# Dividers

Hewn Tool Library · https://hewn.build/tools/dividers

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: "Dividers." Hewn Tool Library, hewn.build/tools/dividers. Reviewed 2026-08-01.

---

## identity

- **name:** **Dividers** (also "compasses" in period sources and much everyday usage — see *modern task-boundary distinction* immediately below; historical catalogs and object records often use the two terms loosely or describe combination instruments, but that doesn't make present-day task boundaries interchangeable); specific forms carry their own names: wing dividers, spring dividers (spring-bow dividers), trammel points (beam compass), proportional dividers [DV3][DV20].
- **modern task-boundary distinction (per the reviewer's 2026-08-01 review):** three related instruments now do three distinct jobs. **Dividers** normally carry two points and are used to compare or transfer spans and step intervals. A **drawing compass** normally carries one locating point and one pencil/pen point, and is used to draw or mark a line. **Inside/outside calipers** are formed to compare internal or external dimensions where straight divider points may not reach or seat correctly. This entry uses that modern boundary except when directly quoting a period source that uses the older, looser vocabulary.
- **family / category:** Measuring & layout → the transfer branch — a two-pointed pivoted instrument that moves a physical distance from one place to another without reading or recording it as a number. A sibling of the corpus's own [[Scratch Awl]] (the awl's stepped-layout partner, per Scratch Awl's own citation) and a close cousin of the drafting/pencil compass, which is mechanically the same tool body with one point swapped for a marking lead [DV29].
- **one-line:** A hinged pair of steel-pointed legs that steps off, transfers, and repeats a physical distance directly, without an intermediate numerical reading — the tool that lets a shop divide a line into equal parts, scribe a circle, or fit a board to an irregular surface by comparison rather than arithmetic.

---

## ✦ how NOT to use it

- **not-for:**
  - Reading an absolute dimension the way a ruler, tape, or caliper does — plain dividers carry no scale [DV19].
  - A non-committing or erasable layout line — the points mark permanently.
  - A span beyond the legs' own reach — that's the trammel points' job, not a standard pair's [DV28].
  - An internal recess or another inaccessible feature where straight points can't seat — that's inside/outside calipers' job.

- **dangerous:**
  - **Honestly held at a lower confidence tier than most of this corpus's DANGEROUS entries — a real risk-shape, but extrapolated, not woodworking-shop-sourced.** The twin sharp points are a genuine puncture/eye-injury geometry — well-established enough that the classroom-compass industry manufactures dedicated "safety compass" products specifically to eliminate the point as an alternative to "traditional metal compasses with dangerous points" [DV22]. General eye-safety guidance separately notes that "running or playing with a sharp object in hand is a common cause of eye injuries" [DV22-sim]. **This run found no OSHA or woodworking-shop-specific injury account naming dividers directly** — the sourcing here is drawn from classroom/drafting-compass safety-product context and applied by reasonable extrapolation from the same sharp-twin-point geometry, not from a documented shop incident. **Per the review: this absence of a located injury account is a research result, not a risk assessment** — it does not establish dividers as low-risk. Basic handling controls (see *safety notes* below) apply regardless of whether an injury account exists.

- **damaging:**
  - **Over-tightening the pivot rivet on old/forged dividers can fracture the iron — a named, first-person account, not general maintenance guidance.** Stephen Shepherd, the entry's named professional-woodworker/restorer voice (see *human side*): *"Old iron can be brittle and efforts to tighten the pivot rivet can cause metal failure, I know this from experience... Rosin or brewers pitch in the joint will also cause enough resistance to solve loose leg problems"* [DV14]. **Per the review, this is Shepherd's own named repair for a specific old-iron pivot joint, not a general maintenance recommendation this entry endorses as universal.** Identify the joint and material first; for a valuable or unfamiliar instrument, consult the maker, a qualified tool repairer, or a conservator rather than defaulting to a friction-packing fix. A second named participant in the same thread corroborates the risk of over-adjustment producing inconsistent drag through the joint's range [DV14].
  - **A documented historical design weakness in early wing dividers — a design-history fact, not a live misuse rule for a tool in hand.** *(Reframed per the review, which cautions against turning a patent's account of prior art into an unsupported misuse rule.)* Starrett's 1889 patent names an actual failure mode in earlier wing dividers: unreinforced designs "have heretofore been no stiffer than the adjusting-spring" [DV6]. This documents what could go wrong with that class of tool at that time; it is not verified guidance about how a specific, well-made modern wing divider will behave under load. State handling guidance based on the actual pattern and lock present in the tool at hand, not on an inference from 19th-century prior art.
  - **Honest gap:** no source located this run documents dividers being bent or damaged through prying/screwdriver-style misuse — a plausible analogous failure mode given the legs' thin, tempered section, but not a claim this entry ships as sourced. The sourced DAMAGING example is the pivot-rivet over-tightening above.

- **beginner-trap:**
  - **Inconsistent stepping-off / disturbing the setting mid-walk.** The "walking dividers" convergence method (see *correct use*) only works if the setting stays fixed through repeated contacts; a beginner who nudges the setting with uneven grip pressure between steps will drift the whole layout without realizing it until the final step misses [DV16].
  - **Fiddling a friction-joint (firm-joint) setting by push-and-pull alone is a real, named, practiced skill gap.** A first-person forum account describes "tight joint" dividers as tricky precisely because "you can spend a lot of time going back and forth between measurements" trying to nail a setting without a screw adjuster to help [DV14].

- **lineage-misuse:**
  - **Confusing dividers with a drafting/pencil compass — the tool's single most common naming and technique confusion, directly documented, not inferred.** Britannica states plainly: "The terms compass and divider are often interchanged, for each instrument can be used to draw circles, mark divisions... or simply mark a distance" [DV20]. A woodworker reaching for "the compass" from a drafting set and expecting divider-grade repeatable-transfer accuracy from a pencil point — or, in reverse, expecting a divider's twin steel points to produce a clean, erasable layout line the way a pencil compass does — is applying the wrong sibling tool's expected behavior. See the *modern task-boundary distinction* under *identity*.
  - **A period-specific version of the same confusion, secondhand-sourced and flagged as such.** A named commenter ("Jennie," a period-tool researcher) responding to Christopher Schwarz's Pompeii-dividers post states that "17th Century records seem for the most part to distinguish between the two. Compass legs are not fixed, divider legs are... Dividers are of course much more accurate and chosen for repeating dimensions as well as drawing circles," citing *Audels Carpenters and Builders Guide #1* (1923) and Charles Hummel's *With Hammer in Hand* [DV21]. **Held at lower confidence** — a blog commenter's citation of two books not independently verified by Haft this run, not a primary-source-confirmed claim.
  - **Treating dividers as a precision absolute-measurement instrument the way a machinist treats a vernier caliper or [[Micrometer (Outside Micrometer Caliper)|micrometer]]** — a category error the tool's own anatomy invites (both are hinged, pivoted measuring instruments) but that the tool's own lack of a scale makes categorically wrong; see *not-for* and *matching → cannot replace* [DV19].

---

## correct use

- **what it's genuinely good at:** transferring a physical distance directly, without an intermediate numerical reading — stepping off equal divisions along a line, scribing a circle or arc, checking that two *accessible* openings or gaps genuinely match, and copying an irregular profile (a wall, a stone foundation) onto a board [DV15][DV24]. **Bounded, per the review:** this does not make the transfer "exact" or "error-free" — the result still depends on point placement, a stable joint setting, surface condition, and operator control, and repeated stepping can accumulate error rather than eliminate it.
- **the core mechanic — "walking" or "stepping off."** Set the dividers to a best-guess span, place one point at the starting reference, swing the tool end-over-end along the line, and read the final position against the target; adjust the setting by a fraction of the observed error and repeat until it lands on target — an iterative, math-free convergence, not a single measured setting [DV16].
- **scribing an irregular surface (the classic fit-to-wall/foundation use) — corrected procedure (the review).** First fix the workpiece in its intended plumb or level position; the technique depends on that reference staying still through the whole scribe. Set the divider span to the controlling (widest) gap in the irregular edge. Then hold the tool at a **constant orientation** — traditionally level or square to the work, not re-angled at every bump — as one point follows the irregular reference surface and the other scratches the offset line onto the workpiece. Rotating the tool to stay "perpendicular to the surface" at every change in contour can shift the projected offset and is not what the traditional technique calls for [DV24]. This produces a scratched transfer line, which is a different kind of mark from a compass-and-pencil drawn line (see the *modern task-boundary distinction*).
- **dovetail spacing — the tool's best-documented single project use.** Set the dividers to the desired pin/socket spacing and step off the layout directly on the tail board or pin board, per the Peters/Cosman/Schwarz lineage above; Schwarz states plainly: "I always use dividers to lay out my dovetails — mostly so I have to think only once about the setup" [DV8][DV9][DV10].
- **lock the setting and don't disturb it mid-task.** Every sourced technique above depends on the divider holding its span through repeated contacts — a wing or spring divider's whole design purpose (see *lineage*) is to make this possible.

---

## ✦ matching — for the app's read

- **does:** steps off a measurement and transfers a distance without converting it to a number [DV15]; scribes a circle or arc [DV32]; divides a line into equal segments by "walking" the dividers end-over-end [DV16]; checks two *accessible* gaps or openings for equal spacing [DV32]; scribes/copies an irregular surface (a wall, a stone foundation, an uneven post-to-plinth joint) onto a workpiece for a tight fit [DV24].
- **suits:** dovetail layout — the best-documented, most concrete project match, with a real named teacher-to-teacher technique lineage behind it (Peters → Cosman → Schwarz) [DV8][DV9][DV10]; circle and arc layout, including polygon/octagon layout on a square blank [DV32]; fitting a board to an irregular wall, foundation, or existing structure [DV24]; general cabinetmaking and furniture layout wherever a story-stick-style, direct-comparison transfer is preferred over a ruled measurement [DV17][DV18]; timber framing's scribe-rule tradition, where dividers are the primary transfer instrument for an entire layout method, not a secondary check [DV11][DV12]; boatbuilding/lofting, chiefly via the trammel-point descendant for large-radius work [DV13].
- **choosing among the family (per the review — subtype-first, not one blanket recommendation):**
  - Choose **dividers** when two accessible points define the span, a direct transfer is preferable to recording a number, repeated intervals or geometrical division are needed, and a fine scratch mark is acceptable.
  - Choose a **compass or scriber** when one leg must follow a datum while the other leaves a visible pencil or scribed line.
  - Choose **inside/outside calipers** when the feature's geometry prevents straight divider points from seating on the relevant internal or external surface.
  - Choose **trammels** when the span or radius exceeds ordinary divider capacity.
  - Choose **proportional dividers** when the task is to enlarge or reduce by a maintained ratio.
  - **Avoid or change method** when the surface will be damaged by points, the work can't be secured, the setting won't hold, or the required tolerance exceeds what a hand transfer can reliably achieve.
- **kit-fit:**
  - **pairs with:** the **[[Scratch Awl]]** — the corpus's own strongest, founding citation: dividers transfer the measurement, the awl pricks the point into the board [SA1-cross]. Also the **marking gauge** (both are point/edge layout tools used in the same setup sequence, per the corpus's own [[Marking Gauge]] lineage note on "compass/divider run against a fingertip fence") [MG-cross]; a **compass saw or coping saw**, to cut a scribed arc.
  - **substitutes for:** a rule or tape, specifically for *transferring* (not reading) a distance — this is the tool's whole reason for existing [DV17][DV18]. This substitution is bounded, not general-purpose: dividers don't supply a recorded dimension, preserve traceable calibration, or replace a graduated tool when a numerical value or a specified tolerance is actually required.
  - **cannot replace:** a rule or tape measure, for reading an absolute dimension — dividers carry no scale of their own and produce no numeric reading; a distance stepped off with dividers must still be transposed onto a ruled scale if an actual number is ever needed [DV19]. A pencil compass, for a non-committing or erasable layout line — dividers' steel points prick or scratch a permanent mark, where a pencil compass proposes a line that can be redrawn (a direct structural parallel to the corpus's own Scratch Awl "pencil vs. awl" framing, offered here as Haft's own reasoned inference from the tools' shared anatomy, not a directly sourced warning) [SA-inference]. A precision machinist caliper or [[Micrometer (Outside Micrometer Caliper)|micrometer]], for a true absolute-dimension reading — see *not-for*.
- **not-for (project-keyed):** don't reach for plain dividers when the job needs a directly-read absolute dimension against a printed scale — that's a rule's or a caliper's job [DV19]. Don't reach for them when the layout needs a non-marking or erasable line — the points commit a mark the way a pencil does not. Don't reach for a standard pair when the required radius exceeds the legs' span — that is the trammel points'/beam compass's job [DV28]. Don't reach for straight divider points on an internal recess or an inaccessible span — that's inside/outside calipers' or a [[Story Stick]]'s job instead (see *choosing among the family* above).
- **access:** opening, closing, and pricking a divider is mechanically trivial for a beginner. But **"walking" the dividers accurately to hit an exact equal division is a real, sourced skill with a genuine learning curve** — even the best-documented technique lineage describes it as iterative trial-and-error, not a one-shot setting: "unless you're really lucky, you won't hit right at the end of the board" [DV16]. A first-time user should expect several passes before a step lands cleanly.
- **revival:** the tool is inexpensive to acquire secondhand — Fine Woodworking's own buying guidance notes this directly ("you can pick up a pair at any flea market for just a few bucks") [DV15] — **but low price is not itself a fitness check (corrected per the review).** A used pair should be inspected before relying on it: points aligned and undamaged, legs unbent, the pivot smooth without uncontrolled play, the lock or adjuster functional, and no cracks or serious corrosion — verified by testing on scrap that a set span actually holds. The dovetail-spacing technique is a genuine, actively-taught modern revival of a pre-industrial, arithmetic-free layout method — currently published and demonstrated by Lost Art Press, Popular Woodworking, and Rob Cosman alike [DV8][DV9][DV10] — a real case of a simple hand tool staying in daily use precisely because it solves a problem digital calculators don't (repeatable physical transfer, not a number). **Honest gap:** no single named source was found tying dividers explicitly to a "museum piece picked back up" framing the way some other corpus entries carry it — this revival case rests on the tool's ongoing, documented, everyday teaching use rather than a rediscovery narrative.

---

## anatomy

- **legs (2)** — matched, tapering steel legs hinged at one end [DV20][DV32].
- **pivot / joint** — either a **firm joint** (rigid friction pivot, no spring or wing — "faster to set" per one manufacturer's own comparison copy) or a **spring/wing joint** [DV32].
- **points / tips** — hardened steel, occasionally independently pivoted or detachable; US Patent 479,112 describes a construction with points riveted to one leg and pivoted on the other, with a set-screw-secured detachable-point variant [DV27]. Dull, chipped, rolled, or mismatched points are named directly as degrading precision, unreliable, and in need of dressing or resharpening [DV24-sim].
- **wing / quadrant** (wing dividers) — the curved locking arm a thumbscrew clamps to fix the angle; the 1245-dated medieval refinement [DV3].
- **spring / bow** (spring dividers) — a C-shaped spring-steel arc at the head. **Corrected mechanism (the review):** the spring supplies a separating (opening) force on the legs; a screw or nut works against that spring to close the legs to the desired setting and hold them there — the spring does not "hold the legs closed." Starrett's own current product copy for its 277-series names the construction directly: "hardened fulcrum stud, strong and flexible bow spring, and centrally located leg adjustment" [DV25]. **Note on the source text:** the listing's own descriptive paragraph contains a typo ("bow string"); its feature/spec field states the correct term, "Bow spring is strong and flexible" — cited here for the mechanism, with the typo flagged rather than repeated as if it were the manufacturer's intended term.
- **thumbscrew / adjusting nut / locking nut** — the fine-adjustment mechanism on wing and spring dividers; Starrett's 1889 patent documents this assembly's own stated rationale (see *lineage*, including the caveat that the Petroski framing is Haft's reading of the patent, not the patent's own claim) [DV6].

---

## variations (living)

*Identify the pattern before adjusting it — plain friction-joint, wing, spring, proportional, and trammel dividers are each adjusted differently (per the review); see below and *anatomy* above.*

- **Plain / firm-joint dividers** — a solid, rigid pivot, no spring or wing; the presumptive oldest form (the geometry of the Roman Met and Pompeii examples), and the geometry of a compass before a pen/pencil leg was ever added [DV2][DV1][DV29]. Adjusted by hand friction alone. Trades quick-setting speed for a setting that isn't locked [DV32].
- **Wing dividers** — the most common modern woodworking form; a quadrant/wing locks the angle via a thumbscrew. Named directly by a current manufacturer as producing "durable, locked, and stable settings," with "hardened sharp points" that "pierce wood, making them perfect for scribing and marking measurements" [DV32].
- **Spring dividers (spring-bow dividers)** — a C-shaped spring supplies opening force; a screw or nut works against it (see *anatomy* for the corrected mechanism). More associated in current commercial framing with the machinist/toolmaker trade (Starrett's own 277-series copy addresses "toolmakers and all good mechanics") but sold and used in woodworking too [DV25][DV32].
- **Trammel points / beam compass** — adjustable heads riding a separate beam for spans beyond a fixed pair's leg length; diameters "from a few inches to several feet" [DV28]. Live current products (General, Taytools) exist alongside a dated, confirmed retired one: Woodpeckers' "Deluxe Trammel System," a OneTIME Tool retired **August 5, 2019** [DV26][DV31].
- **Proportional dividers** — an adjustable, slotted pivot between two point-pairs lets a transferred measurement reproduce at a different, fixed ratio; more a drafting/machinist scaling tool than a general woodworking layout tool. The Smithsonian's own collection holds a dated Japanese brass example ("before 1876") [DV23].
- **Dividers vs. compass — same body, swapped leg.** "A compass is essentially a divider in which one of the points has been replaced by a pencil or other marking device" [DV29]. Modern drafting sets still sell interchangeable divider-needle and pencil-lead points for a single compass body [DV29-sim]. See the *modern task-boundary distinction* under *identity* for how this entry treats the two terms today.

---

## trades

- **Cabinetmaking / fine furniture / hand-tool woodworking** — dovetail layout, general story-stick-style transfer work, the tool's dominant modern home in the sourced record [DV8][DV9][DV10][DV17][DV18].
- **Timber framing** — the scribe-rule tradition's primary layout instrument, a largely arithmetic-free pre-Square-Rule American building method built around dividers and a plumb bob [DV11][DV12].
- **Boatbuilding / shipbuilding — lofting** — chiefly via the trammel-point/beam-compass descendant for large-radius, full-scale layout work [DV13].
- **Machinist / toolmaking** — spring dividers and precision calipers are a documented core of this trade's own kit (Starrett's own current marketing addresses "toolmakers and all good mechanics" directly) [DV25], and the corpus's own *not-for* boundary against the machinist's caliper/micrometer is drawn from this trade's own tool distinctions [DV19].
- **Drafting / technical drawing** — the compass sibling's home trade; the shared-body, swapped-point relationship between drafting compass and shop dividers is the entry's central lineage-misuse finding [DV20][DV29].

---

## lineage

**Ancestors**
- **The bare point, shared with the [[Scratch Awl]]** — dividers are, mechanically, two scratch awls hinged together at a fixed pivot so their two points hold a repeatable relationship to each other. The corpus's own Scratch Awl entry already frames dividers as the awl's "partner in stepped layout" [SA1-cross].
- **No confirmed single point of invention** — dividers and compasses are attested in both Greek and Roman antiquity, with Euclid restricting his geometric constructions to "an unmarked straight edge and rudimentary compass" [DV4], but this run found no earlier, more specific point of origin than "known to both the Greeks and Romans" [DV3].

**Descendants — different design pressures, not one universal failure.** *(Corrected per the review: the earlier framing treated the whole family as "a chain of fixes to one failure — the plain pivot's setting doesn't hold." That's true of the wing and spring lines specifically, but proportional dividers solve a scaling problem, trammel points solve a range problem, and sheet-metal construction solves a manufacturing-cost/weight problem — four distinct pressures, not one.)*
- **The wing (quadrant) divider** — addresses the plain pivot's holding problem. Britannica: a divider "with a circular sector, or wing, connecting the two legs was sketched in 1245" [DV3]. The wing locks the included angle with a thumbscrew instead of relying on pivot friction alone.
- **The rigid-lock wing divider.** LaRoy S. Starrett's US Patent 411,537 ("Calipers or Dividers," filed May 4 1889, granted Sept. 24 1889) states the failure it fixes in the inventor's own words: earlier wing dividers "have heretofore been no stiffer than the adjusting-spring." His threaded-wing-rod-plus-external-adjusting-nut-plus-internal-locking-nut construction, the patent states, makes "the arms, when locked, perfectly rigid" [DV6]. **Haft reads this patent as a Petroski "form follows failure" case** — the inventor names a weak point in prior wing dividers and proposes a stiffer lock — but that reading is Haft's own interpretation, not a framework the patent itself invokes. A second, later Starrett patent in the same class (US 539,759, with Charles P. Fay, 1895) is recorded in DATAMP's index but not independently read this run [DV7].
- **The spring (bow) divider** — trades range for fine, repeatable micro-adjustment: a C-shaped spring-steel bow supplies a separating (opening) force on the legs, and a screw or nut works against that spring to set and hold the leg angle (see *anatomy* for the corrected mechanism). Because the maximum setting is capped by leg length, this is "a narrower but more controllable tool for small, precise work" [DV15]. J.W. Strange's US Patent 30,009 (Sept. 11, 1860, "Spring Dividers") documents a sliding-yoke/screw arrangement applied to spring dividers and calipers, and is the earliest specifically-named spring-divider patent located this run — but the patent itself refers to "usual forms" and existing adjusting arrangements, so it documents an improvement to an already-established mechanism, not the invention of spring dividers as a class [DV7].
- **The lightweight sheet-metal divider — a second, distinct design pressure: manufacture, weight, and cost, not pivot rigidity.** US Patent 479,112 (William A. Bernard, New Haven CT, granted July 19, 1892, "Dividers") states its own purpose directly: to let dividers/compasses/calipers "be made of sheet metal... very much lighter and cheaper than it has been possible to make them as heretofore constructed," because prior tools of this class were "forged or cast solid stock," "clumsy and heavy," and required costly milling and counterboring to fit the pivot joint [DV27]. Haft reads this as its own, separately-motivated Petroski case — again, Haft's reading of the patent, not a claim the patent makes about itself.
- **The proportional divider — a distinct design pressure: scaling, not holding or range.** A slotted, slidable pivot lets a distance transferred at one end of the legs reproduce at a fixed, different ratio at the other. Roman proportional dividers had a single fixed pivot (one ratio only); Leonardo da Vinci is credited by a secondary source with the adjustable-pivot refinement, and Federico Commandino is credited with adding a micrometer screw in 1568 [DV3][DV28] — **held at lower confidence**, this specific attribution chain was not independently verified against a primary source this run. The Smithsonian's own collection includes a dated Japanese brass proportional divider, "before 1876," used by engineers and surveyors "to mark off equal segments on a line or the circumference of a circle" and to "proportionally enlarge or reduce a portion of a drawing" [DV23].
- **Trammel points / beam compass — a distinct design pressure: range, not holding or scaling.** Fixes the range limitation a hinged pair inherits from its own leg length: adjustable heads ride a separate beam, so the working span is bounded only by the beam's length, not the legs' [DV28]. Current, live commercial examples exist (General, Taytools) alongside a dated, confirmed retired one (Woodpeckers' "Deluxe Trammel System," a OneTIME Tool retired Aug. 5, 2019) [DV26][DV31].

**Contemporaries**
- **The drafting/pencil compass** — the exact same instrument body, with one steel point swapped for a pencil, pen, or lead holder. Britannica states the two names are routinely interchanged, "for each instrument can be used to draw circles, mark divisions... or simply mark a distance" [DV20]. Wikipedia adds the historical reverse: "before the 18th century, when one leg was modified to take a pen or pencil point, compasses had two sharp points like dividers, and the user would scratch the writing surface... and then ink the scratches" [DV29]. See *lineage-misuse* and the *modern task-boundary distinction* above — this is the tool's single most common naming/technique confusion.
- **The [[Scratch Awl]]** — dividers' documented layout partner in this corpus's own founding citation: Moxon's joiner "transfers a measurement with dividers, then pricks the point into the matching board" [SA1-cross].
- **The precision machinist caliper (vernier/micrometer)** — a genuine cousin in form (both are hinged/pivoted measuring instruments) but a different category of tool entirely: a caliper reads an absolute numeric dimension off a scale, and dividers carry no scale of their own [DV19]. See *not-for*. (The [[Micrometer (Outside Micrometer Caliper)|Micrometer]] entry, built Run 571, documents this cousin directly — including its own founding 1848 Palmer patent and Brown & Sharpe's re-engineering.)

---

## history

- **Ancient origin, anchored by dated museum objects, not assumed.** Roman bronze dividers were found at Pompeii (destroyed AD 79); originals are held at the Munich Residenz, with a copy displayed at the Deutsches Museum — Christopher Schwarz photographed and described the surviving pair directly: a friction-only pivot, no locking mechanism, with cast decoration on the legs [DV1]. Independently, the Metropolitan Museum of Art holds two dated Roman bronze compass/divider fragments, both acquired via the Rogers Fund in **1917**: object **17.230.33** (6⅝″, titled by the Met's own catalog as "Compass or divider, fragment") and object **17.230.63** (7½″) [DV2]. 1917 is the acquisition/credit-line date, not a manufacture date — this entry treats it as such. This is the entry's strongest deep-antiquity anchor — a named institution, named accession numbers, and a confirmed acquisition date, independent of the Pompeii pair. **Honest gap, added this run:** both objects are cited here as a pair via a single museum record; 17.230.63's own separate object page was not independently re-verified this run.
- **Egypt — geometric-construction evidence, not yet a dated physical artifact.** Ancient Egyptians are documented using compass-and-straightedge geometric constructions roughly nine centuries before the Greek Oenopides of Chios (5th c. BC) described comparable methods, evidenced by an "Overlapping Circles Grid" on a 14th-century-BC Egyptian artifact [DV30]. **Honest gap:** no specific dated physical Egyptian divider/compass object (as opposed to the geometric-construction evidence) was located this run — this is named as an open gap, not converted into a claimed artifact.
- **The Moxon citation, and a residual worth naming plainly.** The corpus's own [[Scratch Awl]] entry already carries the load-bearing citation: *"Moxon's joiner transfers a measurement with dividers, then pricks the point into the matching board"* [SA1-cross]. This run independently checked Moxon's *Mechanick Exercises* (1703, 3rd ed.) illustrated joiner's-tool plate as catalogued by the Smithsonian's Peter C. Welsh — **dividers/compasses do not appear among the 30 lettered tools on that plate** [DV5]. This is not a contradiction of the Scratch Awl citation — Moxon's prose text and his illustrated plate are two different kinds of evidence, and the SA1 citation reads as coming from the prose rather than the plate — but this run could not independently re-locate the exact prose passage against a primary scan to confirm the wording. **Named as an honest residual**: the existing corpus citation is not disputed, but it is not yet independently re-verified against Moxon's primary text either, and the plate itself is silent on the tool.
- **A dated, itemized American shop inventory confirms dividers and spring dividers as distinct, separately-priced tools by 1827.** Welsh's Smithsonian survey quotes **Amasa Thompson's 1827 Middleborough, Massachusetts carpenter's-shop inventory** directly: *"1 pr. dividers or compasses — .71"* and, separately, *"1 pr. Spring dividers — .92,"* alongside a "Case Mathematical Instruments $3.62½" [DV5]. This is a genuine, dated, itemized primary-source record — not an inference — that by 1827 an ordinary American shop stocked plain dividers and spring dividers as two distinct, differently-valued tools.
- **The wing divider's medieval refinement, and its 19th-century rigidity fix.** Britannica dates the wing (sector-locked) divider's sketch to 1245 [DV3]; Starrett's 1889 patent (detailed under *lineage*) states the prior failure and its own fix in the inventor's own words, and is discussed there together with the caveat that the Petroski framing is Haft's own reading [DV6].
- **English cabinetmaker's and carpenter's dividers, dated and named at object level.** Welsh's Smithsonian survey (Figs. 15–19) catalogs an 18th-century English cabinetmaker's divider pair with Renaissance-style decorated legs and a heart-shaped slide-arm stop (private collection); a second pair **dated 1783 and marked "T. Pearmain,"** Sheffield-made, plain and functional in contrast to the decorated pair; and an 18th-century English carpenter's divider, **accession 319557**, described as "a form that changed very little, if at all, until after 1850" — a claim independently corroborated by a period illustration (the frontispiece of Edward Shaw's *The Modern Architect*, Boston, 1855) showing carpenter's dividers "unchanged in form" from the 18th-century example [DV5].

---

## ✦ human side

- **Shipwrights and boatbuilders — lofting.** "Trammel points clamp onto a stick or batten to make large-diameter compass marks for rounding frame corners or other curves. Trammels do the accurate work of full-sized layout in lofting. Their use as giant dividers helps avoid the mistakes made by misreading rulers" [DV13]. This is the trade where the tool's range problem (fixed leg length) is most acute, and where the trammel-point descendant does its heaviest documented work.
- **Timber framers — Scribe Rule / French Scribe layout, largely arithmetic-free.** *(Corrected per the review: the original quote here — "entire cathedrals using only a pair of dividers and a plumb bob... no math and no reading of measurements at all" — is rhetorical and historically implausible as written, so it is replaced with the bounded version the review supplied.)* Dividers, straightedges, squares, cords, and plumb/level references allowed craftspeople to construct and repeat geometry without expressing every operation as a written numerical measurement [DV12]. More technically: "Timbers are supported by cribbing, aligned with the plan using a plumb-line, vertical distances are transferred with dividers... A 1:12 scale model, for instance, would require 12 compass steps on the actual timber" [DV11]. This is the strongest documented case of dividers as a *primary*, not secondary, layout instrument for an entire trade tradition — pre-Square-Rule American timber framing runs on the tool.
- **Stephen Shepherd — professional woodworker/museum restorer, 40+ years, named and credentialed** (author, *Shepherds' Compleat Early Nineteenth Century Woodworker*, 1981/2004; *Hide Glue: Historical & Practical Applications*, 2009). Active, first-person forum voice on a 2010 dividers thread: *"Like Jack, I use mine all the time and always looking for more"* [DV14]. His own directly-quoted, named account of a real maintenance failure is the entry's DAMAGING anchor — see below.
- **Alan Peters → Rob Cosman → Christopher Schwarz — a real, named, dated teacher-to-teacher technique lineage.** Christopher Schwarz traced the dividers-based dovetail-spacing method he learned to **Rob Cosman**, who learned it from English craftsman **Alan Peters** [DV8][DV9]. Schwarz independently located a related description in Bernard E. Jones (ed.), *The Practical Woodworker* (Waverly Book Co., London, 1920s), p. 227, and quotes it directly: *"With the compasses or dividers divide the intervening space between these marks into as many parts, plus one, as there are to be whole pins... Any slight inaccuracy at this stage will affect the appearance but not the fit of the dovetails"* [DV8]. Schwarz notes Jones's method marks centerlines while the Peters/Cosman method marks the extents — two distinct historical technique lineages using the same tool, both real and named, neither flattened into "the" method.
- **The "reduces certain error classes" philosophy — well corroborated, not pinned to one primary quote.** *(Corrected per the review: rewritten from "no cumulative error" — see fix 1 below.)* Multiple independent sources converge on the same principle without a single, citable originating sentence: dividers/trammel use "help avoid the mistakes made by misreading rulers" in lofting [DV13]; story-stick and divider-based layout is framed as avoiding the error that stacks when a measurement is repeatedly read off a ruled scale [DV17][DV24-sim]; and an unattributed Sawmill Creek forum consensus states plainly, "when working with dividers you don't have to talk about fractions, and can divide spaces evenly and accurately without needing to calculate fractional measurements" [DV18]. **Held honestly as a multi-sourced philosophy, not a single named author's claim** — no directly-quotable Schwarz or Sellers sentence pairing "dividers" and "cumulative error" by name was located this run. And per the review: this principle describes avoiding *transcription/scale-reading* error specifically — it does not mean dividers are immune to error generally; repeated stepping, an unstable joint, or inconsistent point placement can still accumulate error of their own.
- **The "walking" / stepping-off technique, described in a real forum how-to.** "Set dividers to your best guess of the desired division... start at one end... step off that many paces, then adjust the dividers by approximately 1/5 of the error and step off again" — an iterative, math-free convergence method [DV16].

---

## ✦ proof of existence

- **Two dated, accessioned Roman bronze compass/divider fragments, Metropolitan Museum of Art** — objects **17.230.33** (6⅝″, Met's own catalog title: "Compass or divider, fragment") and **17.230.63** (7½″), both acquired via the Rogers Fund in **1917** — an acquisition date, not a manufacture date [DV2]. The entry's strongest single proof-of-existence anchor: a named institution, named accession numbers, a confirmed acquisition date, and a Roman-era dating for the general class of bronze divider, independent of the separately-documented Pompeii pair [DV1]. **Honest gap:** 17.230.63's own separate object record was not independently re-verified this run; both objects are cited here as a pair via one shared museum record.
- **A named, dated, itemized 1827 American shop inventory** — Amasa Thompson's Middleborough, Massachusetts carpenter's-shop inventory itemizes "1 pr. dividers or compasses" and, separately, "1 pr. Spring dividers," each at its own price [DV5]. A genuine primary-source object of proof — not a museum accession, but a dated, named, itemized commercial record.
- **Two foundational US patents, both directly read this run.** LaRoy S. Starrett's US 411,537 (1889, rigid wing-lock construction) and William A. Bernard's US 479,112 (1892, lightweight sheet-metal construction) [DV6][DV27] — each documents its own specific claimed improvement, not the invention of its whole tool class (see *lineage*). DATAMP's own index cross-references at least ten further 19th-century American divider patents (1848–1895) not independently re-read this run [DV7].
- **Current market — an access-dated snapshot, not historical evidence (corrected per the review).** Lee Valley's own current dividers (made in France for Lee Valley; steel, ground and hardened tips; solid brass fittings; spring-loaded quick-adjust) — 4″ = $34.50, 8″ = $38.00, 12″ = $51.50, sold in matched sets alongside inside/outside calipers of the same size, directly fetched July 2026 — and Starrett's current 277-series "Toolmakers' Spring-Type Calipers and Dividers," a Made-in-America line from a manufacturer founded 1880 in Athol, Massachusetts [DV24][DV25]. **The company's own founding date does not establish uninterrupted production of this specific divider line or model across that span** — no continuous-production record for the 277-series itself was independently verified this run, and the removed "146 years of continuous production" claim should not be read back in.
- **A dated, verifiable retired-product history for the trammel-point descendant**, matching this corpus's own established proof-of-existence pattern for other entries: Woodpeckers' "Deluxe Trammel System" OneTIME Tool, retired **August 5, 2019** [DV26].
- **Honest gaps, named directly:** no Mercer Museum or Victoria & Albert Museum accession-level divider record was located this run, despite both being plausible collection homes for this tool class; no specific dated physical Egyptian divider/compass artifact was located (only Egyptian geometric-construction evidence dated to the 14th c. BC); the exact primary-text Moxon passage behind the corpus's own Scratch Awl SA1 citation was not independently re-located against a scanned original this run (see *history*); 17.230.63's own separate Met object record was not independently re-verified; and the DATAMP divider-patent-class's full count was not exhausted (roughly ten of a class confirmed to run well beyond that were directly read).

---

## references

| Tag | Source | Kind |
|---|---|---|
| DV1 | Christopher Schwarz, "Early Dividers from Pompeii," Lost Art Press, 2013-06-25, https://blog.lostartpress.com/2013/06/25/early-dividers-from-pompeii/ | named-maker blog |
| DV2 | Metropolitan Museum of Art, Roman bronze compass/divider fragments, objects 17.230.33 ("Compass or divider, fragment") and 17.230.63, Rogers Fund 1917 (acquisition date, not manufacture date), https://www.metmuseum.org/art/collection/search/250596 | museum (directly fetched) |
| DV3 | Encyclopaedia Britannica, "Hand tool — Compass, divider, and caliper" (Joseph A. McGeough), https://www.britannica.com/technology/hand-tool/Compass-divider-and-caliper | encyclopedia |
| DV4 | Smithsonian, "Dividers & Compasses — Introduction," https://www.si.edu/spotlight/dividers-compasses/introduction | museum overview |
| DV5 | Peter C. Welsh, *Woodworking Tools, 1600–1900*, Smithsonian Museum of History and Technology Paper 51, full text via Project Gutenberg #27238, https://www.gutenberg.org/files/27238/27238-h/27238-h.htm | museum monograph (primary catalog descriptions, 1827 shop inventory, Moxon tool-plate figure) |
| DV6 | LaRoy S. Starrett, US Patent 411,537, "Calipers or Dividers" (filed 1889-05-04, granted 1889-09-24), https://patents.google.com/patent/US411537A/en | patent (directly read) |
| DV7 | DATAMP (Directory of American Tool and Machinery Patents), "All Patents of Type: Dividers," https://www.datamp.org/patents/search/xrefType.php?cat=90&class=2100&type=2 | patent index (partial, not exhausted) |
| DV8 | Christopher Schwarz, "Spacing Dovetails with Dividers – a Little History," Popular Woodworking, 2012-01-30, https://popularwoodworking.com/techniques/joinery/spacing-dovetails-with-dividers-%E2%80%93-a-little-history | named-maker blog, incl. direct Bernard E. Jones (c.1920s) quote |
| DV9 | Rob Cosman, "My Dovetail Evolution," https://robcosman.com | named-maker page |
| DV10 | Lost Art Press, "Dovetail Layout with Dividers," 2022-03-25, https://blog.lostartpress.com | named-maker blog |
| DV11 | Michael Langford, "Timber Frame Layout Methods," Timber Frame HQ, https://timberframehq.com/timber-frame-layout-methods/ | trade reference |
| DV12 | "Timber Frame Joinery: Square Rule & Scribe Rule Layout," The Year of Mud, 2013-01-26, https://theyearofmud.com/2013/01/26/timber-frame-joinery-tools/ | trade blog |
| DV13 | Sandy Point Boat Works, "Lofting Basics," https://sandypointboatworks.com/boat-building-articles-journals/boat-building-articles/lofting-basics; WoodenBoat School, "Lofting: How to Loft a Simple Boat Plan," https://skills.woodenboat.com/guides/lofting-how-to-loft-a-simple-boat-plan/ | trade reference |
| DV14 | Stephen Shepherd et al., "Forged dividers" thread, WoodCentral Forums, 2010-02, https://www.woodcentral.com/forums/topic/3144852/forged-dividers | forum, incl. named-maker posts |
| DV15 | Steve Latta, "How to Use Dividers," Fine Woodworking #241, Jul/Aug 2014, https://www.finewoodworking.com/project-guides/hand-tools/how-to-use-dividers | trade publication |
| DV16 | "How to use dividers" thread, LumberJocks, https://www.lumberjocks.com/threads/how-to-use-dividers.41635/ | forum |
| DV17 | "Dividers." thread, Sawmill Creek Woodworking Community, https://sawmillcreek.org/threads/dividers.209525/ | forum (community consensus, not individually named) |
| DV18 | Same as DV17 — the "no fractions" stepping-off quote | forum |
| DV19 | Reid Supply, "Micrometers vs Calipers: Understand the Key Differences," https://www.reidsupply.com/en-us/industry-news/micrometers-vs-calipers-key-differences; Hobby-Machinist forum consensus, https://www.hobby-machinist.com/threads/tool-makers-calipers-dividers-vs-spring-calipers.79215/ | trade reference + forum |
| DV20 | Encyclopaedia Britannica, same entry as DV3 — "terms compass and divider are often interchanged" | encyclopedia |
| DV21 | "Jennie" (named commenter, period-tool researcher), comment on DV1, citing *Audels Carpenters and Builders Guide #1* (1923) and Charles Hummel, *With Hammer in Hand* — neither independently verified this run | blog comment (secondhand citation, lower confidence) |
| DV22 | Westcott Safety Compass product page, https://westcottbrand.com/products/westcott-safety-compass-for-the-classroom-draftsmen-crafters-and-artists-00516-parent; Nationwide Children's Hospital, "Preventing Eye Injuries in Children," https://www.nationwidechildrens.org/conditions/health-library/preventing-eye-injuries-in-children | consumer product + general safety reference (classroom/drafting context, not woodworking-shop-specific) |
| DV23 | Smithsonian National Museum of American History, Proportional Dividers object group, Japanese brass example ID MA.261313, dated "before 1876," https://americanhistory.si.edu | museum (directly fetched) |
| DV24 | Lee Valley, Dividers product page, directly fetched July 2026 — an access-dated market snapshot, not historical evidence, https://www.leevalley.com | product (directly fetched) |
| DV25 | Starrett, 277-series "Toolmakers' Spring-Type Calipers and Dividers," directly fetched, https://www.starrett.com — note: the listing's own descriptive paragraph contains a typo ("bow string"); its feature/spec field states "Bow spring is strong and flexible," the term used in this entry | product (directly fetched); company founded 1880, Athol MA — founding date does not establish continuous production of this specific line |
| DV26 | Woodpeckers, "OneTIME Tool: Deluxe Trammel System (2019)," retired product page, https://www.woodpeck.com | product (directly fetched, retired, dated retirement 2019-08-05) |
| DV27 | William A. Bernard, US Patent 479,112, "Dividers" (granted 1892-07-19), https://patents.google.com/patent/US479112A/en | patent (directly read) |
| DV28 | Wikipedia, "Beam compass," https://en.wikipedia.org/wiki/Beam_compass, citing Ephraim Chambers, *Cyclopædia* (1728) | encyclopedia |
| DV29 | Wikipedia, "Compass (drawing tool)," https://en.wikipedia.org/wiki/Compass_(drawing_tool) | encyclopedia |
| DV30 | arXiv, "Overlapping Circles Grid Drawn with Compass and Straightedge on an Egyptian Artifact of 14th Century BC," https://arxiv.org/pdf/1603.08536 | scholarly paper |
| DV31 | Taytools, "Trammel Head Beam Compass Set, 57″ Diameter Range," https://taytools.com; Highland Woodworking, General Adjustable Trammel Point Set, https://www.highlandwoodworking.com | product (current market) |
| DV32 | Lie-Nielsen Australia, "Why Use Dividers?" https://www.lie-nielsen.com.au | trade reference (manufacturer comparison copy) |
| SA1-cross | [[Scratch Awl]] (this corpus), founding citation: "The dividers / pricker pair — Moxon's joiner transfers a measurement with dividers, then pricks the point into the matching board [SA1]; the awl is the dividers' partner in stepped layout, not a replacement for the rule." | corpus cross-reference |
| MG-cross | [[Marking Gauge]] (this corpus), passing citation: "compass/divider run against a fingertip fence — the 'finger gauge' technique" | corpus cross-reference |
| SA-inference | Structural parallel to [[Scratch Awl]]'s own "pencil vs. awl" contemporaries framing, offered as Haft's own reasoned inference, not a directly sourced claim | corpus-derived inference |
| DV-gap | Claims searched for but not sourced this run (dividers bent through prying misuse; a single primary-quotable "dividers vs. cumulative error" sentence; Mercer/V&A accession records; the exact Moxon prose passage; 17.230.63's own separate object record) — flagged inline at point of use | honest gap marker |

---

## ✦ safety notes *(new section, added per the reviewer's 2026-08-01 review — the original entry had no dedicated safety-notes block)*

- Treat both points as puncture and eye hazards even when the tool is not being actively used and even when the points read as "dull" — geometry, not sharpness alone, is the hazard.
- Keep hands, and any hand supporting the workpiece, out of the points' travel direction; secure the work before scratching or stepping a span.
- Never point the tool toward the body or toward another person, open or closed.
- Close the legs when the tool isn't in use if practical; use point caps, a sheath, or a fitted case for transport and storage rather than carrying an open pair loose in a pocket or tool bag.
- Keep points only as sharp as the task actually requires — chipped, rolled-over, mismatched, or excessively needle-like points can slip, tear fibers, or give an unreliable setting.
- Test the lock and pivot on scrap before relying on a setting for real layout — a setting that moves under ordinary hand pressure isn't trustworthy.
- Inspect a secondhand or antique pair before use per the checklist in *revival* above; don't assume an old pair is sound because it still opens and closes.

---

## The reviewer's corrections (8)

- `UNGRADED` — 1. Replace every unqualified “exact,” “error-free,” and “no cumulative error” formulation with a bounded statement: dividers transfer a span directly without an intermediate numerical reading, but the result still depends on point placement, stable adjustment, surface condition, and operator control. Repeated stepping can accumulate error
- `UNGRADED` — 2. Establish a modern working distinction before discussing historical overlap:
- `UNGRADED` — 3. Correct the spring-divider anatomy and operation. Do not say the bow “holds the legs closed” unless a cited design actually does so. In the common spring pattern, the spring supplies separating force and the screw/nut controls the legs against that force. For Starrett’s current 277 listing, correct “bow string” to **bow spring**; the listing’s descriptive paragraph contains a typo, while its feature field says “Bow spring is strong and flexible.”
- `UNGRADED` — 4. Rewrite the irregular-profile scribing procedure. The work must first be fixed in its intended plumb/level position. Set the span to the controlling gap, then maintain a constant tool orientation—traditionally described as level or square to the work—as one point follows the reference and the other scratches the offset line. “Keep both points perpendicular to the irregular surface” is misleading because rotating with every change in the surface can change the projected offset. Also distinguish a two-point scratch line from a compass/pencil line
- `UNGRADED` — 5. Remove or qualify “any garage-sale/flea-market pair works.” A used pair is suitable only after inspection: points aligned and serviceable, legs unbent, pivot smooth without uncontrolled play, lock/adjuster functional, and no cracks or serious corrosion. Verify on scrap that the setting holds before relying on it
- `UNGRADED` — 6. Replace the “entire cathedrals using only dividers and plumb bob, no math” claim. It is rhetorical and historically implausible as written. A defensible version would say that dividers, straightedges, squares, cords, and plumb/level references allowed craftspeople to construct and repeat geometry without expressing every operation as a written numerical measurement
- `UNGRADED` — 7. Change “a compass or coping saw, to cut a scribed arc” to **a compass saw or coping saw** if a cutting tool is intended. A drawing compass marks; it does not cut
- `UNGRADED` — 8. Do not call current asking prices historical or durable evidence. If retained, label them as an access-dated market snapshot. Likewise, the company’s age does not prove 146 years of uninterrupted production for one divider model or product line