A 1949 whale song proves sound archives deserve fossil-grade care

What is the most valuable thing a scientist can record: the thing they understand, or the thing they don't? The 1949 humpback recording suggests the second, and it should change how we treat old audio.

Here is what happened. As WHOI announced on February 10, 2026, researchers and archivists at the Woods Hole Oceanographic Institution identified what may be the earliest preserved recording of a whale: a humpback captured on March 7, 1949, near Bermuda. It sat on an audograph disc, an office dictation format that etched sound onto thin plastic, while researchers aboard the R/V Atlantis tested sonar and other acoustic systems for the U.S. Office of Naval Research.

Note the hedge: "may be." This is an institutional announcement, not a peer-reviewed paper, and the comparisons it points to have yet to be published, as far as the material we reviewed shows. Treat it as a promising find, not a result.

The silence is the real discovery

The song is lovely, but the stronger prize is the water around it. Marine bioacoustician Peter Tyack told Folha that the recovered recordings tell us not only about the whales but also about what the ocean's soundscape was like in the late 1940s, something he said is very hard to reconstruct any other way. His colleague Laela Sayigh put the problem bluntly: data from this period "simply don't exist in most cases."

That is why we think "baseline" is the right word. You can't measure how much louder the ocean has become if you have no quiet reference. Modern hydrophones, gliders and buoys generate huge datasets, but they all record the noisy present. A single recording from before the expansion of shipping and sonar is worth more than a thousand new ones, because it can't be replaced. Nobody can go back and record 1949.

There is a caveat. One recording near one island in one March is an anecdote, not a trend. Its value depends on context, and WHOI is working with Ocean Alliance, which holds more than 2,400 recordings of whale and ocean sounds from the 1950s through the 1990s. That series is what could turn a curiosity into a time line. We'd bet the series matters more than the headline disc, though that is our judgment.

Nobody knew what they were hearing, and they kept it anyway

The most instructive part is human. Ashley Jester, WHOI's director of research data and library services, explained that the scientists didn't know what they were hearing but chose to record and preserve the sounds anyway. They even made recordings with their own boats deliberately silent, to hear as much as possible. Nobody in 1949 could have asked how humpback songs change over decades, yet they saved the evidence that lets us ask it.

Compare this with how we normally justify data retention: a clear hypothesis, a funded project, a deliverable. That logic would have thrown this disc away. Most recordings from the era were lost as tapes deteriorated; these survived, in WHOI's words, because of their material and careful preservation. Survival was partly luck, and luck is a poor conservation strategy.

Treat sound like fossils

We make the argument plainly: a field recording is a specimen. Paleontologists don't know which bone will answer a future question, so they keep the bones. Audio deserves that same reflex. It degrades, it depends on obsolete machines, and it's often undocumented. The disc was labeled only "whoop whoop," according to Conexão Planeta.

The price of rescue is also telling. WHOI's archives recently received a $10,000 award from the National Recording Preservation Foundation to digitize the audograph collection. In our view, that is a modest sum for the chance of a reference point like this. If that is what a baseline costs, the question is how many other discs sit unlabeled in other institutions, on formats nobody can play.

Our verdict: the 1949 humpback is a promising reference point, not yet a proven one. The real lesson is policy. Funders and institutions should pay for digitizing and cataloging old acoustic collections now, before the playback equipment and the people who know it disappear. The cheapest time to save a baseline is before anyone knows they need it.