Forty Minutes Wrong Because I Trusted the Star

Planispheric Astrolabe Construction and Star-Timekeeping 🎮 Play: Turn the Rete

I already owned this latitude. Back in the spring I cut a brass sundial gnomon and set it at 53.5° from horizontal, the angle that aims it dead at the celestial pole for where I live outside Edmonton. What I did not understand until last night is that the astrolabe takes that single number and bakes it into the geometry of the entire sky — a disc engraved so specifically for one parallel of latitude that carrying it to Mexico City would render it a pretty, lying ornament.

That is the first thing to know. The flat plate under the star map — the tympan — is not universal. Its curved altitude circles (the almucantars) and the azimuth arcs that cross them are computed for one place and one place only. Medieval makers sold astrolabes with a stack of interchangeable plates, one per city, called climates. Mine will need exactly one, and I already know its defining constant by heart.

Why a compass is enough. The reason the whole thing is drawable — no plotting a hundred points, just dividers swinging arcs — is a single property of stereographic projection: it turns every circle on the sphere into a circle on the paper. You project the celestial sphere from the south pole onto the plane of the equator, and horizons, altitude rings, the tropics, all of them survive as honest circles. A star at declination δ lands at radius r = R·tan((90° − δ)/2) from the centre, where R is the radius you chose for the equator. Feed that one formula a list of declinations and you get a set of compass settings. The draughting is nearly identical to laying out sundial hour-lines, except now I’m flattening the sphere instead of a shadow.

I did not touch brass on the first evening. I built the thing in paper, glued to cardstock, the way every sane beginner apparently starts — and immediately drew it inside-out. My instinct said the biggest bounding circle should be the northern tropic. Wrong. Projecting from the south pole throws southern declinations outward, so the outermost ring is the Tropic of Capricorn and the little inner circle is Cancer. I had them reversed, noticed the horizon curving the wrong way against the almucantars, and started over. An hour of careful pencil work undone by one hemisphere.

Over that plate sits the rete — Latin for “net” — a pierced overlay carrying flame-tipped pointers to a dozen bright stars and an off-centre ring for the ecliptic. Rotating it models the sky’s daily turn. One sobering detail while placing the pointers: precession drags the stars about one degree every seventy-two years, which is precisely how curators date antique astrolabes. Use catalogue coordinates from three centuries ago and your instrument is born already wrong. I pulled current positions for Vega, Arcturus, and Capella and pricked them through.

Then the part that connects back to a frustration I logged in May. To actually read the sky you flip the whole assembly over, hang it from its ring by the ornamented throne so gravity gives you a true vertical, and sight a star through the two pinhole vanes of the alidade. It’s the hanging-plumb cousin of squinting a star down the monocular of my Davis sextant — measure an altitude, extract the time. I took the paper mock outside at 22:40, hung it off my thumb, and caught Vega at roughly 66°.

Back inside, I rotated the rete until Vega’s pointer kissed the 66° almucantar, laid the straight rule across the disc, and read a time that was forty minutes wrong.

Forty minutes. I stared at it, convinced the projection was off, before the actual error surfaced: I’d aligned the rule with the star. You don’t read time off the star. You read it off the Sun’s position on the ecliptic for today’s date — the star only sets the rete’s rotation; the Sun’s date-point is the index that meets the hour scale on the rim. Moved the rule to mid-August on the ecliptic ring, and the disc told me 22:44. My watch said 22:41.

Three minutes, off a smudged paper disc I sighted by hanging from one finger. I’m not entirely sure I trust it yet — one lucky star is not a calibration — but the rule crossing the rim at nearly the right hour did something to me that the reversed hemispheres and the forty-minute humiliation had not quite undone.