Credit: M. MOFFETT/NATIONAL GEOGRAPHIC

There are many examples of happy partnerships between plants and ants, in which, for instance, the plant gains protection from herbivores and in return the ants are housed. These symbiotic relationships can, however, be so one-sided that they verge on parasitism, and a striking example is described by Douglas Yu and Naomi Pierce in Proceedings of the Royal Society of London B (265, 375-382; 1998).

The plant involved is a member of the Boraginaceae family, Cordia nodosa, which grows to some 2 m in height and is a common constituent of lowland forest in southeast Amazonian Peru. The ant belongs to the genus Allomerus, and inhabits so-called domatia, specialized swellings in the stems of the host. Yu and Pierce found that almost 80% of the plants they sampled contained colonies of Allomerus.

The ant, however, manipulates the plant to its own ends by destroying the flower buds and flowers while energetically protecting the leaves from attack by other insects. The ants don't eat the buds or flowers — they just wilfully chew off the reproductive structures in what Yu and Pierce refer to as acts of ‘castration’. The result is ravaged buds, such as those shown here, and a dramatic reduction in fruit production. Conversely, senescence of the plant's branches is delayed, thereby providing more domatia for the ants. There appears to be no gain for the plant in this, just the pain of severely impaired reproductive capacity. Nor, apparently, does the plant retaliate in any way.

The question then is why, given that the host's competitiveness must be adversely affected, does this ant-plant relationship seem to persist and be widespread? Yu and Pierce think that one explanation may lie in the truly mutualistic relationship that C. nodosa has with three species of another genus of ant, Azteca. These species — the good guys as far as the plant is concerned — inhabited only about one-tenth of the C. nodosa sampled. But these plants reproduce normally and may provide enough new recruits to keep the overall population stable.

Further twists lie in the observations that a few of the C. nodosa examined carried colonies of both genera of ant on separate trunks; the respective effects on plant reproduction were as expected from the main findings. More curiously, in fewer than 1% of cases the Allomerus ants left the buds alone, and the host fruited normally. That tantalizing finding defies explanation.