Spirit Level

Cahoon claims a combination of vials in one stock, not a vial. Write down the two readings a combined level must give without being reset: level, and plumb.

Lay the steel ruler on the board and look along it against the light. Any gap is error. You cannot build a level on a bent stock.

Rest the workshop level on the board. Note which way the bubble sits. This is your baseline.

Mark a 100 mm seat along the centre of the board's top edge, square to the face.

A perfectly straight tube gives a bubble that slams end to end with no middle position. The slight curve is what makes the reading proportional. Sketch both cases.

Bend a 120 mm steel wire into a very shallow arc, rise about 1 mm over its length. This is the curvature a vial needs.

Place the surface plate down and rest the workshop level on it. Record the bubble position. A surface plate is the flattest thing in the workshop.

Turn the level 180° in place and read again. If the bubble reads differently, the level is out of true, not the surface. This reversal test is the whole art.

Slide one 0.10 mm feeler leaf under one end of the plate. Read the bubble shift.

Add leaves in 0.10 mm steps to 0.50 mm, recording bubble travel each time. Plot it on graph paper. The slope IS the instrument's sensitivity.

Angle = rise over run. 0.10 mm over a 300 mm plate is 0.0003 radians, about 166 minutes of a degree — sorry, 1.1 minutes of arc. Compute yours and check it with the protractor sketch.

Stand the level against a door frame. A combined level must read plumb from the same stock — that is Cahoon's claim in use.

Cup a hand over one end of the vial for a minute. The spirit expands and the bubble shrinks and moves. Temperature is a real error source, not a footnote.

US 35,298, “Improvement in Combined Spirit-Levels”, Alvin Cahoon Jr., Harwich, Massachusetts, granted 20 May 1862.

The spirit level itself is far older — it is generally credited to Melchisédech Thévenot in the 1660s, and this blueprint does not claim Cahoon invented it. What the 1862 patent addresses is the practical nuisance of nineteenth-century site work: carrying several single-purpose levels, or stopping to re-seat a vial to check plumb after checking level.

Why alcohol. ‘Spirit’ is literal. Ethanol has a much lower freezing point than water, so the tool survives an unheated workshop, and its low surface tension lets the bubble slide cleanly instead of sticking — a water bubble jerks and lies.

The reversal test in step 8 is the important idea, and it generalises far beyond levels. It is a way to separate the error of the instrument from the error of the work using nothing but the instrument itself. The same trick calibrates straightedges, squares and, in Whitworth's three-plate method, the surface plate this blueprint stands on. When you cannot appeal to a higher standard, you appeal to symmetry.

Note that the level needs no reference object at all. Every other instrument in this batch is ultimately traceable to a physical artefact kept in a vault somewhere. The level is traceable to the planet.