By Douglas McElwain
The Antikythera Mechanism was discovered in a shipwreck off the Greek island of Antikythera (just south of the Peloponnesus) between 1900-02. Over a century of study by researchers has determined that it is the remains of a two thousand year old astronomical computer. It is believed that the analog device was about the size of a shoebox and used thirty-seven finely crafted bronze gears. Even after a century of study it is not known who designed it or exactly when – estimates range from 205 BCE to a little before 100 BCE – or where it was built. In this brief article is a proposal about how we might determine its place of origin.
At first researchers believed that the mechanism was a type of astrolabe. Further study revealed that it was a complex mechanical device used to calculate the positions of the sun, moon and five planets. The mechanism also calculated the phases of the moon and the timing of both lunar and solar eclipses. There was also a dial displaying the four-year Greek Olympiad Cycle.
An astrolabe was different. “The primary purpose of the astrolabe is to show the user the positions of the Sun and selected stars at a specific place at a given local time. … Most astrolabe problems were solved using the front of the instrument. The fixed parts represent time scales and a view of the sky seen from a specific latitude. The rotating parts simulate the daily rotation of the sky. To use an astrolabe, you adjust the moveable components to a specific time and date. Once set, much of the sky, both visible and invisible, is represented on the face of the instrument. This allows a great many astronomical problems to be solved in a visible way. … The astrolabe is intended to be used for both observation and computation. For observation it is fitted with a ring so the instrument can be hung vertically while the position of the Sun or star is measured using movable sights and a scale on the back of the instrument.” 
Both the Antikythera Mechanism and astrolabes are geocentric devices. That is, the devices are designed with the assumption that the Earth is at the center of the solar system. As such, the devices are only accurate near the latitude where they were designed. A thought experiment illustrates this. Move either of these devices from the Mediterranean to a position along the same longitude line but to a latitude of 70° N or greater. Both devices would indicate that the sun should be visible. However, the sun isn’t visible at these latitudes part of the year. Therefore, the information the devices provided would be incorrect. This also implies that the errors increase the further one moves from the latitude of design.
The astrolabe overcomes this limitation by using a series of plates called tympans. Tympans display a stereographic view of the sky for different latitudes. The plates used in most configurations used a 2° almucantar (altitude arc) resolution.  As astrolabe users moved north or south they would swap in a tympan for the next latitude. Therefore, tympans were used to make astrolabes useful across a wide range of northern latitudes. While astrolabes were used to represent the sky in the present, the Antikythera Mechanism could display the aforementioned astronomical objects in the past, present or future. In the reconstruction pictured below this was accomplished by turning the knob on the side of the mechanism.
There are several theories as to how the Antikythera Mechanism was used. However, there is no consensus as to what that was. It was probably not used for shipboard navigation. First, as indicated above, it was small. The smaller the mechanism, the less accurate the sightings taken with it would be (i.e., angles can be read with more precision on larger devices). Astrolabes used for navigation were more accurate the larger they were made. Second, the Antikythera Mechanism would be difficult to use on a rocking ship. Lastly, the Antikythera Mechanism doesn’t have any way to measure the altitude of astronomical objects. Therefore, for it to be useful for shipboard navigation, the Antikythera Mechanism would have needed to be used in conjunction with other instruments.
The Antikythera Mechanism could have been useful on land. Astrologers, priests, astronomers or just a wealthy individual might find it useful. Astrologers could use it to draw astrological charts for customers. Priests could appear prescient and prepare their followers for unusual astronomical events such as eclipses. Astronomers could prepare for observations and wealthy individuals could impress their friends.
Identifying Its Place of Origin?
The mechanism itself might be used to help identify its own origin. I make the assumption that the latitude where the astronomical observations were taken and used to create the device is the same latitude where the mechanism was designed and built. This is the latitude of origin.
In order to accurately project the lunar information over long periods (such as the Callippic cycle) the mechanism must have had a high degree of precision. Earlier, I suggested that the Antikythera Mechanism produced an error in its readings as it was moved away from the latitude where it was designed. These errors increase in size the further the mechanism is from this latitude of origin. In turn, by mapping these errors, the latitude of origin can be identified.
Conceptually, the process would look like this. Choose an astronomical object for which to take measurements; possibly the timing of a lunar or solar eclipse. Next choose a longitude in the Mediterranean Sea, say halfway between Syracuse and Corinth. Take a reconstructed model of the Antikythera Mechanism and move north along this longitude line. Every 10 km take timing readings from the model until reaching 67° N. At the same time take similar time readings from a modern astronomical source. Compare the two readings for each location. Repeat this process moving south until reaching 7° N. Map the absolute value of the difference between the two readings at each point along the chosen longitudinal line. Next, fit these error points to a curve. The latitude that runs through the minimum point on the curve should be the latitude of origin for the Antikythera Mechanism.
It might be argued that the Antikythera Mechanism is not precise enough for the suggested methodology to be effective. However, any error arising from the Antikythera Mechanism will be applied equally to all the error data making this a moot point. If the timing of the eclipse cannot be determined to any level of precision, the precision of the physical models can be increased in two ways. Neither alters the Antikythera Mechanism’s functionally. First, make a larger model of the mechanism; possibly three or four times as big. A larger model would have dials that are bigger and allow more accurate readings. Second, extend the pointer for the astronomical object being measured to a new, larger and more finely graduated dial constructed further from the center of the device. Alternatively, depending on the accuracy of the digital Antikythera Mechanism models, it might be possible to pull the data directly from them. We do not yet know the itinerary of the shipwreck but some possibilities are included below.
Practically, the data for the time of the eclipse (for every tenth of a degree of latitude over the 60° range suggested above) can be acquired from a modern astronomical database. This information can be then compared to a single reading from a physical Antikythera Mechanism model.
From various sources, the most likely design and production locations are (in alphabetical order):
Location Latitude Reason
Alexandria, Egypt 31.2000° N Library of Alexandria. Possible ship port .*
Babylon, Mesopotamia 32.5364° N Unlikely yet probable locus of prior sky maps.
Corinth, Greece 37.9333° N Doric Greek writing on mechanism.*
Rhodes, Greece  36.1667° N Hipparchus’ home. Possible ship port .*
Siracusa, Sicily  37.0833° N Archimedes’ home. Doric Greek writing.*
* Note the language on the Antikythera Mechanism is Doric Greek, found in places like Corinth and its Sicilian colony where Archimedes lived. A bit later, about the likeliest time of the shipwreck between the late third century BCE and the late 2nd century BCE, Hipparchus in Rhodes seems to also know of such possible astronomical devices.
If the minimum latitude error line (i.e., latitude of origin) crosses any of the locations from the table above, it would be indicative that the Antikythera Mechanism was designed and built there. As can be seen, there are roughly two major groupings of latitude. Because of the closeness of some of the potential latitudes it might be difficult to distinguish between sources.
In conclusion, it appears there is a way to identify the origin of the Antikythera Mechanism. Doing so will allow researchers to focus future investigations of the mechanism in more geographically relevant locations.
Jo Marchant. 2010. Decoding the Heavens: A 2,000-Year-Old Computer – And the Century-Long Search to Discover Its Secrets. Cambridge, MA: Da Capo Press.
James E. Morrison. 2007. The Astrolabe. Tallmadge, OH: Good Place Publishing.
Rob Rice. “The Antikythera Mechanism: Physical and Intellectual Salvage from the 1st Century BC” 11th USNA Naval History Symposium 1985. (http://ccat.sas.upenn.edu/rrice/usna_pap.html)
John Seabrook. “Fragmentary Knowledge: Was the Antikythera Mechanism the World’s First Computer?” New Yorker, May 14, 2007, Department of Archaeology (http://www.newyorker.com/magazine/2007/05/14/fragmentary-knowledge)
Ker Than.“Scientists Unravel Mystery of Ancient Machine” Live Science. November 29, 2006.
The Antikythera Mechanism Research Project, 2006-14
 Morrison, 2
 Morrison, 58
 Marchant, 265
 Marchant, 288
 Marchant, 77-80
 Marchant, 288