2.1.1 Plant-plant interaction experiment
We conducted the plant-plant interaction experiment in a greenhouse located at the Botanical Garden Leipzig, Germany, in May 2017. We recorded an average temperature of 22.6°C and an average relative humidity of 51.6% for the time of the experiment in the greenhouse. We used 2 L microcosms (rose pot 2.0 L, Hermann Meyer KG, Rellingen, Germany) filled with autoclaved (twice at 134°C for 20 min) 50:50 sand-peat (Floradur B Pot Clay Medium, Floragard, Oldenburg, Germany) mixture. We flushed each filled microcosms with water twice to remove pulsed nutrients and toxins prior to transplanting seedlings (Alphei & Scheu, 1993; Trevors, 1996). To allow for similar soil conditions between the plant-plant interaction experiment and the plant-soil interaction experiment (see below), we chose to use a commercial sand-peat mixture as it was not possible to retrieve enough soil from the field site in Jena, Germany. We established the following plant diversity levels and communities: (1) monocultures of each species, (2) the three possible two-species mixtures, and (3) the three-species mixture (Appendix Table A1 ). We transplanted twelve similarly developed seedlings in each microcosm, and each plant community was replicated ten times (total number of microcosms: 70). The relative proportion among species was equal, i.e., six seedlings per species in the two-species mixture and four seedlings per species in the three-species mixture. In the two-species mixture, we transplanted the species in an alternating pattern, while we randomized the position of each seedling in the three-species mixture. All microcosms were randomly placed on tables in the greenhouse and covered with net cages to prevent unwanted herbivory. We watered all microcosms three times per week and randomized the position on the tables every 7 days. We fertilized all microcosms with 250 mL Hoagland solution after 5 weeks to counteract any loss of nutrients and ensure optimal growth.
After 7 weeks of growth, we harvested five microcosms per plant diversity level (see below). The next day, we infested two randomly selected plants per species and microcosm of the remaining microcosms with three 2nd instar Spodoptera exigua larvae each. We covered and closed each plant just above the soil with an organza net to ensure that the larvae could not escape. To ensure similar development of the larvae (eggs purchased from Entocare Biologische Gewasbescherming, Wageningen, the Netherlands), we maintained a laboratory colony on artificial diet in a growth chamber (25°C, 12 h light, 45% relative humidity). After 7 days of herbivory, we harvested the remaining microcosms (see below).