The Victorian pumping station was a product of the combined disciplines of engineering both civil and mechanical, and architecture. In order to understand the architectural form which these buildings took, it is essential to have some knowledge of the engineering problems to which they had to provide the solution.

Throughout the nineteenth century the steam engine was the only reliable source of power, and when large quantities of water (or sewage) had to be pumped, the type of engine almost universally adopted was the beam engine. Although self-contained beam engines could be built, in which all the working forces were contained within a rigid metal frame, these were only suitable for applications with a fairly low power requirement: larger engines would have required a frame of totally unmanageable proportions, and these were therefore ‘house-built’. As the name implies, a house-built beam engine actually used the building which surrounded it to hold the various parts of the engine in their true relationship, and the building thus had to be capable of resisting some of the working forces of the engine. In practice most of these forces were transmitted direct to the foundations, which had to be of massive construction, but some were carried by the side walls and floor beams of the engine-house itself. There was also another sense in which these engines were ‘house-built’: their major components, cylinders, beam and flywheel, were too large to pass through the doors or windows of the engine-house, and they were thus introduced as construction of the building proceeded, leaving final positioning and alignment to be carried out at a later stage (Fig 3).

The two main types of beam engine used in waterworks were ‘rotative’ and ‘Cornish’ or ‘non-rotative’. Although the rotative engine was a development which came after the basic Cornish pumping engine, itself a successor to Newcomen’s engine of 1712, both types were used during almost the whole of the period we are discussing.

Fig 4

Rotative Beam Engine, as installed at

Heigham, Norwich (1878-81)

Fig 4 shows a rotative beam engine, as installed at Norwich in 1878-80. It will be seen that, as steam is admitted alternately above and below the piston in the cylinder, forcing it down and up, the beam is caused to rock about its central bearings, and through the connecting rod and crank the flywheel is caused to rotate. The flywheel serves to maintain a steady speed (usually 10-12 revolutions per minute), and the crank effectively limits the stroke of the piston in the cylinder. Additionally the engine valves may be driven from the crankshaft through gearing. The flywheel plays no part in the pumping action - various pump rods attached to the beam itself move up and down to operate pumps directly.

The most important difference between rotative and Cornish engines is that the Cornish engine is ‘single acting’. Steam is admitted above the piston and a vacuum is applied below it, driving the piston downwards and raising the pumps attached to the other end of the beam. The pumps are then allowed to fall under their own weight, returning the piston to the top of the cylinder ready for the next cycle. The two types of engine, although fundamentally different in operating principle, are however, superficially similar in appearance and make similar demands on the buildings which contain them. The Cornish engine does, however, lack the connecting rod, crankshaft and flywheel - all the elements giving rotation - and is thus ‘non-rotative’: the speed and stroke of the engine are controlled by valves which are operated directly from the beam.

Both types of engine require a ‘parallel motion’ mechanism on each end of the beam, since the beam ends move in an arc whereas the piston rod and the pump rods both have to move in a true vertical line. One part of this mechanism, based on the pantograph, has to be anchored to a fixed point, which is normally provided by the building if the engine does not have an independent frame.

There were variations on these types, for instance the crankshaft of a rotative engine could be used to drive different sorts of pumps through gearing, but the two types described above were the most usual for large waterworks applications, and since the rotative engine was the type most commonly employed, the requirements of an engine-house for this type will be described in some detail.

In order to reduce the amount of excavation required, the massive foundations needed for the engines themselves were usually only partially below ground level. This had the effect of raising the engine-house on a plinth, with the floor level being expressed externally by a weathered offset. Access into the engine-house was almost invariably by means of a short flight of steps leading up to double doors in the centre of the entrance front. The engine-house would normally be built to accommodate either one or two beam engines. Where only one was to be installed, as in the 1859 and 1906 buildings at Heigham, its cylinder, beam and pumps would be on the centre line of the engine-house, with its connecting rod and crank immediately facing the doorway. Where there were two engines they would be placed symmetrically, leaving a clear passage between them in line with the entrance (Fig 6). In both cases the engine flywheels, typically 20 to 25 feet in diameter, would be close to the engine-house wall. There was a practical advantage in this: occasionally the engine had to be turned over by hand for maintenance purposes, or if it had stopped on a dead centre position from which it was impossible to restart. To enable this to be done a heavy cast iron plate with multiple notches was fixed to the wall behind the spokes of the flywheel, and using a crow-bar wedged in one of the notches and bearing on one of the flywheel spokes, the engine could be laboriously turned. The position of this ‘flywheel-rack’ was one factor which limited the size and position of the lower engine-house windows. The engine crankshaft was usually located just below floor level, and its bearings were placed one on each side of the flywheel. The outer bearing was thus virtually in the engine-house wall, and access to it from inside for maintenance or lubrication purposes was impossible. This problem was overcome by providing an opening in the engine-house plinth, through which the crankshaft was allowed to protrude, and specially shaped ‘crankshaft doors’ gave access to the bearing from outside (Plate 35). These crankshaft openings are the least vandal proof part of an engine-house, and are usually bricked up if the building is to be put to any other use.

Internally the engine-house was divided in two by a screen of columns supporting an entablature, the outer ends of which rested on heavy piers built inwards from the walls (Fig 6, Plate 37). This entablature (always referred to by this name, even by engineers who had no idea of the architectural significance of the term) supported the main beam bearings, and was as integral a part of the engine as it was of the engine-house. The columns were of cast iron, and were firmly anchored to the foundations. Each engine had either a single column below the beam (e.g. Norwich 1906), or a pair of columns symmetrically placed under the beam bearings. Although either arrangement was a mechanically sound way of transmitting the main vertical forces direct to the foundations, lateral support had to come from the building itself. The heavy internal piers were thus essential, and were normally expressed externally by pilasters or stepped buttresses, depending on the architectural style chosen for the exterior.

The entablature also carried the main longitudinal joists for the upper floor (the ‘beam floor’). These too were usually of cast iron and frequently incorporated cast pockets to accept the secondary wooden joists which supported the actual boarding of the beam floor (Plate 38). These main joists ran either side of the engine beam, with a hole between them to accommodate the motion of the beam and its attachments. They too formed part of the engine as well as part of the architecture: they gave additional stiffness to the entablature by bracing it to the end walls, and they also provided anchorage points for the ‘parallel motion’ mechanism (Plate 36). Where the engines were of the Cornish type, these joists were expected to arrest the movement of the engine and prevent damage caused by over-travel in an emergency. Additionally they provided stable anchorage points for winches on the beam floor, so that heavy engine and pump components could be lifted for maintenance.

Fig 5

Roof Construction of a typical Beam Engine-House

It was this requirement for heavy lifting facilities which caused the roof structure of engine-houses to be often their most impressive feature. Trusses were formed from timbers of massive scantling, 12 x 10, 15 x 10 or even 18 inches square elaborately jointed and strapped together (Fig 5), with diagonal ‘dragon’ beams braced into the corners of hipped roofs. The trusses were positioned so that lifting eyes, fixed to large spreader plates, could be located more or less over the major components of each engine (Plate 10). In fact, the roof itself may have been fairly lightly constructed: in the 1859 engine-house at Heigham the rafters rest on relatively slender purlins which lie on the backs of the lifting trusses, and the rafters themselves support the ridge-piece several inches clear of the apex of the trusses.

Inside the engine-house, the only other architectural feature of note was the ‘packing stage’. This took the form of an intermediate or mezzanine floor, between the main floor and the beam floor, covering about a quarter of the total area. This floor was just below the top of the engine cylinders, which it usually surrounded, and gave access to the top cylinder cover for lubrication and maintenance of the piston rod packing. It was often supported on columns which mirrored those supporting the entablature, and it served as a landing on the stairs leading up to the beam floor, if these were inside the main engine-house building.

The proportions of the engine-house were determined by the engines it contained and this normally meant that the side elevation was almost square, its height being only slightly less than its length. The need to support the engine entablature normally precluded windows in the centre of the side elevation, dictating an even number of bays, either two or four. In Hawksley’s very early stations the entablature was not across the centre of the engine-house, and this permitted a three-bay elevation. The main floor of the engine-house was usually almost twice the height of the upper, or beam floors and this difference in height was expressed in the height of the windows.

The amount of heavy timbers in an engine-house roof made it impossible to have any form of rooflight, and roof ventilators were therefore provided since the heat given off by the engines could have made the engine-house oppressively hot in the summer. The ventilators were either in the form of a central lantern, or small dormer or gablet vents. These could be opened and closed by cords led down to the main floor level. These vents, particularly the lantern type, form a very distinctive feature of nineteenth century engine-houses: the introduction of overhead travelling cranes, either hand or steam powered, eventually led to the use of clerestory roofs or ‘jack roofs’ which gave improved lighting and ventilation, and caused a radical change in external appearance.

As this description indicates, many aspects of the shape, fenestration and general construction of a beam engine house were dictated by the engines it contained. Despite this, however some designers were able to work within these constraints and still produce buildings which were well-proportioned and attractive, amply justifying the use of the term architecture. Hawksley, and the architects he employed as his firm grew, managed to do this consistently from about 1840 onwards; others were no doubt as good, and deserve equal study - some however were very poor indeed.

The engine-house was clearly the most important building in the pumping station complex; occasionally it may have been the only one Hawksley’s were called upon to provide. The 1859 extensions at Heigham seem to have consisted, above ground, solely in a new engine and engine-house. Boilers existed from the earlier works erected under the supervision of James Lynde in 1851, and extended in 1856, and these may have provided adequate steam raising capacity for the new engine. Normally, however, an engine-house would need to be accompanied by a boiler-house and a chimney at least. Some undertakings seemed content to store their coal outdoors, and an area of paved ground close to the boiler-house would be marked ‘Coal Store’ on Hawksley’s plans. Often, however, a covered coal store was called for, and this could take a number of forms. The practical requirements of these various components and their effect upon the architectural form of the station will be discussed below, together with a brief consideration of some of the other elements which were, from time to time, incorporated in the plan.

The boilers used for large pumping stations were of either the ‘Lancashire’ or ‘Cornish’ type. Both types were cylindrical, typically 25 to 30 feet long and six to eight feet diameter. They were fired from one end, and the hot gases, after passing through tubes in the body of the boilers were conveyed back and forth along the outside of the boiler by brick flues, and the boilers were normally encased in brick or other insulation except at the front end. Access to the boiler-house from the engine-house was normally onto the top of the brick cladding, from which steps led down to ground level, the difference in floor heights being accounted for by the plinth of the engine-house. Additionally the boilers had to be below the level of the engines to ensure that any condensate in the pipes drained back into the boilers and not into the engine cylinders.

Maintenance of boilers was carried out in situ; there was thus no need for the lifting facilities which were such a notable feature of the engine-house. Boiler-house roofs were fairly low, light in construction, with slender wrought iron bar trusses and extensive glazing incorporating many opening lights. One frequent requirement was that the stoker should be able to see the top of the chimney without straying far from his post -- a smoking chimney is a sign of inefficient firing, and had to be corrected as soon as possible. Normally a single roof truss spanned two boilers; where there were more than this the intermediate valleys were supported on cast iron columns and beams.

Boiler-houses had large doors to enable tubs of coal to be wheeled in, and tubs of ashes to be wheeled out. Sometimes these ran on a simple railway track which connected the boiler-house with the coal store and the ash tip.

In order to obtain an adequate draught through the labyrinthine flues of a Lancashire boiler, a fairly tall chimney was required. The chimney at Heigham, built as part of the 1878-80 extensions (Fig 15), was 115 feet high, excluding the ornamental cresting rails which added another five feet or so. Another reason for this height was the provision, at Norwich of an ‘economiser’. This was a device for using the waste gases after they had passed the boiler, to give a certain amount of preheat to the boiler feed water, and its effect of course, was both to cool and to slow down the waste gases necessitating an increase in chimney height. By no means all Hawksley’s stations were fitted with economisers; where they were used it was often convenient to have the chimney separated from the boiler-house with the economiser house in between (as at Ormesby). The incorporation of an economiser in a compact and strictly symmetrical plan as at Norwich must have called for considerable ingenuity in the design of brickwork for the flues.

As mentioned above some undertakings were content to store their coal in the open. Others were lucky enough to be able to run a railway siding direct into the pumping station yard (Blagdon and York (Fig 17) are the only examples known among Hawksley’s clients, but the arrangement was not uncommon). More often covered coal storage was called for. Two coal stores, as at Heigham and Nottingham (Bestwood) for example, not only helped in the achievement of a symmetrical plan form, but had the practical advantage that one could be run down completely before starting on the other, thus avoiding the dead corners which are inherent in a single coal store.

Coal store roofs were generally constructed in the same way as the boiler-house roof, but without the glazing. Frequently the hipped or gabled roofs of the boiler-house and coal stores would range together with no external differentiation (Plate 8, Plate 9 & Plate 22). Doors were not always fitted on coal stores; indeed Bestwood had open three-bay arcades, the intermediate stone columns having finely cut capitals, each one different.

The pumping station was often intended to be a more or less self-contained unit. Many were located some distance from the towns they served, and a fair amount of maintenance was expected to be done on the site. The lifting facilities in the engine-house were not there for show. Pumps had to be laboriously drawn up out of the wells at regular intervals (or in emergency if the valves failed) to be fitted with new leathers, and corroded parts would have to be replaced. Pump rods, or ‘spears’ were frequently of wood, strapped and plated with wrought iron, and again replacements were sometimes required. Most stations of any size had a Smith’s Shop, with forging hearth capable of handling fairly big jobs, as well as making new ash rakes and prickers for the stokers. Some also had a Fitter’s Shop, with a few basic machine tools driven by a small steam engine.

When the Board of the Great Yarmouth Waterworks Company agreed at a special meeting in February 1882 that their new plant at Ormesby was to comprise ‘Engine, Boiler and, Economiser Houses, Chimney 100ft high, a good Coal Store and a sufficient Workshop⁽²⁵⁾ , they neatly listed the commonest set of features that a pumping station complex should contain. The range of ‘extras’ that were sometimes required include a w.c. (by no means a standard feature), offices for the station superintendent, stores of one kind or another, a weighbridge for receiving and ‘tareing-off’ coal deliveries, and even, as at York, a ‘Gardener’s Tool Place’. Two other items do require to be mentioned, even though they were not always provided.

Cooling ponds at pumping stations served the same purpose as cooling towers at a modern power station. Large steam engines, like modern turbines, make use of the vacuum produced by condensing steam in order to increase their power output. Steam is condensed by using cold water; when it and the condensed steam are discharged as ‘engine waste water’ they are warm, and hence a pond of reasonable area was required if the same water was to be recirculated without its temperature getting unacceptably high. The cooling pond, or ponds, formed an important feature of the landscaping of the site. The pond at Bestwood was irregular and informal, with a wooded island near one end; others were more formal, circular or rectangular with semicircular exedrae and complex corners, and sometimes with a central fountain (Fig 14 & Fig 16).

Stations drawing their water from surface sources such as rivers could discharge their waste water into the river below the main water intake, and had a readily available source of cold water for condensing. However, surface water needs to be filtered before being pumped into the mains or to the service reservoir, and filter beds at these locations were sometimes treated almost as ornamentally as cooling ponds at the deep well stations.

Water which has been filtered, or is otherwise fit for drinking purposes, is either fed direct into the mains or stored in a service reservoir. As Hawksley suggested in his Report for Worcester, it is advisable for water to pass into the mains ‘without further exposure to atmospheric influences’, and service reservoirs were hence usually covered, or ‘vaulted’. However, in some areas, such as Sunderland, the service reservoirs were left open, and these too were treated as landscape features.

It was common for pumping stations to have one or two houses on the site: initially, when working was part-time and the station staff was small, for the ‘engine worker’ and the stoker, and latterly as staffs grew larger on a shift system, for the superintendent or engineer and his assistant. Some also had stables - the turncook needed transport to get around his district. It appears that these houses were not always designed by the Consulting Engineer - perhaps a local architect could be found to do the job more cheaply - but many plans for ‘cottages’ are to be found in Hawksley’s files. These are usually substantial and well designed dwellings, though not infrequently they are in a style which differs markedly from that of the main pumping station buildings: e.g. at Spon End, Coventry, the neo-Classical engine-house was accompanied by a cottage in an ornate ‘Tudor’ style, with mullioned windows and tall octagonal brick chimneys. Some of these cottages look as if they may be taken straight from the pattern books, such as Loudon’s Encyclopaedia, but further research would need to be done to establish this. Occasionally the larger cottage incorporates a ‘Board Room’.

Hawksley’s seemed to set great store by careful landscaping of the whole pumping station site. Site plans are highly detailed, the sweep of the drives carefully set out, trees laboriously delineated, and the exact slopes of all the various changes of level all indicated. Although, as will be discussed later, it is unlikely that many undertakings were in a position to challenge Hawksley’s designs for the major engineering work of a pumping station, his extravagant plans for site layouts were sometimes more than his clients were prepared to accept.

On 24th June 1859, at a meeting of the Committee of Management of the Norwich Waterworks Company:

‘Mr. Ayris (Manager of the N.W.C) laid before the committee plans prepared by Mr. Hawksley for laying out the grounds at Heigham, building a Porter’s lodge etc. etc. and was instructed to make Mr. Hawksley acquainted with the surprise the Directors expressed at the extent and the expensive character of Mr. Hawksley’s plans which, considering the Locality of the Works and the financial position of the Company they are not prepared to carry out. The Directors feel the necessity of keeping the expenditure within the narrowest limits possible, and they only desire to erect a lodge at the entrance gate and a coal shed and shed for the stores, all of which should be of an inexpensive character.’

The Norwich minutes do not record the outcome of Ayris’s conversation with Hawksley on this occasion, but on 4th April 1860, the directors considered and approved ‘Plans prepared by the Manager of the Gate Lodge and office ...’. This almost certainly refers to the substantial building which still survives on Waterworks Road (Plate 12), but this is so close in style to the engine-house which Hawksley’s were commissioning at the same time that it is hard to believe they are not from the same hand.