The Species Black List
Tracking down Life Overlooked
Tracking down Life Overlooked
Reptiles described since 2015
Taxonomists have named more than two thousand new reptile species in the past decade. Most have never been evaluated for extinction risk, so no listing protects them and no conservation plan counts them. Nine in ten are endemic to a single country, which is the pattern that best predicts a species turns out to be threatened when someone finally looks.
And the same question, asked of amphibians
The reptile list rests on the Reptile Database; this one rests on AmphibiaWeb's own new-species listing: 1,721 species described from 2015 onward, each checked individually against the Red List by the same method and graded on the same three-point scale.
Amphibians are the better-assessed class: … of those 1,721 already carry an assessment, where most recent reptiles do not. The Global Amphibian Assessment has been run twice and reptiles have no equivalent. What is left is still … species with no evaluation at all, and … of them are endemic to a single country.
All four classes on one map, in the same encoding as the lost-species map so the three can be read against each other: countries shaded by how many of these species they hold, and a circle where a place is not a country. The table below lists all four classes together.
A backlog is a pile you can work through. This is a tap that has been left on. Clearing every species on both lists tomorrow would leave the same number arriving the year after, so the question is not when will somebody catch up, but what would it take to keep up at all.
Where a class's source states a range, the extent column grades it; where it does not, the column says so rather than guessing. AmphibiaWeb gives a country and never a locality, and the Mammal Diversity Database gives a country list and no state-level range, so most amphibians and mammals read "not stated". Every species here was checked against the Red List's own API by exact name; where a name resolves only approximately, we treat it as unmatched rather than borrow a neighbouring species' answer.
Four classes, four checklists. Each is the standard reference for its group, and each is compared against the Red List to leave the species it has no record of.
| Reptiles | Reptile Database, June 2026, against IUCN v2026-1 |
|---|---|
| Amphibians | AmphibiaWeb, its own new-species listing |
| Birds | IOC World Bird List v15.2 |
| Mammals | Mammal Diversity Database |
Comparing names alone would be unreliable. A species often gets moved into a different genus, so its first name changes while the animal stays the same, and the two organizations do not always agree on where it belongs. So reptiles are matched either on the full scientific name or on who first described it and in what year. Those are details that stay with a species through every change of name. That caught 169 species a name-only comparison would have wrongly reported as missing.
Birds and mammals are asked of the Red List's own API one species at a time rather than matched against a downloaded copy of it. GBIF's backbone runs about two years behind the mammal database, so every mammal named in the last two years is missing from it, and a comparison built on GBIF would have called all of them unassessed without checking.
Most of these species have no common name, and that is not a small detail. … of the … here carry an English name in any of the four checklists, iNaturalist or GBIF. A frog described in 2019 usually has only a Latin one. It matters because searching for a species means searching for its name: when we asked every conservation organization we could find which of these species they mention, this list scored lowest of the four, and part of that is our vocabulary rather than the world's neglect. The names that are missing are mostly waiting in the papers the species were described in.
The four are not equally rigorous, and the page does not pretend they are. The Reptile Database publishes a full checklist with authorities and years, which is what makes the author-and-year match possible. AmphibiaWeb's new-species listing is a specialist source but a thinner one, so amphibian rows carry a reference and little else. Mammal figures count species-country pairs rather than species, because the mammal database lists every country a species occurs in and 76 of 372 list more than one. That is right for a shaded map, where the animal really does occur in all of them, and it is why the mammal label reads "records" where the others read "species".
Failing to find a name in a search does not prove nobody ever assessed the species, because it might be filed under a different name. So every species here was looked up individually in the Red List's own database, and every name it did not return was checked again against GBIF's list of alternative names for the same species. That found 51 which had been assessed after all, and they were removed: an error rate of 2.9% in the list before checking.
The evidence is graded, because it is not equally strong for every species:
Extent is taken from each species' own published distribution: the country, state or landform it is recorded from. That is the measure the evidence supports : species endemic to a single country turn out to be threatened roughly twice as often as those crossing borders (27.2% against 13.2%, p = 0.00005, tested on 600 reptiles the Red List has assessed).
A finer, coordinate-based scale was tried first and abandoned. Its top class, living only on an island, turned out to predict nothing at all : island species were no more likely to be threatened than any other (odds ratio 0.92, p = 1: as close to no effect as a result gets). The reason is that “island” taken from a written description covers a one-square-kilometer rock and Hainan's 34,000 km² alike. Geocoding the locality text could not rescue it either: measured against species whose descriptions print coordinates, it was off by a median of 53 km and in the worst case 5,469 km.
The “finer detail” column counts how many records carry map coordinates, where GBIF has any. It is descriptive only and has not been validated against Red List outcomes. For a species described last year, few records may mean a small range, or may only mean nobody has looked twice.
This page gives each species' distribution and no more. Exact type localities and coordinates are deliberately absent, including from the downloadable data.
These are narrowly restricted reptiles the Red List has never listed, so no legal protection follows from a listing, and island endemics among them are exactly what the reptile trade targets. Scattered across separate descriptions their localities are hard to exploit; gathered into one file they become a collecting list. Describers know this, and obscuring a type locality in a new reptile description is common practice for the same reason. Researchers who need locality data should go to the original descriptions, where the authors decided for themselves what to disclose.
The list asks which assessed species IUCN looked at and returned we don't know about, and its ranking is tested against IUCN's own later reassessments. This page asks who is missing from the Red List entirely. It draws on a different source and uses a different instrument, so the two lists must never be added together.