6. Discussion

6.1 Cementation furnaces

The records of the two excavated cementation furnaces add valuable detail to a corpus of excavated Sheffield cementation furnaces. These include those excavated during the Inner Relief Road project, at Green Lane (Elmet Archaeology 2014) and adjacent to the preserved Doncaster Street example (Wessex Archaeology 2018b).

The two furnaces were generally preserved below a level just above the base of the refractory chambers. Below this level, preservation had been impacted by a series of later concrete footings and associated walls, however the furnaces were easily readable and enough detail was preserved across the two furnaces for a complete suite of below-ground features and structures to be observed. Preservation was better than at comparable sites.

As stated in the metallurgical report, the cementation furnaces at Hollis Croft are the first time a furnace has been excavated in Sheffield with parts of the sides of the chests preserved. This has confirmed the presumption that the crozzle lining of the refractory chambers continued up the interior sides of the chests.

The excavation at Hollis Croft was also the first time that a longitudinal section of the base of furnace chests has been seen. It has been revealed that the base of the chests were made up from several rectangular slabs generally 0.22m across, rather than two or three larger slabs. The use of smaller slabs created more joints, which would have had the unwanted potential to allow air into the chests during firing. These joints may have been sealed with a similar clay/slag-like crozzle as the main refractory chamber.

The same longitudinal section revealed that the Hollis Croft cementation furnace chests had been rebuilt on top of earlier chests. The earlier chests may have been replaced due to inadequate provision of flues. The refractory chambers had a limited lifespan and required periodic replacement (Belford 1998).

Another unusual feature of the furnace chests at Hollis Croft were the faint longitudinal grooves and ridges on the surface of the crozzle layer in the base of the refractory chambers. The grooves between the ridges were the same width as the type of iron bars that were normally converted in cementation furnaces, so it seems likely that the contents of the chest had sunk into the crozzle layer, either during loading or possibly when the furnace was being fired. To the authors' knowledge, the ridges are not something that has previously been seen in furnaces excavated in Sheffield.

Analysis of the slag-like layer at the base of the furnace chests has supported previous analyses from other cementation furnaces in Sheffield in suggesting that the raw material used to create the layer was a specific by-product (known as 'wheelswarf') of the cutlery and edge tool grinding trade in Sheffield, rather than a slag by-product of the steelmaking process, and was used by steelmakers to help keep the chest airtight during firing.

Dating

Cementation furnaces are recorded in documentary sources at Hollis Croft in the 18th century (Belford 1998). A structure which is likely a cementation furnace is visible on early mapping (Fairbank plan of 1781; Figure 20), however no archaeological remains of this furnace were identified. Photographs of cementation furnaces at Hollis Croft also exist (Plate 1, Plate 2 and Plate 3); these furnaces are thought to have been in the east of the development area and must be later than the 1853 map. The excavated remains in Area E/F relate to none of the above furnaces.

The excavated furnaces correlate with two dashed circles depicted on the first Ordnance Survey map of 1853 (Figure 30). The identification of the circles on the maps with the furnaces would suggest a date for construction prior to 1853.

Dating of the cementation furnaces is fairly reliably available from the recovered pottery and clay tobacco pipe assemblage. Both clay pipes and pottery were present within the red brick outer structure of the chest of the north cementation furnace (2031). Two sherds of Sponge-Printed Ware were secondarily burnt and date from later than c.1840. Two sherds of TP Whiteware (a spout and a handle terminal) date from the mid- to late 19th century. The clay pipe from this context (2031) comprised two plain stems which have been given as probably deposited between 1800 and 1850. Dating of plain stem is unreliable on a small sample size (here the sample size is two), and is particularly problematic for 19th century pipes. The early- to mid-19th century date of these stems should be treated with caution but does at least support the general picture given by the pottery, which on a sample size of four suggests a mid- to late 19th-century date.

The gaps between the outer chimneys and outer structures of the chests were filled with a variety of sand and clay deposits. These deposits exhibit heat transformation due to the operation of the furnaces and had therefore likely been deposited during the construction of the furnaces. Deposit 2011 contained pottery dating from the mid- to late 19th century alongside other sherds dating back to the 17th century. A clay tobacco pipe bowl found in deposit 2011 was probably deposited between 1870 and 1920, and a pipe from deposit 2016 was probably deposited between 1830 and 1860. Deposit 2017 contained a sherd of 19th-century or early-20th-century pottery. Ignoring the clay pipe bowl from 2011 (deposited 1870–1920), the rest of the evidence is consistent and suggests a mid-19th-century date for the construction of the cementation furnaces, probably between 1840 (on the basis of a single pottery sherd) and 1853 (on the basis of identification of the furnaces with dashed circles on the Ordnance Survey map).

The materials used in the construction of the furnaces (including handmade red brick and lime mortar) are consistent with this mid-19th-century date. It is likely that the furnaces were constructed when Burgin and Wells took over the plot in the early 1850s or shortly before.

Comparison of the two furnaces

The two furnaces were almost identical as might be expected of two contemporary adjacent structures. The contemporaneity of the furnaces is most definitively demonstrated by the central wall which was keyed into, and formed part of, the outer conical chimneys of both furnaces. Identical materials and designs were used for both furnaces. Some differences in the level of heat transformation of firebrick structures were noted; these differences are likely due to minor operational differences in the flow of hot gases through the furnace flues. Two stoke hole doors were recorded and were noted to be of different sizes; it is possible that one or other of the doors was replaced during the operational life of the furnace. Other than these minor differences the furnaces were remarkably uniform, suggesting that they were built at a time when the design of cementation furnaces had been standardised and was well understood.

Phased development and repair

The cementation furnaces generally appeared to have been largely constructed in a single phase of activity. However, at least one of the refractory chambers had been rebuilt. The earliest phase of the chests consisted of a refractory chamber with no underlying ganister flues, although it is possible that the underlying firebrick structure may have been intended to be penetrated by hot gases. This lack of flues is curious in an otherwise well-designed furnace and may have been the reason for the reconstruction of the chests. The refractory chambers of cementation furnaces had a limited lifespan and it is possible that they were rebuilt more frequently than the furnace superstructures. The later phase chests were constructed directly on top of the crozzle base of the earlier phase and incorporated a more typical ganister flue system resembling examples recorded elsewhere. Evidence for the rebuilding of the chests was only observed in the most southerly of the four chest structures, however it is assumed that the other chests were all identical and that each contained relicts of the earlier phase.

It is likely that the majority of structures associated with the furnaces were constructed during the earlier phase of activity, and that only the inner, upper chest structures were rebuilt. The dating evidence outlined above therefore relates to the earlier phase, with the later rebuild of the refractory chambers post-dating this mid- to late 19th-century date. No dating evidence directly relating to the rebuilt chests was identified.

A remnant of a firebrick wall (2103) to the north of the cementation furnaces likely represents a fragment of some former furnace, boiler or a structure related to some other hot process, and may be contemporary with the cementation furnaces. It is possible that it is an element of a further lost cementation furnace, although the presence of early straight wall 2163 tied in to the north-east corner of the cementation furnace structures suggests that there was never an additional contemporary furnace immediately north of the excavated pair. Similarly, the presence of undisturbed natural close to the south-west corner of the furnaces precludes the presence of a further furnace immediately to the south. A ganister 'sleeper' seen as rubble (2084) was likely derived from a demolished cementation furnace. This may have derived from an old chest removed from the two recorded furnaces or may have originated in some furnace at a distance from the recorded pair. Although photographic evidence (Plate 1, Plate 2 and Plate 3) demonstrates that there were additional (probably later) cementation furnaces at Hollis Croft, the excavated pair of furnaces in Area E/F stood together without immediate neighbours.

Decommissioning

Access to the north-west stoke hole of the cementation furnaces from the cellar network was blocked by the construction of an unmortared sandstone wall (2048). The cementation furnace cellars had been linked in to the later cellars immediately to the north and it is likely that the cellars continued to be used after the end of the furnaces' lives. The ash pits had also been intentionally decommissioned with the introduction of a sandstone block (3010) blocking the south-east stoke hole. Two sherds from a child's alphabet mug were also present in upper fill 3008 of the ash pit. The deposition of these sherds may have been meaningful.

6.2 Crucible furnaces

Furnace 1746–1750

A limited view was obtained of two crucible furnaces. The better studied of these (contexts 1746–1750) was poorly preserved, with only a single ash pit surviving from what was presumably a fully-developed row. The ash pit was identical in form to other recorded examples such as at Hoyle Street (Powell 2014) and Blonk Street (Wessex Archaeology 2012), and can therefore be considered to be an example of the well-developed form of furnace most typically associated with the crucible steel process.

Furnace 1757–1764

The second crucible furnace (contexts 1757–1764) was better preserved but was identified only during the late stage watching brief and could not be investigated thoroughly. It is possible that elements of this furnace will survive construction on site and may be preserved in situ. Two charging holes were recorded, but it is certain that there were more, and likely that the lower structures of the furnace were also well preserved. The two charging holes exhibited strong heat transformation indicating that the furnace had undergone sustained and intensive use. Perhaps most interesting were details relating to the decommissioning of the furnace. These included the blocking up of the ash pits (observed by suspending a camera in the charging holes) and the insertion of a metal plate (and possibly a brick wall) in front of the furnace. At Blonk Street (Wessex Archaeology 2012) a final firing with special metals may have taken place; it is possible that behaviour surrounding decommissioning of furnaces took on a ritualised aspect linked to steelmaking secrets (e.g. fluxes, crucible manufacture) and to the status of the furnace at the heart of the steel works.

Dating

Close dating of the crucible furnaces is not possible. The available historic maps (not reproduced) are not detailed enough in these areas to show the furnaces, and the Goad Fire Insurance Plan, so useful in other areas of the city, does not cover the site. The materials used in the furnaces (handmade bricks, firebricks and lime mortar) are consistent with a broadly 19th- (or 18th-) century date. The decommissioning of furnace 1746–1750 is linked to two sherds of pottery: a porcelain cup or mug of broadly 19th-century date and a TP Whiteware plate rim dating to the mid- to late 19th-century. The decommissioning of the crucible furnace must be later than the date of manufacture these sherds.

Workflow

The cementation furnaces were part of the Burgin and Wells works. Documentary evidence attests to the presence of rolling mills at the works contemporary with the cementation furnaces. The workflow of the 'Sheffield Method' was to convert iron to steel in a cementation furnace, to refine the steel in a crucible furnace and then work the steel, for example in a rolling mill. It is almost unthinkable that a works with a cementation furnace and rolling mill would not also have a crucible furnace. Later directory entries for Burgin and Wells are explicit that steel refining occurred at Hollis Croft. The presence of a crucible furnace on the Burgin and Wells plot contemporary with the recorded cementation furnaces is therefore almost certain. The location of the crucible furnace and of the rolling mill are unknown.

Similarly, a three-step process including conversion, refining and manufacturing of machine knives and other products can be assumed at the adjacent Fearnehough works. The two identified crucible furnaces were both on this plot. The location of a cementation furnace feeding these crucible furnaces is unknown, as is the nature and arrangement of machine knife manufacture.

Although it is possible that unrefined steel was passed from one works to another, it is more likely that workflow was internal to each works and there is no reason to suspect that blister steel from the Area E/F cementation furnaces was refined in the Area D crucible furnaces.

Later use

The ash pits of both crucible furnaces would have been accessed by cellars. The walls of the crucible cellar associated with furnace 1746–1750 were recorded. It is likely that the other crucible furnace (1757–1764) was associated with nearby cellars 1734, 1735 etc. The continued importance of the cellars through time was evidenced by remodelling using very modern bricks and cement (1740, 1742, 1753 and 1754). It is tantalising to realise that these crucible cellars were accessible only a few decades ago, cold though the furnaces were by that point. The personal anecdotes and recollections of the Jewitt family (Footprint Tools) heard during the course of this work underline the recent date of the archaeology. A great loss of information sometimes occurs only shortly before archaeologists arrive on site.

6.3 Flue system

Immediately north of the cementation furnaces was a flue system covering an area of almost 12m by around 6.5m (similar in size to each of the cementation furnaces). This flue system belonged to a later phase of construction than the furnaces, and was generally bonded with black ash mortar compared to the furnaces' lime mortar. The flues appear to post-date the Ordnance Survey map of 1853 (Figure 30) although they cannot be closely dated. The flues are probably related to a chimney depicted in 1890 (Figure 31). Preservation of these ground-level flues was poor; generally only the bases of the flues survived, and those only in certain areas.

The fragmentary remains appear to comprise three separate flues converging on a concrete base (2172). From this base, the main widest flue (0.6m internal width compared to 0.5m for the other flues) ran west, before turning north and then east before it was lost. The surviving end of the flue points at an unmortared firebrick surface (2095) which was present in the location where the pattern of cellar vaults might have continued. This firebrick surface must have been of some importance for the designer of the building to have sacrificed a whole cellar for it. Returning to concrete base 2172, two flues ran away to the east. The northerly of these flues contained soot and overlay heat affected natural and is therefore interpreted as having carried hot exhaust gases. This flue turned back around and returned to the concrete base (2172). The remaining flue also overlaid heat-transformed natural, but had a clean interior, suggesting that the hot gases it contained were clean. This flue curved away from the concrete base to the north-east, passing out of the area of excavation. Both of these subsequent flues contained 'guillotine'-style ferrous doors that could be lowered to cut off flow. It is likely that the exaggerated arrangement of the flues was designed to place these two doors in the same location so that they could be controlled from a single position.

Evidence relating to these flues is slight and interpretation cannot be conclusive. Complexes of flues were required for a variety of purposes. These flues may have formed part of a gas regenerator, designed to take exhaust gases and process them for re-use as fuel. The cycling of exhaust gases around concrete pad 2172 may support identification of the structures as part of a gas regenerator. Alternatively, a somewhat similar arrangement of flue was interpreted as the base of a brick kiln at Kelham Riverside (Dransfield 2016 [PDF]). It is possible that the Hollis Street flues relate to a later form of steel furnace such as a Bessemer converter (Dinah Saich pers comm).

6.4 Summary of other structures including domestic housing and public houses

Domestic courts were targeted for excavation. In Area A, an excavated probable toilet complex was depicted on the first 1853 Ordnance Survey map (Figure 5). A series of probable Area I represented several phases of development which can be related to a fair degree with different phases of development depicted on historic maps (Figures 43–45).

Worker's housing was directly excavated at the rear of The Orange Branch public house in Area K and in watching brief Test Pit 18. In Test Pit 18, the rear wall of a range of back-to-backs had somewhat elaborate entrances suggesting there may have been a change in level between the interior and exterior of the houses. The archaeological remains of these small worker's houses comprised walls, thresholds and surfaces. No staircases or cellars associated with dwellings were encountered.

Two public houses were investigated. An outbuilding and exterior courtyard divisions associated with The Cock public house were recorded, along with a soakaway or rubbish pit (1033). In Area K, excavated cellars and exterior surfaces related to the former public house The Orange Branch. The cellars were well-preserved and a sequence of repair, redevelopment and blockage attested to alterations and use of the cellars over time. Later structures were recorded to the north of The Orange Branch remains.

In Area D, a complex of brick walls dividing small areas (below 1m²) defies secure interpretation, but may represent another range of toilets. These structures can be shown by artefacts in sealed demolition layers to have gone out of use by the mid-19th century. The area was later surfaced with sandstone setts.

During the final watching brief, cellars and other structures were recorded in several locations across the site, indicating the potential for preservation of further structures which may (or may not) survive development. The majority of these structures were associated with the Thomas R. Ellin/Footprint Works. A wall containing sockets for floor joists in Area A and a cellar recorded in Test Pit 23 were probably associated with the former Toledo Works to the west.

6.5 Relationship of domestic and industrial areas

These houses, yard and public houses comprise the domestic structural remains of the communities that supplied workers to the various works that occupied the site. The inhabitants of the back-to-back housing seen in Test Pit 18 would probably have worked at the associated British Works (Electro Plating) and some of the drinkers that frequented The Cock and The Orange Branch would have spent the wages they had earnt working in the works, including at the recorded furnaces. Although the technical details of the furnaces are intrinsically interesting, a fuller picture of life and community at Hollis Croft can be built by supplementing the industrial results with the domestic. The close proximity of residential buildings to works and furnaces speaks of lives defined by work and lived in close association with steelmaking. See the catalogue of the census records from 1851 (Appendix 10 in archive report [PDF]) supplemented with almost contemporary directory entries.

It is known elsewhere that 'cottage industry' (industrial processes undertaken in a domestic setting) formed part of the workflow of the great workshop of Sheffield. One example is bone knife handle manufacture which was sometimes undertaken in cellars, sheds or elsewhere at home as an additional supplementary or full income. Such activity underlines the interconnectedness of the industrial and domestic life of post-medieval Sheffield. There is no direct archaeological evidence to support 'cottage industry' at Hollis Croft, although it is probable that such activity took place. Some evidence of cottage industry can be seen in the activities of female dressmakers recorded in the 1851 census. This activity was likely undertaken at home and appears to generally have formed a supplementary income earned by a wife or daughter.

6.6 Conclusion

Documentary research has shown that development of the site began earlier than expected, with workshops and steel furnaces present on the site at least by the mid-18th century. Remains from the 18th century were fragmentary, and primarily comprised dwellings and pits in Areas A and B and The Orange Branch public house in Area K. Small fragments of wall probably survived from John Kenyon's 18th-century works. The impact of 18th-century activity at Hollis Croft was perhaps best preserved in the outline of the various plots.

The most significant result was the remains of two well-preserved mid-19th-century cementation furnaces. The cementation furnaces were built in a uniform style indicating that they followed a well-developed design at a time when cementation furnaces were well understood. The refractory chambers ('chests') of the furnaces had been replaced; regular maintenance of the refractory chambers was a feature of the operation of cementation furnaces. Details of the furnaces were recorded, including the stoke hole entrance doors and the arrangement of 'fire bars' upon which fires were set in the underlying ash pits of the furnaces. It is likely that the cementation furnaces were constructed by Burgin and Wells when they took over the John Kenyon plot (later occupied by Footprint Tools) in the early 1850s or slightly before.

Metallurgical analysis supported the view that the refractory lining of the chests (the 'crozzle') was derived from 'wheelswarf' produced by edge tool grinding. For the first time, it was confirmed that this crozzle extended up the interior sides of the refractory chamber. A further new observation is that of the impression of the ferrous bars in the surface of the crozzle layer.

Two crucible furnaces were identified, although one was badly preserved and the other was only identified under watching brief conditions at a late stage in the works. The crucible furnaces could not be closely dated. The crucible furnaces were part of separate works (W. Fearnehough Ltd.) and there is no evidence to relate them to the cementation furnaces. Nonetheless, the presence of both types of furnace in close proximity underlines the interconnectedness of industrial trades in the great workshop of Sheffield.

To the north of the cementation furnaces was an area of slightly later development characterised by the use of black ash mortar rather than lime mortar. This area included extensive cellars supporting a network of flues. Speculatively, the flues may be related to a gas regenerator, to later methods of steelmaking or to some other activity.

Further industrial activity was recorded in Areas G and H in an area of expansion of the Thomas R. Ellin/Footprint Tools works. Area H contained the remains of flues and a machine base. Structures in Area I to the north of Hollis Croft were enigmatic but may be industrial in nature.

Other areas relating to worker's housing and public houses were investigated. Remains in these areas were generally limited to walls, surfaces and drains although there was sometimes good correlation with historic maps allowing for the identification of some structures as outbuildings and others as more substantial developments. Residential properties were recorded in close association with the various works, and attests to lives closely intertwined with the work of the steel industry.

Acknowledgements

Wessex Archaeology is grateful to Johnson Associates (UK) Ltd who commissioned the work on behalf of GL Europe RE2 Holdings SARL and Newmark Developments (Watkin Jones Group). Gavin Johnson (Johnson Associates Ltd), David Elias (Johnson Associates Ltd), Stephen Rugby (BSRE Group) and Michael Edgar (DLP Consultants) are thanked for their assistance throughout the project so are Peter Garratt, Mike Flynn and Mike Roberts from Watkins Jones Group who provided on-site support. The assistance of Dinah Saich, principal archaeologist for South Yorkshire Archaeology Service, is gratefully acknowledged. The cooperation during the watching brief of the client's sub-contractors Gill Demolition and 4-D Construction was appreciated. Thanks are extended to the Jewitt family (Footprint Sheffield Limited), particularly Penny Jewitt, for their invaluable knowledge of the 20th-century Hollis Croft, steel production and for making the Footprint Tools archive available.

Ashley Tuck and Milica Rajic wrote the text with Sam Bromage and Emma Carter. Illustrations were by Ian Atkins incorporating work by Chris Breeden, Joanna Debska and Jack Fox Laverick. Specialist reports were provided by C.G. Cumberpatch (pottery), Dr S.D. White and Dr D.A. Higgins (clay tobacco pipe), Dr R. Mackenzie (metallurgical residues, crucibles and metal items), Dr P.S. Quinn (petrographic and SEM-EDS analysis of furnace chest residue and crozzle), L. Higbee (animal bone),Q. Mould (leather), L. Mepham (all other finds), I. López-Dóriga (environmental samples) and Ashley Tuck (archive research).

Fieldwork was directed by Emma Carter and undertaken by Ifigeneia Klopa, Jamal Bingham, Philip Maier, Michael Howarth, Amy Derrick, Adam Fraser, Gwen Naylor, Stuart Pierson, Ben Bazeley, James Wright, Ashley Tuck, Otis Gilbert, Owen Jenkins, Daniel Webster, Justyna Dekiert, Max Higgins, Andy Swann, Ciaran O'Neill, Andrea Goodinson and Caroline Thornhill. The samples were processed by Liz Chambers, Stavroula Fouriki, Elizabeth Tooke and Dora Olah. The flots were sorted by Liz Chambers and Nicki Mulhall and assessed by Inés López-Dóriga. The project was managed for Wessex Archaeology by Milica Rajic and Richard O'Neill.