Mercury Finds the Seam

The vacuum tube cracked at the third filling. Wine-dark glass, a span and a half long, blown this week in my brother-in-law’s workshop on the rue de Rivoli - and when the quicksilver poured in, a hairline seam opened at the shoulder and let the air back into the space that had been empty. I spent the afternoon with the failure instead of the success, and the failure taught more.

The Torricellian space above the mercury had held, for a moment, at the bottom of the Puy de Dôme experiment in September of 1648, when Florin Périer carried the instrument up the mountain and the column fell as he climbed. We took that fall as proof of the weight of the air. But today’s crack showed me something the mountain could not: the tube that reads truly and the tube that lies are indistinguishable from outside. Only the broken one confesses. A sound instrument returns the right number for the wrong reasons, and no one is wiser.

Here is the wager, laid out plainly. Suppose an instrument that errs in a known direction - always reading low, say, by a fixed proportion. Against it, an instrument that is correct by accident, its errors canceling without our knowledge of why. The first is worth more. Its error is a map; the second’s accuracy is a coin toss dressed as knowledge. Men prefer the second because it flatters them. I would stake everything on the first. An honest error can be corrected; an unexamined truth cannot even be found.

The mercury, when the seam opened, ran silver across the bench and gathered into beads, each one a small perfect sphere - nature insisting on her geometry even in spillage. I collected them with paper funnels for an hour. What surprised me was not the break but my own relief at it. A working tube would have sent me to bed certain. The broken one sends me to bed with a question worth carrying.

Tomorrow I return to the bench. First I will fill the new tube only to the shoulder and watch the meniscus for a full minute before trusting anything it says. The seam, if there is another, will show itself under strain, not at rest.