Spirit Level

identity

Name: Spirit level (the name refers historically to spirit-based vial liquid; modern vial fills vary by maker and model and can include proprietary hydrocarbon or mineral-oil formulations, [manual, relayed] from the reviewer's reading of Starrett, Johnson and a Kincrome SDS). Also called a bubble level, or simply "a level." Regional/trade names for specific forms include torpedo level, box level, mason's level, line level, post level, bull's-eye level, and — in older trade usage — box level / gradefinder / inclinometer. Spirit level – Wikipedia; Bull's eye level – Wikipedia

Family / category: Layout/measuring tool — an instrument that indicates whether a surface is horizontal (level) or vertical (plumb), via a sealed, liquid-filled vial with a captive air bubble, mounted in a rigid frame.

One-line: A sealed, curved glass or acrylic vial holding a liquid and a trapped air bubble, aligned to one or more manufacturer-designated measuring surfaces. When the bubble is centered, the instrument indicates level or plumb within that model's stated accuracy, provided its reference surface and vial remain true and the model is used in a rated measuring position. It made many local level and plumb checks portable; water levels and plumb bobs keep their own reach (around corners, between points without line of sight, long vertical drops).

✦ how NOT to use it

not-for. Do not choose a level for machine-tool setup (lathe bed, mill table, surface grinder) from its category label: select the instrument's sensitivity and range from the machine maker's alignment procedure, a distinction discussed on the Practical Machinist forum. Practical Machinist, "Can someone explain precision levels specs to me?" Do not substitute a hand spirit level for a laser or rotary level on long-distance, multi-point site leveling — a Garage Journal poster's plain framing: "If you build a shop, having a rotary laser level is well worth it," implying the hand level is simply the wrong tool once runs get long. Not a powered-tool cutting guide. A spirit level is not a saw or router guide unless the exact model's maker explicitly rates it for that operation and the saw or router manual permits the setup; otherwise use a purpose-built, securely clamped cutting guide or track. A level may have removable end caps, magnets, handles, coatings or a profile that cannot be clamped securely or cleared by the saw; a shifting guide can ruin the work or contribute to loss of saw control, and blade contact damages the guide (the review). Forum reports of tradespeople doing it for years do not establish safe compatibility. Do not strike a level that lacks manufacturer-designated strike zones. Do not set a shower or drainage fall from a fraction of a bubble.

dangerous. Absence of a located incident report is not a safety classification. The hazard changes when the tool is struck, placed overhead, used near a flame, leaks, carries strong magnets, or is brought into a powered-saw setup. Controls (the review, [manual, relayed]): inspect before use; keep damaged or leaking levels out of service; for leaked fluid follow the product's SDS or maker on ignition control, ventilation, contact and aspiration (at least one current vial-fluid SDS lists a flammable isoparaffinic hydrocarbon; another maker uses mineral oil; do not generalize); protect eyes and hands from glass or acrylic fragments during cleanup; keep strong-magnet models controlled around ferrous pinch points and sensitive equipment; do not work below an unsecured level; strike only designated strike zones; never use it as a powered-tool guide unless rated (see not-for). Heat exposure near an open flame: A named forum poster (KnurledNut, The Garage Journal) warns directly: "Just don't get a torch too close to a Stabila vial. Ask me how I know" — nearby welding, brazing, or pipe-sweating heat can compromise the vial's fluid or seal, a jobsite-plausible scenario for the trades (plumbing, HVAC) that carry a level alongside torch work. The indirect risk worth naming honestly: relying on an out-of-calibration level for safety-critical structural work (scaffolding plumb, ladder footing, deck ledger attachment) compounds error across a run — plausible and consistent with the tool's known failure modes, but no source this run ties a specific documented injury or collapse to it; flagged as inferred risk, not a confirmed incident.

damaging. ★ Drop damage is real but genuinely inconsistent, and the inconsistency is the actual lesson, not a clean rule. On one Garage Journal thread alone: one poster's three Empire levels went out of true simply from "falling over and/or being dropped," while another dropped a 6" Empire "about 40 feet off a stud block onto concrete" and it stayed true, and a third recounts an uncle's 4' Stabila surviving a 10-story scaffold drop with no loss of accuracy. Multiple named posters (readhead, KnurledNut, Stuart in MN) converge on the same working rule regardless of brand reputation: re-check calibration after any hard drop — don't assume either way. The Garage Journal, "Lets talk levels" Heat exposure near the vial (see Dangerous) is also a damaging failure mode, not just a hazard. Forum posters also report that a level used regularly as a saw fence "won't be a true level for very long"; that is one more reason not to use it as a guide (see not-for), not a guide to doing so. A level that is dropped or gives conflicting readings comes out of consequential layout until it passes the repeated check; do not automatically turn it into a screed or straightedge, since the same event may have bent the reference surface.

beginner-trap. ★ Never checking a level's own accuracy is the single most-repeated finding across every source this run touched, and chile_head's first-person account (Fine Homebuilding) is its clearest anchor: a first-time homeowner tiling a basement floor with a cheap level didn't realize "I could lift one end of the level about a quarter inch before the bubble even started to move" — the level, not the floor, was the actual problem, and he didn't find out until the tile job was already visibly wrong. The fix, named consistently across three separate forums (Sawmill Creek, Fine Homebuilding, The Garage Journal) as the "flip test" or "read flip, turn flip" (EricPaulson's shorthand). The exact procedure is the maker's (see Correct use): for the level vial, end for end back onto a marked line; for a plumb vial, side for side, not end for end; each plumb vial separately; repeat; judge within the model's stated tolerance. A second, closely related trap is assuming a new or expensive level is automatically accurate off the shelf — multiple named posters (Jim Becker, Dave Sabo) describe checking five levels on a store shelf and finding at least one out of spec, even from reputable brands, and recommend checking in-store every time rather than trusting price or brand alone. A third, more technical trap: parallax error — reading the vial from an angle instead of straight-on, which shifts the apparent bubble position; confirmed as a named error class in surveying/leveling instructional literature, and most likely to bite in confined-space work (inside a cabinet, under a sink) that forces an off-angle read. Calendar UK — parallax error in levelling

lineage-misuse. ★ Over- or under-trusting the tool relative to its precision cousins runs in both directions, and a named machinist forum poster states it as a two-way rule rather than a one-way warning: a 0.0001"/10" precision machinist level is "slightly overkill for masonry or carpentry work," while an ordinary carpentry-grade level is inadequate for leveling a machine-tool base — "one man's level is another's useless long stick depending on the application." [The Garage Journal — Mr. T] A bowed profile on any level is treated as damage or an unverified condition: run the maker's accuracy check before relying on it. (A single forum account of intentionally bowed mason's levels was removed; no manufacturer specification supports it.) A third, more speculative case worth naming honestly: trusting a smartphone accelerometer-based level app with the same confidence as a physical spirit level — this is a plausible descendant-trust risk implied by the tool's own lineage (see Matching), but this run did not locate a strong, directly sourced forum debate specifically comparing app accuracy to vial accuracy; flagged as a real gap rather than manufactured texture.

correct use

Genuinely good at: checking level (horizontal) and plumb (vertical) trueness on a single surface, and showing gaps or relative deviation over its contact span; marking a level reference line along a measuring edge intended for marking; checking a 45° angle as well only on a model with an actual 45° vial, or a calibrated readout rated for that angle (vial count alone does not establish it: STABILA's 196-2 K, for example, has one horizontal and two vertical vials (relayed from the review, not re-read here; a direct fetch from this session failed) STABILA, Type 196-2 K).

How to use it well (short version — the workbench holds the deep operation): read the bubble straight-on, not from an angle (parallax error shifts the apparent reading). Check the instrument before consequential work and after a drop, by the maker's procedure. STABILA's (read this pass): for the level vial, mark the surface at one end, read the bubble, turn the level 180° end to end back onto that mark, read again; for a plumb vial, mark beside the top vial and "flip the level side for side (not end for end!)"; test the second plumb vial separately; "do it again, regardless of the reading." Clean the surface first, and compare within the model's stated tolerance (STABILA notes two in-spec levels can differ visibly). After a failed check, inspect for a bent profile or debris and follow the maker's evaluation, calibration or replacement route. Choose a length that spans the feature relevant to the task: a longer verified reference surface bridges more of the work and makes cumulative slope easier to see; a shorter level can reveal local conditions and fit confined spaces. For specified straightness or flatness tolerances, use a rated straightedge or appropriate metrology method.

✦ matching — the workbench's first read

does: checks level (horizontal) and plumb (vertical) via a centered bubble · checks 45° angles only on a model with an actual 45° vial or a calibrated readout rated for that angle (not inferred from vial count) · marks a level reference line along a measuring edge intended for marking · shows gaps or relative deviation over its contact span (not a flatness certification) · (line levels) gives an approximate level reference on a taut line, from which grade is set by measured offset · (digital models) reads a precise numeric angle or % slope.

suits: wall and roof framing · hanging cabinets, shelves, and picture frames · installing doors and windows · deck and fence building · tiling (with a pitch vial, measured elevations or a calibrated digital level for any specified fall) · checking appliance or furniture footing · machine-tool setup (lathe/mill/grinder beds, precision tier only).

kit-fit:

• Pairs with: a straightedge (many box levels double as one) · a framing square (level + square is a core layout pairing) · a Laser Level for long-distance or multi-point leveling across a room · a plumb bob for old-school vertical reference · a line level + mason's string line for long masonry or landscape runs.
• Substitutes for, or is substituted by: a Laser Level replaces a spirit level for projecting one continuous level plane across an entire room or to multiple people at once; a smartphone level app can substitute for quick, low-precision checks, though tradespeople on forums are consistently reluctant to trust an app's sensor calibration for structural or finish work (a real, if thinly documented, trust gap — see Lineage-Misuse); between carpentry and machine-tool work, substitution is decided by the instrument's sensitivity and range against the task's procedure, not by its category label (see How NOT, not-for): a level too coarse for a machine maker's alignment procedure cannot stand in for one that meets it, and a very sensitive precision level can make a coarse carpentry setup impractical.
• Cannot replace: a Laser Level's ability to project a level plane across a whole room to several people simultaneously; a surveying transit or Theodolite's site-scale range and precision for grading work (Haft Run 561, 2026-08-28, closed this citation gap with a dedicated entry); a precision level whose sensitivity and range are selected for a named machine-tool procedure.

not-for (project-keyed): precision machinist work (lathe/mill/grinder setup) with a standard carpentry-grade level — the tolerance gap is real and sourced (see Lineage-Misuse) · long-distance, multi-point site leveling that actually needs a laser or rotary level · a powered-tool cutting guide: do not use a spirit level as a saw or router guide unless the exact model's manufacturer explicitly rates it for that operation and the saw or router manual permits the setup; otherwise use a purpose-built, securely clamped cutting guide or track. Johnson Level's "Straight Edges and Cutting Guides" page (read this pass) describes purpose-built straightedges clamped at both ends with c-clamps and the blade lined up with the saw off; it does not authorize a spirit level as a guide. Johnson Level & Tool — Straight Edge and Cutting Guides · striking any level without manufacturer-designated strike zones · inferring a numeric slope from a fraction of a generic bubble.

access: operation-specific. Reading a clean, checked level on a local surface is beginner-accessible. Shower slope, drainage grade, structural plumb over long runs, precision machine setup and any use beside a powered tool need their own method, tolerance and safety controls. Reading it accurately has a real learning curve: avoiding parallax error, understanding that a centered bubble only proves the vial (not necessarily the whole tool or the surface) is level, and knowing to self-check the tool's own calibration before trusting a reading on anything that matters.

revival: a genuine "pick this up" case, not manufactured nostalgia: an old brass-bound wood level (mahogany or rosewood body, brass edge guards) can still be a usable shop tool, not merely a wall-hanger, but a true vial is not enough: it needs a flip test in each rated orientation plus a straightness check of its measuring face against an appropriate reference, since the body may be bowed, twisted, worn or swollen and old vial fluid or seals may have changed — several named forum users (Mark Wooden, Sawmill Creek) report 20+ years of daily use from a single wood level bought once specifically to be "the last level I'd ever buy." That combination — genuinely functional AND beautiful — is rarer among heritage hand tools than the corpus usually finds, and it appeals directly to hand-tool traditionalists as an alternative to both throwaway plastic torpedo levels and pure smartphone-app leveling.

anatomy

• Frame/body — wood (traditional: mahogany, birch, or laminated strips), cast iron (older trade levels), aluminum (dominant modern material, extruded box or I-beam profile), or plastic/polymer (torpedo and budget models). Stabila Standard Spirit Levels
• Vial(s) — a curved (barrel-shaped, not straight) glass or acrylic tube, slightly bowed so the bubble centers at the high point when level, holding a liquid whose composition varies by maker (proprietary "vial spirit," isoparaffinic hydrocarbons in at least one SDS, mineral oil in Starrett machinist vials, [manual, relayed]); for broken-vial handling follow the product's SDS or maker, not an assumption of alcohol. A general construction level has at least one level vial; many have one or more plumb vials, and some add fixed-angle or pitch vials. Count, visibility, reverse-use rating and accuracy vary by model (STABILA's 80 AS has one horizontal and one vertical vial, the 80 AS-2 adds a second vertical, [manual, relayed]). Johnson Level — Spirit Level Vials; Harmonyglass — vial liquid
• Vial mounting — consumer levels typically seal/epoxy-lock the vial permanently (prevents fogging/leaking but cannot be recalibrated); precision machinist levels instead house an adjustable main vial in a knurled brass tube on a precision-machined base, built specifically for recalibration. [Stabila vial installation technology]; Starrett Machinists' Levels
• Measuring surface(s) — the manufacturer-designated measuring/reference surface; whether it is milled, powder-coated, magnetic, grooved or rated for reverse use is a model fact, not a class rule (STABILA describes both coated and milled surfaces, [manual, relayed]). A level's edge may serve as a pencil-marking edge only if it is straight, clean, undamaged and intended for marking.
• End caps — shock-absorbing rubber or molded polymer, protecting the frame ends and working edges from drop damage.
• Magnetic strips — rare-earth magnets along one edge, for hands-free use on steel studs, ductwork, conduit, or pipe; some pair a magnetic edge with an integrated V-groove for seating on round pipe.
• Handles/grips — longer box and I-beam levels (4–8 ft) sometimes add molded grip zones for two-handed carrying.

variations (living)

• Torpedo level — ~6–12" (9" most common), tapered ends, usually 3 vials, often magnetic — for tight spaces, outlet boxes, cabinet install, pipe/conduit work.
• Box level / carpenter's level — 2–4 ft, hollow rectangular ("box") extruded profile, commonly the more rigid design — the general framing/cabinetry default.
• I-beam level — often 4–8 ft, open "I" cross-section, often lighter and cheaper, and commonly less torsionally stiff than a box profile — used where weight/portability matters more than absolute rigidity.
• Mason's level — commonly 4–6 ft, box-beam; some specific models have manufacturer-designated strike zones for gentle trowel taps (e.g. STABILA 196-2 K, [manual, relayed]); a mason's level without them is not to be struck.
• Line level — very small (~3"), hangs on a string via hooks, no frame edge at all — gives an approximate level reference over a reasonable span; Johnson says to attach it "to the middle of the string" and that "even the slightest dip in the string will negatively affect your reading." Grade is created by a measured elevation offset from that reference, not read from the bubble.
• Post level — small molded plastic wrap-around form with 3 vials, straps around a post for hands-free plumb-checking on two faces at once — fence and deck post setting.
• Digital/electronic level — adds an LCD angle readout (degrees, % slope, in/ft) alongside or instead of a vial; typical accuracy ±0.05–0.1° at 0°/90°, wider at other angles.
• Machinist's precision level — short (4–12"), cast iron or steel base, knurled brass vial housing, sensitivity varies widely by instrument: Starrett's No. 98 series is not one figure: per Starrett's 98-series FAQ, the 98-4 is 0.015 in/ft and the 98-6, 98-12 and 98-18 are 0.005 in/ft (the 0.005 figure is roughly 80–90 arc-seconds), [manual, relayed] (relayed from the review, not re-read here; a direct fetch from this session failed) Starrett, 98 Series Levels FAQ; and 0.0005 in/ft is about 8.6 arc-seconds. Select an instrument whose sensitivity and range fit the machine maker's procedure and the surface variability; too sensitive a level can make a coarse setup impractical.

Note on box vs. I-beam: a closed box profile commonly resists twisting better than an open I-beam, at some cost in weight. This is a design tendency, not a universal hierarchy: alloy, section dimensions, ribs, openings, length and manufacture decide the actual performance of a given model.

trades

Carpentry and framing (walls, doors, windows) · masonry and bricklaying (course leveling; striking only a model with manufacturer-designated strike zones) · tiling (set and verify the specified fall per unit run with measured elevations, a model-specific pitch vial or a calibrated digital level; follow the governing code, system instructions, drain geometry and waterproofing plan; do not infer a numeric slope from an uncalibrated fraction of a generic bubble) · cabinetry and finish carpentry · deck and fence building · machinist/engineering work (machine-tool bed leveling, a wholly separate precision tier from construction trades). Tile Coach — Shower Pan Sloping

lineage

Direct ancestors.

• The Egyptian plumb-rule / "A-frame" level — a wooden isosceles-triangle frame with a plumb bob hung from the apex, aligned to a center mark on the crossbar, used to level surfaces with no liquid at all. A surviving example is held by the Grand Egyptian Museum. Grand Egyptian Museum — Plumb Line Level; Ancient Egyptian Survey Tools – DT Online (The bare plumb bob itself — the simpler, older, standalone ancestor tool — now has its own dedicated entry, Plumb Bob, built Run 450.)
• The hydrostatic (water) level — a long tube or trench of water, using the water's naturally flat surface as the horizontal reference; its use on monumental Egyptian construction is unresolved (the Grand Egyptian Museum record supports a plumb-line level, not a water trench or tube, and no archaeological source was located); per several tool-retailer histories, used by Roman builders. The Roman "libella" — a straight bar with a plumb line hung from its center — is repeatedly cited as a direct precursor. Flagged honestly: this libella claim traces to tool-trade blog content (Haus of Tools, alsyedconstruction.com), not to an archaeological paper or peer-reviewed source independently located this run — carried as plausible received trade wisdom, not verified history. Haus of Tools — History of Levels; alsyedconstruction.com — Origins of the Level

What the bubble vial fixed (Petroski lens). Relative to the plumb bob and the water level, the sealed vial is portable (no trench or tube of water needed on site), reads both horizontal and vertical from one small instrument once properly mounted, settles and reads faster than a swinging plumb bob, and is far less disturbed by wind or vibration than a hanging weight. It does not replace a communicating-water level between points without line of sight or a plumb bob over a long vertical drop. This reasoning is stated consistently by modern tool-history sources (Kapro, Hultafors, Haus of Tools) but is inference from the tool's known advantages, not a claim traceable to a surviving document written by the inventor himself — flagged as sourced-but-secondary reasoning, not a primary-document quote.

Contemporaries. The plumb bob remained (and remains) in parallel use for pure vertical work, especially where a level's vial isn't practical (deep excavations, very long drops). The Laser Level today runs alongside the spirit level as a modern high-tech parallel tool for long-distance work rather than a full replacement — see Matching.

Descendants. Torpedo levels → machinist's/engineer's precision levels → digital/electronic levels with numeric angle readouts → laser levels (line and rotary) → smartphone accelerometer-based level apps. Spirit level – Wikipedia

history

Origin — a well-attested 1661 invention, not just manufacturer folklore. Melchisédech Thévenot (c. 1620–1692), a French scientist, librarian to Louis XIV, and diplomat, is credited with inventing the bubble/spirit level — a sealed alcohol-filled vial mounted on a stone ruler with a viewing lens — some time before 2 February 1661. Per Wikipedia's account, which cites Turner (1992), the date comes from his correspondence with Christiaan Huygens and within about a year he circulated the design to Robert Hooke in London and Vincenzo Viviani in Florence. Neither the letters nor Turner's text were read for this entry. Melchisédech Thévenot – Wikipedia

The scholarly basis is Turner, Anthony J. (1992). "Melchisédech Thévenot, the bubble level, and the artificial horizon." Nuncius: Annali di Storia della Scienza, 7(1), 131–145 (DOI: 10.1163/182539192X00064): a peer-reviewed history-of-science study that analyzes seventeenth-century correspondence and the level's early development. It is secondary scholarship, not a primary document; the reviewer consulted its bibliographic record and abstract. The report that Adrien Auzout recommended the Académie Royale des Sciences bring "levels of the Thévenot type" on its 1666 expedition to Madagascar is attributed to the same secondary account; the expedition record was not inspected.

Honest caveats, named rather than smoothed over. The popular retelling of this story (repeated across tool-trade sites) traces, per Wikipedia's own sourcing, to a secondary 1887 Scientific American article (Munn & Co., Aug 27 1887) rather than to a primary 1661 document review — Turner 1992 is the better-grounded secondary citation. Both Huygens and Hooke are recorded as having later claimed the invention for themselves, within their own countries — the "sole inventor" framing carries real historical contestation, not just modern doubt. And the earliest surviving physical examples of bubble levels are generally dated to the early 18th century — the 1661 date is documentary (correspondence), not artifactual; no source found bridges that ~50-year gap with a surviving object.

Key evolutions, sourced.

Honest gaps named rather than papered over. No primary/archaeological source located for the Roman libella beyond tool-trade blogs. The Hosmer patent is now resolved as US 39,906 through the 1877 case record (above); the earlier "36,906" was wrong. The Stratton Bros. 1870/72 brass-bound-level patent number is likewise unconfirmed against a primary patent source. DATAMP itself (the Directory of American Tool and Machinery Patents) was not directly queried this run — a clear next step for pinning down 19th-century wood/brass level patents (Poole, Stratton, Davis, Stanley/Hall & Knapp).

human side

Traditions & trades. Used across nearly every hands-on trade: carpentry and framing, masonry and bricklaying, tiling (setting a specified fall), cabinetry and install work, deck and fence building, and — at a completely different precision tier — machinist/engineering work leveling lathe, mill, and grinder beds. Named brand traditions include Stabila (German precision-level maker, repeatedly named by tradespeople as the brand they trust without re-checking), Empire Level, Crick (wood levels, favored specifically by masons), Johnson, Swanson, Sola, Kapro, and the older wood/brass American makers above (Poole, Stratton Bros., Davis).

Named hands / quotes, real and sourced.

• Jim Becker (Sawmill Creek moderator): "The bubble centered only tells you that the vial that the bubble is in is level... it doesn't tell you if the 'level is level.'" Sawmill Creek, "Levels: Curiosity?"
• Larry Edgerton, a working pro on the same thread: "Over the years I have had them all, and all I have now is Stabila. They just stay true. They are not bothered by bad weather or humidity swings... as a pro they are a one time buy."
• Ray Grundhoefer's pole-shed story: helping a carpenter cousin who couldn't get poles plumb, Grundhoefer flipped the cousin's level 180° and it read correctly instead — the level itself had been out of true. The cousin, per the account, later became a building inspector.
• On-Wheel, a mason (The Garage Journal): "In masonry building leads you have LPS: level, plumb and straight edge. Leveling I beat the level with rubber handle of [an] Estwing hammer" — a practitioner account of tapping the level itself (not the individual brick), echoed in a Contractor Talk masonry thread about striking a level's trowel bumper. Kept as practitioner texture only, not as a use: only strike a masonry level if the exact model has manufacturer-designated impact or strike zones, using the specified tool, location and force (STABILA's 196-2 K permits gentle trowel taps on two purpose-built strike zones, [manual, relayed] from the reviewer); never strike the vial, an ordinary frame, an end cap or a non-rated level. The Garage Journal, "Lets talk levels"; Contractor Talk, "Masonry Torpedo Levels"
• chile_head (Fine Homebuilding forum), a first-time homeowner tiling a basement floor with a cheap level: "I couldn't figure out why the floor didn't look level until I realized that I could lift one end of the level about a quarter inch before the bubble even started to move." Fine Homebuilding, "Level accuracy"
• readhead, 40 years in the trade (The Garage Journal): "I had to teach every guy I hired how to check a level... you need to know how to check your level no matter how much you spent to know if it is telling you the truth."
• A recurring jobsite joke, quoted by PugetDude: asked what the best use for an out-of-whack level is, the answer given by his mentor carpenter was "a [expletive] concrete screed" — demoting a level that's failed calibration to straightedge/screed duty rather than throwing it away.

Manufacturer accuracy-warranty culture. Multiple named posters (DieselNut88 and others, The Garage Journal) report Empire Level replacing out-of-true levels for free on request, no return required — real evidence that accuracy drift is treated by at least one major manufacturer as a known, warrantied failure mode rather than a rare defect.

Honest gaps. No named, verifiable forum accounts were found for plumbing-trade use or casual DIY picture-hanging use specifically — the human-side record here skews carpentry/masonry-professional. Recorded as "not found by this search," not as evidence the record doesn't exist.

✦ proof of existence

Earliest dated/documentary example: the 1661 Thévenot correspondence (Huygens, Hooke, Viviani circulation) — documentary, not artifactual. Earliest surviving physical bubble-level objects are generally dated to the early 18th century per Wikipedia's own sourcing on the Thévenot article; no source bridges the roughly 50-year gap between the documented 1661 invention and the earliest surviving physical vial.

Museum / collection records:

• Smithsonian National Museum of American History holds a No. 101-A Leveling Instrument (L.S. Starrett Co., Athol, MA) — a sighting-tube level with a hanging bubble vial and graduated semicircle on a 4-legged tripod base. NMAH — Level
• NMAH also holds a Laser Level artifact: Spectra-Physics model 910, date made 1985 (read 2026-10-16), documenting the electronic-descendant lineage — now built as its own entry, Run 478. NMAH — Laser Level
• Grand Egyptian Museum holds a Plumb Line Level artifact — the pre-bubble ancestor. GEM — Plumb Line Level

Patents verified directly (real numbers, checked on Google Patents):

• US 2,211,201 — William B. Fell, "Machinist's precision level" (filed May 16 1938, granted Aug 13 1940): circular-bubble, spherical-surface level with crossed graduations. Google Patents, read 2026-10-16
• US 2,316,777 — William B. Fell, "Machinist's precision level" (filed Aug 1 1940, granted Apr 20 1943): circular-bubble level, mounting and temperature compensation. Google Patents, read 2026-10-16
• US 39,906 — Thomas N. Hosmer (Nov 11, 1862), vial securing and adjustment; held void in Stanley Rule & Level Co. v. Davis (1877). Case record, read 2026-10-16
• US 6,173,502 — Dane Scarborough, "Magnetically Engageable Level Sensing Apparatus And Standoff" (magnetic torpedo level, issued 2001).
• British Patent 1,259,531 — Quenot & Cie, "Improved Spirit Level Device" (published 1972).

Patents flagged as unverified, not repeated as fact: the Stratton Bros. c.1870/72 brass-bound level patent is referenced only via a dealer listing, not a primary patent-office source.

Honest gap: DATAMP (the Directory of American Tool and Machinery Patents) was not directly queried this run — the clear next step for pinning down 19th-century American wood/brass level patents (Poole, Stratton, Davis, Stanley/Hall & Knapp) with real patent numbers.

references

The reviewer's corrections (20)

• BLOCKER — 1. — Exact claim: “use as a cutting straightedge on tools/models not built for it — Johnson Level's own guidance states most levels ‘are not designed to act as a saw guide’ ... though box-beam levels tolerate this better than I-beam ones and several named tradespeople do it successfully for years, so this is a real …
• BLOCKER — 2. — Exact claim: “masonry and bricklaying (course leveling, using the level itself as a course-striking tool)” and the human-side account of beating a level with a hammer handle or trowel.
• BLOCKER — 3. — Exact claim: “tiling (shower-pan slope, typically read at ‘1/3 bubble’ rather than dead-center as the trade convention).”
• SHOULD-FIX — 4. — Exact claim: “the ‘spirit’ is the alcohol-based fluid in the vial” and “filled with a colored alcohol-based fluid.”
• SHOULD-FIX — 5. — Exact claim: the bubble centers “exactly when the frame's working edge is level or plumb.”
• SHOULD-FIX — 6. — Exact claim: “replacing a plumb bob or a trench of water with one portable, instantly-readable tool” and the Egyptian monumental water-level claim.
• SHOULD-FIX — 7. — Exact claim: “a genuine peer-reviewed source ... the strongest single piece of evidence located for any origin story in this corpus to date” and later “a stronger primary anchor.”
• SHOULD-FIX — 8. — Exact claim: “1859 — Stanley enters the level business by acquiring Hall & Knapp.”
• SHOULD-FIX — 9. — Exact claim: “Thomas N. Hosmer 1862 US patent (36,906)” and “Patents flagged as unverified.”
• SHOULD-FIX — 10. — Exact claim: “Fell ‘All-Way’ precision level, US Patent 2,211,201 ... first successful American bull's-eye level ... accuracy 0.0005 in/ft (~5 arc-seconds)” and later “down to 0.0005 in/ft, i.e. ~8.6 arc-seconds.”
• SHOULD-FIX — 11. — Exact claim: “1968 first dedicated construction laser,” “1973 first self-leveling laser level,” and “1989 digital levels introduced.”
• SHOULD-FIX — 12. — Exact claim: “Most models carry a horizontal vial and two vertical vials (one for each direction the tool is held).”
• SHOULD-FIX — 13. — Exact claim: “Working edge(s) — the milled/machined true bottom edge” and “a box level's 4-sided closed profile resists twisting and bending better than an I-beam's open profile.”
• SHOULD-FIX — 14. — Exact claim: “use the longest level the job allows ... since a longer working edge is inherently more accurate over distance” and “reveals a surface's overall flatness/consistency.”
• SHOULD-FIX — 15. — Exact claim: the flip test: “flip the level end-for-end (or side-to-side for the plumb vial) on the same surface; if the bubble reads identically both times, the vial is true.”
• SHOULD-FIX — 16. — Exact claim: “machinist's precision level ... 0.005 in/ft down to 0.0005 in/ft” and “a true machinist precision level's sub-arc-second sensitivity.”
• SHOULD-FIX — 17. — Exact claim: “long mason's levels are sometimes deliberately built with a slight intentional bow.”
• MINOR — 18. — Exact claim: “line levels check slope/pitch over long distances via a string line.”
• MINOR — 19. — Exact claim: “an old brass-bound wood level ... is still a fully usable daily-shop tool if its vials remain true.”
• MINOR — 20. — Exact claim: “No source located this run documents a direct physical injury ... genuinely thin,” followed by a torch anecdote as the one real hazard.