The Glimmery

The fourth of the “-ery” sheets, after The Hatchery’s twelve cosmic eggs, The Quillery’s eight Esmx and The Bow-ery’s ten bows. Same rule as all three: every drawing carries one checkable thing, with its source under it. The subject is the word that Glimmers is named for, taken in its plainest sense — a small light, at or near the edge of being seen — and followed into ten places where somebody has measured one.

The one thing to know before the plates

The block of dots in the corner of every plate is the same block, at the same size, on all ten. It is the human visual threshold as Hecht, Shlaer and Pirenne measured it in 1942: about five hundred rod cells under a small brief flash, of which 5 to 14 need to absorb one quantum each before you see anything at all. Fourteen are drawn lit, which is the top of that range, and the caption on every plate says so.

That block is doing the job the Bow-ery’s dashed 42° ghost does: it is the reference the plate is being read against. Plate 1 is nothing but the block, drawn large, and the other nine carry it small.

What this sheet cannot show: how faint these are relative to each other. That is the scale we wanted and could not build, and the reason is at the foot of the page rather than buried here.

What this is, and what it is not

These are drawings, not photographs, and not data. Where a number is doing real work it is printed on the plate and sourced under it — 87 km, 15 ms, 48.6°, 82% against 3%, ×1.3. Everything else is drawn to be legible: the thickness of a layer, the exact shape of a brightness curve, the number of dust grains in a cone. If a figure is not printed on the plate, do not measure it off the plate.

Nothing here is drawn dim. That was the first idea and it is a bad one twice over. On a screen, faintness is a lie you tell about a device whose brightness the reader sets; on paper, where this collection expects to be printed, the whole sheet inverts and a faint mark becomes a nearly invisible one. Faintness on this page is carried by numbers and by the rod block, never by low-contrast ink. That is the house rule about opacity arriving from an unexpected direction, and it is the right rule.

The word is borrowed, and it is contested. Glimmer in the sense our Glimmers collection uses it is Deb Dana’s, from 2018, and it arrived inside polyvagal theory, which is influential and disputed. Helen Edgar’s reading — simply the opposite of a trigger — is the working definition there, and No. 80 is the piece that argues it. This sheet does not adjudicate any of that. It takes the older, literal meaning and goes looking for measurements.

The ten

The Threshold

No. 1 · the plate the other nine are drawn against

In 1942 Selig Hecht, Simon Shlaer and Maurice Pirenne measured the least light a human being can see. At the cornea it is between 54 and 148 quanta of blue-green light. Then they took it apart: 4 per cent is reflected by the cornea, almost exactly half is absorbed by the ocular media, and at least 80 per cent gets through the retina. What is left is 5 to 14 quanta actually absorbed — and because they were spread over some 500 rods, no rod could plausibly have caught two. One quantum, in each of 5 to 14 separate cells, and you see something.

That is the block drawn here, and a small copy of it sits in the corner of every other plate on this sheet. It is what the other nine are being compared against.

But the finding this page is really built on is the fifth point of their summary, and it is not about light at all. The assumption had always been that the stimulus is constant and the organism variable — that when a person sees the flash on one trial and misses it on the next, the person is what changed. Hecht and his colleagues showed the opposite: at the threshold it is the stimulus which is variable, arriving in a countable number of quanta that fluctuates trial to trial, and the person’s inconsistency is the honest readout of a world that is not sending the same thing twice.

Seventy-four years later Tinsley and colleagues went one further and asked whether a person can report a single photon. They could, at 0.516±0.010 against a 0.5 baseline — and we should say plainly that the headline P value is 0.0545, which is on the wrong side of the line the same papers usually draw at 0.05. The result that carries real weight is the one restricted to trials where subjects said they were confident: 0.60±0.03, P=0.0010. The eye reaches the floor. The claim about it arrives hedged, and we are not going to unhedge it.

Sources: Hecht, S., Shlaer, S. & Pirenne, M. H., “Energy, Quanta, and Vision,” Journal of General Physiology 25(6):819–840 (1942) — the 54–148 figure, the three corrections, the 5–14 absorbed, and the summary’s fifth point, read from the paper itself rather than from a textbook’s version of it. Tinsley, J. N. et al., “Direct detection of a single photon by humans,” Nature Communications 7:12172 (2016) — both P values are the paper’s own.

Airglow

No. 2 · there is no such thing as a dark sky

Take away the Moon, the towns, the planets and every star, and the sky still is not black. The air is glowing on its own account. A. B. Meinel found the bands in 1950 with a spectrograph pointed at the night sky from Yerkes, and they turned out to be hydroxyl — OH — sitting in a layer conventionally given as 87 km up, about 8 km thick.

The reaction is the plain one printed on the plate: O₃ + H → OH* + O₂, releasing 3.34 eV, and the excited molecule sheds it as light. This is chemistry, running in the dark, with no sunlight falling on it. The sky’s own faint light is a slow reaction between ozone and hydrogen atoms, eighty-seven kilometres over your head, all night, every night.

We wanted to tell you it is the brightest natural component of a moonless sky. We have seen that said in several places and could not put a primary source under it, so it is not on the plate. What is on the plate is the height, the width and the reaction, each of which we could.

Sources: Meinel, A. B., “OH emission bands in the spectrum of the night sky,” Astrophysical Journal 111:555 (1950). Height, width and reaction: Wüst, S., Bittner, M., Espy, P. J., French, W. J. R. & Mulligan, F. J., “Hydroxyl airglow observations for investigating atmospheric dynamics: results and challenges,” Atmospheric Chemistry and Physics 23:1599–1618 (2023), which gives 87 km and an 8 km FWHM after Baker & Stair (1988).

The Zodiacal Light

No. 3 · sunlight, hours after sunset, off something solid

A faint cone leaning up from where the Sun went down, along the ecliptic. It is sunlight scattered off dust — the flattened interplanetary cloud that fills the inner solar system, grains shed by comets and ground off asteroids. It emits nothing. Every photon in it came from the Sun and bounced.

It is also the reason the phrase empty space is a figure of speech. The cloud is of “extremely low optical thickness” and you can still see it with your eyes, from the ground, if you get far enough from streetlights and look west after the last of twilight has gone.

A date we could not settle, and are not going to pretend we did. Cassini’s work on this is usually dated 1683 in popular accounts; the review we are citing lists it in its own bibliography as Cassini, G. D. (1693), Découverte de la lumière celeste qui paroist dans le zodiaque, Mémoires de l’Académie Royale des Sciences. Observation and publication are two different years and the Mémoires were reissued more than once. The plate prints 1693 because that is what the source we actually read prints.

Sources: Lasue, J., Levasseur-Regourd, A.-C. & Renard, J.-B., “Zodiacal light observations and its link with cosmic dust: a review” (arXiv:2005.07480, 2020) — “The first scientific work on the study of the zodiacal light was published by Cassini in the 17th century,” and the 1693 bibliography entry. We have deliberately left out the frequently repeated claim that Joshua Childrey gave the first description in 1661: it may well be right, and it is not in the source we read.

The Gegenschein

No. 4 · the sky is brightest where the Sun is not

Follow that same dust away from the Sun and the glow falls off steeply, flattens into a broad minimum — and then, at exactly 180°, goes back up. A faint oval directly opposite the Sun, in the direction of your own shadow, cast out into space.

The measured size of the rise is about 1.3 times the ecliptic minimum. It is backscatter: dust grains throwing light back the way it came, the same effect that makes dew-covered grass glow around the shadow of your head. Surveys find it always sits at the anti-solar point, and the grains doing it are dark — an albedo of about 0.06, which is darker than asphalt.

Almost black dust, lit from behind you, is the brightest thing in that part of the sky.

Sources: Lasue, Levasseur-Regourd & Renard (2020), as above — the 1.3× figure is the review’s. Anti-solar positioning: Buffington et al. (2009), cited there. Albedo 0.06: Ishiguro et al. (2013), likewise. The brightness curve on the plate is drawn to show the shape — steep fall, flat minimum, small rise — and only the 1.3× ratio at the antisolar point is a measured quantity.

Noctilucent Clouds

No. 5 · still in sunlight when everything below is not

The highest clouds there are. They were first written down in the summer of 1885 — Backhouse in the Meteorological Magazine, Leslie in Nature, Jesse in Germany — and eleven years later Jesse triangulated them at about 83 km, roughly eight times higher than the cloud deck of an ordinary rainy afternoon.

That height is the whole phenomenon. After sunset the shadow of the Earth climbs; ordinary clouds go dark because they are inside it. These are far enough up that the Sun, already below the observer’s horizon, is still above theirs. So they hang there lit while everything underneath them has gone out.

They are made of ice on almost nothing. At 83 km the air is around a hundred-thousandth of sea-level pressure, and it is the coldest place in the atmosphere. The water freezes onto motes of meteoric dust — the same population of debris the zodiacal light is made of, two plates back, arriving at the end of its journey.

Sources: Fiedler, J. & Baumgarten, G., “Solar and lunar tides in noctilucent clouds as determined by ground-based lidar,” Atmospheric Chemistry and Physics 18:16051–16061 (2018), whose introduction carries the 1885 first records (Backhouse, Met. Mag. 20:133; Leslie, Nature 32:245; Jesse, 1885) and the ~83 km optical triangulation (Jesse, Astronomische Nachrichten 140:161–168, 1896).

Foxfire

No. 6 · a light that keeps time, and is addressed to somebody

Rotting wood that glows. It has been noticed for as long as there have been people in forests at night, and the obvious modern assumption — that it is a by-product, waste light from a metabolism doing something else — turns out to be wrong twice over.

First, it is on a clock. Kept in constant darkness, the mycelium of Neonothopanus gardneri goes on brightening at night and dimming by day, driven by a temperature-compensated circadian rhythm in the luciferase, the reductase and the luciferin together. Nothing outside is telling it what time it is. It knows.

Second, it is calling. The light attracts insects, and the insects carry spores — which matters most exactly where this fungus lives, on the floor of a forest, under a canopy where there is not enough wind to do the job. The clock makes the call efficient: shine when there is something to be seen by.

A glimmer, in the most literal sense available: a small light, kept deliberately, meant for someone else.

Sources: Oliveira, A. G., Stevani, C. V., Waldenmaier, H. E., Viviani, V., Emerson, J. M., Loros, J. J. & Dunlap, J. C., “Circadian control sheds light on fungal bioluminescence,” Current Biology 25(7):964–968 (2015) — the temperature-compensated clock, the cycling of the luminescent system’s components, and the insect-attraction result.

The Spark

No. 7 · touched, and it answers before you can blink

Disturb the water where dinoflagellates are and it lights up: a bow wave, an oar, a foot. Latz and colleagues put single cells through a microfluidic channel and timed it properly. Mean latency from stimulus to light: 15 milliseconds. The fastest measured: 12. The flash itself runs about 70 ms, and a cell can fire again with as little as 5 ms between flashes.

The mechanism is worth having. The cell carries its luciferase in small packets called scintillons, hanging into a big central vacuole. Mechanical stress opens a voltage-gated proton channel; protons pour from the vacuole into the scintillon; the pH drops; the enzyme switches on. The trigger is not a chemical signal. It is a change in acidity, arriving as fast as electricity.

We have put it on this sheet because of what kind of light it is. Airglow does not care that you are there. This does. It is a response — and a fast one, on a timescale where the interesting comparison is not to other lights but to your own reflexes.

Sources: Latz, M. I., Bovard, M., VanDelinder, V., Segre, E., Rohr, J. & Groisman, A., “Bioluminescent response of individual dinoflagellate cells to hydrodynamic stress measured with millisecond resolution in a microfluidic device,” Journal of Experimental Biology 211(17):2865–2875 (2008) — 15 ms mean latency, 12 ms minimum, ~70 ms emission, 5 ms interflash interval. The proton channel: Rodriguez, J. D. et al., “Identification of a vacuolar proton channel that triggers the bioluminescent flash in dinoflagellates,” PLOS ONE (2017).

Cherenkov Blue

No. 8 · faster than light, and nothing broken

The blue in a reactor pool. Pavel Cherenkov found it in 1934 as a graduate student, looking at what pure liquids do under gamma rays and refusing to accept that the faint glow was fluorescence.

Nothing here exceeds the speed of light. What is exceeded is the speed of light in water — with a refractive index of about 1.33, that is three-quarters of c, and a fast electron has no difficulty beating it. The light the particle has already made cannot keep up with the particle, so the wavefronts pile into a cone behind it, exactly as a sonic boom does. The geometry is fixed by the ratio: the shock front stands at 48.6° to the track, and the two lines on the plate are tangent to the circles by construction rather than by eye.

An attribution worth getting right. The 1958 Nobel went to Cherenkov with Ilya Frank and Igor Tamm, who worked out the theory. In Russian the effect is usually the Vavilov–Cherenkov effect, after Sergei Vavilov, who supervised the work and pushed the interpretation. He died in 1951, and the prize is not awarded posthumously. The naming and the prize disagree, and both are telling the truth about something.

Sources: Cherenkov, P. A., “Visible glow of pure liquids under gamma irradiation,” Doklady Akademii Nauk SSSR 2:451–457 (1934). The Nobel Prize in Physics 1958, awarded jointly to Pavel Cherenkov, Ilya Frank and Igor Tamm “for the discovery and the interpretation of the Cherenkov effect” (NobelPrize.org). The 48.6° on the plate is computed from cos θ_c = 1/n at n = 1.33 with β = 1, not measured.

Synchrony

No. 9 · the signal only works if it is shared

Photinus carolinus males flash together, in bursts, across a whole hillside. For a long time the interesting question was taken to be how — what oscillator locks them. Moiseff and Copeland asked what it is for, and answered it from the female’s side.

They built an array of LEDs producing the species’ own flash pattern, put females in front of it, and varied only whether the lights were in step. Synchronous: females responded to 82% of flashes. Asynchronous: as few as 3%.

The reason is visual clutter. A female is trying to pick one male’s pattern out of a forest full of other males flashing the same pattern at different times, and out of step the signal is destroyed — not weakened, destroyed, by a factor of twenty-seven. In step, the whole hillside becomes one legible thing.

It is the plate on this sheet that is least about physics and most about what a light is doing. The individual flash is unchanged in both conditions. What changes is whether the others are with it.

Sources: Moiseff, A. & Copeland, J., “Firefly synchrony: a behavioral strategy to minimize visual clutter,” Science 329(5988):181 (2010) — the 82% and 3% are the paper’s. The plate’s rasters show eight males over five cycles, which is a drawing of the design rather than the published stimulus.

The Milky Way

No. 10 · the one we switched off

Every other plate on this sheet is about light that is hard to see because it is faint. This one is about light that is easy to see and is not there any more, for most people, because of light we made.

Falchi and colleagues built the atlas from satellite radiance calibrated against thousands of ground measurements. More than 80% of the world’s population lives under light-polluted skies, and the Milky Way is hidden from more than one third of humanity. Not dimmed. Below the threshold the first plate on this sheet is about.

The band is still there. Nothing has happened to the galaxy. The drawing shows one continuous band across the frame, with a wash coming up from one side, because that is the honest picture: the same sky, twice, and the difference is entirely at our end.

It is the one glimmer on this page with a switch on it. We think that is the hopeful reading rather than the bleak one, and we would rather say so than pretend the plate is neutral.

Sources: Falchi, F., Cinzano, P., Duriscoe, D., Kyba, C. C. M., Elvidge, C. D., Baugh, K., Portnov, B. A., Rybnikova, N. A. & Furgoni, R., “The new world atlas of artificial night sky brightness,” Science Advances 2(6):e1600377 (2016). The two figures quoted are the ones we could confirm against the paper’s own abstract; the widely repeated regional splits (60% of Europeans, nearly 80% of North Americans) come from the same work but we read them only in secondary coverage, so they are not on the plate.

What we left out, and the scale we could not build

The plan was a brightness scale, and it was wrong. Every other “-ery” sheet holds one thing constant — one egg outline, one figure, one angular scale — and the obvious constant here was luminance: put all ten on a shared logarithmic strip and let the reader compare faintness by eye, which is the exact pleasure the Bow-ery gives you with radii. It cannot be done, and finding out why was the most useful part of making this page.

Only some of these are the same kind of object. Airglow, the zodiacal light, the gegenschein and the Milky Way are surface brightnesses: diffuse sky, quotable in magnitudes per square arcsecond, genuinely comparable. A noctilucent cloud is usually reported as an albedo or against the twilight sky, not as a sky surface brightness. A single photon is not a brightness at all. And a mushroom, a dinoflagellate and a reactor pool are objects at arm’s length whose apparent brightness depends entirely on how far away you are standing. Four of the ten belong on one bar. Putting the other six there would have required inventing numbers, and a plate that invents a number is worse than a plate that admits it has none.

So the constant moved from the source to the detector. If the ten lights cannot be compared to each other, they can all be compared to the same thing — the floor of human vision, which is published, which is a real measurement, and which every one of these lights is either above or, in the last plate’s case, no longer above. Hold the eye still and let the sky vary. That is also, as it happens, exactly what the 1942 paper found: the observer was never the unreliable part.

What we could not source, and left off rather than rounding. Three things. That airglow is the brightest natural component of a moonless sky — widely said, and we could not put a primary under it. That Joshua Childrey gave the first description of the zodiacal light in 1661 — likely true, absent from the review we actually read. And the regional splits in the light-pollution atlas (60% of Europeans, nearly 80% of North Americans) — the paper’s own, but reaching us only through news coverage, so the plate prints the two figures we confirmed instead. Three small holes, named, rather than three sentences that read fine.

One number on the sheet is weaker than it looks, and it is the headline. The single-photon result carries P = 0.0545, which is on the wrong side of the threshold those papers conventionally use. We have put it on the plate in that form rather than reaching for the stronger confidence-restricted figure and calling it the result. A sheet about the edge of detectability has no business being confident about where the edge is.

What is not here. The aurora, because it is bright, well drawn elsewhere on this site and not a glimmer. Deep-sea bioluminescence, which deserves a field guide rather than a plate. Phosphenes and the light astronauts report seeing with their eyes closed, which we could not source to our own standard in the time this took. And the green flash, which is faint and brief and is a Bow-ery subject — refraction and dispersion, not a light of its own.

Printed in grey. The colours on this sheet are flattened for legibility, and two plates carry an argument in colour that the paper copy states as a label instead. The rod block, every printed number and every label are unaffected — which is the whole comparison. The colour version is at starstuff.earth/glimmery.

Why a fourth “-ery”, and why it is an Easter egg. The Hatchery came from the small egg at the foot of the front page; The Quillery from the mascot beside it; The Bow-ery from a catalogue that could not show you how big anything was. This one comes from a collection named after a word for faint light that had never once been drawn as faint light. An egg is not a lower standard here — same sourcing, same fact-check ledger, same eight gates, same changelog entry as anything else on this site. Only the door moves.