Such congestion is excellent for producing stop signals.
The key was that although dancers often restarted dancing after receiving a stop signal, they performed fewer dance circuits when they received stop signals.
Sometimes the effect on dancing was slight and difficult to detect, but after recording hundreds of dances and stop signals, it became unmistakable.
I noticed that our bees produced many stop signals after being attacked, and we began studying them instead.
Stop signals inhibit waggle dancing when continued recruitment becomes unprofitable or dangerous.
My path in science began, in part, because Donald Griffin broke his foot. Famous for his pioneering work on animal sounds, he had been trying to insert a microphone into a beaver lodge to eavesdrop on the homeowners. That summer, the beavers kept their privacy, and I was left largely on my own in Concord, Massachusetts, with a honey bee colony we had planned to use to study the sounds produced by waggle-dancing honey bees.
After returning from a good food source, a honey bee forager can perform a waggle dance that communicates the source’s distance and direction to nestmates, recruiting them to leave the nest and seek the same food. Studying these dance sounds was the goal of our project. But as I scanned a microphone over the waggle dance, I was soon distracted by short piping sounds directed at the dancer. Dance followers track the dancer to learn the information that she is trying to communicate, and, oddly, they sometimes butt their heads against the dancer and produce a fairly loud “beep” that causes the dancer to briefly freeze. Older scientific literature described this sound as a food-begging signal, but the interpretation did not fit what I saw. The signaling bee rarely received food, and so I decided to call these “stop signals.”
Because I was working alone, I could not videotape the bees inside the hive while also controlling the number of bees recruited to my syrup feeder. My feeder rapidly became overcrowded and ran out of food. I had not anticipated this problem, but I was fortunate. Such congestion is excellent for producing stop signals. When crowding was greatest, I recorded the most signals and found that they tended to inhibit waggle dancing. The key was that although dancers often restarted dancing after receiving a stop signal, they performed fewer dance circuits when they received stop signals. More dance circuits recruit more nestmates and thus stop signals inhibit colony recruitment. Sometimes the effect on dancing was slight and difficult to detect, but after recording hundreds of dances and stop signals, it became unmistakable. The harder question was this: why would followers try to stop dance communication? After all, followers are trying to learn where the dancer is telling them to go.
This question was in the back of my mind for over 20 years. I did not figure out the answer that summer, but the “failed” experiment gave me a better question, not the one I had intended to resolve. That summer marked the beginning of my career and gave me a passion for decoding animal communication.
Scientists rarely speak publicly about this side of research. Perhaps “failure” is even too strong a term. Research takes different and unexpected paths, but when something goes fundamentally wrong, it can feel like a punch in the gut. Unlike the way we are sometimes portrayed (and we are a bit complicit in this), scientists are not white-coated acolytes in a temple of reason. We have blood, sweat, and tears invested in our craft.
The questions worth asking are often difficult, and success can therefore be hard-won. Much of science relies on a determination to see an experiment through, come hell or lightning-spiked water.
Sometimes the answer is persistence. In Panama, I tried to train foragers of the stingless bee Melipona panamica across a stretch of water to see if they used odor trails to guide recruits. Such trails cannot be laid on water. We placed the feeder in a canoe and slowly rowed to the opposite shore. On the day we finally trained them across the water gap, a Jesus lizard perched on the hive entrance and made an expensive snack of the foragers we had been painstakingly training for weeks. I installed a slippery barrier. A few days later, a large toad breached it and enjoyed another bee buffet. Later attempts brought torrential rain, frequent lightning, an angry wasp nest, and bullet ants (aptly named for their stings) marching along the rope we used to guide our canoe. On roughly the eighth attempt, the experiment worked. These bees did not use odor trails!
Persistence, however, is not always the right response.
Years later in Brazil, we spent weeks preparing experiments on communication in the stingless bee Melipona rufiventris. Highly aggressive Trigona spinipes foragers repeatedly discovered our feeders, recruited large numbers of nestmates, attacked the Melipona, and made the planned work impossible for an entire field season, a disaster since we only had a limited time window. Were the invaders attracted to the feeder color, scent, or location? We changed all three, but still the Trigona arrived.
After three weeks of repeated failures, it became clear that we were not simply unlucky, or perhaps we had not recognized how lucky we were. The invaders appeared only after Melipona began feeding. Once we stopped treating them solely as a nuisance, a new question emerged: how were the T. spinipes finding the food? We abandoned our Melipona experiment. Our new experiments showed that scouting T. spinipes were attracted to scent marks left by M. rufiventris. They were eavesdropping on their competitors and using the intercepted information to take over the resource. The obstacle had become the phenomenon, revealing a fascinating strategy in stingless bee warfare.
Many years later, I was again training honey bees to a feeder when bees from other colonies attacked our foragers and made our planned experiment futile. I noticed that our bees produced many stop signals after being attacked, and we began studying them instead. Finally, the question from Concord began to make sense. Stop signals inhibit waggle dancing when continued recruitment becomes unprofitable or dangerous. Later experiments showed the same response when predators attacked bees at rich food sources. Ironically, some of the experiments I did not plan turned out to be the better ones, because they followed biological phenomena strong enough to swamp our careful plans.
These experiences changed how I think about scientific resilience. It is not simply persistence. Sometimes resilience means trying again because the question remains worth asking. At other times, it means recognizing that an apparent failure is pointing toward a different question.
But there is a danger in celebrating scientific resilience. It should not become a polite way of asking scientists to abandon their questions or simply make do with less. Much essential science requires substantial and sustained investment, particularly in the people we are training to become the next generation of researchers. Resilience cannot substitute for that support. When resources are constrained, however, a good question can still guide us toward the science that remains possible. The work may be smaller, less expensive, or simply different from what we had planned, but that does not necessarily make it less important. Sometimes a different path leads us to questions we did not know to ask.
I am heartened to see scientists around the world making the case for why research matters. Part of that case, I think, lies in showing how science is actually made: we try things that may not work, persist when the question remains compelling, and change direction when the evidence leads somewhere unexpected. Research does not always go according to plan, but that unpredictability is part of what makes discovery possible.
At Concord, I could not rescue the experiment I had planned. Eventually, I stopped trying and listened to the bees. The feeder ran dry that summer, but the questions it raised continue to inspire me. More than 35 years later, I’m still listening.
James C. Nieh is a Professor in the Department of Ecology, Behavior, and Evolution and an Associate Dean in the School of Biological Sciences at the University of California, San Diego. He is a Fellow of the Royal Entomological Society and a recipient of the Hambleton Award for his bee research and the Springer Nature Editor of Distinction Award.
