Some particles facilitate the process, however, through a process called nucleation. A little mineral crystal can act as a template, coaxing water molecules on its surface to organize into the hexagonal lattice of an ice crystal. There is plenty of microscopic junk that allows the water in a puddle in your backyard (or even in your ice-cube tray) to freeze just below 32 degrees. But a cloud droplet that is just slightly wider than a red blood cell may contain only one such particle. In order for this particle to nucleate ice crystals, it needs to have just the right shape and give off minute attractive and repulsive atomic forces in just the right places so that the H’s and O’s in those H₂O molecules stick to the particle in the right hexagonal pattern.

Clouds at –30 or –35ºF are often entirely liquid because they do not contain any efficient ice-nucleating particles. And yet scientists also see clouds that are much warmer, even 10ºF, that are full of ice and gushing out rain or snow. These clouds obviously contain something that efficiently nucleates ice at those much warmer temperatures.

For decades scientists could not put their finger on what that mystery particle was. In the mid-1960s Gabor Vali, a Hungarian-born Ph.D. student in physics at McGill University in Montreal, devoted most of his waking hours to looking for it. Vali spent months collecting snow and rainwater by the gallon. He brought it to the lab and squeezed it by syringe, drop by drop, onto sheets of aluminum foil, 100 drops per sheet. He cooled the sheets by 2ºF per minute, taking a photo every 30 seconds. Later he projected the photos on a wall and looked to see the temperature at which each drop had frozen.

Vali did 10 experiments a day, five or six days a week, for several years—“many hundreds of thousands of drops,” he says. Each drop froze at a different temperature depending on what specks were floating inside it. A few drops froze at 23ºF, but not frequently enough to explain how efficiently ice crystals form in warm clouds. Vali was missing something.

Then one day on a whim, he grabbed some muddy snow from under his kids’ swing set. It was sloppy science: Researchers around the world were already testing powdered minerals for their ability to form ice, but they used minerals that were sterile and pure. Vali’s quick and dirty experiment produced an unexpected result.

Water drops from the dirty snow froze at 23 degrees much more often than his cleaner samples had. When Vali added rotting leaves to the mix, some drops formed ice at 28 degrees, the highest temperature he had ever seen. The leaves’ potency increased as they turned brown, suggesting that something growing on them was causing ice formation. By this time Vali had finished his degree and moved to the University of Wyoming. There, in 1972, a graduate student named Richard Fresh managed to identify Pseudomonas syringae, a bacterium from the rotting leaves that was triggering the ice.

The discovery had big implications for agriculture. Billions of dollars of crops worldwide are damaged by frost each year. Steven Lindow, a Ph.D. student at the University of Wisconsin who independently discovered P. syringae’s ice-nucleating capability around the same time Vali and Fresh did, found that frost damage is not caused by temperature alone; syringae bacteria on leaves actually use ice crystals as crowbars to rip open plant cells and grab their nutrients. To prevent this, some growers of fruit trees and vegetables now routinely apply bactericides to crops before frosts.

During the 1980s Lindow managed further to isolate the gene that allows syringae to form ice. That gene codes for proteins that spread out on the bacteria’s surface in just the right arrangement to pull water molecules into their hexagonal crystal formation. Since then, four other bacteria have been found that nucleate ice at extremely warm temperatures. Although they are not all related, they share the same gene for ice nucleation.

Even as the importance of biological ice nucleation was being recognized by agricultural scientists, it still wasn’t embraced by atmospheric scientists, who stuck by the traditional view that soot, or sea salt, or some as-yet-unidentified mineral in dust was seeding ice in clouds. There was no real sense at the time of the number and diversity of microbes living in the atmosphere. Vali tried to collect ice-forming cells from the lowest few hundred feet of the atmosphere but failed to find anything. And no one was systematically looking thousands of feet up, where clouds actually form. People doubted that enough cells could get high in the atmosphere to make rain or snow.

“The field collapsed,” says David Sands, a bacteriologist now at Montana State University in Bozeman, who began working with syringae in the 1960s. “The people who were in it couldn’t get grants anymore.” Vali shifted his focus back to cloud physics. Sands continued to study syringae but concentrated on crop diseases, where he could still get funding. The idea that a microscopic, floating biomass was influencing the world’s weather was just too weird for most atmospheric scientists.

In 2005 Sands decided the time was ripe to reopen the investigation and find out whether bacteria really do manipulate the weather. Genetic techniques had advanced considerably in the last two decades, so that it was possible to identify even a few ice nucleators like syringae out of thousands of microbes that might inhabit a cloud. Sands contacted Brent Christner, a postdoc at Montana State who had cut his teeth searching for traces of life in 100,000-year-old Antarctic ice, and invited him to expand his search for life into the clouds.

Christner, now at Louisiana State University, started with a simple experiment. He collected snow from mountains around the world and repeated Vali’s frozen-drop experiment to look for invisible particles that spawned ice. Once Christner found the drops of water that froze above 20ºF, he treated them with an enzyme that kills bacteria but leaves other particles alone. When he cooled those droplets back down, up to 85 percent of them no longer froze (pdf).

It was indirect but tantalizing evidence that bacteria were helping the water freeze. But it left an open question: Were airborne microbes forming ice in the clouds, or were they merely getting mopped up by snowflakes as they fell to Earth? The only way to find out was to fly into a cloud to see what was happening from moment to moment.

Prather unwittingly answered the question in late 2007, soon after Christner finished his experiment but before he published it. Prather had spent 15 years developing a device that could suck in microscopic particles from the air and analyze their chemical composition in a few thousandths of a second. She intended the device to study air pollution. But that year she teamed up with Paul DeMott, a cloud physicist from Colorado State University, to see if they could use the machine to identify the particles that form ice in clouds.

Prather and DeMott made a dozen flights through clouds as high as 25,000 feet over the mountains of Wyoming, Colorado, and Montana. The results caught the researchers off guard. They saw the expected soot, sea salt, and plenty of local dust inside the liquid droplets, but those things rarely had ice associated with them. No, the icy patches in clouds contained something else: 50 percent of the particles were high-titanium dust from the deserts of Asia or possibly Africa, and another 30 percent seemed to be biological based on their carbon, nitrogen, and phosphorus content.

Until then, no one had seen life in a rain cloud. “We were not going in purposely looking for bacteria,” she says. But she has since seen similar results in two more cloud-sampling campaigns. The Sierra flights in early 2011 found that most ice crystals contained dust from deserts on other continents, along with what seemed to be bacteria. When Prather and DeMott sampled clouds over the Caribbean island of Saint Croix in July 2011, they again saw signatures of both dust and cells in the ice. The dust itself seemed to come from North Africa.

DeMott amassed a library of 100,000 ice-forming particles collected from these flights. He has only begun to identify them, but in electron microscope images he finds that some of them resemble bacteria. Franc plans to start testing them soon, looking for the DNA sequences involved in ice nucleation.

Despite these results, some researchers remain skeptical that microbes could be important for cloud formation. Corinna Hoose, a cloud physicist at the Karlsruhe Institute of Technology in Germany, thinks that the number of ice-forming cells that Christner found in snow is too low to make much difference. She estimates that airborne microbes contribute less than 1 percent of all cloud ice. “Dust and soot particles are much more abundant,” she says. “They are much more important than biological particles.”

That has been the dominant view for decades, and dominant views do not die easily. But Prather, Christner, and Sands are chipping away at it. Lab studies have yet to turn up a common mineral that triggers freezing as effectively as bacteria like syringae do. “These organisms,” Christner says, “are able to catalyze ice formation at a temperature warmer than any other naturally occurring particle.”

And Prather may actually have underestimated the abundance of ice-forming biological particles in her samples. Even though she saw lots of dust in the icy patches of clouds, 60 percent of those dust grains also contained carbon, nitrogen, and phosphorus, suggesting that they are not just minerals. They might have cells lurking inside them—or at least the guts of dead cells splattered on their surfaces. It’s an issue that Prather chats about with her colleagues plenty these days. “I don’t think it’s the dust itself” that’s forming ice, she now says. “I think it’s the bio.”

While Prather and others explore the environmental impact of airborne bacteria, Christner is trying to understand this mysterious high-altitude ecosystem as a whole. Working not far from his base at LSU, he is setting out to learn how much is alive in the sky and how far up it lives.

On a cold autumn morning, Christner guides his pickup down a two-lane highway in central Louisiana, speeding past pine thickets, swamps, and dirt roads that don’t show up on his truck’s navigation system. “I’m not sure where the hell we’re going,” he admits.

Christner’s movements are at the mercy of the shifting winds thousands of feet overhead. He and five other vehicles are following a weather balloon that was launched several minutes ago. “It’s at 5,000 feet and still ascending—we’re about two miles ahead of it on the road,” says a voice over the radio.

The balloon’s 12-pound payload, including a GPS transponder, is held together with flimsy string and Styrofoam—designed, per FAA regulations, to disintegrate if it encounters the turbines of a passenger plane. Between 10,000 and 30,000 feet, a door will spring open. The device will collect drifting cells in a coat of gooey grease, just as the front grill of Christner’s truck is collecting insects. At 35,000 feet the germ catcher will cut its string, drift down on a parachute, and hopefully land where it can be found.

The device has previously come to rest in swamps and rice paddies and tangled itself 60 feet up in a longleaf pine. Today the scientists find it in a dense forest, seven-tenths of a mile past a locked gate adorned by two deer skulls with metal stakes driven through their foreheads.

Christner’s Ph.D. student Noelle Bryan will grow the bacteria the balloon collects in order to identify them. Later she’ll look to see if they carry the gene for ice nucleation. “My guess,” Christner says, “is there are probably a whole slew of organisms out there that have this capability that we simply haven’t identified.”

Conditions in the stratosphere resemble those on the surface of Mars. Levels of damaging ultraviolet radiation are 1,000 times as high as those at sea level, threatening to chisel a cell’s DNA into haiku-length snippets. And the atmospheric pressure is only half a percent of what it is at sea level, threatening to shrivel cells into freeze-dried corpses. Christner sees his studies in the stratosphere as a test run for looking for life in other worlds. “What we are doing,” he says, “is not so different from what will happen in the future when we send missions to Mars to take samples to bring them back to Earth.”

Analyzing what, if anything, Christner’s team collected on that 122,000-foot flight will take months. But they have promising results from earlier launches to lower levels of the atmosphere.

Back at their lab Christner and Bryan examine two dozen petri dishes containing bacteria collected from flights that peaked between 10,000 and 80,000 feet. They have yet to sequence and ID the microbes’ DNA, but just looking at the petri dishes gives them some clues. Bright-colored colonies dot many of the dishes—reds, oranges, pinks, and yellows. “Those are natural sunscreens,” Christner says. Colored carotenoid pigments (similar to those in many plants, including the carrots that the compounds are named after) can neutralize damaging ultraviolet light.

Christner holds up a petri dish containing colonies mottled with beige, black, and white: an indication that these bacteria produce dehydration-resistant spores that could help them survive at extreme altitudes. They are probably some sort of Actinomyces (a group of bacteria that live in soils and include species that make streptomycin and other antibiotics), but the species could be new to science. “We don’t know,” Christner says, looking at a splotch. “That might make an antibiotic that nobody’s ever seen before.”

Searching the skies for high-living microbes may also lead to insights concerning some species that we already do know about. In the 1950s researchers funded by the U.S. military tried to sterilize canned meat by blasting it with radiation. When they opened the cans, they were surprised to find the meat rotten: It had been fermented by a bacterium, now called Deinococcus radiodurans, that is exceptionally resistant to radiation. The species carries a muscular set of enzymes that stitch its DNA strands back together as quickly as radiation splinters them apart. Deinococcus can survive 5,000 times as much radiation as human cells, but no natural environment on Earth comes close to those levels of irradiation. “So what the hell has an organism like this evolved tolerance for?” Christner asks.

John Battista, who studies the microbe at LSU, just upstairs from Christner, thinks Deinococcus’s DNA repair enzymes primarily help it survive dehydration in its native desert environment. But radiation tolerance could also allow the bacterium to join that 10-mile-high ecosystem. Windstorms in places like the Gobi Desert could easily pick up Deinococcus and propel it around the world. “If it managed to get into the upper atmosphere,” Battista says, “it has all the tools it needs to survive.” Deinococcus and other similarly hardy microbes may be lurking in the samples Christner’s team is culturing.

Just as the same genes that allow Deinococcus to thrive on the ground may give it the ability to survive at high altitudes, the ice-nucleation gene may originally have given syringae and bacteria like it an advantage other than rainmaking. The nucleation gene appears to be unrelated to any of the more than a million genes that have been sequenced to date from various organisms. And the gene seems to have arisen only once in the course of evolution; after that, it passed from one species to another, changing little along the way. No one knows how long ago the gene emerged, but its appearance may have marked a pivotal moment in Earth’s history. It may have provided a new way for life to modify the planet’s environment.

Ice nucleation might have emerged as an ecological handshake between bacteria and the plants they lived on. Many wild plants (unlike most cultivated crops) are frost tolerant. They can survive as long as the freezing happens slowly, giving the plants time to activate their defenses. By causing frost to set in at higher temperatures—at 25ºF, say, rather than 15—ice-nucleating bacteria would have caused freezing to happen more slowly, helping protect the plants they lived on.

Later on, the talent for forming ice may have found other uses. Syringae uses ice crystals to rip open the cells of plants that are not frost tolerant, so it can devour their nutrients. Microbes like syringae may also exploit ice nucleation to parachute down in raindrops or snowflakes, ensuring they do not remain stuck at high altitudes when swept up by storms.

Ice-nucleating bacteria might even influence the entire landscape. By triggering rain, Sands says, “they cause more plants.” Just as humans farm wheat, syringae might cultivate leafy ecosystems that can sustain the bacteria once they reach the ground. Those ecosystems would then spawn more bacteria, some of which would return to the sky.

The realization that bacteria could have such profound impacts adds one more twist to the already convoluted connection between human activity, weather, and climate. Forests may make their own local rain by releasing bacteria and other organic compounds into the lower atmosphere. Deserts may trigger precipitation thousands of miles away when their dust and bacteria collide with water-rich masses of air. What effect, then, of deforestation or desertification?

Researchers have studied desert dust for decades, tracking its serpentine trajectory around the globe and trying to understand its environmental impact. Now it seems that dust might have been a decoy, hiding the bacteria that could be the real directors of much of our planet’s weather. “When I look at what physically forms the ice in clouds, I’d say 80 percent of it has some sort of biological signature,” Prather says. “The dust by itself doesn’t explain it.”

Microbe Rainmakers

Four decades ago, scientists discovered that the bacterium Pseudomonas syringae triggers the formation of frost on plants. Since then, some researchers have proposed the ice-making bug and others like it might be creating ice crystals in clouds that result in precipitation. It’s not clear yet if airborne microbes really influence the weather, but that hasn’t stopped some optimistic scientists from studying the bacteria as a tool to increase rain and snow.

Harnessing bacteria to make precipitation could be big business. Dozens of states and countries run cloud-seeding programs using artificial ice-nucleating compounds. In California the effort is especially urgent. The Sierra Nevada snowpack, which provides about 65 percent of the state’s water, has been declining since 1950, and a quarter of the snow is projected to disappear by 2050.

Montana State University bacteriologist David Sands imagines using syringae to bring more rainfall to places like California. “I’m an agriculturalist,” he says. “I don’t like droughts.” Microbiologists have identified strains of syringae that form frost without damaging their hosts. It could be possible to plant large tracts of land with plants that harbor these strains. The microbes might then get into the air, form ice crystals in clouds overhead, and pull more rain from them.

No one knows if this is practical. Clouds would have to be hit at just the right time, when their humidity and temperature are ideal for ice formation. And the areas of land that have to be planted might also be prohibitively large. But Sands is plugging ahead, working with researchers in Syria and other countries to find varieties of wheat and barley that preferentially harbor the right strains of syringae. “We might be able to introduce new varieties with the bacteria on the seed,” he says. “We might be able to capture 10, 20, 30 percent more rainfall in some areas.” 

Douglas Fox is a freelance science writer based in California. His work has appeared in The Best American Science and Nature Writing.