The mystery of the Tasmanian Tiger Apextalk

Or again, should those words be “was” and “carried” and so on—that is, should that paragraph have been written in the past tense? Because the last Tasmanian Tiger died many decades ago. Seen as pests and a threat to livestock on the island of Tasmania, they were widely hunted. The last of the species died at the zoo in Hobart, Tasmania, in 1936. Pest or not, the thylacine is such a cultural icon in Australia that the country observes National Threatened Species Day every 7 September, the day in 1936 when the thylacine went into extinction.

Or did it? Reminiscent of the Loch Ness Monster, there have been plenty of reports of sightings of thylacine since 1936.

For example, one night in 1982, a wildlife ranger called Hans Naarding was sleeping in his car in a forest in a remote part of Tasmania. Heavy rain woke him at 2 am. He turned on his torch and swept the beam around him. To his astonishment, the beam caught a thylacine, about six or seven metres away. It was an adult male, he reported later, “with 12 black stripes on a sandy coat”.

Then in July 2019, a hiker was climbing up to Sleeping Beauty Mountain, west of Hobart. He saw a footprint that he later Googled. He was convinced it was made by a thylacine.

These were just two of over 1,200 reported sightings of Tasmanian Tigers all over Tasmania between 1910 and 2019. Barry Brook, a mammal ecologist at the University of Tasmania, has compiled all these reports into what is known as the Tasmanian Thylacine Sighting Records Database.

Why am I telling you about this possibly extinct animal, and this database, in a column that’s ostensibly about mathematics?

Because Brooks and his colleagues used that database to estimate the thylacine’s date of extinction. The date they suggest is a serious surprise. More about that in a bit. But intriguingly, they came to their conclusion after a statistical analysis of the database (Extinction of the Thylacine, Barry Brook et al., bioRxiv, 19 January 2021, https://bit.ly/3zTeHMz).

As they write in their paper, they “collate and characterize the type, quality and uncertainty” of all those reported sightings in the database. They dated each one and located it on a map. They broke them down into categories, like those that were actual specimens of the animal, or just signs of its presence (like the Sleeping Beauty footprint). They classified the sightings according to the person reporting it: was she an expert in the field —a trapper, a scientist?—or just a casual observer like a hiker, or you or me? They looked for “spikes” in the reporting numbers; one they found, in 1970, probably happened because of “media attention linked to a well-publicized expedition” to search for the thylacine. Brooks referred to this as “recency bias”.

Put simply, all the sighting reports in the database are not equally reliable. As Brooks and his team commented, trying to decide the reliability of the sightings is “the most vexing difficulty with a scientific mystery like this”.

For example, clearly an actual sighting of the animal—like Hans Naarding reported—should carry more weight than a mere footprint. And in fact, Brook’s team gave the Sleeping Beauty sighting a low rating, meaning it was unlikely to be true. Then possible “copycat” sightings, or recency bias, should be rated lower than others. But again, there are some clusters of sightings that are close in both space and time, with similar descriptions of the animal, but that may not be copycats. If these were multiple sightings of the same thylacine by different people, that must warrant a higher rating.

With detailed and careful analysis, Brook and his co-researchers were able to give each report in the database a rating.

To more fully comprehend what they were attempting, think about what we mean when we say a particular species went extinct at a specific time. Such a conclusion follows an essentially statistical process. The sightings grow rarer, and after a point stop altogether. But if a reported sighting doesn’t involve the actual capture of the animal, then there is a question about how much weight it should carry. In effect, what is the probability that that sighting was authentic? These probabilities feed into certain mathematical techniques, known as “extinction date estimators” (EDE).

Brook and his team used their ratings with two different EDEs. Both are able to work with “cases where sighting frequency declines over time and (there are) sparse sighting records”.

First, though, they fed the EDEs of only those sightings that reported a physical specimen of the Tasmanian Tiger. No surprise, the dates that popped up were pre-1940, matching the 1936 death of the last-known thylacine.

But then they opened up the EDEs to “unverified expert reports and other opportunistic sightings by the public”. While of course these were not as reliable as physical specimen sightings, they represented a range of ratings. What the EDEs then reported was a great surprise: “The extinction window spans a later period, from the 1980s to the present.”

That is, this “uncertainty modelling” suggests that the thylacine died out much later than 1936, possibly as recently as the early years of the 21st century. But if that came as a great shock to wildlife conservationists, there was something still more astonishing this analysis threw up.

Brooks thinks his model offers a chance —if a small one—that in some remote wilderness area of south-western Tasmania, the thylacine lives on even today.

Mathematics that tells us things about extinction: It’s enough to make me want to go searching for the Tasmanian Tiger in its island homeland. One of these days.

Once a computer scientist, Dilip D’Souza now lives in Mumbai and writes for his dinners. His Twitter handle is @DeathEndsFun.