The dimensions of fear: Can we link genes to ecosystems?

Background/Question/Methods

Predator effects on prey demography have traditionally been ascribed solely to direct killing in studies of population ecology and wildlife management.  Predation risk of course affects the physiology and behaviour of prey but such effects have largely been ignored in wildlife conservation and management because there remains no ready means of quantifying whether and to what extent predation risk affects prey numbers.  We have conducted two large-scale field experiments conclusively demonstrating that predation risk alone (i.e. fear itself) can affect the demography of free-living wildlife.  In both cases we assessed effects on both physiology and behaviour.  We know the effects on physiology and behaviour were due to predation risk and we know they were associated with effects on demography, because we conducted manipulative experiments.  Our results answer the theoretical question of whether fear itself is powerful enough to affect wildlife population dynamics, and consequently have conservation implications. However, to be of practical value we need to be able to say that this or that measure of physiology or behaviour is most predictive of there being significant fear effects on demography, and that ignoring fear effects thus risks underestimating the total impact of predators.

Results/Conclusions

We have assayed the effects of predation risk on a variety of physiological measures, including glucocorticoid and corticosteroid binding globulin levels, oxidative status and antioxidant capacity.  Measuring glucocorticoids has proven particularly unproductive because it is possible to interpret any change in glucocorticoids as either beneficial or deleterious.  To overcome such vagaries in the interpretation of any given index, our approach has been to measure multiple physiological variables, and use multivariate statistical procedures.  Measuring multiple ‘risk factors’ is common practice in medicine and seems the most sensible means of evaluating the likelihood that fear affects demography.  However, other results from our experiments suggest this may not be as cost-effective as assaying crude measures of condition, e.g. fat score, or measuring behaviour, rather than physiology.  Part of the challenge lies in trying to link changes in physiology and behaviour that may vary from minute-to-minute or hour-to-hour, with demographic processes such as reproduction or death from malnutrition, that take days or weeks.  One solution may be to assay measures of lasting damage, such as oxidative damage to DNA or the attrition of telomeres.  We are presently testing whether predation risk causes such lasting genetic damage, and hence whether we can link fear effects on genes to effects on ecosystems.