In 1912, Henrietta Leavitt at Harvard looked at photos of a far-away cloud of stars, taken on glass night after night. Some of the stars blink: they get bright, fade, and get bright again, over and over.
You will look at her photos, turn a dial to find how long each star takes to blink, and put the star on a chart. Then you look at the chart. Is there a pattern?
Now move the dial, and watch the dots.
Each glass photo shows a far-away cloud of stars on one night. Every dot is a star. Small and pale pink: a faint star. Big and dark: a bright star. The key in the corner shows the scale. The star in the big ring is the one you are timing. Stars you have already timed keep a small ring, with the blink time you found, so you can compare their dots with the new one.
Each photo becomes a dot in the box below it, the same size as the star was that night. Further right means a later night. Higher (and bigger) means the star was brighter that night. At first the dots look messy, because the photos were taken on random nights.
The blink time is how long the star takes to blink once: a quick jump up to bright, then a slow slide down. Move the dial to guess it. The dots jump to new places: now they show where each night fell inside one blink, and the box shows two blinks side by side.
A thin line joins the dots in order. Wrong guess: the line is a messy scribble. Right guess: it turns into a neat zigzag, a quick jump up and a slow slide down, twice. The top dial goes from 1 day to 128 days. The lower dial is for fine tuning.
Careful: if the zigzag jumps up twice in each half of the box, your guess is twice too long.
Slide the top dial slowly from left to right. Nearly everywhere the line is a scribble; at a few spots it snaps into a tidy shape. Stop there and check: you want one hump in each half of the box, a quick rise and a slow slide. Two humps in each half means your time is twice too long, so halve it. Then use the fine dial until the dots hug the line as tightly as they can, with no stray dots. The tightest picture is the answer.
Once the chart shows a pattern, use it: a big dark star means start far to the right, a small pale one far to the left.
When the dots line up, press Write it down. Then you see Leavitt’s own blink time for that star, in violet, and her blink drawn over your dots.
Each star goes on the chart as a line from its faintest to its brightest, in the same colours as on the glass. Left to right is the blink time you wrote: slow blinkers on the right. Each step to the right doubles the time. Up is brighter. Watch what the lines do.
For the seventh star you guess its blink time from its brightness, before its photos arrive. The glass shows all seven stars at their brightest in a row, faintest to brightest, with the blink times you found. The new one sits between two of them: guess a time between theirs. A dashed line shows where your guess would put it on the chart. Then its photos come in and you time it as before, starting from your guess.
Everything in violet is Leavitt’s. You only see it after you have written down yours.
Built from Henrietta Swan Leavitt, “Periods of 25 variable stars in the Small Magellanic Cloud”, Harvard College Observatory Circular 173 (1912), 1–3, signed by Edward C. Pickering, 3 March 1912. Bibcode 1912HarCi.173....1L. Read it at the NASA Astrophysics Data System. Game built from this paper; the paper is the authority where the two disagree.
Leavitt had a catalogue of 1,777 variable stars in the two Magellanic Clouds, all faint, on a limited stock of long-exposure plates. For 25 in the Small Cloud she found the period. Set side by side, the brighter ones had the longer periods; plotted against the logarithm of the period, the stars at their brightest fell on one straight line and at their faintest on another, the logarithm rising by about 0.48 for each magnitude. Since the stars are probably all at nearly the same distance, she wrote, their periods go with how much light they really give out. She hoped for spectra and for the parallaxes of some such stars, and for a study of all the variables in the Clouds.
You time seven of her twenty-five from glass plates, by folding the plates on a trial beat until they line up; each goes on a chart as a stroke from its faintest to its brightest at the beat you wrote. The seventh you first set from its light alone. Then her other eighteen and her two lines appear. The chart can be spaced by doublings of the beat, as her Figure 2 (two straight lines), or by days, as her Figure 1 (two curves).
Kind: reconstructed. The game runs a constructed set of plates that reproduces Table I’s periods, brightest and faintest magnitudes and epochs, with a light curve shaped as the circular describes it (a slow fade, a long time near the faintest, a very quick rise to a brief maximum).
The game demonstrates that a period can be found from uneven plates by folding them, and that her twenty-five stars, set by period and brightness, fall on her straight lines.
The game assumes the plates themselves: their dates, how many (26 a star, covering four to six blinks; hers covered years and many hundreds), the estimating error (about a tenth of a magnitude), and the exact shape of the light curve. It also assumes the fixed time-lapse at which the ringed star blinks on the glass.
The paper establishes the 25 periods and brightness ranges, and the straight-line relation between brightness and the logarithm of the period, on a provisional scale of magnitudes. It argues, but does not show, that the stars are at nearly one distance, and it measures no distance. Using the relation to measure distances came later, once the brightness of a few such stars nearby could be fixed.
Table I was read by eye off the scan; no transcription was found. A check recomputes the printed residual columns from it and agrees for every star but one: for Harvard 1374 the faintest-magnitude residual is printed as −0.3 where its own numbers give +0.3, a slip in the 1912 table. The game uses the magnitudes as printed. No magnitude appears on screen: brightness is a dot’s size and a stroke’s height, and the 0.48 is said in words (a blink about three times as long, about two and a half times the light).
The comparison stars and the magnitude scale; the epochs; the light curves of brighter stars such as UY Cygni, and the questions about spectra and the limits of mass the circular raises; the other two thousand variables.
Built with Claude under the paper-to-game skill. Type: Spectral (Production Type) and Josefin Sans (Santiago Orozco), SIL Open Font License, embedded. Source, design memo and calibration check: github.com/steffenhuck/paper-to-game.