Swope & Stein: Soil type mediates indirect interactions between Centaurea solstitialis and its biocontrol agents
ESM Table 1 Soil nutrient status (ppm) at the two study sites at McLaughlin Natural Reserve (mean ± 1 SD)
Soil typeNPKCaMgCa/MgpH
Non-serpentine3.3330.67288.172046.83720.833.076.28
±1.03±17.36±39.48±340.05±219.93±1.12±0.23
Serpentine6.908.68253.55675.982617.500.286.98
±0.93±6.67±20.32±73.14±876.69±0.10±0.13
ESM Table 2 The purpose of this mini-experiment was to determine if the seed predators (Eustenopus villosus and Chaetorellia spp) were choosing among the available inflorescences randomly or if they were selecting inflorescences based on size. At 10 – 14 day intervals, we observed insects in the field ovipositing on an inflorescence which we then sprayed with Ortho Systemic Insect Killer (formerly Isotox®) to kill the egg/larva; unattacked inflorescences were also sprayed with Isotox. Isotox does not affect pollinator visitation (Garren and Strauss 2009). Inflorescences were open to natural pollination and allowed to mature in the field. We collected inflorescences just prior to seed dispersal and dissected them in the lab to count the total number of ovules per inflorescence.
Note that we used total ovules and not viable seeds per inflorescence because the insects lay their eggs prior to petal expansion and pollination (i.e., they may be using total ovules as a cue when selecting oviposition sites but they can’t be using viable seeds as there aren’t any at the time of oviposition). For the number of viable seeds per inflorescence, see ESM Table 5.
SourcedfMSFP
Inflorescence type*1100.3640.6060.55
Soil type166.13.950.19
Pathogen infection12.9690.0190.89
Error1068165.625
* Inflorescence type refers to the following categories: unattacked inflorescences, inflorescences selected as an oviposition site by E. villosus, and those selectedas oviposition sites by Chaetorellia spp.
Mean ovules per inflorescence (Centaurea solstitialis) that were and were not selected as oviposition sites by the biocontrol seed predators Eustenopus villosus and Chaetorellia spp.
Ovules per inflorescence (mean ±SD)Uninfected / +Puccinia
Unattacked / E. villosus / Chaetorellia spp / Unattacked / E. villosus / Chaetorellia spp
Non-serpentine / 36.4 ± 11.84 / 34.91 ± 12.68 / 36.07 ± 13.03 / 31.24 ± 13.57 / 34.35 ± 14.61 / 32.81 ± 14.03
Serpentine / 32.95 ± 13.75 / 27.48 ± 11.82 / 32.64 ± 13.01 / 34.42 ± 15.52 / 29.54 ± 9.98 / 37.01 ± 14.72
Garren JM, SY Strauss (2009) Population-level compensation by an invasive thistle thwarts biological control from seed predators. Ecological Applications 19:709-721.
ESM Table 3 Plant calcium content on the two soil types (mean ± 1 SE)
Plant tissue Ca ug g-1
(mean ± 1SE)
SourcedfMSFP Non-serpentineSerpentine
Soil type12.248e9317.8070.000111920.733985.47
Error1687073399.1± 603.07± 136.31
ESM Table 4 Percentage of leaves with pustules on the two soil types (mean ± 1 SE)
Percentage of leaves infected
(mean ± 1SE)
SourcedfMSFP Non-serpentineSerpentine
Soil type10.1018.0560.00526.730.9
Error1470.013± 1.1±1.0
ESM Figure 1 Degree of Puccinia jaceae solstitialisinfection onCentaurea solstitialis plants growing on non-serpentine and serpentine soils (F1,146=156.067, MS= 156.067, P=0.0001). Error bars represent ± 1SE
ESM Table 5 Number viable seeds per unattacked inflorescence (mean ± 1 SD)
Non-serpentine Serpentine
Uninfected+PucciniaUninfected+Puccinia
No. of seeds per inflor.25.7826.2426.8026.66
±10.56±14.05±8.89±10.30
No. of viable seeds per inflorescence in the absence of seed predation
SourcedfSSMSFP
Soil type10.2130.2131.2340.27
Pathogen infection10.6260.6260.0040.95
Interaction1393.34393.342.2790.15
Error19733303.84172.559
Note that the number of ovules per inflorescence (ESM Table 2) differs from the number of viable seeds per inflorescence (presented here). In the mini-experiment presented in EMS Table 2, we focused on the number of ovules per inflorescence because the insects lay their eggs prior to petal expansion and pollination (i.e., they may be using total ovules as a cue when selecting oviposition sites but they can’t be using viable seeds). Here we focused on viable seeds (pollinated ovules) because it is the appropriate comparison when estimating both the pathogen’s impact on the plant and seed consumption using our effect size calculation.
ESM Figure 2 The effect of soil type and pathogen (Puccinia jaceae solstitialis) infection on the proportion of seeds consumed (per inflorescence) by the biocontrol seed predators (A) Eustenopus villosus (soil: F1,316=16.302, MS=3.553, P=0.0001; pathogen: F1,316=24.451, MS=5.329, P=0.0001; interaction:F1,316=0.028, MS=0.006, P=0.87); (B) Chaetorellia australis andC. succinea(sample size too small to analyze); and (C) Urophora sirunaseva (sample size too small to analyze). Note the different scale on the Y-axes. Data were arcsine transformed for analysis; untransformed data are shown. Error bars represent ± 1SE
1