| dc.description.abstracteng | Soil food webs are an essential part of terrestrial ecosystem functioning and characterised by a high degree of cross linkage between the members of a highly diverse soil community. Centipedes are abundant predators in the litter and soil layers of temperate forests. They are assumed to be generalist predators, feeding on a wide range of prey such as collembola and earthworms. However, knowledge of their feeding ecology is scarce, as the opaque habitat, the high diversity of prey, and extra oral digestion hamper analysis of their feeding behaviour. Molecular gut content analysis, however, allows their trophic interactions to be studied even under these unfavourable conditions and consequently allows assessment of prey choice and of the strength of the predator-prey interactions. I therefore used group and species-specific PCR assays to track the DNA of abundant prey in guts of lithobiid and geophilomorph centipedes to illustrate feeding interactions. Based on these results I examined the effects of a variety of biological and environmental factors on centipede prey choice and the strength of predator-prey interaction. In addition, I conducted experiments to study the effect of two factors on prey DNA detection success which could lead to over-or underestimation of feeding link strength.
In Chapter 2 I report on the design and optimization of group- and species- specific PCR assays to screen soil predators for twelve abundant prey including extra- and intraguild prey. Cross-reactivity tests against a wide spectrum of non-target animals and prey DNA dilution tests confirmed high specificity and sensitivity of assays. An initial screen of the gut contents of 50 lithobiid individuals revealed that centipedes feed on many prey species though showing preferences for collembolan and dipteran prey.
Prey DNA detection success can be affected by a variety of factors some of which lead to over- or underestimation of the strength of predator-prey interactions. In Chapter 3 I tested whether the success with which prey can be detected in a predator’s gut is positively correlated with decreasing predator body mass and with prey quality. I used singleplex PCR and quantitative real-time PCR to analyse the time-dependent reduction in DNA detection success for three qualitatively different prey (Collembola, Diptera, Lumbricida) which were fed to lithobiid predators. Likewise I analysed singleplex detection rates of collembolan prey in three centipede size classes. Contrary to my expectations, prey DNA detection success is not affected by predator body size nor prey quality but by PCR assay sensitivity and prey DNA copy number. This suggests that both DNA concentration and assay sensitivity need to be considered when assessing prey quality effects on prey DNA detection success.
Land-use change can affect feeding ecologies of soil predators by changing resource availability and habitat structure. In Chapter 4 I therefore studied prey choice of staphylinid and centipede predators along a land-use gradient in forest ecosystems of two regions. Predators were screened for collembolan, dipteran and lumbricid prey and the results were tested against a set of environmental and biological parameters. In fact, forest management does not affect prey choice but the depth of the litter layer and soil pH are important factors. Likewise, trophic interactions varied with prey abundance. In all cases I observed that the direction of effect depends on predator body mass. Large predators feed more in low-structured habitats and at high prey densities, while the opposite is true for small predators.
Functional response models are used to predict the strength of trophic interactions, which can be affected by predator-prey body mass ratios. In Chapter 5 I analysed the feeding behaviour of lithobiid and geophilomorph centipedes and compared the results with the body-size-dependent functional response for eight different prey groups. I showed that calculated feeding rates of most prey are significantly correlated with prey DNA detection rates for lithobiid predators. Intraguild prey and lumbricids however correlated negatively, indicating that prey-specific traits must be taken into account to fully explain feeding interactions in soil food webs.
Overall, I show that molecular gut content analysis resolves trophic interactions between centipedes and their prey. These interactions are defined by a variety of factors of which body size is the most important. | de |