The principal clue to the location of the uphill fortress underlying the later colonia (Chapter 7 below) was the discovery of early pottery and artefacts, some of military association, in the area of Westgate and in particular at the Water Tower built in 1910 (ON 237, Webster 1949). Webster was able to identify the line of the fortress’ northern and western defences during excavations in the 1940s (Fig. 6.6), and Thompson (1956) and Petch (1960) subsequently confirmed that eastern and southern lines also lay beneath those of the colonia.
Little progress could be reported, however, in eluci-dating the internal arrangements – except close to the northern part of the western rampart, where Webster was fortunate to find legionary-period deposits close to the modern ground surface. Generally the fortress’
slight remains are deeply buried, difficult of access and in places already destroyed. Whitwell’s exca-vations close to the Bailgate Methodist Church in 1967–
8 revealed a limited number of wall trenches for timber structures, but not sufficiently extensive or well-preserved to identify their functions (ON 261 – unpublished but noted in Whitwell 1970, 21). Similarly slight hints of early buildings were also noted by Petch during work on the nearby colonia baths (CP 56), a site
which yielded much early pottery – presumably from rubbish pits and demolition deposits of the fortress structures. Remains of timber buildings on the site suggest that the fortress baths could not have covered quite the same area, and baths did occupy different locations in some fortresses – as, for instance, at Exeter.
The relationship between baths and other fortress structures and their replacements or equivalents in the colonia period needs further exploration. The position of the colonia baths at Lincoln is, however, similar to those at the 9th Legion’s fortress at York (Ottaway 1993, 31–3, fig.9), and to those at Caerleon (Zienkiewicz 1986). At Exeter and Usk, the baths were situated in the range to the rear of the principia (Henderson 1991;
Manning and Scott 1989, 169). It was extremely rare at this period for legionary fortresses to have extra-mural baths, in contrast to auxiliary forts, where such a location was normal (Johnson 1983). The unfinished Flavian fortress at Inchtuthil is one exception; here a baths-suite was provided in the officers’ temporary compound (Pitts and St Joseph 1985, 215–8). The scale of the building at Inchtuthil suggests a ‘restricted clientele’, and it may have been intended to construct a full-scale balneum in due course. Alternatively, the extra-mural location at Inchtuthil could have been connected with the problem of water-supply, and the same may have been the case at Lincoln.
The space available inside the defences identified by Webster, Thompson and Petch measured c.440m east–west by c.360m north–south. The postulated plan would allow for barrack blocks, plus the width of an adjacent street, to measure up to 300 Roman feet, although many contemporary examples were shorter (Maxfield 1986, 63). Two cohorts could be accom-modated to the north and south of the via praetoria in the blocks closest to both west and east gates, giving a total of eight cohorts. The first cohort – not yet double at this date (Frere 1980, 58) – might occupy some of the area to the south of the principia, and possibly some of the space to the north of it, fronting on to the via principalis. Part of what may have been the east–west street north of the principia was noted south of the famous length of standing Roman masonry known as the Mint Wall in 1979 (WB 80). There would then still be room for the final, tenth cohort. But the arrangement could be more complex, as at Exeter, where the barracks were shorter in length (c.200 feet), inside a fortress that was more elongated in shape than Lincoln, and for which a surveyor’s blue-print has been proposed (Bidwell 1997, 32, fig. 16). Two cavalry alae also seem to have been fitted in at Exeter, and it is quite possible that in these early years of the Conquest the garrisoning of all bases had to be flexible. At Usk for example, there were fabricae (workshops), rather than first cohort barracks, adjacent to the principia (Manning and Scott 1989, 166–70).
Another metrological approach to the planning of the Lincoln fortress has been suggested, involving the use of proportions (Jones 1975, 54–60) and square Fig. 6.6. Section through the Roman legionary ditch and, to
its left (east), the front palisade trench for the rampart. A view (from the north-west) of the excavations undertaken in Westgate in 1945–6 by Graham Webster – the photograph was taken by I.A. Richmond (photo and copyright Lincolnshire County Council, Lincolnshire Museum Service).
roots, based on principles of geometry exemplified in the mitre square found at Canterbury (Ball and Ball 1988). By this analysis, the praetentura (the area east of the via principalis) at Lincoln would be of the proportion 1:2 (or the square root of 4), whilst the retentura would be 1:1.6216 (probably correctly 1.618, or the ‘golden section’). Evans (1994) has also dis-cussed military building techniques, with special reference to the application of modules, but Lincoln is too poorly understood as yet to test this hypothesis.
Most recently there has been an impressive attempt to estimate the quantities of materials and manpower required in order to build a fortress (Shirley 2001).
While it is accepted that precision is impossible, this careful analysis confirms that the Lincoln fortress’
construction would have involved a good proportion of the Legion for a few years. In addition, a large team would be needed to supply them with building materials and with food.
The discovery and identification of the principia (below) indicates that the via principalis ran north–
south (roughly along the line of modern Bailgate/
Steep Hill), so that the legionary gates would have underlain those of the colonia. The remains of the east gate (RENO 76, Thompson and Whitwell 1973) would then represent those of the porta praetoria. Certainly this gate was a double one, while the west gate to the rear of the principia, the porta decumana, was only a single carriageway, at least in the colonia period (Ibid.,194–200). This seems to make perfect sense as far as it goes, but the absence of evidence for other identifiable structures or definite streets except those adjacent to the defences means that we can establish little else of the internal layout. The area covered by the fortress, at c.17 hectares (c.43 acres) is only some 80% of the normal size of later examples, but it is not clear whether it housed a full legion, or only part, since none of the barrack blocks has been investigated to any large extent – it may be that there were fewer ancillary buildings (Manning and Scott 1989, 161). A reconstruction (Fig. 6.7) can be attempted, however, using standard measurements as proposed by Crum-my (1985; 1988), even though the pitfalls of this approach have been spelt out by others (Millett 1982).
It seems probable that the fortifications and the streets were provided early in the construction process, while the soldiers were in temporary accommodation, either within the area of the fortress, or outside it. To the west, evidence of early structures in the grounds of the Lawn (L 86) might represent an earlier base rather than extra-mural occupation contemporary with the fortress.
More details of both the defences and the internal arrangements have been discovered since 1970. Infor-mation on the legionary fortifications up to 1979 – notably at Westgate School in 1973 – has already been collated (W 73, Jones 1980). The rampart was timber-fronted, although a turf or clay revetment may have been intended originally, and only a single ditch was
provided. The use of timber fronts was comparatively rare in Britain and its use might have been occasioned by the friable nature of the soil, but it is equally possible that those who chose to use it had previously used timber in the Rhineland (Jones 1975, 82–8). Towers projecting beyond the original line were added subse-quently (Jones 1980, 48–9). Subsequent work on the legionary defences took place at East Bight in 1980–81 (EB 80) and Chapel Lane in 1985 (CL 85), and the results from these sites are described elsewhere (Steane et al.
2003). In every case, although several sites produced quantities of pottery from the Neronian period, the amount of new information on building layout was modest. We can at least now confirm the presence of so-called ‘rampart-buildings’, between the rampart and the via sagularis on the line of the northern defences at North Row – evidenced here by waste dumps (Jones 1980, 30–1) – and at East Bight in 1964–6 (Whitwell 1980, 6–9) and in 1980–1 (EB 80). In some cases, these may have been ‘cookhouses’, possibly including bread-ovens (Marvell 1996, 71–3). At East Bight (EB 80), the 1980–1 excavations produced metalworking refuse from associated deposits, whilst at the earlier site (EB 66) copper alloy fragments were discovered, although these have yet to be analysed. It is possible that this rubbish was derived from a workshop undertaking repairs to metal equipment.
At both the Westgate 1973 and East Bight 1980–1 sites, the excavations extended inside the via sagularis into the fringes of the adjacent buildings, presumably barracks. The evidence from the Westgate site (W 73) appears to indicate up to three rooms of a block running east–west, and presumably therefore the centurion’s quarters – although the function of none of the individual rooms is clear (Jones 1980, 29–30, fig. 37).
Some good quality glass of 1st-century date came from this site, partly from residual contexts, it must be admitted, but perhaps representing something of the centurion’s lifestyle (Hoffman 1995; Price et al. forth-coming). At East Bight, building construction tech-niques were similar, but it was impossible to determine if the structures ran east–west along the intervallum road or north–south. There were probably two phases of timber building here, both presumed to belong to the military occupation since they were sealed by what appears to be the military demolition dump. The dump contained a quantity of early pottery as well as many fragments of military equipment, including objects associated with cavalry: its function may have been for recycling copper alloy waste. The most notable object was a dagger-scabbard, with decorated panels of silver inlay (Scott 1985; Webster 1985b) (Fig. 6.8). The group includes the largest collection of armour from any site in the city.
Hints of other fortress structures came from a number of the other sites investigated since 1970. At Chapel Lane, there were two successive phases of timber buildings, with differences in internal arrange-ments, although both might still have been barracks
(CL 85). The use of posts not set in wall trenches in its first phase may indicate a verandah. The earlier structure here was dismantled, the later burnt, per-haps at the end of the legionary occupation. Close to Chapel Lane, at West Bight in 1976 (WB 76), the demolition debris beneath the make-up for a colonia-period building included some wattle and daub, rendered prior to being given a plaster surface. It was associated with much 1st-century pottery. Remains of early stone buildings in similar locations to the legionary baths at Exeter and Usk were too slight to suggest a structure of the scale of a bath-house, and may actually represent the first colonia phase there.
Similar demolition material came from nearby (WB 80) and the other side of the Mint Wall (MW 79) to its south, all three sites lying immediately north and
north-west of the principia and likely to represent, therefore, structures other than barracks.
With the exception of the principia, the constructional details of the legionary fortress buildings at Lincoln appear to indicate two different types of construction;
continuous wall-trenches, or intermittent postholes (possibly indicating different functions), and white-painted wattle and daub walls. We cannot yet tell if the ground had to be cleared of trees, turf or other vegetation before building could begin. Several sites show traces of rebuilding, while repairs to the rampart front were noted at East Bight, and at Cuthberts Yard, to the north of the Westgate School site. The evidence for demolition, in some places involving fires, at the end of the fortress’ life seems more definite (Steane et al. 2003).
Fig. 6.7. Reconstruction of layout of Neronian fortress (sources, Jones 1988 and others – drawn by Dave Watt, copyright English Heritage).