There's a scene in Pixar's Finding Nemo when Dory, a yellow-finned regal tang, injures herself in a tug-of-war over a snorkel mask. A tiny plume of blood curls away from Dory's face into the water around her, where it is sucked into the nostrils of Bruce, a "vegetarian" shark who immediately recants his no-sushi policy. (Fortunately, Dory escapes.) Scientists have known for some time that many ocean predators relish the scent of an injured fish, whereas fish that are more likely to end up as a meal flee from the same scent. Now, researchers think they have pinpointed the key chemical in fish skin that warns nearby fish of danger—a chemical related to a supplement some people take for joint pain.

In the 1930s Austrian animal behavior scientist Karl von Frisch accidentally injured a minnow in a tank. He noticed that the other fish in the tank began alternately darting back and forth and freezing in place—classic predator-evading behavior. Subsequent experiments established that the frightened fish were responding to chemicals released from the skin of their injured peer—a cocktail dubbed schreckstoff, which is German for "scary stuff."

For decades, the chemistry of schreckstoff remained unknown. In the 1970s and '80s some scientists discovered that exposing fish to a chemical known as hypoxanthine-3-N-oxide (H3NO) frightened them in the same way as schreckstoff, albeit to a lesser degree. H3NO, they concluded, was probably the active compound in schreckstoff. But there was a problem with that idea: scientists had never reliably detected H3NO in fish skin. Instead, some researchers proposed, H3NO may mimic the genuine active compound.

Now, Suresh Jesuthasan and Ajay Mathuru of Neuroscience Research Partnership in Singapore, as well as Rainer Friedrich of the Friedrich Miescher Institute for Biomedical Research in Switzerland, and their colleagues think they have isolated the key ingredient in schreckstoff—a sugarlike molecule named glycosaminoglycan (GAG) chondroitin. Their findings appear online in the February 23 issue of Current Biology.

Friedrich and his colleagues studied schreckstoff extracted from ground-up zebra fish, which are closely related to minnows and catfishes. The researchers performed a series of tests on the schreckstoff to identify the active compound. For example, destroying all the proteins in the solution did not diminish schreckstoff's ability to frighten fish, suggesting that the active compound was not a protein. In another test, the researchers dipped a column of wheat germ agglutinin—a sugar-binding protein that protects wheat from insects and pathogens—into schreckstoff and introduced fish to the compounds the wheat germ had soaked up. The fish behaved as though they were escaping a predator: they stayed near the bottom of the tank and rushed from one spot to another or they swam very slowly for a short time and then shot off in another direction over and over again. Perhaps the active compound in schreckstoff was similar to the sugars sponged up by the wheat germ column.

As expected, Friedrich and his colleagues detected glycosaminoglycans in zebra fish schreckstoff, in particular glycosaminoglycan chondroitin. Glycosaminoglycans are long, sugar-like molecules that exist in various forms in living creatures, most notably in cartilage and other connective tissue. When the researchers broke down the glycosaminoglycan chondroitin in schreckstoff with enzymes or rendered chondroitin inert with antibodies, they robbed schreckstoff of its power to frighten. Further, exposing fish to chondroitin alone triggered darting, slow swimming and bottom-dwelling—but not as effectively as undiluted schreckstoff.

"It's really nice to have biochemists seriously looking at the chemistry of chemical alarm cues in fish," says Brian Wisenden of Minnesota State University Moorhead, who has studied schreckstoff in minnows but was not involved in the new study. "We've basically been working with a mystery substance."

Schreckstoff also appears to activate a specific region of the olfactory bulb, the part of the brain that processes odors. Anatomically unusual neurons called crypt cells are woven into this part of the olfactory bulb—neurons whose function remains a mystery. Friedrich and his colleagues suggest that these cells are specifically dedicated to sensing schreckstoff. When they exposed fish to other odors, such as bile acids and amino acids, the neurons in this part of the brain did not respond—they were only interested in schreckstoff.

Ever since the discovery of the scary stuff, scientists have wondered why fish would evolve to release a chemical alarm when injured. Sure, it is a nice gesture toward other fish, but how does the injured animal benefit? Warning siblings and cousins of danger helps protect one's own genes—a strategy known as kin selection—but schools of fish are not necessarily populated by members of the same family.

Another explanation is that there is no real benefit to the injured fish. Instead, evolution has favored fish that recognize the chemicals naturally released by injured comrades—just as some plants have evolved to recognize chemicals released by neighbors under attack. In both these situations, injured organisms are not intentionally releasing chemicals. In fact, they would rather keep those useful substances to themselves. The volatile compounds that escape from injured plants are usually toxins stored to defend against herbivores. And some evidence suggests that the GAGs in schreckstoff play an important role in the skin's immune system. As a natural consequence of physics, these molecules are always released into the water when a fish is injured. In this way, certain compounds are consistently associated with injury—and any fish that recognizes that link has a better chance of escaping predators and surviving. Over time, evolution establishes these compounds as reliable signals of danger. Today, all zebra fish—and indeed many fish species—are born with brains wired to learn that a whiff of schreckstoff is their cue to hightail it.