A resonator amplifies whatever is inside it. The volume between Earth's surface and the ionosphere rings with electromagnetic waves in general [7], so the boost that would raise a dark matter signal near 8 hertz raises everything else living in that band. That puts the weight of the result on the cleaning step: the team stripped artificial sources out of the record, looked for a steady signal concentrated in a very narrow frequency range, and ran statistics on what was left [12]. Dark photons need no magnetic field to make electromagnetic waves [13], so the same record serves both searches, with two different expected signatures.
The theory work is what opened the band. Widening the reliably modelled range by roughly thirty times [2] matters less than where the widening landed: the cavity's amplification sits inside the mass range the team wanted to probe, which is what corresponding author Atsushi Taruya means in calling the Earth-ionosphere cavity a natural resonator [6]. In that window the two candidates, both leading proposals for the unidentified quarter of the universe's energy content [3], are something like 19 to 21 orders of magnitude lighter than an electron [4], far below the reach of a laboratory magnet of workable size [5].
Two orders of magnitude [1] is the honest size of the gain reported in the Kyoto University announcement of September 5, 2026 [1], and it is measured against the previous best ground-based experiment [14]. Against the bounds inferred from X-ray observatories such as Chandra and NuSTAR, the new limit is competitive rather than superior, with the difference that those bounds lean on theoretical assumptions [15]. The thing this does not tell you is anything about axion masses outside the particular range examined [14].
The interesting asymmetry is geographic. The model predicts axion signals that vary with location and peak in Southeast Asia, while dark photon signals should arrive at nearly the same strength everywhere on the planet [10]. The decade of data came from Eskdalemuir [11], which is to say the axion bound was set a long way from where the model puts the strongest axion signal, and the several unexplained candidates [16] fall into the category the model says should be reproducible anywhere. One magnetometer cannot tell a planet-wide line from a local one that happens to be narrow and persistent.
Which is what makes those candidates testable rather than merely odd. If a dark photon is producing them, a second long geomagnetic record from a station far from Eskdalemuir should show the same frequencies at close to the same amplitude [10]. If it does not, they belong to the site or the instrument, and the bound is the whole of the result.
