Research Note

First, there is a large element of 'it depends' associated with scarf joints in general. Length of joint depends on a lot of factors, including who is building what for whom, by whose rules or directions, where, when, with what materials, and for what purpose.

Second, scarf joints have been in use for millennia. Ship builders have reached conclusions about what's good and what's not so good based on observation, experience and practice leading to custom, tradition, and standard. Hence length guidance along the line of Chapelle's "...not less than three frame spaces in any case" (1) and Lloyd's present rule of "...not less than 6 times the moulding [vertical dimension]..." (2) for keel timbers. In addition, Chapelle observes that "The line of the scarf is usually less than 13 degrees to the top or bottom of a timber." (3) A 13 degree slope yields a length to moulding factor of about 5 which is less than Lloyd's current minimum length. A 10 degree slope yields the factor of 6 that Lloyd's now considers a minimum for keel timbers.

Third, scarf joints with lips (perpendicular cuts at each end of the joint) 1/3 the breadth of the joint found in some cut kit parts for keels might be more for the kit maker's convenience than for historical accuracy. None of the photos and other images I've seen (modern and historical, some from archaeological recoveries) show such deep lips, or lips angled into the long portion of the joint. To the contrary, most appear to be relatively shallow. Lloyd's current rule states "The depth of the lips is to be about 1/4 to 1/7 of the moulding." (4) Keel scarf joints on plans from three AL kits (see picture) show deep cut lips that are angled – not perpendicular – and short joint lengths relative to the vertical dimension of the keel (about twice keel height compared to the minimum 6 to 1 ratio currently required by Lloyds) with surprising consistency over a span of about 300 years. And only the Mayflower has keel scarfs oriented so that forward movement through the water tends to compress, tighten, and hold the joints together.

Elsewhere, Harold Hahn's plans for the Sir Edward Hawke of 1767 show a stem to keel scarf joint three times the keel's height in length. (5)

Fourth, the smaller the cross-section dimensions of the timber the longer the scarf joint tends to be. This is based on observation and examination of photos. I'm uncertain about how widespread this relationship may be. It may, in fact, be peculiar to certain applications. I've seen it, for example, in the strip wood used to build kayaks where scarf lengths 8-10 times the width of a strip seem to be common.

Finally, Clayton Feldman, citing tabled data from "Steel's Naval Architecture" (1805) [see the Message Thread below], gives the example of 28 gun frigates that incorporated several keel scarphs measuring 3'8" in length with a moulded depth of 1'9" and 4" lips – a length:depth ratio of 2.1 and lips 1/5 the moulded depth of the timber. SOURCES:

1. Boatbuilding, by Howard I. Chapelle, W.W. Norton, 1994, Chapter Two, Lofting, pages 106-107
2. Rules and Regulations for the Classification of Yachts and Small Craft, Part 2, Hull Construction, Lloyd's Register of Shipping, 1979, Chapter 4, Wood and Composite, Section 4, Centreline Structure, 4.1 Wood keel
3. Boatbuilding, by Howard I. Chapelle, W.W. Norton, 1994, Chapter Three, The Backbone - Setting Up, on page 204
4. Rules and Regulations for the Classification of Yachts and Small Craft, Part 2, Hull Construction, Lloyd's Register of Shipping, 1979, Chapter 4, Wood and Composite, Section 4, Centreline Structure, 4.1 Wood keel
5. Ship Modeler's Shop Notes, Merritt Edson, Editor, Nautical Research Guild, 1979, Chapter 7, Hull Construction, pages 96-97.

Source
Source
Source Approximate length:width ratio of 10:1 --- Amistad replica keel
Source Square hole at center is for opposing wedges used to tighten and secure the joint
Source
SOURCE: Rules and Regulations for the Classification of Yachts and Small Craft, Part 2, Hull Construction, Lloyd's Register of Shipping, 1979, Chapter 4, Wood and Composite, Section 4, Centreline Structure, 4.1 Wood keel

To start your own research

See the web site here for photos and diagrams of joinery used in large timber framing of buildings, including several intricate and complex joints.

Message Thread

Longridge's 'Anatomy of Nelson's Ships' is a good starting place. If you are thinking of building a model accurately representing all the scarphs and chocked joints, you could check out: this site
{ David Antscherl}

go to the website http://www.bonhrichard.com/grosventre1en.html and click the april2004 issue. This newsletter has an article on axial timbers, where the cutting of the scarph joints is explained. Also on their forum is some explanation :

http://forum.aceboard.net/p-15916-1747-3925-0.htm
{Rudi Welter}

It appears that the cuts perpendicular to the length of the plank each account for one-third of the plank's width, with the middle third being given over to the long angled cut linking the perpendicular cuts at either end of the joint (see sketch below). How is the distance between the perpendicular cuts determined? Is it based on a multiple of the plank's width and/or thickness? Or the length of the plank? Does it vary from one kind of wood to another?

                                               ___________   ___
                                               |
                                               |             1/3
                                               |             ___
                                 _____.....-----
                  _____.....-----                            1/3
______.....-------                                           ___
|
|                                                            1/3
|__________________________________________________________  ___

|<--------------    How Long?    -------------->|
                      How Determined?

In Boat Building (pg 107), Howard Chapelle states, "The length of the scarf should be five to eight times the depth of the timber, or not less than three frame spaces in any case". The book is intended for small wooden boat design and construction, so this info may not apply to larger warships and merchantmen.

In addition to the length to width ratio of the diagonal of the scarf, I would like to know the orientation of the leading edge of the joint (closest to the bow). For the keel, is it always on top and adjacent to the hull? Also for horizontal scarf joints, like for waterways, covering boards, and cap rails, was the leading edge of the joint on the outboard or inboard side of the board? I hope I'm making myself understood here.

Looking forward eagerly to the responses,
{Roy Ozaki}

_______________________________________________________
|_____________________________ __ __ __ __ __ __ __ __|
|_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ |______________________|
|_____________________________________________________|

In addition, the two halfs were probably "coaked" - the "end" of each timber joint would have a "groove" cut lengthwise, and the matching part of the other timber would have a "tongue" to provide even more interlocking against bending movement.
{John O. Kopf}

This shows the different methods and types of construction.

If similar information is not there already then by all means add this drawing to the FAQ's
{Dave M.}

1. lengths of the angled portion of scarf vary from as little 2-3 times the width of the plank for things like siding and fence work to as much as 12 times plank width.
2. scarfs vary in form from simple, single cut, all taper to complex and intricate forms of joinery.
3. scarfs are used extensively in timber framed construction of buildings.
4. Lloyds apparently has mandated scarf lengths of 8 times plank width for current ship contruction.

For a simple, all-taper scarf joint (no cuts perpendicular to the length of the plank):

This site says scarf length should at least 8 times the thickness of the plank for an unstressed joint and 12 time plank thickness for a stressed joint

This site says 12 times thickness for thin planks and 8 times thickness for planks more than half an inch thick - plywood being the wood under discussion

This site says 8 times thickness for plywood 1 inch thick or more - this specifically for boat building

This site says 7 to 12 times the thickness of the plank - the context is rub rails

More complex scarfs ---

This site where a simple single-hooked scarf is shown in a kayak keel with the angled cut about 8 times the plank width in length

This site shows a "...stopped splayed scarf joint with undersquinted butts" being cut for a barn that looks much like a hooked scarf with an angled cut length of about 5 or 6 times the width

Genuine 'gold' paydirt at: This site where shear clamp planks are shown being scarfed together during 2003 construction work on the replica of the 1916 pilot schooner Virginia. One of the photo captions notes that "The length of joint is based on 8:1 ratio per building regulations from Lloyds." It appears that the perpendicular cuts are about 1/6 the width of the plank instead of 1/3.

Really complex scarfs ---

And then there's the wedge set hooked scarf used in large timber construction; see: This site wherein the long angled portion appears to be 4 to 5 times timber crossection.

And even more intricate, the splay stopped scarf with bridled butts at This site also used in house construction - this with diagrams and photos of the joint.

And farther afield there are the ALLOC and UNICLIC scarf joints used for flooring materials; see: This site and scroll down.
{Mike Graff}

The quotes of 8 and 12 to 1 is for scarf joining panels or planks by there thickness not there width.
{Dave M.}

Opinions expressed about the strength of such joints vary from the joints being stronger than unscarphed timber to such joints weakening the structure in which they are used.

I have yet to see in any of this, whether the discussion is of scarfs or scarphs, an example of a joint in which perpendicular cuts extend 1/3 of the way into the timbers at each end. The deepest cuts seen in all of this looking through web sites appear to be 1/4 at most and, more often, something on the order of 1/6 or less.

I'm coming to the conclusion that, setting aside the mentioned Lloyds modern requirement that such joints have a length to depth ratio of 8 to 1, scarf/scarph joints and their dimensions fall into the broad category of 'it depends' when it comes definitions and/or standards.

Given the amount of detailed information available about mast and yard dimensions and shapes, it seems unusual that there is not comparable guidance available concerning the scarfing together of both major and minor structural members of ship's hulls. Am I (are we) looking in the wrong places?
{Mike Graff}

If you like formulas, Goodwin developed them for all sorts of things from plans and direct observations for the classical era of sail. The keel scarphs formula can be found on pages 6-7 of his "The Sailing Man of War 1650--1850". The results come out similar to the Steel tabular data.
{Clayton A. Feldman, MD}

The scarf length for large timbers is usually 5x to 8x the depth of the timber (note Lloyd's requirement of 8x, and Chapelle's comment of "not less than three frame spaces in any case".) For planking and other thin materials the scarph length is usually 12x the material thickness. Scarfs in keel timbers must be stopwatered if the scarf cuts across the planking line. When the scarf is horizontal, the 'exposed tip' (bottom) should point aft. However, be aware that keel scarfs (especially in 18th cent construction of large ships) are often vertical, with the scarph profile is visible on the top and bottom of the keel, not the sides. Such a joint must be well fitted and probably buttered with tar, but not, of course, stopwatered.

Roy asks about scarfs on covering boards and cap rails, etc. Follow the same guideline... the outside or exposed part of the scarf should point backwards, which can minimize damage in case of impact.

There are a great number of variations in scarf (scarph) joints. The very simpliest scarf is simply to cut long angles on both pieces, and join them with glue and/or mechanical fasteners. This form, usually at 12:1 slope, is commonly used today for glued scarphs of plywood panels. The 'nibbed scarph' pictured by Mike is a simple scarf with the ends terminated in blunt ends. The nibbed ends avoid a feather edge which is difficult to fasten and provides resistance to compression .. the parts won't slide past each other, though they can still be pulled apart. The hooked scarf that John Kopf describes addresses expansive stress, with the hook in the middle of the long angled part preventing the parts from pulling apart.

A further refinement of this leaves a small gap at the hook, into which contrasting wedges are driven to tighten the joint. A slightly reversed angle on the end cuts helps secure the tips, locking them from slipping under the pressure of the wedges, and also provides some resistance against vertical pressure on the joint.

Such a joint is still weak to sideways stress, and the simplest solution to this problem is as John mentioned - to use tables or coaks within the mating surfaces. There are other more complicated solutions, including putting a broad point on the end cuts (which places resistance to lateral stress at the extremities) and other complicated interlocking mortices within the mating surfaces.
{ David Hill}

--- Mike Graff wrote: Am I (are we) looking in the wrong places?

The document includes a great deal of information about materials, fastenings, and how things are to be done. It is something of an education in boat/ship building all by itself.

If there is an interest in seeing 'the whole thing' I could post a copy (briefly) in the SSL Files section. I definitely want to avoid sending copies out to individual requestors.
{Mike Graff}

Rules and Regulations for the Classification of Yachts and Small Craft, Part 2, Hull Construction, Lloyd's Register of Shipping, 1979, Chapter 4, Wood and Composite

Section 4, Centreline Structure, 4.1 Wood keel

4.1.3 When the length, L, does not exceed 10 m [meters] the wood keel is to be in one length. In larger craft, the keel should, where possible, be in one length but when a scarph is necessary in the centerline structure it is to have a length, L, not less than 6 times the moulding, M, of the item. The scarph is to be of the hooked or tabled type if bolted (see Fig. 4.4.1), or plain type without lips if glued. The depth of the lips is to be about 1/4 to 1/7 of the moulding. [Moulding (M) is shown as the vertical dimension of the keel timber, and length (L) is shown as measured along the long dimension of of the keel timber, NOT along the length of the angled portion of the scarph.]

4.1.5 Scarphs in the keel and hog are to be at least 1.5 m apart, and the keel scarph, where fitted, is to be clear of engine seating and, if practicable, of the mast step.

Section 5, Framing, 5.4 Grown Frames

5.4.1 Grown frames are to be cut to shape from timber having the required curvature of grain.

5.4.3 Grown frames may be butted or scarphed. Scarphs are to be glued and have a length not less than 6 times the siding. Where frames are butted, the butts are to be close fitted and side clamps arranged. The clamps are to have a sectional area not less than that of the frame and a length not less than 12 times the frame siding. The clamp is to be through fastened to the frame by not less than three fastenings on each side and is fay closely to the planking.

Section 9, Hull Planking, 9.2 Single Skin

9.2.7 The length of a scarph is to be not less than four times the thickness of the planking. The scarph is to be positioned on the frame, and glued and fastened to it.

9.5 Plywood Planking

9.5.3 Butts and seams are to be scarphed or strapped where necessary. The length of a scarph is not to be less than 8 times the hull thickness. The scarph is to be glued and, if made in situ, fitted with a backing strap of a width not less than 10 times the hull thikness. The strap is to be glued and fastened to the hull with two rows of fastenings of the size given in Table 4.9.4 and spaced about 8 times the hull thickness.

Boatbuilding, by Howard I. Chapelle, W.W. Norton, 1994, Chapter Two, Lofting, pages 106-107 says,

If the plans call for a keel in a single length and this cannot be obtained, it is permissible to scarf the keel. The location of this should be chosen with care so that the scarf bolts will not interfere with the bolts in the floor timbers or those for the ballast keel. The length of the scarf should be five to eight times the depth of the timber, or not less than three frame spaces in any case.

The line of the scarf is usually less than 13 degrees to the top or bottom of a timber.

The Oxford Companion to Ships and the Sea, Peter Kemp, editor, Oxford University Press, 1976 and 1988, page 756 says,

SCARF or SCARPH, the joing of two timbers by bevelling off the edges so that the same thickness is maintained throughout the length of the joint. In the construction of a wooden ship, the stem and sternposts are scarfed to the keel. A scarf which embodies a step in the middle of the joint, so preventing the two parts from drawing apart, is called a lock scarf. It is a joint of great antiquity, having been used by the early Egyptian and Phoenician shipbuilders.

Just be careful in applying the Lloyd's data- It's twentieth century stuff:

Rules and Regulations for the Classification of Yachts and Small Craft, Part 2, Hull Construction, Lloyd's Register of Shipping, 1979, Chapter 4, Wood and Composite

Probably similar to the prior century's required practice, but you need different sources for the classical era of sail.
{Clayton A. Feldman, MD}

Back?