# Workbench Hewn Tool Library · https://hewn.build/tools/workbench Corrected: An independent reviewer read this entry and found corrections. They have been applied to the text, and the reviewer re-checked the correction that mattered most for safety and found it done; the others have not been re-checked yet. The list at the end is the record of what changed. Cite as: "Workbench." Hewn Tool Library, hewn.build/tools/workbench. Reviewed 2026-09-09. --- ## identity **Name**: Workbench (also "work bench," "bench") **Regional / alternate names**: French *établi*; the heavy Continental form Roubo draws is sometimes called the "German workbench" in his own Plate 279 caption, though modern scholarship reads it as more likely Scandinavian in origin [W7] (a naming question the corpus should not resolve here — it belongs to the bench-type literature, not to this entry's own claim). The lighter English joiner's form is retroactively called the "Nicholson bench" by the modern hand-tool revival. Per Schwarz's terminology history (not checked against the 1812 plate itself in this pass), Peter Nicholson documented standard period shop practice in *Mechanic's Companion* (1812), the bench appears on Plate 12, and "Nicholson bench" is a later retronym [W4]. **Family / category**: Workholding → the host fixture. Every other entry in the corpus's workholding family — [[Holdfast]], [[Planing Stop]], [[Leg Vise]], [[Bench Dog]], the vises and stops — is *incomplete without this tool*, and this entry is incomplete without them: the bench is the platform, they are its infrastructure. **One line**: A stable work-support and workholding platform, sized and configured for the operations it must resist. Fixed joiner's benches, lighter knock-down benches, low benches and portable folding work centres solve different combinations of horizontal thrust, downward impact, clamping, space and transport. Mass can help, but frame geometry, bracing, sound joints or engineered fasteners, secure footing and appropriate anchorage determine whether a bench racks or walks. For many designs the top, stops, dogs and an integrated vise form one holding system, so the bench is a host for holding mechanisms and sometimes part of them. --- ## ✦ how NOT to use it **not-for**: heavy or repeated metal-on-wood hammering, hot work, welding and grinding (not assessed; see *matching*); a ladder, scaffold, step stool or standing platform, ever. **dangerous**: no documented bench-collapse or racking-related injury account was located; that is a research result, not a risk assessment, and the hazards below stand regardless. Predictable hazards of this tool: tipping or walking from unlevel footing, unlocked legs or an unbalanced or overloaded bench; racking under lateral force; a fall from standing on the bench; a workpiece releasing from a poorly clamped or cracked vise; long stock levering a vise; a struck vise jaw or hammered handle; striking concealed metal supports when cutting or drilling into a portable bench's top; flying chips or tool strike to the eyes; and heat damage to the top and clamping from torches or heat guns. Controls: the stability and workholding gate in *correct use* [W18][W19]. Practitioner texture (not evidence of frequency): a builder's bench racked under planing force [W12]. **damaging**: wood moving as it dries toward shop equilibrium can cup, bow, crook or twist a laminated top. Read the grade stamp, measure representative boards with a moisture meter, and acclimate material toward the expected shop equilibrium before final flattening and assembly. USDA Forest Products Laboratory guidance, as the reviewer relays it, defines dry softwood lumber at a maximum of 19 percent moisture content (some grades lower) and recommends matching moisture content to service conditions `[manual, relayed]` [W20]; no universal arrival moisture range is asserted for construction lumber. Flatsawn lumber tends to cup toward the bark face during drying (a tendency, not a guarantee), and bow is a separate lengthwise warp, not the opposite-face counterpart of cup `[manual, relayed]` [W21]. Benchtop delamination from glue-starved joints, uneven clamping pressure or boards not flat-jointed before glue-up is mentioned across woodworking forums at search-snippet level only. **beginner-trap**: an undersized or underbuilt first bench that looks stable standing still and only reveals itself under real lateral force — the Sawmill Creek racking thread is one practitioner example [W12]. Mistaking hardwood mass for a prerequisite: Sellers's own polemic is that beginners are pushed toward expensive hardwood benches as a completeness/status signal, when a bench "beyond many new woodworkers to build straight off" becomes an obstacle to starting at all [W10]. Wrong bench height for the user's own body is a commonly cited ergonomic beginner issue, but this run did not directly fetch a first-person account of resulting strain — carried at lower confidence, a reasoned claim rather than a sourced anecdote. **lineage-misuse**: treating a portable folding bench (Workmate-class) as a full substitute for sustained heavy hand-tool work, or treating a heavy traditional bench as necessary for light/portable site work — no first-person forum account of either specific confusion was located despite direct, targeted searching (Workmate-as-ladder/scaffold searches returned only generic ladder-safety material, not woodworking-specific misuse accounts). **This class is stated as a genuine, named gap, not filled.** The component-level equivalent (treating a leg vise like a quick-release vise, treating a bench dog like a holdfast) is already fully sourced in the corpus's own [[Leg Vise]] entry via the Klausz/Kirby confrontation and does not need re-sourcing here. --- ## correct use **What it's genuinely good at**: providing a flat, dead-still reference surface and a mounting platform for whatever workholding infrastructure ([[Holdfast]], [[Bench Dog]], vises, stops) the work in hand actually needs — hand planing, sawing, chopping mortises, chisel paring, and glue-ups all load the bench differently, and a bench sized/braced for one can feel wrong for another. **How to use it well (short version — the workbench holds deep bench-work operation)**: match bench mass and rigidity to technique, not to tradition for its own sake. Heavy hand-plane and mortising work loads a bench differently from lighter joinery and machine-assisted work. An underbuilt or improvised bench (sawhorses plus a door or plywood sheet) can walk or rack under every stroke. **Stability, not body weight:** use a stable base, lock or brace the legs, secure the bench to its base or floor where appropriate, or add low, mechanically retained ballast within the bench's rated load. Never stand on the bench or portable work centre. Forum accounts of a sawhorse-and-door bench that walked until anchored are search-snippet level only [W14] and are not relied on here. **Stability and workholding gate (check before use).** Drawn from CCOHS *Hand Tools: Vises* and the Black & Decker WM525 manual, both read in this pass [W18][W19]; for any other model the exact manual governs: - Level footing; legs fully open and locks or levers engaged and checked (WM525: confirm the Transition Lever and Toe Plate Slider are locked) [W19]. - The workbench firmly secured to its base [W18]; loads balanced, not uneven, since an uneven load could tip the unit [W19]. - Respect model-specific load limits (WM525: under 450 lb in Workmate mode, under 200 lb in Hand Cart mode; these are that model's numbers only) [W19]. - Inspect joints, top, vise and fasteners; check a vise for cracks or damage before use and do not use one with even a slight crack; keep screws clean and oiled; replace a bent handle and worn jaw inserts [W18]. - Choose a vise with enough capacity and do not open jaws wider than designed [W18]. - Clamp the work with the full jaw surface supporting it, close to the jaws; support extra-long stock separately (adjustable stand, saw horse or box) instead of straining the vise [W18]. - Apply even pressure and tighten both crank handles equally on a Workmate [W19]. - Keep the cut or drill path clear of the jaws and of metal supports under the top; the WM525 manual warns that drilling or cutting into the work surface can strike the metal support beneath [W19]. - Wear eye protection: safety glasses or goggles, or a face shield with them, when using striking or power tools on held work [W18][W19]. - Never use the bench as a step, ladder or standing platform ("Do not step on or stand on Workmate") [W19], and never use vise jaws as an anvil [W18]. --- ## ✦ matching **does**: provides a flat, rigid reference surface for stock prep and layout · resists lateral/racking force under planing and sawing strokes via mass and/or braced joinery · absorbs downward chopping and mortising impact without moving · hosts dog-hole and holdfast infrastructure for downstream workholding tools · sets a fixed working height for the user's body and stance. **matching classes (three rows, not one).** Fixed hand-tool joiner's bench; lightweight or knock-down shop bench; portable folding clamping work centre. They differ in mass, anchorage, clamping geometry, load rating, cut and drill clearance, and suitable duration and intensity of work. Identify the operation and the direction of force first (planing thrust, sawing, chopping and mortising impact, glue-up, assembly), then choose the class. Carving, chairmaking, green woodworking, site work, assembly and machine outfeed do not share one optimum surface; low benches, shaving horses, assembly tables, portable supports and vises designed to clamp to another stable surface are legitimate answers for some operations. **suits**: hand planing, joinery, chopping mortises and hand-cut dovetails reward a rigid, well-braced bench; the Schwarz and Sellers camps both hold that a construction-lumber bench can be a fully capable working bench [W8][W9][W10] (practitioner testimony, not comparative engineering proof). Power-tool-forward shops often lean on clamps, jigs and machine tables for part of the holding load. A first bench project can be a construction-lumber or viseless build [W8][W9][W11] (the Krueger system is carried at search-snippet level), always subject to the stability and workholding gate above. **kit-fit**: pairs with the corpus's workholding family — [[Holdfast]], [[Planing Stop]], [[Bench Dog]], [[Leg Vise]], [[Tail Vise]], [[Shoulder Vise]], [[Wagon Vise]], [[Moxon Twin-Screw Vise]], [[Doe's Foot]] — for the operations where they apply; some of them need a stable surface to bore into, clamp to or brace against. **Substitutes**: sawhorses plus a door or plywood sheet can serve casual, occasional hand-tool work once stability is added by the means in the gate above (never by standing on it) [W14, search-snippet only]. A portable folding work centre can serve site work, clamping and light assembly within its manual's rated load; no sourced comparison against a fixed shop bench for sustained heavy hand-tool work was found, a named gap. **What cannot be substituted** is stability under the intended load; a bench that cannot resist the expected horizontal and vertical forces without racking or walking does not fit the operation, whatever its joint type. **not-for (project-keyed) / not assessed:** metalworking, welding, grinding, forge work, hot work and heavy repeated metal impact are **not assessed here and require a purpose-rated bench and fixture system** with model-, load-, fire- and process-specific controls. Reasons from the sources: CCOHS says not to use vise jaws as an anvil and requires the vise and bench to be securely mounted [W18]; the WM525 manual warns that heat guns, torches and soldering irons can damage the work surface and reduce clamping grip [W19]; welding, grinding and hot work add fire, spark, load and electrical-grounding questions that a wooden joiner's bench entry does not assess. The earlier adjacent-trade forum consensus is withdrawn as a source. Also not a scaffold, ladder or step stool (see the gate above). **access**: the primary match is woodworking and light, non-hot assembly. A simple bench is accessible to a beginner (construction-lumber or viseless build, one vise) **only with the stability and workholding gate above applied and a competent source of instruction for the operations**; "safe from day one" is an outcome claim this entry does not make. The sourced record does not support the idea that real hand-tool work requires a heavy traditional bench first; that claim is contested between named authorities [W8][W9][W10][W11]. **revival**: the workbench is the one tool in this corpus a maker doesn't so much "pick up" as build — and building your own first bench, at whatever mass and complexity your skill and space actually support, is itself the entry ritual of the whole hand-tool tradition; Sellers's own point is that a bench beyond a beginner's ability to build immediately becomes an obstacle to starting the craft at all, not a milestone within it [W10]. --- ## anatomy - **Top** — the flat, thick working surface. Historically a single massive plank (Roubo, French units: 5–6 inches thick, 20–25 inches wide, see *history*) or, in the modern revival, a laminated slab of narrower boards glued face-to-face. Hosts the dog holes, holdfast holes, and vise mounting points documented in the corpus's [[Bench Dog]], [[Holdfast]], and vise entries. - **Legs and stretchers** — the frame that resists racking. Leg and stretcher sizing, bracing, joint condition and fastener design all bear on racking; the Sawmill Creek thread is one practitioner discussion of this, not a ranking of causes [W12]. - **Height** — Roubo gives an ordinary height of 30 French inches (about 32 modern inches, a conversion) and the rule that it should not exceed the top of the worker's thighs [W2b]; a widely repeated modern figure of ~34–36" is treated in this entry as **unverified folklore, not fact** — multiple blog sources attribute it to Delta's Unisaw/contractor-saw tooling decisions setting a de facto industry standard loosely matched to mid-20th-century average male height, but no primary Delta document, trade catalog, or standards-body citation was located or fetched this run. Landis reports that the Tarule reproduction was built taller (34½") for a modern practitioner's comfort [W2]; his 31¾" figure is not Roubo's stated number. - **Dog holes / holdfast holes** — see [[Bench Dog]] and [[Holdfast]] for the mechanism-level history; this entry notes only that their earliest confirmed appearance together is the Löffelholz Codex (Nuremberg, 1505). - **Integrated vise mounts** — see [[Leg Vise]], [[Tail Vise]], [[Shoulder Vise]], [[Wagon Vise]], [[Moxon Twin-Screw Vise]], [[Cast-Iron Quick-Release Face Vise]] for the mechanism-level entries this bench hosts. --- ## variations (living) *These are functional configurations (dog-and-vise, stop-and-holdfast, viseless, portable clamping work centre), not clean evolutionary branches.* - **The Continental/German-Scandinavian dog-and-vise bench** — heavy, tail-vise-and-dog-row system; already documented via [[Bench Dog]] and the vise entries. - **The French/Anglo joiner's stop-and-holdfast bench (Roubo, Nicholson)** — simpler, no squeeze system, matched to long straight-stock joinery work [W2][W4]. - **The Anarchist's/construction-lumber bench (Schwarz)** — buildable from dimensional construction lumber, with a heavy drawbored mortise-and-tenon approach [W8]; one documented design philosophy, not a universal requirement. - **The minimalist five-foot bench (Sellers)** — softwood, one quick-release vise, no dogs, no holdfasts, no tail vise [W9][W10]. - **The viseless bench (Krueger)** — holdfast + toothed stop + batten + doe's foot + crochet, no vise of any kind [W11]. - **The portable folding clamping work centre (Workmate and descendants)** — sawhorse-height, clamping top members, folds flat, travels [W5][W6]. Follow the exact model's manual. - **The engineered lightweight/torsion-box bench (Paulk-style)** — rigidity from box-section construction rather than raw mass; named in a forum thread title only [W16]. --- ## trades Primarily woodworking (joinery, cabinetmaking, furniture-making) and light, non-hot assembly. **Metalworking, welding, grinding and forge work are not assessed in this entry** and need a purpose-rated bench and fixture system (see *matching*). This entry does not claim cross-trade research. --- ## lineage *This entry sits underneath, not alongside, [[Holdfast]], [[Planing Stop]], [[Leg Vise]], and [[Bench Dog]] — each of those entries already carries a slice of workbench history (the Saalburg benches, the Löffelholz Codex, Roubo's Plate 279, Nicholson's Plate 12) because each needed the bench to explain its own mechanism. This entry is the first to treat the bench as the subject rather than the backdrop, and it deliberately does NOT re-source the individual vise/dog/holdfast/stop mechanisms already built as their own entries — see those files for the mechanism-level patent and provenance trail.* **Ancestors** - No confirmed pre-Roman physical bench evidence was located this run despite direct search for Bronze/Iron Age carpentry-bench archaeology — a real, stated gap, not a filled one. Claims of Egyptian "notched rocks" as a proto-bench surface only in tertiary aggregator material and are not carried into this entry as fact. - The general concept of a raised, stable work surface (as opposed to ground-level or lap work) is older than any woodworking-specific record; the corpus does not claim to have traced that origin. **Descendants / the fork the corpus's own workholding family already lives inside** - **The Continental/German-Scandinavian dog-and-vise bench** (Löffelholz Codex, Nuremberg 1505, onward) — already fully sourced in [[Bench Dog]]. Answers the stop's one limitation (it only resists thrust in one direction) with a squeeze system: a tail vise and a row of dog holes trap work between two points instead of pinning it against one. - **The French/Anglo joiner's stop-and-holdfast bench** (Roubo 1769, Nicholson 1812) — did NOT adopt the squeeze system, and the Fine Woodworking/Landis piece is direct on why: this was not a failure left unfixed, it was a different trade's different task profile. Roubo's subjects were architectural joiners working long, straight stock; they didn't need the vise-trapped grip a cabinetmaker working short, irregular pieces needed [W2]. **This is a case where forcing Petroski's "form follows failure" onto the whole bench-type axis breaks down** — the simpler bench wasn't behind, it was matched. - **The portable folding clamping work centre** (Ronald P. Hickman's Workmate, US 3,615,087, 1971) — a functional configuration aimed at space and portability for the site or tradesman's worker, not a woodworking-technique fix [W5][W6]. A 1971 patent is a landmark in this folding, integrated-vise lineage; it is not claimed to be the first whole-bench patent (see *history*). The failure it addresses is logistical. - **The engineered lightweight bench** (Paulk-style torsion-box construction, contemporary) — a live, named community question (a Sawmill Creek thread titled "Maybe a dumb question? Engineering vs pure mass for workbench (Paulk vs Roubo)") treats mass-based and engineered-rigidity benches as two approaches to the same racking problem [W16] — thread title only; the thread itself was not read, so no claim is built on it. **Contemporaries** — every mechanism in the corpus's workholding family ([[Holdfast]], [[Planing Stop]], [[Leg Vise]], [[Bench Dog]], plus the already-built [[Tail Vise]], [[Shoulder Vise]], [[Wagon Vise]], [[Moxon Twin-Screw Vise]], [[Emmert Patternmaker's Vise]], [[Cast-Iron Quick-Release Face Vise]]) is a contemporary infrastructure choice bolted onto or bored into this tool, not a separate lineage branch of the bench itself. --- ## history *Petroski's "form follows failure" lens is applied only where the sources support it; for the Roubo/Continental fork it breaks down, stated plainly rather than forced.* **The deep record — Roman benches.** Christopher Schwarz's first-person account of the two Saalburg Roman fort benches (recovered from well 49, excavated 1901) supplies details beyond what the corpus's [[Holdfast]] and [[Planing Stop]] entries carry: solid oak, low enough to straddle, roughly 11" to 12" wide; the legs on both were added after excavation, so the leg placement and standing height of the surviving museum objects is a reconstruction, not as-found; one bench was repaired; odd dovetail-shaped notches on one edge remain unexplained [W1]. **Dating, corrected:** per Schwarz's later Saalburg chapter as the reviewer reports it, one similar bench has a date of about 187 CE, while the bench examined was not directly dated in the cited account `[manual, relayed]`. An excerpt of that chapter read in this pass confirms the oldest-surviving-bench claim at "approximately 187 C.E." and the deposition in a Roman well, but does not itself state well 49, 1901, or an undated second bench [W1b]. The idea that the benches were put in the well to hide them in an attack, or as the fort was abandoned around 260 CE, is one hypothesis for deposition, not a demonstrated end-of-service date for either bench `[manual, relayed]`. No interval of service is asserted. As the corpus's [[Planing Stop]] entry already establishes, the benches show mortises consistent with a stop; no confirmed holdfast holes exist in them. **Roubo names the bench "the first and most necessary tool," and gives it real numbers.** André-Jacob Roubo's *L'Art du Menuisier* (Paris, 1769–1775): *"The bench is the first and most necessary of the woodworker's tools"* [W2]. From the Lost Art Press translation (read in this pass) [W2b], in the translation's French units: top thickness 5 to 6 inches, width 20 to 25 inches, length 6 to 12 feet (most commonly 9); ordinary height 30 inches, with the rule that the bench should not be higher than the top of the worker's thighs. The translator's note gives one French inch as about 1.066 modern inches, so these are about 21.3 to 26.7 inches wide and roughly 32 inches high in modern measure; those modern figures are conversions, not Roubo's numbers. The earlier "20–24 inches" and "31¾ inches" came from a secondary article (Landis) and are not used as Roubo's dimensions. The bench Roubo describes carries several holdfast holes in the top (typically eight or nine), three leg-holdfast holes in each front leg, rails and lower planking, a drawer at one end for small tools, a rear tool rack, a stop and crochet, and an optional front vise [W2b]. What the illustrated joiner's bench does not show is a tail vise or a dog row. A hands-on modern test of this spec is reported by Landis: timber-framer Rob Tarule built a reproduction for Plimouth Plantation and found it functional for period joinery without a vise [W2]. Landis, in his words: *"much of the bench's evolution since the mid-18th century has been icing on the cake — perhaps even superfluous"* [W2]. **Nicholson documents the English joiner's bench (1812).** Peter Nicholson (1765–1844), a Scottish architect, mathematician, and trained cabinetmaker, published *Mechanic's Companion* in 1812 (later republished as *The Mechanics Companion*, 1831); the bench appears on Plate 12 [W4]. The corpus's [[Planing Stop]] entry refers to this bench as "Nicholson's bench"; "Nicholson bench" may be a modern retronym, which is Schwarz's terminology history, and the 1812 plate itself was not quoted here [W4]. **Hickman's Workmate patent, and why it is not "the first whole-bench patent."** The earlier claim that US 3,615,087 was the first and only patent for a bench as a whole system is withdrawn. Earlier U.S. patents titled "Work-bench" cover whole benches: **US 581,829** (May 4, 1897, Stephen Peggy) describes a double work-bench whose central frame also serves as a tool chest, with front sections that slide inward so the bench can be narrowed for a doorway [W5b]; **US 1,112,729** (published Oct 6, 1914, Theodore Swanson) covers a carpenter's bench for planing board edges, with a vertically adjustable support bar and clamping hangers [W5c]. Both were read in this pass. No claim of priority is made here without a defined jurisdiction, claim taxonomy, search method and complete prior-art review. **US 3,615,087 A, "Workbenches," Ronald P. Hickman** (filed 1969-03-05, granted 1971-10-26, 14 claims) is a landmark in the folding, portable, integrated-vise Workmate lineage: a sawhorse-height bench with two opposed top members, one fixed and one movable, forming a vise [W5]. The bibliographic facts are corroborated across sources; the claim text was not read line by line, and the patent page was not re-fetched in this pass. **Litigation record:** the Science Museum Group's catalog record for a Sears "Craftsman Companion II" combined vice/workbench (object 2020-127) documents that Sears's version (built by Emerson, marketed as "Work Buddy") was found to infringe the Hickman/Black & Decker Workmate patents in US litigation, that Sears cited prior art including "an eighteenth century work bench by Dominy" and a 1928 jeweller's-chuck patent, and that Sears's appeal was thrown out in 1982 [W6] (carried from the original entry; the record was not re-fetched in this pass). This supports the commercial importance of the integrated-vise design, not first-in-class status. --- ## ✦ human side **A live, named, unresolved argument between two of the most credentialed authorities in the modern hand-tool world — and this entry does not pick a side.** Christopher Schwarz (Lost Art Press) is the central modern authority on bench history and mass-based design; his *The Anarchist's Workbench* (2020, free PDF) argues for a design "buildable from construction lumber with only mortise-and-tenon joinery, no metal fasteners," a form Schwarz traces spreading across Europe in the 1500s [W8]. A Lost Art Press blog post, "Clarifying the Correction on Wood Movement," is reported (search-snippet level only, not fetched) to correct a leg-lamination photo and caption in the book [W13]; no claim here depends on it. **Paul Sellers is the clearest, most quotable dissent, directly fetched in his own words.** From "Democratising Workbench Logic" (paulsellers.com, 2026): *"No one really needs anything bigger than my five footer, and I have never used a tail vise in my life… I don't use any kinds of bench dogs or holdfasts either. My woodworking life as a maker of fine furniture and every kind of joinery has been highly successful without any of this stuff. You don't need it, either."* [W9] Sellers built his own bench from construction-grade softwood/spruce and reports 61 years of daily professional use with just a five-foot bench and one quick-release vise. From an earlier 2013 post: *"we no longer need such mass… I would defy anybody to say that pine and spruce are any the less resilient for the task they perform… than hardwoods."* Sellers attributes historical bench mass to hand-cut joinery loads now largely offloaded to machines, and names a real social dynamic worth carrying into the matching block: heavy hardwood benches are sometimes built as **"workshop furniture"** — status objects rather than a beginner's first working tool — and that framing becomes an obstacle to a new woodworker actually starting [W10]. **A genuine point of agreement, not just conflict, between the two named camps.** Schwarz's construction-lumber Anarchist's bench and Sellers's construction-lumber five-footer converge on the same conclusion from opposite philosophical starting points: neither treats a softwood/construction-lumber bench as a compromise substitute for a hardwood bench — both treat it as a fully capable working bench in its own right [W8][W9][W10]. This is a rare, real point of consensus worth stating directly rather than only reporting the disagreement. **Rex Krueger — the named authority for the accessible, viseless entry point.** Krueger's "Joiner's Bench" and "Viseless Workholding" teaching (rexkrueger.com, Patreon) builds a complete no-vise system from holdfast, toothed planing stop, batten, doe's foot, and crochet — explicitly a cheap, low-material-investment entry point for a beginner, distinct from both Schwarz's mass thesis and Sellers's minimalist-but-still-vised bench [W11]. **A named practitioner thread on why one bench racked.** A 2010 Sawmill Creek thread ("why does my workbench rack?", #151571) discusses one builder's bench racking under lateral hand force. Contributors named undersized legs and stretchers (Gary Curtis, on his own small bench: *"If you are going to plane by hand, you need mass/mass/mass"*), inadequate tenon shoulder depth, bolted knock-down joints (Jeff Bratt: *"bolted together joints will never be as sturdy as glued and wedged through tenons"*), and poor floor contact (John Coloccia: uneven contact "will rack much easier… even a nice, heavy bench") [W12]. These are practitioner opinions about particular benches, not a controlled structural comparison, and they are not carried as universal rules. The neutral requirement: the bench must resist the expected horizontal and vertical loads without racking, walking or joint movement, which means assessing frame geometry, bracing, joint condition, fastener design, mass, footing and anchorage together; well-made knock-down or bolted frames can be rigid, and badly made wedged joints can fail. --- ## ✦ proof of existence - **Earliest surviving physical example**: the two Saalburg Roman fort oak workbenches, well 49, excavated 1901, one similar bench dated to about 187 CE and the examined bench undated in the cited account; no service interval asserted (see *history*) [W1][W1b]. No formal public accession number for the two benches was located this run — the Saalburg Museum's own English-language catalog page was referenced by Schwarz but not independently fetched, a real gap consistent with the same institutional-query gap the corpus's [[Leg Vise]] entry already names for other Roman-era objects. - **Museum/collection record**: Science Museum Group Collection, object **2020-127**, "Craftsman Companion II combined vice and workbench" — read in the original run; not re-fetched in this pass. Documents US patent-infringement litigation (Sears's "Work Buddy," built by Emerson, found to infringe the Hickman/Black & Decker Workmate patents; Sears's appeal thrown out 1982) and names prior art cited in that litigation, including "an eighteenth century work bench by Dominy" (not independently verified) [W6]. Supports the commercial importance of the integrated-vise design, not first-in-class status. Carried from the original; not re-fetched in this pass. - **Patents**: US 581,829 (1897, "Work-bench") and US 1,112,729 (1914, "Work-bench") are earlier whole-bench patents [W5b][W5c]; US 3,615,087 A, "Workbenches," Ronald P. Hickman (filed 1969-03-05, granted 1971-10-26, 14 claims) is the Workmate, a landmark in the folding, portable, integrated-vise lineage, not the first whole-bench patent [W5]. - **No pre-Roman physical or accessioned example located.** No Smithsonian-specific period joiner's or cabinetmaker's bench object record was located this run — a direct catalog search session, rather than general web search, is the honest next step, the same pattern the corpus's [[Bench Dog]] entry already flagged for the Mercer Museum. --- ## references Access status: "read 2026-10-16" means fetched in the fix pass; "original run" means read in Run 527 and not re-fetched; "relayed" means described by the reviewer; "snippet" means search-snippet level only. [W1] Christopher Schwarz, "2 Roman Workbenches at Saalburg Museum," Lost Art Press blog, 2017-06-10, https://blog.lostartpress.com/category/roman-workbenches/page/4/ — blog; original run (not re-fetched). [W1b] Schwarz, Saalburg chapter excerpt, *Ingenious Mechanicks*, https://blog.lostartpress.com/wp-content/uploads/2024/12/Ingenious_Mechanicks_4.pdf — read 2026-10-16 (excerpt only; truncated); remaining detail relayed from the reviewer. [W2] Scott Landis, "Throwback: 18th-Century Workbench," Fine Woodworking (orig. issue #65; republished 2024-10-03) — secondary article; original run; no URL recorded in the original. [W2b] Roubo workbench translation, Lost Art Press, https://blog.lostartpress.com/category/to-make-as-perfectly-as-possible-roubo-translation/page/39/ — read 2026-10-16. [W3] Löffelholz Codex, Nuremberg 1505 (Ms. Berol. Germ. Qu. 132, Jagiellonian Library) — carried via the corpus's [[Bench Dog]] entry; not re-fetched; no URL. [W4] Christopher Schwarz (as "fitz"), "Origin Book of the 'Nicholson Bench'," Lost Art Press blog, 2021-01-28 — blog; original run; no URL recorded. [W5] US Patent 3,615,087 A, "Workbenches," Ronald P. Hickman, https://patents.google.com/patent/US3615087A/en — patent; bibliographic facts only; not re-fetched this pass. [W5b] US Patent 581,829 A, "Work-bench" (1897), https://patents.google.com/patent/US581829A/en — read 2026-10-16. [W5c] US Patent 1,112,729 A, "Work-bench" (1914), https://patents.google.com/patent/US1112729A/en — read 2026-10-16. [W6] Science Museum Group Collection, object 2020-127, "Craftsman Companion II combined vice and workbench," https://collection.sciencemuseumgroup.org.uk/objects/co8641434/craftsman-companion-ii-combined-vice-and-workbench — museum record; original run; not re-fetched. [W7] Roubo, *L'Art du Menuisier*, Plate 279 — carried via the corpus's [[Leg Vise]]/[[Bench Dog]] entries; not re-fetched; no URL. [W8] Christopher Schwarz, *The Anarchist's Workbench* (2020), https://lostartpress.com/products/the-anarchists-workbench — book/product page; snippet; not fetched. [W9] Paul Sellers, "Democratising Workbench Logic," paulsellers.com, 2026-02-23, https://paulsellers.com/2026/02/democratising-workbench-logic/ — blog; original run; not re-fetched. [W10] Paul Sellers, "Question on Workbenches – Are They Becoming Fanciful," paulsellers.com, 2013-01-22 — blog; original run; no URL recorded. [W11] Rex Krueger, rexkrueger.com / Patreon, viseless workholding system — snippet; no URL recorded; no claim depends on it alone. [W12] Sawmill Creek forum thread #151571, "why does my workbench rack?", 2010 — forum; original run; no URL recorded; practitioner texture. [W13] Lost Art Press blog, "Clarifying the Correction on Wood Movement" — snippet; no URL recorded; no claim depends on it. [W14] LumberJocks / RenaissanceWoodworker forum account, improvised sawhorse-and-door bench — snippet only; not relied on. [W15] Lost Art Press blog, "A Backlash Against the French Bench" — title only; not cited in the body. [W16] Sawmill Creek forum thread, "Maybe a dumb question? Engineering vs pure mass for workbench (Paulk vs Roubo)" — thread title only; not fetched. [W17] Wikipedia, "Workmate" — tertiary; no longer carries any claim in this copy; replaced by [W5], [W6] and [W19]. [W18] CCOHS, *Hand Tools: Vises*, https://www.ccohs.ca/oshanswers/safety_haz/hand_tools/vises.html — safety authority; read 2026-10-16. [W19] Black & Decker WM525 Workmate manual, https://www.toolservicenet.com/i/BLACK_DECKER/GLOBALBOM/QU/WM525/1/Instruction_Manual/EN/588974-03_WM525.pdf — manufacturer manual, model-specific; read 2026-10-16 (summary of warnings). [W20] USDA Forest Products Laboratory, *Drying and Control of Moisture Content*, https://www.fpl.fs.usda.gov/documnts/fplgtr/fplgtr190/chapter_13.pdf — fetch failed 2026-10-16; relayed from the reviewer. [W21] USDA Forest Products Laboratory, *Drying Hardwood Lumber*, https://www.fpl.fs.usda.gov/documnts/fplgtr/fplgtr118.pdf — not fetched; relayed from the reviewer. --- ## The reviewer's corrections (9) - `BLOCKER` — 1. **Standing on a bench is not an acceptable ballast method — Exact claim: “Anchor an underbuilt or improvised bench … or add body-weight mass” and “weighted by standing on it.” - `BLOCKER` — 2. **Do not generalize a woodworking bench directly to metalworking or hot work — Exact claim: “the raised, rigid, fixture-hosting work surface generalizes directly to metalworking (with a dedicated anvil/hammering surface substituted or added).” - `BLOCKER` — 3. **Add the minimum stability and workholding safety gate — Exact claim: “a beginner can build and use a simple bench … safely and productively from day one.” - `SHOULD-FIX` — 4. **Withdraw the first/only whole-bench patent claim — Exact claim: “The first — and, this run's research suggests, only — patent for the bench as a whole system” is US 3,615,087 (1971). - `SHOULD-FIX` — 5. **Correct the Saalburg dating and separate object fact from abandonment theory — Exact claim: “the two Saalburg Roman fort oak workbenches … in service roughly 187–260 CE.” - `SHOULD-FIX` — 6. **Repair the Roubo dimensions and anatomy summary — Exact claim: Roubo specifies “20–24 inches” width and “31¾ inches” height, and the drawn bench carries “only a stop, a wooden hook, and an optional leg vise.” - `SHOULD-FIX` — 7. **Stop treating forum diagnoses as universal structural laws — Exact claim: leg/stretcher sizing and joint quality are “the single biggest factor,” bolted joints are categorically inferior, and “a well-jointed softwood bench with glued-and-wedged through-tenons beats a poorly-jointed hardwood one.” - `SHOULD-FIX` — 8. **Correct the wood-moisture and warp explanation — Exact claim: “construction lumber often arrives at 14–18% moisture content” and “the bark side of a board cups, the heart side bows.” - `SHOULD-FIX` — 9. **Tighten the source ledger and scope the matching claims — Exact claim: snippet-only W11/W13/W14/W16 and tertiary W17 support beginner safety, wood movement, substitute-bench behavior, engineered-bench taxonomy, and patent history; the matching block says “any hand-tool-forward project” benefits and several workholding tools “cannot” function without this bench.