Colonization History and Alternative Community States in Experimental Microcosms

Using s suite of disparate experimental systems, three tests of the effect of variation in community history on community states were performed. The first test explored the effect of species invasion order on the structure and invasibility in soil microbial communities. Microcosm communities were assembled by augmenting an existing soil community with sequential introductions of three bacterial strains under three alternative sequences. Assembled communities were then probed with a genetically engineered bioremediative bacterium to test the relative vulnerability of these communities to this strain. Results indicated that variation in invasion order resulted in the production of alternative community states with distinct vulnerabilities to invasion. The second test explored the effect of the interaction of variation in invasion sequence with varying productivity level on community composition and media chemistry properties. Detritus- based aquatic communities consisting of bacteria and protists were assembled under two alternative sequences on a gradient of five nutrient concentrations. In total, five unique community states were found to emerge from the interaction of species order and productivity level. Alternative states arising from sequence effects were found at two of the five nutrient levels tested. In addition, sequence effects were found to produce unique biologically-mediated changes to media chemistry. Notably, such effects were not necessarily reflected by observable changes at the species compositional level. The third test evaluated trends in species turnover data for evidence of convergence in community composition among a suite of 15 artificial pond microcosms established at the same location. Ponds were arranged in five clusters of three. Here, an explicit manipulation of invasion sequence was not performed; rather, microcosms are assumed to possess similar histories and similar environments based on spatial proximity. Evidence of convergence was sought using five alternative compositional classification schemes. No evidence for convergence for the overall study site was found. Results of multiple analyses indicated a weak degree of convergence at the cluster level. Disturbance arising from multiple heavy rainfall events, however, had a strong disruptive effect on this convergence. Stronger evidence was found for a divergence in composition between two sets of microcosms that were independent of spatial proximity.