17th, 18th and 19th Century Clamping

I’ve received a lot of questions about clamps, and specifically, how they clamped things in the 18th century, so I thought I’d do a post on it. While I don’t know for sure everything that they used clamps for 250 years ago, I can offer a somewhat educated guess based on some historical information and my own work habits in the shop. It’s an interesting topic indeed, since today, we tend to have the mentality that he who has the most clamps wins. However, 250 years ago, it may not have been so.

There is historical evidence that clamps made of iron and or wood have been used since at least the 17th century (I haven’t researched the topic any further back than that). Several 17th, 18th and 19th century texts and images speak of or picture some kind of clamp. It is, however, unlikely, that clamps as we know them were as heavily relied upon as they are in today’s modern shops. At least not for the tasks they tend to be most used for today.


From what I can gather from the few hitorical sources I’ve had the opportunity to browse, clamps were used more for work holding and gluing panels rather than assembling pieces. Hand screw clamps like the one pictured above are invaluable in the shop for myriad work holding tasks. They’re capable of exerting bone crushing pressure and hold abnormally shaped pieces well as the jaws may be angled some. If you notch the jaws, they also hold round work very well, and they can also be held in a bench vise for shaping or carving tasks.

Traditional joinery, such as drawbored and/or wedged mortise and tenon, and dovetails, make the use of clamps unnecessary for assembly tasks. In fact, these types of joinery can essentially be assembled without glue (and typically were in 17th century joined work), if they are well fit, and will stay tight indefinitely. This is because the mechanical advantage of these joints is their primary strength. The glue is just a secondary measure.

By contrast, many modern joinery methods have no such mechanical advantage. Joints like dados, biscuits, dowels, and cope and stick have no mechanical advantage. These joints rely on the glue alone to hold things together, which creates the need for dozens of clamps to hold things together until the glue cures. The problem is, when the glue fails (and one day it will), these joints will fail as well as a result.

In my own experience, I have found this to be generally true. On occasion I may use a clamp as a third hand to help steady a long or complex assembly until I can drive the pegs or wedges, but for the most part, I rarely use clamps for assembly tasks in my own shop. In fact, I’ve sold most of my clamps as I found having so many of them unnecessary. All they did was take up valuable space in my already tight shop. I haven’t missed them at all.

Where I do use clamps more often is for panel glue-ups. Small, two board panels and edge joints are easily joined with no clamps using hide glue and a rub joint. However, longer panels, and those glued up from more than two boards, benefit from a few clamps to hold things in place. The purpose of the clamps, however, is simply to hold the boards in place, not to close a bad joint. If your joints have gaps when dry fit, clamps won’t solve your problems. The joinery must fit well.


The simplest of these clamps is the iron staple or pinch dog, pictured above right. When they were mentioned at all in a period inventory, they were typically referred to as staples (at least that’s my interpretation). The wedge shape of the pointed legs serves to pull the two adjacent boards tightly together when driven into the end grain. These are best used on shorter panels or in conjunction with a different kind of clamp as they only serve to hold the ends of the boards and do little at the center of a long panel.


Roubo pictures several other types of clamps in his 18th century treatise. While I cannot read a lick of French, I can understand the picture, which pretty much speaks for itself. I would hazard a guess that the top clamp would be made of iron by a smith, and tightened with a tap of a mallet, similar to using a holdfast. The bottom clamp would be made of wood by the joiner/cabinetmaker, and tightened using wedges. Roubo even pictures a panel glued up from four different boards in this clamp. I’m not sure, but the middle engraving may be a close up of the shaft of the top clamp? Then again, it could be something different altogether. If anyone who can read French knows for sure, please chime in.


Des Principes de L’Architecture pictures clamps almost identical to those in Roubo’s work, on the shelf above the workbench. Again, the one on the left would be made of iron and the one in the middle of wood. The one on the right bears strong resemblance to the double screw pictured and described in Moxon. The rear jaw of this appliance would be fastened to the bench top with a holdfast or two, as per Moxon’s description. However, there is no obvious way of attaching this appliance to the front edge of the bench, as it is pictured and described in Moxon. Perhaps the Moxon engraver got it wrong? I think of this double screw clamp as the precursor to today’s hand screw clamps. Peter Follansbee just wrote about this in much more detail. My special thanks to Peter Follansbee as well, for graciously allowing me to “borrow” his image of Felebien’s plate.


The cabinetmaker’s shop in Williamsburg has yet another type of clamp. This kind of clamp is probably the most familiar looking to most of us as it’s basically a wooden version of a bar or pipe clamp. I’m not sure where they found the reference to this style of clamp, but if it’s in their shop, it was likely available in some form in the 18th century. I personally find these clamps particularly appealing as I think the wooden screws just have a high coolness factor. I think I’m going to need to make myself some of these.

Interior of the Carpenter's Shop at Forty Hill, Enfield ?exhibited 1813 by John Hill circa 1780-1841

The last picture I have is a painting called “The Carpenter’s Shop” and was painted by John Hill in 1813. At first glance, it doesn’t seem to show much in the way of clamps. The picture is a little dark, but look at the enlarged version, at the extreme lower left corner. Leaning up against the window is what appears to me to be a panel clamp with a screw, similar to the ones used in the Williamsburg cabinetmaker’s shop.


10 thoughts on “17th, 18th and 19th Century Clamping

  1. Bob,
    I must admit I like a little bit of “spring” to my edge joints; a slight gap between the boards in the middle to compress the ends. It’s easy to do and adds a level of security, particularly when the moisture in the stock is in question. At any rate, it certainly minimizes the number of clamps; clamp the middle and the ends are tight. The staples line things up well though.

    • Rick,
      If the spring joint works for you, that’s what’s important. I agree that it does reduce the number of clamps needed. However, for my own work, I don’t completely buy into the theory.

      My take on sprung joints is that they solve a problem that we really don’t have with today’s kiln dried wood. I’ve never experienced any separation at the ends of a panel glue up, and I don’t spring any of my panels. I understnad the theory for times like you mention when the moisture content is questionable. However, I think it’s a bandaid type fix. In my opinion it’s a solution that doesn’t address the true root cause of the problem, which is excess moisture in the boards. If a joint is opening up over time, it’s likely because those boards weren’t dried completely before use. In properly dried stock, this should not happen. In my opinion, the solution is not to spring the joint, it’s to allow the wood more time to dry. But again, I don’t think modern kiln dried lumber really has this problem if it was dried properly.

      I think a very mild spring in the joint is ok, as long as it can be closed by hand. But if the spring is such that it requires clamp pressure to close the joint, in my opinion, the joint has too much stress in it and there is risk of failure later in the panel’s life. At some point in time, that glue will begin to degrade. Stress in the joint could serve to accelerate the joint’s failure. Of course YMMV, and this is all speculation and opinion.

      One important note though is that rub joints cannot be sprung. In order for a rub joint to work, the two mating surfaces must come together completely flat. Any gap will weaken the bond. So I typically just plane all my panel glue ups with flat edges. I’ve never had a problem.

  2. Hi Bob,

    Enjoy the blog. I speak a little french, but I couldn’t read the script on the Roubo plate that you needed a hand on. I guess the resolution is low on my screen?

    If you post a higher resolution image or email it to me, I might be a be able to take a stab at it for you.

  3. Like Dallas, I can’t read the French in the diagram but could tell you what it says if I could 🙂 My guess, however, is that the middle graphic is an example of those “Williamsburg clamps” for gluing up panels.

    Do you know whether the screw clamp in your first graphic has the base screw tapped all the way through both blocks or if it stops/seats in the second block (under the hand plane)? I note that it is not double/reverse threaded like modern versions of these clamps.

    • Thanks for the offer to translate. Unfortunately, the image I have is very low res and just isn’t clear. I have a file that has several pages of text, but that plate in not included in that file. I don’t have the full Roubo volumes unfortunately, not that I could read them even if I did. Hopefully the translation that Chris Schwarz, et. al. are working on will include these sections.

      As for the Williamsburg hand screws, the way these work is that the middle screw is a “pull” screw with the jaw closest to that screw’s handle being bored as a through hole and the jaw furthest from the middle screw’s handle is threadded in that position. At the back screw (furthest from the tapered end), the action is a “push”. The jaw closest to the back screw’s handle is tapped, and the jaw furthest from the back screw’s handle is bored only part way through the jaw so that the screw pushes the back of the jaws, providing lots of clamping force at the front.

      • I didn`t know Chris Schwarz was working on a translation. That`s nice to hear. You hear so much about Roubo (and you can watch Roy run breathlessly through the french terms on his show – pretty funny), but it would be cool to see what his texts are like in their entirety.

  4. Hello Bob, I really enjoy your blog. Thanks for taking the time to share your experiences with us.

    My take is that the middle engraving you mentioned is a precursor to the Williamsburg clamps, only that wedges would have been employed at the toggled end. The end of the Roubo version has a slight hook (I believe to enhance the “grab”) where as the Williamsburg version has a screw. Again, this is just my guess. I won’t even offer it up to being an educated guess. I’s ain’t edgukated all dat mutch.

  5. Bob,

    I hope you have your blog set up so you can read comments on older posts … my question belongs here, I think.

    As a woodworking newbie (who already knows she wants a hand tool only shop), I’m curious about what you think is an essential, minimal set of clamps? Looking at some other podcasts, I see thousands of dollars hanging on the clamp rack alone. I know you got rid of many of your clamps, I wonder what you kept and why.

    • Liz,
      I think the need for thousands of dollars in clamps stems from inferior joinery methods, like dowels, biscuits and cope and stick. These forms of joinery have little to no mechanical connection and therefore require a lot of clamps to hold assemblies together while the glue dries. If you design and build your pieces using traditional joinery like dovetails and pegged mortise & tenon, very few clamps are required.

      As I mentioned, I use clamps for two main purposes in my shop; workholding and large panel glue ups. I do not clamp dovetailed assemblies or pegged or wedged M&T joints. They simply don’t require clamps. Nor do small edge joints, joined with a rub joint and hot hide glue.

      For workholding, nothing beats a pair of handscrew clamps like the one in the top picture. I recommend a pair of 12″ hand screws and a pair of 6″ or 8″ hand screws for holding smaller stuff. Cut a “V” shaped notch in the jaws of one clamp from each pair to provide a great way for holding odd shaped and round parts. You can see how useful this feature is in Quick Tip #5 (part of the Techniques podcasts).

      For large panel glue ups, I’d recommend four pipe or bar clamps, all about 3′-4′ long. I think you can get by with nothing more than that. I occasionally will use an “F” style clamp or two for securing something to the workbench or saw bench, but you can also use the handscrews for this, or a holdfast, so the “F” clamps really are not a must have in my opinion.

      One other thing to not overlook are nails. Nails were used extensively in period pieces, for attaching case backs, drawer bottoms, moldings, or holding things like knee blocks on cabriole legs in place while the glue dried. A few strategically placed nails can sometimes replace an entire wall full of clamps. And the nails don’t always need to stay in place either. They can be used much as a clamp might be, and removed after the glue has dried. They are also frequently used in place of glue when we want to allow for wood movement (e.g. attaching cross grain moldings or drawer runners/kickers). Nails have an important place in my shop.

Comments are closed.