Ecology of microarthropods in arable soil

Soil microarthropods are all free-living mites and collembolans, living in the soil. The study presented in this thesis formed part of the Dutch Programme on Soil Ecology of Arable Farming Systems, an integrated multidisciplinary research programme, focused on the functioning of two differently managed agro-ecosystems: a conventional and an "integrated" system. To understand the mechanisms that control litter decomposition and mineralization of nutrients, knowledge of the role of the microarthropods in the soil food web is indispensable. In this thesis I have studied the effect of microarthropods on the decomposition of organic matter and the mineralization of nitrogen in arable soil. A reliable outlook on the subject could only be achieved, however, by including in the research all possible relationships with other organisms in the soil food web. Therefore, the soil fauna is ordered into functional groups: organisms that share a common resource and a common mode of resource utilization. Microarthropod functional groups are: predatory mites, predatory Collembola, nematophagous mites, omnivorous collembola, cryptostigmatid mites, non-cryptostigmatid mites and bacterivorous mites. Food web interactions and the effect of microarthropods on decomposition and mineralization were studied at three different resolution levels.

In a microcosm experiment, the separate and combined effects of an amoeba ( Acanthamoeba spec.) and a bacterivorous mite ( Histiostoma litorale ) on C- and N mineralization were studied with sterilized silt loam soil, reinoculated with a bacterial solution and with lucerne meal as substrate for the microbes. The aim of the experiment was to separate the direct contribution of the mites to the mineralization from possible indirect effects, such as enhancement of microbial grazing by protozoa. In spite of some problems with the gnotobiosis of the microcosms it was evident from the results that the amoebae enhanced nitrogen mineralization whereas mites had no measurable direct or indirect effect on the nitrogen mineralisation. No grazing effects on the microbial activity were found, since there were no significant differences in the oxygen consumption between the separate treatments.

In a litterbag experiment, the effect of soil microarthropods and enchytraeids on the decomposition of buried wheat straw was studied by selective admission and exclusion. Litterbags with 20m mesh size did not exclude nematodes, but did keep out microarthropods until after 27 weeks of incubation, when a major colonization, mainly by fungivorous collembola and mites, occurred. When litterbags with a complete microarthropod community were compared to litterbags with strongly reduced microarthropod numbers, no differences between decomposition rates were found. In colonized fine mesh bags, a reduced decomposition rate was found compared to the coarse mesh litterbags, probably due to overgrazing of the fungal population by large numbers of fungivorous microarthropods; these large numbers were probably the result of absence of predators. The results indicate that the decomposition rate is predator-controlled. Extraction of microarthropods as well as enchytraeids and nematodes from the coarse mesh litterbags showed a distinct succession during decomposition. The process was dominated in the first phase by bacterivorous nematodes, nematophagous and bacterivorous mites, and in the later phase by fungivorous nematodes, fungivorous and omnivorous mites and collembola, and predatory mites. This succession is indicative for a transition from bacterial to fungal dominated decomposition of the buried organic matter.

Microarthropod population dynamics in arable soil were observed in a two year soil sampling scheme in a conventional and an integrated wheat field as part of the Dutch programme on Soil Ecology of Arable Farming Systems. The most abundant functional groups were omnivorous collembola, omnivorous non-cryptostigmatid mites and predatory mites. Management practice affected the dynamics of most groups, especially soil fumigation, but these effects could not be observed from average yearly biomass values. In the food webs studied in the Netherlands (Lovinkhoeve), in Sweden (Kjettslinge) and in Georgia, USA (Horseshoe Bend), dependence of the abundance of the microarthropods on their food source could not be identified from the annual mean biomass of both. The sampling results further indicated that the microarthropod populations were able to recover rapidly from harsh management practices.

In a comparative field sampling program, soil microarthropods were sampled every three months for one year at ten sites in the northern Netherlands, varying in soil type and land use. Biomass-C per ha of the seven functional groups of microarthropods was calculated for the top 10 cm soil layer. Here, the four overall principal functional groups were: cryptostigmatid mites, non-cryptostigmatid mites, predatory mites and omnivorous collembolans. Possible relations between the mean annual biomass of these groups and soil type, land use or soil organic matter were studied. Microarthropod biomass was larger in sandy than in loamy soil, and generally larger in meadows than in wheat fields, the mineral layer of forest soils being intermediate. Non-cryptostigmatid mite, omnivorous collembola and predatory mite biomass showed strong positive correlations. Cryptostigmatid mite biomass correlated with lower organic matter input quality, while omnivorous collembola and non-cryptostigmatid mites showed a positive correlation to the amount of input. Omnivorous collembola were negatively affected by a discontinuous input of organic matter to the soil. Relationships between functional group biomass and either soil organic matter density fractions or soil pore size distribution were only found, when the grassland sites were analyzed separately. Both analyses showed correlation patterns for cryptostigmatid mites to deviate from those of the other three main functional groups. Cryptostigmatid mites showed a positive correlation with the lightest organic matter density fraction, while the non-cryptostigmatid mites and omnivorous collembola were correlated to the heavier fractions. The cryptostigmatid mites correlated with the 6-90m pore size class, while the other three groups showed strong correlation to the 1.2-6m size class and the largest (> 90m) size class. Both observations lead to the conclusion, that omnivorous collembola and non-cryptostigmatid mites are related to fungal growth (in the largest pores and on the heavy organic matter fraction), while the cryptostigmatid mites show a more saprovorous feeding mode.

Although in none of my experiments a significant effect of the microarthropod community on the decomposition of litter or the mineralization of nitrogen, either positive or negative, could be assessed, a number of conclusions on the ecology of the microarthropod functional groups, their mutual relationships, as well as their relationships with other faunal groups in arable soil can be drawn:

1. In undisturbed situations, negative effects of microarthropods on the decomposition of organic matter are balanced by positive effects.
2. Top-down control of predatory mites on fungal browsing microarthropods and collembolans precludes the negative effects of overgrazing on litter decomposition. Among the microarthropod community, tri-trophic interactions are most important in respect to litter decomposition and nitrogen mineralization.
3. Only the most efficient microbial grazers will directly affect nitrogen mineralization. In most cases these are not the microarthropods, but protozoa and nematodes.
4. Food web interactions are limited by the soil pore size distribution: grazing and predation by microarthropods will only occur in the larger soil pores; competition can only exist between groups or species of comparable body width.
5. The soil food web of agro-ecosystems can be divided into three spatially and temporally separate webs, based on three different carbon sources: the fresh organic matter web, the "old" organic matter web and the rhizosphere food web.
6. Agricultural land use directs the distribution of the main life history tactics of the microarthropods of the "old" organic matter food web towards omnivory, sexual reproduction and/or arrhenotoky and semelparity and/or seasonal iteroparity. The fresh organic matter food web is characterized by highly phoretic specialists.
7. Field samplings are useful in studying the "old" organic matter and the rhizosphere food webs and litterbags are a good tool in studying the fresh organic matter food web.