Madness of Crowds: Single Ants Beat Colonies At Easy Choices


Virtually every article or documentary about ants takes a moment to fawn over their incredible collective achievements. Together, ant colonies canraise gardens and livestockbuild living raftsrun vaccination programmesoverpower huge preydeter elephants, and invade continents. No individual could do any of this; it takes a colony to pull off such feats.

But ants can also screw up. Like all animal collectives, they face situations when the crowd’s wisdom turns into foolishness.

Takao Sasaki and Stephen Pratt from Arizona State University found one such example among house-hunting Temnothorax ants. When they need to find a new nest, workers spread out from their colony to search for good real estate. In earlier work, Sasaki and Pratt have shown that, as a group, the ants are better at picking the best of two closely matched locations, even if most of the workers have only seen one of the options. It’s a classic example of swarm intelligence, where a colony collectively computes the best solution to a task.

But Sasaki showed that this only happens if their choice is difficult. If one nest site is clearly better than the other, individual ants actually outperform colonies.

Written By: Ed Yong
continue to source article at


  1. The rogue ant wandering around, not following any trail is of interest to me. I frequently wonder if the are somehow biologically different from the swarm – maybe a little brighter or have better sensors (?)

  2. The title suggests there’s a huge discrepancy between colony and individual performance, but if so, the implied finding is exaggerated. In actuality, the difference between the two only seems to be around 10%, and both found the better site more often than not in any case. It said so later in the article:

    As the light difference between the nests got bigger and the task became easier, the ants, whether as individuals or colonies, made more accurate choices. The team expected as much. But to their surprise, the single workers showed the greatest improvements and eventually outperformed their collective peers. In the easiest tasks, they chose the darker nest 90 percent of the time, while the colonies peaked at 80 percent accuracy.

    And I think comparing eusocial insects with human societies, as some commentators are doing, is a bit unwarranted. Insect societies are heavily based around immense kinship ties, whereas human committees involve unrelated individuals negotiating the stakes for a collective benefit. I think there’s far more scope for groupthink-type errors in such committees than in a giant family, if only because the need for cooperation despite differences puts more pressure to reach a consensus on the former than on the latter, where cooperation comes a little more readily due to the common genetic stake.

Leave a Reply