For 125 years, the small bronze object catalogued as Fragment D has sat in the National Archaeological Museum in Athens, considered an incidental piece of the Antikythera Mechanism, the famous Hellenistic calculator pulled from a wreck off the Greek island of Antikythera in 1901.

It was always a little odd. It did not fit, mechanically, with the main fragments. Its gear pitch was wrong. Its dimensions were slightly off.

A team led by physicist Eleni Markou at the University of Glasgow has now used high-resolution neutron tomography, performed at the ISIS facility in Oxfordshire, to image the interior of Fragment D in a way that earlier X-ray studies could not.

The result, published in the Journal of Roman Archaeology on April 28, is that Fragment D is not a fragment of the main mechanism at all. It is a separate, simpler instrument, possibly a portable parapegma, used to track the rising and setting of fixed stars.

The neutron scans revealed two complete gear trains, one with 19 teeth and one with 47, arranged in a configuration that has no analog in the main Antikythera Mechanism.

Markou's team proposes that Fragment D was a companion instrument, likely owned by the same Hellenistic patron, that performed a subset of the main mechanism's calculations in a more portable form.

The 19-tooth ratio strongly suggests a Metonic cycle calculation, used to align the lunar and solar calendars over a period of 19 years. The 47-tooth ratio is harder to interpret but appears to track a stellar phenomenon.

Markou's team proposes that Fragment D was a companion instrument, likely owned by the same Hellenistic patron, that performed a subset of the main mechanism's calculations in a more portable form.

If the interpretation holds, it doubles, at a stroke, the number of known ancient Greek geared astronomical instruments.

The wreck off Antikythera, dated to roughly 60 BCE, was a Roman cargo ship carrying luxury goods from Asia Minor, probably bound for Italy. Among its cargo were bronze and marble statues, glass vessels, amphorae, and at least one, and now apparently two, geared computers.

What kind of patron commissioned such instruments has long been a matter of speculation. Cicero, writing in the same period, describes a similar device built by Archimedes and brought to Rome by Marcellus.

The new finding does not name a patron. It does, however, sharpen the picture of a workshop tradition, somewhere in the eastern Mediterranean, capable of producing instruments of remarkable precision in small numbers.

Markou's team has identified at least three distinguishing features of the Fragment D gearwork that may, with further study, allow comparison with future finds.

The implication, quietly stated in the paper's conclusion, is that other ancient bronze "lumps" in museum drawers around the Mediterranean may not be what they have been catalogued as.

The National Archaeological Museum has begun a systematic re-imaging of all Antikythera-associated bronze objects, using neutron tomography where possible.

Markou is also collaborating with colleagues in Naples and Istanbul to survey unassigned bronze fragments from other Hellenistic shipwrecks.

There is something humbling about the idea that the question has not been, all these years, how to read what we have. The question has been which of our small green lumps are instruments and which are debris.

The answer, in at least one case, has turned out to be neither what was assumed nor what was hoped, but something more interesting than both.