For a decade, codebreakers have tried—and failed—to solve the case of two crumpled papers that were found stuffed in a secret pocket under the bustle and inside the seams of a silk dress from the 1880s. Over the years, the mystery even earned its own moniker: The Silk Dress Cryptogram.

Now, the code itself has an answer—and it came from unlikely sources: the National Oceanic and Atmospheric Administration (NOAA) and a data analyst at the University of Manitoba.

The modern mystery starts with the dress itself: an 1880s, silk bustle dress with a fully concealed pocket. By the time Sara Rivers Cofield bought it at an antique shop in Maine in 2013, the coded sheets had been forgotten inside. What is still not known is who placed them there, or when.

That’s when the fun starts. Rivers-Cofield posted a blog at the time about discovering not only the completely concealed secret pocket, but also what appeared to be a code. “I’m putting it up here in case there’s some decoding prodigy out there looking for a project,” she wrote on her blog.

For a decade, the code remained, well, encoded. Questions lingered over what “Bismark omit leafage buck bank,” “Calgary Cuba unguard confute duck fagan,” and “Spring wilderness lining one reading novice” could have possibly meant.

There were also numbers between the lines, each marked off with a different color, along with time-like notes in the margin. Wayne Chan, a data analyst from the University of Manitoba and hobby codebreaker, later transcribed those notes more cautiously: one sheet appears to read “101 PM,” while another carries “1115 PM” and “1124 P.” Those marginal times remain harder to pin down than the weather code itself.

The theories proliferated: illicit gambling codes, dress measurements, spy talk. The better lead turned out to be less cloak-and-dagger and more practical: telegraphic shorthand, the compression system that kept word counts down when telegraph companies charged by the word.

Chan, working in the university’s Centre for Earth Observation Science, scoured 170 telegraphic codebooks. Nothing.

The Silk Dress Cryptogram rose to become one of the top 50 unsolved codes in the world, but Chan wasn’t done. He eventually stumbled upon the old book “Telegraphic Tales and Telegraphic History,” and read about weather codes used by the U.S. Army Signal Corps—the group that served as the national weather service during the late 1800s. He drew a parallel between the code from the dress and the examples in the book of code employed to send weather observations.

Chan had a thread to pull. He yanked. Soon, he was working with the NOAA’s Central Library in Maryland and finding un-searched-through weather telegraph code books, including one published in 1892 that proved he was on the right track. Combining resources, Chan decoded the text and published his findings in the journal Cryptologia.

As NOAA explained in a post, each line on the crumpled paper indicates weather observations at a given location, coupled with the time of day. That information was telegraphed into a central Signal Service office in Washington, D.C.

The format begins with the unencoded station location, followed by codewords for temperature/pressure, dew point, precipitation/wind direction, cloud observations, and wind velocity/sunset observations. So, “Bismark, omit, leafage, buck, bank” really means the weather in Bismarck, then in the Dakota Territory, had an air temperature of 56 degrees Fahrenheit with a barometric pressure of 30.08 inHG (omit), a dew point of 32 degrees (leafage) on a clear day with no precipitation and wind from the north (buck), all with a clear sunset and a wind velocity of 12 miles per hour (bank).

There you have it. The famous nonsense phrase was a weather report.

The code was tricky because only government officials creating weather maps used it. Adopted in 1887, the code allowed for six words to provide an entire report for a location. Adopted in 1887, the system allowed a station to compress an entire weather report into a handful of words. The caveat: Chan’s paper treats the Canadian station messages with more caution, saying they did not exactly follow the U.S. format and had to be reverse-engineered from the available evidence. They may still need archival confirmation.

Using the decoded observations and available weather records, Chan argued that the messages most likely point to May 27, 1888.

Even with the code broken, there’s still plenty of questions about the dress, and its owner. Rivers-Cofield found the word “Bennett” written on a label inside the dress, but Chan couldn’t connect that name to a woman on the clerical staff of the U.S. Army Signal Service office in Washington, D.C. Later reporting narrowed the possibilities—Chan identified possible women connected to the predecessor weather organization, and others raised a possible telegraph-worker-household theory—but none of those leads has confirmed the owner or proved that the Bennett label belonged to her.

Rivers-Cofield has pointed out that the dress has roughly 130 years of unknown history. The label could have been added later. The garment could have moved through laundry, donation, resale, or a household where someone else handled telegraphed weather reports. While the cipher has been translated, the object’s path hasn’t.

Solving one mystery leads to another. But at least this one doesn’t have a mind-bending code to crack.