Ecological Archives E095-167-A1

Bitty A. Roy, Kenneth Hudson, Matt Visser, Bart R. Johnson. 2014. Grassland fires may favor native over introduced plants by reducing pathogen loads. Ecology 95:1897–1906.

Appendix A. Experimental design.

This appendix illustrates the experimental design in idealized state (Fig. A1, and as it was in reality, on the ground (Figs. A2–A4).


Fig. A1. The sampling design was hierarchical: plants within plots within treatments. We used a stratified random burn and sampling pattern to ensure that both treatments and controls were applied across the full gradient of vegetation types. Within each treatment cell a 5 × 10 m macroplot was randomly located and permanently marked. For pathogens and herbivores, we sampled only the lower left 1-m² plot in each macroplot. To census damage in the 1-m² plots, we divided them into sixteen 25-cm² cells with string. Within each 25-cm² cell, we found the nearest "target" plant host species to the center of the cell. All target hosts were searched for in all 25-cm² cells, but were not always present, thus the number of plants sampled for each species in each plot ranged from 0–16.



Fig. A2. Experimental design mapped onto a 1995 U.S. Government air photo. Figure by B. R. Johnson.



Fig. A3. The field site shortly after the burn on 9 October 2002. Two of the eight burned cells are visible in this picture.; Note the fire reduced thatch. Photograph by B. A. Roy.



Fig. A4. One of the macroplots the spring after the burn, May 2003. Photograph by B. A. Roy.

[Back to E095-167]