Section 1 · Stereographic Projection of Equator, Tropics & Ecliptic¶
Planisphaerium, Section 1 (Heiberg 1907, Vol. II, pp. 227–229)¶
Exordium to Syrus & Construction of the Equatorial Plane (Heiberg II, p. 227)¶
p. 227 Since it is possible, O Syrus (Jesure), and frequently necessary, that the circles occurring upon the solid sphere be represented in a flat plane just as if it actually were flat, it seemed advisable in the truth of science that one who wishes to know this should set forth a demonstrating rationale whereby it is proper to assign the oblique circle, the circles parallel to the equinoctial circle, and the circles known through the meridian circle, and whatever is intended to be adapted to that which appears in the corporeal sphere.
This intention requires that in place of the meridian circle we use straight lines, and between the parallel circles to the right circle, that first a great circle be made assigning the oblique circle, reaching on this side and that the parallels to the right circle at equal distance on both sides, which it bisects in this manner:
We describe the equinoctial circle with letters a, b, g, d around center e, whose diameters orthogonally cutting one another are ag and bd. We shall understand one diameter to be the meridian circle, and point e to be the northern pole; nor indeed is it fitting that the other pole be placed in the plane with respect to this one, as will be established in what follows. And since the northern pole perpetually appears in our region, it is more suitable in the plane whose assignment is ours.
Projection of Parallels from the South Pole (Heiberg II, pp. 227–228)¶
p. 228 It is therefore necessary that of the circles parallel to the right circle, the northern parallel be described internally, and the southern parallel externally. That this may be done correctly, we produce line ag in both directions, and from circle abgd on both sides of g we cut off two equal arcs—above gh, and below gn. We join straight lines from d with both marks, in such manner that dh reaches line ag and marks point k, while dn where it touches line ag marks point t.
This being done, with center fixed at e at distance ek, a circle is made upon diameter km (the Tropic of Capricorn); and without moving, another circle is consequently made at distance et of the line upon diameter tl (the Tropic of Cancer).
Theorem: The Projected Oblique Ecliptic Circle Bisects the Equator (Heiberg II, pp. 228–229)¶
Then, segment tm being bisected at its midpoint r, a circle is described around point r with radius equal to the half of tm.
I say that those two circles are parallel to the equinoctial at equal distance on both sides, and the third circle around center r is the oblique circle, which line tm cuts into equal halves until it reaches both of them—the one at mark m and the other at point t—and that it bisects the equinoctial circle, which it intercepts at the two opposite points b and d.
Demonstration. That this may be established by reason, a straight line dm joins d with m, crossing the equinoctial circle at point z. Since therefore arc az is equal to arc gh, which is given equal to arc gn, it is necessary that arc zdn be half the whole circle; whence it follows that angle ∠mdt is a right angle (angulum mdt rectum esse, by Eucl. III.31).
Since therefore the circle described upon line tm circumscribes right-angled triangle mdt and passes through point d, it must also pass through point b; consequently it cuts the equinoctial circle into equal halves.
p. 229 Hence it is established that between circles parallel to the right circle, when we duplicate equal arcs on either side of point g, their quantity measures the arc of total declination. When we join their ends by straight lines with point d, the straight lines they cut off from line ag give the distances of the circles described around center e—so that internally is the Tropic of Cancer, and externally the Tropic of Capricorn, with the zodiac circle mbtd touching both and cutting the equinoctial into equal halves.
Our determination measures both arcs ng and gh as $23^\circ 51'$ parts out of the $360^\circ$ measuring the whole circle abgd, which is the distance of both tropics from the equinoctial circle on the solid sphere.