P. aeruginosa displaced the resident community
To determine how P. aeruginosa affected resident populations, we first tested invasion and pulse effects on resident density (Fig. 4; Table 1). Resident density was significantly affected by the main effect of pulse and by invasion (pulse frequency: likelihood ratioχ 2 =364.1, d.f.=3, p<10-4; invasion likelihood ratioχ 2 =57.5, d.f.=1, p<10-4). Furthermore, the negative effect of pulse on resident density was significantly greater in the invaded treatments compared to the non-invaded treatments (interaction between pulse and invasion: likelihood ratio χ 2 =62.1, d.f.=3, p<10-4; Fig. 4A). In contrast, although the total density (resident + invader CFU) decreased with pulse frequency (likelihood ratio χ 2 =242.3, d.f.=3, p<10-4), invasion only had a marginal, non-significant, effect on density (likelihood ratioχ 2 =2.9, d.f.=1, p=0.09). Crucially, we did not find a statistically significant interaction between pulse frequency and invasion on total density (i.e., resident plus invader; likelihood ratio χ 2 =0.75, d.f.=3, p=0.86). Taken together, these results imply that the pulse frequency determines the density of the total community but that some of the residents are displaced by the invader (Fig. 4B).
To determine if P. aeruginosa displaced each resident species equally, we used a multivariate approach to test how the densities of each of the five resident species were affected by pulse frequency, invasion, or their interaction (Fig. 5, Fig. S2). The density ofA. agilis was not affected by the invasion (p ≥ 0.12 for all covariates). The densities of P. corrugata , S. rhizophila , and V. guangxiensis , however, were greatly reduced by invasion at high pulse frequencies (P. corrugata in the 2 and 4-day treatments, V. guangxiensis in the 2, 4, and 8-day treatments, and S. rhizophila in the 2-day treatment only), to the point where they were undetected in 2-day disturbance frequency samples (Fig. 5). O. daejonense followed a similar, but less severe, trend. This demonstrates that P. aeruginosa reduces the relative proportions of some species rather than reducing the density of each species equally (Fig. 5; Fig. S2). We see that S. rhizophilais resistant to displacement by the invader in all but the highest disturbance treatment, showing that this effect differs among residents. We therefore demonstrate that high pulse frequency facilitates the establishment of P. aeruginosa , which negatively affects residentP. corrugata , S. rhizophila , and V. guangxiensisdensities to the point of local extinction, and this results in reduced diversity.