Trophic-level dynamics in microbial-based soil food webs

The aims of the thesis were to test current trophic-dynamic models and to find essential features of microbial-based soil food webs in order to formulate a verbal trophic-dynamic model for them. To achieve the aims I carried out experiments in microcosms filled with sterilised litter-humus mixture. Into microcosms I established simplified soil food webs of various structure containing bacteria and fungi (1. trophic level), bacterivorous and fungivorous nematodes (2. level), and a predatory nematode (3. level). I found that the biomass of microbivores was limited by predation in these food webs, i.e., it was top-down controlled, whereas grazing had a varying and substantially smaller effect on microbial biomass. Fungal biomass appeared to be donor-controlled, i.e., limited by resources only, while bacterial biomass was affected by both resources and grazing. The limitation of microbivore biomass by the predatory nematode did not affect microbial biomass or production, i.e., trophic cascades of biomass or production control did not occur. Microbial respiration was, however, increased when grazers were added, and mineralisation of C and N was loosely correlated with microbivore biomass. Increased microbial production tended to increase biomass at all trophic levels, meaning that all levels were bottom-up controlled. Refuges for microbes and microbivores in soil possibly caused the strong bottom-up control of the two lowest levels. Differences in microbivore diversity of three-level food webs affected significantly trophic interactions and led to idiosyncratic, or unpredictable, changes in trophic-level biomasses and nutrient mineralisation. The idiosyncracy of changes between the diverse and simple food webs followed from differences in microbivore characters - in their efficiency in resource utilisation and vulnerability to predation and competition. Trophic-level dynamics also differed between the bacterial and fungal channels, for bacterial and fungal biomasses responded differently to both grazing and energy addition. An omnivorous nematode, feeding on both bacteria and fungivore, did not substantially change trophic-level dynamics when compared with a food web containing a bacterivore. The results contradict trophic-dynamic models that predict limiting factors - resource availability and predation - to alternate at adjacent trophic levels, and support models predicting all levels to be limited by both resources and predation. Moreover, the results give evidence of discernible energy channels in soils, weak effects of omnivory on trophic-level dynamics and difficulties in predicting the consequences of declining species diversity.