Phys90: Introduction to Astronomy
Ethno & Archeo Astronomy:
The Antikythera Mechanism

In 1900 a group of sponge divers blown off course in the Mediterranean discovered an Ancient Greek shipwreck dating from around 70 BC.

Lying unnoticed for months amongst their hard-won haul was what appeared to be a formless lump of corroded rock. It turned out to be the most stunning scientific artefact we have from antiquity. For more than a century this 'Antikythera mechanism' puzzled academics. It was ancient clockwork, unmatched in complexity for 1000 years - but who could have made it, and what was it for? Now, more than 2000 years after the device was lost at sea, scientists have pieced together its intricate workings and revealed its secrets.
In order to understand this ancient computer we unearths a diverse cast of remarkable characters -- ranging from Archimedes to Jacques Cousteau -- and explore the deep roots of modern technology not only in ancient Greece but in the Islamic world and medieval Europe too. At heart an epic adventure story, this quest challenges our assumptions about technology transfer over the ages while giving us fresh insights into history itself.


The history of the Mechanism

The Antikythera Mechanism is the name given to an astronomical calculating device, measuring about 32 by 16 by 10 cm, which was discovered in 1900 in a sunken ship just off the coast of Antikythera, an island between Crete and the Greek mainland. Several kinds of evidence point incontrovertibly to around 80 B.C. for the date of the shipwreck. The device, made of bronze gears fitted in a wooden case, was crushed in the wreck, and parts of the faces were lost, "the rest then being coated with a hard calcareous deposit at the same time as the metal corroded away to a thin core coated with hard metallic salts preserving much of the former shape of the bronze" during the almost 2000 years it lay submerged. The quotation is from Derek de Solla Price's monograph Gears from the Greeks ... in the 1974 Transactions of the American Philosophical Society (Volume 64, part 7).

General plan of all gearing, composite diagram
from De Solla Price, Transactions of the American Philosophical
Society Vol 64 No 7 (1974). Reproduced with permission.
It is hard to exaggerate the singularity of this device, or its importance in forcing a complete re-evaluation of what had been believed about technology in the ancient world. For this box contained some 32 gears, assembled into a mechanism that accurately reproduced the motion of the sun and the moon against the background of fixed stars, with a differential giving their relative position and hence the phases of the moon. It is enough to know that there is no trace of anything like it until around 1000 A.D., and that when it was first published there were serious suggestions that it had been dropped into the wreck at a much later date or even that it was the work of alien astronauts.

The general plan of the gearing could only be drawn after 1971 when, on De Solla Price's instigation, the remnants of the mechanism were viewed with gamma rays, which could penetrate the calcareous block in which the gears were embedded. Then careful counting of teeth, and examination of the way the gears meshed, showed that "the gear ratios could be associated with well-known astronomical and calendrical parameters" (Price) and allowed the almost complete description of how the device must have functioned.

The Sun-Moon Assembly

Images by Bill Casselman.

These two images give a schematic version of De Solla Price's general gear plan; on the right the color picks out the sub-assembly connecting the sun's motion to that of the moon.

The sun marker and the moon marker were driven by the two central gears (the moon axis threaded through the sun's), exactly like the hour and minute hands on a modern clock. The train of gears linking the sun's motion to that of the moon can be described by the meshing pattern and the numbers of teeth.

The sun gear has 64 teeth. It meshes with the smaller of a 38,48 gear pair. The 48 meshes with the smaller of a 24,127 gear pair. The 127 meshes with the 32 teeth of the moon gear. The ratio of angular speeds can then be calculated as
  64   48   127   254
  -- X -- X --- = --- = 13.36842..
  38   24    32    19
which is an excellent approximation of the astronomical ratio 13.368267.. .

Since the sun-moon linkage involves an odd number of meshings, the two gears will turn in opposite directions. For the display to be realistic the sun and the moon must move the same way. In the device this was accomplished by a vertical "contrate" gear (marked A in De Solla Price's plan) linking the sun gear to an identical gear above it, which thus turned at the same speed but in the opposite direction.

This image can be java-animated. It shows the operation of the Sun-Moon assembly, with a somewhat fanciful simulation of the display. In the actual device, the zodiac constellations were represented by their Greek names ("Libra" and the end of "Virgo" are decipherable in the relic). The constellation-schemata used here are imitations of the more accurate versions in Find the Constellations by H. A. Rey, Houghton-Mifflin Co., Boston, 1988.

Java animation: GO
(by Bill Casselman, University of British Columbia).

Return to Main Menu