last development it was not a true sphere on which could be represented the starry constellations, but an arrangement of circles forming a sort of imaginary sphere, the circles being intended to represent the relative position of the principal celestial circles. This instrument seems, at first, to have been suspended, when in use, but later was made to rest upon a base, the whole adjusted to revolve around an axis and within a graduated horizon circle. In the earliest examples, the earth at the center of the circles, it represented the Ptolemaic system (Fig. 9); in the later examples, having the sun at the center, it represented the Copernican system.
Fig. 9. Armillary Sphere according to Ptolemy.
It is expressly stated by Ptolemy that a celestial globe was constructed by Hipparchus, who is reputed to have been the founder of spherical trigonometry,41 and Pliny tells us that Hipparchus was the inventor of the astrolabe,42 which statement probably means that he greatly improved the simple armillae used at an earlier date as an instrument for astronomical calculations.
Ptolemy, in his ‘Syntaxis,’ or ‘Almagest’ as it was called by the Arabs, devoted a chapter to the method of constructing, and to the use of the astrolabe, which must have closely resembled the armillary sphere, describing therein, in terms not altogether easy of comprehension, its several rings and cylinders, and the method of adjusting the same for purposes of determining the latitude and the longitude of celestial bodies. He tells us also how to construct a representation of the sphere of the fixed stars by means of a solid ball, how to place thereon the several constellations, and how to use the same in the study of astronomical problems. Such a globe, he says, “should be of a dark color, that it might resemble the night and not the day.” His description is detailed as to the proper method of procedure in marking the position of the celestial circles on this globe, in arranging the movable rings of “hard and well polished material,” in graduating the rings and adjusting them to move about an axis which is likewise an axis of the globe proper. In marking the position of the fixed stars, we are told that the proper method is to commence at some constant and invariable point of a certain constellation, and he suggests that the best starting point is the fixed star in Canis Major, that is, the so-called dog star, or Sirius. “The position of the other fixed stars, as they follow in the list, could easily be determined,” he says, “by making the globe to turn upon the poles of the zodiac, thus bringing the graduated circle to the proper point of each. The stars could be marked with yellow or with such other color as one might choose, having due regard for their brilliancy and magnitude. The outline of each of the constellations should be made as simple as possible, indicating with light strokes, differing but little in color from that of the surface of the globe, the figures which the stars in the several constellations represent, preserving in this manner the chief advantage of such representation, which should be to make the several stars very prominent without destroying, by a variety of color, the resemblance of the object to the truth. It will be easy to make and to retain a proper comparison of the stars if we represent upon the sphere the real appearance or magnitude of the several stars. While neither the equator nor the tropics can be represented on the globe, it will not be difficult to ascertain the proper position of these circles. The first could be thought of as passing through that point on the graduated meridian circle which is 90 degrees from the poles. The points on this meridian circle 23 degrees 51 minutes (sic) each side of the equator will indicate the position of the tropics, that toward the north the summer solstitial circle, that toward the south the winter solstitial circle. With the revolution of the globe from east to west, as each star passes under the graduated meridian circle, we should be able to ascertain readily its distance from the equator or from the tropics.”43
That the Romans especially interested themselves in globes, either celestial or terrestrial, is not at all probable, because of their very practical inclinations. There is evidence, however, that in the time of the emperors celestial globes were constructed, especially in the studios of sculptors, but these were made largely for decorative purposes, having therefore an artistic rather than a scientific value. In the year 1900 there was found in a villa at Boscoreale, not far from Pompeii, an interesting fresco (Fig. 10), this being acquired by the Metropolitan Museum of New York in the year 1903. It has been referred to as a sundial, but was clearly intended to represent, in outline, a globe exhibiting the prominent parallels and a certain number of the meridians. It is not at all improbable that such subjects were frequently selected for wall or floor decoration.44 It appears that astrologers at times made use of globes in forecasting events.45 It may further be noted that on certain early Roman coins there may be found the representation of a globe (Figs. 11, 12), which perhaps had as its prime significance the representation of universal dominion.46
Fig. 10. Bosco Reale Roman Fresco, ca. 50 AD
Fig. 11. Greek and Roman Coins.
Fig. 12. Roman Gems.
Not until the day of the Byzantine Emperors do we meet with a real scholar who made a particular study of such astronomical apparatus, apparatus which he describes in a special treatise. Among historical scholars the work of Leontius Mechanicus seems not to have found the recognition which it deserves.47 He appears to have been a practical man, very active within the field concerning which he wrote, and from his remarkably detailed description we are able to learn something of the extent to which globe technique was carried in the days of the early Eastern Emperors. We at any rate learn from him that globes were constructed in his workshop, which globes, in all important respects, were like those in use at the present time, being, for example, provided with a meridian circle adjusted to move through notches in a horizon circle. The information given us by Leontius, which here follows, is in free translation or paraphrase of his treatise, the whole being condensed. He appears to have been a student of astronomy, as represented by Aratus, for he tells us that he had endeavored to construct a globe on which the constellations and the circles could be made to conform to the records of the ancient poet astronomer. He tells us further that he constructed this globe for Elpidius, an estimable man of letters, and one full of zeal for study; that at the time of its construction, though he had the leisure, he did not prepare a description of the globe, but on the insistence of his friends such description he now proposed to write. This appears to be the raison d’être for his treatise. The importance of adhering closely to the statements of Aratus he insists upon, though admitting that writer’s errors, being convinced that most of the globes of which one had knowledge in his day agreed neither with him nor with Ptolemy. Leontius first directs attention to Aratus’ threefold plan in describing the several constellations, in which description that author speaks first of the relation which part bears to part in each; second, of the position of each constellation relative to the celestial circles, as, for example, to the tropics, and third, its position in the heavens relative to the constellations
