Using Artec 3D scanners to discover the true size of the now-extinct thylacine


By Matthew McMillion

Millions of years of evolution had transformed the thylacine into a specialist of survival. As the largest carnivorous marsupial in existence, the thylacine had carved out its own unique ecological niche. But then the first settlers showed up.

Lead researcher Douglass Rovinsky scanning a thylacine skull with Artec Space Spider at the Queen Victoria Museum & Art Gallery

Just a few decades later, starting in the early 1830s and continuing into the next century, thousands of thylacines, often referred to as “Tasmanian tigers,” were deliberately exterminated by farmers and hunters in a widespread bounty scheme. Until there was only one left.

He was captured in the wild and spent the last three years of his life locked away in a zoo. Then, on a late winter’s morning, Benjamin, Tasmania’s last known thylacine, died after being locked out of his shelter for the night. That was September of 1936. Only 59 days earlier did Tasmania finally grant protected status to the thylacine.

Digital recreation of a thylacine by researcher and digital artist Damir Martin

For many years leading up to that morning, all kinds of false, highly-exaggerated stories about the thylacine’s supposed viciousness arose and spread like wildfire, including stories about it having berserker-like strength, being able to withstand shotgun blasts, and even draining its victims entirely of blood, etc.

At the same time, while thylacines were still with us, roaming across the Tasmanian landscape, scientists failed to study them in any real depth. Even today, less than a century after Benjamin’s passing, despite a significant level of interest in the animal, there are facts and details about it that scientists simply don’t know.

Digital recreation of a thylacine pair by Damir Martin

Details such as the thylacine’s diet, its mating habits, the way it hunted, the way it moved, how it interacted with its landscape, and even how much the thylacine weighed, its body mass. The body mass of an animal is one of the most foundational aspects to determine if you want to study and gain a scientific understanding of that animal. This comes down to basic chemistry and physiology.

How much energy did it burn? Did it retain heat well? How quickly did it digest food? And how often did it need to eat? What kinds of animals did it hunt? How well did it thrive in its environment? To know these things with a high level of confidence, we first need to understand how much the thylacine weighed.

Science is always built upon the bedrock of other science, which means that unless we’re able to achieve an accurate and truthful understanding of “Concept A,” not only will we repeatedly misunderstand “Concept A,” but we also won’t be able to make sense of everything that relates back to that concept as a foundation.

Digital recreation of a thylacine pair by Damir Martin

Throughout decades of thylacine research, several scientists have remarked about the lack of an accurate body mass estimate of the animal. But after that, nothing was done to rectify the situation. There was an often-quoted assumption that thylacines weighed in at around 25 to 29 kilograms (55-63.8 lb). But no one was certain.

Another serious aspect of the thylacine that needed to be determined was whether or not it violated what’s termed, “The costs of carnivory.” This is essentially an energy budget threshold that defines how carnivores around 14 kg (30.9 lb) or smaller, out of necessity, tend to feed on prey much smaller than themselves.

Whereas carnivores around 21 kg (46.3 lb) or larger, such as wolves and jaguars, will usually go after equally-sized prey or larger. For carnivores that fall into the 14-21 kg (30.9-46.3 lb) range, for example, foxes and wildcats, they usually focus on smaller prey, but can occasionally take down larger animals as well.

Digital recreation of a thylacine by Damir Martin

Without an accurate understanding of the thylacine’s true body mass range, this aspect, along with handfuls of others, were simply unable to be scientifically determined with any degree of confidence.

Recognizing the enormous need for this, Monash University PhD candidate Douglass Rovinsky and a team of three other researchers, including Dr. Justin W. Adams, embarked on a project that would change that.

The details of the project have been published in the scientific journal Proceedings of the Royal Society B, under the title, “Did the thylacine violate the costs of carnivory? Body mass and sexual dimorphism of an iconic Australian marsupial.”

Rovinsky explained the importance of the project, “The more we know about the thylacine, like all extinct animals, the more we can understand how extant animals, the ones alive right now, will respond to changes in the environment, which are happening at a staggering rate both globally and locally.”

Their research would take them around the world, where in museums and institutions, including the Smithsonian and many others east and west, they would work directly with the few remaining thylacine specimens in existence.

Rovinsky and his team understood early on that solely relying on the traditional body mass estimation method of linear regression, made via measuring an animal’s teeth, was more often than not highly inaccurate when it comes to extinct animals.

To properly do this, there must be a highly-similar living relative of the animal. In the case of the thylacine, the closest living relative is the numbat, a 1-pound (.5 kg), furry little creature that feeds on termites. To make any proportional comparisons between the two animals is simply out of the question.

Rovinsky and his team decided to combine the results from multiple body mass estimation methods, including: linear regression of the teeth as well as the upper arm/upper leg bones, building a convex hull around a 3D-scanned thylacine skeleton, then sculpting lifelike representations of a thylacine over a digital scan, and finally, digitally weighing both scanned taxidermied mounts as well as the sculpted thylacines (over the same scanned skeletons that the convex hulls were made from).

Top row from left: thylacine skeleton scan with Artec Leo & convex hull volumetric model constructed over the skeleton for digital weighing
Bottom row: Sculpted volumetric 3D thylacine models over the skeleton (with and without texture) used for digital weighing

In order to properly carry out this immense project, prior to the multi-phase spectrum of analyses being done, hundreds of thylacine specimens would need to be 3D scanned and turned into extremely-precise 3D models suitable for unquestionably accurate measurement.

Some of the challenges faced by Rovinsky and his team were due to logistics: all the specimens were spread across 18 different museums and institutes, from the local museum in Melbourne, various museums in Australia and Tasmania, several across the ocean in the US, and others throughout Europe and the UK.

To digitally capture the thylacine specimens for the body mass analyses during these trips, Rovinsky and Adams used a lightweight, submillimeter-precise handheld 3D scanner, the Artec Space Spider. The two researchers separately visited these museums and institutes and scanned hundreds of specimens: bones, skulls, full skeletons, taxidermied thylacines, in addition to a preserved full-body thylacine.

Dr. Justin W. Adams scanning a preserved female thylacine specimen with Artec Space Spider at the Swedish Museum of Natural History

Even before the project had commenced, Rovinsky and Adams decided against the traditional measurement methods of calipers and measuring tools, as well as more modern methods such as GDI and photogrammetry. Time was limited at each museum visited. As were unnecessary risks to the specimens from excessive handling.

Those previous methods would require extensive time to carry out, not to mention tying up the museum staff for hours each day as they supervised. And the results would doubtfully even match, much less exceed, those achieved via 3D scanning with the Artec Space Spider.

The risks of damage while using the earlier methods arise from having to repeatedly pick up and reposition the specimens during the measurement process. With the Artec Space Spider, data capture is accurate up to .05mm, dozens of times faster than these other methods, with little-to-no handling required.

Dr. Justin W. Adams at the Swedish Museum of Natural History, capturing a preserved female thylacine specimen in submillimeter 3D with Artec Space Spider

In contrast with their earlier experiences, museum staff were, in Rovinsky’s words, “ecstatic” with the brief scanning times needed, coupled with the extremely low risk to the specimens, both of which meant they weren’t distracted from their normal work schedule for more than a few minutes. Even so, museum curators and staff, usually brimming over with curiosity, would frequently stay to observe Space Spider in action.

They’re so impressed to see how fast and easily the Space Spider works. Because, compared to other visiting researchers who aren’t using a Space Spider, we don’t need a lot in order to do our work. We don’t need a big space; we don’t need controlled lighting; and we don’t need to take up much of their time,” said Rovinsky.

Referring to how surprised museum staff can be prior to seeing a Space Spider in action, Rovinsky shared, “They’ll say to us, here’s your box of 12 specimens, and we reply to them, okay I’ll be done with this in about 2 hours, so when can I get the next 12? Let’s just say they’re used to something that takes a lot longer to gather the data.”

In terms of how Rovinsky captures animal specimens with Space Spider, he’s refined his process over time. When he’s scanning something simple, such as an upper arm bone of a wolf or thylacine, he said, “It’s relatively easy and fast, and takes just a few minutes for each specimen. I usually capture everything in 3 scans. So, I put the object on my portable little turntable and scan it for a couple of revolutions. I aim for something like 400 or so frames per scan pass. Then I turn over the object and do it again.”

Lead researcher Douglass Rovinsky scanning a thylacine skull with Artec Space Spider at the Queen Victoria Museum & Art Gallery

When it comes to scanning skulls, some are quite straightforward, while others are more difficult to capture in their entirety. “Especially for the larger skulls, which can take up to 9 scans to get everything. The challenges with those are the hard-to-reach undercuts, which often require multiple scans to fully capture all of the jaw bone, the cheekbones, the orbits, etc.,” said Rovinsky.

As for processing the scans in Artec Studio, Rovinsky always takes care of this after leaving the museum, “All of the editing and fusing are things that I do at later points. I’ve gotten used to doing it that way, to be as minimally invasive to the museum staff as possible. When I get to the collection, I just scan everything I need. Then I go away. And all of the scan processing is something I do after the fact.”

He continued, “I can just take the Space Spider in and over the course of one afternoon, I’ve fully scanned 20 or 30 specimens. This means that when I go on a data collection trip, a few weeks later, I literally come back home with hundreds of specimen scans that are incredibly sharp and accurate, with more than enough detail for us to do our work.”

Rovinsky explained his process in Artec Studio: “First I manually align my scans, and then do a global registration with geometry only. After that, I do a sharp fusion, and the resolution I choose depends on the size of the object. Basically, for objects the size of a dog’s head and smaller, I fuse them at .1, and if it’s bigger than that, I fuse it at .2, generally speaking.”

He went on, “Then I run one or two passes of the smoothing algorithm. Following this, I do a fast mesh simplification down to 600,000 triangles. Ultimately, even 600,000 triangles is more than enough in terms of triangle count. Sure, there are resolution differences between a 600,000-triangle skull and a 1.5-million-triangle skull, but whether or not those differences are meaningful for paleontology purposes is another story.”

For a subsequent phase of the project, Rovinsky worked with Artec 3D Gold Certified Partner Thinglab in Melbourne to scan full thylacine specimens in the Tasmanian Museum and Art Gallery using Artec Leo, a 100% wireless 3D scanner with a built-in touchscreen. These scans captured several of the museum’s thylacine skeletons and full-body taxidermied mounts. The digital skeletons were then used to construct both the convex hull models and to serve as starting points for the digital sculptures.

Thinglab’s Ben Myers scanning a thylacine skeleton with Artec Leo in the Tasmanian Museum and Art Gallery

In the words of Ben Myers, Thinglab’s director of 3D scanning, “We love the Leo and have utilised it for a range of projects, as have our customers. It has many features that we here at Thinglab have been impressed by. The VCSEL light technology makes a huge difference in terms of registering surfaces that in the past were problematic.”

3D model of a thylacine skeleton scanned in the Tasmanian Museum and Art Gallery with Artec Leo

He continued, “The tracking is incredibly impressive, which makes it possible to move around objects, even those with challenging geometries, and capture every single surface with ease. Of course, being wireless and having an inbuilt display makes the entire scanning process far, far easier and more efficient.”

Thinglab’s Ben Myers scanning a taxidermied thylacine specimen with Artec Leo in the Tasmanian Museum and Art Gallery

For one of the final phases of the project, Rovinsky turned to Damir Martin, the digital artist member of the team. Martin, adept at meticulously depicting the paleo world, set out to create the most true-to-life 3D models of the thylacine ever made.

It must be noted that even before this project, Martin had already studied the thylacine in depth and had created numerous breathtaking images of the creature in its natural environments throughout Tasmania.

Martin used the Leo scans of the mounted skeletons and taxidermied mounts as a basis for his work. Then, with his consummate understanding of the thylacine, as well as input from others, including Rovinsky and Adams, step by step, he digitally sculpted each creature’s musculature and exterior appearance to life in ZBrush. Once the models were ready, they were digitally weighed, and the data from these measurements were carried forward along with the other body mass estimates.

Anatomically-precise 3D model of a thylacine, created by Damir Martin

Martin commented on his work: “I’ve always had a deep interest in extinct life. The thylacine is a special animal in so many ways, with such a mysterious and unique appearance. Unfortunately, most restorations and reconstructions that have been done, from an artistic standpoint, failed to capture the subtle nuances of the living animal.”

“But we actually know how the thylacine looked in real life, thanks to some rare photos and footage that survived. So, the very challenge of the project was another appealing factor for me.”

With all the data from the body mass estimation calculations (via 93 thylacines using 207 scans) in hand and brought together, Rovinsky and his team determined that the actual weight of the thylacine was approximately 19 kg (41.9 lb) for males and 14 kg (30.9 lb) for females. That means that males were about 30% larger than females.

In comparison with the decades-long assumptions about the thylacine’s body mass estimates, Rovinsky and his researchers’ body mass estimations for mixed-sex thylacines are around 55% of the earlier body mass conclusions. Consequently, it was obvious that the thylacine didn’t violate “the costs of carnivory,” since it was in the middle-range of carnivores, firmly within the 14-21 kg (30.9-46.3 lb) range that usually preyed on smaller creatures than itself, while occasionally taking down larger prey.

This successful project is one facet of a larger research focus on the evolutionary biology of the thylacine, including this present study on body mass and extending into more detailed investigations of their diet, locomotion, and overall biology using globally-gathered samples of 3D data.

Adams shared his views on how crucial 3D scanning is for such work: “Paleontologists are finally starting to realize that three-dimensional methods create a far more accurate body mass estimation across-the-board. They simply work better, because even if you’re measuring this digitally in the computer, you’re measuring the mass of the shape of the object, as opposed to trying to plot two variables on a regression line.”

3D technology is revolutionizing our understanding of estimating the size of extinct animals. I think that, moving forward, we’re going to see these technologies being used for this a lot more often. And when you consider the ease and speed with which the Artec Space Spider or Leo allows for this type of precise data to be gathered, you just can’t underestimate that.”

Rovinsky commented on the pivotal role that data played, “The difficulty comes with the fact that you absolutely need enough data in order to generate a conclusive 3D representation of the organism, which for most paleontologists, you just don’t have.”

“I couldn’t have attained this degree of accuracy if there had been just a limited number of specimens to measure, and if I didn’t have access to full skeletons to work from. Most paleontologists don’t even have a complete bone to work with, let alone a full skeleton. That said, the Artec scanners made it easy for us to collect the massive amounts of specimen data we needed.”

Rovinsky elaborated on the project’s significance, “The more that we can get our understanding of the thylacine closer to the actuality of it, the reality of the amazing animal that it was, all the better. Because at the end of the day, all data, all observations, all of our understanding, it’s all secondhand at best, since we, unfortunately, don’t have the thylacine with us any longer.”

He continued, “All of our research and data is filtered through our observations, our interpretations, our measurements, etc. That makes it even more crucial for us to know something as fundamental as body mass. Because this one seemingly-minor aspect of this animal affects countless volumes of research that we’re building off our understanding of this factor.”