Evolutionary stability of mutualism between yuccas and yucca moths

Abstract

INTERSPECIFIC mutualisms inherently possess a conflict of interests between the interacting species in that fitness increases of one species occur at the expense of the other. This holds for mutualisms as diverse as plant associations with mycorrhizal fungi or nitrogen-fixing bacteria, animals and endosymbionts, and obligate plant-pollinator associations1–6. Prevailing models of interspecific cooperation predict that mutualistic interactions are evolutionary stable only when both interacting species possess mechanisms to prevent excessive exploitation3–6. In light of this, it is paradoxical that some of the classical examples of coevolved obligate mutualism seemingly do not meet this criterion. In mutualisms involving seed parasites that actively pollinate their hosts, such as yucca/ yucca moth and fig/fig wasp interactions, there is no apparent means of retaliation on behalf of the plant. Predictions from theory suggest that a cryptic mechanism, such as selective abortion of flowers with heavy egg loads, could stabilize these interactions4,6–9. Here we present the first empirical evidence that such a mechanism in fact exists in the yucca/yucca moth interaction. A strong nega-tive effect exists between moth egg number and probability of flower retention. Furthermore, we show a strong positive effect between the number of pollinations received and the probability of flower retention. Selective maturation of fruit with low egg loads and high pollen loads provides a mechanism to increase the quantity and possibly quality of seeds produced, and simultaneously select against moths that lay many eggs per flower or provide low-quality pollinations4,6,8,10. Not only can these results explain the stability of this type of interaction, but selection for high-quality pollination also provides a mechanism to help explain the evolution of active pollination among yucca moths.