The Torch Magazine,
The Journal and Magazine of the
International Association of Torch Clubs
For 87 Years
A Peer-Reviewed
Quality Controlled
Publication
ISSN Print 0040-9440
ISSN Online 2330-9261
Winter 2014
Volume 87, Issue 2
The
Antikythera Mechanism
by
Karl R. Schaefer
Over the past century, several
researchers have tried to decipher the
structure and purpose of a mechanical
device built more than 2000 years ago.
This paper traces the growth of
understanding about the Antikythera
Mechanism and the role played by
various personalities in the pursuit
of that understanding.
Some 2000 years
after the eclipse of the classical
Greek Empire, reminders of Hellenic
artistic and architectural skills are
still evident. Remains of impressive
buildings are found in locations
across the Mediterranean; statues,
pottery, and jewelry may be seen in
any number of museums. From this
evidence, and surviving writings of
Greek philosophers, mathematicians,
and the like, we have sufficient
reason to characterize ancient Greeks
as clever and sophisticated. In terms
of technology, however, the evidence
is much slimmer. Greek inventions like
"Greek fire" and the "Claw of
Archimedes" are well known, but in the
mechanical arts, the Greeks are
generally considered to have been more
theoreticians than practitioners.
This perception
was upended early in the last century
when a Greek sponge diving boat sought
shelter from a storm near an island
called Antikythera, lying in the
straits between the Peloponnesian
Peninsula and the island of Crete.
After the storm, the divers decided to
try their luck in the bay. One man
went down in about sixty meters of
water and surfaced a short time later
bearing part of a bronze statue and an
account of a shipwreck containing
statuary, wine jars and other items.
The captain took the find to the Greek
authorities, who expressed interest in
the artifact and engaged the same
divers to return to the site. Over the
next couple of months, statues,
amphorae (tall, two-handled jars),
weapons, coins, and jewelry were
recovered. Among the finds was a
corroded clump of wood and bronze on
which toothed gears were visible.
The artifacts
were transported to the National
Archaeological Museum, where they were
photographed, catalogued, cleaned,
and, where possible, restored or
repaired. For several months, the
strange cluster of wood fragments lay
neglected in a corner. Priority was
given to the other objects. Meanwhile,
the device, now exposed to the
atmosphere, began to deteriorate. The
parts crumbled, breaking into four
main sections.
In 1902 and
1908, preliminary studies established
fairly conclusively that the shipwreck
had occurred about 80-70 BC. Although
the bronze gear fragments had not
initially elicited much attention, by
1907 several articles speculating on
the purpose and origin of the
mechanism had been written. The first
examiners of the enigmatic chunks
surmised that they had the remains of
an astrolabe or a navigational
instrument. The first attempts at
reconstructing the device date only
from 1928. But those results were
rudimentary.
Derek de Solla
Price, Avalon Professor of the History
of Science at Yale, closely followed
the scholarship about the device,
developing a deep interest in it. He
realized that the mechanism held great
significance for his studies of
ancient scientific instruments. In
1951, he secured funding which enabled
him to travel to Greece and undertake
a first-hand study of the Antikythera
Mechanism, as it was coming to be
known.
Using x-rays to
pierce the accretions covering much of
the four fragments, he identified some
thirty gears, two large ones and 28
smaller. He counted the number of
teeth on the larger gears and
estimated the number on others. From
this, Price determined that the device
was no mere astrolabe, but a much more
complex apparatus used to track
several astronomical phenomena. Most
importantly, the mechanism proved to
be a device for displaying the monthly
phases of the moon according to two
criteria: the sidereal month, one
complete lunar revolution around the
Earth, and the synodic (lunar) month,
the phases of the Moon as it appears
in the night sky (1).
Given the number of
gears and their axes, the mechanism
appeared capable of displaying both
lunar and (apparent) solar movements.
Price was also
able to read fragments of text from
the front and rear panels of what had
been the box that contained the
gearing. One of these fragments
related to a dial mounted on the back
of the machine. Preserved sections of
dials and geared rings bore
fragmentary words indicating that the
machine could calculate astronomical
events in relation to the Egyptian
solar year, the most accurate calendar
in ancient world. Each month was
thirty days in length; an epagomenal
(intercalary) five days was added to
the end of the year to keep it
synchronized with the earth's annual
orbit about the sun. A second scale,
inscribed on a ring mounted inside the
one bearing the months, indicated the
zodiacal signs in which the sun
appeared.
Price was
convinced that by counting the teeth
on the various gears, their
configuration, and their interactions,
he could solve the riddle of the
device. Besides displaying the solar
year and the sidereal and synodic
cycles of the lunar year, he
theorized, the gearing also calculated
the so-called Metonic cycle of
astronomical time (equal to nineteen
solar years, 235 lunar cycles
(lunations), and 254 sidereal
revolutions of the moon). From this
evidence, De Solla Price constructed a
model of the device that now resides
in the Archaeological Museum in Athens
(2).
Price died in
1983 without completely, or completely
accurately, unraveling all the secrets
of the Antikythera Mechanism. He
thought it had probably been made on
the island of Rhodes by someone who
knew the theory of planetary motion
attributed to Hipparchus, a Greek
astronomer (second century BC). Price
believed that the device was meant to
display even more elaborate celestial
events than its remains indicated, but
he was unable to prove it.
Meanwhile,
Price's work came to the notice of a
curator in the Science Museum of
London. Michael Wright's area of
responsibility there was engineering;
he combined an interest in how things
worked with an understanding of how
mechanical principles were conveyed
from the construction of one mechanism
to another. He realized that the
Antikythera Mechanism constituted an
early stage in an engineering
tradition that might be traced back
even further into history. After
reading everything he could about the
mechanism, he realized that there was
more to it than Price had uncovered
and became convinced that several of
Price's theories about the machine
were mistaken. It became clear to him
that Price had, in some instances,
made arbitrary determinations about
the number of teeth on a gear, for
example, in order to fit his
conception of how the device worked.
Wright also concluded that Price's
reconstruction of the rear face of the
mechanism was too simplistic, given
the complex gear train found in the
remains. He embarked on a life-long
endeavor to reconstruct the device in
its original form.
Wright's
progress was slow, since his museum
was unwilling to allow him to travel
to Athens to study the device
first-hand. Matters improved
significantly when he made the
acquaintance of Allan G. Bromley, an
astrophysicist from Australia. The two
men shared a deep love of mechanical
devices, which led to Wright
introducing Bromley to the Antikythera
Mechanism.
Bromley's
fascination with the artifact prompted
him to secure permission from the
museum in Athens to work on it.
Wright, believing that Bromley's
position as an academic had enabled
him to gain access to the device, felt
betrayed. Despite his bitterness over
this development, he asked Bromley to
take him as his assistant. Bromley
agreed. Over one month, in 1989, they
worked on the device every day. It
became clear to Wright that Price had
missed a great deal.
Upon Wright's
return home, he met a doctor with an
interest in horology. The physician
suggested that linear tomography (a
radiographic technique for
representing cross sections of solid
objects) might be a useful tool for
revealing more details of the interior
of the device. Wright built a crude
tomography machine in his basement and
returned with it to Athens; the result
was 700 images that revealed parts of
the mechanism that had lain hidden for
two millennia. However, Wright's
delight was cruelly deflated by
Bromley's announcement that he was
taking all the images back with him to
Sydney, where a student was prepared
to scan them onto a computer. Wright
was returning to England empty-handed.
The next several years of Wright's
life were fraught with personal and
professional difficulties, but he
thought about the mechanism daily.
In 2003,
Bromley was dying of cancer in
Australia, and he wanted to see Wright
before he died. Wright traveled to
Australia. Bromley eventually allowed
Wright to take many of the images back
to England. Armed with them, Wright
was able to add valuable information
to what was known about the design of
the device. Most important was the
discovery that it employed epicyclic
gearing to replicate the perceived
irregularities in planetary movements.
Price's claim that a differential gear
had been used was wrong. Instead,
Wright proposed, a pin and slot
arrangement allowed planetary motions
to be accurately represented. This
theoretical reconstruction also helped
to explain why such a large gear was
needed. Not only were the movements of
the sun and moon replicated by this
piece, but the movements of two (or
more) of the five known planets as
well. Such reasoning explained the
presence of the large four-spoke gear
that, he argued, was unnecessary to
portray the movements of the sun and
moon alone.
His hypothesis
was supported, in part, by the remains
of Greek inscriptions that mentioned
the planets by name and their periodic
"stationary points" in the heavens.
Details revealed in the x-ray images
suggested this kind of arrangement,
particularly a square center mount to
prevent the wheel from slipping
against the resistance generated by
the meshing of the other gears.
By the end of
2001, Wright had assembled a working
model of the mechanism based on his
observations of the remains of the
gearing and adhering to 2nd century BC
Greek astronomical theory, which
posited, among other things, that the
earth stood at the center of the
universe, the sun and all the other
planets orbiting about it.
Shortly before
he finished his model, Wright was
contacted by Michael Edmunds, an
astronomer from Cardiff University,
and Tony Freeth, a mathematics PhD and
documentary filmmaker. Edmunds and
Freeth were interested in examining
the Antikythera Mechanism, too. Wright
feared that, as with Bromley, his own
work would be overshadowed. Wright had
put in over 500 hours of work on his
model, composed of both bronze and
wood elements, like the original. His
anxiety increased when, after a phone
conversation with Edmunds—during which
he shared what he knew and suspected
about the device—Edmunds and Freeth
published an article in which Wright's
contributions were given short shrift.
Wright,
aided by his son, a PhD candidate in
medical imaging, had been able to
transfer all of Wright's and Bromley's
x-ray images onto a computer and was
thus able to manipulate the images,
achieving a much greater level of
accuracy than he had been able to do
before. He could now measure the rings
on the back of the mechanism,
discovering that they formed two
spirals, rather than concentric rings.
One displayed the Metonic cycle of 19
years/235 sidereal months. The second,
smaller dial represented three
divisions of the so-called Callipic
cycle of 76 years (4 x 19) that
closely approximated the actual solar
year of 365 ¼ days.
Mechanically,
the device was revealing itself to be
very complex. Wright discovered not
only the functions of various gears
that de Solla Price and others had
mistaken or overlooked, but also the
method by which the mechanism was able
to represent the perceived variations
in the speeds of the sun and planets,
and the so-called "retrograde motions"
by which the planets periodically
appeared to be moving backward in
relation to the background stars. The
pin and slot arrangement on certain
wheels allowed the planets to move at
varying speeds, replicating that
apparent motion on the face of the
main dial.
Wright's
presentation of his model at a
conference in Athens in 2005 was
overshadowed by the announcement that
a fifth piece of the artifact had
recently been discovered in a museum
storage area. It contained a piece of
the second dial that had been mounted
on the rear face of the mechanism.
Freeth and
Edmund's investigation had the
assistance of two brand new
technologies: computed tomography
(CT), which provided a
three-dimensional image of the
interior of the artifact, and
light-mapping, a computer graphic
innovation that reflects light off a
surface from multiple angles in
sequence and allows a composite image
to be made from the images that
result. Freeth was insistent that
Thomas Malzbender, the developer of
the imaging process, be brought along
on his team. X-Tek, an x-ray company,
shipped an experimental CT machine to
the museum and, after much
bureaucratic wrangling, permission was
given to carry out examinations of the
device.
The results
exceeded everyone's expectations and
the data, comprised of 3000 digital
images, were loaded onto a computer.
Details of the mechanism were now
clearer; many lines of Greek text
could be seen and read. To date some
two thousand characters (out of an
estimated original 20,000 character
text) have been deciphered, providing
crucial clues to the functions of
specific elements in the mechanism.
Wright again
chafed under the realization that
Freeth and his team were to receive
the lion's share of the credit for
having decoded the Antikythera
Mechanism, but he nonetheless attended
a conference in Athens in 2006 at
which Freeth's findings were made
public. Wright displayed his working
model, this time including
improvements and corrections that
brought the mechanism much closer to
the "ideal" reconstruction that Freeth
was proposing.
About 2005, the
Hellenic Ministry of Culture
established the Antikythera Mechanism
Research Project, a joint effort
involving the University of Cardiff,
the National and Kapodistrian
University of Athens, the Aristotle
University of Thessaloniki, the
National Archaeological Museum of
Athens, X-Tek Systems UK, and
Hewlett-Packard USA. A more
cooperative approach to the study of
the device was now in force.
Information is shared electronically
and can be studied in detail remotely.
The object itself is very fragile and
is never exhibited outside the museum.
The story does
not necessarily end here. We may never
be certain that what we have concluded
about the machine, based on what
remained of it after 2000 years, is
100% accurate. It would help, of
course, if we were to find an example
of Greek gear works of any sort from
that time. Is it possible that another
ship carrying a similar device also
found its way to the bottom? Unlikely.
However, it is likely that if other
such devices did exist, they were
probably melted down at some point and
turned into other objects. Such was
the fate of innumerable bronze statues
and other metal artifacts from the
classical period.
What is clear is
that the Antikythera Mechanism was
made to portray a number of celestial
events relating to the solar year. The
front dial showed the apparent motion
of the sun across the sky in relation
to the Egyptian solar year. Remains of
inscriptions indicate that Greek names
for the Egyptian months were used on
the calendar wheel. The front dial
also showed the zodiacal signs through
which the sun passed as the seasons
changed. This was achieved by
offsetting the center of the zodiacal
cycle by a mathematical adjustment so
that the zodiac could be made
congruent with the sun's apparent
motion through the sky. Monthly moon
phases were represented by a small
sphere moved by a gear mounted at a
right angle to the face of the device.
Pointers attached to the gears showed
the location of the sun and the phase
of the moon in relation to the
calendar. Irregularities in the
orbital motions of the sun and the
moon were also incorporated. Wright
and others contend that the motions of
the five known planets—Mercury, Mars,
Venus, Jupiter and Saturn—were also
illustrated by pointers connected to
the gears corresponding to their
orbits. However, with one exception,
the gear train that would have allowed
planetary displays is either missing
or never existed. Wright's latest
model incorporates this feature,
though.
On the rear
face, multiple displays are found. Two
large dials allow the tracking of more
important phenomena. One pointer on a
spiral dial shows the Metonic cycle,
the congruence of 19 solar years and
235 lunar months, about 6940 days. The
spiral is marked with the month names
of a local civic calendar. This
allowed investigators to deduce that
there was a connection to the city of
Corinth, on the isthmus between the
Peloponnesus and Attica. The
second large dial was designed to
predict solar and lunar eclipses
according to what is known as a Saros
Cycle. This is a period of
approximately eighteen years, after
which period a nearly identical
eclipse will occur. A smaller dial
here displays three Saros cycles (= 54
years), the so-called exilegmos, which
allows for a more precise calculation
of nearly identical eclipses. Two
other dials are also mounted here, one
showing the cycle of four panhellenic
games, the Olympia, Isthmia, Nemea and
Pythia, which took place over four
years. The device thus allowed one to
determine when specific games were to
be held. The second shows a
calculation of four Metonic Cycles (76
years) (2).
It is
no doubt ultimately futile to attempt
to recreate the mechanism as
originally constructed and as
originally intended to function. Too
much of it has been destroyed.
Moreover, attempts at reconstruction
require modern investigators to think
like antique Greeks—as daunting a task
as ever existed.
The Antikythera
Mechanism is the result of long term,
painstaking observations of the
heavens, a tradition of astronomical
recordkeeping dating back to the
Babylonians, and a highly developed
set of skills in applying mathematics
to engineering problems. This analog
computer, the oldest one known to
exist, is not an anomaly and was
probably not unique. We may never know
the name of the engineers who built
the mechanism and can only hypothesize
about what its real capabilities were.
What it does tell us though, is that
the ancient Greeks were capable of
much more than we give them credit
for.
Footnotes
(1) The
sidereal month is the time the Moon
takes to complete one full revolution
around the Earth with respect to the
background stars. However, because the
Earth is constantly moving along its
orbit about the Sun, the Moon must
travel slightly more than 360° to get
from one new moon to the next. Thus,
the synodic month, or lunar month, is
longer than the sidereal month. A
sidereal month lasts 27.322 days,
while a synodic month lasts 29.531
days. Return to Text
(2) Reconstructed
dimensions: approx. 32 x 16 x 9 cm.
(according to De Solla Price) Return to Text
Antikythera
Mechanism Readings
Edmunds,
M.G. "An Initial Assessment of the
Accuracy of the Gear Trains in the
Antikythera Mechanism," Journal
for the History of Astronomy 42:3
(August 2011), 307-320.
Freeth,
Tony, "Decoding an Ancient Computer,"
Scientific American 301:6 (Dec.
2009), 76-83.
Freeth,
Tony, Alexander Jones, John M. Steele,
Yanis Bitsakis, "Calendars with
Olympiad Display and Eclipse
Prediction on the Antikythera
Mechanism," Nature 454:7204
(July 31, 2008), 614-617.
Marchant,
Jo. Decoding the Heavens.
Philadelphia: Da Capo Press, 2009.
Price,
Derek de Solla, "Gears from the
Greeks," Transactions of the
American Philosophical Society (New
Series) 64:7 (1974), 1-70.
Fragment of the
Antikythera Mechanism in the
Archaeological Museum, Athens.
Photo by Giovanni Dall'Orto, via
WIkiCommons.
Biographical Note
Karl R. Schaefer is a native of the
Hudson Valley in New York State. He
received a BA in Political Science
from Union College in Schenectady,
New York (1971), and a Masters and
PhD in Near Eastern History from New
York University (1985). He also
holds a Library Science degree from
the University of Oklahoma (1992).
Since 1995, he has been a reference
and instructional librarian at Drake
University in Des Moines, Iowa. He
has published several articles and a
book on medieval Arabic block
printing. He has been a member of
the Des Moines, Iowa Torch Club
since 2005.
