Mike Carrell wrote:
Both Jed and Steve Lawrence responded to my comments about Chinese
navigation and longitude measurement. Rather than annotate each post, I will
comment on both, based on Menzies' book.
Longitude is based on measurement of time from some reference point.
Time can be measured by observing the location of the moon against the
background of the fixed stars. Similarly, with accurate tables of lunar
eclipses, the time is known precisely at the moment of the eclipse. As
far as I know, all methods of "lunar navigation" are equivalent to
determining the time from an observation of the moon relative to other
bodies, and then determining longitude by (simultaneously) observing the
location of either the moon or another celestial body relative to the
horizon.
The details get messy, of course, as with all celestial navigation, and
the fact that somewhere in there the celestial time is being determined
and then used may be hidden under the mass of detail.
In the
contest between Harrison and the British astronomers which I have read
about, reference was made to a complex mathematical method using lunar
observation. I don't know details.
Menzies notes, and many agree, that the Chinese were master astronomers who
knew the sky very, very well. They made precision 'sundials' and water
clocks which were calibrated with the sundials. At a landfall on a distant
shore, an observatory could be built which would accurately define the
zenith and the local meridian. An eclipse of the moon gave precise moments
which could be related to the position against the fixed stars. This event
was simultaneously observable from the Chinese reference observatory. When
sailors returned it was possible to accurately determine the angular
relationship on the surface of the earth, and thereby the longitude.
The Polynesians also navigated across long stretches of ocean, based on the
night sky and the rhythm of the waves as refracted from distant islands, and
other factors.
Lunar eclipses aren't often convenient. The transit of Jupiter by one of its
moons is a much more frequent event, simultaneously observable at two
points, and the relative longitude established back home.
And the location of the moon versus the fixed stars is something that's
always available, and as noted above it also provides a celestial
"clock". The drawback, as noted in previous posts, is that it's hard to
get a _precise_ time by observing the moon against the stars: it moves
too slowly relative to the fixed stars. In consequence, lunar
navigation is inherently less accurate than clock-based solar
navigation, assuming the same instruments are used to measure the angles
in both situations. I don't recall off hand exactly what the typical
precision of a lunar fix was, but what I do recall is that even using a
modern sextant, it's a whole lot worse than a clock-based solar fix.
Occultations of particular stars by the moon could provide better
precision than measuring angles against stars at positive separation,
but then you're dependent on the precision of the tables of
occultations; that's also likely to introduce substantial errors.
Lunar eclipses can be used as you describe without the use of
precomputed tables. For navigation at sea you'd need the tables, of course.
