Does the moa still live in New Zealand's forests? From a blurry photo to bone dating

The moa were nine species of wingless birds that lived from the coast to the mountains of New Zealand; the largest could reach up to about three metres. All of them most likely became extinct around 1400, after a few centuries of hunting and habitat change. Reports from the European period and a blurry photograph from 1993 have kept alive the hope that a few individuals remain in the South Island forests. But bone dating, archaeology, ecology and the absence of fresh traces all argue against survival into modern times.

The moa was not a single species

Te Papa recognises nine species in three families, from the little bush moa, about the size of a large dog, to the giant moa. In some species the females weighed twice as much as the males, which once led to the bones of the two sexes being classified as different species. Ancient DNA and morphology resolved the confusion. The moa was unique in losing its wing bones entirely, not merely reducing the wings.

Old reconstructions usually set the neck upright like an ostrich. Skeletal research indicates a low head posture with the back nearly horizontal, closer to a cassowary. Getting the posture wrong affects how sightings are assessed: a witness expecting a tall, long-necked bird may overlook the real shape, or conversely turn a deer standing in the scrub into a moa.

Where does the extinction date come from?

Hunting sites, bones with cut marks, cooking ovens and eggshell show that Polynesian people hunted moa intensively after arriving in Aotearoa. Large birds breed slowly, had not evolved alongside human hunters, and were easily exploited. Te Papa summarises that all of them probably disappeared around 1400. A few later dates have been proposed, but they require checks for contamination and context.

Bones in caves can preserve skin, feathers and tissue for hundreds or thousands of years, looking surprisingly fresh. "Not fossilised" does not mean the animal died recently; cold, dry conditions preserve the material. Radiocarbon dating and stratigraphy matter far more than how something looks to the eye.

The 1993 photograph and the sightings

In 1993, three people walking in the bush at Craigieburn said they saw a large bird and took one photograph from behind. The blurred image shows brown legs and a body; supporters identify a moa, others see a red deer or an object. No feathers, droppings or well-preserved tracks were recovered. A single frame without the head and without scale cannot settle the identification.

Nineteenth-century reports often appeared when forest areas were still barely surveyed, but many are second-hand. If moa survived for a few generations after 1400, that does not mean they survived to 1993. Each source has to be placed on a timeline, rather than stringing every account together into one population lasting six hundred years.

Is the rediscovery of the takahe a precedent?

The takahe was once thought extinct until Geoffrey Orbell found a population in the Murchison Mountains in 1948. That case proves science can be wrong and that flightless birds can live unnoticed. But the takahe is far smaller, its population sat within a narrow habitat, and a search following actual signs produced observable individuals. A precedent gives a reason to survey; it does not transfer its probability directly to the moa.

Large moa ate a great deal of vegetation and created trails, droppings and eggshell. New Zealand has an extensive network of pest control, hunters, cameras and conservation staff. Deer, pigs and people moving through the forest leave traces continuously; a moa population would be hard to keep out of that data.

Could environmental DNA end the argument?

eDNA from soil, water and droppings could look for moa sequences, provided the design avoids ancient DNA from the sediment. That is a particular challenge here: New Zealand is full of moa bones and past moa DNA, so a positive result would have to demonstrate new material rather than ancient particles washed into a stream. RNA, fresh droppings, undegraded cells and repeated observation would carry far more weight than a short DNA fragment.

To date there is no modern evidence. The strangeness does not require a surviving individual: that humans met an entire radiation of nine species and made it disappear in a short span is haunting enough. The blurred photographs should remind us to search carefully, while the bones and the dates remind us not to let hope stand in for evidence.

"Still alive" is a hypothesis that has to pay for itself in evidence

A species that once existed is quite different from a creature that has never had a specimen. Bones and historical footage give us an identification standard, but they also create a bias: the observer knows exactly which silhouette they hope to see. When a species has become extremely rare, survival for a while after the last specimen is genuinely possible. The further the hypothesis is stretched towards the present, the more it has to explain why generations have left no carcass, droppings, hair or clear photograph.

The "extinction" date in a book is usually the date of the last specimen, the last accepted sighting, or the date an agency made a declaration - three different markers. The last biological individual may have died somewhere between them. Research should say which marker it is using, so that an administrative year does not turn into a mystical instant when a whole species vanished.

Distinguishing a failed survey from evidence of absence

Finding nothing on one short trip says very little. The value rises when a survey covers the right habitat, in the right season, using methods with a known detection probability, and is repeated over several years. A camera mounted high will miss a ground bird; a microphone on the wrong frequency will miss the call; eDNA taken from a watercourse not connected to the habitat can come back negative even if the animal is there.

When the design is sensitive enough, a run of negative results becomes evidence in itself. The probabilities matter: if an individual has a 20% chance of being recorded in each campaign, ten independent empty campaigns are a great deal less likely. "Absence of proof is not proof of absence" is logically correct, but it does not mean absence can never weaken a hypothesis.

How have camera traps changed the hunt?

Infrared cameras run for months and store the time, the location and a sequence of frames. They have recorded many species more often than people have met them directly. Night images, though, distort motion, produce eyeshine and shift colours; a single frame easily turns a familiar animal into something strange. Consecutive video gives far better information about gait and proportion.

The data has to be archived in its original form, together with the configuration, the mounting height and a reference image taken at the site. If only a cropped picture is shared, the viewer has no idea how far the subject was from the lens. A good system uses two crossed cameras, collects DNA along the trail, and lets local experts make the identification before anything goes online.

eDNA is powerful, but it is not magic

Animals leave cells, hair, droppings and fluids in soil and water. Metabarcoding can identify many species from a single sample. Detection depends on the primers, the reference database, temperature, water flow and density. A negative result has to come with the number of samples and the controls; a positive one has to rule out contamination from the laboratory, a museum or food products.

For an extinct species with many bones in the ground, ancient DNA can generate a false positive for modern life. Fragment length, chemical damage, the sediment layer and RNA all help tell them apart. A very short fragment that is "98% similar" is not enough; raw data and independent replication are what turn a signal into a discovery.

Which way does size in a photograph usually go wrong?

Without anything for scale, a viewer assumes the subject and the background are the same distance away. A cat on a wall close to the lens can look the size of a sheep far away; a bird near the lens looks broader than an aeroplane; waves in a line join into one long back. Digital zoom blurs the edges and the brain completes the shape. Size is only trustworthy when the focal length and distance are known, or when there is a structure in the same plane.

A witness's estimate is also anchored by the question. Asking how tall the animal was after calling it a beast produces a larger number than asking which branch it reached. A good interview uses maps, sight lines and landmarks before showing the teller any pictures of candidates.

A population cannot consist of a final individual alone

The story usually imagines one old animal living discreetly. But for a species to persist for decades you need males, females, young and genetic diversity. More individuals means more road crossings, more deaths, more animals caught in traps and more traces left. A tiny population also faces inbreeding, natural disasters and swings in the sex ratio.

An energy analysis asks how many animals the environment can support. A large predator needs prey and leaves bite marks; a bird that feeds in trees needs food trees and nest sites; a primate needs routes to move along. Ecological traces can be looked for even before a body is seen, and they are usually harder to fake consistently than a photograph.

Why do "Lazarus" species not vindicate every legend?

The coelacanth, the takahe and a number of other rediscovered species show that the scientific record is incomplete. But before a rediscovery there is usually a plausible habitat, a new specimen or clear signs; afterwards, verifiable individuals appear. Each case has its own probability. A small species surviving unnoticed on an island does not give a super-shark in every ocean the same chance.

Precedent should teach us how to search, not serve as a card that exempts anyone from evidence. If a hypothesis makes predictions and can be surveyed, then survey it. When the results come back empty, the belief has to be updated rather than claiming the species has learned to avoid every piece of technology.

Legend and biological classification call for two ways of reading

A local name may refer to an animal, an ancestor, a spirit, a place or a moral rule, depending on the sentence. Translating it as "monster" and then using it as a zoological report can be both wrong and hurtful. Sources published by the community itself, the original language and the rights over images should take priority.

Science can still test a photograph or a sample said to be an animal. The result does not "refute" a sacred meaning. Conversely, respecting a culture does not require treating every detail as anatomy. The two readings can coexist as long as an article states its question clearly and does not use one as a prop for the other.

The media turns many causes into a single character

A memorable name lets the press bundle footprints, cries, dead livestock and blurry photographs together. After the first story, people pay more attention and reports increase; pranksters join in as well. Later accounts are not independent, because everyone has seen the same illustration. Maps and timelines make it testable whether one animal could have moved that way at all.

When one piece of evidence is explained, the legend does not necessarily collapse, because the other pieces continue to hold each other up. This is the structure of a "flexible file": any counter-example is called a fake case, while the real case is still out there somewhere. To avoid it, every claim needs criteria for success and for failure set out before the test.

A checklist for assessing a new find

Ask where the original file is; whether the time and coordinates are verified; who holds the sample; whether there is a wide shot; the chain of custody; whether the expert works in the right field; whether the data has been peer reviewed; whether an independent team reproduced it; and by what method local species were excluded. "I don't know it" is not the same as "science has never known it".

Then ask what evidence would have to exist if the hypothesis were true: fresh teeth, fresh droppings, a nest, a carcass, young, predation marks or DNA. Focusing on predictions keeps an article open without being endlessly vague. A real discovery grows stronger under testing; a myth that survives only on weak data keeps retreating from the standard.

What are the pictures in this article for?

The illustrations reconstruct the landscape, the methods and the way a silhouette might appear; they are not photographs of an event. The creature is kept distant or indistinct so that no false "evidence" is created. The captions say what the picture explains; they do not confirm that a historical encounter took that exact form.

A beautiful image easily replaces the memory of the original data. Readers should therefore open the sources, look for specimens and consider the limits of the equipment. The sense of mystery comes from the distance between what is known and what is not; it does not need a sharply drawn scene standing in for an answer.

References

• Te Papa: an overview of the nine moa species and their extinction
• Te Papa: moa species and DNA data
• Department of Conservation: the story of the takahe rediscovery