Although the engineering history of Glasgow’s busy underground passenger railway is little known, the project contained may novel and interesting features when it opened on 14 December 1896, and most of these survive today. In 1935 there was an official name change from Subway to Underground, but Glaswegians still refer to it affectionately as the Subway.
The most interesting feature of the Subway’s early days was its method of traction, which was by an endless cable running on sheaves between the rails. It was not the first cable-operated underground railway in Britain, as cable haulage was used in 1870 by the Tower Subway Company in London on a short-lived crossing under the Thames. Cable haulage had also been widely used for street tramways and mines. The Subway nevertheless soon became the world’s only cable-operated underground railway until its electrification 1935.
By the 1880s, Glasgow, then second city of the British Empire, was seriously congested. Horses and carts, vans, omnibuses and cabs jostled for road space with horse-drawn trams. Glasgow already had, or was constructing, two long stretches of mainline railway running in tunnel under the city from east to west – the Glasgow City & District Railway was opened in 1886 and the Glasgow Central Railway a few years later. These lines set a precedent and spurred the idea of further underground extensions. The Glasgow District Subway Company was formed in 1887 by a group of prominent Scottish industrialists to promote a Bill in that year for Parliamentary powers to construct an underground railway. Earlier tunnelling ventures had revealed that the geology of Glasgow was far from attractive for underground railway construction. The Clyde was another obstacle.
The promoters of the Subway chose a small local Glasgow firm, Simpson & Wilson, to undertake the civil engineering work. This may have been because the senior partner, Alexander Simpson (1832-1922) had a long connection with the mining industry in Lanarkshire or, more likely, because the firm had a track record in tunnel construction for the City & District Railway (1886), the planning of the Harbour Tunnel (1890-95) and in railway construction both in Scotland and abroad.

Early proposals
Several schemes for ringing Glasgow by underground railway were considered, of which two were promoted by Bills in Parliament in 1887 and 1888 and rejected.
The main reason for rejection was that the Clyde Trustees argued that the ground cover to the tunnels where the Subway passed under the river was insufficient and would limit further deepening.
Adjustments were made and a third Bill received Parliamentary sanction in 1890. The approved scheme (Figure 1) was for an underground railway 10km – 54km in length in double tunnels, wholly underground, to be worked by any means other than steam locomotives. The gauge was to be 1.435m, but a later Act of 1894 altered this to a minimum of 1.067m. The chosen gauge was 1.129m in 3.35m diameter tunnels.
The total cost amounted to £1.59M, of which £1.47M was spent constructing the tunnels, stations, track and cable haulage system.
About £11,000 of the cost arose from the lines of the tunnels under private property. The company was forced to buy all such property to obtain wayleave and avoid disputes. The route was therefore chosen as far as possible to lie under streets, which offered free wayleave.

Method of Traction and its influence
The layout of the Subway involved steep gradients of 1 in 20 at the river crossings and a short length of 1 in 18 elsewhere. These gradients had led to reservations about the necessary adhesion between wheels and rails being adequate if the trains were self-propelled. Also, the motors available for electric traction at the time were heavy and bulky and could not be accommodated easily under car floors.
Cable haulage on the other hand was well understood and was adopted only after an exhaustive survey of systems of traction both in Britain and the US. It had the advantages of being clean and free from fumes. The endless cable to which the trains attached themselves revolved at a constant speed of about 22km/h and trains going downhill assisted those going uphill. It allowed maximum space for passengers in the cars, owing to the absence of mechanical plant. The main disadvantage was that considerable skill was required by the train driver (the ‘gripman’) in picking up and releasing the cable and achieving smooth starting and stopping of the trains.
The vertical curve of the cable was limited to 762m radius at the steepest gradients in order that the cable under tension would not lift off the sheaves. The adoption of cable haulage and the acceptance of steep gradients also allowed the adoption of ‘humps’ at the stations. The 15 stations are an average of 700m apart and braking and acceleration is assisted at several by gradients of 1 in 20 on each side. The wholly underground nature of the Subway, with no rail connection to the surface, led to the requirement of a special access pit and overhead crane for servicing of the cars. They were lifted off the track and deposited in the car shed above. This unusual arrangement, worked well for 80 years before modernisation.

Tunnelling
Table 1 shows the ground encountered. About 58% of the route – including all ground south of the Clyde – consisted of soft material and the remaining 42% was mainly through rock.
The soft material engendered the most difficult part of the construction, particularly the river crossings. Cut and cover rather than face tunnelling was used where the depth from ground level to tunnel formation was less than around 9m.
Tunnelling in rock was relatively straightforward, apart from influxes of water and a flood caused by a buried quarry near Glasgow Street in Hillhead. There are many hidden filled-in quarries in the north and west of Glasgow and the Subway was perhaps fortunate in encountering only one of them.
Construction began at St Enoch Station in March 1891 and involved excavation in the area’s water bearing sand.

Cut and cover
Two rows of timber sheet piling 8.4m apart were driven to formation level and the ground between excavated to the level of the soffit of the tunnels’ double mass concrete arch (Figure 2). The bottom of this excavation was shaped to the profile of the arch soffits and spaces cut in the piling at intervals to provide a bearing for the arch haunches, which were then cast. The top of the arches was covered with two coats of asphalt for waterproofing. The excavation was then filled and the surface reinstated.
The ground between the piles under the concrete arches was then excavated, one tunnel at a time. One half of the double invert was laid and one side wall and one side of the centre wall built in brickwork. After a 4.6m length had been completed the second tunnel was built in a similar manner.
Sometimes in vacant ground the excavation was carried down to formation level, the piles being strutted apart and the arches of the tunnel roof constructed with centring. Where the ground was water-bearing sand, the material entered the excavation through joints in the sheet piling. When this happened, settlement of the buildings at street level began to occur.
However, it was discovered almost by chance that if the piling was tight and water alone entered the excavation without carrying sand with it, then no settlement took place.
In some places the material was water-bearing and so soft that at the level of the invert the ground could not bear the weight of the concrete. Air pressure was then employed to force back the water and dry out the formation allowing the invert to be constructed.

Tunnelling in alluvial material
Where the tunnels were in water-bearing strata at a depth greater than 9m, or the ground surface could not be opened up (as at the Clyde crossings), the construction required a watertight lining of bolted iron rings and use of a shield under compressed air.
The shield (Figure 3) was a cylindrical shell, 3.72m in external diameter and 1.98m in overall length. One end of the shell was fitted with the cutting edge and about 300mm back from this was a bulkhead with sliding door to give access to the face. The shield was controlled by six hydraulic rams located round the circumference. Any one of the rams could be worked independently of the others, altering the direction of travel as required. Each shield with its fittings weighed 6t. The rams pressed against the rings of the completed tunnel, forcing the shield forward as the work advanced. Each 457mm wide ring was built up inside the shield from nine segments of 1.25m circumferential length and a narrow crown segment (Figure 4) and were grouted as the shield moved forward. All segment joints were packed with softwood sealed from inside the tunnel with timber wedges.
The ground ahead of the shield was excavated for a length of about 3m using poling boards closely fitted round the circumference of the tunnel and grouted (Figure 5). The shield was then moved into this space. It was found that the shield was of very little advantage when using this method under compressed air and some of the contractors abandoned its use. Working in sand required poling boards and the compressed air stiffened the ground, rendering the shield to some extent superfluous.
Difficulties were experienced however at the river crossing at Custom House Quay. Here the cover to the top of the tunnel was a mere 4.3m of sand and silt, increasing to 8.8m across the width of the river. Before the western tunnel had advanced 24m under the river, the bed had blown up no less than ten times. The worst burst created a hole 7m square and 5m deep, fortunately without loss of life. The contractor asked to be relieved of the work and a second contractor, George Talbot, was appointed. He carefully regulated the air pressure according to the state of the tide and successfully completed both tunnels. The 125m eastern tunnel was completed in 14 weeks, including time spent in restoring the working after one blow-out.

Tunnelling in clay
The clay met with was impervious to water and occurred mainly at places where some settlement of the surface ground could be tolerated. In this situation, ordinary brick tunnel lining was used, being cheaper than iron rings, though slower to construct.
The brick tunnels are circular both internally and externally. The lowest quadrant was constructed in mass concrete and the remainder in four rings of brickwork. Construction was done in 2.7m lengths.

Tunnelling in rock
A long stretch of the Subway on the north side of the Clyde, about 4.46km, is in rock, chiefly sandstone and shale. Here the work was laborious rather than difficult. The material was removed by explosives and the tunnel interior was then trimmed and lined with two to four rings of brick work.
One of the contractors, Robert McAlpine & Co, chose to use mass concrete rather than brickwork and thickness of this varied from 229mm to 457mm.
The River Kelvin was crossed by first building a cofferdam across half its width and excavating a trench in the rock of the river bed of sufficient depth and width to allow rings of iron lining to be laid in place. The cover to river bed was approximately 600mm and this was filled with concrete. The other half of the distance, a little over 15m, was crossed in the same way.

Stations
According to the depths of the tunnels, stations were constructed either in tunnel, cut and cover, or in the open between retaining walls.
The shallowest station, Kinning Park, was only 4.3m from ground to platform level and was constructed in the open between retaining walls. The deepest station, 12.2m, is Buchanan Street.
In built-up areas entrances to stations were often through shop premises converted for the purpose. Others had plain architect-designed small buildings in brick.

Civil engineering contractors
There were eight main contractors involved in the civil engineering work of the Subway between 1891 and 1896. Most were well-known and highly experienced public works contractors of the day (see Table 2).

Electrification
Plans for electrification had been considered before the First World War and were raised again in 1922, but it was 1933 before tests were completed sufficiently for a final decision to be taken. The existing rolling stock was changed from cable to electric traction and conversion was finally completed by December 1935. The track remained ballasted with tile drains.

Modernisation
By the 1960s the Subway had become archaic and worn-out. Frequent breakdowns were commonplace and despite valiant work by maintenance staff it was becoming obvious that it was nearing the end of its useful life. Plans for modernisation were prepared and the Subway finally closed in May 1977 for a comprehensive programme to begin, this was completed in 1980.
The modernisation transformed the Subway’s former stations and buildings, somewhat cramped in space and dingy in appearance, to well-lit, pleasing and attractive buildings with excellent amenities for passengers and staff. The rolling stock has undergone a similar transformation and new life has been breathed into the whole system. It seems that the Subway is set to roll on its way for a second 100 years of existence, maintaining its unique service to the city of Glasgow.

Related Files
Figure 3
Figure 4b
Figure 1
Figure 4
Figure 5
Figure 2