1st Person · Photographic Pursuit Record · Garnett, Kansas · April 26, 1884

The First Tempest Photographer

He tracked weather over southern Colorado for years before anyone had a name for a man who did that. On April 26, 1884, he predicted a tornado, followed it from Colorado into Kansas, and photographed it west of Westphalia — minutes before A. A. Adams photographed the same funnel from near Garnett. Adams became the record. Caleb Sloane did not. His plate showed too much.

Pueblo, Colorado Territory / Westphalia, Kansas · 1884–1926 — Attributed to Caleb Sloane, Weather Observer & Railroad Photographer

I am going to set this down because I am old enough now that the people who dismissed it are dead, and the ones who replaced them do not know it happened, and I find I prefer the record to the silence. My name is Caleb Sloane. I was born in Pueblo in 1837. I have spent the better part of fifty years watching the sky over southern Colorado, keeping records on what I found there, and occasionally following weather that seemed worth following. In the spring of 1884, I followed a tornado from Colorado into Kansas. I made a photograph of it. The photograph was rejected. I believe I understand now why it was rejected, and that understanding took me the better part of forty years to arrive at, and it is not the explanation I would have preferred.

The Work

The weather over southern Colorado does not behave the way it behaves in the eastern states. The mountains change the rules. They redirect storms, funnel them, amplify them, and create conditions that the Signal Corps in Washington had no framework for and therefore often predicted wrong. I knew this because I had been keeping daily observations since 1868 — temperature at three intervals, barometric pressure morning and evening, humidity by sling psychrometer, wind direction and estimated velocity, cloud formation by the Howard classification system with my own extensions for what I called contact cloud: the kind that forms along the front face of the mountains and comes in low and fast in ways that do not match the northern European models the Signal Corps was trained on.

I was a railroad photographer by trade. The Santa Fe line paid me to document depots, bridges, churches, and settlement towns — the visual evidence that the frontier was becoming permanent. I did that work reliably and I do not want to diminish it. But the work I cared about was the sky.

I carried an aneroid barometer in the field at all times. At my Pueblo station I kept a barograph — a self-recording instrument that produces a paper roll showing pressure over time — and I retained every roll dating to 1871. I had a thermometer array at three elevations in the Beulah corridor. I had a sling psychrometer, an anemometer, a wind vane, a rain gauge, and a set of Howard's cloud charts with twelve years of my own marginal annotations. I had informal correspondents along the Santa Fe line, men who sent me conditions when the weather turned interesting: a stationmaster at La Junta, a farmer near Rocky Ford, the telegraph operator at Trinidad. From this I built tables — pressure-drop rates against storm type, wind shifts preceding violent weather, cloud-base height against hail probability.

I submitted these observations to the Signal Corps. They acknowledged receipt. They did not ask follow-up questions.

I kept a prediction ledger. I recorded my forecast before each event and the outcome after. My accuracy on severe weather in the Pueblo corridor, over a fifteen-year period, was roughly three in four by my own count — and I gave myself no credit for storms that arrived half a day late or moved north of my predicted track. The Signal Corps, in the same period, predicted roughly one in three. I know this because I kept their bulletins alongside my own.

I say this not to make myself sound capable beyond reason, but to establish the correct context for what follows. I did not pursue the Garnett tornado by accident. I did not stumble into Kansas and happen to find a funnel in the sky. I predicted the conditions, tracked the system for four days, and moved east because my instruments told me to move east. That is the fact of the matter.

The Prediction

The system that produced the Garnett tornado was visible in my barograph rolls from the 22nd of April onward. A pressure depression moving northeast across the territories, deepening faster than the Signal Corps bulletins allowed for. By the 24th, the dew point along the Arkansas River corridor had risen to a level I had learned to associate with severe convective development. By the morning of the 25th I had a wind shift at Trinidad — south-southeast at approximately nine miles per hour, backing to east-southeast by noon — that matched the pattern I had documented preceding three previous significant tornado events in this territory.

I left Pueblo on the morning of April 24th by rail east on the Santa Fe line. I had my field kit: the aneroid barometer, the sling psychrometer, the camera and dry plates, the logbook. I was not racing. I was following a probability at a measured pace. The sky to the southwest, as I departed, was already beginning to build in the way I had learned to read. When the air has that quality — pressurized, still, with the cloud base showing that particular greenish underside — I have found it wise to be moving toward it rather than waiting for it to arrive at my door. Weather observed from your own doorstep is weather that has already chosen its terms. I preferred to choose mine.

East

By the evening of April 25th I was in eastern Kansas, working west of Garnett on a wagon hire. The sky to the west was doing what I had predicted it would do. The pressure had dropped eleven-hundredths of an inch in six hours. The cloud base was low, green at the margins, moving with a deliberateness I have never found a better word for than intent. The air had the specific stillness that precedes violent movement — a stillness I had documented enough times to recognize without looking at an instrument.

I made camp that night knowing the storm would come the following day. I had three unexposed plates. I intended to use them.

I want to be clear about the fear, because I have been asked about it since, and the answer matters. Yes, there was fear. Any sensible man is frightened by a thing that can kill him without effort. But the fear was not the primary condition. The primary condition was attention. When you are following weather with a camera and a barometer, you are not thinking about danger the way a man thinks about it when it surprises him. You are calculating distance, angle, light quality, exposure time. The fear runs underneath all of that and does not interrupt it.

The Exposure

The funnel formed approximately fourteen miles west-southwest of Garnett at around eleven in the morning. It was a rope tornado — narrow, elongated, moving northeast at perhaps fifteen miles per hour. It was visible for a considerable period. Long enough for me to set the camera, assess the light, and make three exposures.

I was positioned near a crossroads west of Westphalia, closer to the rope than the town. I had a fence line to my left and open field to the right. The storm light was extraordinary — that particular yellow-green cast that comes before violent weather, the quality that makes everything look as though it is being documented even before the camera opens.

I exposed the first plate at approximately eleven-fifteen. The second, five minutes later. The third near the end of the rope's visible phase, as it began to thin and lift toward the base of the larger storm. I believed, in that moment, that I had made the first tornado photograph. I had no way of knowing that A. A. Adams was doing the same thing from near Garnett, some minutes behind me — photographing the same funnel as it continued northeast toward town. I did not know Adams then. I learned of him later, when his plate was accepted and mine was rejected, and I want it on the record that I do not hold Adams responsible for what happened. He photographed a tornado and the photograph was good. That is not a fault.

I caught the weather being remembered. That is all I know for certain. That, and that I was there first, and that the record preferred otherwise. — Attributed, C. Sloane — late-life deposition, Pueblo, Colorado, 1926

What I Found in the Development Room

I developed the plates in Pueblo two weeks later.

The first plate: a partial exposure, blurred at the edges, the funnel at distance. Usable but not exceptional. The second plate: better. The rope funnel clear and centered, the storm light correct, the Kansas horizon flat beneath it. This was the plate I intended to submit.

The third plate was the one that ended the matter.

I developed it in sequence, in my usual manner, with no expectation of anything unusual. The funnel was present, somewhat thinner than in the second plate, the storm light still correct. The fence line was present. The road was present. The figures were present.

My first response was damage. The glass was cracked. The emulsion had failed. There had been a light leak. These are the explanations a photographer reaches for immediately because they are the explanations that fit inside a world that makes sense.

I examined the plate under a loupe.

A light leak floats on an emulsion. It creates soft-edged brightness that does not interact correctly with the underlying image. These figures had hard edges. They cast shadows in the correct direction, consistent with the storm light. Two of them were standing behind the fence line — behind it, meaning the fence posts stood in front of them, partially occluding their bodies, which is what real objects do to real people. They were inside the exposure, not on top of it. Damage floats. These figures stood.

I counted seven.

Plate Annotations — Third Exposure — April 26, 1884 — C. Sloane, Pueblo

Figure — Road, Center LeftA man standing on the road with what appeared to be a camera on a short support. The support was wrong — too low, wrong proportions. The camera itself had no bellows visible and no focusing cloth. It was an object I did not recognize as a camera except by the posture of the man operating it. I have not identified the instrument.

Figure — Road, Right of CenterA woman in a garment I cannot identify by material or construction. The cut was unlike anything I had seen in any catalogue or in any town I had photographed. The material caught the storm light differently than wool or cotton catches light. Something about the way it held its shape in the wind was wrong for 1884.

III

Figure — Road, ForegroundA boy, perhaps ten years of age, standing in the road between the fence line and my camera position. He was not looking at the tornado. The natural position for anyone standing in a Kansas road with a rope funnel to the west is to be looking at the rope funnel. He was not. He was looking at my camera. I have spent considerable time deciding whether that distinction constitutes a meaningful difference. I believe it does.

Figure — Field, Near Road EdgeA man holding a flat rectangular object in both hands. Small — smaller than a book. The object was producing its own light. Not reflected storm light. Not a lamp or lantern. The light came from within the object itself and was directed upward toward the man's face. I have turned this particular detail over for forty years. I have not arrived at an explanation consistent with 1884.

V & VI

Two Figures — Behind the Fence, LeftTwo figures standing behind the fence to the left of the road, partially obscured by the fence posts in the correct way — meaning the posts stood in front of them, not the reverse. They were looking toward the tornado. This was the most ordinary behavior on the plate, and I found it the most disorienting because of that. Two people watching a storm is a thing that makes sense. Everything else about the plate does not.

VII

Figure — Far Right Edge, Road Meets FieldAt the far right edge of the frame, at the place where the road meets the open field: a person. I cannot determine sex or age. They were standing still. They were not looking at the tornado. They were not looking at the fence. They were not looking at the other figures. They were looking directly at the camera. At me. I have looked at this figure many times across many years. It has not changed its posture. It has not looked away.

The Submission and Its Outcome

I submitted the second plate — the clean exposure — to the Kansas photographic society through correspondence. I included a written description of the third plate and an offer to forward it for examination. The response characterized the figures as evidence of either plate damage or deliberate manipulation. Both plates were returned without further inquiry.

I wrote to the Signal Corps. I described the figures in the third plate in detail. I asked whether any other observers had documented anomalies of this kind in storm photography. The response acknowledged receipt and stated that the Corps had no record of such phenomena, and that the most probable explanation was chemical degradation of the emulsion combined with stress from field conditions.

I want to note, for the record, that A. A. Adams's plate was accepted in roughly the same period without, as far as I have been able to determine, anyone questioning whether it constituted a first. His plate showed a tornado. My plate showed a tornado and impossible people. A man whose plate shows a tornado is an observer. A man whose plate shows a tornado and impossible people is something else. That is the arithmetic of the record, and I have made my peace with it, and I want that peace on the record too.

What I Believe Now

I am eighty-nine years old. I have kept weather logs for more than fifty years. I have been wrong about storms and I have been right about storms, and I know the difference between the two well enough to know which category the Garnett prediction falls into. The prediction was correct. The plate was not damaged.

What I have come to believe, after a long time with the question, is this: the Garnett tornado was the first storm to be photographed. The act of photographing it — the act of making it a fixed record — created something. Not a rupture. Not a mystical event. Something more like what happens when you mark a place on a map. The place was always there. The mark gives it a different kind of existence.

The people on my plate were not traveling through time. They were looking back at it. Every person who would one day stand in a field with instruments and watch a tornado and document it — they were present in the light, gathered at the moment when the first such record was made. My camera did not photograph the future. It photographed every future eye that would come looking back at this moment.

The tornado did not open time. The photograph did.

Whether that is an explanation or a different kind of confusion, I cannot say with certainty. I am a weather observer. I have always worked from what the instruments show. The third plate is an instrument. It shows what it shows. I have reported it accurately, and the rest is not my problem to solve.

Obscura Note A. A. Adams's photograph of the April 26, 1884 Garnett, Kansas tornado is recognized by the National Oceanic and Atmospheric Administration as the oldest confirmed tornado photograph. Adams documented the rope funnel from near Garnett minutes after Sloane made his first exposure west of Westphalia — the same funnel, the same morning, moving northeast toward town. The recognized modern storm-chasing lineage begins considerably later: Roger Jensen began pursuing severe storms in 1953; David Hoadley, widely regarded as the pioneer of organized storm chasing, began in 1956. Sloane predates both by nearly seventy years. Sloane's second plate was submitted and returned. His third plate, which he described in two letters of submission dated 1884 and 1885, has not been located. His weather documentation archive — twelve logbooks, illustrated cloud studies, barograph rolls dating to 1871, and a draft almanac manuscript — was held by his family until approximately 1951 and has not been definitively traced since. The Obscura designates the matter: open.

Obscura Classification — 1st Person Desk — Photographic Pursuit Record

The account attributed here to Caleb Sloane is reconstructed from archive fragments held in the recovered correspondence files, including two letters of submission (1884–1885) and a late-life deposition recorded by his granddaughter in Pueblo, Colorado, 1926. Sloane was eighty-eight or eighty-nine years of age at the time of dictation. The deposition runs to eleven pages. The Obscura holds a transcription; the original is not in the collection.

Sloane's weather prediction claim is plausible on its face. A skilled observer using period instruments — aneroid barometer, sling psychrometer, barograph, thermometer array, telegraph correspondence — could reasonably track a developing severe-weather system from Pueblo east across the Arkansas River corridor and into Kansas. The instruments he described were in common use among Signal Corps observers by the 1880s. His prediction methodology, as reconstructed from the deposition, reflects a systematic approach that would not be formalized in the meteorological literature until decades later.

The third plate has not been found. The second plate was returned to Sloane and is also not in the archive. Sloane's name does not appear in any Signal Corps roster, Kansas photographic society record, or Santa Fe railroad payroll from this period. This is consistent with the position of an independent contractor and informal correspondent, which is how he describes himself. The absence from the official record does not constitute an absence from the field. That rule applies here as it applies everywhere in this archive.