Keels, Ballast and Adjustable Draught

The hull’s immersed shape affects its resistance and support. Beneath it, a keel can add useful side area, carry ballast or do both. Those jobs must reach the structure that keeps the boat together.

The structural keel is a lengthwise backbone within the hull’s connected construction. An underwater keel is a projecting part of the immersed body. Calling both a keel does not make a timber backbone into a ballast weight, or make an external fin capable of supporting itself. Their different jobs meet at the hull, where the projecting part must find its support.

Put the mass inside or outside

Ballast is mass carried to influence the boat’s balance and stability. Internal ballast lies within the hull shell. Its weight reaches the boat through the structure beneath it, and its restraints prevent it shifting as the hull heels or accelerates. The enclosing shell keeps the surrounding sea out; it does not, by itself, secure a heavy object inside. A low, firmly restrained installation uses interior space and needs capable supports. ABS, permanent ballast, section 5.13.

External ballast lies outside the main hull shell, commonly in the projecting keel. Its weight must cross an attachment or enclosing structural connection to reach the hull. A ballast keel is an external keel carrying substantial mass. It usually supplies underwater side area as well, but its shape and its mass distribution answer different questions.

Internal and external ballast can share one boat. The Albert Strange Association’s record of the 1909 design for Tally Ho specifies both a ballast keel and internal ballast. That is a record of the design’s two locations, not a measurement of every later reconstruction. In the separately documented 2022 rebuild, Sampson Boat Co describes aluminium-bronze bolts passing through hardwood and lead. The later work shows an actual external attachment; it does not retrospectively establish the original fastener material. Albert Strange Association, Appendix III; Sampson Boat Co, Episode 123, 9 April 2022.

In either location, the chosen ballast mass has to fit into the boat’s available space. Lead is dense, so a useful ballast mass can occupy a comparatively small volume. Iron, often cast for a ballast body, occupies more space for the same mass. A lead keel identifies lead as a substantial part of the ballast; an iron keel identifies a different material choice. Neither name specifies the keel’s shape, its complete supporting structure or the strength of its attachment.

Take a declared, hypothetical mass of 1,000 kilograms. Using indicative densities of 11,300 kilograms per cubic metre for lead and 7,100 for grey cast iron, the lead occupies about 0.0885 cubic metres and the iron about 0.141. The iron body needs roughly 1.6 times the volume. These are reference material values; an actual casting’s composition changes its density. Equal volumes would therefore give different masses, while equal masses leave different amounts of space to accommodate. OIML R 111-1, page 58.

That extra volume may be distributed through a longer body, a thicker section or some other revised shape. It need not all become additional draught. The designer must fit it into the chosen water-flow shape or interior space. Density alone says nothing about the metal’s resistance to corrosion, the strength of a fastener or the hull’s ability to receive the load.

A denser material can fit the chosen ballast mass into less space, perhaps making a lower or more concentrated position practical. The same mass retains the same weight. Moving it changes the leverage, which is the next comparison.

Give depth a stated condition

Consider the ballast placement at one fixed heel angle, keeping the hull shape, total mass and other loading unchanged. Lowering the same ballast relative to the hull lowers the whole boat’s centre of gravity. When the boat is heeled, that movement also shifts the ballast sideways in the earth-fixed view. Its downward weight can then act through a larger horizontal righting lever relative to the buoyant support.

For an illustrative comparison, lower our 1,000-kilogram ballast by half a metre along the boat’s upright vertical direction. At 30 degrees of heel, that displacement has a horizontal component of a quarter of a metre. Taking gravity as 9.81 metres per second squared, the unchanged 9,810-newton weight adds about 2,450 newton-metres of righting moment in this fixed condition. It is the change due to that moved ballast, not the boat’s entire righting moment. Changing ballast mass or hull immersion would require a new whole-boat comparison.

A bulb keel concentrates much of its ballast in an enlarged body near the bottom of a narrower fin. The fin supplies much of the side-force surface and carries the lower body’s loads upwards; the bulb provides a compact place for mass. Bulb describes the enlarged form, not a promise that the ballast is lead or that every bulb has the same shape. Depth and the position of the bulb’s mass must still be stated.

The deeper arrangement also reaches farther towards the seabed. Harbour access, shoal passages and the place where the boat can lie become part of the design brief. If a deeper fin permits a smaller fore-and-aft surface for the required side force, that creates a hydrodynamic opportunity. Its actual resistance still includes surface friction, pressure effects and the flow around its ends. The complete shape and sailing condition determine the result.

Changing draught need not move ballast. The centreboard already taught rotates about its pivot, with the centreboard pin transferring load into the case and the surrounding hull. Raising it reduces the projecting side area and opens shallower water, but does not automatically lift a heavy ballast body. A daggerboard slides instead: its case supports the board along the guided contact region, without supplying the same pivot mechanism.

A leeboard obtains its support outside the shell. Its bracket and pivot transfer the board’s force into the side of the hull, leaving the interior free of a central case. The museum’s reconstruction of Mary, the Dutch yacht given to Charles II in 1660 for pleasure sailing and fleet visits, shows a shallow hull using leeboards. The model was made in 1949 from historical drawings and comparable plans; it is an interpretation of that earlier arrangement. It illustrates how shallow hull access and adjustable side area can be combined, without making the boards synonymous with external ballast. Royal Museums Greenwich, Mary.

The fixed ballast keel and these movable boards therefore trade different things. A board can reduce draught by withdrawing a side-force surface; a fixed ballast keel keeps its mass and projecting shape in place. A movable system that carries substantial ballast must also support and control that mass. We will examine that added machinery later, after following the fixed attachment’s loads.

Spread the keel along the boat or concentrate it

Look at an underwater profile before choosing a family name. How far does the keel extend fore and aft? How far does it project below the hull? Where does it join the shell, and is its after end connected to the rudder?

A long keel has a long fore-and-aft footprint. Full keel commonly describes a keel extending along much of the underwater hull, often blended into it and associated with a rudder at its after end. Usage overlaps, including between regional descriptions; there is no universal percentage at which long becomes full. Here both names identify that extended family, and we will specify any cutaway or separated part when it matters.

The Rustler 36 gives a concrete long-keel example: its forward region is cut away at the forefoot, while its rudder is carried at the keel’s after end. The immersed body therefore supplies side area over a substantial length rather than putting nearly all the projecting area into one short fin. The builder also describes long-chord fins on other models. Chord is the fore-and-aft extent of the fin; it does not by itself specify how deep the fin projects. Rustler, keel arrangements.

A fin keel is a comparatively short fore-and-aft projecting surface beneath the hull. Its span is its extension away from the hull; its chord runs from leading edge to trailing edge. A fin can contain distributed ballast, carry a bulb below it or combine several structural materials. Removing length from the footprint changes the available attachment region as well as the immersed shape.

The lateral plane helps compare those outlines, but it is only a side projection. A long keel, the hull and an attached rudder can distribute their water forces over a long region. A separated fin and aft rudder place substantial surfaces at distinct positions. Their areas, angles to the local flow and interactions determine the actual resultant force and its centre of lateral resistance. Moving the centre of a drawn area is not a complete calculation of that centre of force.

For a friction comparison, count the actual wetted surface of each complete assembly: hull, keel and rudder, including both faces and any fixed skeg. Compare at the same loading, heel, trim and speed through water, with comparable surface finish. Replacing an extended keel with a shorter fin creates an opportunity to reduce that total; a deeper fin or extra support also adds surface. If the average resisting force per square metre were equal, the assembly with less wetted area would have less frictional resistance. Changed shapes also change the local flow and pressure forces, so a smaller side silhouette cannot establish either wetted area or total resistance. ITTC, frictional resistance, page 51.

An early, explicit experiment joined these jobs in Dilemma, Nathanael Herreshoff’s wooden gaff sloop launched in 1891. His later account describes a metal plate fin with a lead body along its lower edge. Riveted angles at the upper edge connected the plate to vertical fastenings through floor timbers and sister keelsons. The low lead and the thin upper plate had distinct functions, while the timber hull received their combined loads. The surviving accounts differ in their iron and steel descriptions, so metal plate is the useful common description here. Herreshoff catalogue, Dilemma and the 1934 account.

The 1967 America’s Cup defender Intrepid supplies a later separation of fixed and steering surfaces. Its shortened keel and trim tab were followed farther aft by a separate rudder associated with a skeg, a fixed projection in front of the moving blade. The trim tab is a small movable surface at the keel’s trailing edge; the rudder is the separate steering surface aft. The skeg provides a fixed body ahead of that rudder and part of its support arrangement. Its complete bearing and steering-load route comes later. America’s Cup, Intrepid.

That underwater separation did not require a composite hull. Intrepid was built with double mahogany planking on oak frames. Model testing helped select the combination of keel and steering surfaces; the chosen forces still entered a connected wooden structure. Later keel and timber alterations also mean that a surviving yacht’s present fabric cannot simply be equated with its 1967 state. 12 Metre Yacht Club, Intrepid.

These examples do not retire the long keel. Long forms, short fins and intermediate arrangements continue to answer different draught, steering, accommodation, construction and operating requirements. Separating surfaces gives the designer more freedom to assign their positions and shapes. It also demands distinct, capable attachments at those positions.

Put supports to either side

Number and position form another comparison. A twin-keel sailing arrangement has two keels, commonly one to port and one to starboard. In a typical drying-out arrangement, their lower ends provide separated contacts on the ground. Each keel must transfer both its sailing loads and its share of the supported boat’s weight into the hull. Standing successfully also depends on the actual ground and the stability of that support arrangement.

Those outboard appendages are often called bilge keels, locating them near the region where the bottom turns into the sides. In one arrangement the two keels themselves carry ballast. In another, a central ballast keel remains and two outboard fins provide additional support contacts. The second boat has three possible ground-contact regions, despite the presence of a pair of bilge fins. Dave Gerr’s transverse comparisons show why counting and locating the actual parts is more useful than assuming a shared name fixes their ballast role. Gerr, twin-keel arrangements, page 1.

There is a further usage boundary. A ship’s bilge keels can be relatively small, narrow fins extending fore and aft near the bilge to damp rolling. Their movement through the water opposes part of the rolling motion. That job does not establish that they contain ballast or can support the vessel on the ground. The label needs its vessel and purpose attached. ITTC, keel and bilge-fin terminology.

The twin-keel sailing family also has documented development beyond a static outline. Mystic Seaport’s Arthur Robb papers list the 1961–1963 correspondence, design and specification material for the steel auxiliary cruising ketch The Bluebird. Designer Michael Kasten describes the Riverdale twin-keel work and his own 1995–1996 refit of Bluebird of Thorne. These records place continued twin-keel investigation within twentieth-century cruising design. They do not establish one inventor or make every paired arrangement equally effective. Mystic Seaport, Robb papers; Kasten, twin-keel research.

Two outboard keels change more than drying-out convenience. Their inclined or vertical surfaces meet different local flow as the boat heels, and their roots introduce loads at separated places in the shell. Added contact area also changes friction. The comparison therefore needs the actual shapes, immersion and force balance in the sailing condition.

Add wings to a keel

A wing keel has sideways-projecting wings, usually near the lower part of the principal keel. These are extensions of the underwater body, not a second pair of hull-mounted keels. Their geometry alters the flow around the lower end and can distribute ballast sideways and low down. Both the pressure forces on the wings and any ballast they contain must pass through their connection into the keel above.

Australia II, the 1983 America’s Cup winner, is a documented test programme rather than proof that wings improve every fin. Peter van Oossanen’s account of the 1981 development describes early winglets increasing side force and drag together. Their extra wetted area was one reason to revise the keel’s chord. Subsequent work also changed the hull, ballast distribution, rudder and winglet shape. The opportunity emerged from the combination and its testing, not simply from attaching two plates. Van Oossanen, wing-keel development; Australian National Maritime Museum, the 1981 programme and 1983 victory.

The names now describe different axes. Fin and long concern the principal surface’s extent; twin concerns number and placement; bulb concerns an enlarged lower ballast body; wing concerns sideways extensions. Lead and iron concern material. A keel can legitimately combine several of these descriptions. Each combination still needs sufficient water depth and a structure that can receive its particular forces.

Follow the fastening into timber

An external keel’s root is the region where its projecting structure meets the hull. A keel bolt is a fastening connecting that external body to the receiving structure. Several bolts can share the loads, but their spacing, end connections and surrounding material determine how they do so. A visible row of nuts is only one part of the assembly.

In a long timber arrangement, a ballast keel can introduce weight at several stations along the backbone. At a given station, an upper nut and bearing surface pass the fastening’s pull into substantial timber. A keel floor or floor timber spans across the lower hull, linking the keel region to the frames on either side. Keel timber supplies the lower lengthwise member, while the keelson above and any sister keelsons alongside connect several stations. The fastened planking and stringers carry the connection onward along and around the shell.

Herreshoff’s 1927 construction rules document one period solution: enlarged floors receive distributed ballast fastenings, with arrangements for the upper nuts and, in some cases, straps linking outside lead to the timber and planking. This is the same causal need visible in Dilemma’s shorter fin attachment, although the detailed connections differ. Once the fastenings deliver the load into the timber, that timber must carry it onwards. Herreshoff, floor timbers and lead fastenings, pages 8–9.

The hog timber, keelson and other backbone members remain structural pieces within that route. Deadwood can fill out and connect a deeper external shape, but its volume does not automatically make it dense ballast. Likewise, a false keel or keel shoe can take wear beneath a timber backbone without becoming the main ballast attachment. Their fastenings and positions decide which contact load reaches which member. A long dark strip in profile could represent several different jobs.

Internal ballast joins the same boat by another entrance. It bears on its supports within the shell, which pass load into floors, frames and the backbone; restraints prevent movement across those supports. Its weight does not first travel through an external keel bolt. Both locations eventually load the connected hull, but they do not share every intermediate fitting.

Spread a shorter root through the hull

A short fin reduces the fore-and-aft region available to introduce its load. The hull must spread that more concentrated input into a larger structure. A keel stub is a short hull projection to which an external keel is attached. A keel sump is a low internal recess in that region, often forming a place where bilge water collects. The stub describes projecting structure; the sump describes the internal space. Neither word names a second ballast body.

The J/105’s 1992 owner’s guide describes a locally substantial solid-glass stub with the lead keel bedded in epoxy and through-bolted to it. Follow the route from lead through the fastenings and their bearing regions into that structural projection, then into the hull. Bedding fills and joins the intended contact surfaces; it does not make the bolts or the receiving laminate unnecessary. The inside of the projecting region also explains the associated sump. This is one documented construction, not the detail behind every similarly shaped fin. J/105, 1992 construction details.

A keel matrix is a connected reinforcing grid that spreads keel loads into the surrounding hull. The keel grid and floor grid already encountered describe the connected members doing this work: transverse floors carry it across the root, and longitudinal girders carry it fore and aft. Their joints prevent the root being treated as an isolated patch of skin.

In a moulded construction, such a structural grid may be made separately and bonded to the hull. Then the load must cross both the grid’s own members and its secondary bond to the shell. A moulded liner serves as part of this route only where it is designed and connected to be structural; a surface that merely finishes the accommodation cannot be credited with the same job. A smooth interior can conceal the very interface that matters.

The Cheeki Rafiki investigation includes original drawings and a structural model of a lead keel with embedded bolt assemblies, upper washers, transverse floors and longitudinal girders. The fastenings transfer keel loads into the receiving laminate and grid; the bonded connection carries them into the wider shell. The model boundary includes only a selected region of that larger boat. It illustrates the route without making the rest of the hull disappear. MAIB, Annexes C and D.

Continue beyond that local region. The hull skin carries distributed shear and tension or compression between stiffeners. Floors, frames and longitudinals give it supported boundaries and carry load into adjoining regions. A structural bulkhead connected across the hull can participate through its attachments; the name bulkhead alone is insufficient. Where composite tabbing makes that connection, its fibres and bond transfer load across the junction. The load has to enter the tape, cross the corner and leave through the bond to the other member.

This is comparable to a chainplate bulkhead or chainplate laminate receiving a concentrated stay load higher in the boat. There the metal fitting introduces rig tension into reinforcement and surrounding structure. At the keel root, ballast and water forces enter through a different fitting and location. The comparison is the need to spread concentrated load across a capable interface, not an assumption that a chainplate component is hidden beneath every keel.

Keep the sandwich field separate from the attachment

The surrounding hull may use sandwich construction even where the keel attachment requires a different local structure. Across an ordinary sandwich panel, the inner and outer skins carry much of the bending tension and compression, while the balsa or foam core separates them and carries shear. Their core bonds make the layers act together. Neither balsa nor a light foam core automatically provides a suitable bearing surface under a highly loaded keel fastening.

If a concentrated clamp load simply squeezes such a core, core crushing can reduce thickness and clamping support. A designed solid laminate region, or an appropriate dense local load path, lets the bearing force enter capable material. A sleeve can prevent a fastener squeezing the core; a backing surface spreads bearing. The details must connect into the skins and reinforcement, rather than merely placing a stronger object in a weak hole. The ABS local-fitting drawings show these distinct duties; they are not a keel construction recipe. ABS, local sandwich attachments and connected structure.

Follow the load away from that solid region. Core shear still helps a surrounding sandwich field work as a panel. A debond between core and skin interrupts their cooperation; delamination within a skin separates layers that were meant to carry load together. At the edge of a stiff bonded member, bending can also create peel, pulling an interface apart rather than loading it only in shear. The transition into the surrounding hull is therefore part of the root design, even when the bolt itself passes through solid laminate.

Change the direction of the load

Start with a deliberately simple upright case. If the ballast’s weight acts vertically through the centre of an upright root, that weight chiefly pulls along the attachment. Greater depth alone does not give this aligned vertical force a greater bending lever. Sailing water forces and the boat’s motions can add other loads; we have temporarily set them aside to see the weight’s direction clearly.

Now heel the same boat. The ballast’s weight still acts towards the earth, while the keel and root have rotated. Relative to the hull, the force has both an along-keel and an across-keel component. The offset line of action creates bending at the root. Lowering the same ballast can increase that bending lever in the heeled condition, alongside its useful righting effect on the whole boat.

Water’s side force on the fin adds another load below the hull, with its own distribution and lever. A bulb’s concentrated weight and a wing’s distributed pressure need not act at the same place. A long or full keel spreads its connection farther fore and aft; a short fin introduces load over a smaller region. Twin ballast keels introduce it at two outboard roots, while a central ballast keel with separate bilge supports assigns weight and support differently. The chosen geometry changes the route and loading, not the requirement for a complete route.

Lead and iron return here as attached materials. Equal ballast mass gives equal weight in the same gravitational field, but a larger iron casting can require a different contact shape and fastening arrangement. A lead bulb beneath a narrow fin sends its weight through the bulb-to-fin connection before the fin reaches the hull. Internal ballast sends its load through internal supports instead. Neither equal mass nor a familiar external profile establishes equal root strength.

A grounding load path follows the contact force from the place where the boat meets the bottom into its supporting structure. If a keel strikes an obstruction while moving forwards, a rearward contact force below the hull creates fore-and-aft bending at the root. One region of the connection bears more heavily while another tends to open, increasing tension in the fastenings there. The floors, girders, bonds and shell must then carry the changed distribution onward.

A vertical contact under the keel supplies a different load direction. Planned drying out on two capable keels puts upward ground reactions into two supports; a dynamic strike, uneven bottom or sideways contact changes that situation. A wing or bulb may make the first contact at a different place from the fin’s centreline. The impact’s magnitude also depends on motion and how abruptly the boat stops. A keel intended to support a stationary yacht is not thereby certified for every grounding.

The MAIB annex analysis explicitly separates a heeled ballast case, vertical impact and longitudinal impact. That separation is useful even without adopting its numerical design checks: one test direction cannot represent all the others. A grounding can also damage a hidden grid-to-shell bond away from the visible external joint. The collision’s load path continues after the first piece has survived. MAIB, structural analysis; MCA, MGN 613.

Let the whole hull bend and distort

The connected hull responds to more than a local keel load. Its weight is distributed along its length, and the water’s upward support is distributed along the immersed shape. Where those distributions fail to balance locally, neighbouring regions must transmit the difference through the hull structure.

Hogging is longitudinal bending that raises the middle relative to the ends. A wave supporting the middle more strongly while the ends are less supported can produce that tendency. Sagging bends the middle down relative to the ends; relatively strong support towards the ends with more downward load between them can produce it. These describe deformation of the hull, not merely the boat pitching as a rigid body.

The deck, sides, bottom and lengthwise reinforcement work together as the earlier hull-beam analogy suggests. Their joints and openings determine how continuously they can transmit tension, compression and shear. Keel weight is one input to that system, but not its only cause. Mast compression pressing down amidships and fore-and-aft stays pulling at separated hull attachments can add a sagging tendency. Wave response and rig loading can therefore combine in the same structure. Greene, SSC-403, pages 35–37.

Racking is a different distortion. Imagine a transverse frame enclosed by the deck and bottom. If the top shifts sideways relative to the bottom, its corners change angle: an initially rectangular structural box tends towards a leaning parallelogram. A real curved hull need not become that exact shape for its transverse connections to undergo the same kind of angular distortion.

Unequal loads around the section, including reactions from the keel region and rig, must be carried through frames, deck connections, shell and effective structural bulkheads. Their shear stiffness and joints resist racking. Rotating the whole intact section in heel leaves its internal angles unchanged; racking changes those angles. It is also distinct from hogging and sagging, which describe the hull’s lengthwise bend. Park, racking deformation, pages 514–515.

Keep the complete attachment capable

Keel attachment integrity is the condition of the whole system that keeps the keel connected: fasteners, bearing surfaces, root, reinforcement, bonds and receiving hull. A sound-looking nut cannot establish the condition of a buried bolt or an inaccessible bond. Nor does a strong original design establish that the present structure retains its intended capacity.

Water ingress makes that distinction practical. Water entering an external joint or fastening passage can reach metal or bonded interfaces that are normally hidden. On a susceptible stainless fastening, an oxygen-poor wet crevice can promote crevice corrosion; pitting removes material locally and creates stress concentrations where cracks may begin. A small damaged region can matter more than its share of the total visible surface suggests. MCA, keel inspection and maintenance, sections 2.2–2.9.

Galvanic corrosion has the additional requirements already established: dissimilar metals, an electrical connection and an electrolyte. An iron keel, metal fastenings and seawater therefore require attention to the actual coupled materials and protection. An appropriately connected sacrificial anode can be part of that protection, but cannot restore missing bolt section, close a laminate crack or remake a debonded matrix. General corrosion, galvanic attack, crevice attack and pitting name related but distinct conditions; a stain alone does not identify which is present.

Composite deterioration also needs its own mechanism. Hydrolysis can damage susceptible resin chemistry after water exposure; osmotic blistering concerns water transport and pressure associated with soluble material. Neither is another name for an impact debond. Ultraviolet exposure can degrade susceptible exposed resin, whereas a keel-root interface hidden inside the hull is not receiving sunlight through the laminate. The location and material matter before the visible symptom is given a name.

Those distinctions identify different failure modes: loss of fastener section, a crack growing at a stress concentration, crushed bearing material, separation within a laminate or separation at a structural bond. Several can interact. Once one fastening or bond region loses its share of load, the surviving parts receive a different distribution; their original individual margins do not simply remain unchanged.

Serviceability includes keeping movement, cracking and leakage within the intended limits, before outright detachment occurs. A safety factor compares a stated capacity with a stated demand under stated assumptions. It is not a reserve that can be read from a nut’s size, or a number that certifies an unknown damaged assembly. Access to upper fastenings and nearby structure improves maintainability, but some of the most consequential surfaces remain beneath them or behind a liner.

The MCA guidance notes that fastening pressure can clamp a debonded region and make simple tapping misleading. The Cheeki Rafiki investigation likewise distinguishes the modelled original structure from the uncertain condition and failure sequence of the lost hull. Understanding these limits explains why the whole attachment matters; it does not turn this chapter into a method for clearing a damaged boat for service. MCA, MGN 613; MAIB, findings and analysis.

The ballast’s weight has now reached the wider hull through its attachments. Its righting contribution, however, was compared at just one fixed heel angle. As that angle changes, the buoyant support moves too. The next comparison must follow the changing relationship between weight and buoyancy.