Non‐pollinating cheater wasps benefit from seasonally poor performance of the mutualistic pollinating wasps at the northern limit of the range of Ficus microcarpa

1. Species interactions in tightly bound ecological mutualisms often feature highly specialised species' roles in which competitive exclusion may preclude multi‐species coexistence. Among the 800 fig (Ficus) species, it was originally considered that each was pollinated by their own wasp (Agaonidae). However, recent investigations show that this ‘one‐to‐one’ rule often breaks down, as fig species regularly host multiple agaonids but in ways suggesting that competitive processes still mediate biodiversity outcomes. 2. A phenological survey was conducted of the fig–fig wasp pair, Ficus microcarpa and its associated pollinating wasp, alongside its sister species, the cheating wasp, in Xishuangbanna, China. 3. Reproductive output underwent extreme seasonal variation. Seed and pollinator production fell markedly during cooler, drier months, although high levels of fig production continued. However, this resource was predominantly utilised by the cheater species, which offers no pollination services. Pollinators and cheaters rarely co‐occur, suggesting that temporal coexistence is constrained by competition for access to figs. 4. The overall findings indicate periodic rearrangements of mutualism dynamics, probably resulting from a strongly seasonal environment. Sympatric co‐occurrence may result from a window of opportunity for a functionally divergent agaonid, potentially due to constraints on the main pollinator in adapting to variable year‐round conditions that prevent competitive exclusion.     

Control in mutualisms: Combined implications of partner choice and bargaining roles

When two species form a mutualistic association, the degree of control that each has over the interaction may be pivotal in determining the relative benefit each obtains. We incorporate the capacity for partner choice into a model of mutualism based on the exchange of goods and/or services, where one guild of mutualists plays the role of proposer (proposing a price at which the goods and/or services will be exchanged) and the other plays the role of responder (accepting or rejecting the deal). We show how the payoff structure in this scenario and other closely related ones correspond to the ultimatum and demand games of economics. In the model, there are both costs and benefits to a guild whose players have control over interactions. Control over interactions in the sense of being able to exercise partner choice can benefit a guild by selecting for mutualism in its partners, but is most effective in selecting against moderately exploitative partners, and so can give highly exploitative partners an advantage. This can generate dynamics similar to taxon cycles or those seen in models with competition-colonization tradeoffs, wherein increasingly more mutualistic partners (acting as superior competitors) are selected for up to a tipping point, at which highly exploitative strategies (akin to superior colonizers) gain an advantage. Control over interactions in the sense of being able to appropriate ‘surplus’ payoffs in each interaction, which is selected for within-guild and is equivalent to playing the role of responders, selects against high demands (and so for mutualism) in the guild with control. Combining the two mechanisms, a high degree of mutualism in both guilds and coexistence of more mutualistic and more exploitative strategies within each are both consistent with control over the interaction being highly skewed toward one side that does what is in its own short-term interests.     

Testing the out-of-Florida hypothesis on the origin of cheating in the yucca-yucca moth mutualism

Mutualistic interactions can be exploited by cheaters that take the rewards offered by mutualists without providing services in return. The evolution of cheater species from mutualist ancestors is thought to be possible under particular ecological conditions. Here we provide a test of the first explicit model of the transition from mutualism to antagonism. We used the obligate pollination mutualism between yuccas and yucca moths to examine the origins of a nonpollinating cheater moth, Tegeticula intermedia, and its pollinating sister species, T. cassandra. Based on geographic distribution and ecological factors affecting the pollinators, previous research had indicated that the cheaters evolved in Florida as a result of sympatry of T. cassandra and another pollinator species. We used mitochondrial DNA (mtDNA) sequences and amplified fragment length polymorphism (AFLP) data to investigate the phylogeographic history of the pollinator-cheater sister pair and to test whether the cheaters arose in Florida. Contrary to predictions, phylogenetic and population genetic analyses suggested that the cheaters evolved in the western United States and subsequently spread eastward. Western populations of cheaters had the most ancestral haplotypes and the highest genetic diversity, and there was also significant genetic structure associated with a geographic split between eastern and western populations. In comparison, there was evidence for weak genetic structure between northern and southern pollinator populations, suggesting a long history in Florida. The western origin of the cheaters indicated that the pollinators have more recently become restricted to the southeastern United States. This was supported by AFLP analyses that indicated that the pollinators were more closely related to the western cheaters than they were to geographically proximate cheaters in the east. Shared mtDNA between pollinators and eastern cheaters suggested hybridization, possibly in a secondary contact zone. The results negate the out-of-Florida hypothesis and reveal instead a long, complex, and disparate history for the pollinator-cheater sister pair.     

Interactive effects of builders and exploiters on environmental quality and the outcome of competition between the two

The implications of spatial and temporal structure for the maintenance of mutualism, altruism, and niche construction or ecosystem engineering have been explored by many theoretical models. Part of what these models have shown is that organisms that give up some amount of potential short-term gain in order to improve the quality of their environment can, in a variety of scenarios, persist in the face of more exploitative competitors if structure in environmental quality allows the former to preferentially benefit from their investments. The models presented here consider the additional implications of interactions between competitors in their effects on their environment (recently documented in multiple systems). Relative to when competitor types were additive, synergistic effects promoted coexistence and antagonistic effects promoted founder effects (but favored the less exploitative type when both had equal initial frequencies). Spatial and temporal patterns of patch quality and occupancy also differed markedly between scenarios, even where all three scenarios generated the same qualitative outcome. These models show that understanding both the scale over which organisms affect their environment and the degree to which organisms interact in such effects are important for interpreting patterns in environmental quality, predicting the effects of organism-environment feedback on competition, and explaining the persistence of mutualistic traits.     

STABILIZING MECHANISMS IN A LEGUME–RHIZOBIUM MUTUALISM

Preferential rewarding of more beneficial partners may stabilize mutualisms against the invasion of less beneficial, that is cheater, genotypes. Recent evidence suggests that both partner choice and sanctioning may play roles in preventing the invasion of less-beneficial rhizobia in legume–rhizobium mutualisms. The importance of these mechanisms in natural communities, however, remains unclear. We grew 12 Medicago truncatula maternal families with a mixture of three rhizobium strains from their native range for three plant generations and estimated the symbiotic benefits (nodule number and size) conferred to each rhizobium strain. In this experiment, the majority of M. truncatula genotypes formed more nodules with more beneficial rhizobium strains, providing evidence for adaptive partner choice. We also found that three generations of symbiosis resulted in an increase in the relative frequency of rhizobium strains that were most beneficial to plants—suggesting that partner choice affects rhizobium fitness. By contrast, we found no evidence that plants differentially rewarded rhizobia postnodulation via sanctioning leading to differences in nodule size. Taken together, our data suggest that plants have evolved to recognize beneficial rhizobial signals during the early stages of symbiosis, and that signaling between plants and rhizobia may be subject to coevolutionary pressures.