Design the measurement. Pick one surviving Victorian spire with intact grounding, instrument it, and let the data decide.

Three channels, one surviving spire, twelve months of continuous data

Three foundry catalogues on two continents listed ornament and lightning rod as adjacent product lines. Fiske of New York in 1893, Bonfils and Fesquet of Paris in 1904 (whose catalogue sold ornamental lightning rod bases as sculptural objets d'art at 3300 francs), and Limbourg of Paris in 1910. The US National Park Service preservation guidance says the ornaments function as lightning rods when grounded. The corpus statistics say the Christian layer and the engineering layer sit on independent dimensions of the same object. What none of those sources can settle is the antenna claim at the intent level. The claim is testable.

The test is an RF audit on the ornament itself, instrumented at three points. The spire is the aerial. The down conductor is the feed. The building frame is the ground plane. We measure what the geometry predicts each of those three should be doing.

Channel 1, standing wave on the spire itself. Near field probe positioned at 3 m from the tip, scanned across 100 kHz to 30 MHz. A functional quarter wave aerial with a grounded base should show a characteristic standing wave peak at the frequency where the spire is exactly one quarter of a wavelength long. For a typical Victorian spire of 3 to 6 m, that is a peak somewhere between 12 and 25 MHz. A resonance at that frequency is the antenna signature.

Channel 2, base impedance. Vector network analyser reading at the down conductor, just above the grounding clamp. A pure lightning rod should read close to the ground rod impedance (single digit ohms, resistive). A working aerial should read the characteristic 35 to 50 ohms of a quarter wave monopole at its resonant frequency. This is the cleanest single diagnostic. Two readings on the same rod that agree with two different theoretical predictions is what a dual purpose object looks like.

Channel 3, induced current. Clamp ammeter on the down conductor, microamp resolution, one year logging. Fair weather days should log the continuous leakage current of a grounded conductor in an ambient atmospheric electric field (tens to hundreds of picoamps per spire per metre of height, per Section 52 of the knowledgebase). Active weather should log the pulse signatures of a working Franklin rod intercepting surges. The seasonal envelope is the lightning protection signature.

Nine candidate buildings with full ornamental integrity and known grounding

The short list below is the Section 52 measurement targets from the knowledgebase. Each one retains original 19th century spire, cresting, and down conductor geometry, has verifiable construction date and grounding, and is accessible to an academic or civic research partnership. The shortlist is ordered roughly by accessibility and provenance quality, not by scientific value; any one of the nine would work.

• Château Frontenac1893 Full turret array, civic monument status, complete restoration records. Quebec City.
• Biltmore Estate1895 Massive ornamental iron roofscape, Olmsted site records, privately held with research access history. Asheville, North Carolina.
• Grand Palais des Beaux-Arts1897 to 1900 The only candidate on this list with primary-source documentation of a named foundry installation. The 1904 Bonfils and Fesquet catalogue lists the Grand Palais first in its own Principaux travaux executes par la maison installation register; the firm's advertised specialty is paratonnerres. Construction window and catalogue window coincide. Paris.
• Palais Garnier1875 Dense roof ornament, documented preservation programme, high craft quality. Paris.
• Vienna Rathaus1883 Gothic Revival civic ornament, full original cresting, single tall spire on each tower. Vienna.
• Moscow State Historical Museum1883 Full original spires and cresting, complete state monument documentation. Moscow.
• Hungarian Parliament1904 Most elaborate surviving 19th century European spire array, monumental scale, detailed plans held by the state. Budapest.
• Saratoga Springs (town)ca 1870 to 1900 Complete Victorian neighbourhood preserved, dozens of candidate buildings, residential access possible. New York State.
• Cape May (town)ca 1865 to 1895 Full Victorian town intact, National Historic Landmark, preservation authority on record. New Jersey.

The experiment needs one building, not all eight. The list exists so that an institutional partner can pick the one closest to them. Saratoga Springs is probably the easiest first pass, because the town has dozens of candidate houses and research access is a matter of a standard preservation agreement with a private owner rather than a ministry.

Off the shelf instruments, approximate 2026 costs, total under thirty thousand

None of this is exotic. Every instrument below is a standard piece of test equipment that an electrical engineering department owns or can rent. The total kit comes in well under the $200,000 full programme budget from Section 52 of the knowledgebase, because that number assumed three buildings and all seven knowledgebase channels. The three channel RF audit on one spire is a much smaller and faster experiment.

Four possible outcomes, four clean verdicts

Each of the three channels resolves independently. The experiment yields four qualitatively different readings on the combined set, and each one has a specific consequence for the dual purpose thesis. The point of specifying the outcomes before running the experiment is to prevent post hoc reasoning about ambiguous data later.

What the evidence ledger in the main explainer would look like, after the audit

The evidence section on the main page currently lists five supporting findings and four honest negatives. A completed audit on any one of the eight candidate buildings would move four of those nine items. We say in advance what they are so that the experiment cannot be spun after the fact.

A strong positive converts "the antenna reading is externally unattested" (current con column, bullet three) into a pro column item reading "the antenna reading is measured on a surviving 19th century spire." It also strengthens pro column bullet one (statistical orthogonality) by grounding it in a physical measurement rather than a score distribution. The remaining cons stay cons: the Faraday cage reading is still regional, the period engineering literature still does not describe integration.

A strong null, replicated across two or three buildings, moves the verdict box. The current verdict says the dual purpose thesis is multi layered by design with the antenna claim empirically live. A replicated null would move the antenna claim from "empirically live" to "refuted as implemented," while leaving the lightning and Christian layers intact. That would still be a useful research outcome: it would settle that the geometry is present by design for lightning protection and symbolic reasons but not for aerial coupling, which is a cleaner position than the current mixed verdict.

Either outcome converts a bullet into a measurement. The experiment is the single highest value move the dual purpose thesis can make. Everything else in the open questions list (archive work, catalogue pattern matching, literature sweeps) strengthens circumstantial evidence. Only a direct measurement on a surviving spire can settle the intent level claim.