Field experiments
The experiments were conducted for three years (late March 2016 to early September 2018). Three adult individuals of each species with similar diameter at breast height (DBH) were selected. Due to the size difference between the two species, the DBH of selected adults forO. semicastrata was > 20 cm and for C. patelliormis was > 60 cm. Three treatments were applied to each tree species: warming (two levels: OTC warmingversus control), distance away from the selected adults of each tree species (three distances: 1, 10, and 20 m), and pesticide treatment (two levels: watering with pesticides versus without pesticides). At each distance from a selected adult, an OTC (size 1.6×1.6×1 m) was set up to passively warm the soil (monthly average warming magnitude: 0.89 ± 0.42 °C, Figure 1b). In total, nine OTCs were established for each focal tree species. The OTCs were made of 4 mm thick transparent Plexiglas boards with > 92% light transmission at wavelengths 250–800 nm (Evonik Degussa, Essen, Germany). A control plot with the same size (1.6×1.6 m) was established next to each OTC (ca . 1.5 m away). Control plots were enclosed with a 5 mm mesh steel screen with a 0.5 m height to protect the seedlings from animal trampling or grazing.
Each OTC (and the control plot) was planted with seedlings of three tree species, i.e., the focal species of that OTC (O. semicastrata orC. patelliormis , depending on the parent trees under which OTCs were placed), a congeneric of this focal species, and the other focal species (C. patelliormis or O. semicastrata ) (Figure 1b). The congeneric species of O. semicastrata was O. pinnata , and the congeneric species of C. patelliormis was C. hui . The congeneric species were used for testing whether soil-borne pathogens were specific to their host tree species.
From October to November 2015, seeds of the two focal tree species (O. semicastrata and C. patelliormis ) were collected directly from the selected adult trees to control the possible intraspecific variation in susceptibility to soil-borne pathogens among seed provenances (Liu et al. 2012; Eck et al. 2019). The seeds were germinated in sterilized sand in March 2016. Subsequently, two-week old seedlings were transplanted to two columns of six 40×40 cm quadrats (with planting density of nine seedlings per quadrat) in the paired OTC and control plot (Figure 1b). One column (consisting of three 40×40 cm quadrats) was randomly selected for pesticide treatment, while the other column was treated with the same volume (1 L) of distilled water without pesticides. The pesticide treatment included two selective pesticides (Celest Gold and Ridomil Gold, Syngenta Ltd, Basel, Switzerland). Celest Gold is effective in killing fungal pathogens (e.g., Fusarium spp. and Rhizoctonia spp.) that cause seedling damping-off, a major causative agent of seedling death (Liuet al. 2015), whereas Ridomil Gold is for killing oomycete pathogens (e.g., Pythium spp.) that also cause seedling damping-off (Bell et al. 2006). These pesticides were sprayed every two weeks for the first two years and every six weeks in the third year. The control quadrats were sprayed with the same volume of distilled water. Seedling survival/mortality was recorded at the time when pesticides or distilled water was sprayed.
Soil temperature and moisture were monitored within OTCs and control plots, starting one week prior to seedling transplantation in March 2016 until early September 2018 using microclimate data loggers (Em50-5TM; Decagon Devices Inc., Pullman, WA, USA). The logger sensors were inserted at 10-cm soil depth and data were recorded every 4 hours. Monthly average temperature and moisture data were used for analysis in this study.