Report — Field Note
A Forest Made of Missing Light
Hundreds of dark lines in the light of one quasar. A name borrowed from a walk in the woods, the places it stops being true — and why the most untestable theory in physics now depends on it.
A quasar is the core of a young galaxy, lit by matter falling into a massive black hole and bright enough to be seen across most of the observable universe. Its light travels to us along a single line. On the way it crosses long stretches of hydrogen — thin gas drawn out along the filaments of large-scale structure, with no edges worth the word.
Pass that light through an instrument that spreads it by wavelength, the way a prism spreads sunlight, and you get a band running from blue to red. Uninterrupted, it would be smooth.
Hydrogen absorbs at one particular ultraviolet wavelength — 1216 ångströms, called Lyman-alpha after Theodore Lyman, who measured that family of lines early in the last century. A patch of hydrogen along the path removes a thin slice of light there and leaves a dark line behind. One patch. One line.
The universe expands, so light stretches as it travels. A patch nearer to us leaves its line at one position; a patch farther away leaves its line at another. Hundreds of patches leave hundreds of lines. In 1970 Roger Lynds saw this in the light of a quasar catalogued as 4C 05.34 and called it a forest. The name has never been replaced. Notice what the picture is made of: light that was taken out.
It is right in one respect. Stand at the edge of woods, look straight ahead, and trunks at many depths each block part of your view along a single line of sight — which is the arrangement exactly. Then it slips. Trees are solid things with edges; these are diffuse. A wood can be walked into and sampled in any direction; this is one needle-thin core, and a second line needs a second quasar. And a forest stands in the present, while reading this band from one end to the other is reading across billions of years.
Those lines record how clumpy matter is at fine scales. The cosmic microwave background — the leftover glow of the early universe — records how clumpy it is at very large ones. Given a theory of what matter is, the two have to agree. They do not entirely agree, and the gap has sat there for years, too small to be a scandal and too stubborn to dismiss.
String theory proposes that the smallest things in nature are not points but tiny vibrating strings, and that the universe has extra dimensions folded up too small to see. Its mathematics is vast and consistent, and this May it got stronger: physicists at Caltech, NYU and Barcelona showed that its signature ladder of ever-heavier particles follows from two basic assumptions about how particles collide. But the theory allows an enormous number of possible universes — roughly one for every way of folding those extra dimensions — and it has never made a prediction an experiment could check. Its critics, Peter Woit and Sabine Hossenfelder among them, have said so for twenty years. The charge is not that it is false. It is that nothing it says can be told it is wrong.
In August, Mudit Jain, David Marsh and Keir Rogers at King's College London went looking for a place where the theory touches data. Almost every way of folding string theory's extra dimensions produces axions: hypothetical, extremely light particles, and a leading candidate for dark matter, the unseen mass whose gravity holds galaxies together. There are far too many foldings to check one at a time, so the team used machine learning and statistical weighing to survey millions of them at once, asking of each what axions it predicts and whether they fit the forest. The software did not discover anything. It sorted a pile too large for people to sort — and a small group of models came out fitting the gap between the forest and the microwave background.
The models that fit are the smaller half of the result. The larger half is this: if an axion is ever detected above a certain mass, whole classes of string models are eliminated outright. That is a friction point — conditional on a detection nobody has made, dependent on assumptions about dark matter, and leaving much of the landscape standing. Astrid Eichhorn's objection hangs over all of it: whether the calculations behind string theory still make sense once gravity itself is quantum. Still, for the first time in decades a measurement could come back and say no, and it would come back written in missing light. The name was tested at the start of this piece; the theory is tested at the end. The words to look up are Lyman-alpha forest and axion — a particle named in 1978 by Frank Wilczek after a laundry detergent, because it cleaned up a problem.
Sources: King's College London, Aug 5 2026 · Jain, Marsh & Rogers, Physical Review D (2026) · Lyman-alpha forest — overview · W. Keel, University of Alabama — the Lyman-alpha forest · NASA Astronomy Picture of the Day, Feb 24 1998 · Lynds, Astrophysical Journal Letters (1971) · Cheung, Remmen, Sciotti & Tarquini, Physical Review Letters, May 19 2026 · Quanta Magazine, Mar 23 2026. Diagram is schematic: absorbers sit blueward of the quasar's own emission line, but line spacing, model counts and the threshold are drawn for legibility, not computed.
note: the line positions here are schematic. Real absorbers all sit blueward of the quasar's own emission line, which the drawing respects, but the left-to-right spacing is drawn for legibility rather than from a redshift calculation
Sources: KCL · Lyman-alpha forest (Wikipedia) · Keel, U. Alabama — the Lyman-alpha forest · NASA APOD, 1998 · ScienceDaily · Quanta, Mar 23