Walk through a grocery store and you are looking at two different food systems sitting side by side. The bread, rice and cereal aisles do not need an insect. The apples, almonds, blueberries and squash do.
A 2007 review in Proceedings of the Royal Society B put primary data from 200 countries behind that split. Of the leading global food crops, 87 depend to some degree on animal pollination and 28 do not. Then the authors weighed the same crops by tonnage, and the picture inverted. Sixty percent of global production volume comes from crops that need no pollinator at all. Thirty-five percent comes from crops that do.
Both facts are true at once, and holding them together is most of the skill in this subject. Pollinators are not what stands between you and hunger. They are what stands between you and a diet of grain.
What follows sorts the biodiversity figures you are likely to meet into two piles: what somebody counted, and what somebody modeled. That difference matters more than the size of any single number.
Managed honey bees get counted, and the count surprises people
USDA’s National Agricultural Statistics Service surveys beekeeping operations with five or more colonies and publishes the results quarterly. Its report of August 1, 2025 put 2,633,700 colonies in the country on January 1 of that year, down about 1 percent from a year earlier. By April 1 the figure had climbed to 2,987,310.
Losses in the same report are large and routine. Beekeepers lost 267,260 colonies between January and March 2025, about 10 percent of the total, and 292,630 between April and June, also about 10 percent.
Here is what trips people up. Colony counts and colony losses can both be high at the same time, because a beekeeper can split one strong hive into two and rebuild the number within a single season. The survey measures a beekeeping industry, not the health of bees as such. Read it that way and the long arc makes sense. USDA’s Agricultural Research Service reports managed colonies falling from roughly 5 million in the 1940s to about 2.66 million today, a slide shaped as much by the economics of honey and labor as by anything happening inside a hive.
ARS also attaches a dollar figure to what those hives do: honey bee pollination adds at least $18 billion a year to the value of American agriculture.
Colony Collapse Disorder, meanwhile, has a narrow technical meaning that rarely matches how the phrase gets thrown around. The Agricultural Research Service defines it as “a dead colony with no adult bees and with no dead bee bodies but with a live queen …” Dead bees on the ground are not CCD. A weak, dwindling hive is not CCD. ARS says no scientific cause for it has been proven, and the agency closes the door on one popular theory outright: cell phones and cell towers have not been shown to be connected to CCD or to poor honey bee health.
Wild bees are mostly not counted at all
The honey bee is a managed farm animal, imported and kept in boxes. It is also one species among thousands. The U.S. Geological Survey reports more than 20,000 known bee species worldwide, 4,000 of them native to the United States, and roughly 10 percent of American bees still unnamed and undescribed. You cannot track a population that has no name yet.
That gap is why nearly every statement about wild bee trends traces back to a model rather than a count. The clearest American example appeared at the end of 2015 in the Proceedings of the National Academy of Sciences. The researchers combined national land-cover data with a spatial habitat model and formally quantified expert judgment, then estimated wild bee abundance county by county. Between 2008 and 2013, modeled abundance fell across 23 percent of US land area, concentrated where natural habitat had been converted to row crops. They identified 139 counties where thin bee supply sat next to heavy planting of pollinator-dependent crops, together covering 39 percent of the country’s pollinator-dependent crop acreage.
That is careful work and it is the best national picture available. It is also a model. The 23 percent figure comes out of assumptions about which kinds of land support which bees, not out of anybody counting bees on 23 percent of the country.
What does get counted, one species at a time, is listing under the Endangered Species Act. The U.S. Fish and Wildlife Service listed the rusty patched bumble bee as endangered in 2017, and reports that since 2000 the species has turned up in only 13 states and one Canadian province. Franklin’s bumble bee was listed as endangered in August 2021. Its entire known range covers about 13,000 square miles in the Klamath Mountains of southern Oregon and northern California, and the last confirmed sighting was a single bee netted during a 2006 survey.
A listing decision rests on a formal evidentiary record and a public rulemaking. That is a different kind of certainty than a map produced by a habitat model, and the two are worth keeping separate.
What threatens pollinators
Researchers are considerably better at naming the stressors than at ranking them. Any article that hands you a tidy priority order is adding confidence the evidence does not have.
For managed honey bees, one stressor does dominate the reporting. Varroa destructor, a parasitic mite that feeds on bees and spreads viruses between them, was the number one stressor reported in every quarter of 2024 in the NASS survey. It affected 29.9 percent of colonies in the first quarter of 2025 and 27.3 percent in the second.
For wild bees the leading explanation is about food and ground rather than parasites. The U.S. Geological Survey states it plainly: “Loss of plant diversity is the primary cause of native bee decline.” USGS adds that 30 to 50 percent of native bees are highly specialized, drawing on a narrow set of plants, so a bee tied to one wildflower goes when the wildflower goes. That is also the mechanism the PNAS model identified behind its declines.
Pesticides belong in the picture, and the picture there is genuinely mixed. For the rusty patched bumble bee the Fish and Wildlife Service points to a probable combination of an introduced pathogen and exposure to insecticides and fungicides, while saying outright that the exact cause of the decline is unknown. ARS reports its own experiment exposing bees to both the insecticide imidacloprid and Nosema fungi, which found sublethal effects on individual bees and no adverse effect on overall colony health. Dietary residues and household pesticide products each get their own treatment elsewhere in this library.
A 2016 review in Nature, written by the team that assembled the international pollinator assessment, groups the threats as land use change and management intensity, climate change, pesticides and genetically modified crops, pollinator management and pathogens, and invasive species. Their summary of the trend evidence is carefully hedged: declines are well documented in some wild and managed pollinators in several regions of the world. That is not a global collapse, and it is not nothing.
Crop diversity narrowed, and also widened
FAO’s State of the World’s Biodiversity for Food and Agriculture, published in 2019 from reports submitted by 91 countries and 27 international organizations, carries the figures that circulate most. More than 6,000 plant species have been cultivated for food. Fewer than 200 make a major contribution to food production at a national, regional or global level. Nine account for 66 percent of total crop production.
Livestock show the same shape with a bigger hole in the middle of the data. FAO counts 7,745 local breeds still in existence and reports 26 percent of them at risk of extinction. The figure almost nobody quotes is the next one. The risk status of 67 percent is unknown. Only 7 percent are confidently not at risk. When two thirds of a category is unassessed, the headline share is a floor, not a measurement of the problem.
Now the complication, and it is a real one. A 2014 analysis in PNAS tracked 50 years of national food supplies and found something that cuts against the simple story. Inside individual countries, food supplies became more diverse. More crop commodities contributed, their contributions evened out, and the dominance of the top few fell. What changed globally is that national diets converged on each other. The authors describe the outcome as a food supply that is relatively species-rich within any one country and species-poor across the world as a whole.
Your supermarket almost certainly stocks more kinds of produce than your grandparents’ did. Supermarkets in different countries now stock more of the same things as each other. Both statements come out of the same measurement, and an article that gives you only one of them is selling something.
Inside the national seed collection
USDA’s Germplasm Resources Information Network describes more than 600,000 active accessions of plant genetic resources in the National Plant Germplasm System. An accession is one distinct sample with a record attached: a named variety, a farmer’s landrace, a wild relative of a crop, a breeding line somebody thought worth keeping.
It is not one vault. The NPGS repository list runs to roughly 30 entries, with collections held at sites including Geneva in New York, Corvallis in Oregon, Davis and Riverside in California, Griffin in Georgia, Brownwood in Texas, and Miami, Hilo and Mayaguez. The division of labor is practical. Seeds that tolerate drying and freezing go into cold storage. Apples, citrus, berries and other crops that cannot be stored as seed are kept as living trees and vines, or frozen as cuttings in liquid nitrogen.
This is a working library rather than a museum. Plant breeders and researchers request material from it, and USDA’s own documents are candid about the strain on the system. The 2018 Farm Bill directed the department to develop and implement a national strategic germplasm and cultivar collection assessment and use plan, with reducing backlogs that keep germplasm and its associated data out of public reach named among the purposes.
The 1970 corn epidemic, and what it really shows
In 1970 a fungus took an estimated 16 percent of the US corn crop. A 2024 perspective article in the journal Phytopathology puts the loss at roughly $8 billion in current terms.
The mechanism is the textbook case for genetic uniformity. Nearly all the hybrid corn seed planted that year carried Texas male sterile cytoplasm, a trait seed companies had adopted because it spared them the labor of detasseling plants by hand. A new race of the fungus turned out to be extraordinarily virulent against precisely that cytoplasm, and because almost every field shared it, almost every field was vulnerable at once.
The authors of that perspective, one of them at USDA’s Agricultural Research Service, push back on how the story usually gets told. They read the epidemic not as reckless breeding but as an unlucky consequence of a real efficiency gain, and they put as much weight on what happened afterward. The epidemic lasted a single year. Public-sector researchers worked out what was happening, and the seed supply shifted back to normal cytoplasm within a season.
Hold both halves of that. Genetic uniformity is a measurable and expensive risk. A functioning breeding system with a genebank behind it is also a measurable and fast defense. The 1970 corn crop is evidence for both claims, and it usually gets cited for only the first.
What biodiversity loss means for you, honestly
This is where most coverage oversells. Biodiversity loss is not a personal exposure. No clinician tests you for it, no threshold shows up in your bloodwork, and nothing about it behaves like lead paint or secondhand smoke. The route from a declining pollinator population to your body runs through yields, grower costs, prices and what the store stocks. It runs slowly, unevenly, and mostly out of sight.
The nearest thing to a direct health estimate is a 2015 modelling analysis in The Lancet, and reading it closely is a lesson in itself. The researchers assembled supply data on 224 foods across 156 countries, then asked what would follow from the complete removal of animal pollinators, holding calorie intake constant by substituting staple foods. In that scenario global fruit supplies fell 22.9 percent, vegetables 16.3 percent, and nuts and seeds 22.1 percent. An estimated 71 million people in low-income countries became newly deficient in vitamin A, and annual deaths from non-communicable and malnutrition-related disease rose by about 1.42 million. A scenario of 50 percent pollinator loss produced roughly 700,000 additional deaths a year.
Nobody forecasts the complete removal of pollinators. That number is a bound, constructed to show the shape of the dependency, and the modeled harm lands overwhelmingly on populations already short of micronutrients. It is not a prediction about a US reader with a well-stocked store eight minutes away.
The measured American finding is smaller and much more concrete. A 2020 study in Proceedings of the Royal Society B tested seven crops at 131 locations across major US growing regions and found evidence of pollinator limitation in five of them, meaning yields would have been higher with more pollination than the crops received. Wild bees and honey bees supplied comparable amounts of pollination for most of those crops, even in intensively farmed country. The authors valued wild pollinators’ contribution to just those seven crops at more than $1.5 billion a year.
Translated into something you can act on: this shows up as hive rental fees, grower margins and produce prices, not as a nutrient gap in your kitchen.
The one direct health risk in this subject
Bees do pose an immediate medical risk, and it has nothing to do with biodiversity. A sting reaction that spreads beyond the sting site is an emergency. Trouble breathing, swelling of the face, lips or throat, hives appearing somewhere other than where you were stung, repeated vomiting, or dizziness and faintness after a sting all warrant calling 911 and using an epinephrine auto-injector if one has been prescribed. Anyone with a known sting allergy should not wait to see how the reaction develops.
If a swarm settles on your property, a county extension office or a local beekeeping association will often collect it, which is safer and usually cheaper than spraying. If anyone feels unwell after a pesticide application, Poison Help answers at 1-800-222-1222.
Four questions for any biodiversity number
Counted or modeled? A colony census and a habitat model are both legitimate, and they carry different uncertainty. The NASS colony figures were collected from beekeepers. The 23 percent decline in wild bee abundance was produced by a model. Neither is fake; they are not interchangeable.
What is the unit? Colonies, species, land area, production volume and dollars all tell different stories about the same system, and swapping units mid-argument is the most common move in bad coverage. The 87 pollinator-dependent crops and the 35 percent of production volume in this article’s opening come from the same 2007 review.
What is the baseline, and did the method change? Colony counts from the 1940s came out of a different agricultural economy with different reporting. A long series that crosses a methodology change is not a trend line.
Is it a measurement, a projection, or a bound? The Lancet’s pollinator figures are a bound: a deliberately extreme scenario used to size a dependency. A bound presented as a forecast is the single most misleading thing in this field.
The distinction worth keeping
Three different systems get stirred together under one word, and almost every confusing headline about biodiversity and food comes from the mixing.
Managed honey bees are livestock, counted reasonably well, under real pressure from a mite, and partly buffered by beekeepers who can rebuild numbers. Wild pollinators are thousands of species, most of them unmonitored, with the best national evidence coming from models that point to habitat conversion. Crop genetic diversity is a third thing again, held partly in half a million accessions in federal genebanks and documented well enough that FAO can tell you both how narrow the global crop base is and how much of the livestock picture remains unassessed.
Keep those three apart and the numbers stop contradicting each other. Collapse them into one story and you will believe something wrong no matter which side of the argument wrote it.
Keep exploring: right to repair, household water use and sustainable packaging.
Sources
- U.S. Department of Agriculture, National Agricultural Statistics Service. “Honey Bee Colonies, released August 1, 2025.” nass.usda.gov
- U.S. Department of Agriculture, Agricultural Research Service. “Colony Collapse Disorder: Questions and Answers.” ars.usda.gov
- U.S. Geological Survey. “How many species of native bees are in the United States?.” usgs.gov
- U.S. Geological Survey, Eastern Ecological Science Center. “Native Bee Inventory and Monitoring Lab.” usgs.gov
- U.S. Fish and Wildlife Service. “Rusty Patched Bumble Bee (Bombus affinis).” fws.gov
- U.S. Fish and Wildlife Service. “Franklin's Bumble Bee (Bombus franklini).” fws.gov
- U.S. Department of Agriculture, Agricultural Research Service. “Germplasm Resources Information Network (GRIN).” ars-grin.gov
- U.S. Department of Agriculture, Agricultural Research Service. “National Plant Germplasm System Collections and Repositories.” ars-grin.gov
- U.S. Department of Agriculture, Agricultural Research Service. “National Plant Germplasm System: National Strategic Germplasm and Cultivar Collection Assessment and Utilization Plan.” ars-grin.gov
- Food and Agriculture Organization of the United Nations. “The State of the World's Biodiversity for Food and Agriculture, 2019.” fao.org
- PubMed. “Klein AM, Vaissiere BE, Cane JH, Steffan-Dewenter I, Cunningham SA, Kremen C, Tscharntke T. "Importance of pollinators in changing landscapes for world crops." Proceedings of the Royal Society B, 2007. PubMed. DOI: 10.1098/rspb.2006.3721.” doi.org
- PubMed. “Reilly JR, Artz DR, Biddinger D, et al. "Crop production in the USA is frequently limited by a lack of pollinators." Proceedings of the Royal Society B, 2020. PubMed. DOI: 10.1098/rspb.2020.0922.” doi.org
- PubMed. “Koh I, Lonsdorf EV, Williams NM, Brittain C, Isaacs R, Gibbs J, Ricketts TH. "Modeling the status, trends, and impacts of wild bee abundance in the United States." Proceedings of the National Academy of Sciences, 2015. PubMed. DOI: 10.1073/pnas.1517685113.” doi.org
- PubMed. “Smith MR, Singh GM, Mozaffarian D, Myers SS. "Effects of decreases of animal pollinators on human nutrition and global health: a modelling analysis." The Lancet, 2015. PubMed. DOI: 10.1016/S0140-6736(15)61085-6.” doi.org
- PubMed. “Potts SG, Imperatriz-Fonseca V, Ngo HT, et al. "Safeguarding pollinators and their values to human well-being." Nature, 2016. PubMed. DOI: 10.1038/nature20588.” doi.org
- PubMed. “Khoury CK, Bjorkman AD, Dempewolf H, Ramirez-Villegas J, Guarino L, Jarvis A, Rieseberg LH, Struik PC. "Increasing homogeneity in global food supplies and the implications for food security." Proceedings of the National Academy of Sciences, 2014. PubMed. DOI: 10.1073/pnas.1313490111.” doi.org
- PubMed. “Balint-Kurti P, Pataky J. "Reconsidering the Lessons Learned from the 1970 Southern Corn Leaf Blight Epidemic." Phytopathology, 2024. PubMed. DOI: 10.1094/PHYTO-03-24-0105-PER.” doi.org