A larger hidden world of bees
The world may contain substantially more bee species than scientists have formally named. A global analysis published in Nature Communications estimates a lower bound of 24,705 to 26,164 species, compared with 20,934 recognised species in the study’s reference taxonomy.
That difference amounts to an estimated 3,771 to 5,230 additional species, or 18% to 25% more than the recognised total. It does not mean those species were discovered in the study. Instead, statistical models indicate how many may remain undescribed.
The result matters because bees pollinate wild plants and agricultural crops, while many conservation decisions depend on knowing which species exist and where they live.
The researchers also found that gaps in scientific knowledge are unevenly distributed. Asia, Africa and the Americas appear especially important for future taxonomic work.
How scientists estimated species they have not seen
James B. Dorey and colleagues combined a global bee occurrence database with taxonomic names, national checklists and information extracted from scientific literature.
Their inputs included approximately 8.3 million cleaned occurrence records, roughly 21,000 valid species names, about 24,000 synonyms and approximately 66,000 species-country combinations.
A synonym occurs when multiple scientific names refer to the same recognised species. Accounting for these names is essential to avoid counting a bee twice.
The team used statistical richness estimators, including iChao1 and an approach based on Hill numbers, to infer species that existing records are likely to have missed.
They generated estimates across global, continental and national scales, including 186 countries, and repeated parts of the sampling process 100 times.
These calculations draw information from how often different species appear in records. Rarely recorded species can indicate that an inventory is incomplete.
The methods do not provide a catalogue of the unnamed species. They instead quantify a likely lower boundary for the size of the hidden diversity.
The key numbers
The researchers’ principal global interval was 24,705 to 26,164 species. Against the recognised baseline of 20,934, that corresponds to 3,771 to 5,230 additional species.
Two estimation approaches produced somewhat different central values. The iNEXT-based global estimate was 24,985, while the iChao-based estimate was 25,929.
The researchers presented uncertainty intervals around these estimates rather than claiming a single exact total. This is particularly important because incomplete sampling can distort estimates.
The analysis also showed why the rate of scientific naming matters. Since 1960, valid new bee names accumulated at approximately 117 species per year.
At that historical pace, describing the estimated gap could require roughly 32 to 45 years. The authors stressed that this is likely conservative.
The projection is not a deadline or a forecast of when the last species will be found. It simply illustrates the scale of taxonomic work still required.
Where the missing diversity may be concentrated
The analysis identified Asia as having the largest observed and predicted bee richness, together with a substantial gap between known and estimated species.
The authors also highlighted underdescribed diversity in Africa and the Americas. However, country estimates depend partly on the quantity and quality of records available.
For Turkey, the estimated taxonomic gap was 843 species, equivalent to a 46% increase on its recognised total in the study.
China had an estimated gap of 637 species, or 47%, while Israel’s gap was estimated at 553 species, or 52%.
These country figures are estimates, not confirmations that those exact numbers of unknown bees inhabit each location.
The analysis also found that island nations tended to have higher bee species richness per unit area than continental nations, after accounting for area.
This pattern underscores the conservation value of islands, where distinctive evolutionary histories can generate concentrated biodiversity.
Why Africa deserves attention
Africa emerged as an important region for potential undescribed bee diversity, but the authors also noted limitations in available records.
Uneven sampling can make a region appear less species-rich simply because fewer specialists have surveyed it or fewer specimens have been entered into accessible databases.
South Africa has extensive natural habitats and diverse flowering plants, yet national and regional inventories still depend on sustained taxonomic expertise.
For conservation planners, the implication is not that every unstudied site necessarily contains a new bee species. It is that uncertainty should be acknowledged when prioritising research.
Collaborations between museums, universities, local researchers and international specialists could help strengthen inventories and reduce reliance on incomplete external datasets.
Training taxonomists and supporting specimen collections are therefore part of biodiversity protection, not merely academic activities.
The role of research capacity
The researchers tested how estimated taxonomic gaps related to socioeconomic and geographic variables.
Their analysis considered gross domestic product per capita, tertiary education, observed richness, occurrence-record coverage, country area, elevation and distance from roads.
Observed species richness was consistently associated with the estimated number of undescribed species. Places already known to support many bees may also contain many species yet to be named.
The number of cleaned occurrence records and the completeness of digital checklists were also informative indicators of the estimated gap.
Such associations should not be mistaken for proof that national income directly causes biodiversity to be overlooked. Research investment, accessibility and historical collecting practices are interconnected.
The practical point is that scientific capacity affects what becomes visible in global biodiversity databases.
Expanding local expertise and making reliable records accessible can improve both estimates and the decisions that depend on them.
Why counting species is difficult
Many bee species are small, seasonal or difficult to distinguish from closely related organisms.
Two bees can look nearly identical yet represent different evolutionary lineages. Conversely, differently coloured or sized individuals may belong to the same species.
Modern taxonomy can combine external anatomy with genetic information to resolve these ambiguities.
The study cited a striking example from Fiji, where four morphologically described Lasioglossum bee species increased to 13 after additional systematic work, with molecular evidence suggesting at least 28.
That example does not establish that every region will experience a comparable increase. It illustrates how traditional inventories may underestimate diversity.
Taxonomic revisions can also reduce species counts when multiple names are shown to describe the same organism.
Consequently, species totals remain provisional as evidence accumulates and classification improves.
What the estimates mean for conservation
A species cannot be assessed accurately for extinction risk if researchers do not know that it exists.
Unknown species may inhabit habitats undergoing agricultural expansion, urban development or climatic change before they are formally described.
However, the study did not directly measure extinction rates or demonstrate that the estimated unnamed bees are all threatened.
Its main contribution is to identify where inventory work could be especially valuable and to provide a quantitative framework for planning it.
Conservation organisations can use these estimates alongside habitat condition, land-use pressure and existing protected-area coverage.
Protecting diverse habitats while improving inventories can reduce the risk of waiting for complete taxonomic certainty before taking action.
The analysis also highlights the value of well-maintained collections, open data and transparent statistical tools.
Important limitations
The estimates are lower bounds, not a definitive census of all living bee species.
Occurrence databases contain geographic biases because some places are visited, sampled and digitised more frequently than others.
Countries with very few records were excluded from some analyses, and the researchers acknowledged that estimates can change with input data.
Statistical estimators also rely on assumptions about the distribution of common and rare species.
Rare or cryptic species may be especially difficult to detect, potentially leaving even the new estimates below the true total.
The researchers cautioned that taxonomic concepts, limited molecular analysis and inconsistent sampling could influence their results.
The 32-to-45-year calculation assumes historical naming rates and does not account fully for future changes in funding, technology or the difficulty of finding increasingly elusive species.
These limitations do not erase the evidence of a gap. They explain why the reported numbers should guide research rather than be treated as exact inventories.
What happens next
The study’s computational framework offers a way to revisit estimates as new records become available.
Researchers can use improved checklists, museum collections, field surveys and genetic analyses to refine the global picture.
The same general approach may also help estimate hidden diversity in other poorly documented groups of organisms.
For the public, the result offers a reminder that familiar animals can still contain substantial scientific unknowns.
For policymakers, it reinforces the need to invest in the people and institutions capable of documenting biodiversity.
The clearest conclusion is that the recognised inventory of bees is incomplete by thousands of species, and that closing the gap will require sustained, geographically equitable scientific work.
Source Information
Study: Estimating global bee species richness and taxonomic gaps
Authors: James B. Dorey, Amy-Marie Gilpin, Nikolas P. Johnston, Damien Esquerré, Alice C. Hughes, John S. Ascher and Michael C. Orr.
Journal: Nature Communications, volume 17, article 1762
Published: 24 February 2026
Methods: Statistical richness estimates from 8.3 million cleaned bee occurrence records, taxonomic databases, checklists and literature data, covering 186 countries.
