In early 1900, a sponge diver diving off the coast of Antikythera – a small Greek island between Crete and the Peloponnese – discovered the remains of an ancient, wrecked cargo ship. Dated to between 200 and 100 BCE, amongst the ship’s surviving contents of bronze and marble sculptures was a curious piece of rock with an embedded gear wheel.
Intriguing archaeologists, the rock was examined and found to be a heavily corroded and encrusted devise consisted of numerous gears which had been broken when the cargo ship sank. Christened the Antikythera Mechanism, it was the oldest geared device ever found and has been the focus of study and speculation for over a century.
There are certainly allusions to similar devices existing in the Classical Greek world, but until the discovery of the Antikythera Mechanism these were thought to be simply conjecture and rumours. The Roman philosopher and writer Cicero in particular mentions that a device “of Archimides deserved special attention because he had thought out a way to represent accurately by a single device for turning the globe those various and divergent movement with their different rates of speed”, along with two enormously complex “planetariums” brought back to Rome as booty from war.
The German philologist Albert Rehm was the first scholar to claim that the device was an astronomical calculator; an extraordinary claim as mechanisms of a similar complexity were not seen again for another 1000 years. From the 1930s, the Antikythera Mechanism was the subject of a number of reconstructions to attempt to find out exactly how it worked and what information it could impart.
In the 1950s, the English scientist and historian Derek de Solla Price researched and built the first substantial model attempting to demonstrate the intricacies of how the mechanism worked. Having established it was the first known calculator and a precursor to mechanical clocks, his model used all known gear trains but added two extra gears that he believed were missing from the original wreck. However there were problems with Price’s model; not only could it be easily simplified by removing some of the gears, but some of the tooth counts were rounded and therefore inaccurate to the original.
Attempting to improve on Price’s model, Associate Professor Allan Bromley – an Australian computer historian – continued to research the Antikythera Mechanism and proposed an alternate model. Working with Michael Wright, Bromley used linear X-ray tomography to create 2D slices of the gears still encased in rock. These images showed that Price’s reconstruction was inherently flawed. With the assistance of Sydney clockmaker Frank Percival, Bromley initially trialled his new construction in Meccano before building a complete model in 1989.
Crucial to Bromley’s new model was the incorporation of the Saros cycle, which predicts eclipses of the Sun and Moon. He also changed the use of the crank which powers the whole device making one rotation equal to exactly a day, the simplest astronomical division of time. Bromley’s model was not without its faults. While it correlated best with the new X-ray evidence it was difficult to predict the future or past astronomical events, which would have been a large part of its purpose.
Bromley’s model, which was acquired by MAAS in early 2017, is representative of the understanding of the Antikythera Mechanism at a particular moment in time, and prior to his untimely death in 2002, Bromley had become the world authority on the device.
Akin to an 18th century clock in its complexity, the Antikythera Mechanism represents an incredible understanding of astronomy and engineering at a time only 600 years after the end of the Greek ‘Dark Ages’. Similar technology would not be known again for a millennium and the tragedy of its initial loss at sea has meant modern archaeologists and historians have had a glimpse in an ancient and long forgotten knowledge. Bromley’s model in particular demonstrates the use of new technologies in re-examining old discoveries, and the wider contribution of Allan Bromley in understanding the history of early computing.
Written by Lauren Poole, MAAS Intern 2017.