Science1 distinct publisher3 min readUpdated
Michigan State and its partners scaled Poweshiek skipperling releases from 102 in 2022 to more than 1,300 in 2025. What that demonstrates, and what the account still does not report.
The Scientist · Science desk

Compiled by The ScientistSomething wrong?How this is made
Researchers at Michigan State University and a group of zoo and agency partners have moved the Poweshiek skipperling from the edge of extinction in the United States to a captive rearing operation that put more than 1,300 butterflies into restored prairie in 2025 [1][3]. The interesting figure is not the total but the slope: 102 released in 2022 and more than 1,300 three years later, roughly a thirteenfold increase in output for a species that had nine documented individuals in the entire country that same year [3][4][16].
The partner list is unglamorous and probably load-bearing: John Ball Zoo, the Minnesota Zoo, the U.S. Fish and Wildlife Service and the Michigan Department of Natural Resources [2]. At one point the program was holding more Poweshiek skipperlings in captivity than existed in the wild anywhere on Earth [5]. That is a statement about how thin the margin was. Early on, the objective was not to found new colonies but simply to keep the butterfly from vanishing from the last few sites in Michigan [6]. Now some releases seed new populations and others reinforce the remnants [7].
Captive rearing on its own would be a holding pattern, and the habitat side of the ledger explains why Michigan is where this worked. The species once ranged from Iowa and the Dakotas to Michigan and southern Canada, then declined with unusual speed in the early 2000s [8]. It now survives at only a few sites in southeast Michigan plus one population in Manitoba [9]. Michigan holds most of the world's remaining prairie fens, the butterfly's habitat, and restoration work by The Nature Conservancy and other land managers is described by the team as the key to its survival [10][18]. Against that, less than 1 percent of the prairie habitat that once existed remains, according to MSU's David Pavlik [11].
The wider trend has not turned. Nick Haddad of MSU says a third of Midwestern butterfly species have declined about 40 percent over 20 years, and none have increased [12]. Habitat loss, wetland destruction and changing agricultural practice are the named causes, with neonicotinoid pesticides adding drift beyond their intended targets [14].
Two cautions about calling this a template. First, the published account reports releases, not outcomes: there are no survivorship figures, no establishment rates for the new populations, and no cost per butterfly [19]. The strongest claim on record is that releases are helping wild populations begin to thrive, which is qualitative [7]. Second, the enabling conditions are specific. Michigan's Great Lakes geography and its rare prairie fens made it a refuge, and John Ball Zoo's Bill Flanagan calls the state a stronghold for these species [15][18]. A program that depends on being the last place the habitat still exists is not trivially portable.
What is portable is the structure, and the team is already testing that. The Haddad lab and John Ball Zoo are now rearing and reintroducing two more endangered butterflies in Michigan, the Mitchell's satyr and the Karner blue [13]. Those two are the actual test of replicability: whether the rearing protocols, the partner arrangement and the restored acreage transfer to species that were not the original project.
Watch for release counts and, more usefully, post-release monitoring numbers for the satyr and the Karner blue; watch whether the Manitoba population responds to anything Michigan does; and watch whether restored prairie fen acreage keeps up with a rearing pipeline that has grown faster than habitat can be rebuilt [13][9][11].
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.
Researchers at Michigan State University and their conservation partners have helped pull the Poweshiek skipperling, one of the world's rarest butterflies, back from the brink of extinction, with releases onto restored prairie in Michigan.
Partners in the recovery effort include John Ball Zoo, the Minnesota Zoo, the U.S. Fish and Wildlife Service and the Michigan Department of Natural Resources.
David Pavlik, research assistant at MSU's W.K. Kellogg Biological Station and John Ball Zoo, said: "We were actually able to save that species from extinction. We released 102 butterflies in 2022 and more than 1,300 butterflies in 2025."
By 2022, researchers documented only nine Poweshiek skipperlings in the entire United States.
At one point during the recovery effort, researchers were caring for more Poweshiek skipperlings in captivity than existed in the wild anywhere on Earth.
Early in the recovery effort the goal was not to create new populations but to keep the butterfly from disappearing from the last few populations left in Michigan.
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.
Single-source, participant-reported counts with no outcome data
The cluster rests on one publisher carrying one institutional account. The core figures are specific and directly attributed to named program participants (102 and 1,300+ releases, nine documented U.S. individuals in 2022, one-third of Midwestern species down ~40%, <1% prairie remaining), which supports the input-side facts. But there is no independent count, no peer-reviewed dataset, and no survival or establishment rate to test the recovery conclusion the account draws from those inputs.
Real, repeated deployment at growing volume but narrow geography
Adoption is genuine and observable rather than announced: releases have run for at least four seasons at rising volume, five named institutions participate including federal and state agencies, and the protocol has already been transferred to two additional species with releases completed this summer. It is capped by geographic concentration - a few southeast Michigan sites plus one Manitoba population - and by the fact that sustained, self-supporting wild populations are asserted rather than demonstrated.
Definitive 'saved from extinction' framing outruns reported outcomes
The account's strongest statements - that a species was saved from extinction and that wild populations are beginning to thrive - are outcome claims, but every number offered is an input count. No post-release survival rate, establishment rate, current wild census or cost figure appears, and the species still occupies only a handful of sites. The gap is moderate rather than severe because the release counts and the 2022 nadir are concrete, specific and attributable.
Program participants describing their own program in an institutional release
All substantive statements come from three people employed by the institutions running the effort - an MSU/John Ball Zoo research assistant, an MSU professor whose lab runs the work, and John Ball Zoo's conservation manager - carried by an outlet that republishes institutional research communications. The organizations have direct funding, reputational and mission incentives to present the effort as a success, and no independent or adversarial voice appears in the cluster.
Internally consistent but unreplicated across publishers
Confidence is limited by cluster structure rather than by internal contradiction: one publisher, one source, no corroborating account. Within the source the figures are consistent and repeated by two independent speakers (the nine-butterfly 2022 nadir is stated in the narrative and again by Flanagan), and the recency is strong, so the input facts can be relied on more than the outcome interpretation.
science
Deep-time extinction predictors flip sign in the present, and 770 data-deficient mammals ride on it1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 14, 2026