Interactions between cacti and cactus -feeding insects: Causes and consequences of variation

I examined sources of variation in the abundance and dynamics of herbivorous insects, and evaluated the consequences of this variation for the fitness and population dynamics of their host-plants. As a model empirical system, I focused on the tree cholla cactus (Opuntia imbricata) and specialist cactus-feeding insects in the Chihuahuan desert of central New Mexico. My work on variation in cactus-insect interactions was conducted at two spatial scales. First, at the local scale, I found that variation among co-occurring plants in the abundance of the cactus bug (Narnia pallidicornis), a flower-bud feeding insect, was driven by differences in host-plant meristem allocation between growth and reproduction, an important plant performance trait that has not been previously recognized for its influence on higher trophic levels. Plant allocation to flowering had positive effects on insect fecundity and weakened the strength of density-dependence. Experimental reductions in predation and inter-specific competition had no strong effects on cactus bugs, implicating the bottom-up influence of host-plant allocation as the dominant force in cactus bug dynamics. The linkage between cactus reproductive effort and herbivory led to an increase in the fitness costs of plant reproduction. Field data and dynamic programming models indicated that herbivore-mediated costs of reproduction (feedbacks between reproduction and herbivory) can play an important selective role in the evolution of plant life history strategies. I suggest that these costs warrant greater attention in the context of life history theory. Second, at a landscape scale, I found strong and systematic variation in herbivore pressure across an elevational habitat gradient, which was likely driven by spatial variation in the strength of ant-plant protection mutualism. Tree cholla density also varied across the gradient, and was inversely related to herbivore pressure. A combination of field experiments and demographic models showed that insect herbivores have negative impacts on cactus population dynamics, and that systematic spatial variation in these impacts limits the distribution of these plants across the elevational gradient. My results provide new insight into the linkages among mutualism, herbivory, and plant population dynamics and distribution, and the role of spatial variation in mediating the outcomes of these interactions