Books (Science): Eclipse

I think I've commented before on the recent trend of popular science books that focus on one very narrow aspect of science (or science history), such as the quest for an accurate way to measure longitude, or the development of the magnetic compass, or the science of shadows. Duncan Steel's Eclipse: The Celestial Phenomenon Which Has Changed The Course of History is the latest one of these that I've picked up. After finishing Use of Weapons I wanted another non-fiction book, and Paul Davies' name on the Forward caught my eye. I picked up the paperback second edition, in which Steel updated text which had referred to the 2000 solar eclipse in the future to referring to it in the past.

The book isn't quite as narrowly focused as the title suggests: Steel actually covers not only solar and lunar eclipses, but occultations (eclipses of stars or planets by the Moon or nearer planets), transits (passages of Mercury, Venus, or various astroids in front of the sun), and other, more exotic sorts (Pluto and Charon eclipsing each other, binary star eclipses, etc).

Steel's main interest is on how the study of ecilpses has us knowledge far beyond just the ability to predict eclipses. The set of circumstances necessary for a solar or lunar eclipse to happen depends on the relative lengths of months and years as measured in several ways, and by studying eclipses to determine these lengths, people were able to discover the length of the equinox-to-equinox year very precisely -- precisely enough to know that a leap year every four years (as in the Julian calendar) is one too many every four hundred years, a problem that the current Gregorian calendar fixes.

The book also covers such eclipse-related science as the discovery of helium in the solar corona (only observable during a solar eclipse), the discovery of the Martian atmosphere and Uranian rings during occultations of stars, and of course Sir Arthur Eddington's famous expedition to photograph star displacement during the 1919 eclipse as evidence in support of General Relativity.

My enjoyment of the book was very slightly marred by having read Tony Rothman's Everything's Relative a few months back, though. Rothman's book takes on many traditional stories in science -- such as Eddington's expedition -- and demonstrates how the standard tellings are often very misleading, simplifying matters to the point of distortion and making science look much tidier than it really is.

In the case of the Eddington eclipse, the traditional story tells how Eddington photographed stars around the during the 1919 solar eclipse and measured their displacement against their normal positions, finding it to be 1.75 seconds of arc -- exactly as Einstein had predicted. In actual fact things weren't quite that clean: Eddington only got one or two usable photographs, not enough to be sure of the value. Newton's theory of gravity predicted a deflection as well, of about 0.8 seconds of arc, and Eddington's photographs were not good enough to exclude the possibility that the deflection was closer to Newton's value than Einsteins.

In Eclipse, Steel uses the traditional, simplified versions of a number of stories, and Rothman's book kept intruding in my mind, pointing out that it didn't really happen that way. Nevertheless, the book is fun and engaging, and pretty informative.

One odd thing that bothered me about the paperback edition was in the typesetting -- not something you usually bring up in a discussion of books, but ever since reading Elements of Typographic Style by Robert Bringhurst I've paid a lot more attention to such things.

The book is set in a slightly antique-feeling font (I don't know the name), which is very nice to read, but the numbers were a problem. First of all, the fact that I noticed antyhing odd about them means that they weren't blending into the rest of the font properly!

The first problem was that the numbers were fixed-width -- the "1" took as much horizontal space as the "8", which means that all the "1"s looked like they were spaced away from the other digits. Fixed-width numbers are essential when typesetting columns of numbers, but there aren't any in the book. The "1"s also looked very similar to the "I"s -- instead of a single beak on the top left, they had serifs on each side, with only a slight lower-left to upper-right slope distinguishing them from the "I"s.

Lastly, the antique feel of the book font really suggests the use of lower-case numerals rather than full-height ones. This weblog uses lower-case numerals, which have ascenders and descenders and generally have the same visual weight as lower-case letters: 1234567890. When you have a page full of dates and measurements in the text, full-height numerals can be very distracting, like BLOCK CAPS scattered through the text; lower-case numerals help keep the overall colour of the page even.

Of cousre, the typesetting probably has very little to do with Steel, and doesn't matter to the quality of the book! It gave me another nice few days spent reading interesting stuff about the universe.