Caterpillars of the Japanese oakblue butterfly produce sugary droplets that ants seem to enjoy. But these secretions are actually behavior-modifying drugs, and after they imbibe the ooze, the ants leave their nest duties behind to guard the caterpillar. The findings were published in Current Biology this week. 

We used to think that Narathura japonica caterpillars offer the secretions as a treat to Pristomyrmex punctatus ants who might help protect the butterfly larvae from predatory spiders and wasps. It was mutual… or was it? Researchers started noticing that the caterpillars were always attended by the same ant individuals. “It also seemed that the ants never moved away or returned to their nests,” Kobe University’s Masaru Hojo tells New Scientist. They seemed to abandon searching for food, and were just standing around guarding and defending the caterpillar.

So, Hojo and colleagues collected butterfly eggs, young caterpillars, and three ant colonies in Kyoto and Okinawa. The caterpillars were raised on Japanese blue oak (pictured above), while the ants were raised in nest boxes supplied with tasty items like mealworms and maple syrup. Then in a series of experiments, some of the two species were allowed to interact. To the right, you can see attendant ants standing on and around their caterpillar.

The team found that reward secretions from the caterpillar’s dorsal nectary organ are manipulative drugs that reduce the locomotory activities of its attendant ants, making them less active. Furthermore, ants who consumed these caterpillar secretions show aggressive responses when the caterpillar appeared to be alarmed (which for them means they display an eversion of their tentacle organs). 

When the team analyzed the brains of the drugged ants, they found a significant decrease in levels of the chemical messenger dopamine, compared with unrewarded control ants. This likely increased the ants’ cooperation with and fidelity to the caterpillar. 

Unless manipulated ants also receive a net nutritional benefit from the secretions, the authors write, these findings suggest that some reward-for-defense interactions that seem mutualistic may in fact be parasitic.

Center figure from M.K. Hojo et al., 2015 Current Biology