreciation has been shown in these places for the preservation of local antiquities.
2. Although the walls were really running north-east, south-east, south-west, north-west, I retain for convenience Whitaker's description, and call them, therefore, north, east, south, west.
3. See Memoirs of the Manchester Literary and Philosophical Society, vol. v., part ii., p. 676. 1802
4. It measures 16 inches by 12 inches and is of millstone grit, and was found near the tower, at the eastern angle of the fort. See Rev. Watson's description of Melandra. Archæologia, vol. iii. 1771.
5. The general and larger section, showing all the trenches, is in the Manchester Reference Library.
6. See section.
7. I have kept samples of the various kinds of mortar and a specimen of the tough clay with the boulder stones imbedded in it for reference and as an ocular demonstration of the foundations of Roman Manchester.
8. Vol. i., p. 20, ed. 1771.
9. The situation is indicated on his "ground plan of Mancenion," taken August 8th, 1765, and corresponds to Knot Mill, facing Gaythorn Street, Hewitt Street, and Commercial Street.
10. See vol. i., pp. 223, ed. 1771.
11. See vol. i., p. 21.
12. See General Section in the Reference Library, King Street. B
13. I had saved one of the blocks for later examination, but on my return the men had removed it to my great regret. Each stone must have weighed between 30 pounds and 40 pounds.
14. The section has been photographed, and is in the Reference Library.
15. Whitaker, vol. i. p. 37, 1771; see also his "Ground Plot of Roman Mancunium."
16. See the "Excavations of Birrens, a Roman Station in Annandale." — Proceedings of the Society of Antiquaries of Scotland, vol. xxx., 1896, pp. 81-199, and plans, p. 96.
17. The contents of this pit will be described in the appendix.
18. He also notices such another arch, vhich was visible in 1769, a little to the "west of the south-east angle, and the crown just rose above the ground.
19. Asphalt was also used along with the usual mortar.
20. This is the present Crown Inn.
21. The situation is made clear from a reference to Green's Map and Raffald's Directory of 1772. John Markland was a large check manufacturer, who lived at No. 35, Deansgate.
22. Another considerable deflection has been noted by Binney, and is proved by finds of Roman colonial coins between Cambridge Street and River Street, where the Medlock runs respectively 100 yards and 58 yards more westerly at present.
23. Found on the exact line of the road to Chester and a few hundred yards to the southwards from the station at Castlefield, viz., near Jackson's Lane, or Great Jackson Street. It is known to have been usual for the Romans to erect, without the boundaries of the stations, where they were in garrison, votive altars, and centurial and other stones to the honour of favourite deities and in commemoration of events. The idea that all the above objects should have been carried to the Hulme side, viz., 400 yards to 500 yards away, is preposterous. One of them resembles No. 169, page 79, described in Professor Haverfield's Catalogue of the Roman and Inscribed Stones in the Grosvenor Museum, Chester, 1900, which at the present he considers safest to class as sepulchral. The three stones occurred close together in the same spot, the tile tomb is in the same place, so that we must infer them to be all sepulchral commemorative stones. The full inscription and tenour of the Frisian centurial stone, also found here, has never been made out. (See Plates.)
24. The construction of this rather prolonged stem seems to be due to the interposition of the slanting, rock-cut, eastern fosse. The Porta Prætoria was placed on this (the east) side (facing the Brigantian enemy). The roads to Cambodunum and Condate split off from this common stem.
25. In 1292 called "Wodarne forde." See Manchester Guardian, February 15th, 1887. J. P. Earwaker.
26. Two third Roman brasses have been found at Ordsall Lane, near this Roman road, of Gallienus (258–268) and Constantinus II. (337–340).
27. The Roman road from Wagstaff Fold (Failsworth) to Street is well indicated on William Johnson's "Plan of the Parish of Manchester, 1820."
28. By the time the road to Buxton was made and the northern suburb erected, the fosses on the northern wall were already filled up and levelled. The ground plan and the perspective view of Mancunium (see engraving), which includes a view of the suburb, also show the northern fosses which, of course, should have been omitted there, but are partly indicated just to give an idea and to save the expense of another special plan.
29. These were Humphrey's gardens.
30. Whitaker seems to be quite right in this latter remark, for, just at this point, the levels to the west of Quay Street sink rapidly away, and are marked by a line extending from Water Street to Canal Street, where the land lies rather low, which we may express by the 93 feet contour or inundation line, and the ground must have been very swampy.
31. Up to Hanging Ditch.
32. See also the article "Roman Britain," Edinburgh Review, April, 1899, p. 369–390.
33. Viâ Wilderspool and Warrington.
34. For the course and direction of the various Roman roads, see William Harrison's excellent map in the Archæological Survey of Lancashire.
35. See footnote in his History of the Siege of Manchester.
36. Since apparently confirmed.
37. The centurial stone of the First Cohort of the Frisians, found in the rubbish at the eastern gateway, along with coins of Trajan and Hadrian, is supposed by Dr. Holme to point to the destruction and subsequent repair of this entrance.
38. Mr. C. F. Bell, assistant curator of the Ashmolean Museum, has kindly taken a fine squeeze of the altar in question, which is now placed in the Free Reference Library.
39. In the York Museum is an altar dedicated to the Dea Fortuna. This goddess was specially worshipped in connection with baths, and inscriptions have been found to her under the title of Fortuna Balnearis. See Orelli Inscr., 5796–7. See York Museum Catalogue, 1891, p. 32.
40. It slants right away here from the wall at an angle at a good distance, instead of running parallel, and I cannot account for this peculiar deviation and construction unless it was fed with water direct from the Medlock in some manner.
41. A Roman moulded capital, of millstone grit, is still lying, half concealed, in the timber-yard at Duke Place, near Castlefield.
42. A first brass of Hadrian has been found close to the remaining piece of the wall, and is still extant.
43. Or, more exactly, 44 yards away from the east gate.
44. See Harland's Mamecestre.
45. A kiln has since been discovered at Wilderspool (Veratinum). See "Discovery of a Roman Potter's Kiln in a Sandpit at Stockton Heath," by Ths. May. The Antiquary, September, 1900, pp. 258–9.
46. See Harrison's Archæological Survey of Lancashire.
47. See "The Mersey as known to the Romans." Hist. Chesh., vol. ii., p. 187.
48. These form now part of the "Charles Potter collection" presented to the Grosvenor Museum, Chester, which is the finest and most complete set ever got together from Dove Point.
49. See Antiquities from the Sea Coast of Cheshire, by Rev. A. Hume, 1863.
50. The drainage of the whole country west of the great watershed of the Pennine chain finding its way over these gradual slopes and undulating flats must of necessity, says Mr. De Rance, have been much obstructed, and prepared the way to the formation of the mosses.
51. Eight canoes were found in Merton Lake (=Martin Meer), one of which had some plates of iron on it, two skin boots, a skin cap, and a bronze celt, and between Blackpool and Poulton three coracles and a bronze axe in the morass in Sawick. About these and other finds in the Fylde see Natural History of Lancashire, &c., by Chs. Leigh, 1700.
52. See Speed's map and the Polyalbion.
53. But compare such names as Segobodium (=Seveux), Segobriga (=Segorbe), Segodunum (=Rodez), also Segia, Segosa, Segisamo.
54. The secondary meaning is thieves, depredators, grassatores.
55. We have also the Hill towns: Segobriga, Nemetobriga, Flaviobriga, Brigantium, Brigiosum, &c., &c.
56. See Numismatic Chronicle, vol. xvii., part iv., p. 294–301, G. F. Hill's paper "On the Coins of Cartismandue."
57. Also Deæ Nymphæ, Brig, Castlestead (about a.d. 203); Deæ Brigantiæ Sacrum, South Shields; Brigantiæ S. Amandus, &c., a statuette found at Birrens (Annandale).
58. In the Horæ Britannicæ, by John Hughes, London, 1818, vol. i., p. 132, the Brigantian war-dance, chwareu Brigant, is alluded to. In Owen's Welsh-English Dictionary, 1793, it is described as follows: In the British war-dance, the parties sometimes move in a circle, and at other times they clap their hands in particular passages of the music; after having taken one of the party, who is called brigant, he is made to pass under their arms, and then he is sent to some work, as a tailor or shoemaker, to denote his being a slave. He, being reconciled, falls to work and sings a song, whilst the dance continues.
59. Upon the etymology of the local river names, such as the Irwell, Irk, Tib, Corn, Gore, Medlock—all of Brythonic origin,—I shall return perhaps at a future occasion.
60. The wild bull continued in the extensive woods of Blakeley as late as the fourteenth century (Leland, vol. vii., part i., p. 42, Hearne).
61. See On Vessels of Samian-ware. By K. Syer Cuming. British Arch. Assoc., December, 1891.
62. See Excavations in Bokerly and Wansdyke, and their bearing on the Roman occupation of Britain, by Pitt-Rivers, Lanc. and Chesh. Antiq. Soc. vol. viii., p. 17., 1890.
63. A F lingulate.
64. Querns, in situ, imbedded in a setting of clay and fringed round with small boulder stones, have been found at Melandra (Hamnett). See also "Roman and Gallo-Roman Flour Mills," Scientific American Supplement, October 6th, 1900, p. 20,714, and illustration, figure 5.
65. An ancient quarry has been discovered there, near the Roman road (see Palatine Note-Book, vol. iii., p. 172, 1883). At Melandra Castle the millstone grit seems to have been obtained from Hargate Hill, a few miles away from Glossop (R. Hamnett).
66. The old "Stone Delph" is indicated on Green's Map, and was owned still in 1793 by Sir John Parker Mosley, the lord of the manor.
67. Date fixed to a.d. 146, see Roman Britain, by H. M. Scarth.
68. Figured by Noel Humphreys in his second volume.
69. All from Castlefield.
70. See also detailed account in Manchester Weekly Times, March 18th, 1898, "Notes and Queries," No. 2,227, "Roman Coins and Roman Stations, by Fred. L. Tavaré.
71. Toad Lane is derived from the word toad; it corresponds to the German equivalent Kröten Gasse, and its habitat points to the former ditched character of the place.
72. This fine, bedded sand had long thin streaks of phosphate of iron (of blue colour), indicating the presence of decaying animal remains.
73. See Hibbert-Ware's History of the Foundations of the Ancient Parish Church of Manchester.
74. His MSS. volumes are now the property of the corporation and incorporated in the Reference Library.
75. Although scarcely necessary, I may say that the assumption of the mortar having been carried here from Castlefield is absurd, considering the circumstances I have described. Leland, a casual visitor of Manchester in 1538, by the by, only speaks of the stones of the ruins of Man (Castle having been translated towards making of bridges for the toune." Salford Bridge chapel was repaired in 1505. The other small bridges were at Hunt's Bank, Millbrow, and Scotland. The dressed red sandstones may have been used for the foundations of the Deanery. Hanging Bridge, Chetham College, which are all made up of stones of mixed sizes, but no blocks of millstone grit can be discovered amongst them; they may have been used for Salford Bridge particularly.
76. See his paper in our Transactions, vol. xiv., p. 65, 1896.
77. Iter Boreale. Ed. 1776. By William Stukeley.
78. See Manchester Guardian, 5th Dec, 1883. Note by J. P. Earwaker, who quotes from some MSS. of the Rev. John Watson (+1783) now in the Bodleian Library, Oxford.
79. The Rev. J. Watson, who was for some years curate of Ripponden and lived in the neighbourhood for near twenty years and who first identified Cambodunum, does not speak of this paved road, then plainly visible, as a Roman road. See also J. E. Bailey's letter to the Guardian, December 15th, 1883, on the whole question of the old and new Blackstone Road.
80. See also City News, May 22nd, 1886, and Green's Map. It was later on euphonised into Cold House
81. See "Broughton Place-name," City News, No. 4,868, November 18th, 1887, by J. P. Earwaker.
82. Ibid, No. 4,814, November 11th, 1887, Joshua Bury.
83. See Harrison's Archæol. Survey of Lancashire.
84. There is also the place-name of the Barrow brook, near the Paradise, in 1628 (see Ancient Court Records of Salford).
85. The name occurs in 1320 as Burghe-ton, and Burgh-ton de Salforth in 1322 (see Harland's Topographical Gazeteer).
86. See City News; Lanc. and Chesh. Antiq. Soc., vol. iv., 1886; and also Manchester Guardian, 20th April, 1887.
87. See Joshua Bury's "Broughton Place-names," City News, November 19th, 1887.
88. See Lanc and Chesh. Antiq. Soc., vol. v., 1887, p. 327.
89. A neolithic factory was discovered by me thirteen years ago on Kersal Moor, and a fine collection of cores, flakes, knives, &c., from that locality, made of chert and flint, has been presented by me to Peel Park Museum.
90. In 1473 mention is made in the Rental of "a fishing in the water of Irk, late in the tenure of John Huntington [the warden], 2/- yearly," see Mamecestre.
91. As shown in the list of the earlier and lost Roman finds.
92. Manchester Mecury, May 28th, 1771, communicated to me by Mr. John Owen; Whitaker, Principal Corrections, page 14-19, vol. i., 1771.
93. Incomplete and front broken off.
94. Compare also plate p. 24 in Roman Inscriptions in Britain, 1888-1890, by F. Haverfield, in Archæological Journal, vol. lxxvii., p. 229, of a leaden graffitti found by me at Chester, 1886, with C O I I scratched on, in duplicate, and read by him as Coh. II.