# My 2018 Mathematics A To Z: Kelvin (the scientist)

Today’s request is another from John Golden, @mathhombre on Twitter and similarly on Blogspot. It’s specifically for Kelvin — “scientist or temperature unit”, the sort of open-ended goal I delight in. I decided on the scientist. But that’s a lot even for what I honestly thought would be a quick little essay. So I’m going to take out a tiny slice of a long and amazingly fruitful career. There’s so much more than this.

Before I get into what I did pick, let me repeat an important warning about historical essays. Every history is incomplete, yes. But any claim about something being done for the first time is simplified to the point of being wrong. Any claim about an individual discovering or inventing something is simplified to the point of being wrong. Everything is more complicated and, especially, more ambiguous than this. If you do not love the challenge of working out a coherent narrative when the most discrete and specific facts are also the ones that are trivia, do not get into history. It will only break your heart and mislead your readers. With that disclaimer, let me try a tiny slice of the life of William Thomson, the Baron Kelvin.

# Kelvin (the scientist).

The great thing about a magnetic compass is that it’s easy. Set the thing on an axis and let it float freely. It aligns itself to the magnetic poles. It’s easy to see why this looks like magic.

The trouble is that it’s not quite right. It’s near enough for many purposes. But the direction a magnetic compass points out to be north is not the true geographic north. Fortunately, we’ve got a fair idea just how far off north that is. It depends on where you are. If you have a rough idea where you already are, you can make a correction. We can print up charts saying how much of a correction to make.

The trouble is that it’s still not quite right. The location of the magnetic north and south poles wanders. Fortunately we’ve got a fair idea of how quickly it’s moving, and in what direction. So if you have a rough idea how out of date your chart is, and what direction the poles were moving in, you can make a correction. We can communicate how much the variance between true north and magnetic north vary.

The trouble is that it’s still not quite right. The size of the variation depends on the season of the year. But all right; we should have a rough idea what season it is. We can correct for that. The size of the variation also depends on what time of day it is. Compasses point farther east at around 8 am (sun time) than they do the rest of the day, and farther west around 1 pm. At least they did when Alan Gurney’s Compass: A Story of Exploration and Innovation was published. I would be unsurprised if that’s changed since the book came out a dozen years ago. Still. These are all, we might say, global concerns. They’s based on where you are and when you look at the compass. But they don’t depend on you, the specific observer.

The trouble is that it’s still not quite right yet. Almost as soon as compasses were used for navigation, on ships, mariners noticed the compass could vary. And not just because compasses were often badly designed and badly made. The ships themselves got in the way. The problem started with guns, the iron of which led compasses astray. When it was just the ship’s guns the problem could be coped with. Set the compass binnacle far from any source of iron, and the error should be small enough.

The trouble is when the time comes to make ships with iron. There are great benefits you get from cladding ships in iron, or making them of iron altogether. Losing the benefits of navigation, though … that’s a bit much.

There’s an obvious answer. Suppose you know the construction of the ship throws off compass bearings. Then measure what the compass reads, at some point when you know what it should read. Use that to correct your measurements when you aren’t sure. From the early 1800s mariners could use a method called “swinging the ship”, setting the ship at known angles and comparing what the compass read. It’s a bit of a chore. And you should arrange things you need to do so that it’s harder to make a careless mistake at them.

In the 1850s John Gray of Liverpool patented a binnacle — the little pillar that holds the compass — which used the other obvious but brilliant approach. If the iron which builds the ship sends the compass awry, why not put iron near the compass to put the compass back where it should be? This set up a contraption of a binnacle surrounded by adjustable, correcting magnets.

Enter finally William Thomson, who would become Baron Kelvin in 1892. In 1871 the magazine Good Words asked him to write an article about the marine compass. In 1874 he published his first essay on the subject. The second part appeared five years after that. I am not certain that this is directly related to the tiny slice of story I tell. I just mention it to reassure every academic who’s falling behind on their paper-writing, which is all of them.

But come the 1880s Thomson patented an improved binnacle. Thomson had the sort of talents normally associated only with the heroes of some lovable yet dopey space-opera of the 1930s. He was a talented scientist, competent in thermodynamics and electricity and magnetism and fluid flow. He was a skilled mathematician, as you’d need to be to keep up with all that and along the way prove the Stokes theorem. (This is one of those incredibly useful theorems that gives information about the interior of a volume using only integrals over the surface.) He was a magnificent engineer, with a particular skill at developing instruments that would brilliantly measure delicate matters. He’s famous for saving the trans-Atlantic telegraph cable project. He recognized that what was needed was not more voltage to drive signal through three thousand miles of dubiously made copper wire, but rather ways to pick up the feeble signals that could come across, and amplify them into usability. And also described the forces at work on a ship that is laying a long line of submarine cable. And he was a manufacturer, able to turn these designs into mass-produced products. This through collaborating with James White, of Glasgow, for over half a century. And a businessman, able to convince people and organizations to use the things. He’s an implausible protagonist; and yet, there he is.

Thomson’s revision for the binnacle made it simpler. A pair of spheres, flanking the compass, and adjustable. The Royal Museums Greenwich web site offers a picture of this sort of system. It’s not so shiny as others in the collection. But this angle shows how adjustable the system would be. It’s a design that shows brilliance behind it. What work you might have to do to use it is obvious. At least it’s obvious once you’re told the spheres are adjustable. To reduce a massive, lingering, challenging problem to something easy is one of the great accomplishments of any practical mathematician.

This was not all Thomson did in maritime work. He’d developed an analog computer which would calculate the tides. Wikipedia tells me that Thomson claimed a similar mechanism could solve arbitrary differential equations. I’d accept that claim, if he made it. Thomson also developed better tools for sounding depths. And developed compasses proper, not just the correcting tools for binnacles. A maritime compass is a great practical challenge. It has to be able to move freely, so that it can give a correct direction even as the ship changes direction. But it can’t move too freely, or it becomes useless in rolling seas. It has to offer great precision, or it loses its use in directing long journeys. It has to be quick to read, or it won’t be consulted. Thomson designed a compass that was, my readings indicate, a great fit for all these constraints. By the time of his death in 1907 Kelvin and White (the company had various names) had made something like ten thousand compasses and binnacles.

And this from a person attached to all sorts of statistical mechanics stuff and who’s important for designing electrical circuits and the like.