Counting the Uncountable: A New Look at Global Insect Diversity
by Ritu Pandey and Shashank P. R.
For centuries, scientists have been fascinated by insects due to their incredible diversity and essential ecological functions. They facilitate crop pollination, nutrient recycling, pest control, and support vast food webs. Despite their significance, a fundamental question in biology has remained unanswered for over forty years: what is the actual number of insect species on Earth? For a long time, entomologists have estimated around six million species, a number that appears in textbooks, policy reports, and popular literature on biodiversity. However, a study published in PNAS in June 2026 by Robert Colwell, Laura Melissa Guzman, and colleagues challenges this estimate, suggesting it could be off by a factor of two to three. They propose that the real number may range from at least 14 to 20 million species.
Why Is Counting Insects So Difficult?
Unlike birds or mammals, insects are small, extremely numerous, and often difficult to distinguish. Many species live in remote tropical forests, emerge only at certain seasons, or inhabit specialised habitats such as forest canopies that are rarely sampled. Despite decades of collecting, taxonomists still discover species that have never been documented before.
This indicates that merely counting insects during field surveys does not fully capture the true insect diversity. Many species go unnoticed in any given sample, so researchers rely on statistical techniques to estimate the species that were not detected.
Why Was Six Million the Consensus?
Entomologist Terry Erwin's early guess in 1982 suggested there were about 30 million terrestrial arthropod species, sparking decades of debate and re-evaluation. Since then, most careful estimates have lowered that number to around six million. Counting insect species differs significantly from counting birds or mammals because many tropical insect species have yet to be formally identified. Taxonomists now often use DNA Barcode Index Numbers (BINs, genetically distinct clusters of specimens) as a practical proxy for species identification. Previous global estimates, mostly around six million, were based on extrapolations from known diversity, assumptions about host-plant relationships, or expert opinion. An exception was Li and Wiens' 2023 estimate of 21 million species, which challenged the consensus by accounting for cryptic species detected through DNA methods. Colwell and colleagues approached the problem differently by building their estimate from detailed sampling of one of the most intensively studied tropical insect communities.
What distinguishes the new study from previous rough estimates, dating back to Erwin's 1982 figure, is its scale and rigour: it analyses 1.6 million DNA-barcoded insect specimens, provides a comprehensive census of a highly diverse group of parasitoid wasps, and utilises statistical richness estimators adapted from epidemiology.
The new estimate bridges the gap between the current consensus and earlier high estimates, highlighting the potential extent of Earth’s undocumented insect diversity.
The Costa Rican Foundation: Malaise Traps and Caterpillar Rearing
The new research takes place in the Área de Conservación Guanacaste (ACG) in northwestern Costa Rica, where Daniel Janzen, Winnie Hallwachs, and a group of local parataxonomists have conducted one of the world's longest and most extensive insect surveys. Using fifteen core Malaise traps that ran nonstop for a total of 69 trap-years across dry forest, cloud forest, and rainforest, they gathered over 1.6 million barcoded insect samples, identifying 53,945 species (BINs).
To put that number into perspective, the paper notes that a separate nationwide barcoding project in Canada, which sampled over 1 million specimens across all insect orders, identified about 47,000 species nationwide. In comparison, the 53,945 species found in just 15 traps within a small Costa Rican conservation area already exceed that national total, an indication of how extraordinarily concentrated tropical insect diversity can be.
Despite sampling more than 1.6 million barcoded insects, the authors found that even this scale of effort failed to capture the full extent of species richness within a single conservation area.
Why a Wasp Subfamily Became the Measuring Stick
The main contribution of this paper is rooted in a statistical approach rather than in a purely taxonomic one. Rather than attempting to directly count every undetected insect species, an impossible task, the authors used the highly diverse parasitoid wasp subfamily Microgastrinae (family Braconidae) as a benchmark to assess undersampling.
Microgastrinae were extensively sampled using three independent approaches: 15 core Malaise traps, 15 additional 'peripheral' Malaise traps distributed more broadly across ACG, and 11,373 specimens reared from approximately 1,500 parasitised caterpillar species collected by hand. Notably, these methods showed little overlap. About 75% of the 1,414 microgastrine species identified appeared in only one sample set, indicating that even this thorough, well-funded effort remains limited in scope, capturing only parts of the hidden diversity. Despite pooling 21,669 specimens across all three methods, 403 out of 1,414 species (nearly 30 per cent) were singletons, represented by only one specimen each. This indicates that rare species still went undetected even at this scale.
Quantitative Venn diagram illustrating the number of Microgastrinae species (BINs) exclusive to reared specimens, core Malaise traps, and peripheral Malaise traps, as well as the shared counts among them. Source: Colwell et al., 2026.
Using Chao's "one-step" sample coverage estimator, a statistical tool originally created to estimate unrecorded disease cases in epidemiology, the team integrated data from three sources to produce a conservative lower-bound estimate of the wasp subfamily's actual diversity in ACG: 2,394 species. In comparison, only 388 species were found in the core traps alone. This discrepancy results in an undersampling ratio of approximately 0.16, indicating that the core traps likely captured just about 16% of the true microgastrine biodiversity.
The most significant assumption is that the same undersampling ratio is applicable to all of ACG's insects, not just wasps. Applying this ratio to the 53,945 observed species indicates that the true insect diversity in this forest could be around 332,846 species, including those that exist but were not sampled.
From One Forest to the Whole Planet
The final step estimates ACG's insect richness globally by using the ratio of the total number of tree species on Earth to those in ACG as an "upscaling factor." Trees are selected as the reference group because, like insects, they inhabit nearly every terrestrial ecosystem and have well-documented global and regional richness data. Using this method, the authors estimate that there are approximately 20.3 million insect species worldwide. The four reference groups-trees, amphibians, mammals, and saturniid moths- together yielded estimates from 14.2 to 20.3 million species, averaging 17.3 million.
A Conservative Estimate
The researchers explicitly state that their figure is a lower-bound estimate rather than a definitive count. Their sampling used Malaise traps placed exclusively at ground level, which do not effectively capture canopy insects or species best collected using methods such as pan traps, Winkler extractors, or canopy fogging. The authors acknowledge that accounting for these limitations and the likelihood that tropical insect ranges are smaller than those of reference groups would probably increase the estimated total number of insect species rather than decrease it.
Why Does This Matter?
Understanding how many species exist is more than just an academic exercise. Every undescribed insect species embodies a distinct evolutionary history and could be crucial for healthy ecosystems. Insects serve as pollinators, decomposers, predators, and parasitoids, naturally controlling pest populations and being vital for agriculture and environmental health.
Precise estimates of insect diversity assist scientists in measuring biodiversity loss, setting conservation priorities, and understanding the impacts of climate change and habitat destruction on still largely uncatalogued ecosystems.
Looking Ahead
This study demonstrates how DNA barcoding, combined with statistical tools borrowed from fields as diverse as epidemiology, is transforming our understanding of biodiversity. It also serves as a reminder that despite centuries of scientific exploration, much of Earth's insect diversity remains unknown.
As sampling techniques and genetic technologies advance, future research is likely to yield even more accurate estimates. At present, this study offers some of the most compelling evidence to date that Earth's insect diversity is significantly higher than previously believed, and that our cataloguing efforts are still in the early stages.
Reference
Colwell, R. K., Guzman, L. M., Steinke, D., Chao, A., Janzen, D. H., Hallwachs, W., Baker, A., Fernandez-Triana, J. L., Hebert, P. D. N., Joyce, F., Puschendorf, R., Quicke, D. L. J., Rougerie, R., Smith, M. A., Zamora, N., & Sharkey, M. J. (2026). Constructing a lower-bound estimate of the global number of insect species on a hyperdiverse empirical foundation. PNAS, 123(27), e2524283123. https://doi.org/10.1073/pnas.2524283123
About the authors:
Ritu Pandey, Research Scholar, Division of Entomology, ICAR- Indian Agricultural Research Institute, New Delhi, India 110012.
Shashank P. R., Senior Scientist, Division of Entomology, ICAR- Indian Agricultural Research Institute, New Delhi, India 110012.
Email: spathour@gmail.com
Disclaimer: The contents, style, language, plagiarism, references, mention of any products if any, etc., are the sole responsibility of the author.
Comments