# Mawashibiki

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

Reviewed: An independent reviewer read this entry and found corrections. None of them is a way for a person to get hurt, so it is published now. The corrections are listed at the end and are not yet applied to the text: read it with them in hand.

Cite as: "Mawashibiki." Hewn Tool Library, hewn.build/tools/mawashibiki. Reviewed 2026-09-03.

---

## Identity

**Name.** *Mawashibiki* (廻し挽き / 回し挽き, "turning cut") and *hikimawashi* (引き廻し / 挽き廻し, "pull-turning") are the two current orderings of the same compound. Modern sellers and reference sources use both terms for closely related Japanese narrow-bladed turning saws, though usage is not perfectly uniform across every catalogue; the encyclopedic dictionary form is *hikimawashi* (挽廻鋸). [M1][M2] Western functional equivalents are usually sold as **keyhole saw**, **compass saw**, or **turning saw** — overlapping trade names that are not perfectly interchangeable in every Western catalogue either. [M5][M12][M23] A long, narrow Japanese hand saw, cutting on the **pull**, whose blade tapers from a wider heel to a point and carries **no stiffening spine**, made to **follow a curved line and enlarge an interior opening** — the Japanese functional equivalent of the Western keyhole/compass saw. [M1][M5][M21]

**Family / category.** Japanese saws (*nokogiri*); a specialty member of the **pull-saw root** of the saw family ([[_Japanese Pull-Saw Root — Lineage Tree]]), one of the [[Kataba]] type-form's modifications — specifically the kataba **narrowed to turn.** The Nipponica encyclopedia lists 挽廻 (*hikimawashi*) as one of the **nine canonical carpenter's-saw types** (9種12丁 — nine kinds, twelve pieces), alongside *maebiki*, *kagari*, *yokobiki*, *ryoba*, *anabiki*, *dōzuki*, *azebiki*, and *hosobiki*. [M1]

**One line.** *The narrow, spineless, pointed pull-saw that follows curves and enlarges interior openings — the Japanese keyhole saw, and the one node where the pull root and the Western tree reach the same answer.*

---

## ✦ How NOT to use it

### Not-for (jobs it should never do)
- **Long, precision straight cuts.** The narrow, flexible, unspined blade has neither the width to track straight nor a spine to resist lateral flex, so sustained straight runs tend to wander — though a short straight relief or joinery cut is fine. For length and precision, reach for a [[Kataba]], [[Ryoba]], or [[Dozuki]]. [M1][M22] *(Mechanical from the sourced geometry; the deflection mechanism is sourced [M22].)*
- **Sustained work in thick stock.** Thin (1 mm), short (120–180 mm), moderate-pitch blades cut thick material slowly and at higher risk of binding, wandering, and poor chip clearance — not categorically impossible, just a poor tool for it. Use a wider saw when straight tracking, stiffness, or chip clearance matters more than turning radius. [M6][M7]

### Dangerous (misuse that injures the user)
- **The springy unsupported blade and the exposed point.** Unspined and flexible, the blade can spring laterally if it binds and is suddenly freed, and it ends in a sharp tip near the working hand. The documented hazards are the exposed teeth and point, hand placement close to the cut, binding followed by sudden release, blade kinking or breakage under force, and the loss of control that can follow any of these — not a fixed "force always travels toward the body" direction, which in fact depends on stance and work orientation. Secure the work, keep the free hand outside the projected blade path, begin with light controlled strokes, and stop if the blade binds; excess force or twisting can kink or break the narrow blade and may cause a sudden loss of control. General keyhole-saw guidance: "do not apply excess pressure … the blade may become jammed." [M_kh] *(Inferred from blade geometry — see the honest flag below.)*
- **Honest flag — no first-person mawashibiki injury account was found.** A dedicated sweep of the pull-saw forums surfaced no documented laceration or near-miss specific to this saw. The hazard above is a mechanical inference and a documentary gap, **not** a sourced injury — the same discipline the [[Ryoba]] (Run 126) and [[Azebiki]] (Run 131) held. I will not manufacture an injury the record does not contain.

### Damaging (misuse that ruins the tool or the work) — *the defining failure*
- **Kinking / buckling the blade by forcing too tight a radius.** This is **the** mawashibiki failure and the exact price of the narrow blade: drive it around a curve tighter than the blade can pivot and the back jams the kerf wall; force it and the thin unspined plate takes a permanent lateral bend (a kink) it can never track straight through again. "Narrow blades break easily." [M21] The accumulated-kink death is documented on a real blade: bind, bend, repeat, until the steel is "a patchwork quilt of bumps and hollows." [M20]
- **Twisting in the kerf.** Torquing the handle to force a sharper turn while the blade is buried loads the steel at the heel, where it enters the handle — the place a kink most often starts.
- **Snapping the impulse-hardened kae-ba blade.** The hardness is concentrated at the tooth tips, not the whole plate — impulse-hardened tips (above ~RC 67) are brittle and not file-sharpenable, so a chipped or dulled edge is replaced, not restored, while the softer blade body can still bend under a kink rather than crack. [M22] The fully-hardened artisan blade is a different, more brittle case end-to-end — hardened through the whole plate, not just the teeth: "breaks with improper handling … broke into three pieces." [M13]

### Beginner-trap (looks fine until it isn't)
- **Turning too sharp, on the wrong part of the blade.** Tighter radii belong on the **tip-end**; a beginner who turns hard on the wide heel binds and kinks. (Correct technique uses the toe for tight curves; for a really small radius a [[Bow Saw]] is the better tool.) [M_fww]
- **No start hole.** The mawashibiki is not a plunge-starter — that is the [[Azebiki]]. Try to rock-start it on a flat surface and the tip skates and scores the work. Drill (or azebiki) first. [M16][M21]
- **Pushing a pull saw.** Habit from Western saws puts the thin unspined blade into compression; it buckles, won't cut, and risks a kink. [M21][M22]
- **Letting it wander off the line.** No spine means no lateral guidance; without continuous sighting and micro-corrections the flexing blade drifts before you notice. [M22]

### Lineage-misuse (treating it like its relatives or ancestors)
- **Treating it like the [[Azebiki]] (the critical one).** The azebiki **starts** an interior cut; the mawashibiki **follows / enlarges** one. They are a sequence, not substitutes — use the mawashibiki to plunge-start and you skate and score the panel; use the azebiki to chase a long curve and you've brought the wrong half of the pair. [M1][M16]
- **Treating it like a rigid Western compass / coping saw.** Western compass saws are stiffer and often push-stroke; bring their force levels and stroke to a Japanese pull blade and you overdrive it into a kink. Even within the Western family the two differ — "the keyhole saw … a narrower blade with less taper than the compass saw." [M23]
- **Treating it like a stiff backed [[Dozuki]].** The dozuki's spine buys straight tracking; the mawashibiki gave that spine up on purpose. Expect dozuki-straightness from it and it wanders. [M14]
- **Treating it like a scroll-saw blade.** A scroll blade is far thinner and held in constant tension between two clamps — that is what buys its tiny radius. A mawashibiki, wide at the heel and far less flexible, kinks if you ask scroll-saw radii of it.

---

## Correct use
*(Brief — the workbench holds the deep operation.)* Following a curved or scrolled line; **enlarging a drilled or plunged hole** into a shaped interior opening; cutting out fretwork and the pierced patterns of *ranma*; clearing waste from a mortise or joint between drilled relief holes; reaching into small, awkward interior spaces a framed or backed saw can't. [M5][M16][M17] The standing sequence is **start hole first, then turn**: a drill or an [[Azebiki]] opens the surface; the mawashibiki's narrow blade enters and follows the line. [M16][M21] The narrow blade is the capability — it pivots where a wide blade would jam — and also the whole liability, which is the next section.

---

## Matching


**Does.** Follows a curved or scrolled line; enlarges a drilled or plunged hole into a shaped interior opening; cuts fretwork and the pierced patterns of *ranma*; clears waste from a mortise or joint between drilled relief holes; reaches into small, awkward interior spaces a framed or backed saw can't.

**Suits.** Tateguya (shoji/fusuma/ranma work); sashimono-shi (cabinetmaking — keyholes for hardware, shaped interior cuts); general finish carpentry doing interior cutouts started from a drilled hole; makers who already own a drill or an [[Azebiki]] to open the start hole. Named-maker forms (Mitsukawa) suit conservation-minded buyers wanting a hand-forged, fully-hardened blade; consumer forms (Kakuri, Suizan, Hida) suit general shop use.

**Kit-fit.** Pairs with a drill (or the [[Azebiki]]) as the opening tool — the mawashibiki doesn't plunge-start its own hole. Complements a [[Kataba]], [[Ryoba]], or [[Dozuki]] for the straight cuts this saw can't do well. Sits alongside a [[Bow Saw]] / [[Frame Saw]] as the *other* turning answer (frameless vs. framed) for tighter radii than this blade can take.

**Not-for (project-keyed).**
- Long, precision-critical straight runs — reach for a [[Kataba]], [[Ryoba]], or [[Dozuki]] instead.
- Sustained cutting in thick stock — a wider saw will track straighter and clear chips better; the mawashibiki can do it, just slowly and at higher risk of binding.
- Projects where a resharpenable edge matters long-term — modern impulse-hardened kae-ba teeth are not file-sharpenable; budget for blade replacement, or choose a fixed-blade traditional form if resharpening service is available.

**Access.** Consumer fixed- and replaceable-blade (kae-ba) forms (Kakuri, Suizan, Hida, Gyokucho/Razorsaw, Z-Saw) are the accessible, low-barrier entry point, with kae-ba keeping ongoing cost down. The hand-forged, fully-hardened artisan form (Mitsukawa, Miki) sits at the premium/harder-to-access end and is explicitly not beginner-tolerant per its own dealer description.

**Revival.** Not endangered — currently produced across multiple active mills (Gyokucho/Razorsaw est. 1969, Z-Saw/Okada est. 1943, Nakaya est. 1907, Kakuri, Suizan). The narrower thread worth tracking is the named-master hand-forged line (Juntaro/Daizo Mitsukawa) against the more common industrial consumer forms.

---

## Anatomy

| Part | What it is / does |
|---|---|
| **Blade (*ha* / 鋸身)** | Long, narrow, **tapering from a wider heel to a point** — the taper is what lets the toe turn inside a kerf a wider blade would jam. Cuts on the **pull**. [M5][M9] |
| **No spine** | Deliberately **unspined** — the enabling decision. A back spine (as on the [[Dozuki]]) would stiffen the blade and forbid both the flex needed to follow a curve and the insertion through a small hole. Encyclopedia form-class: *鋸身自立形式*, self-supporting blade. [M1][M14] |
| **Point / toe (*kissaki*)** | The narrow tip that enters the bored or plunged start hole and leads the turn. The tip-end is the part used for the tightest radii. [M5] |
| **Teeth** | Pull-cut, general/"universal" pitch; modern consumer product lines (Kakuri, Hida) run roughly **13–15 TPI**, though the fully-hardened artisan form runs far finer (22 TPI) — TPI is a product-line variable, not a fixed family spec. [M6][M11][M13] Modern mass-production teeth are commonly **impulse-hardened** (tips above ~RC 67 — harder than a file, so **not resharpenable**; the body left softer and flexible); hand-forged artisan blades are hardened through the whole plate instead. [M22][M13] |
| **Handle (*e* / 柄)** | Wood (beech, oak) on modern consumer models; straight or — a named ergonomic variant — curved, to angle the wrist into confined work. [M6][M11] |

**Representative dimensions (manufacturer-direct).** Kakuri *hikimawashi* 120 mm: 253 mm overall, 120 mm blade, 14 TPI, 1.4 mm kerf, **1.0 mm blade thickness**, beech handle, carbon steel. [M6] Kakuri 180 mm: 316 mm overall, 180 mm blade, same 14 TPI / 1.4 mm / 1.0 mm. [M7] The Wood Works lists a 180 mm blade tapering **6 → 14 mm** in width at 1.0 mm thick. [M9] fine-tools.com lists a 180 mm / 14 TPI / 1.2 mm form. [M10] The fully-hardened Mitsukawa artisan form is finer and thinner: 70 mm, 22 TPI, 0.5 mm. [M13] These figures describe specific modern product lines, not a fixed family-wide spec — the Mitsukawa form shows how far an individual maker's blade departs from the consumer norm. Note the blade body, at 1.0–1.2 mm, is on the **thicker** end of the Japanese range (a fine [[Dozuki]] runs ~0.3–0.5 mm) — the extra body is what lets an unspined blade resist twisting in a curving kerf. [M10][M14]

---

## Variations (living)
- **Pull *hikimawashi* (引き廻し)** — the dominant form; blade tension on the cutting stroke keeps the thin blade tracking. [M2][M6]
- **Push *oshimawashi / tsukimawashi* (押し廻し / 突き廻し)** — the push-stroke form; encyclopedia-confirmed as a real type (引（突）回し鋸) [M2]; coarser teeth, chosen for wall-hollowing where chips must clear forward. [M8]
- **Fixed-blade (traditional)** vs. **kae-ba (替刃) replaceable-blade (modern)** — both current in the market. [M6][M11]
- **Double-edged mawashibiki** — a ryoba-style two-edge variant. [M21]
- **Sokomawashibiki / cooper's saw** — sub-form for the circular bottoms of wooden buckets. [M12]
- **Straight- vs. curved-handle** — the same blade offered with either grip. [M11]
- **Blade-length range** — roughly 100 mm (4″) compact to ~210 mm (8″); the 120 mm and 180 mm sizes are the market core. [M6][M7]

---

## Trades (where it appears, with trade-specific notes)
- **Tateguya (建具師)** — *shoji*, *fusuma*, and the pierced **ranma** transoms: curve-cutting and interior-opening work where the tool is commonly used. [M18]
- **Sashimono-shi (指物師)** — cabinetmaking: keyholes for hardware, waste cleared from joints, shaped interior cuts. [M18]
- **Miyadaiku (宮大工)** — temple/shrine restoration: one named maker's dealer copy places his hand-forged saws here and describes industrial saws as prohibited on significant work; not yet confirmed against an institutional or conservation source. [M13]
- **General daiku / finish carpentry** — any interior cutout started from a drilled hole. [M16]

---

## Lineage

### Ancestors
A [[Kataba]] **narrowed** — the type-form with its blade taken down to a tapering point so it can pivot inside a tight kerf. It belongs to the encyclopedia's "self-supporting blade" form-class (*鋸身自立形式*) — the unspined family the kataba heads — as opposed to the spine-reinforced ([[Dozuki]]) and frame-tensioned ([[Bow Saw]], [[Frame Saw]]) form-classes. [M1] A secondary source (JA Wikipedia) attributes an early-18th-century reference to the *Wakan Sansai Zue* (和漢三才図会, completed 1712 / prefaces dated 1713), placed in the "Hundred Crafts" volume as recording the tool under 引き回し — but Run 137 could not confirm this specific entry against the primary text, so it is carried here as a research lead, not a verified date. [M4] (See Honest edges.)

### Contemporaries (sibling tools that do a similar job differently)
- **The [[Azebiki]]** — the root's other short specialty saw and the one most confused with it. The azebiki **starts** a cut in the middle of a flat surface (rocks down on its convex edge); the mawashibiki **follows / enlarges** an opening already begun. A functional pair in sequence, not rivals (Finding 3). [M1][M5][M16]
- **The Western keyhole / compass / pad saw** — the convergent twin from the push tree. Same narrow tapered blade, same job, reached independently (Finding 1). Wikipedia's own family separates them: "the keyhole saw is characterized by a narrower blade with less taper than the compass saw." [M23]
- **The [[Bow Saw]] / [[Frame Saw]] (narrow-blade form)** — the Western tree's *other* turning answer: hold a narrow blade in a tensioning frame. The mawashibiki and the keyhole saw are the *frameless* turning answer (narrow the blade, accept the flex); the bow/frame saw is the *framed* one (narrow the blade, tension it straight). [M23]

### Descendants / modifications
- **Oshimawashi / tsukimawashi** (押し廻し / 突き廻し) — the **push-stroke** turning saw. Confirmed at encyclopedia grade: the Sekai Daihyakka writes the type as 引（突）回し鋸 — the turning saw in both 引 *(hiki, pull)* and 突 *(tsuki, push)* forms. [M2] Kakuri makes a 120 mm *oshimawashi*; coarser (10 TPI vs. the pull form's 14), and chosen "when hollowing out a wall," where the push stroke throws chips clear of the cavity behind. [M8]
- **Kae-ba (替刃) replaceable-blade form** — the modern impulse-hardened snap-in blade; coexists with the fixed-blade traditional form (Hida Tool's Maruoto is explicitly "replacement blade: not available"). [M6][M11]
- **Sokomawashibiki / cooper's saw** — a specialised sub-form named for cutting the circular bottom of wooden buckets. [M12]
- **Double-edged mawashibiki** — a less common ryoba-style variant carrying two tooth configurations. [M21]

---

## History
Read through Petroski's lens — *form follows failure* — the mawashibiki is the pull root's answer to a failure that is older and simpler than the buckling failure the rest of the root is organised around: **a wide blade cannot turn inside its own kerf.** Set a blade wider than the radius you are trying to cut and the back of the blade jams against the outside wall of the curve; force it and the kerf widens, the line wanders, or the blade kinks. Every tradition that ever cut a curved hole met this wall, and every one solved it the same way — take width off the blade until it can pivot. A secondary source attributes an early-18th-century reference (c.1713) for the 引き回し to the *Wakan Sansai Zue*, though Run 137 could not confirm the specific entry against the primary text [M4] — so this is held as a research lead, not a settled date; if it holds, it would place the solution in Japan's tool vocabulary well before the late-Meiji recombinations that produced the [[Ryoba]] and the double-edged [[Azebiki]]. What the mechanical argument establishes on its own, independent of that date, is simpler: the turning saw is **old**, because the turning problem is old.

What the mawashibiki could *not* do is start its own hole — and here the history runs straight into the [[Azebiki]]'s. The encyclopedia draws the line cleanly: the azebiki's "curved tooth-path takes effect when sawing out from the middle of a flat surface" (*平面の途中から挽き出す*), while the *hikimawashi* "has a narrowly-made blade and can cut curves" (*鋸身が細くつくられており、曲線を挽くことができる*). [M1] One opens the surface; the other walks the line. The standing workflow — draw the line, **drill or plunge a start hole, insert the narrow blade, then follow** — is the same one Kakuri publishes today [M16] and the same one a working joiner describes for cutting waste out of a mortise (drill the waste, run the mawashibiki to each end, pare to the line with a chisel) [M17]. The narrow blade is the whole tool; everything else is the price it pays for it (the misuse layer, below).

---

## ✦ Human side

**Whose tool it is.** The mawashibiki is used in **joinery, cabinetmaking, fitting, and decorative interior cutouts**, including work associated with the **tateguya** (建具師 — the makers of *shoji*, *fusuma*, and the pierced openwork transoms called **ranma** 欄間), where curved cutouts and shaped interior openings are common work, and the **sashimono-shi** (指物師 — cabinetmakers), where keyholes for locking hardware and waste cleared from joints want a narrow blade that starts from a bored hole and turns to shape. [M18] A dealer source describing Daizo Mitsukawa's traditionally-forged saws also places them in temple and shrine restoration under the **miyadaiku** — "for the restoration of significant shrines and temples … the use of an industrially manufactured saw is prohibited out of respect for the historical object" [M13] — but this is maker/dealer copy, not an institutional or conservation source, so it is carried here as a named maker's claim about his own saws' use, not a confirmed general miyadaiku practice.

**Named hands.** The fine artisan mawashibiki comes out of **Miki** (三木), Hyōgo — Japan's saw town. **Juntaro Mitsukawa** (三塚順太郎), a government-certified traditional craftsman and third-generation saw-smith, and his son **Daizo Mitsukawa** (三塚大蔵, b. 1961) make a hand-forged, fully-hardened mawashibiki in a small craftsman's series (70 mm blade, 22 TPI, 0.5 mm thick). [M13] The furniture and *shoji* maker **Brian Holcombe** names his own hikimawashi as a Juntaro Mitsukawa saw, and reaches for it "for small detail work such as cutting in small awkward areas … even cutting the side grooves of *dai* (plane bodies), though there's a purpose-built saw for that." [M14] The volume end of the family runs through **Gyokucho / Razorsaw** (玉鳥, Miki, est. 1969 — Japan's first replaceable-blade handsaw maker), **Z-Saw / Okada** (岡田金属, Miki, est. 1943 — first in Japan to impulse-harden saw teeth), **Nakaya** (中屋, est. 1907), and the consumer brands **Kakuri** (角利, Sanjō, Niigata) and **Suizan** (Tsubame-Sanjō). [M15][M6] A documented antique survives in dealer hands: a c.1960s mawashibiki by the maker **Senkichi**, laminated carbon steel, 14″ overall. [M24]

**The texture — what it feels like to get it wrong.** The pull-saw forums do not record the mawashibiki's failures under its own name, but they record the failure of the thin Japanese blade with unusual precision, and the mawashibiki is the thinnest-tracking of the family. Wiley Horne, on why a thin blade wanders: *"the blade will flex just a little in the kerf, and the teeth … will start heading for where the blade flex is telling them to go, not where you're wanting them to go. Don't get a death grip on the saw … let the saw fall through the work."* [M19] Robert Rozaieski, on the kink: *"If you flip that blade around [to push], you'll kink or break the dozuki in a heartbeat"* — and the mawashibiki, narrower and unspined, is more fragile still. [M19] Dieter Schmid, a saw dealer of thirty years, on the fully-hardened artisan blade: *"These saws are not 'beginner-tolerant.' The blade breaks with improper handling … My first fully-hardened handmade saw … broke into three pieces … I was so attached to it I worked with the remaining two-thirds for years."* [M13] And the slow death by binding, documented on a kinked blade under a waterstone: *"the cut bound up and the saw bent just a little bit. Over the years, the cumulative damage manifests itself as a patchwork quilt of bumps and hollows."* [M20]

**The texture — and what it feels like to get it right.** The same thinness that punishes a heavy hand is the whole virtue when the work is fine. The Norwegian joiner **Walter Egon**, making top-hinged windows in oak and mountain pine, bought three mawashibiki specifically to clear dovetail waste in thick stock where a bowing fretsaw blade was eating into his pins — and ran a head-to-head on white oak and pine: *"The second one struggled to get through even the pine; skipping and binding but seldom sawing nicely. The third saw had the thinnest, most flexible blade and cut both woods very well … I've found a new friend among the slender saws and my workshop is just a little bit improved."* [M26] That same thinness is also why the kink-from-forcing failure above is the price the tool is built to risk — Egon's result reads less as a universal rule for the whole tool class than as a confirmation that, for his specific dovetail-waste task, the more flexible blade was the better choice.

---

## ✦ Proof of existence
- **Dictionary / documentary (verified directly).** The Nipponica encyclopedia (*日本大百科全書*, Akao Kenzō, 2021) lists 挽廻 (*hikimawashi*) as one of the **nine canonical carpenter's-saw types** and defines it verbatim: 「挽廻鋸は、鋸身が細くつくられており、曲線を挽くことができる」 ("the hikimawashi has a narrowly-made blade body and can cut curves"). [M1] The Sekai Daihyakka (*改訂新版 世界大百科事典*, Narita Juichirō) lists the type as 引（突）回し鋸 — confirming **both** the pull (引) and push (突) forms exist — and grounds the whole pull tradition: 「日本の現在のものはすべて引く形式」 ("all current Japanese saws are the pull type"). [M2] These establish the form's existence and canonical standing at primary-reference grade.
- **Earliest textual record — claimed, with the metadata firmed but the entry itself NOT yet verified (Run 137).** The *Wakan Sansai Zue* (和漢三才図会) is the Osaka physician **Terajima Ryōan**'s illustrated encyclopedia; Run 137 firmed its bibliographic facts against library catalogues — **completed Shōtoku 2 (1712)**, prefaces dated Shōtoku 3 (1713), which is why both years circulate; the carpentry/craft tools sit in the **百工具 ("tools of a hundred crafts")** section, candidate **vol. 24 (巻24)**. [M4][M27] **But the specific claim that it records the turning saw as 引き回し could NOT be confirmed against the primary text** — the digitized NDL/Waseda facsimiles are image-only (no machine-readable text), and no secondary source quotes or cites the saw entry; the claim still rests on JA Wikipedia alone. So this is held as a *plausible-but-unverified* earliest-text claim, **not** a proof anchor — the form's existence is carried entirely by the dictionaries [M1][M2] and the physical antique [M3], which is why narrowing this claim does not lower the entry's confidence. (Route to close: a Japanese reader viewing NDL vol. 24 directly.) *(See Honest edges.)*
- **Physical museum object — rendered directly this run (Run 137).** **Powerhouse Museum (MAAS), Sydney — Object No. H986** (collection system ID **258547**), "Japanese saw (mawashibiki-nokogiri)." Object statement, read verbatim from the museum's own record: *"Japanese saw (mawashibiki-nokogiri), metal / wood, maker unknown, Japan, 1875-1892."* **Purchased 1892 from Mr C. J. T. Browne** (who ran a "Japanese Curios and Oriental Art Fabrics" store in Sydney); **acquisition date 21 June 1892.** [M3] A dated, accessioned, provenanced Meiji-era antique, acquired the year it entered the collection. The museum's **physical description confirms the form firsthand**: *"Thin, narrow saw blade with row of small sharp teeth. The teeth are triangular and follow the nezumi-ba style … a combination of ripping and crosscutting teeth, [sharpened] like knife points. The blade tapers from heel to toe and from cutting edge to top edge. The blade is tanged in a[n] oval shap[ed] wooden handle."* [M3] The recorded **width is 20 mm** (the only dimension given; no overall length recorded). The first pass held this object at search-confirmed because the page is client-rendered; **Run 137 rendered the page in a browser and read the full item-level record directly**, closing that residual — the blade's double taper (heel→toe *and* edge→edge) independently corroborates the anatomy section, and the nezumi-ba "knife-point" teeth match a general-purpose narrow saw. This confirms one dated, provenanced exemplar of the form firsthand; it does not establish the standard tooth geometry, dimensions, trade ownership, or an early-Edo origin for the mawashibiki family as a whole — those rest on the dictionary entries [M1][M2] and the modern manufacturer record, not on this object. *(Object No. H986 is the museum's object number; 258547 is the collection-system record ID — two identifiers for the one object, kept distinct, not conflated. The record was catalogued by intern Emma Johnston, 2011, citing Toshio Odate's* Japanese Woodworking Tools *and the Takenaka Carpentry Tools Museum.)*
- **Named-maker objects (dealer-grade).** A c.1960s laminated-steel mawashibiki by the maker **Senkichi**, 14″ overall [M24]; the hand-forged Mitsukawa (Juntaro / Daizo) artisan mawashibiki out of Miki. [M13]

---

## references
- [M1] **Kotobank → 日本大百科全書 (Nipponica), "鋸," Akao Kenzō, 2021** — the nine carpenter's-saw types (9種12丁) incl. 挽廻; verbatim def. "挽廻鋸は、鋸身が細くつくられており、曲線を挽くことができる"; the 鋸身自立形式/補強形式 form-classes; the azebiki "平面の途中から挽き出す" distinction. https://kotobank.jp/word/%E9%8B%B8-112123 · {kind: museum/dictionary; **verified directly this run**}
- [M2] **Kotobank → 改訂新版 世界大百科事典, "鋸," Narita Juichirō** — "日本の現在のものはすべて引く形式…外国ではすべて押してひく形式"; type list incl. "細身の引（突）回し鋸"; "曲線を挽く引（突）回し鋸." Same URL. {kind: dictionary; **verified directly this run** — sources the tsukimawashi push form}
- [M3] **Powerhouse Museum (MAAS), Sydney — Object No. H986 (collection system ID 258547), "Japanese saw (mawashibiki-nokogiri)," metal/wood, maker unknown, Japan, 1875–1892; purchased 1892 from C. J. T. Browne; acquisition date 21 June 1892.** Physical description verbatim: nezumi-ba (rat-tooth, combination rip/crosscut) triangular knife-point teeth; "blade tapers from heel to toe and from cutting edge to top edge"; tanged in an oval wooden handle; width 20 mm. Catalogued by intern Emma Johnston (2011), citing Toshio Odate & the Takenaka Carpentry Tools Museum. https://collection.powerhouse.com.au/object/258547 · {kind: museum; **rendered directly in-browser, Run 137** — the client-rendered page read in full this run; first pass had it search-confirmed only}
- [M4] **Wikipedia (JA) — 日本の鋸**: claims Wakan Sansai Zue (和漢三才図会) records 引き回し. https://ja.wikipedia.org/wiki/%E6%97%A5%E6%9C%AC%E3%81%AE%E9%8B%B8 · {kind: reference; **the saw-entry claim is UNVERIFIED against the primary — Run 137**, see [M27]; Wikipedia is its only source}
- [M27] **Wakan Sansai Zue (和漢三才図会) — bibliographic metadata (Run 137):** compiler Terajima Ryōan (寺島良安); completed Shōtoku 2 (1712), prefaces Shōtoku 3 (1713); 百工具 craft-tools section = candidate vol. 24 (巻24). Confirmed via NIJL Kokusho DB (https://kokusho.nijl.ac.jp/biblio/200018257), Waseda Kotenseki catalogue, and NDL Digital Collections (facsimile image-only, not machine-readable). · {kind: reference/library-catalogue; **metadata confirmed; the turning-saw ENTRY itself NOT confirmed in the text**}
- [M28] **Takenaka Carpentry Tools Museum — saw development history (鋸の発達史)**: Edo-period gains in tool-production capacity and occupational specialization subdivided saws by use, assembling "most of the forms of traditional saws seen today." https://www.dougukan.jp/tools/12 · {kind: museum}
- [M5] **Japan House LA / Takenaka Carpentry Tools Museum — hikimawashi-noko exhibition entry**: "keyhole saw … curves that penetrate through the material … 7–8 inches … cuts on the pulling stroke." https://www.japanhousela.com/exhibitions/masters-carpentry-melding-forest-skill-spirit-woodworking-daiku-takenaka-dougukan/tools/ · {kind: museum}
- [M6] **Kakuri — Hikimawashi (keyhole) pull-cut 120 mm**: 253 mm overall, 120 mm blade, 14 TPI, 1.4 mm kerf, 1.0 mm thick, beech/carbon steel. https://kakuritools.com/products/kakuri-keyhole-saw-pull-cut-120-hikimawashi · {kind: maker}
- [M7] **Kakuri — Hikimawashi pull-cut 180 mm**: 316 mm overall, 180 mm blade, 14 TPI, 1.4 mm, 1.0 mm. https://kakuritools.com/products/kakuri-keyhole-saw-pull-cut-180-hikimawashi · {kind: maker}
- [M8] **Kakuri — Oshimawashi (keyhole) push-cut 120 mm**: 10 TPI, 1.5 mm kerf; "when hollowing out a wall, difficult for chips to fall to the back side." https://kakuritools.com/products/kakuri-keyhole-saw-push-cut-120-oshimawashi · {kind: maker}
- [M9] **The Wood Works Book & Tool Co. — Mawashibiki #712613**: blade 180 × 6–14 × 1.0 mm, pitch 3.4 mm, overall 310 mm. https://www.thewoodworks.com.au/ · {kind: blog/dealer}
- [M10] **fine-tools.com — Japanese Keyhole Saw 180 mm / 14 TPI / 1.2 mm.** https://www.fine-tools.com/Japanese-Keyhole-Saw-180mm-14-TPI-Blade-Thickness-1.2mm/312007 · {kind: blog/dealer}
- [M11] **Hida Tool — Maruoto hikimawashi**: 13 TPI crosscut; "straight or curved handles … for the same saw blade"; replacement blade not available. https://hidatool.com/item/2621 · {kind: blog/dealer}
- [M12] **Japan Woodcraft Association — Saws**: kataba style; "softer steel … more flexible when cutting round corners and less at risk of breaking"; sokomawashibiki/cooper's-saw note. https://japanwoodcraftassociation.com/traditions/tools/saws/ · {kind: blog}
- [M13] **fine-tools.com — Mawashibiki, Craftsman Series, Daizo Mitsukawa (Miki, b. 1961; succ. Juntaro Mitsukawa)**: 70 mm, 22 TPI, 0.5 mm, fully hardened; "not 'beginner-tolerant' … broke into three pieces"; temple/shrine restoration use. https://www.fine-tools.com/Mawashibiki-Saw-Keyhole-Saw-Craftsman-Series-by-Master-Daizo-Mitsukawa-Blade-Length-70-mm-22-TPI/313535 · {kind: named_maker/dealer}
- [M14] **Brian Holcombe — Brian's Guide to Japanese Saws**: hikimawashi "small detail work … cutting in small awkward areas … side grooves of dai"; his by Mitsukawa Juntaro; dozuki-spine contrast. https://brianholcombewoodworker.com/brians-guide-to-japanese-saws/ · {kind: blog/named_maker}
- [M15] **Working by Hand — Exploring Japanese Saws IV: Saw Makers**: Gyokucho/Razorsaw (1969), Z-Saw/Okada (1943, impulse-hardening), Nakaya (1907), Hishika, Suizan. https://workingbyhand.wordpress.com/2024/03/02/exploring-japanese-saws-iv-saw-makers/ · {kind: blog}
- [M16] **Kakuri — How to use a Japanese saw**: draw line → drill/punch start hole → insert hikimawashi → follow line. https://kakuritools.com/blogs/how-to-use-japanese-saws/how-to-use-a-japanese-saw · {kind: maker}
- [M17] **Adrian Preda — Japanese saws**: drill the waste, run the mawashibiki to each end, pare to the line with a chisel. https://www.adrianpreda.com/blog/japanese-saws · {kind: blog}
- [M18] **Japan Woodcraft Association — Disciplines**: tateguya (shoji/ranma), sashimono-shi. https://japanwoodcraftassociation.com/traditions/disciplines/ · {kind: blog}
- [M19] **Sawmill Creek — "Pros and cons of Japanese saws"**: Wiley Horne (thin-blade flex/wander); DeCristoforo & Wilbur Pan (brittle teeth, breakage from too much force); Rozaieski ("flip that blade … kink or break … in a heartbeat"). https://sawmillcreek.org/threads/what-are-the-pros-and-cons-of-japonese-saws.59700/ · {kind: forum}
- [M20] **My Peculiar Nature — Japanese saw tune-up: straighten out**: accumulated kinks from binding ("the cut bound up and the saw bent … patchwork quilt of bumps and hollows"); Mark Grable / Yataiki straightening. http://mypeculiarnature.blogspot.com/2015/05/japanese-saw-tuneup-straighten-out.html · {kind: blog}
- [M21] **Daitool — Japanese saw types**: "Narrow blades break easily"; double-edged variant; "cut holes … started with a drill"; azebiki "only option when cutting into a board's surface without using power tools." https://daitool.com/blogs/daitool-blog/japanese-saw-types-choose-the-best-pull-saw-for-you · {kind: blog/dealer}
- [M22] **Sawyer Design (Nick Sawyer) — The Japanese pull saw**: impulse-hardened teeth not file-sharpenable; thin blade deflects under lateral pressure. https://sawyerdesign.io/blog/tools-trade-2026-05-05/ · {kind: blog}
- [M23] **Wikipedia — Keyhole saw**: "narrower blade with less taper than the compass saw"; retractable-blade form retracts blade "to prevent unwanted flex." https://en.wikipedia.org/wiki/Keyhole_saw · {kind: reference}
- [M24] **Shogun's Gallery — c.1960s Senkichi mawashibiki-nokogiri**, laminated carbon steel, 14″ overall / 7″ blade. https://shogunsgallery.com/products/antique-japanese-c1960-s-steel-mawashibiki-nokogiri-key-hole-saw-14 · {kind: dealer/antique}
- [M26] **Walter Egon (Norwegian Woodworker) — "Selecting a keyhole saw," 2024-09-24**: professional joiner's first-person comparative test of three mawashibiki on white oak and pine for clearing dovetail waste; the thinnest/most-flexible blade won — "I've found a new friend among the slender saws." https://walteregon.substack.com/p/selecting-a-keyhole-saw · {kind: blog/named-practitioner; firsthand-read}
- [M_kh] **HowStuffWorks — keyhole saw**: general guidance, "do not apply excess pressure … the blade may become jammed." https://home.howstuffworks.com/keyhole-saw.htm · {kind: reference; general keyhole-saw, applied by extension}
- [M_fww] **Fine Woodworking forum — sharpening compass & keyhole saws**: use the tip-end for tighter radii; for a really small radius a bow saw is better. https://www.finewoodworking.com/forum/sharpening-compass-and-keyhole-saws · {kind: forum}

## The reviewer's corrections (12)

- `BLOCKER` — 1. Remove the mandatory drilled-hole / azebiki relay
- `BLOCKER` — 2. Correct the Mitsukawa identity and kanji
- `SHOULD-FIX` — 3. Recast the convergence and direct-descent thesis as interpretation
- `SHOULD-FIX` — 4. Correct the claim that turning failure is “not directional”
- `SHOULD-FIX` — 5. Make the identity and stroke fields variant-specific
- `SHOULD-FIX` — 6. Correct the Mitsukawa product specifications
- `SHOULD-FIX` — 7. Separate blade construction and resharpening by model
- `SHOULD-FIX` — 8. Bound materials, trades, and wall-cutting claims to products
- `SHOULD-FIX` — 9. Replace universal radius and failure claims with model guidance
- `SHOULD-FIX` — 10. Add the maker’s complete safety controls
- `MINOR` — 11. Reframe current availability and “not endangered”
- `MINOR` — 12. Remove process and model narration