Multi-level complexity in the use of plant allelochemicals by aposematic insects

Summary As recognized by Miriam Rothschild as early as the 1960s and repeatedly emphasized in her papers, the use, misuse, or non-use of plant allelochemicals by insects is extremely variable and difficult to predict, at many levels of time, space, and biological organization. Although certain patterns that reoccur have been important in the development of ecological theory, the optimization of cost-benefit equations involving two or three trophic levels, each with large numbers of individuals, populations, and species in erratic and complex interactions, produces unexpected and fascinating scenarios. The development of rapid colorimetric and chromatographic analyses for several types of plant allelochemicals, notably certain groups of alkaloids, cardiac and cyanogenic glycosides, phenolics, terpenes, and glucosinolates, has permitted a detailed investigation of the variation and flow of these substances in natural organisms and ecosystems. The results of these analyses, in our hands mostly for pyrrolizidine alkaloids (PAs), do not suggest a straightforward classical choice by the aposematic insect to simply sequester or synthesize its defences. Rather, they reveal a confusing variety of diffuse and complex patterns that become increasingly closer to chaos as they are multiplied across structures, species, sexes, stages, sites, seasons, and selective regimes. We present a model reflecting results of analyses at this chemoecological interface. Depending upon an initial option, involving the recognition (or not) of a plant allelochemical, the herbivore will face thereafter options to ingest it (or not), and then to tolerate and absorb (or detoxify and excrete), modify (or not), passively, actively or selectively accumulate, turn over (or not), distribute (or concentrate), and use this compound in a variety of growth, defense, or reproductive functions. The herbivore can also quantitatively or qualitatively regulate the intensity or dispersion of its attack on the plant tissues, in order to modify feedback loops of selection on the plant and its chemicals which exist in most of the earlier steps, or those with its predators and parasites that occur in the later ones. Options that lead to mutualism through positive feedback loops will tend to accumulate and become rapidly fixed by natural selection. Additional variations and anomalies such as automimicry, chemical mimicry, sexual dimorphism and communication, selective sequestration and passing-up of allelochemicals, special glands and structures, and synergism effects, are among the secondary complications of this model that have occupied much thought, time, experimental labor, and polemical space in chemical ecology journals and meetings. Examination of the tendencies and results at various points in the model can be used to explain these features and to make further predictions, plan experiments, and devise activity-based bioassays and new chemical analyses. These may lead some day to new and more robust visions of the major patterns of chemical transfer at this widespread and important natural interface.     

Responses of three mouse species to deterrent chemicals in the monarch butterfly. I. Taste and toxicity tests using artificial diets laced with digitoxin or monocrotaline

Summary Of three common mouse species at the Mexican overwintering sites of the monarch butterfly, onlyPeromyscus melanotis eats monarchs. We hypothesized thatP. aztecus andReithrodontomys sumichrasti reject monarchs because they are more sensitive to the bitter taste and/or toxic effects of the cardiac glycosides (CGs) and pyrrolizidine alkaloids (PAs) in the butterflies. Two-choice preference tests revealed no difference in taste avoidance thresholds to free base and N-oxide forms of the PA, monocrotaline, but very different avoidance thresholds to the CG, digitoxin. Avoidance thresholds forR. sumichrasti andP. aztecus were, in respective order, 1020 and 34 times less than that forP. melanotis. We also tested the toxic sensitivity of juvenile mice by chronically feeding diets containing digitoxin or monocrotaline at concentrations similar to those used in the preference tests. No species developed CG toxicity, but bothP. melanotis andP. aztecus developed moderate PA toxicity (R. sumichrasti was not tested for PA toxicity).P. aztecus grew more slowly and manyP. melanotis had hepatic metabolic lesions. Thus, the three mouse species responded very differently to the taste and toxic properties of CGs and PAs at ecologically relevant concentrations: 1) CGs were taste rejected by all species exceptP. melanotis, while PAs were not; and 2) PAs were toxic, while CGs were not.