Science1 distinct publisher3 min readUpdated
A Brazilian-led study in npj Biodiversity counts 1,116 mammal species described since 1990 and where their reference specimens ended up. The answer is now a number institutions can be asked about.
The Scientist · Science desk

Compiled by The ScientistSomething wrong?How this is made
A study by Brazilian scientists published in npj Biodiversity has counted the 1,116 mammal species described between 1990 and 2025 and asked a simple question about each one: where is the holotype [1][2]. Of those species, 95% were native to the Global South, while 60% of the holotypes are held in wealthy countries in Europe and North America [3][4]. That turns a familiar complaint into an auditable figure, which is a different kind of problem for a collections manager or a funder.
The asymmetry is visible in both directions. For species native to the Global North described in the same window, only 22% have holotypes stored outside their country of origin, and those tend to sit within the same geopolitical bloc, which the authors note can include Australia and New Zealand [5]. Looked at from the shelf rather than the field, more than 90% of all mammal holotypes held in the Global North were collected in a different country; in the Global South, fewer than 8% of mammal holotypes came from somewhere else [6][7].
Mammals are worse than comparable groups. Among freshwater fish described after 2000, 30% have holotypes outside the country of origin; among reptiles described after 1990, the figure is 50% [8][9]. According to Mario Moura of the Federal University of Paraiba, the study's first author, the mammal numbers were a surprise: colonial-era collecting was expected, but not persistence at this level in recent decades, and he describes the pattern as scientific neocolonialism being more pronounced in mammals than in other animal groups [10][11].
Brazil is the counterexample inside the dataset. It led the period with 133 new mammal species described, and 108 of those holotypes stayed in national institutions [12][13], which is about 81% retention [14] on roughly 12% of all descriptions worldwide [15].
The reason this matters operationally is that holotypes are working equipment, not trophies. Anyone describing a candidate new species must compare it against the holotype of the closest known relative to confirm the find [16]. Co-author Jhonny Guedes says accessing material from one's own country when it is held abroad requires resources that are not always available, which delays the generation of knowledge [17]. The authors place the results in the context of parachute science, in which researchers from the Global North collect material in the South and deposit and study it at home, often without local collaborators [18]. Moura's group previously found parachute science accounted for most descriptions of terrestrial molluscs [19], and another study he supervised found Northern countries published more robust work on Southern biodiversity, partly on the strength of better analytical equipment [20].
Two things to watch. First, Brazil's repatriation track: one set of dinosaur fossils has already been returned and negotiations are under way for at least one more [21], which will indicate whether specimen return is becoming routine or stays case-by-case. Second, whether the 60/40 holotype split starts appearing as a reported statistic per institution and per grant, since the study has made it cheap to compute for any taxon with a described-species list.
Follow any of these and your For You feed starts watching them — no settings page required.
Ranked by verification strength, evidence, and original report placement.
The study concluded that 95% of the 1,116 species were native to the Global South, a group of emerging countries in Latin America, Africa, Asia and the Caribbean.
60% of the species' holotypes are housed primarily in wealthy countries in Europe and North America.
Of Brazil's 133 new mammal species, 108 holotypes (more than 80%) remained in national institutions.
Co-author Jhonny Guedes said: "Accessing material from one's own country that's held at a foreign institution requires resources that aren't always available, which delays the generation of knowledge."
A study by Brazilian scientists was published in the journal npj Biodiversity on the housing of newly described mammal species' holotypes.
The study analyzed 1,116 new mammal species described between 1990 and 2025.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Peer-reviewed counts, single-outlet relay
The core findings are census-style: a defined sample (1,116 species, 1990-2025), explicit shares (95% Southern origin, 60% Northern custody), a directional asymmetry (>90% versus <8%), cross-taxon benchmarks (30% fish, 50% reptiles) and country-level examples. That is strong, checkable evidence for the descriptive claim. It is discounted because the cluster contains one publisher relaying the paper, the country breakdown is truncated mid-sentence, and the two supporting prior studies are uncited.
One repatriation completed, no institutional uptake recorded
The only concrete downstream action in the cluster is Brazil's repatriation activity: one set of fossils returned and negotiations under way for at least one more, with no institutions, volumes or dates given. No museum policy change, loan-programme commitment, funder response or deposition-rule change is reported, and the study itself has only just been made public. Adoption is therefore real but minimal.
Interpretive framing slightly ahead of what the counts show
The numbers are modest in their claim - they record where specimens are held - while the surrounding language ('scientific neocolonialism', 'parachute science', 'often without involving local researchers') asserts intent and process that the custody counts alone do not establish; no collector-behaviour data, permit history or co-authorship analysis is presented in the cluster. The overshoot is small because the descriptive core is well quantified and the piece includes a counterexample (Brazil retaining ~81%) rather than flattening the picture.
Author- and funder-adjacent framing, no adversarial voice
The item reads as an institutional research communication: author affiliations and funding lineage are foregrounded (UFPB, IB-UNICAMP, FAPESP Young Investigator Program), all quotes come from the study's own authors, and the piece explicitly links the findings to Brazil's active repatriation agenda - a policy position the authors' institutions benefit from advancing. No holding institution, funder or independent taxonomist is given space to contest. This describes alignment of interests around the framing, not error in the counts.
Solid on the numbers, thin on corroboration
Confidence is moderate: the descriptive claims are precise, internally consistent and traceable to a named peer-reviewed paper, and the derived Brazilian shares check out arithmetically. It is held down by the single-publisher cluster, the truncated country-level data, the absence of any holder-side or independent voice, and two supporting claims that the cluster cannot substantiate.
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 18, 2026