Introduction: Geological and geographical setting

An Inventory of the Historical Monuments in London, Volume 3, Roman London. Originally published by His Majesty's Stationery Office, London, 1928.

This free content was digitised by double rekeying. All rights reserved.

'Introduction: Geological and geographical setting', in An Inventory of the Historical Monuments in London, Volume 3, Roman London, (London, 1928) pp. 10-18. British History Online [accessed 20 April 2024]

3. The Geological and Geographical Setting.

Attempts have been made on two occasions to establish a permanent metropolis elsewhere than at London; the supremacy of Colchester, however, was short-lived, while Winchester was always a dynastic rather than a national capital. Only in ecclesiastical administration—by the diversion of the archiepiscopal see to Canterbury at the beginning of the 7th century—has the metropolitan authority of London found a permanent rival. With this exception, the concurrence of peculiarly favourable geological and geographical conditions that determined the origin of the city has equally decided her destiny; and an introductory note on the nature of these conditions is, therefore, doubly necessary to any discussion of her early history. For fuller information the reader is referred to the Geological Survey Drift-maps (sheets 256, 257, 270 and 271), issued by the Ordnance Survey, and to the Memoirs on London and the adjacent districts issued since 1921 by the Geological Survey of England and Wales.

The primary importance of the Thames estuary as the main gateway into Britain from the Continent is manifest at most periods from the Bronze Age onwards and requires no emphasis. We may pass, therefore, immediately to the more local problems of the city and its environs. London stands somewhat S. of the centre of the great chalk-basin which extends from southern Hertfordshire on the one side to the North Downs of Surrey and Kent on the other. N. of the Thames, the chalk last emerges in the neighbourhood of Watford, whence it dips south-eastwards beneath an accumulation of Tertiary deposits, to re-appear S. of the river in the neighbourhood of Sutton, Croydon, Farnborough and Dartford. Of the deposits which fill this basin, superficially the most extensive is the London Clay. This is occasionally capped (as at Hampstead) by patches of light Bagshot sand, and overlies alternating sands and clays which come to the surface sporadically S. of the river; but of more importance than these in the present context are the deposits of river-gravels which have here and there been formed upon the Clay and have covered it locally with a clean, dry metalling that was to determine the extent of urban habitation here until the end of the 17th century.

The patches of gravel with which we are particularly concerned are three in number: (1) a large triangular area, with a base extending from the former mouth of the Fleet river at Blackfriars to the alluvium of the Lea at Blackwall, and with its apex at Lower Clapton; (2) a smaller area extending westwards from the Fleet to Westminster; and (3) a large, irregular stretch extending southwards from Southwark to Camberwell. All these regions were occupied in Roman times, but one of them offered special attractions. The south-western corner of the first-mentioned takes the form of two small hills rising to a maximum height of about 50 ft., and divided by the shallow bed of the former Walbrook, which joined the Thames on the W. of the present site of Cannon Street Station. These two hills were both commanding and accessible; but their determining advantage was that they lay opposite to the third of our gravel oases at Southwark. Here is the lowest point in the Thames valley at which two readily traversable areas are thus confronted; here may have been the approximate tidal limit in Roman times; and here, under appropriate conditions, was the obvious focus of land-borne, river-borne and sea-borne traffic.

The appropriate conditions were two-fold. In order to appreciate them it is necessary to reconstruct the general aspect of the district before its natural vegetation was swept away by the hand of man. The tracts of gravel may be supposed to have supported an open tree-growth with light scrub, and with alder and willow along the banks of streams such as the Walbrook or the Fleet. (fn. 1) The surface would drain readily, but ample fresh water would be found at the junction of the gravel with the clay. In short, these gravel islands were from every standpoint suitable for human occupation. On the other hand, beyond them to the N. and N.W. stretched many miles of the heavy London Clay, ill-drained and closely covered with oak-forest, with a dense shrub-layer of hazel, sloe, hawthorn and brambles—a jungle impossible to clear save by maintained and intensive labour, and relieved only by rare patches of heath or light birchwood on the sands at Hampstead or Harrow.

Thus, gripped by forest but easily accessible by water, London looks naturally southwards and seawards. The first condition for a prosperous London is, therefore, maritime trade. In itself, however, this condition is insufficient to explain the growth of the city; for innumerable objects, not infrequently of foreign type, found in the bed of the river and along its banks show that throughout the Bronze and Early Iron Ages the Thames estuary had been already the main inlet for overseas commerce and immigration. But the earlier settlers, concentrating upon the easy river-highway, had had no special reason for staying their progress at the future site of London. Rather had they found their way upstream to the more friendly reaches of the river at Hammersmith and Mortlake, where long stretches of gravel flank the comparatively narrow waterway on both sides. The site of London only came into prominence when a further condition became insistent—when regular distribution towards the hinterland to the N. and the S. became as important as up-river traffic, and when the lowest convenient crossing of the Thames was called into use as the natural point of disembarkation for this overland trade. Such trade, again, implies a strong and wealthy central authority, capable of making and maintaining long roads through the uninhabited forest-lands of the interior. The two historical conditions requisite for developing the natural potentialities of the site were, therefore, regular foreign intercourse and effective centralized government. When these conditions, and with them London, came into being will be discussed in the next section.

It has been suggested by Haverfield and others, though without emphasis, that the southern rather than the northern bank of the Thames may have borne the first Roman settlement. Certain it is that early Roman pottery, together with Roman buildings and burials of various dates, has been found in Southwark. The historical position of these discoveries will be considered later, but the physiographical problems which they raise are appropriate to the present section. Southwark has been included, above, amongst those regions, rendered specially suitable for human habitation by deposits of gravel. Nevertheless, to-day almost the whole of the tract lying between the river and a line drawn roughly E. and W. from Deptford and New Cross to Lambeth—a tract about 4 miles long and 2½ miles deep—is 5 or 6 ft. below Trinity High Water Level. (fn. 2) Within this tract the only area which rises definitely above the tide is at Bermondsey, the name of which proclaims the site to have been formerly an island. Here the Trinity High Water contour encloses a space about one-third of a mile long and half-a-mile wide. Apart from this, two other areas may have been comparatively dry: Horsleydown, between Bermondsey and the river; and St. George's Fields, represented now by St. George's Circus and Borough Road, nearly a mile W. of Bermondsey. It is clear, however, that these limited and scattered refuges are not sufficient to account for the distribution of Roman occupation in London S. of the Thames, and the explanation must be sought elsewhere.

Two explanations have been offered. The first, put forward with a wealth of valuable detail by Mr. Codrington, (fn. 3) is that the Romans built a riverside embankment about 7 miles long and 12 ft. high, extending from Deptford to Vauxhall, and thus excluding the tide from the area in question. Some such embankment is known to have existed at least as early as the 13th century, and there is evidence that the Roman engineers were quite capable of draining marshes and erecting sea-walls. (fn. 4) On the other hand, Mr. F. C. J. Spurrell and others have pointed to a gradual change in the relative level of tide and land-surface along the Thames estuary since Roman times and have maintained that "of banks against the tide in the district below Purfleet there are none surviving of the Roman period, while above that place none or but the slightest ones were needed, and no signs of any can be found." (fn. 5)

Of the two views, the latter is unquestionably correct. At Tilbury the structural remains of Romano-British huts can still be seen in situ on the foreshore at a depth of about 13 ft. below Trinity High Water, (fn. 6) and many other evidences of occupation in this period at a correspondingly low level have been recorded, along both banks of the river, from Westminster downwards. (fn. 7) The gradual lowering of the land-level relatively to sea-level is indeed a well-known phenomenon along the more southerly shores of Britain and the adjacent shores of the Continent, and has been proceeding intermittently since early neolithic times. It is marked by submerged layers of peat, silt, clay, and forest, and has inspired countless legends of "engulfed cities" and the like. One of these sunken forests is noteworthy in the present context. "In the marshes of Long Reach and nearer London, the upper layer of the great mass of peat supported a forest of birch, elm, hazel and yew, with many others including oak. The yew-forest is a remarkable feature—as the yew is intolerant of water and cannot live in salt—yet the yew-forest stretches across the whole marsh at Dartford, Dagenham, Rainham, Erith, and Plumstead (as well as elsewhere)." (fn. 8) The stubs of the trees, in some cases upwards of 18 inches in diameter, were found in situ "on both sides of the river bank in Long Reach" at a depth of about 12½ ft. below Trinity High Water—i.e. at about the same average depth as that of Roman relics from the estuary. No absolute equation of the two is possible on the present evidence, but the extent and level of the sensitive yew-forest at least indicate that the tidal-flow was comparatively restricted in the Thames estuary at a time which is not likely to have been remote from that of the Roman occupation. In other words, the level of high tide must have been upwards of 13 ft. (probably at least 15 ft.) below its present level in relation to the existing land-surface. (fn. 9)

If this datum be applied provisionally to the Southwark area, it at once becomes apparent that the greater part of the original surface stood safely above the former high-tide level. Roman remains are recorded to have occurred there on twenty sites at an average depth of 9½ ft. below Trinity High Water, and only in one instance, where the depth is given as 17 ft., does it exceed 15 ft.—a depth which is itself abnormal. These maxima must have been quite close to the old high-tide level—a fact which would adequately account for the frequent discovery of alluvial deposit and of pile-substructures in association with Roman remains there, but at the same time indicates that there was no necessity for any extensive system of embanking in Roman times. (fn. 10)

One further point arises from these considerations. The submergence of the land-surface since Roman times has not only changed the relative levels of land and tide; it has affected the whole floor of the estuary and led to its gradual encroachment westwards. The action would be progressive submergence, leading to greater tidal flow and this in turn to increased scour and consequent widening and deepening of the channel. Allowing for a 15 ft. submergence since Roman times, tidal action at the site of London Bridge must have been very much less than at present and the conditions much more favourable for a crossing-point. The range of tidal action must have ceased a good deal below its present limit at Teddington, and it is not unlikely that, as in the case of other Roman settlements at the heads of estuaries, Roman London was established at, or only a short distance below, the tidal limit at the time.

To turn now to the topography of the actual site of the City. At the outset it is necessary to bear in mind the fact that the natural surface—normally gravel but occasionally brick-earth or alluvial clay—lies at a depth of from 10 to 30 ft. below that of the existing streets. This difficulty has hitherto prevented any closely accurate survey of the contours of the site, and the drastic levelling-influence of streets and buildings is not easy to appreciate. (fn. 11) The main features, however, appear to have been as follows. At a distance of rather less than three quarters of a mile from each other lay the summits of the two hills (marked now by Leadenhall Market on the E. and St. Paul's Cathedral on the W.) at a height of 50 ft., more or less, above Ordnance Datum. Between them lay the valley of the Walbrook, steep towards its mouth at Dowgate (on the western side of Cannon Street Railway Station) but sloping more gradually from the present site of the Mansion House northwards. To this stream, forming as it did the central landmark of the ancient city, it will be necessary to return. On each side of it the hills fell sharply southwards to the river with an average drop of 1 in 16. A slightly more gradual slope, where Ludgate Hill and Snow Hill now lie, carried the western bluff to the banks of the Fleet or the Turnmill Brook (now Bridge Street and Farringdon Street), which probably received a small tributary from a re-entrant on the site of Smithfield Market. This re-entrant marked the N.W. limit of the city and influenced the line of the town-wall at this point (but see also p. 79). Similarly, the northern stretch of the wall followed approximately the flank of a shallow depression falling slightly from W.N.W. to E.S.E. through Finsbury, but partially interrupted by a small natural or artificial salient at Aldgate and a rather broader one at Bishopsgate. The south-eastern corner of the city was bounded by a further re-entrant dividing Tower Hill from the corresponding knoll on which the Mint now stands. To the E. and N.E. the gravel plateau, at a maximum height of about 40 ft., undulates towards the wide and shallow valley of the Lea.

It has been seen that the natural axis of the Roman site was the Walbrook, and this small stream played so great a part in the subsequent development of the city that a somewhat detailed consideration of its course and physical history is essential to the proper understanding of the subject. The information accumulated by a series of observers, from Tite and Kelsey onwards, and here collated by Mr. A. W. Clapham, is enough to make it unlikely that any subsequent evidence will throw much fresh light on the subject.

The following account, though based upon the evidence contained in the various papers cited at the end of the section, will be found not always to agree with the conclusions arrived at by the observers themselves. This is in part due to the fuller light cast on the subject by later discoveries and in part to the conflict of opinion which must always arise in a discussion of this nature. It can only be suggested that the account here put forward provides the simplest explanation of the facts observed.

The original course of the Walbrook, before the building of the Roman city, has been observed and noted in various places. Its sources consisted of a number of small streams rising to the N. of Hoxton and Shoreditch. Five of these rivulets, according to Tite, still existed in sewers in the district of Finsbury, where they united to form the Walbrook. Between this point and the Thames the united stream flowed at the bottom of a gradually broadening depression, washed out by storm-water in the surface of the natural gravel.

At the northern limit of the city this depression was about 150 ft. wide; farther S., near the Bank of England, it had increased to 250 ft., and at its outfall it was perhaps some 300 ft. wide. The depth of the depression below the banks also increased towards the S., being about 8 ft. at London Wall and over 15 ft. near the Bank of England. In its passage through the city the main stream was fed by at least two considerable tributary channels, one flowing in a little S. of the wall and the other making a junction at the E. end of Poultry. It has often been asserted that the lower part of the Walbrook channel formed a tidal inlet, but whether this was the case in Roman times is an open question. That the broad depression formed by the stream remained without any considerable deposit on its bed and banks before the establishment of the Roman city is proved by the uniform presence of Roman antiquities in the lowest stratum of the deposit which subsequently filled it, and also by the fact that the structure of the town-wall, where it crosses the depression, is accommodated to the fall of the ground in such a way as to prove that the original depression was still largely unencumbered when the wall was built, the culvert being at or near the bottom of the depression.

One curious feature contemporary with the stream-bed in its early state deserves particular notice. On five sites, within and without the. town-wall, large quantities of human skulls have been found, accompanied by very few other human bones. Wherever these discoveries have been scientifically noted, the skulls were found to lie on the actual gravel banks or bed of the channel, and one was partly built over by the town-wall. The presence of the skulls alone, or with only a minimum quantity of other bones, may perhaps imply that they had been deposited along the banks by the action of storm-water, the rounded shape of the skull accounting for its greater mobility as compared to the other bones. This implies a very large deposit of human remains higher up the course of the stream, at a date either preceding or very early in the Roman occupation.

Another explanation is suggested by a passage in Geoffrey of Monmouth [Hist. Britonum, V, 4] describing the capture of the city by Asclepiodotus: the defenders being all killed except one legion, its commander Gallus "surrendered himself and his men to Asclepiodotus, who was disposed to give them quarter; but he was prevented by a body of Venedotians, who rushed upon them, and the same day cut off all their heads upon a brook within the city, which from the name of the commander was afterwards called in the British tongue Nantgallim and in the Saxon, Gallembourne." This stream is identified by Speed with the Walbrook. It is certainly remarkable that a circumstance related by the romancer Geoffrey should be so closely illustrated by an archæological discovery. The place and the circumstances exactly equate and it is only in the date that the account does not fit in with the facts. Is it possible that some remote traditional echo of an actual event had been preserved by Geoffrey's informants and that Boudicca and not Asclepiodotus was responsible for a wholesale execution ?

The establishment of the Roman city involved an immediate alteration in the conditions of the stream-channels within the limits of habitation. The earliest method of confining the stream appears to have been the erection of enclosing walls on the top of the banks of the depression. Remains of these walls have been found lining the depression of the tributary stream near the foot of Coleman Street. This method, even if it were universally used within the city, and of this there is no evidence, can have subsisted only for a short time; for the filling up of the stream-channels began quite early in the Roman period, as is indicated by the date of the objects and coins found in the lowest deposits. The rapidity of accumulation of this deposit near the main bed of the Walbrook is shown by the discoveries of General Pitt-Rivers on a site to the S. of London Wall. Here the deposit, composed almost entirely of black earth, was some 8 ft. thick, the whole belonging to the Roman period, as was proved by the discovery of Roman leather shoes, etc., up to the top. This accumulation can only be accounted for on the supposition that the broad depression of the stream was used by the Romans from the first as a repository for refuse, the deposit being increased and solidified by the surface water drainage from the streets and house-roofs on the adjoining slopes. (fn. 12) That a central channel, probably of quite small dimensions, was kept open is indicated by the discovery of two or more culverts discharging into open ditches, found in the neighbourhood of London Wall and designed initially to carry the water of the various streams under that barrier. (fn. 13) The area thus recovered from the broad stream-valley on both sides was early built over, some of the timber-structures recorded by Pitt-Rivers having piles driven into the original gravel-bed. At a rather later date more substantial buildings were erected, and at numerous points on both sides of the stream remains of tessellated and other pavements have been found, proving that towards the close of the Roman period the stream was confined to a quite narrow channel even as near its mouth as Cannon Street.

Fig. 1. Diagram showing deposits in bed of the Walbrook. From Archæologia LX, by permission.

We have so far considered only the conditions S. of the city-wall; to the N. of this barrier a rather different course of development seems indicated by the investigations of Mr. F. W. Reader on the W. side of Blomfield Street. Here the total Roman deposit was about 8 ft. thick, much the same as at the corresponding site studied by General Pitt-Rivers within the wall, but its composition was quite different. Only the top 3 ft. consisted of black earth, the remainder, down to the original gravelbed, being composed entirely of river sand and silt. This sand would appear to represent the deposit brought down by the stream while its course was still open, though confined, first by the culverts which conducted it under the wall, and secondly by the timber-structures built on the sloping banks of the depression which formed its bed. These constructions would appear to have been erected some time after the first obstruction, caused by the wall and its culverts, for the following reasons:— (a) The piles penetrate into the original gravel to a depth not sufficient of itself to ensure stability; and (b) the enclosures formed by the camp-sheathing were filled with made earth to form a platform and this made earth rested on a deposit of sand and silt about 1½ ft. thick above the gravel; the surface of this deposit must thus have been the surface of the ground when the pile-structures were first built. That this date was hardly later than Trajan is indicated by the consistent evidence of the pottery found with these structures.

Much the same conditions were observed in the shaft sunk by the Society of Antiquaries immediately outside the wall in 1905. Here, however, the silt was a foot less in thickness than in Blomfield Street and was immediately surmounted by a band of black marsh deposit.

It would seem probable that the conditions represented by the deposits of silt and the construction of the pile dwellings continued until well on in the Roman period; that is to say, that the culverts were kept open and were sufficient to deal more or less effectively with the normal flow of water, the silt being presumably the result of storm-water which temporarily choked the culverts. The deposit of black earth or mud on the top of the silt, however, points to a different state of affairs. There is ample evidence that the culverts at last became entirely blocked and that the city-wall was thus transformed into a vast dam holding back the waters of the Walbrook and eventually producing the fen or marsh of Moorfields, which subsisted throughout the Middle Ages. The evidence already adduced points to the beginning of this state of affairs towards the close of the Roman period, and with this the subsequent evidence is entirely in accord. It is clear that once the culverts became permanently blocked the first place of deposit would be the depression, now greatly reduced in depth, of the original stream-bed. The marsh conditions would spread more gradually over the surrounding area, the water eventually finding its way under the foundations of the city wall and producing a not very dissimilar state of affairs on the S. side of that boundary also. The 3 ft. of black earth, containing Roman remains, found by Mr. Reader, represents the first stage of this process—the filling-up of the streamdepression and its immediate neighbourhood. The investigations of Mr. F. Lambert in Finsbury Circus show that the Roman deposit was confined to the surface of the gravel or to a reed-layer 1 ft. thick, both levels being overlaid by the black marsh deposit which contained no Roman remains. This seems conclusive evidence that the great Marsh of Moorfields was not formed, except to a limited extent neighbouring on the town-wall, before the close of the Roman period, more especially as the objects found by Mr. Lambert below the marsh deposit almost all belonged to the later Roman period [Roach Smith in Arch., XXIX, 152; Tite in Cat. Antiq. Roy. Exch., xxxi; Pitt-Rivers in Anthrop. Rev., V (1867), lxxi; Reader in Journ. Roy. Arch. Inst., LX, 137; Norman and Reader in Arch., LX, 169; Lambert in Arch., LXXI, 75.]


  • 1. The areas shown as gravel on the Geological Survey Drift Maps include in fact occasional patches of overlying brick-earth which are too small in extent to appear on the one-inch scale. For example, brick-earth is found on the gravel at St. Paul's (where the Romans appropriately established kilns) and in the neighbourhood of King William Street (see p. 32). These patches would tend to increase locally the density of the natural vegetation; and, similarly, other small local variations in the subsoil which are not indicated on the map would slightly modify the general statements made in the text in regard to the ancient distribution of heath and forest.
  • 2. Trinity High Water is 12½ ft. above Ordnance Datum.
  • 3. T. Codrington, "London South of the Thames," Surrey Arch. Col., XXVIII, 111 ff., with map.
  • 4. Tacitus, Agricola 31, and the Roman centurial stone found in the sea-wall of Monmouthshire, have been cited in this context.
  • 5. Journ. Royal Arch. Institute, XLII, 269 ff.
  • 6. Roy. Com. on Hist. Mons. (Eng.) Essex S.E., 38–9. For the depth, I am indebted to Mr. H. J. Melliss, who, in connection with the building of a jetty, made a borehole within a few feet of the huts in 1923. The huts may well have extended to a yet greater depth, since erosion has obliterated the lower limit of the settlement.
  • 7. See Spurrell, loc. cit., 274 ff. Compare also the submergence of some of the Essex "red-hills." See Reader and Wilmer, Report of the Red Hills Exploration Com. (Society of Antiquaries of London), 3.
  • 8. Spurrell, loc. cit., 270. This statement contains an inaccuracy in that peat does not actually "support" forest trees.
  • 9. No discussion need be attempted here of the various factors (such as subsidence, the deposit of silt, the shrinkage of reclaimed land flanking the river, the possibility that the concentration of the tidal wave between river-walls may have tended to raise the high-tide level, etc.), which may have contributed to this result. Only, in regard to the shrinkage of reclaimed land, it is worth noting that between 1850 and 1906 the construction of low-level sewers in Southwark, by draining the subsoil water, lowered the surface by more than two feet over a considerable area.—Codrington. loc. cit., 161.
  • 10. The entry in the Anglo-Saxon Chronicle under 1114 has been cited as indicative of a comparatively dry Southwark as late as the 12th century: "In this year also was so great an ebb-tide everywhere in one day, as no man remembered before, so that men went riding or walking over the Thames eastward of London Bridge."
  • 11. A useful sketch-map of the present contours of the City is included by Mrs. H. Ormsby in London on the Thames (1924), which also contains a concise account of the physical features of the site and its environs.
  • 12. In some places there may have been a definite levelling up of the stream-depression. Thus on the Bank of England site (see p. 106) an unstratified deposit of black occupation-earth, 8 ft. thick, was observed overlying the earliest remains of Roman occupation; it contained a certain amount of pottery, none of it necessarily of later date than the 1st century. This deposit was apparently of the same nature as that observed by Pitt-Rivers and there is at least a probability that it was the result of the destruction of the City of Boudicca.
  • 13. These culverts (see p. 87–9) were carried, in the form of built sewers, to an indeterminate distance (20 ft. or more) beyond the wall, both within and without the city. The discoveries on the main channel of the brook perhaps indicated that, at a later date, it became necessary to pierce a second culvert, five feet above the original opening.