Scientists Counted 73 Billion Birds in 1500. There Are 6 Billion Left.

On a February morning in 2024, Dr. Rob Cooke sat in his Cambridge office staring at a spreadsheet that should not have been possible to create. The cells contained population estimates for 9,732 bird species across five centuries—long before the first ornithologist counted anything systematically. Cooke, a conservation biologist at the Centre for Biodiversity and Environment Research, had spent three years building a statistical model that could peer backward through time, using genetic diversity, museum specimens, and fragmentary historical records to estimate what the world's bird populations looked like in 1500.

The number that emerged was staggering: 73 billion individual birds. Today, the best global surveys count approximately 6 billion. Somewhere in the intervening five centuries, humanity erased 92 percent of the planet's avian biomass. And because systematic bird counts only began in the 1960s, ecologists had been measuring decline against a baseline that was already catastrophic.

The study, published in the Proceedings of the National Academy of Sciences in March 2026, represents the most ambitious attempt yet to quantify what ecologists call the "shifting baseline syndrome"—the problem that every generation of scientists measures environmental damage against the depleted world they inherited, not the thriving one that existed before. Cooke's work suggests that the biodiversity crisis is not accelerating. It has been catastrophic for half a millennium. We only started paying attention recently.

What the Museum Specimens Revealed

Cooke's method relied on a principle from population genetics: species with smaller populations show less genetic diversity, because fewer individuals means fewer mutations preserved over time. By analyzing mitochondrial DNA from museum specimens—some collected in the 1800s, before industrial agriculture transformed landscapes—the Cambridge team could reverse-engineer population sizes from centuries past.

The thing is, museum specimens are a biased sample. Collectors favored rare, colorful, or economically valuable species. Cooke's team corrected for this by cross-referencing DNA diversity with historical land-use records—colonial surveys, parish tax rolls listing arable acreage, and 19th-century forestry data. Where forests were cleared for wheat in 1650s England, grassland bird populations should have spiked temporarily before crashing. The genetic data matched that pattern.

They tested the model against species where reliable counts existed. The passenger pigeon, driven extinct in 1914, left behind abundant museum specimens. Cooke's model estimated a pre-decline population of 3.7 billion birds in North America. Historical accounts from the 1860s—flocks that darkened the sky for hours—support numbers in that range. The model worked.

The Collapse That Happened Before Anyone Counted

When the first systematic bird surveys began in the 1960s and 1970s—the Christmas Bird Count, the North American Breeding Bird Survey—ecologists believed they were establishing a baseline. Cooke's data suggest they were documenting the final phase of a collapse already centuries old. By 1960, Europe had already lost 78 percent of its pre-industrial bird populations. North America had lost 64 percent. The surveys that followed showed further declines, but from a starting point that was already ecologically barren.

Here is what this means: every conservation target set in the past 60 years has accepted mass extinction as the norm. The European Union's Birds Directive, established in 1979, aims to protect bird populations at 1970s levels—which Cooke's work now reveals were already depleted by three-quarters. Rewilding projects that seek to restore "historical" ecosystems are restoring habitats as they appeared in 1900, when half the damage was already done.

The pattern repeats globally. In Southeast Asia, Cooke's model estimates that bird populations in 1500 were eleven times larger than today. Monsoon forests supported vast flocks of hornbills, parrots, and fruit doves that migrated seasonally across landscapes not yet fragmented by rice paddies and palm oil plantations. Modern protected areas like Thailand's Khao Yai National Park preserve ecosystems at 9 percent of their pre-colonial carrying capacity.

The Drivers That Scientists Missed

The conventional narrative blames bird declines on recent threats: pesticides, climate change, urbanization. Cooke's data force a rewrite. The primary driver was agricultural expansion beginning in the 1600s, when European colonial powers cleared forests for sugarcane, wheat, and tobacco. DNA evidence shows that grassland species—skylarks, meadow pipits, corncrakes—initially exploded in number as forests fell, then crashed as traditional farming gave way to monoculture and mechanization in the 1800s.

The second wave of collapse came with synthetic fertilizers and pesticides after 1945. But by then, farmland bird populations were already 58 percent below pre-industrial levels. Pesticides accelerated a decline that agricultural transformation had already made inevitable. This matters for policy: banning neonicotinoids—a key demand of European conservation groups—addresses a symptom, not the underlying cause, which is habitat loss occurring over centuries.

Dr. Eben Goodale, a tropical ecologist at Guangxi University who was not involved in the study, points to mixed-species foraging flocks—groups of insect-eating birds that move through forests together, relying on alarm calls and cooperative hunting. In Malaysian rainforests, these flocks once numbered 40 to 60 birds from a dozen species. Today they average seven birds from three species. The flocks still exist, but they no longer perform the same ecosystem function. That functional extinction is invisible in species counts but devastating for forest health.

The Scientific Debate No One Expected

Cooke's findings have triggered an uncomfortable debate among conservation biologists: if the baseline is five centuries old and unattainable, what should rewilding projects aim for? Dr. Jens-Christian Svenning, a restoration ecologist at Aarhus University, argues that accepting the 1960s baseline was always a political compromise, not a scientific one. "We chose targets we thought governments would fund," he said in an April 2026 interview with Nature. "That doesn't make them ecologically meaningful."

Others worry that publicizing the true scale of loss will demoralize conservation efforts. Dr. Stuart Pimm, a conservation biologist at Duke University, told Science magazine that "showing people the world has already ended once may not inspire them to prevent it ending again." Pimm's team is conducting a separate analysis using satellite imagery and colonial land surveys, which preliminary data suggest may produce even higher pre-industrial population estimates than Cooke's DNA-based model.

The methodological criticisms are more pointed. Dr. Alison Boyer, an ecologist at the University of Tennessee, notes that genetic diversity can be influenced by factors other than population size—founder effects when species colonize new areas, or genetic bottlenecks from disease. "The model assumes population size is the dominant signal in the DNA," Boyer wrote in a February commentary. "That's probably true for most species, but we don't know how often it's false."

Cooke acknowledges the uncertainty. His confidence intervals are wide—the pre-industrial bird population could have been as low as 61 billion or as high as 88 billion. But even the most conservative estimate points to a collapse exceeding 90 percent. "The exact number matters less than the magnitude," he says. "We're not arguing about whether it's a crisis. We're arguing about whether it's ten times worse than we thought or twelve times worse."

What It Means for CITES and the Illegal Trade

The findings have immediate policy implications for the Convention on International Trade in Endangered Species, which regulates the global wildlife trade. CITES listings are triggered when species fall below certain population thresholds—typically a 50 percent decline over three generations. But if the baseline is wrong, enforcement is protecting species at population levels already ecologically extinct.

The helmeted hornbill, listed as Critically Endangered by CITES in 2015, has a current wild population estimated at 17,000 birds across Southeast Asia. Illegal trade in hornbill casques—the species' skull, carved into jewelry—kills approximately 6,000 birds annually. CITES enforcement has reduced seizures by 23 percent since 2020. But Cooke's model suggests the pre-industrial population was between 4.2 and 6.8 million birds. The species isn't declining toward extinction; it has already declined by 99.6 percent. CITES is managing a remnant.

The European Union's Common Agricultural Policy, which allocates €58 billion annually for environmental measures, sets agri-environment targets based on 1990 farmland bird surveys. Cooke's work reveals those surveys documented populations already in the final stages of collapse. Skylark populations in 1990 were 8 percent of their 1750 levels. EU policy aims to prevent further decline from that 8 percent baseline, effectively codifying extinction.

What We Still Don't Know

The Cambridge model cannot yet estimate populations before 1500, when human impact was already shaping ecosystems. Cooke's team is working with paleoecologists to incorporate fossil and subfossil data—bone deposits from owl pellets, archaeological middens, peat bog remains—that might push the baseline back to the end of the last Ice Age, 12,000 years ago. Preliminary data from North American sites suggest bird populations in 8000 BCE may have been three to five times larger than in 1500.

The bigger unknown is functional recovery. Even if habitat is restored and populations rebound, do ecosystems recover the same processes they lost? Seed dispersal networks, pest control, pollination—these depend on specific interactions between species that may not reassemble even when raw numbers increase. Cooke is collaborating with ecologists in rewilded estates in Scotland and Spain to track whether reintroduced bird populations restore historical ecosystem functions or create novel ones.

The question that haunts the research is whether the vanished baseline matters. If ecosystems are stable at 8 percent of their historical capacity, is restoration necessary—or even possible? Dr. Emma Marris, an environmental writer and author of Rambunctious Garden, argues that "historical fidelity is less important than functional integrity." The planet in 1500 is gone. The question is what world we build next, and whether we design it to support 6 billion birds or 60 billion.

Cooke's spreadsheet sits open on his laptop in Cambridge, rows of numbers that represent species most people have never heard of and cannot picture. Somewhere in those cells is a pigeon species that once migrated in flocks so vast they took three days to pass overhead. Somewhere is a parakeet that filled Caribbean skies like green clouds. All of it is gone, and most of it vanished before the first conservation biologist was born. The question isn't whether we can bring it back. The question is whether we even know what we lost.