The evolution of plant-insect interactions: Insights from the tertiary fossil record

Abstract

Plant-feeding insects are the most species-rich group on the planet today. Models have been proposed to explain this diversity, but few use the fossil record to evaluate hypotheses. I conduct studies in modern systems to examine (1) taphonomic biases in insect preservation and how this may affect our understanding of insect diversity trends through time and (2) patterns of herbivory in modern ecosystems to improve the comparability of fossil and modern datasets. I then use the Cenozoic fossil record to examine the history of ecological associations between insects and plants and how these interactions respond to environmental change. I conducted an actualistic study on the preservation of beetles in Willcox Playa, an ephemeral lake in SE Arizona. I compared the insect death assemblage in shoreline sediments to the living beetle assemblage. The sediments captured 56% of the live-collected beetle families, and 28% of the live-collected beetle genera. The relative abundances of living beetles were not reflected in the death assemblage. Beetle diet, feeding habitat, and size influenced the composition of the death assemblage. Necrophagous, ground-dwelling and smaller beetles were over-represented in the death assemblage. Such biases should be considered in insect paleoepology and in studies of diversity change. Annual variation in herbivory was compared within and between two lowland neotropical forests Costa Rica. Herbivory did not vary significantly within sites between years, but was significantly different: between sites. Modern herbivory data collected with discrete sampling techniques is compartable to herbivory data from fossil forests. Herbivory data from one-time collections of leaf litter are most suitable for comparison with fossil herbivory. I compared herbivory between two Eocene floras and between the Eocene floras and six modern floras. A decline in levels of herbivory corresponds with a decline in temperature from the middle to the late Eocene. Fossil herbivore damage was significantly lower than modern herbivore damage. This pattern may result from taphonomic bias, environmental differences between the fossil and modern sites or evolutionary change.