Between Seven and Ten Vaquitas Were Seen in 2025. That Is Not the Same as How Many Exist.

On 28 October 2025, Mexico's environment ministry (SEMARNAT), navy (SEMAR), protected areas commission (CONANP) and Sea Shepherd Conservation Society jointly reported the results of the 2025 vaquita monitoring effort in the Upper Gulf of California.

They confirmed sightings of between 7 and 10 vaquitas, and the birth of new calves.

The sentence everyone gets wrong

"7 to 10 vaquitas were observed" does not mean "7 to 10 vaquitas exist."

It means: over a defined survey window, in a defined area, with defined effort, observers identified somewhere between seven and ten distinct individuals.

An animal outside the search area was not counted. An animal that surfaced when nobody was looking was not counted. The vaquita is small, elusive, and — as Undersecretary Marina Robles García noted — some individuals turned up that had not been seen for years, evidently spending time in areas that were not being monitored.

A full 2025 abundance estimate is expected in 2026. That is the number that will attempt to say how many exist. It does not exist yet.

Almost every headline collapsed these two things. We are not going to, because for a species this small the difference between an observed count and an abundance estimate is the difference between a fact and a guess dressed as one.

What "67% probability of 7 to 10" actually means

The figure comes from Expert Elicitation — a formal method in which facilitators work with the cruise's observers to reach a structured consensus on the frequency distribution of animals seen.

This is a legitimate, standard technique for exactly this situation: too few animals for conventional statistics, high-quality observations, deep expertise. It is also, unavoidably, structured expert judgment. It is not a headcount. Naming the method is not a criticism of it — it is what the method is.

Here are the three years, stated exactly as reported:

Read that table before accepting "the population increased."

Several outlets reported 2025 as a slight increase over 2024. Against 2024, the range did shift up. But:

• The probability levels differ. 2024's range was given at 75% confidence; 2025's at 67%. A wider range at lower confidence is not straightforwardly "more."
• The ranges overlap. 7–10 and 6–8 share 7 and 8.
• 2023 was higher than 2025. 8–12 exceeds 7–10 at the same 67% confidence.

So "up from 2024" is true and "down from 2023" is also true, and neither supports a trend claim. What these three years jointly support is the far more careful statement the officials actually made: no substantial decline for three consecutive years, and the species is still reproducing.

That is what the data says. It is a smaller claim than recovery, and it is a large claim compared to what came before.

The number that gives it weight

The population once dropped 45% in a single year.

Against that, three years without a crash is not a plateau — it is a change in the shape of the curve. And the first survey, in 1997, estimated roughly 567 individuals. The vaquita has fallen ~98% in under three decades.

Stopping a 45%-per-year decline is the only thing that could possibly precede a recovery. It is not a recovery.

Frida

At least one or two calves were observed. The mother of one is known as Frida — she has a bent dorsal fin, and is named for Frida Kahlo.

Frida was also seen with a calf in 2023 and in 2024.

Three calves in three years, from one identifiable female, in a population of roughly this size. A species does not persist on sentiment, but it is worth being precise about why this detail is data and not decoration: a female calving on a regular interval in a population of ~10 is a measurable fraction of the entire species' reproductive output.

The effort behind the number

The scale of work required to see fewer than a dozen animals is itself the story:

• Acoustic monitoring: 254 acoustic encounters between 24 May and 29 September, from 1,228 detectors deployed across 497 sites.
• Visual monitoring: 2–30 September, from Sea Shepherd's vessels Seahorse and Bob Barker, each carrying three pairs of 25× "Big Eyes" binoculars, crewed by scientists whose combined careers total more than 400 years of cetacean research.
• Local participation: local fishers collected acoustic data to establish where vaquitas spend time. 17 young people from San Felipe were trained in visual monitoring methods and took part.

That last point matters more than it looks. The vaquita's decline is driven by gillnets — set illegally for totoaba, whose swim bladders fetch extraordinary prices — and the fishing communities of the Upper Gulf are the people whose livelihoods any solution runs through. A survey that trains seventeen local teenagers to spot the animal is doing something a survey conducted entirely by outsiders cannot.

What we are not saying

We are not saying the vaquita is saved. Nobody involved said that.

We are not saying the population grew — the elicited ranges do not support it.

We are saying the decline stopped for three years running, the animals are breeding, and the abundance estimate that could turn observation into a population figure is due this year.

Why this is on AEDO1

The vaquita is the most endangered marine mammal on Earth, and it is being counted by consensus among people looking through binoculars, because there are too few left for statistics to work normally.

That is what the end of the line looks like as a measurement problem. Every number in this article has a boundary around it, and we have tried to draw each one where the evidence actually stops.