Chapter II: Part 2

The smaller inset plan on Fig. 22 is a very visionary reading of the possibilities. A street in line with Fish Street Hill and the Bridge, which I will call _Axis Street_, may not have pointed to the centre of this great building, but rather by its west end as suggested (X). If this is too far west for the _Axis Street_, then we must suppose that it was directed towards some point in the south front of the Forum. (It is desirable that all the walls found in this locality should be accurately laid down on a plan.) A parallel street to the east, which I will call _North Gate Street_ (Y), would not be in continuation with _Axis Street_. The question whether Bishopsgate and Gracechurch Street represent a Roman street from the Bridge to the Gate has been much argued over (see _Archæologia_, 1906), and it seems to have been shown that the line was interrupted in some way. The southern part, however, must, I think, represent the Roman street from the Bridge, although it may later have been bent aside to tend more directly to Bishopsgate. The facts and the fault in the line may be reconciled in some such way as suggested. (Hodge’s drawings are in the old Gardner collection, and it would be interesting to know what other Roman records are included.)

Beyond the statement quoted from Brock no identification of the building is offered in _Archæologia_, and Mr. Bushe-Fox thought that if the walls were contemporary they could not belong to a Basilica. “If there were a nave with two aisles and an apse there would be no reason for the cross wall, nor for the excessive thickness of the side wall. The building had perhaps been a bath; the wall which ended abruptly at the west end was probably a flue for heating the apse, and the large drain would be accounted for” (_Proceedings_, 1914-5). That the building was indeed the civil Basilica of Londinium is proved to my mind by: A comparison of the plan with those of other British Basilicas—notice the way that the apse is within straight external walls, and compare Fig. 21; by the great scale of the work; by its central position in the City; by the scale and character of the construction; by the fact that the only possible alternative seems to be the supposition that it was the great Bath of the City, and for this neither the planning nor the situation seems suitable; by the exceptional wall decorations described below; by the fact that a tile bearing the official stamp PR-BRILON was found on the site (Price). It is a remarkable fact that Leadenhall was the market, and that the Crossing at Cornhill was the carfax of London during the Middle Ages.

We have seen above that Brock said that fragments of painting were found on the site. In the British Museum are four pieces of wall painting, given by Mr. Hilton Price—1 and 2 in 1882, and 3 and 4 in 1883; the first pair are said to be from Leadenhall, the second pair from Leadenhall Market. One and 3 are fragments of large-scale scrolls of ornamental foliage of a grey-green colour; 2 is a piece of large-scale drapery, and 4 is part of a life-sized foot. These four remarkable fragments evidently form one group and came from the Basilica. The large scale of the ornament and figure work differentiates these pieces of painted plaster from all others found in London. At Silchester and Cirencester fragments of marble wall linings have been found on the sites of the Basilicas, and some of the marble fragments in the British Museum may have come from our Basilica, which must have been a handsome, indeed splendid, civic centre. In the Forum would have been statues of Emperors, and in the Basilica some impersonation of Londinium itself (cf. the fragment of such a figure found at Silchester, now at Reading).

FIG. 23.
]

_Houses._—In 1869 a mosaic pavement was discovered in Bucklersbury which is now at the Guildhall (_Builder_, May 15 and 29). It was fully described in a volume by Price. The floor was that of a small round-ended chamber, and belonged to a building on the western bank of the Walbrook, Around the apsidal end of the room which had the mosaic was a wall of stone and chalk, built upon piling; this wall contained the flues of the heating-system, and it terminated in piers at the ends of the semicircle. From the fact that no more walling was found and the evidence of an attached lobby which had a wooden sill around it, we may suppose that the rest of the house was of timber work (Fig. 23). The curved apse would be a strong form in which to build a mass of wall to contain the vertical wall flues; and it is an interesting example of building contrivance. We have already seen that timber and clay construction was frequent in Londinium. Near this building a well was found (built of square blocks of chalk, _The Builder_ says). This building with the mosaic floor must have been a superior house on the bank of the Walbrook. To the west, as we shall see, seems to have been a street possibly of shops; we can thus imagine a little group of buildings and streets, and a bridge over the Walbrook at the end of Bucklersbury.

FIG. 24.
]

The well mentioned above is one of a great number which have been discovered; for instance, in excavating for Copthall Avenue “a pit or well, boarded, and filled with earthenware vessels,” etc., was found (_Builder_, October 5, 1889). Such wells with boarding like a long barrel have been excavated at Silchester. Again to the south of Aldgate High Street two wells were found (_Builder_, May 3, 1884).

The most complete Roman building which has been recovered and planned is one excavated in Lower Thames Street in 1848 and again in 1859 (_Builder_, February 5, 1848, and June 11, 1859). A restored plan was given in the _Journal of the British Archæological Association_, vol. xxiv. (see Fig. 21). The two apsed chambers had hypocausts beneath their floors, supported on little piers built of tiles 8½ in. square, and broken materials. Fig. 25 is reproduced from the illustration of the eastern chamber given in _The Builder_. Several sketches and some notes, by Fairholt, of this building are in the Victoria and Albert Museum. About 4 or 5 ft. of the walls remained in places, all of tiles with mortar joints nearly as thick as themselves.

FIG. 25.—(A, old masonry; B, brickwork with a flue tile; C,
foundations of chamber.)
Foundations discovered in Lower Thames Street in 1859.
]

“The walls were of red and yellow brick in alternate layers composed of 18 in. tiles.” Outside the walls was “a drain of wooden planks, 18 in. deep by 10 wide, running towards the river” (see plan). The walls were erected on piles. The sketches show some of the box-flue tiles which had impressed patterns (see Fig. 25). Some additional information is given on a lithograph by A. J. Stothard (1848). The walls were 3 ft. thick. Above the floor of the south room, which was of coarse red and yellow tesseræ, was a second, about a foot higher in level; this was “a layer of red concrete 2½ in. thick, hard, and the upper surface almost glazed” (compare a floor found in Eastcheap, “concrete stuccoed over and painted red.”—_V.C.H._). This building was doubtless a house; at the time it was found it was called a bath, but it seems too small to have been even a secondary public bath. As Thos. Wright says: “Many writers have concluded hastily that every house with a hypocaust was a public bath” (cf. the plan of a house at Lymne, _The Roman, etc._, p. 160). The stoke-hole of the hypocaust was at F, and there were flues up the middle wall and the western apse. The large room was 23 ft. square; some tiles of 2 ft. square were found here, also window glass and an iron key. The plan lay square with the south City Wall (Fig. 24), and the building can hardly be earlier than this wall. It may thus be accepted as a late fourth-century house, and we may further infer that box-tiles with impressed patterns were a characteristic of this century. On two sides of the house were lanes about 10 ft. wide. As in so many cases modern walls seem to have been laid out on the same alignment as the Roman building.

The house just described had two apses, and the Bucklersbury house also had an apse. This was in agreement with general custom. As Thos. Wright remarked: “One peculiarity which is observed almost invariably in Roman houses in Britain is that one room has a semicircular alcove, and in some instances more than one room possesses this adjunct.” In the plan given in _Archæologia_ of the Roman walls and floors found in and about Lombard Street in 1785 two apses seem to be indicated; thus we have evidence for five in the scanty records; altogether there must have been scores in the city.

Within the walls of the City were many large houses of the villa type as well as minor dwellings and streets of shops. Roach Smith speaks of such great houses about Crosby Square; he also describes a mosaic floor under Paternoster Row which extended 40 ft.; a second important floor on the site of India House, Leadenhall Street, was at least 22 ft. square, and may have been considerably more; a third large floor which was found under the Excise Office, Broad Street, was about 28 ft. square (probably 30 Roman ft.). All these must have been the floors of the chief central rooms of large houses of the villa type. Tite saw this of the Broad Street floor as his speaking of the “triclinium, other rooms, and the garden” shows. This Broad Street pavement was lying square with more modern walls surrounding it, and it may not be doubted that buildings continuously occupied the site.

The supposition that there were important houses of the villa type within the walls of the City has been fully confirmed by the excavations at Silchester, and I may here quote Dr. Haverfield’s general conclusions as to Roman towns in Britain. “Roman British towns were of fair size, Roman London, perhaps even Roman Cirencester were larger than Roman Cologne or Bordeaux. They possessed, too, the buildings proper to a Roman town—town hall, market-place, public baths, chess-board street-plan, all of Roman fashion; they had also shops and temples and here and there a hotel.... The dwelling-houses in them were not town houses fitted to stand side by side to form regular streets; they were country houses, dotted about like cottages in a village. But in one way or another and to a real amount, Britain shared in that expansion of town life which formed a special achievement of the Roman Empire.” The evidence as to the isolation of the houses is here a little overstated, but in the main the passage gives a true impression. Fragments of wall decorations and mosaics found in Southwark suggest that there were big houses on that side of the river, and doubtless others occupied sites along the Strand and Holborn.

FIG. 26.
]

I give here a little sketch plan (Fig. 26) of a house found about a century since at Worplesdon, Surrey, from a survey at the Society of Antiquaries. This house is interesting as its unaltered plan gives an example of a simple “Corridor House.” It was 62 ft. long by 22½ ft. wide within the foundations, and faced west. The slight foundations of flint, not much more than a foot wide, show that the walls must have been of timbering or wattle work. The rooms and passage had floors of plain coarse tesseræ, except that the outer side of the passage had a simple twist border in mosaic. Possibly there had been some pattern in the central room as the floor was there missing, and a note reads: “Near this place was found the lozenge-shaped tessellated pavement.”

_Baths, Temples, etc._—Remnants of important buildings have been found in Cannon Street from time to time, and London Stone is probably a fragment of one of them. Wren was of the opinion “by reason of its large foundations that it was some more considerable monument in the Forum; for in the adjoining ground to the south were discovered some tessellated pavements, and other extensive remains of Roman workmanship and buildings.” Under Cannon Street a building with one apartment 40 ft. by 50 ft., and many other chambers, is mentioned in _V.C.H._ At Dowgate Hill the foundations of large edifices are listed in _V.C.H._, and of Bush Lane it is remarked: “That there must have been extensive buildings here seems clear.” At Trinity Lane, Great Queen Street, “great portions of immense walls with bonding tiles” have been found (_V.C.H._). There was a house on the south side of St. Paul’s known as _Camera_ or _Domus Dianæ_ which may have taken its name from some Roman monument. In a St. Paul’s deed of 1220 it appears as a messuage or inn, _domum que fuit Diane_.

In December 1921 Mr. Lambert described the foundations of a building by Miles Lane. The plan of this suggested a house of the corridor type facing east. The site seems to have been levelled up by timber walling or wharfing against the river and running back into the sloping ground.

One of the most important public buildings in the City would have been the Public Baths, as those of Silchester and Wroxeter show. At Trèves the great Baths cover acres of ground by the river. Bagford says that after the fire of London some Roman water-pipes were found in Creed Lane “which had been carried round a Bath that was built in a round form with niches at equal intervals for seats.” This suggests a part of important Baths, and Creed Lane does not seem an unlikely situation for the Public Baths. (In _V.C.H._ the site is said to have been in Ludgate Square.)

FIG. 27.
]

The only certain evidence we have for Temples are some inscriptions and sculptures. For the most part they would, like those found at Silchester and elsewhere, have been small square and polygonal structures set on a rather high podium approached by steps. Fig. 27 is a restored plan of the little Temple found at Caerwent. Doubtless here and in most cases, the roof of the cella ran on to cover the podium. At the foot of the steps an external altar would have stood. The column illustrated before (Fig. 1) seems suitable for a temple. Roach Smith, speaking of the group of Mother Goddesses found in Crutched Friars (see _Builder_, October 30, 1847), says: “It is the only instance with the exception of the discovery made in Nicholas Lane in which the site of a temple can with reason be identified” (_Ill. Rom. Lon._, p. 33). The find in Nicholas Lane was part of an important and early inscription which may have been on the chief temple in Londinium. Some sculptures found on the bank of the Walbrook suggest that a cell of Mithras occupied the site. In the fourth century a Christian church would, as at Silchester, have occupied an important site in the City.

A large Theatre or Amphitheatre, or both, would have been necessary in such a town. Roach Smith, who had a wonderful instinct of insight, thought that such a building probably occupied a site against the bank of the Fleet, called “Breakneck Steps.” Lately it has been suggested that the drawing-in of the line of the City Walls at the north-west angle was done to avoid an amphitheatre; more probably, I think, it was to avoid wet ground. There is evidence that gladiator contests and chariot races were popular. For gladiators, compare two small bone figures at the British Museum, evidently from one shop, with the fragment of a little statuette at the Guildhall. The bronze trident-head, also at the Guildhall, really does seem to be a gladiator’s weapon as suggested in the catalogue. For chariot races, see the fragments of glass bowls, which may have been made in London, in the British Museum. I have found an additional little point of evidence on chariot races. Amongst Fairholt’s sketches at the Victoria and Albert Museum, is one of an enigmatical little fragment of a Castor vase, found in Bishopsgate Street, which seems to represent four heads of dogs running neck and neck. Now there is a whole vase in the British Museum (found in Colchester) which was practically a replica of the other, and this shows that the four running animals of the fragment were chariot horses, and the whole represented a race. Above the horses of the fragment is scratched ITALVS, which, I suggest, must have been the name of some favourite “winner” in Londinium.

_Streets._—In his account of the Bucklersbury pavement, Price describes also some walls which were found “about 30 yds. westerly from the pavement” (the position is shown on his plan). “Two Roman walls running nearly in line with Bucklersbury directly towards the Walbrook.” In the space between them had been laid a drain to fall towards the brook with a tile pavement above, and mortar fillets against the walls. The walls were 2¾ ft. thick, and built on three rows of piles, and the space between was 2¼ ft. The tiles are of the usual kind of red and yellow brick. Above these walls were others of chalk and stone 3 ft. apart, of later date. This is one of a great number of instances where we find that mediæval buildings were founded directly on Roman walls. The space Price suggested was “an open passage-way, or it may be of an alley between two buildings.” Comparison makes it certain that the walls were those of neighbouring houses in a street; similar conditions have been found at Caerwent, Silchester, etc. At the former “the shops along the main street were probably roofed with gables; this is substantiated by the finding of a finial in front of a house. The narrow space between the houses would serve to carry away the water which would drop from the eaves” (_Archæol._, 1906). The walls are shown in Fig. 28.

The Bucklersbury paved passage, only just wide enough for a man to get at it, with the underlying drain, is obviously a similar space. The tradition of dividing houses in streets from one another in this way lasted into the Middle Ages (see V. le Duc’s _Dict._, “Maison”) and, of course, occasionally to modern times. By this means party walls and difficult roof gutters are avoided. From the two parallel walls we are justified in inferring a row of houses—possibly shops—and a street running to the west of them; moreover, the example suggests to some extent what continuous streets of houses must have been like. In Southwark a passage-way between houses was found 3 ft. 8 in. wide. A wall with tile paving against the outside, found under the Mansion House, suggests a similar passage.

FIG. 28.
]

It has been mentioned above how in several cases, as is clear even from our imperfect records, that later walls were founded directly on Roman walls. Modern buildings were thus in direct and unbroken succession to Roman ones and maintained the same alignment. In the Archer collection at the British Museum is a drawing of “a Roman pavement and foundations, supposed to be remains of Tower Royal” in Cannon Street; this again was square with modern work. Roman remains have been found under several churches. Massive walls of chalk were found under St. Benet’s, Gracechurch. Roach Smith, speaking of a floor found at the corner of Clement’s Lane, says: “This adds another to the numerous instances of churches in London standing on foundations of Roman buildings.” In 1724 Roman foundations were found under St. Mary, Woolnoth, and “three foundations of churches in the same place” (_Minutes Soc. Ant._, June 17). Even Westminster Abbey and St. Martin’s in the Fields were built on Roman sites, and so probably was St. Andrew’s, Holborn.

This continuity of the buildings from the Roman Age is not only an interesting fact, but it is a strong argument for the general continuity of the street lines as well. The plan of the extensive finds in and about Lombard Street in 1785 shows the building to have conformed very much to lines parallel with, and at right angles to St. Swithin, Sherborne, Abchurch, Nicholas, Birchin and Clement’s Lanes, and I cannot doubt that these lanes are in some degree the successors of Roman streets. In “Lombard Street and Birchin Lane the discoveries are said to have indicated a row of houses” (_V.C.H._). If all the evidence as to the “orientation” of buildings and walls was laid down on a plan, merely marking the direction of the minor ones with a cross, we might build up further results in regard to the direction of the streets. At the same time it would be vain to expect any large and simple scheme of lay-out of the chess-board type, the Walbrook and other streams, and probably the persistence of some earlier lines (Watling Street to St. Albans?) would have interfered with that. The Walbrook seems to have been crossed by two chief bridges, which must have been governing facts in the lay-out. One was at Bucklersbury, the other, Horseshoe Bridge farther south, is recorded from the thirteenth century. Cannon Street, I cannot doubt, represents one east to west street. Thames Street must have been formed when the south City Wall was built. I have spoken of the north-south lines above. Saint Benet “Gerschereche” is mentioned in a charter of 1053 (_Athen._, February 3, 1906).

Wren found a “causeway” made up of stones and tiles by Bow Church (under the present tower). It is suggested in _V.C.H._ that this was an embankment, but _causeway_ was one of the regular names for a Roman road. At Rochester one 5 ft. or 6 ft. thick of hard stuff has been found crossing some soft ground.

The best way now to see again the old Roman City of London is to go to the foot of the hill below St. Magnus the Martyr and then, turning away from the riverside quays of the seaport, to walk up the street which still retains something of the look of a High Street in an old market town. Behind it we may still discern the ghost of the Roman Axis Street. Right and left are narrow streets with red plastered houses separated by little “drangways.” Here at a corner is a small temple with a dedication to the deified emperor. There is the great City Bath. Farther on is the civic centre, the market-place and hall; one, a square piazza containing imperial statues in gilt bronze, and the other a big building having internal ranks of tall Corinthian columns, a wide apse, and an open timber roof—sombre but noble. Round about are many isolated and widespreading mansions, one doubtless being the palace of the Governor of the province. Beyond are the walls and gates which will be next described, and in the background rise the northern heaths and wooded hills now called Hampstead and Highgate.

“On alien ground, breathing an alien air,
A Roman stood, far from his ancient home,
And gazing, murmured, ‘Ah, the hills are fair,
But not the hills of Rome.’”
MARY E. COLERIDGE.

------------------------------------------------------------------------

+CHAPTER III+

WALLS, GATES AND BRIDGE

“Gem of all Joy and Jasper of Jocundity, Strong be thy walls that
about thee stand; London, thou art the flower of cities all.”

WILLIAM DUNBAR.

THE walls, gates and bastions of the City may be traced by the record of early maps such as that of Braun and Hogenberg. The bastions of the east side are particularly shown on a plan of Holy Trinity Priory made in the sixteenth century; the west side from Ludgate to Cripplegate plainly appears in Hollar’s plan after the fire, 1667. There were two bastions between Ludgate and Newgate, then an angle bastion to the north; three more on the straight length to Aldersgate, then one beyond that gate at the angle where the wall turned north again; two bastions occurred between this angle and the bastion at the corner where the wall again turned east, which now exists in Cripplegate Churchyard.

Several of the gates stood until 1760. In an old MS. book of notes I find under the heading “Remarkable Transactions in ye Mayoralty of Sir J. Chitty.”—“In July, ye gates of Aldgate, Cripplegate and Ludgate were sold by public auction in ye council chamber, Guildhall, and were accordingly taken down without obstructing either ye foot or cartway, and their sites laid into ye streets. Aldgate for £157, 10s.; Cripplegate, £93; and Ludgate for £148.” Many old drawings of parts of the wall are preserved in the Crace, the Archer and other collections. The exact line of the wall and positions of the bastions has been verified by modern excavations and discoveries. For full description and a plan, see the _Victoria County History_ and _Archæologia_, lxii. (1912). A good description of what was visible in 1855 is given in _The Builder_ for that year.

FIG. 29. See p. 61.
]

FIG. 30.
]

In September 1903 an important section of the Roman wall was found in excavating the site of Newgate Prison; in some parts it was about a dozen feet high. I saw it in October and noted—“The wall is about 8½ ft. wide. On the outside and inside one or two courses of facing stones were first raised and the core of rubble was then filled in to that height; first there was a thick couch of mortar, then a layer of rubble stones, then another liberal supply of mortar running down between the stones as grout; there were two or three such levellings-up in the heights between the tile bonding courses.” The wall had a rough rubble foundation, then a course of plinth stones on the outside, with three tile courses corresponding to it on the inside of the wall, then followed five courses of the fairly square facing stones on both sides of the wall, then two rows of tile, five courses more stone and two rows of tile, then five more stone courses; above this level the wall had been destroyed. The stones and tiles were set in mortar, and the latter, except for the three courses at the bottom on the inside, which served as a plinth, were carried right through the thickness of the wall; the “tiles” were Roman bricks about 18 in. by 12 in. and 1½ in. thick, laid in what we call Flemish bond. The stone facing courses were a little higher at the bottom than upwards, but all were comparatively small and square; there was a clear distinction between the wrought facings and the rubble filling, which was practically concrete. The “facings” were hard skins adhering to the filling and required by the method of building as described above (Fig. 30).

The mode of construction of the wall is likely to be misunderstood when we speak as we almost necessarily do of facings and filling and of bond tiles. The “facing” stones were small, roughly wrought, and set in much mortar; they formed outer skins to the concreted mass into which they tailed back. The whole was homogeneous. The method was analogous to the facing of concrete with triangular bricks notching back into the core.

The tile courses in the City Wall were doubtless bonds, but they also divided the wall into strata locking up the moisture of the mortar from too rapid absorption and evaporation. I have little doubt that the wall was carried up a stratum at a time over long lengths; it would thus have been available as a defence from an early stage, and scaffolding would not have been required. The building of this wall and casting the ditch about it required a great constructive effort. A strip of ground some 100 feet wide must have been cleared as a preliminary. Then the immense quantity of stone required would have been brought by ships and barges. It is often said that old material was not re-used in the wall, but I can hardly think that two miles of chamfered plinth had to be provided out of new stone at the very beginning of the work. And material from destroyed monuments was doubtless broken up for the small facing stones. The lime-burning, brick-making, stone-cutting, as well as the actual building, called for much labour. It would be interesting to have the quantities taken out and an estimate prepared.

The south wall along the river front is well described in _V.C.H._ Roach Smith, in an article in vol. i. of the _Archæological Journal_, recorded the fact that it had “alternate layers of red and yellow plain and curve-edged (_i.e._ flanged) tiles”; the rest being of ragstone and flint. It was founded on piles. In _The Builder_ (January 19, 1912) it is recorded that in digging for a foundation at No. 125 Lower Thames Street, between Fish Street Hill and Pudding Lane, there was found the base of the Roman wall resting upon long and thick timber balks laid crosswise, with piles beneath them; there were three courses of rough rag and sandstone capped with two courses of yellow bonding tiles, all in reddish mortar; what remained was about 3 ft. high and 10 ft. wide, and was at 24 ft. below the existing pavement. Full evidence of the course of the City Wall along the river front has been found (_Archæol._ xliii.). It may be noticed that in mediæval regulations foreign sailors might not go beyond Thames Street; that is, pass where the wall had been, into the City proper. This south wall, like the bastions, contained remnants of Roman monuments.

The south wall would have been interrupted at the outlet of the Walbrook, which must have been a tidal creek. This was doubtless the original harbour, and there would have been quays within the line of the wall. Daremberg and Saglio’s plan of Bordeaux shows a remarkable parallel to Londinium, standing on the bank of a great river, flanked by a little stream and with a port within the walls (Fig. 29). It seems probable that the strong wall which Roach Smith reports as having been found on the east side of the Walbrook may have been a quay wall. The Thames has been much encroached on where it passes the City. In making the approach to new London Bridge three successive embankments were found, one being of squared trunks of trees. A similar timber wall has just been found in Miles Lane. In Lower Thames Street the Roman house found on its north side was built on piles, “probably on the river bank” (_Athen._, 1848), and the south City Wall was wholly built on timbering. In earlier Londinium, Cannon Street must have been the southern thoroughfare.

_Bastions._—In July 1909, when the angle bastion near Giltspur Street was excavated, I noted that close to it the City Wall was badly fractured, and inclined outwards; there had evidently been a serious settlement here, which was sufficiently accounted for by the nature of the ground—wet clay on the bank of a stream. The wall was taken lower than the ordinary level here, and the bastion was founded at a lower level still. The bastion was not bonded to the City Wall, but merely built against it with a straight joint; it was of horseshoe shape on plan and projected about 27 ft., the masonry was rubble in thin courses, and the whole looked mediæval to me. In the careful report in _Archæologia_ it was said that some evidence for Roman date was discovered in the foundation. The facts suggested to me not only that the bastion had been built against the wall, but that it was probably built at a point of failure in the original wall. It is agreed that the bastions were built later than the wall, and with a straight joint between them and it, and I would suggest that they were built to cover cracks and form buttresses as well as for their additional defensive value, and this may very well have been the general procedure. It would have been impossible to build a wall measured by miles on inferior foundations without bad settlements; the Egyptians provided for them by building such walls in sections with inclined straight joints at intervals.

M. Blanchet, writing of the walls of the cities of Gaul, says: “Often the curtains are not bonded with the towers. This independence reminds one of a precept of Philo’s, which advises that the method should be followed so as to prevent the consequences of unequal settlement between the two. But there is a more simple explanation—the town under immediate danger ensures itself first with the curtain and adds the towers after. Most of the fortifications are those which the Romans built on the approach of the Barbarian invasions. To this period belong the walls of Rome and those of the cities of Gaul.” Choisy again has an interesting account of the towers of the walls of Constantinople, with a diagram of arches in the sides of the towers at the ground level, which were built so that the effective part of their foundations should be kept clear of the wall. Now, the foundations of the London bastions provide evidence of a similar way of thinking.

FIG. 31.
]

In Fig. 31 I give a sketch of this angle bastion made on July 5, 1909. Here is seen the City Wall curving round from the north to the west, and against it the bastion. The Roman wall was badly cracked and leaning outward (A); in the corner by the bastion the plinth and the foundation are seen, and below a sloping bank of wet clay (C), and farther out water (W). The bastion was built of rubble, and was hollow to the base; the form was different below and above (see B). In the sketch the tile courses are seen going through the thickness of the wall.

The bastions which have been most carefully examined are those on the site of the General Post Office, described in _Archæologia_, lxiii. (1912). One is said to have been built in “the usual manner of random rubble”; it was separate from the City Wall, and the foundation was deeper than that of the wall. A second was built in a very soft spot. “Why it should have been selected is not easy to see, as at a little distance either way the builders could have found firm soil.” Its site was an old stream bed, and the conditions might well be the cause of a settlement at the point. This, as suggested above, may have been the reason the bastion was erected just here. (For the bastion by Giltspur Street, see _S. A. Proceedings_, 2 S. xxii. 476.)

FIG. 32.
]

Nothing very definitely Roman was found in these bastions, but one at All Hallows was certainly Roman. This is described as (I condense) “built of stonework which, like the rest, so far as they have been observed, is of random rubble, built principally of irregular pieces and ragstone with portions of Roman tile (none complete) and other material; much of it appears to have done duty in some previous building. A base was formed of large square stones a uniform height of 2 ft.; they had been employed in some former building; several had lewis holes. This base rested on a table of large flat stones 9 in. thick. Most of these seem to have been portions of a cornice. Roman origin was shown by red mortar in which the joints had been set.” The foundation was about 3 ft. below that of the City Wall, and projected into the original Roman ditch. What is called the “table” above was a square-fronted lower base; the back of this base was set in advance of the City Wall; indeed, it was 3 ft. in front of it on the eastern side and “the gravel in this intervening space was undisturbed.” This gap is specially to be noted. The description of the masonry as random rubble must apply mainly to the core of the work, for the illustrations show an approximation to courses on the face; indeed, on the east side, thirteen courses may be counted in the photograph up to a line which seems to be the top of a sloping plinth; these courses averaged about 4½ in. high. The full significance of this account is only brought out on comparing it with Price’s description of what was found in excavating the Camomile Street bastion. This bastion was founded on two deep courses of heavy stones taken from Roman buildings, many sculptured, and having lewis holes in them. These masonry courses were set 1½ ft. in advance of the City Wall, one over the other, forming a straight joint, and leaving a gap “separated from the wall by an intervening space filled with rubble” (Price) which was filled with small stones. This curious and carefully-arranged construction in both bastions was clearly with the object of making the foundations of the bastions take their bearings away from the wall so that they would tend to lean inwards against the wall; it is analogous to the arches of the Constantinople towers. This bastion had a batter or slope at the bottom of about 4 ft. high. Price describes the masonry as “rag rubble walling faced with random courses. The size of the blocks of which the facing was composed varied from 3 in. to 8½ in. thick [high] and from 5 in. to 14 in. long.” This account is supported by the carefully-executed illustrations which show coursed facings of small stones which seem almost identical with the facings of the City Wall. Such masonry of small facing “blocks” with concreted rubble behind is certainly Roman. The masonry at the All Hallows bastion seems to have approximated to the same character; there it may be noticed the courses became narrower upwards. This was certainly not so regular as the masonry of the City Wall, but it may be said to have resembled it (Fig. 32).

At the Guildhall Museum is “a group of architectural remains and fragments of sculptured stones from tombs, public buildings, etc., found in a bastion of London Wall, Duke Street, Aldgate, 1881.” This find is best described in _The Athenæum_ for that year. Mr. Watkins, while excavating in Houndsditch and Duke Street, found the City Wall and a mass of masonry extending 18 ft. outward from the wall; the stones were dressed and weighed from 1 cwt. to 1½ tons. “In the structure he observed a channel 15 in. deep by 18 in. wide, which showed signs of use as a watercourse. It had been filled with concrete composed of chalk and flints. The site was the foundation of one of the bastions composed of sculptured stones in character similar to those previously recorded, upwards of twenty in number.” This was the second bastion east of Bishopsgate. The channel filled with concrete suggests a gap dividing the bastion from the City Wall as already described; but see also account in _V.C.H._

In 1887 Mr. Loftus Brock reported to the British Archæological Association the removal of part of the City Wall on the east side of Wormwood Street. Nearly opposite Bevis Marks Synagogue the foundation of a circular-fronted bastion was found of worked freestones and not bonded into the main wall (_The Builder_, May 28, 1887). A paper by J. E. Price in 1884 (London and Middlesex Archæol. Soc.) referred to the discovery of a bastion containing several sculptured stones in St. Mary Axe (_The Builder_, November 22, 1884, and compare _V.C.H._).

FIG. 33.
]

In 1852 an excavation was made against the outside of the City Wall on Tower Hill, and a number of large wrought and carved stones were found (_The Builder_, September 4, 1852) (Fig. 33). In an account given in the _Journal of the British Archæological Association_ the workmen are said to have discovered a “complete quarry of stones cut in various forms and evidently belonging to some important building ... 125 making 40 cart loads.” Fairholt made an etching of the place while the work was in progress, which shows that the “quarry” was heaped against the external face of the wall like the bases of the other bastions, and that, in fact, it was a ruined bastion Fig. 34 from Roach Smith’s _Roman London_, slightly modified). Another account is given in the Antiquarian Etching Club by A. H. Burkitt, with a plate: “These interesting remains were discovered during the excavations in June 1852, which laid bare the wall to its base. The various portions of stone, which amounted to about forty cart loads, bear evidence of having belonged to an important building. The inscription and band of laurel leaves, which probably formed an ornament above it, indicate a monument of considerable magnitude to the memory of a commander of the Roman Navy. There were found at the same time fragments of frescoes with inscriptions.” (In Fig. 33 the fragment with laurel leaves is represented upside down.)

FIG. 34.
]

The two stones specially mentioned are now in the British Museum. It appears from the accounts and illustrations that this bastion was built against the wall without being bonded to it in the lower part, that its foundation was formed of large carved and moulded stones, and was at a lower level than that of the wall. (The part below the plinth in Fig. 34 on the left is rough foundation.)

We thus have clear record that several of the bastions on the east and north sides of the City were constructed in a similar way. Those farther west near the Post Office were probably rebuilt in mediæval times. These were hollow at the base, not solid like the others.

FIG. 35.
]

The towers of the city wall of Carcassone, described by Viollet le Duc (_Dict._, vol. i.), were so similar in construction that it is plain our bastions were constructed according to general custom. In the illustration we see big stones at the base of the bastion only; large window-like openings closed with woodwork above; and an upper storey rising higher than the wall top. Fig. 35 is a suggested restoration of one of the London bastions, showing the foundation gap A, and an upper storey overlapping the City Wall.

It is probable that most, or all, of the bastions from Tower Hill to Cripplegate were built in the same way as those just described, and there is evidence to suggest that the western bastions were also similar. In 1806 fragments of Roman monuments were found near Ludgate; “these may have come from a later Roman gate or from the adjoining bastion” (_V.C.H._). Allen says: “At the back of the London Coffee-house, Ludgate Hill, a circular tower and staircase was discovered; and about 3 ft. below the pavement some remains of Roman art were found.” An etching of the stones published by T. Fisher in 1807 describes them as “dug out of the foundations of the wall of the City, a few yards north of Ludgate.” Archer, speaking of an inscribed pedestal, says it was found “in extending the premises at the back of the London Coffee-house. It appeared in a bastion of the City Wall, and was built in with the masonry near some remains of a circular staircase” (_Illust. Family Jour._, _c._ 1850). Now, Horwood’s plan of 1799 shows the back of the Coffee-house adjoining the line of the old wall and extending a long way north—apparently much more than sufficient to overlap the bastion numbered 55 on Mr. Reader’s plan. The Post Office excavations recently made down Ludgate Hill show that the natural ground is here only about 10 ft. below the modern level.

The Camomile Street and All Hallows bastions were about 20 ft. wide and projected about 16 ft. In mediæval days the bastions rose above the parapet walk on the main wall, and each formed a round-ended chamber having loopholes. This is well shown on the Survey of Holy Trinity, Aldgate, 1592, which I published about 1900 in _Middlesex Notes and Queries_. (Several round-fronted bastions are planned as well as Aldgate itself.) The mediæval arrangement, I have no doubt, followed the Roman scheme. The openings in the original bastions would, we may suppose, have been wider than mediæval loops, and have had semicircular arches of brick over them. (See Viollet le Duc’s _Dictionary_, vol. i. p. 333.) The walls and bastions which still exist at Le Mans and Senlis more closely resemble those of Londinium than any others I have seen. At Le Mans a long portion fronting, but some way back from the river Sarthe, has three bastions 60 yds. to 70 yds. apart, round on the front about 20 ft. wide, and 15 in. or 16 in. projection. The curtain is about 30 ft. high, and the bastions rise higher—say, to 45 ft.; they rise sloping for some way from the ground (Fig. 36). The bastions at Senlis are very similar, but some of these have two storeys of large openings, three in each.

FIG. 36.
]

For a long time it was argued that the bastions of the Wall of London were mediæval; then very considerable difference of construction from the City Wall has been alleged. It has been said that their masonry was unlike the other, and that there were no tile bands. We only know with any certainty the lower parts of the bastions now recognised as Roman, and there is no reason for asserting that there were no tile bands in the upper parts. The bastion illustrated by Roach Smith from a sketch by Gough had bands of brick, but in the illustration this bastion appears as square, and this is unlikely (see _Archæol._ lxiii.). It is possible, however, that the form is a misreading of a rough sketch. This, I think, is more likely than the suggestion in _V.C.H._ that it was mediæval. An illustration of a round-fronted bastion near Falcon Square given by Thornbury (_Old and New London_), shows two bands of tile. This seems to be bastion 40 of _V.C.H._, which was about 40 ft. high; “in the upper part was a row of tile-brick, probably due to later patching.” There are also some other references to tiles in bastions, and on the whole I conclude that they probably had tile bands more or less like the wall. Both the bastion just mentioned and that of Gough’s sketch had openings below the upper storey, showing that in these bastions there were chambers below the level of the parapet. So there must have been at Le Mans (Fig. 36) and Senlis. Compare also V. le Duc’s _Dictionary_, vol. i. p. 333.

In an article on the City Walls in the _Journal of the London Society_ (November 1922), Dr. Norman says: “Last summer the remains of another bastion were laid bare not far from the west end of the Church of St. Anne and St. Agnes.” This was “the inner angle bastion” near Aldersgate.

It is not exactly known when the City was protected by walls. Stow says: “It seemeth not to have been walled in the year of our Lord 296, because in that year the Franks easily entered London.” He accepted the legend that “Helen, the mother of Constantine, first enwalled this City.” Camden held the same view, and has a note: “Coins of Helena often found under the walls.”

It is now agreed that the walls were built around a late and extended city, for rubbish pits and burials have been found within the walls. A belt of the former occupied the site of St. Paul’s and the Post Office. It was Roach Smith’s impression that the walls were probably built “after the recovery of the province by Constantine, or even later, when Theodosius restored the towns” (_Archæol. Jour._, 1844).

Mr. Lambert, from planning the find-spots of Roman coins, comes to the conclusion that the wall was not in its later position until the fourth century. The type of walling is especially characteristic of the fourth century. Haverfield has pointed out some earlier cases of the use of bonding tiles, but these seem to be exceptional. (See also what is said of Colchester in _J.R.S._, 1919.) Daremberg and Saglio give 309 as the date of the earliest wall of our kind in Rome. (They illustrate an example from Timgad, in North Africa, which closely resembles the wall of London.) I suggest that a point of evidence may be found in the Constantinian coin, which has a city gate or fortification for device, and the inscription PROVIDENTIAE CAESS, with the mint mark of London (Fig. 37). This device was not invented for London, but I cannot think that at such a time it could have been adopted if Londinium still remained an open city—it would have invited too obvious irony after what had happened in 296. This coin was issued between 320 and 324, and I suggest that it may be accepted as a record of the walling of the City, or, perhaps more probably, the beginning of the works. The coins of Helen mentioned by Camden were issued about this time. In the later half of the fourth-century London acquired the title of Augusta, and this change of style probably followed on the change of status of its having then been completely walled. (I find that Mr. Reg. Smith has already made this same suggestion in _V.C.H._) Sir Arthur Evans has recently called attention to a silver coin of Valentinian the Elder as having in an abbreviated form the monetary stamp of _Londinensis Augusta_. “A group of coins shows that the Mint at London, which had been closed since the time of Constantine, was restored by Valentinian in A.D. 368” (_Proceedings, S. A._, 1915, p. 105). I suggest that this is a probable date for the completion of the river wall. Several of the cities of Gaul were protected by walls at a still later time.

FIG. 37.
]

Many of the carved fragments found in the bastions can be little earlier than the year A.D. 300. The important monuments of which remnants have been found must have been destroyed when the long, wide strip required for the original wall and its ditch was cleared, for the bastions themselves did not go beyond this ground. It seems possible that the big stones were reserved for founding bastions; this is more likely than that distant monuments were destroyed to provide foundation stones.

“To put an end to incessant pillage the Gallo-Roman towns sacrificed their faubourgs, and, retrenching their extent, surrounded themselves with strong walls, which were very often supported on sculptured blocks taken from destroyed edifices. Le Mans, like the towns of Senlis, Tours, Autun, Bourges, Fréjus, etc., girded itself with ramparts flanked with round-fronted towers, of which important remains still exist, especially along the river Sarthe. The _enceinte_ of Le Mans enclosed an area about 500 by 200 metres” (A. Ledru, 1900).

FIG. 38.
]

_Gates._—The excavations of 1903 at the Old Bailey revealed some remnants of the Roman gate on the site of Newgate. The most significant of these was a portion of plinth on the City side, with a return at the south end. This, as shown in _Archæologia_, lix., by Dr. P. Norman, when linked up with earlier discoveries made in 1875, allowed of the recovery of the plan of the gate (Fig. 38). The plinth had been removed from its place before I saw it, but the stones were certainly shaped in Roman days; they had a chamfer 8 in. wide, with a square face of similar width below, and they had been strongly cramped together; one had a “return end,” and clearly came from a corner (A and B). A portion of the western plinth was discovered in 1909 (_Archæol._ lxiii.). The gate, with its towers on either side, had a frontage of about 96 ft.—probably 100 Roman feet, as a Roman foot was about 11·60 in. The space between the towers appears to have been about 35 ft., which is not more than sufficient for two large archways. The great gate at Colchester, which was about 107 ft. wide, had two carriage-ways 17 ft. wide, and two small side openings 6 ft. wide as well (see _J.R.S._, 1919). Enough of the walling was found in 1875 to show that the London gate was of stone bonded with tiles; it was erected on a thick platform of “clay and ragstone,” which raised the plinth about 5 ft. above the plinth of the adjoining City Wall. Fig. 39 is a restoration of the front.

FIG. 39.
]