Appendix 2—figure 1. Generating species with sorted $\mathit{r}$ and $\mathit{K}$ parameters prevents intersection of growth curves but not niche flip.

(a) Growth curves of 10 species on a single resource are shown. For these species in a single resource environment, the outcome of competition does not depend on resource concentration or dilution rate. (b) Communities with sorted $\mathit{K}$ per resource still exhibit U-shaped DDR. Shannon diversity of communities vs. mean dilution rate. Color indicates frequency of fluctuations. Calculations performed after 6 days of simulated growth, and averaged across 100 communities with randomly generated and sorted $\mathit{r}$/$\mathit{K}$ parameters. (c) Niche flip occurs in a community with sorted $\mathit{K}$ parameters at high dilution rate. At low dilution rate, ZNGIs intersect the axes and the resource consumption vectors outline a coexistence region. At high dilution rate, species cannot survive on a single resource, so ZNGIs do not intersect the axis. The relative positions of ZNGIs and consumption vectors (i.e. invasion boundaries) have flipped at high dilution rate, consistent with niche-flip.