Coexistence and collapse: An experimental investigation of the persistent communities of a protist species pool

Abstract

1. Using protists in laboratory microcosms we investigated the long-term coexistence of combinations of species from a six species set to address the following questions: (i) What proportion of the possible communities from the species pool are persistent? (ii) From how many initial states are each of these persistent communities obtained? (iii) What are the impacts of individual species on community collapse? (iv) Can the behaviour of the system be characterized by a simple set of rules? (v) To what extent does knowledge of the results from the pairwise species combinations allow prediction of the outcomes of the more species-rich sets? 2. Replicated microcosms were set up with each of the 63 possible combinations of species from a pool of six protist species (three bacterivores, two predators and one omnivore) along with a mixed bacterial flora. Microcosms were maintained, with nutrient replenishment, for over 100 days until a constant species composition was reached. 3. Forty-seven of the 63 starting combinations showed a repeatable collapse to one of eight communities (including the 'empty' community - containing just bacteria). The number of starting combinations leading to each of these persistent communities varied from 17 down to one. Extinction rates were higher in more species-rich systems. 4. The effects of each species in determining the composition of the persistent community were characterized as to whether they were present in the final set or not, and whether their occurrence in the starting combination had any effect on the eventual presence or absence of others. Under this classification the larger species seem to be more influential in shaping the final community. 5. The collapse from the initial species composition to that at the end could be stated in terms of four simple rules in 34 of the 47 starting combinations for which a persistent community could be defined. 6. Using the outcomes from the single-species and two-species starting sets only, the eventual states to which the larger communities collapsed could be correctly predicted in 27 out of the 31 starting combinations containing three or more species for which persistent communities could be defined.