Ecological Archives E091-232-A3

Kirsten M. Prior and Jessica J. Hellmann. 2010. Impact of an invasive oak gall wasp on a native butterfly: a test of plant-mediated competition. Ecology 91:3284–3293.

Appendix C. Using Neuroterus saltatorius foliar damage as a surrogate for density, and assignment of damage categories for butterfly (Erynnis propertius) performance experiment.

Using Neuroterus saltatorius foliar damage as a surrogate for density

We used damage as a surrogate for gall wasp density. In 2006, we estimated density on ten trees at ten sites on Vancouver Island, British Columbia (BC) by counting galls on 15 leaves on five different branches. At each site we also estimated damage levels of 50 randomly chosen trees by assigning each tree to one of four damage categories based on our observations of gall damage on numerous branches at eye level: ‘no galls’ if no galls were observed, ‘low’ if the majority (e.g., greater than ~50%) of surveyed leaves on the tree had less than ten galls or less than ~ 25% of gall-damaged area, ‘moderate’ if gall damage was more extensive, but still spotty (e.g., ~25% to ~75% gall-damaged area), and ‘high’ if large continuous portions of the leaf were covered with damage (e.g., >~75%) (Fig. C1). We calculated average gall wasp density at a site by averaging the total number of galls counted on each tree. We calculated gall wasp damage at a site by calculating the average percent damage of 50 trees using the mid-point of each category (e.g., low is 12.5 %). We conducted a linear regression between average gall wasp damage and average gall wasp density among all sites and found that assigning trees to damage categories is a good estimate of gall wasp density at a site (R2 = 0.66, P = 0.004, n = 10; Fig. C2).

FigC1
 
   FIG. C1. Representative leaves from trees placed in one of three damage categories (damage caused by invasive gall wasp Neuroterus saltatorius): (a) low, (b) moderate, and (c) high. An additional category, ‘no galls’, was used when assessing damage as a surrogate for gall wasp density.

 

FigC2
 
   FIG. C2. Linear regression between percent damage and density of Neuroterus saltatorius at ten sites in BC. Percent damage was measured by calculating the average of 50 trees that were placed in one of four damage categories (see Fig. C1) and taking the average of the midpoints of each category. Density was measured by averaging the total number of galls counted on 15 leaves from five branches on ten trees at a site.


 

Assignment of density categories for butterfly performance experiment

In mid-June 2007, nine trees were chosen at S1 (see map in Appendix A) and placed into one of the three density categories (described above as damage categories) (Fig. C1). To make sure that our coarse assignment of trees to density categories was accurate, we took pictures of ten randomly chosen leaves within each experimental enclosure (n = 3 per tree) at the end of July and used APS Assess (Lamari 2002) to calculate damage (‘percent damage’). We also assessed damage in our control enclosures (1 per tree) by collecting ten leaves, scanning them, and using APS Assess to calculate percent damage. APS Assess is designed to measure the percent of foliar damage on a leaf by detecting different wavelengths of healthy foliage (i.e., green) vs. damaged foliage (i.e., yellow to brown). A nested ANOVA revealed greater differences in percent damage among density categories in experimental enclosures (F2,6 = 35.37, P < 0.0001) than among trees within density categories (F6,18 = 5.26, P = 0.003), showing that our assignment of trees to different density categories was successful (Fig. 1 in Manuscript; Fig. C3). Variation in percent damage within enclosures was highest in the high-density trees (e.g., ranging from 5% to 99%). Damage ranged from 0.1% to 14% in the low-density trees and from 5% to 30 % in the moderate-density trees (Fig. C3). In addition, our control enclosures were relatively good representatives of the distribution of damaged leaves that the caterpillars were foraging on in experimental enclosures (Fig. C3).

Each enclosure contained an abundance of leaves (i.e., in excess of ~100). Caterpillars in these enclosures were not limited by the quantity of food because defoliation of leaf material within enclosures did not occur by the end of the season (also see Hellmann et al. 2008, Prior et al. 2009). Caterpillars are small, sedentary, and live in leaf folds and we think that it would be highly unlikely that they would be able to move among trees. In fact, we think that they rarely move among branches as we often find individuals inhabiting the same leaf fold over multiple sampling occasions. While we found caterpillars in leaf folds on damaged leaves we do not know if they prefer to forage on damaged or undamaged leaves.

During the experiment, N. saltatorius caused the majority of damage on Q. garryana. In previous years high levels of defoliation due to Malacosoma californicum (Packard)(Lepidoptera: Lasiocampidae) and Operophtera brumata (L.) (Lepidoptera: Geometridae) were observed, but these species were observed in low abundance 2006–2008 (K. Prior, personal observation).

FigC3
 
   FIG. C3. Box-plot of damage caused by Neuroterus saltatorius in enclosures. White bars represent the percent damage of leaves in experimental enclosures where each bar represents an enclosure. A pooled version of this figure is Fig. 1 in the manuscript. Gray bars represent the percent damage of leaves in control enclosures that were not stocked with caterpillars, and were used to estimate the nutritional quality of the tree. Percent damage was digitally assessed from (7–10 leaves per enclosure). Top and bottom bars represent the 90th and 10th percentile respectively. The top and bottom of the boxes represent the 75th and 25th percentile respectively, and the horizontal bar represents the median.

LITERATURE CITED

Hellmann, J. J., S. L. Pelini, K. M. Prior, and J. D. K. Dzurisin. 2008. The response of two butterfly species to climatic variation at the edge of their range and the implications for poleward range shifts. Oecologia 157:583–592.

Lamari, L. 2002. Assess: Image analysis software for plant disease quantification. APS Press, The American Phytopathological Society, St. Paul, Minnesota, USA.

Prior, K. M., J. D. K. Dzurisin, S. L. Pelini, and J. J. Hellmann. 2009. Biology of larvae and adults of Erynnis propertius at the northern edge of its range. The Canadian Entomologist 141:161–171.


[Back to E091-232]