Myrmecophytic Piper spp. were re-examined in view of what has become a common expectation in studies of mutualistic ant-plant associations: overtly aggressive behavioral traits in ants that function in anti-herbivore defense. Piper plants provide food and shelter for the Pheidole ant inhabitants. The ants, small and “sluggish”, were previously thought to serve in nutrient procurement for the plant even though their foraging is probably restricted to the plant surface and most of the ants' food is produced by the plant itself.
An alternative hypothesis, that Pheidole ants function in anti-herbivore defense by disrupting herbivores as eggs or early instars occurring on Piper foliage, was tested using the following lines of reasoning: (1) If the youngest, tender leaves are more vulnerable to herbivore attack, or simply more valuable to the plant, then ants may “patrol” these leaves preferentially. More ants were found on young leaves (\(\bar X\)=2.00 ants/leaf) than on mature leaves (\(\bar X\)=0.51). (2) If the ants are functioning in anti-herbivore defense, herbivory should be lower on occupied plants than on plants without ants. Estimates of mean percent foliovore damage per sapling showed that for samples with similar average heights (and presumably, ages), the unoccupied plants had significantly greater damage. A comparison of newest leaves showed a significant trend for damage to decrease with increased ant activity. (3) If ants are attacking hervibores at early stages of development, baiting with eggs should demonstrate this activity. Over 75% of all egg baits on young and mature leaves were taken up by an ant during observation periods of 30 or 60 min. Baits were found significantly faster on young than on mature leaves. In over half of the baiting trials, the egg was taken to the edge of the leaf and dropped to the ground rather than sequestered as a nutrient source.
These data suggest that Pheidole inhabiting Piper plants are important in plant defense from herbivores and that adherance to a classical notion of aggression in similar studies might bias the initial question-asking stages of an investigation.