Dell climbed down slowly. His sandy gray hair lay combed across the thinning place, and his thumbs found his belt out of habit before his hands dropped to his sides.

He looked past me at the west pasture. The ponds were green in a summer that had cracked the stock ponds around the township into plates of mud. Dust from the lane drifted behind his truck and thinned in the hot air.

“I was wrong,” he said.

Three words. He did not add anything to them.

I let them sit between us. Then I turned toward the nearest pond.

“You want to see how it holds,” I said.

I walked him down to the water. The soil at the edge was dark, and the clay showed smooth beneath the waterline. I brought out the weighted string and lowered it into one of the test pipes. Dell watched the string go down, then leaned toward the water and put his hand against the wet clay bottom.

The old pond had held through the drought. The water level had fallen a few inches, not a foot a week. The deep supply was not feeding it. The rain had filled the ponds, and the buried clay had kept the water from draining away.

I explained it in plain terms: the clay shelf was eighteen inches thick, packed while it held the right amount of moisture. Water stayed above it, perched on the clay and separate from the deep water table. The sun could take some. The ground beneath could not swallow the rest.

Dell listened. He had spent nineteen years behind the co-op counter, telling farmers what their land and their numbers could bear. He knew what this drought cost them by the gallon. Now he was standing beside water that had stayed put.

He asked whether he could copy the pages of the 1937 manual and leave a stack at the co-op register. Farmers came to that counter sooner or later, and he thought they ought to see the instructions.

I brought him the pamphlet the next morning. He copied the pages about clay lenses, moisture, and compaction. By fall, farmers who leaned on the counter were leaving with those sheets and the suggestion to visit my ponds.

My father had watched Dell’s truck return down the lane. He stood by the fence with his brass-cased pocket watch turning slowly in his fingers. When I came beside him, he looked toward the water.

“You were right,” he said.

He did not make a ceremony of it. He had never doubted that I could read the ground. He had needed the numbers to settle the question of what we could risk. The numbers had come one pond, one measurement, one dry season at a time.

After that, the method traveled in the way most useful things travel: one farmer telling another what had worked.

Vernon Aldis dug a pond that fall using Dell’s copied pages. It held the next August while his old ponds fell. In 1994, I walked two other farms. By 1995, I was visiting five, and a man came down from Guthrie County because he had heard about the ponds at a sale barn.

I did not charge to walk a neighbor’s ground with the probe. If a pond held, I accepted a day of the farmer’s labor or the goodwill that came from being useful. I would not take money just to tell someone what lay beneath their dirt.

My own place changed slowly. The pasture over the perched water could support more cattle because they did not have to walk to dry troughs. I put the extra into more ground, including forty acres to the south where the probe found another buried channel. I built three ponds there and let the system season.

By 1995, the pasture carried forty more head than it had any business carrying, and the cattle came off it heavier than the county average. I expanded only when I had read the ground first and had the cash for it. The method was not just a way to keep water. It was a way to keep from borrowing against a future the weather could take away.

The outside world found us late and by accident. In the spring of 1997, a writer named Coyle, who worked for a soil and water trade monthly out of Ames, came to look at the ponds. He had heard about a woman building ponds from an old dust-bowl manual and finding water in a dry county.

He spent the day walking the run. He read the levels himself. In the green pasture, with the county dry through June, he shook his head and took notes. The article ran that fall.

The county board of supervisors read it. They also read the letters coming into the county office asking whether anyone was going to do something with the method. Years earlier, the assessor had sent me a notice warning that the fourteen basins might be flagged as a nuisance. Now the board sent me a letter on county letterhead.

It was an invitation.

The county owned low ground along two road ditches and a piece of parkland where a failed recreation pond had dried into a cracked hollow and become a mosquito bed. The board wanted to know whether I would consult on and build a system of perched-aquifer ponds for public stock-water access and wildlife. The amount named for surveying, designing, overseeing, and building the system was $47,500.

I read the letter in the yard. The paper was official, but I had learned what paper could not do.

A letter is not a pond.

I carried it inside, but I did not pretend the invitation had already proved anything. There were sites to test, clay to find, water to hold, and a public project that could make people doubt the method again if I hurried it or cut corners.

In the spring of 1998, I staked the first public pond in a road ditch. A man from the county office came with a plat map to mark the boundary. He was the assessor, or from the same office that had sent the warning. He held one end of my tape while I drove the probe into the ditch bottom.

I called out the depth. He wrote it down.

We did not discuss the old notice. There was work to do.

I built the public ponds as carefully as the ones at home. We reclaimed the failed recreation pond first, dug out its porous bottom, and relined it with blue clay. The clay was hauled and packed in shallow passes at the right moisture. That fall, the pond filled and held. People in town watched the old mosquito bed become a lake.

I oversaw four ponds along the road ditches over two seasons. I checked each load of clay; two came too dry, and I sent them back. A public pond that failed would set the method back twenty years. I would not let speed put the work at risk.

When the county paid the final part of the $47,500, I put all of it into land and cattle. I did not borrow against the future. Through the good years of the late 1990s, I expanded slowly and only on ground I had probed. The figures were steady and unspectacular, the way I wanted them.

The home place carried more cattle every few years. In 2004, the forty acres to the south was joined by another eighty. I read the ground before I bought it, then dug and sealed ponds there before the next dry spell could make the decision for me.

By 2008, I had thirty-one ponds across my own land and the county’s. More than a hundred others had been built to the specifications, across four counties. The copied pages had gone a long way from Dell’s counter.

Then 2012 came.

It was the worst drought the country had seen in eighty years. Corn burned before it tasseled. Rivers ran to sand. Deep wells that had never failed pulled air. The regional aquifer dropped faster and farther than the hydrologists had expected.

My ponds held.

They did not draw from the great deep aquifer. They sat on their own shelves of clay, fed by rain and drained by very little. Every pond built to my specifications remained green and full while the deep wells failed around them. The public ponds held, too. The town filled stock tanks from the reclaimed recreation pond, and wildlife came to the road-ditch ponds from miles of burned country.

The method had worked in the drought of 1988 and again in 1993. In 2012, it held across a much larger stretch of land, through a season when even the deep groundwater could not keep up.

That winter, the county rewrote its water and conservation guidance. The perched-aquifer pond became recommended county practice, with my specifications printed in the appendix under my name.

I was fifty-five that summer, according to the account of those years. I walked less ground than I once had. At the end of a long day, the probe felt heavier in my hand. The drives to farms farther away began to tell on me.

The numbers were in notebooks and manuals by then, but the feel of the ground was harder to pass along. The probe changed when it reached the clay. The resistance traveled through the steel into the bones of my wrist. A person could learn what the manual said, but the reading of the ground had taken me years.

I began to wonder who would carry that part forward.

At sixty-seven, I sat more on clear days. The probe hung by the door, worn bright and smooth where my hand had held it. On the kitchen table, I kept maps of the county with my ponds marked in the recorder’s ink: thirty-one of mine, and more than a hundred built to the same specifications.

My father was gone. He had lived to ninety-one and had seen the 2012 drought prove the ponds right again. Near the end, sitting on the porch and looking toward the water, he said the same thing he had said when we first spread the maps across the kitchen table. He had been waiting for the numbers. They had come.

One Tuesday morning that spring, a young woman drove in from the county road. She was twenty-four, with a soil-science degree and a laptop bag on the seat beside her. She was the granddaughter of my father’s brother.

She sat across from me at the kitchen table and opened the laptop. The screen showed the county in bands of color: elevation, soil types, and the buried channels the glaciers had left behind. The model could trace a watershed from satellite data without driving a probe into the ground.

She pointed to a low corridor beneath failing pasture eight miles east. She thought there was deep clay there, a forgotten channel nobody had dug.

If the model was right, she said, there might be room for forty ponds. Perhaps a hundred.

She wanted me to check the first site with the probe. The computer could read the county at once. It could not feel the clay.

For a long moment, I looked at the colored map and then at the notebook open beside her hand. Years earlier, I had carried my own numbers to this table and waited three days for my father to consider them. The probe was still by the door.

“Get the probe,” I said. “It’s by the door.”

We went out together. The map stayed open on the table, its colored channel running beneath a stretch of pasture neither of us had tested. The old steel handle lay in the light beside the door, waiting for a hand to pick it up.