---
title: "What Congress Got for Paying to Test the Sky"
canonical: "https://anchor.is/docket/issue/58/archive/article/65669/what-congress-got-for-paying-to-test-the"
publication: "The Docket"
publisher: "The Docket"
language: "en-US"
guide: "https://anchor.is/docket/llms.txt"
format: "Paying for Rain"
author: "Ivy Marlowe"
author_url: "https://anchor.is/docket/team/29"
issue: 58
section: "archive"
section_name: "Climate Past"
published: 2026-11-13
updated: 2026-11-14
description: "In 1891 Congress paid to blast the Texas sky for rain, and got no answer. What should today's cloud-seeding districts ask first?"
word_count: 1361
source_links: 12
---

# What Congress Got for Paying to Test the Sky

*By [Ivy Marlowe](<https://anchor.is/docket/team/29/markdown>)*

*November 13, 2026*

[Read on the website](<https://anchor.is/docket/issue/58/archive/article/65669/what-congress-got-for-paying-to-test-the>)

In March 1891, a single federal spending law answered drought in two ways. It set aside [$150,000 for seed](https://www.govinfo.gov/content/pkg/STATUTE-26/pdf/STATUTE-26-Pg1044.pdf) for Western settlers whose crops had failed the year before. Elsewhere in the same law, tucked into a $15,000 line for forestry, was money for trying to make rain with explosives. Agriculture Secretary Jeremiah Rusk let a Washington patent attorney, Robert G. Dyrenforth, spend up to $7,000 of it.

The larger sum covered what the drought had already taken. The smaller one was a bet that the next drought might be called off. That bet is the one I want to follow: what a public body thought it was buying when it paid to test the sky, and what it actually had once the test was over.

## Rain after battle

Much of the idea's evidence came from Civil War soldiers. In a letter dated February 26, 1871, a former Union general named Julius White described the fighting around Virginia's Weldon Railroad in August 1864. He remembered clear mornings and clear evenings, heavy rain at night, and officers arguing over whether their artillery had brought it on.

Picture days of cannon fire you can feel in your chest, then nights of soaking rain. It would be odd if the two hadn't gotten linked. For the tired, wet men who lived through it, the theory explained something they had just endured.

Edward Powers collected letters like White's in [*War and the Weather*](https://books.google.ba/books?id=e0oXAAAAYAAJ). He argued that heavy firing shook the atmosphere into giving up its moisture, and he offered battle after battle followed by rain. His book had no room for everything else: the same weather with no battle, and battles followed by dry nights. Nobody writes to an author to report that it didn't rain. So a list of hits with no record of misses can keep getting more persuasive without getting any more reliable. That habit isn't special to the nineteenth century; it's how most people notice patterns.

Powers lobbied for years, and in 1890 Congress put its first $2,000 toward "experiments in the production of rain fall," [folded into forestry](https://www.govinfo.gov/content/pkg/STATUTE-26/pdf/STATUTE-26-Pg282.pdf). The pressure behind it was real. Settlers had moved onto the High Plains and into the Southwest during wetter years, and the dry years that came after made those settlements look a lot less secure.

## Sixteen days of firing near Midland

The 1891 expedition set up at the C Ranch near Midland, Texas, under Dyrenforth, himself a Union cavalry veteran. The ranch's owner, the Chicago meatpacker Nelson Morris, supplied lodging and other help. Whatever Morris expected, the question "can people make rain?" was being asked on the land of someone with a large stake in the answer.

Firing began on August 9, according to Dyrenforth's [official report](https://tile.loc.gov/storage-services/service/ll/llserialsetpdf/02/90/0_/00/_0/0/02900_00_00/02900_00_00-005-0045-0000/SERIALSET-02900_00_00-005-0045-0000.pdf). He tried several things at once: balloons of hydrogen and oxygen meant to explode in the air, kites meant to carry charges up toward the clouds, and explosives set off on the ground. That kind of variety makes sense for an inventor. For anyone trying to read the results, it meant that if rain came, there was no way to say which device deserved credit, if any did.

Even the amount of rain was in dispute. For August 18, the expedition's meteorologist, George Curtis, recorded two hundredths of an inch. Dyrenforth described long, drenching rain and flooded roads. In that country a storm can soak one pasture and skip the next, so the two men may have been describing different places, and both could be accurate. The trouble is that nothing in the record shows they agreed beforehand on where rain would be measured or what would count as a result.

The last big barrage shows the problem most plainly. On August 25, the weather service forecast showers for northwest Texas. Dyrenforth knew, and argued about how much area the forecast covered. On August 26 it rained. By his account the rain was light at the ranch house, with an estimated inch to an inch and a half over pastures to the north. So the figure that sounded like success came from the place hardest to check, on a day when rain had been predicted anyway.

Dyrenforth called his findings "altogether preliminary," while suggesting that with more money, explosions might produce rain even in unfavorable conditions. The astronomer [Simon Newcomb](https://www.gutenberg.org/cache/epub/4065/pg4065-images.html) objected that a blast compresses air, which warms it, when rain needs moist air to cool. He did leave one door open, though: maybe smoke particles could give droplets something to form on.

## When "does it work?" became "what does it cost?"

In June 1892, William Hatch, chair of the House Agriculture Committee, [told the House](https://www.govinfo.gov/content/pkg/GPO-CRECB-1892-pt6-v23/pdf/GPO-CRECB-1892-pt6-v23-4-2.pdf) that his committee had at first opposed more funding. Then it heard from Rusk, his assistant secretary and Dyrenforth, and changed its mind. Hatch said he believed rain had been produced under some conditions. What remained was whether it could be done at reasonable cost.

The field results were still contested, and the committee was already talking about price, as though effectiveness had been settled. That shift didn't take any cynicism. The witnesses ran the program and described it in terms that let it continue, and in a dry year it's easier to vote for a next step than to sit with an open question.

Congress provided a [separate $10,000](https://www.govinfo.gov/content/pkg/STATUTE-27/pdf/STATUTE-27-Pg74-2.pdf). More trials followed, including one near San Antonio in November 1892 that an observer summed up as plenty of firing and no rain at camp. Then the [March 1893 spending law](https://www.govinfo.gov/content/pkg/STATUTE-27/pdf/STATUTE-27-Pg734.pdf) simply left the rainmaking line out. A little under half of the new money had been spent, and the rest went back to the Treasury. The surviving record doesn't explain the decision.

Paying for a test instead of for rain gave Congress something useful: it owed no farmer anything, and it could walk away. But the test was never designed to produce an answer anyone would have to accept. Believers could point to wet pastures, and critics to the forecast and the mismatched measurements. When the money stopped, the argument was more or less where it had begun.

## Five cents an acre

Public bodies still pay for this, with better science behind it. Texas now [licenses](https://www.tdlr.texas.gov/weather/weatherfaq.htm) weather modification, and the [Panhandle Groundwater Conservation District](https://www.pgcd.us/precipitation-enhancement), a local agency that manages underground water, funds cloud seeding out of its regular budget: $205,182.74 in 2024, about five cents an acre.

Cloud seeding doesn't make clouds. Planes release tiny particles into clouds that already exist, giving moisture something to gather on. Any benefit depends on a chain of things going right. A suitable cloud has to show up, seeding has to make it drop more, the extra has to fall where it helps, and for a groundwater district, some of it has to soak into the ground. The district knows its cost to the cent. The benefit can get smaller at each of those steps.

Something like the record of misses Powers never kept now exists. In Wyoming, researchers used a [random draw](https://doi.org/10.1175/JAMC-D-17-0335.1), across 118 experimental cases, to pick which of two neighboring mountain ranges got seeded. Because chance chose the target, the unseeded range showed what the same weather did on its own, and that comparison couldn't confidently detect an effect. Computer modeling of the same project found a small one, around 5 percent more precipitation from seeded storms, and much less over a whole year. The [World Meteorological Organization](https://wmo.int/content/wmo-statement-weather-modification), the UN's weather agency, puts the strongest evidence in winter clouds over mountains and warns against assuming results carry over anywhere else, including flat country like the Panhandle.

Five cents an acre may well be worth a modest, uncertain gain. What the 1890s suggest is a question to settle before the next promising season instead of after it: what result would lead the district to expand, continue or stop? And when the answer comes back small and hedged, will the next budget take it at that size, or jump ahead, the way Hatch's committee did, to price as the only question left? I'd ask the same of anything a district or a household buys against the weather when the payoff can't be seen in a single year.

## More

- [Climate Past — Issue 58](<https://anchor.is/docket/issue/58/archive/markdown>)
- [Issue 58](<https://anchor.is/docket/issue/58/markdown>)
- [Publication guide](<https://anchor.is/docket/llms.txt>)