The model is a reconstruction of the original fragmentary instrument, believed to date from the early sixth century A.D., which is also on display (Field & Wright, 1985).

The instrument comprises two practically independent parts: a sundial for use at any latitude, and a geared calendrical device showing the phase of the Moon, the day of the month and the places of the Sun and the Moon in the Zodiac.

This type of sundial is attested by a several examples, some inscribed in Latin and some in Greek (Field, 1990). None is securely dated, but the archaeological record suggests that these instruments were widely distributed within the Roman and the early Byzantine empires. The design probably corresponds to that described by the Roman author Vitruvius (late 1st century B.C.) as “pros pan clima” (for every latitude), suggesting a yet earlier Greek origin.

The dial occupies most of one face of the instrument. It comprises a piece that is both shadow-caster and hour-scale, the central pin of which passes first through the circular body and then through the swinging arm at the back to which is jointed a ring by which the instrument is hung upright.

The dial occupies most of one face of the instrument. It comprises a piece that is both shadow-caster and hour-scale, the central pin of which passes first through the circular body and then through the swinging arm at the back to which is jointed a ring by which the instrument is hung upright. Two scales on the body enable the user to adjust the first part to the elevation of the Sun at noon, according to the place and the time of year: the shadow-caster is moved over a double scale of solar declination, marked out with abbreviations of the Julian month-names; and the arm is adjusted according to a quadrant scale of latitude near the rim of the body. The dial is then held up, and rotated until the shadow of the projecting part falls along the curved scale, whereupon the user may read off the morning or afternoon hours. Much of the rest of the face of the dial is taken up with a reference table of place-names and their latitudes.

The known comparable dials are smaller and are based on flat circular discs. In this case, uniquely, a hollow box, which takes the place of the disc, contains a geared calendrical mechanism which is worked by turning a pointer on the face of the dial.

The pointer moves over a circle of seven incised heads representing the seven days of the Judaeo-Christian week. A ratchet inside prevents the user from turning it backwards. Simple gearing in the ratio 7:59 rotates a disc making one turn in 59 days which displays the day of the month (alternately 29 and 30 days in length) and an approximate representation of the phase of the Moon, through openings in the back of the box. Following the description of a similar instrument by al-Bīrūnī (Hill, 1985), the remainder of the mechanism is restored to drive indications of the places of the Moon and of the Sun in the Zodiac. A more elaborate reconstruction might include a display of the Moon’s nodes (enabling the user to predict the possibility of eclipse) or mechanism whereby the position of the shadow-caster and hour-scale is set automatically by the calendar (Wright, 1990); but these possibilities have no historical basis.