The Species Black List
Tracking down Life Overlooked
Tracking down Life Overlooked
What this is
The IUCN Red List is the world's authority on extinction risk, and it has two blind spots. Some species are assessed and come back Data Deficient, meaning IUCN looked and there was not enough information to say. Others are Not Evaluated, meaning nobody has looked at all. In practice both are treated as though the species were safe.
The gap is much bigger than this list. IUCN v2026‑1 records 8,659 vertebrates as Data Deficient, and the number never assessed is larger again; nobody knows it precisely. The Reptile Database alone holds 2,345 reptiles the Red List has no entry for, and a separate count finds 3,088 reptiles, amphibians, birds and mammals described since 2015 that have never been assessed. What follows is a ranked subset of that gap, not a measure of it.
Where this list comes from. One source: Matthew Richardson's Threatened and Recently Extinct Vertebrates of the World (2023). The list inherits that book's scope, vertebrates the book documents, so a Data Deficient species Richardson does not cover will not appear here. That cuts both ways. The book is a compendium of species somebody was already worried about, which is a narrower starting point than "the species nobody counted".
What the record actually says
Neither category is covered by the commitment that counts. Target 4 of the Kunming-Montreal Global Biodiversity Framework binds governments to halting the extinction of known threatened species, wording that excludes both groups by definition. The money follows the same line: under 3.5% of Mohamed bin Zayed Species Conservation Fund awards have gone to Data Deficient taxa, under 1% at the People's Trust for Endangered Species.
IUCN's own guidance says Data Deficient species should receive the same attention as threatened ones. It is not followed. That turns a gap in knowledge into an exclusion written into policy and funding, and it is the reason for compiling this list at all.
What the tiers mean
Every species sits in one of five restriction tiers by how narrow its published record is. That is all the tier measures, and it is the axis the whole list is ranked on: a smaller known range is what qualifies a species as threatened under IUCN Criterion B. Tiers 3, 4 and 5 are the priority stratum, tested against IUCN's own reassessments, and the tested break in risk falls between tier 2 and tier 3.
Every species on the Black List, one mark each
Where these species are in the world is on the map, and the places holding several at once are listed under localities.
Hover a mark to name the species; select one to open its record. How the tiers were built and tested, and what the ranking cannot tell you, is set out in methods and limits.
Held-out test
Between the previous data and v2026‑1, IUCN independently reassessed 108 of these species. The ranking rules never saw those outcomes, so they work as an answer key. Two caveats: only 28 of the reassessed sit in the priority stratum, so the interval matters more than the point; and ranking and assessment read the same primary literature, so this is two readings of one evidence base agreeing, not a blind prediction. Share reassessed as threatened or extinct:
Does that depend on trusting IUCN's “not threatened” calls? A fair question. The backlog is being worked down by specialist consensus rather than new fieldwork, so a species can reach Least Concern because nobody reported a threat. We tested that by throwing out the thinnest calls, and the result gets stronger, not weaker: set aside every call resting on five or fewer specimen records worldwide and 53.6% becomes 62.5%; keep only the reassessments citing new research on that particular species and it reaches 83.3%. The honest concession is that most reassessments cite no new species-specific research at all, on both sides of the ledger, so it does not tilt the comparison. Methods §4a has the tables and the limit of the argument.
Composition
By class
By restriction tier
Every species here was ranked on one sentence about it in Matthew Richardson's Threatened and Recently Extinct Vertebrates of the World (2023). Click any row to read that sentence and see how it was scored.
| Species | Class | Tier | Basis | Locality | Realm | DD since |
|---|
Where museum specimens and field sightings of these species have actually been recorded, for every record that carries map coordinates. The data comes from GBIF, a global archive of biodiversity records. Switch layers to isolate a single restriction tier or view them together. Select any shaded cell or ring to open its species in the Black List. Scroll or pinch to zoom, drag to pan; cells subdivide as you zoom in, so finer clusters resolve rather than the same blocks getting bigger.
A dense patch means records exist there, which is not the same as species being concentrated there. Records pile up around universities, museums, expedition routes and coastlines somebody surveyed. Read the map as a picture of where the evidence sits, not of where the risk is.
These are point records, not range maps, and they come from a different source than the localities on the Localities page, so the two will not correspond exactly. Individual records can also be plain wrong; GBIF's own flagged coordinate errors are excluded, but some survive. That is why this shows density rather than authoritative dots.
The source is organized geographically, so species cluster by place. Where a single locality carries many of them, one protected area could cover the lot, which is the cheapest form of conservation available.
Caveat: these localities are extracted at inconsistent grain. Some are a single named river, some an entire ocean basin. Treat the small, specific ones as actionable and the broad ones as an artifact of the source.
Where a locality has an account of its own in the book, that account's opening sentence is shown, which is what disambiguates a name like Espiritu Santo or Cordillera Central, both of which occur more than once in the world and one of which occurs twice in this book. The rest were recovered from the species' own records, where the place is named in the entry itself; those lines say so. Nothing here is geocoded or inferred. Select a locality to open its species in the Black List.
These are not places. They are habitat categories the source defines and then groups species under: terra firma forest, cerrado, miombo woodland. They are kept because the groupings are real, but no protected area can be drawn around them.