Chapter IV.
TRING: History About Library Contact TRING: Home

Foreword Contents Canal Profile System Map (North) System Map (South) Site Search




“The mere digging or cutting into the earth is so common and obvious an operation that it may seem to require neither skill nor explanation.  This however only applies to small and ordinary operations, for when the work is extensive, as in the formation of canals, reservoirs, tunnels, and the like, many expedients are resorted to that might not occur to common workmen; they have arisen out of experience and are adopted because they economise labour and time, and consequently diminish the expense of executing the work.”

Elements of Civil Engineering, John Millington (1839)

In deciding what route a canal was to follow, its chief engineer first needed to undertake a preliminary survey of the area, for most canals were built in an age before reliable maps were generally available [1] ― they had therefore to be made, and to a standard sufficient for their purpose.  Equipped with level, staff, measuring chain, compass and surveyor’s barometer, [2] the engineer and his assistants set off to acquaint themselves with the landscape.

A surveyor's Y-level by Adams, London, late 18th century.
History of Science Museum, Oxford.

A telescope supported in y-shaped rests and as a result capable of being rotated around its own axis or of being taken out of the supports and turned end for end for purposes of adjustment. The surveyor uses a level to determine elevations. In conjunction with a level, the surveyor uses a ‘level rod’ to read an elevation above or below the level of the telescope. From these observations can be determined the differences in elevation of different points along a route, or an elevation can be transferred from one location to another.


A ‘waywiser’ by Thomas Wright and William Wyeth, London c. 1740.
History of Science Museum, Oxford.

This is an surveying instrument for measuring distance travelled over the ground, including its undulations. As it is moved along, the wheel drives a mechanism in the dial, which clocks the distance moved.

Surveyor's Chain.
Campus Martius Museum.

In the top illustration, the 'chainman (to the right of the surveyor) is holding a surveyor's (or Gunter's) chain, used to establish horizontal distances accurately along compass sight lines. A chain is sixty-six feet or four poles long, and is composed of one hundred links, connected each to each by two rings. On sloping land, the chain has to be levelled by raising one end to the horizontal, so that undulations in the ground do not increase the apparent length of the side or the area of the tract being measured.

In most cases the preliminary survey would identify several possible routes or variations on them.  Other factors would then need to be considered, not least of which were the locations of any estates that the canal needed to traverse.  If an influential landowner did not wish to sell a tract of land to the canal company, objections could be expected during the parliamentary committee hearing, objections that counsel would be employed to express most forcibly by cross-examining the chief engineer on the feasibility of his plan [3] and by embroidering their client’s grievances to best effect.  In practice, canal (and later railway) promoters often placated influential landowners by buying them off at much inflated rates.  While this was usually an effective strategy, it was one that became increasingly expensive as landowners realised that by adopting hindering tactics they could greatly increase the value of their land.

Another aspect of land purchase to be considered was the higher cost of land in urban areas; might it be more cost-effective to select a route around the outskirts of a town rather than through its centre, or to avoid the town altogether?  Traffic levels would also influence a canal’s dimensions and thus the amount of land required; a wide canal could accept barges larger than the conventional ‘narrow boat’, [4] and greater loadings earned more revenue, but it required more land, materials and man-hours to construct, and thus more capital.  Empirical rules were developed to determine the optimum dimensions of a canal for whatever size of craft was most likely to be used:

“Although for the sake of saving expense in aqueducts and bridges, short portions of a canal may be made wide enough for the passage of one boat, only the general width ought to be sufficient to allow two boats to pass each other easily.  The depth of water and sectional area of waterway should be such as not to cause any material increase of the resistance to the motion of the boat beyond what it would encounter in open water.  The following are the general rules which fulfil these conditions:―

Least Breadth at Bottom = 2 x greatest breadth of a boat.
Least Depth of Water = 1½ feet + greatest draught of a boat.
Least Area of Water-way = 6 x greatest mid-ship section of a boat.

The bottom of the water way is flat.  The sides when of earth which is generally the case should not be steeper than 1½ to 1; when of masonry they may be vertical; but, in that case about 2 feet additional width at the bottom must be given to enable boats to clear each other, and if the length traversed between vertical sides is great, as much more additional width as may be necessary in order to give sufficient sectional area.”

A Manual of Civil Engineering, W.J.M. Rankine (1862)

In an age when canal barges were hauled by horses ― mules and donkeys were sometimes employed ― the land purchased had also to be sufficient to permit a towing path to be built, and of a width that would allow two horses to pass. [5]

In making his recommendations to the Board, the chief engineer needed to weigh estimated traffic revenue plus any other income (such as water sales and property rent) against operating expenses plus interest on loans, to estimate the likely yield on capital ― was the canal, as proposed, a worthwhile investment?

“A proper engineer being fixed upon, the adventurers [i.e. the Proprietors] should not tie him down too closely by restrictions as to time; but allow him leisure to consider digest and revise again and again the different projects and ways which will in most instances naturally present themselves to him in an extensive and thorough investigation . . . . The most eligible route for a canal being settled in the engineer’s mind, he will then proceed to make a rough calculation of the quantity of goods of each kind which may be expected to pass upon the line in a given time; he will also examine all the canals and rivers with which the proposed canal is to connect, and ascertain the widths, and depths thereof; the sizes of their locks and of the vessels usually navigating them”.

An Encyclopædia of Agriculture, J. C. Loudon (1871)

The need to solve problems such as these illustrate why the civil engineering profession evolved quickly from its trade roots, such as those of stone mason or millwright, into that of seasoned professionals who could handle large projects rationally.  Promoters of canal (and later on of railway) schemes eventually came to expect their chief engineer to possess the knowledge and experience to plan the entire project, not merely one of its parts however important that part might be.

For many years Edward Gray of Buckingham and Acton
Chaplin of Aylesbury served as Clerks to the Grand
Junction Canal Company, with press notices such as
this frequently appearing above their names.

All factors having been considered and the route decided, a detailed survey was then necessary to determine accurately the form and dimensions of the ground to be traversed and the positions of objects upon it, processes called ‘triangulation’ and ‘levelling’, [6] and to make trial borings which (it was hoped!) would reveal potential problems with geological formations invisible to the naked eye.  This information allowed maps to be prepared, showing the position of the canal in relation to land boundaries and the physical features on the landscape, and cross-sectional diagrams on which were represented the positions of slopes in relation to the horizontal, enabling the chief engineer to examine the gradient at any point along the route.  Engineering works could then be designed to reduce gradients to a sequence of level sections by the construction of locks, cuttings, embankments, aqueducts and tunnels, while reservoirs and pumping stations could be sited where necessary.

The rules for applying to Parliament for a private Act required canal promoters to provide the committee who were to consider their application with the details of the intended route and of the landowners who would be affected.  Collectively, these are known as the  ‘deposited plans’. [7]  If the promoters’ Bill succeeded, thereby becoming an Act, it was this route, perhaps modified by objections made during the committee hearings, which became the authorised ‘Parliamentary Line’ from which the canal company was not permitted to deviate by more than 100 yards.  The Act also authorised the canal company to acquire the land they needed by compulsory purchase, but with the inclusion of an arbitration procedure to be followed in cases where the parties could not agree on valuation.

Having raised the finance, engaged the services of a solicitor (‘Clerk to the Company’) and civil engineer, completed the surveys, posted public notices, obtained the necessary private Act, purchased the land, and agreed terms with earthwork and other contractors, construction could begin.  Activities often commenced with the ceremony of ‘cutting the first sod’, at which the company chairman and other dignitaries attended to turn a token spade-full of earth (rather less than the twenty tons that each navvy would subsequently consider to be a fair day’s work) before sitting down to a sumptuous dinner, which contemporary news reports suggest followed each major achievement during a canal’s construction.




Although this account is of a canal, it is worth mentioning what distinguishes a canal from the other types of navigable waterways that are referred to occasionally in the following text.

A ‘river’ is a natural watercourse that drains water off the land into the sea; in its natural state it is vulnerable to drought and to flooding.  If a river is to be used for navigation, its natural course generally needs to be straightened and deepened by dredging and by the construction of weirs and locks.  Rivers made navigable in this way are termed ‘river navigations’ or ‘canalised rivers’; which term applies appears to be much a matter of scale.  For instance, the extensive Aire & Calder Navigation in West Yorkshire [8] is referred to as a ‘river navigation’, while the more modest section of the Grand Junction Canal between Hanwell bottom lock, where it is joined by the River Brent, and the end of its journey at Brentford Creek is described as a ‘canalised river’.

In contrast, a canal is a man-made trench or channel that is filled with water to a depth sufficient to permit navigation [9] and which, unlike a river, connects predefined locations.  In effect it is a manmade highway laid with water.  If properly managed, a canal should be much less vulnerable than a natural river to drought and flooding, and because it is built in level sections and includes weirs to carry off any excess water, what current there is is negligible, thereby permitting the comparatively easy haulage (or propulsion) of heavily-loaded craft in either direction.  However, as still water freezes more quickly than running water ― especially if the latter is saline ― a canal is more vulnerable to freezing over.

A canal is a manmade highway laid with water.  Here, the Wendover Arm of the Grand Junction Canal is being restored.  This section of the Arm was abandoned in 1904 due to incurable leakage and its water fed into a pipe to supply the Tring summit via Tringford pumping station.  Today’s navigators ensure the canal is watertight with the use of Bentomat© sheeting, seen here being rolled along the bed before being topped with 300mm of earth.  Bentomat sheeting is also laid behind the protective blockwork along the canal banks.



On level ground, a canal was formed by cutting the ground [10] and using the excavated spoil to form its banks.  ‘Cutting’ commenced with the ground being cleared of obstructions.  The chief or resident engineer would then mark out the course of the canal (in accordance with the deposited plans) and the position and height of its banks with pegs and stakes.

Following cutting, it was often necessary to make the canal watertight.  If the ground over which it passed absorbed water ― such as sand, gravel or chalk  ―   an impervious lining was needed to prevent leakage. [11] Although a lining of clay might appear the obvious solution, it is unsuitable, for if the water level in the canal sinks below the upper part of the lining, pure clay, on drying out, shrinks and cracks.  When the water level returns to normal the cracks remain and cause leakage.  The canal builders therefore used a modified form of clay for lining called ‘puddle’, the process of laying it along the canal bed and its banks being known as ‘puddling’.

 The application of puddle (‘A’ in the diagram) to canals in cuttings and on level ground, from . . . .
Illustrated Glossary of Civil Engineering - S. C. Brees (1852).

Puddle clay does not occur naturally.  It is made from loam mixed with sharp sand and water, which is then worked intensively into a plastic state in which it forms an impervious seal.  A civil engineering manual of the age described the process:

“No cheap and common material is found to oppose the filtration and passage of water so effectually as a soft loamy clay when it is well worked or kneaded into a soft paste with water and is not permitted to get dry again.  Even if a little fine gravel or what is called by the navigators in England hoggin, being small sifted gravel no stone of which is larger than a common pea is mixed with it, it seems to hold better but this can only arise from these small stones assisting in the kneading process.”

Elements of Civil Engineering, John Millington (1839)

The puddle was then applied in layers and left to mature ― but not dry out ― as the handbook goes on to inform the budding canal engineer:

“The ground is loosened in the bottom by the scoop [digging tool], but is not thrown out; that done, a pretty copious supply of water is sent into the puddle gutter by buckets or a temporary pump, and the workmen, by pressing down the scoop tool and walking backwards and forwards in the puddle gutter, reduces all the natural soil that has been disturbed into a state of very soft mud, or slush, as it is called.  This is done for the purpose of producing an intimate union and incorporation between the natural soil and the puddling stuff to be afterwards added.

The puddling stuff is now brought in barrows and cast into the gutter, to be treated in the same manner; a copious supply of water must constantly be given and the more the puddling stuff is trod and worked by the feet and scoop the more perfect the puddle will be.  Nothing is found to answer the purpose so effectually as treading with the feet and the layers of puddling stuff should never exceed nine inches in thickness without being trodden and worked.  The stuff should be kept so wet that the feet sink in eight or nine inches at every step and this same operation is continued until the puddle gutter is filled to the top or at any rate to a greater height than that at which the water in the canal or reservoir will stand.  Dry earth is then placed over the top of the puddling to protect it from the sun and air . . . .”

Elements of Civil Engineering, John Millington (1839)

The total thickness of puddle varied between 2ft and 3ft depending on the nature of the underlying soil.  The process of treading it in, by navvies wearing thigh-length boots, must have been extremely tedious and tiring, and it is unsurprising that there is anecdotal evidence of cattle being herded over the puddle to achieve the same effect.  Rushes might then be planted to help consolidate the banks.

Re-puddling a canal bed.

While puddling controlled leakage, water level had also to be limited, for a canal overflowing its banks could result in a breach and serious flood damage to the locality.  ‘Overflow weirs’ take the form of a section of wall built into the canal bank to a height corresponding to the maximum permitted depth; should this be exceeded, the excess water flows over the top of the weir into a channel.  Weirs built at locks channel any excess water around the lock into the section of canal (‘pound’) below it.  Excess water might also be channelled into a reservoir to be pumped back into the canal at times of low water level, or into a nearby stream or river where it might then have been used to drive water mills (disputes between water millers and canal companies over the latter’s use of river water were common) or simply run off to waste.  Temporary dams (‘stop gates’) and sluice gates enabled sections of the waterway to be isolated and drained completely.  Again, the manuals of the day provide the novice civil engineer with sound advice:

“Each reach of a canal should be provided with waste weirs in suitable positions to prevent its waters from rising to too high a level; also with sluices through which it may be wholly emptied of water for purposes of repair; and in a reach longer than two miles, or thereabouts, there may be stop gates at intervals so that one division of the reach may be emptied at a time if necessary.  The rectangular channel under a bridge or over an aqueduct is a suitable place for such gates.”

A Manual of Civil Engineering, W. J. M. Rankine (1862)


This overflow weir at New Bradwell permits surplus water from the Canal’s two summits to spill into the Great Ouse.
 The narrow boat has just crossed the New Bradwell Aqueduct (opened in 1991).

Originally, only the nearside canal bank was protected, usually with dry-stone walling backed with puddle clay, its main purpose being to support the towpath.  When powered craft began to appear late in the 19th century, it was found that the turbulence they caused resulted in erosion of the unprotected bank, resulting in weakening and debris, which subsided into the channel where it caused silting.  Over many years, hundreds of miles of piling ― concrete at first, steel later ― had to be driven into canal banks to provide the necessary bank protection.


New sheet steel piling on the Wendover Arm at Bulbourne Junction.  The interlocking sheets are driven into the canal bed using a pile driver.  To the rear, the piling is held in place with steel rods sunk into the embankment, back-filled with rubble and topped with coir rolls (just visible).

Bridges were needed ― indeed, they were usually specified in a canal’s Act of Parliament ― to carry public roads and to connect the opposite sides of estates that had been divided by a canal.  Bridges were also necessary to carry the towpath across the canal where it changed banks, [12] or where it crossed the entrances to docks or intersections with other waterways.

This attractive turnover bridge is at The Grove on the Grand Junction Canal.
This type of bridge is so constructed that horses could proceed with the tow-rope remaining attached to the boat.



A significant factor to affect the canal engineer’s choice of route was that of ‘gradient’, or the steepness of the ground over which the waterway was to pass.

Sections of a canal need to be built on the level if it is to provide still water for navigation.  In some parts of the country it was possible to build long stretches on the level merely by following the contour of the land.  Such ‘contour-following canals’ tend to be characterised by their long, meandering course.  In contrast, a direct route, while shorter and requiring less land to be bought, generally required more engineering work in the form of locks, cuttings, embankments, tunnels and aqueducts to permit the canal to be built in level sections across whatever gradients lay in its path.

Thus, the choice for the canal engineer was whether to select a longer, flatter route that bypassed rising ground, or to cut across it with engineering work.  A longer slower journey did not matter that much when the first canals were built, for their carrying capacity far exceeded anything that had previously been practicable.  Furthermore, because civil engineering was at the time in its infancy, canals of this period tend to follow the contour to avoid the need for risky and expensive earthworks; the meandering southern section of Brindley’s Oxford Canal is an example. [13]  Later canal engineers took more direct routes, an example being the Birmingham and Liverpool Junction Canal (now part of the Shropshire Union).  Engineered by Thomas Telford and completed in 1835, this was one of the last canals to be built and exhibits the characteristics of the railways that were shortly to follow.  By taking a direct line across country at the cost of building cuttings and embankments, some of which were massive undertakings, [14] Telford’s route saved some 12 miles.

An illustration of how engineering work can be used to reduce the gradient of a road crossing rising ground.
A canal follows the same principle, using a series of level steps, each separated by locks.



The principle of the ‘side cutting’, here applied to
a road traversing the side of an incline.

Cuttings and embankments are two means of eliminating gradient.  In a sense one is the opposite of the other; a cutting is cut through the ground, an embankment is built above it, [15] while the term ‘cut and fill’ describes the process by which, in an ideal situation, the spoil dug out of a cutting matches the quantity needed to build a nearby embankment.

Cuttings come in two types; a ‘side cutting’ is created on one side of a slope and is formed by cutting into the high side of the hill and using the spoil to build up the low side.  Alternatively, the low side might be built up from below.  But more often a cutting is excavated with rising ground on both sides, and the spoil is carried away for use elsewhere, or if there is no use for it, it is spread on the surrounding land. [16]  During the period of canal and railway building, cuttings were excavated manually with the assistance of gunpowder for blasting rock.  Compared with the problems that can be encountered in excavating a tunnel, forming a cutting or embankment might appear straightforward by comparison, but they too present engineering difficulties.

An extreme example of a canal cutting.  The Corinth Canal connects the Gulf of Corinth with the Saronic Gulf in the Aegean Sea.  The limestone terrain permitted its walls to be cut nearly vertical ― they rise 90 metres (300 ft) above sea level at an 80° angle.  Cutting through softer terrain would have required a much shallower angle of repose to prevent the walls slipping into the channel.  As things turned out, the stone proved to be heavily faulted resulting in landslips, and much remedial work has had to be carried out over the years to stabilise them.

The problem with cuttings lies in landslip, or the walls of the cutting slipping or otherwise falling into the channel.  This can result from the walls being built too steeply, perhaps to economise on the amount of land that has to bought.  In the case of cuttings driven through clay, poor drainage can cause the clay to expand as it becomes saturated ― much difficulty was experienced with this particular problem when building the cuttings into Euston on the London & Birmingham Railway.  Unstable rock, particularly when it becomes fractured by ice, can fall away from the walls.

‘Angle of repose’ describes the natural angle at which a granular material, such as earth or sand, will rest without slipping.  A knowledge of this property for different materials plays an important part in the construction of cuttings and embankments when deciding what angle the walls will stand at without other measures having to be taken to prevent slip.  The angle varies for different types of materials, and also whether they are dry, wet or waterlogged.  For example, no matter how much dry sand is added to a pile, it won’t form a slope steeper than approximately 35º, which is the angle of repose for sand in this condition.  In contrast, by making the sand damp, the capillary attraction or cohesion between the sand grains increases, allowing the creation of up to vertical walls.  But as more water is added, the sand begins to act like a plastic substance, and in its waterlogged state the angle of repose drops to about 12º.  From this can be seen the need to build good drainage into the walls of cuttings and embankments, particularly where the terrain is a granular material rather than hard rock.

Some of the techniques that are used to prevent slopes slipping are:

  •     the construction of retaining walls;

  •     putting drains through retaining walls so that water is not trapped behind them;

  •     constructing terraces to reduce the angle of slope;

  •     using grasses or other plants (such as rushes in soft canal banks) whose roots anchor the slope;

  •     sinking piles through unstable debris down to firm bedrock;

  •     inserting bolts (rock bolts) to hold unstable rocks.



Whereas a cutting lowered the level of a canal below that of the surrounding land, an embankment raised it.

Building an embankment.

Ideally, the spoil from a cutting would provide the material from which a nearby embankment is formed, assuming that it is suitable for load-bearing.  For this reason, in planning a route through undulating country, the Chief Engineer attempted to locate embankments and cuttings near to each other.  If there was no cutting near at hand to provide the material to form an embankment, it needed to be excavated from a site nearby, which involved buying land.  And if the material could not be obtained from a site above the level of the planned embankment, there was the added difficulty of building the embankment ‘bottom up’ rather than ‘top down’.

In building a canal embankment, the spoil from which it was formed was conveyed by canal boat, floating in a temporary wooden trough that was extended as far as possible up to the edge of the workings.  Alternatively, a temporary railway would be run up to a point at the edge of the workings where a large baulk of timber was fixed across the rails to form a buffer.  Tipper trucks filled with spoil were then worked up to a good speed, either by gravity or by horse; if the latter, on approaching the buffer the driver would untether the horse, which was trained to move smartly to one side leaving the tipper truck to continue on its way.  On hitting the buffer, the truck would stop dead, allowing its inertia to cause the ‘tipper’ to swivel upright, spilling its contents over the edge onto the workings below.  The track and buffer would be advanced as the work proceeded, which was generally from both sides of the valley simultaneously.

Building an embankment.

Much civil engineering at this time was based on the trial and error of past experience, rather than on proper scientific investigation.  Because the underlying scientific principles were not well understood, canal and railway embankments were made of relatively uncompacted material, [17] which sometimes caused them to settle over time.  Further problems occurred with the gradients of embankment slopes, which were often too steep and poorly drained; both could result in collapse.  The Grand Junction Canal was raised on several substantial embankments, including one across the Great Ouse Valley.  Opened in 1805, this large embankment, which incorporated an aqueduct, sank in the following year (to be followed in 1808 by the collapse of the aqueduct).  And so the wisdom of the age was that . . . .

“Great care is necessary to be taken in making high embankments.  No person should be intrusted with these works who has not had considerable experience as a canal or road maker; for, if the base of an embankment be not formed at first to its full breadth, and if the earth be not laid in regular layers or courses of not exceeding four feet in thickness, it is almost certain to slip. . . . . No doubt, a chief reason for making cuttings and embankments, as is frequently the case, with slopes of one to one [1:1] has been to save expense in the purchase of land and moving earth.  But the consequence of making such slopes is that the earth is constantly slipping, so that in the end the expense is always greater in correcting the original error than it would have been if proper slopes had been made in the first instance.”

A Treatise on Roads, Sir Henry Parnell (1833).



Among the most recognisable feats of civil engineering in the world, the Pontcysyllte Aqueduct comprises a cast iron trunk mounted on iron arches supported by stone piers.  Opened in 1805, it was constructed by Thomas Telford and William Jessop to carry the Ellsemere Canal across the River Dee valley at Froncysyllte near Llangollen.

Navigable aqueducts (as opposed to those used for water supply) are bridge-like structures designed to carry canals over other waterways, valleys, railways and roads.  The first navigation aqueduct to be built in this country was constructed by James Brindley.  Opened in 1761, the Barton Aqueduct carried the Bridgewater Canal over the River Irewell.  Brindley’s biographer, Samuel Smiles, described it thus:

“The Barton aqueduct is about two hundred yards in length and twelve yards wide, the centre part being sustained by a bridge of three semicircular arches, the middle one being of sixty-three feet span.  It carries the canal over the Irwell at a height of thirty-nine feet above the river — this head-room being sufficient to enable the largest barges to pass underneath without lowering-their masts.  The bridge is entirely of stone blocks, those on the faces being dressed on the front, beds, and joints, and cramped with iron.  The canal, in passing over the arches, is confined within a puddled channel to prevent leakage, and is in as good a state now as on the day on which it was completed.  Although the Barton aqueduct has since been thrown into the shade by the vastly greater works of modern engineers, it was unquestionably a very bold and ingenious enterprise, if we take into account the time at which it was erected.  Humble though it now appears, it was the parent of the magnificent aqueducts of Rennie and Telford, and of the viaducts of Stephenson and Brunel, which rival the greatest works of any age or country.”

James Brindley and the Early Engineers, Samuel Smiles (1864) [18]

Aqueducts were not much liked by the early canal builders due to the considerable weight of the water and the clay (needed to keep the trunk watertight) to be supported.  Another engineering problem to solve was how to build-in sufficient lateral strength to overcome the outward (spreading) thrust of the water. [19]  However, cast iron trunk (or trough) aqueducts appeared at the close of the 18th century.  The strength and rigidity of cast iron was sufficient to contain the outward thrust of the water, while bolting the plates permitted a watertight seal to be formed more easily than using layers of puddle clay, hence cast iron trunk construction came to replace masonry.

The Pontcysyllte Aqueduct, which carries the Llangollen Canal across the River Dee, is the most famous example of an iron trunk aqueduct in the UK, or possibly anywhere.  Opened in 1805 and generally attributed to Thomas Telford, although William Jessop probably contributed, the Aqueduct consists of a cast iron trunk, 11 feet 10 inches wide, supported on 18 ashlar stone piers, and reaches a maximum height of 126ft over the lowest point of the Dee valley.  The trunk is made up of ¼-inch thick cast iron plates bolted together along flanges, the joints made watertight with a mixture of flannel, white lead and iron borings. [20]

Canal aqueducts continue to be constructed.  One such example, opened in 1991, is the concrete New Bradwell Aqueduct, which carries the Grand Junction Canal across the Grafton Street dual carriageway at Milton Keynes.



Cross section of Thomas Telford’s Harecastle Tunnel on the Trent & Mersey Canal.
Opened in 1827, at 2,926 yard it is the UK’s forth longest navigable canal tunnel still in use.

Building the early canals required the use of new and sometimes untried engineering methods on a grand scale.  Tunnelling was the most difficult challenge facing the early canal engineers, and one that they avoided so far as possible:

“We have already noticed the evils of this mode of establishing a canal or part of one; sometimes it becomes absolutely necessary however in which case the greatest care is necessary in blasting rocks in consequence of the fissures which explosions will cause from which filtrations [leakage] will inevitably follow.  The excavation may be made with the pick if the soil is soft and cohesive which is the most favourable case for tunnelling.”

An Elementary Course in Civil Engineering, M. I. Sganzin (1837)

In commencing the construction of a tunnel, the over-ground route across the ridge was first marked out using a telescope to secure a straight alignment.  A detailed investigation was then made of the ground conditions through which the tunnel was to pass by collecting samples from boreholes dug along its route.  As these only represented samples at the locations where they were taken, there remained a considerable risk that any strata of hard rock, subterranean springs and pockets of loose gravel and quicksand [21] that lay in between would go undetected ― such was the case with Blisworth Tunnel on the Grand Junction Canal, resulting in serious flooding:

“As tunnels compared with open excavations are an expensive and tedious class of works and as they form inconvenient portions of a line of communication the engineer should study to avoid the necessity for them as far as possible . . . . The nature of the strata through which a proposed tunnel is to pass should be carefully ascertained not only by means of borings and shafts, but in some cases also by means of horizontal or nearly horizontal mines or drifts along the intended course of the tunnel . . . . The most favourable material for tunnelling is rock that is sound and durable without being very hard. Great hardness of the material increases the time and cost of tunnelling but gives rise to no special difficulty.  A worse class of materials are those which decay and soften by the action of air and moisture as some clays do; and the worst are those which are constantly soft and saturated with water, such as quicksand and mud.”

A Manual of Civil Engineering, W.J.M. Rankine (1862)

Surveying techniques were at first basic.  Plumb lines, lowered down shafts from ground level to that of the tunnel, were used to gauge its alignment; the result was that the horizontal sections of the tunnel didn’t always meet up perfectly.  The Saltersford Tunnel on the Trent & Mersey canal (opened in 1777) is far from straight; although only 424 yards long, one end cannot be seen from the other.


The tunnel’s alignment is established at surface level using telescope and staff.  Two plumb lines are then lowered down a shaft sunk to the tunnel’s level, aligned in the same direction as that established by the surveyor at the surface.  The two plump lines then indicate to the tunnelers the direction in which excavation is to take place.

The sinking of shafts would then be repeated at intervals along the tunnel’s alignment. Twenty-one shafts were sunk during the construction of the Blisworth Tunnel on the Grand Junction Canal, giving 44 working faces underground plus the two tunnel entrances.

A horse gin would be employed to work each shaft, raising and lowering men, materials and excavated debris to and from the subterranean workings.

By the age of the railways, tunnelling technique had improved.  A small central tunnel (or ‘head-way’) was first excavated as a further test of the strata to be passed through, to prove correct alignment and to act as a drainage duct for the main tunnelling operation ― the technique is thought to have been first used in the construction of the Blisworth Tunnel on the Grand Junction Canal.  Francis Conder, in describing the excavation of the Watford Tunnel on the London & Birmingham Railway, talks of the completion of the head-way . . . .

“. . . the pupil of the sub-engineer [Conder] . . . had the extreme satisfaction of viewing the red signal lamp, fixed at the north end of the head-way, from the southern extremity over a regular exact line of candles, one close to each shaft.  The question of the direction of the tunnel was thus solved.”

Personal Recollections of English Engineers, F. R. Conder (1868)

After the head-way was complete, the excavation of the tunnel was gradually enlarged.  Miners carried out the work, which commenced from each tunnel entrance.  In a long tunnel, a number of shafts, generally of about nine feet diameter, would be sunk at intervals along its route to the required level and excavation would take place from the bottom of each, working outwards in both directions.  A horse gin (later on, a small steam engine) would be installed at the top of each shaft for the purpose of raising or lowering skips or buckets.  The spoil removed from the workings by this means was carried away for use in the construction of embankments, or if there was no such need it was spread over the adjacent land.  During construction of the Blisworth Tunnel on the Grand Junction Canal, 21 working shafts were sunk which, including the two tunnel ends, gave 44 working faces.

Section showing a horse gin working a shaft (or ‘pit’),  the excavations proceeding in both directions.

Tunnelling in the canal-building era.

Blasting was resorted to when sinking a shaft and driving a tunnel through rock.  The technique was to hand-drill a shallow hole in the rock formation, which was then dried out with oakum, packed with black powder and plugged with clay.  On lighting the fuse, the miners clung, one above the other, to a winding rope and at a signal were hauled some distance up the shaft where they remained until the shots were fired.  They were then lowered into the choking fumes to remove the broken rock and debris, and then to repeat the process.  Unsurprisingly, accidents occurred when the miners were not lifted sufficiently high above the danger zone.

Tunnelling was hazardous in other ways.  Many lives were lost due to subterranean streams or quicksand deposits suddenly bursting through, or the tunnel collapsing due to unstable rock; such problems also meant extra time and cost to remedy.  Leakage of water into the workings was a continual problem and it was not unusual for steam-driven pumps to be required to prevent flooding until the tunnel’s interior could be lined, and sometimes even after that ― the railway tunnels under the Severn (1885) and the Mersey (1886) need to be pumped continually to prevent flooding.

Where flooding or quicksand proved a problem, or the formation of the rock was unstable, the tunnel interior was lined with brick.  Quicksand deposits were very difficult to tunnel through and required heavy brick lining . . . .

“The manner in which the brickwork is laid is of great importance . . . In a quicksand it has been found necessary to lay the lining to the thickness of twenty-seven inches in the sides and top, and eighteen inches in the inverts, Roman cement being used. [22]  This, however, is the greatest strength ever required; and, as the nature of the ground will allow of it, this may be lessened to the point where the material will stand by itself.”

Our Iron Roads, F. S. Williams (1852)

Clay required similar treatment.

It was sometimes necessary to build brick-lining into the sides of the shafts (following construction, some of these were then extended above ground to form small towers, and used for ventilation).  However, lining tunnels with brickwork provided no protection against subsidence, which could be very expensive to repair ― Brindley’s Harecastle Tunnel on the Trent & Mersey Canal became one such casualty, the cost of repair being so high that it was abandoned.

Few early canal tunnels had towpaths ― the Braunston and Blisworth tunnels on the Grand Junction Canal being examples ― so boat horses were led over the ground above and the boatmen, possibly with paid help, would ‘leg’ the boat through by lying on their backs and pushing with their feet against the tunnel roof, or lie on planks extending from the boat‘s sides (‘wings‘) and push with their legs against the tunnel walls.  Legging was a slow and arduous job, often taking two or three hours in a long tunnel and causing considerable bottlenecks, especially if the tunnel was too narrow for boats to pass.  Such was the case with Brindley’s Harecastle Tunnel ― Telford’s later tunnel, which was built to relieve the congestion, did have a towing path.  Some tunnels had ropes or chains connected to the walls with which to pull boats through.  Later, steam or electric tugs were used before powered narrow boats became common.



A flight of locks at Foxton on the ‘old’ Grand Union Canal. These locks have been built as a ‘staircase’ ― because the incline is steep, the upper gate of one lock forms the lower gate of the next.

A narrow boat descending Foxton Locks.  This is a narrow (7ft) lock ― that in the photograph below is broad (14ft 3ins). The man turning the windlass is raising the paddle to drain the lock chamber and lower the boat.


A lock enables a boat to move between sections of canal at different levels.  In this respect it is identical in purpose to a step, which performs the same function for a pedestrian ― moving from one level to another through a small vertical distance.  And as steps are built into flights to permit higher inclines to be negotiated, flights of locks perform the same function for a canal.

If rising ground that could not be bypassed was too substantial for taking the canal through in a cutting, crossing over it using flights of locks was generally considered a safer civil engineering option than tunnelling, assuming that was possible.  However, the choice of which technique to use was not always clear-cut.  Flights of locks slowed traffic, while providing a sufficient water supply to the summit could prove challenging.  On the Grand Junction Canal, the Chief and Resident engineers couldn’t agree on whether to use flights of locks or a tunnel to negotiate Blisworth Hill, so the canal company engaged two eminent canal engineers to decide for them.  Blisworth Tunnel was the outcome, but it has proved exceedingly expensive to maintain over the years.  Further south at Cosgrove, the valley of the River Great Ouse was originally crossed by two flights of four-locks.  These were later replaced by a substantial embankment and an aqueduct, both of which initially gave problems, subsidence in the embankment and structural failure in the aqueduct.  Thus, each approach for negotiating inclines has its pros and cons.

A broad lock.  The square aperture in the lock gate is the ‘paddle hole’.  The hole is blocked by a sliding panel, the ‘paddle’.  When shut, the lock can be filled with water from the upper pound; when lifted (by the ‘paddle gear’ at the top of the lock gate), water can flow out to the lock chamber to the lower pound.  Note the ‘invert’, the inverted arch of brickwork that forms the floor of the lock chamber.  This is built to resist upward and lateral pressures on the lock chamber ― in effect, reinforcement.

A lock consists of a chamber, usually built of brick and masonry, at each end of which are fitted one, or a pair of watertight gates.  Traditionally, lock gates are built of oak or elm and strengthened with ironwork.  The size of lock chambers vary between canals and even within the same canal. [23]  However, many canal locks will accommodate a 72ft by 7ft narrow boat, although narrow boats are now usually built to smaller dimensions to give them a wider radius of operation.  The change in water level provided by a lock is termed its ‘rise’; that for a single canal lock is usually in the range 7 to 12 feet.

A boat ascending from the lower to the higher level of a canal passes into the empty lock chamber and the gates are closed behind it to form a watertight seal.  Water from the higher level of the canal is then admitted to the chamber through valves, gradually raising the water level in the chamber and with it the ascending boat.  When the water level in the lock chamber equals that in the upper level of the canal ― the water pressure across the upper gates then being equal ― the upper gates are opened and the boat passes out of the lock chamber into the higher section of canal.  The same principle applies in reverse to a descending boat.

Using flights of locks to cross rising ground meant that the canal engineer had to solve the problem of supplying the summit of the system with sufficient water to carry the traffic, making due allowance for losses through leakage and evaporation.  Each time a boat crosses a canal summit, it uses two lockfulls of water, one on its ascent and one on its descent.  This quantity of water [24] flows down the canal and has to be replaced if the summit is to remain adequately flooded.  Because water resources become less plentiful on higher ground, a sufficient supply might require the construction of one or more reservoirs to capture and store drainage water from the surrounding area.  If the reservoir(s) was below the summit, steam pumping would be necessary to raise the water to the summit level ― pumping water from bore holes was another possibility to consider.  Thus, if either was required, the Chief Engineer needed to include its construction and operation in his costings (in periods of drought, the cost of pumping could make a significant dent in the canal company’s profits, as will be seen).

[Chapter V.]




Ordnance Survey maps began to appear in 1801, the first being of Kent at a scale of 1 inch to a mile.  The primary triangulation of the UK was completed by 1841.


A mercury barometer with which approximate heights (and thus gradients) along potential routes could be determined.


The railways were originally compelled to bypass Oxford and Eton due to objections raised by influential landowners.


Narrow boats were generally from 70ft. to 72ft. long, from 6ft. 9in. to 7ft. 2in. beam, and drew from 8 to 11 inches of water when empty, loading afterwards about 1 inch to 1 ton, but there was no defined standard.


Because barge horses sometimes fell into the waterway, for instance when a towing rope broke, slipways were built into the towpath at intervals to assist their recovery.


Triangulation is the process of determining the location of a point by measuring angles to it from known points at either end of a fixed baseline, rather than measuring distances to the point directly.  The point can then be fixed as the third point of a triangle with one known side and two known angles.  Levelling is a process by which the relative levels of points along the route were measured and compared with a datum level to calculate their elevations above it.


Among other documents, a prospective canal company was obliged by law to deposit with the Clerk of Parliament (and also clerks of the peace along the line) a scale plan showing the line of the canal in relation to the positions of notable features, such as churches, inns, mansion houses, roads/lanes, streams/rivers and springs.  It might also show the names of owners and occupiers, field boundaries and land use, which before the Enclosure Acts provide a valuable source of information for the local historian.


The Aire and Calder Navigation is an improved river and canal system that runs for 33 miles from Leeds to Goole, with a branch of 7½ miles from Wakefield to Castleford. Until well into the 20th Century it was very successful commercially, and today continues to carry some commercial as well as leisure traffic.


Usually at least 3ft 6ins.


In bygone times canals were usually referred to, even in Acts of Parliament, as ‘cuts’.


Coarse textured soils comprise mainly large particles in which there are large pores that permit water to seep through readily; without the application of an impervious liner a canal built on this type of ground would leak.  On the other hand fine textured soils comprise mainly small particles in which there are small pores that resist leakage, and in this type of ground a lining might be unnecessary.


Known variously as a roving, changeline or turnover bridge.  Such bridges were constructed to allow a horse towing a barge to cross the canal when the towpath changed sides.


The Oxford Canal’s busy northern section was straightened during the 1830s.


The Shelmore embankment took six years to build, while the 1¾-mile Woodseaves cutting is some 70 feet deep.


The rule followed by the civil engineer John Rennie Snr. (1761-1821) was that it was generally more economic to tunnel at depths greater than 45ft. than to excavate a cutting.


Tring Cutting at the summit of the Grand Junction Canal appears deeper than it is due to spoil being deposited along the tops of its walls.


Construction plant, such as mechanical road rollers, was not then available.


Brindley’s aqueduct was dismantled in 1893 to make way for the Manchester Ship Canal.  It was replaced by the impressive Barton Swing Aqueduct.


But unlike a road or railway viaduct, the weight on the piers of an aqueduct doesn’t increase when a boat crosses.


Details according to the Royal Commission on the Ancient and Historic Monuments of Wales.


Quicksand consists of fine sand, clay, and salt water.  The water lubricates the sand particles and renders them unable to support significant weight.


Developed by James Parker in the 1780s.  When made into mortar with sand, it set into a hard cement in five to fifteen minutes, far quicker than other cements of that period.


This lack of standardisation was to hinder canal inter-working, and was one cause of canals going into decline following the arrival of the railways.


The quantity depends on the size of lock.  For the broad locks on the Grand Junction Canal it is of the order of 55,000 gallons per lock.  Water losses in the summit pound through the bed of the canal and its banks, through evaporation, and through leaking lock gates also need to be replaced.