The Drift
LV27 · Michael S. Moniz, after H.P. Lovecraft
Story catalogue · Reading guide
• • •
The caesium does not lie. I want to begin with that, because everything that follows depends on it, and because the temptation — which I have felt and which I am resisting — is to suggest that the instruments failed, that the data was corrupted, that some environmental factor intervened in a manner too subtle for our diagnostics to detect. The temptation is to explain this away. I am a metrologist. Explaining away is not what I do. I measure, and I report what I measure, and what I measured in the autumn of 2027 was a drift in the caesium fountain standard at three remote timing stations that should not have drifted, that had no mechanical or environmental reason to drift, and that drifted in a pattern I did not expect and cannot account for.
My name is Elin Pryce. I am a senior metrologist at the National Physical Laboratory, Teddington. My work concerns the maintenance and verification of SI-traceable time standards — which means, in language that does not require a physics degree, that I make sure the clocks are right. Not wristwatches. Not mantle clocks. The clocks that underpin everything: GPS satellite timing, financial transaction timestamps, telecommunications synchronisation, power grid phase-locking. Every system that requires two events to occur at the same time, or in the correct sequence, or at a known interval, depends on clocks that I and my colleagues verify to an accuracy of approximately one second in three hundred million years.
I mention the accuracy because the drift I observed was within it. Barely. The stations drifted by fractions of a nanosecond — quantities that would be invisible to any instrument less precise than the ones I operate, quantities that have no practical consequence for GPS or telecommunications or any civilian system, quantities that should not have alarmed me and that I cannot explain having alarmed me except by saying that I knew the drift was wrong before I confirmed it with the data, and that knowing has been the defining feature of my professional life, and that I have never once been incorrect when I have known.
I come from a family that has always been good with time. My grandmother, who grew up in a village near Llandeilo in Carmarthenshire, could tell you the hour within ten minutes by the quality of the light, and she could tell you the season within a week by the angle of the shadow on her kitchen floor, and she attributed this to farming and to God and not to anything more complicated, and I loved her for the simplicity of that attribution, and I have spent my career in a laboratory that proves, daily, that the universe is not simple.
The three stations that drifted were geographically dispersed. One in New Mexico, near a decommissioned military facility. One in the Aegean, on an island whose name I will use only the Greek designation for: VS-site Delos. One in mid-Wales, near a village I could not find on the ordnance survey map that covers the region, though the timing station itself is maintained by a telecommunications company and is perfectly well documented.
The drift at each station was not identical but correlated. The New Mexico station drifted first, by 0.4 nanoseconds over a seventy-two-hour window. The Aegean station followed, same magnitude, with a twelve-hour lag. The Welsh station drifted last and least — 0.2 nanoseconds, barely above the noise floor — but the correlation was unmistakable. Three stations, three locations, three drifts of the same character occurring in sequence.
I should explain what causes clock drift, because the explanation is what makes the absence of explanation significant.
Caesium fountain clocks measure time by counting the oscillations of caesium-133 atoms — specifically, the radiation emitted when the atom transitions between two hyperfine energy levels. The frequency of this radiation is, by definition, 9,192,631,770 cycles per second. That number is not an approximation. It is the definition of the second itself. When we say a caesium clock drifts, we mean that the local measurement of that frequency has deviated from the expected value, which means either the instrument has malfunctioned or the atoms have changed.
Atoms do not change. The transition frequency of caesium-133 is a fundamental constant — it depends on the electromagnetic force, the nuclear structure, the mass of the electron. These are not variable quantities. They are the architecture of matter. If they changed — if the fine-structure constant varied, if the weak interaction coupling shifted, if the electron mass altered by even a fraction — the consequences would propagate through every physical system: chemistry, biology, material science, everything. The constants are constant. That is what makes them constants.
The drift I observed is consistent with a localised variation in the fine-structure constant.
I have written that sentence and I am going to leave it there and let you understand what it means. The fine-structure constant — alpha, approximately 1/137 — governs the strength of the electromagnetic interaction between charged particles. It determines atomic structure, chemical bonding, the behaviour of light in matter. If alpha varies — even locally, even by parts per billion — atoms behave differently at that location. Clocks tick differently. Not because the clock is broken. Because the physics is different.
I did not include this interpretation in my report. My report attributed the drift to correlated environmental interference — possible seismic microactivity, possible thermal gradient anomalies, possible electromagnetic contamination from an unidentified source. These explanations are not credible. I know they are not credible. My colleagues will know they are not credible. But a metrologist who reports that the fine-structure constant has varied locally at three geographically dispersed sites is a metrologist who has departed the boundaries of her discipline, and I have not yet decided to depart those boundaries, and the decision is what keeps me writing this instead of filing it.
I ran the geographic correlation because I could not stop myself. This is the part I find most difficult to account for — not the data, which is what it is, but the compulsion. I am not a person who follows hunches. I am a person who measures. But the three stations occupied positions on the globe that my mind — or something beneath my mind, something in the architecture of my attention that I have carried since childhood and that my grandmother carried before me and that I have never named — recognised as related. Not by latitude. Not by longitude. Not by altitude or geology or climate. Related in the way that notes in a chord are related: each one makes the others legible.
I plotted the three stations on a projection. I added four more stations from the historical drift records — stations that had shown similar anomalies in previous years, anomalies that had been attributed to equipment aging and dismissed. Seven points. I plotted them and I looked at the distribution and I felt — the word is felt, and I am aware that it has no place in a metrological report, and I am writing this and not a metrological report — I felt the geometry.
The seven stations were not randomly distributed. They were not clustered. They were arranged in a pattern that had the properties of a lattice — regular spacing, angular relationships that repeated at consistent intervals, a structure that was too ordered to be coincidence and too dispersed to be caused by any single environmental factor. The lattice did not correspond to any geological feature I could identify. It did not correspond to tectonic plate boundaries, to volcanic hotspots, to oceanic ridges, to any physical system that might produce correlated effects at these distances.
It corresponded to something else. Something I do not have a name for but that I recognise the way my grandmother recognised the hour by the light — not through analysis but through a faculty that precedes analysis, that is older than my training, that lives in the part of my attention that has always known when a clock is wrong before the diagnostics confirm it.
The lattice is real. The drift is real. The correlation is real. The variation in the fine-structure constant — if that is what it is, and I believe it is, and my belief is not scientific but it is honest — is localised at points that form a geometric network across the surface of the earth.
I have not filed this. I have not told my colleagues. I have written it here, in a document that is not a report and that will not be submitted, because the writing is what I need — the act of making the measurement legible, even if the only audience is myself, even if the document goes into a drawer and stays there.
A colleague — Dr. Marsh, who works in our geodetic timing section — stopped at my office yesterday. He does this occasionally. He is a quiet man whose questions have a specificity that I have always found slightly misaligned with his stated research interests. He asked whether I had noticed anything unusual in the remote station data. I said I had not. He looked at me with an expression I cannot characterise except to say that it was the expression of a man who knew I was lying, and who was not surprised by the lie, and who would file the lie alongside whatever other data he was collecting about me.
I have been measured. I recognise the experience because I have spent my career doing it to caesium atoms.
The drift continues. I checked this morning. The New Mexico station has accumulated 1.2 nanoseconds of unexplained deviation. The Aegean station is at 0.9. The Welsh station is at 0.6. The magnitudes are small. The trajectory is not. If the drift continues at the current rate — and I have no reason to believe it will not continue, because I have no explanation for why it started, which means I have no mechanism by which it might stop — the deviation will cross the threshold of operational significance within eighteen months. At that point, the GPS corrections at those locations will require adjustment. The adjustments will be small. They will be automated. They will propagate through the timing network without anyone noticing, because the system is designed to absorb corrections, and the corrections will look like routine maintenance.
But they will not be routine. They will be adjustments to a physics that has changed at specific locations for reasons that the system does not contain. And the adjustments will be visible in the correction data to anyone who looks, and someone will look, because the correction data is not classified, and the locations are not random, and the geometry of the lattice is not something that can be explained by equipment aging or thermal gradients or seismic microactivity.
Someone will look. They will see the pattern. They will not know what it means. But they will see it, and they will share it, and the sharing will do what the drift itself is doing to my clocks — it will move the measurement beyond the boundary of what the current framework can contain.
I am a metrologist. I measure. I report. I do not interpret. But the measurement is telling me something that the framework was not built to hold, and the framework is what I have spent my career inside, and the measurement does not care about my framework, and the caesium does not lie.
The drift is 1.2 nanoseconds in New Mexico. It was nothing, three months ago. It is almost nothing now. It will be something soon. And the geometry of the lattice will be visible to anyone with the data and the patience, and I have the data, and I have lost the patience, and I am writing this instead of filing it, and the writing is a clock that I cannot verify against any standard I trust.
My grandmother could tell the hour by the light. I can tell that the light is changing. I cannot tell the hour.
• • •
Fiction by Michael S. Moniz. Produced with AI assistance in writing and editing under the author’s direction. H.P. Lovecraft is the literary influence; references to Lovecraft in source credits identify the inheritance or AI writing persona.
Edition and source
Selected final: LV27_THE_DRIFT.docx. Selection basis: only version. Story prose preserved; reading-edition title and credits standardized, source word-count notes omitted.