Table 1.
Summary of behavioral trials carried out to address each experimental question.
Figure 1.
Tactile response results showing the mean proportion of Camponotus novograndensis workers repelled by the fresh dehiscing floral parts and control parts of each plant species.
Results arranged by increasing repellence from fresh flowers. In the case of Ruellia inudata and Blechum pyramidatum areas of inflorescence carrying trichomes were used. Floral parts of species in bold repelled ants on more than 50% of encounters and those marked * repelled ants significantly more from fresh flowers than control parts used (pollen-depleted flowers or buds). At least 12 replicates were carried out for each plant species. Error bars indicate 95% confidence intervals.
Figure 2.
Tactile ant-repellence, nectar volume and nectar accessibility.
a) Significant differences in ant-repellence occurred between plants with different volumes of nectar (χ2 = 42.9, df = 4, p<0.001) with a significant positive trend between nectar volume and the proportion of ants repelled (J-T = 5.9, df = 4, p<0.001). b) Ant-repellence differed significantly between plants with different levels of accessibility to nectar (χ2 = 26.6, df = 3, p<0.001), although the negative trend between accessibility and repellence was not significant (J-T = −1.5, df = 3, p = 0.13). Significant differences between groups indicated with a/b/c. Error bars indicate 95% confidence intervals.
Figure 3.
Plant species tested for tactile ant repellence with Ectatomma ruidum.
* indicates significant difference between freshly dehisced flowers and old pollen-depleted flowers. Error bars indicate 95% confidence intervals.
Figure 4.
Floral ant-repellence in flowers tested against a range of ant species (including Cephalotes setulifer tested only with its host C. alliodora).
(a) Barleria oenotheroides, (b) Malvaviscus arboreus, (c) Ruellia inudata and (d) Cordia alliodora. * indicates significant difference between freshly dehisced flowers and old pollen-depleted flowers. Error bars indicate 95% confidence intervals.
Figure 5.
Floral ant-repellence in two species of Stachytarpheta tested against different ant species.
(a) Stachytarpheta jamaicensis and (b) S. frantzii. * indicates significant difference between freshly dehisced flowers and old pollen-depleted flowers. Error bars indicate 95% confidence intervals.