The last known captive thylacine died in a Hobart zoo in 1936.
That date became a symbol.
A black-and-white animal pacing behind bars.
Then gone.
Except the sightings did not stop.
Hunters.
Bush workers.
Wildlife professionals.
Drivers.
Residents.
For decades, people kept reporting an animal with:
a stiff tail
striped hindquarters
a strange gait
and a shape unlike an ordinary dog.
The obvious question is irresistible.
What if extinction happened later than we think?
The harder question is:
How much later does the evidence actually support?
The 1936 date is not the same as the extinction date
1936 is the date the last known captive animal died.
The last confirmed wild records were earlier in that decade.
Neither fact tells us the exact day the final wild thylacine died.
Extinction is rarely observed directly.
A species does not announce:
I am the last one.
Researchers usually infer extinction from the pattern of final confirmed records and later uncertain sightings.
That means extinction has an uncertainty window.
For the thylacine, that window may be larger than popular history suggests.
The official position remains presumed extinct
Tasmania's Department of Natural Resources and Environment still lists the thylacine as presumed extinct.
There has been no officially verified wild or captive record for many decades.
Unverified sightings continue.
The difference between those statements matters.
A sighting report can be sincere and interesting.
A verified record requires stronger evidence.
Photograph.
Specimen.
DNA.
An independently confirmed observation.
The species has not crossed that threshold again.
Then researchers built a database of 1,237 records
In 2023, Barry Brook and colleagues published a major reanalysis of thylacine extinction timing.
They assembled 1,237 observational records from Tasmania dating from 1910 onward.
The records were not treated equally.
That is the key.
A physical specimen receives more weight than:
an expert sighting
a public sighting
a track report
or a vague claim.
The researchers then used uncertainty modelling to ask:
Given the quality and timing of the observations, when was extinction most likely?
The result moved the date later
Using high-quality records such as confirmed kills, captures and strong expert observations, the study suggested the species may have persisted for decades after the last captive individual died.
The most likely extinction period based on stronger records fell broadly between the 1940s and 1970s.
When the full database of uncertain sightings was incorporated using probability weights, model outcomes allowed extinction as late as the late 1980s to early 2000s.
That sounds dramatic.
But there is an important final clause.
The probability of persistence into the present was small.
The model did not rediscover the thylacine.
It widened the uncertainty around when the final population disappeared.
This is a probability problem, not a monster story
Imagine three reports.
Report A
A dead animal in hand.
Species confirmed.
Report B
An experienced wildlife professional gets a clear view at close range.
No photograph.
Report C
A driver sees something striped cross a road at night for two seconds.
All three matter differently.
The 2023 study explicitly used that logic.
Uncertain observations were given lower probability weights.
That is a stronger method than either extreme:
"Every sighting proves survival."
or
"Every sighting after 1936 is worthless."
Evidence can be graded.
Why persistence after 1936 is biologically plausible
A species rarely vanishes exactly when the last official specimen is recorded.
Small remnant populations can survive in remote habitat.
Tasmania contains difficult terrain and large wilderness areas.
The thylacine was already rare and heavily persecuted.
If a few animals survived after the 1930s, formal detection could lag behind reality.
That is not extraordinary.
Conservation biology contains many examples where species persisted after experts feared they were gone.
The question is how long a hidden breeding population could remain undetected.
Why survival into 2026 is much harder
A remnant population cannot consist of one immortal animal.
To survive for nearly ninety years after 1936, the species would need:
multiple breeding individuals
successful reproduction
food
territory
and enough genetic viability to continue.
That population would leave traces.
Carcasses.
Scat.
Tracks.
Camera-trap images.
Roadkill.
DNA.
Predation patterns.
Reliable photographs.
The longer the proposed survival window becomes, the greater the expected biological footprint.
This is why "it survived into the 1980s" and "it survives today" are very different claims.
Search technology has improved
Modern wildlife detection is better than it was in 1936.
Camera traps.
Environmental DNA.
Genetic sequencing.
Automated acoustic monitoring.
Drones.
Citizen-science databases.
Large-scale remote sensing.
These tools do not make detection guaranteed.
They increase the pressure on the survival hypothesis.
If a breeding population exists, it has to remain hidden from a progressively stronger surveillance environment.
That is possible in principle.
It becomes harder over time.
Sightings can cluster even when species are gone
This is another subtle point.
Once a species becomes culturally famous, observers are primed to notice animals that resemble it.
Foxes.
Dogs.
Quolls.
Unusual lighting.
Partial views.
A fleeting animal can become a thylacine in memory because that category already exists.
This does not mean every report is wrong.
It means reports after extinction can continue indefinitely.
Folklore can outlive biology.
But false positives do not erase every late report
The 2023 modelling becomes interesting precisely because some post-1936 records came from people with relevant field experience.
The authors did not simply discard them.
They asked what happens when stronger and weaker observations are weighted differently.
That approach creates a more realistic historical picture.
The last captive animal may not have been the last thylacine.
The species could have lingered.
That is a meaningful correction to the standard narrative.
The "rediscovery" examples are real but often misused
Cryptid discussions often list animals once believed extinct and later rediscovered.
The coelacanth.
The Lord Howe stick insect.
The takahe.
This proves an important point:
extinction judgments can be wrong.
It does not prove every claimed extinct species still survives.
Rediscovery requires evidence.
The existence of previous surprises should make us open.
Not careless.
What would change the case instantly?
A modern thylacine confirmation does not require philosophical debate.
A clear tissue sample.
Authenticated DNA.
A carcass.
A living animal documented by independent experts.
Repeated camera-trap evidence with location and metadata.
The evidential bar is high because the claim is biologically important.
But the bar is not mysterious.
If the animal exists, biology can win the argument.
The DarkBrain probability ladder
Established
Thylacines survived at least into the early 1930s, and the last known captive animal died in 1936.
Strongly plausible
Some wild individuals survived beyond the famous 1936 date.
Model-supported possibility
Extinction may have occurred decades later, potentially into the late twentieth century depending on how uncertain sightings are weighted.
Weak current possibility
A tiny population survives today in remote Tasmania.
Not impossible in pure logic.
Poorly supported by authenticated modern evidence.
The DarkBrain conclusion
The thylacine is more interesting than the simple textbook sentence:
Extinct in 1936.
That date marks the last known captive animal.
It may not mark the final wild death.
Modern statistical analysis suggests the species could have survived for decades longer than the traditional story implies.
That is a legitimate scientific revision.
But revision is not resurrection.
The leap from:
"possibly persisted later"
to
"still alive today"
requires evidence we do not yet have.
So the strongest current position is beautifully uncomfortable:
The thylacine may have disappeared later than history remembers, while still being extinct now.
That middle possibility is less sensational than a hidden population.
It may also be much closer to the truth.