Explaining the persistence of less-mutualistic rhizobia despite legume host sanctions

Simultaneous infection by several rhizobium strains per individual plant creates a potential tragedy of the commons.  If the individual cost to rhizobia of fixing N₂ were balanced only by collective benefits (e.g, due to increased host photosynthesis), then “cheating” rhizobia, which invest more in their own reproduction at the expense of N₂ fixation, would displace more mutualistic strains.  The discovery that legumes impose sanctions on nodules that fix less N₂ solves this problem, but creates another: why haven’t sanctions reduced the frequency of cheaters to negligible levels?  Are mixed nodules common enough that cheaters might escape sanctions by sharing nodules with better N₂ fixers? DNA fingerprinting of multiple isolates per nodule showed that mixed nodules are common in annual lupines in the field.  In soybean, intermediate rates of N₂ fixation (as expected with mixed nodules) resulted in partial sanctions.  Sanctions would presumably affect the fitness of all rhizobia within a nodule, but not necessarily equally.  Chemical manipulation of host metabolism could decouple rhizobium fitness from any actual benefit to the host.  For example, we found that rhizobia that interfere with ethylene signaling accumulated more resources in nodules.  Ideally, rhizobium fitness would be measured over their entire life cycle, including the crucial stage of escape from nodules into the soil.  Slow escape and loss to predation can each reduce the number of rhizobia recolonizing the soil well below what would be expected from rhizobium numbers inside nodules.  The question is whether these processes have any differential effect on rhizobia differing in mutualism.