Premise of study- Sparse understory communities, in association with non-native tree species, are often attributed to allelopathy, the chemical inhibition of one plant by another. However, allelopathy is a difficult ecological phenomenon to demonstrate with many studies showing conflicting results. Eucalyptus globulus, a native tree to Australia, is one of the most widely planted trees around the world. Sparse understories are common beneath E. globulusplantations and are often attributed to allelopathy, but the ecological impacts of E. globuluson native plant communities are poorly understood. 

Methods -To assess allelopathy as a mechanism of understory inhibition, we tested volatile- and water-soluble leaf extracts on seed germination of California native plants. We also quantified germination rates and early seedling growth of native plants grown in soil from E. globulusplantations versus soil from an adjacent native plant community. 

Key results- Volatile compounds from E. globulusdid not significantly reduce germination for any species. Inhibition from water-soluble E. globuluscompounds was comparable to that of a native tree, Quercus agrifolia (10%).Eucalyptus globulussoil supported germination and early seedling growth of native species equal to or better than coastal scrub soil, although species responses were variable. 

Conclusions- In contrast to previous studies, our results fail to support the hypothesis that E. globuluschemically inhibits germination of native species. California native plants germinate and grow well in soils from E. globulusplantations, which may have significant implications for management and restoration of land historically occupied by E. globulusplantations. 

Germination in the presence of volatile compounds - We attempted to germinate seeds of native species (Acmispon glaber (Vogel) Brouillet, Eschscholzia californica Cham., Lupinus succulentus Douglas ex K. Koch, Festuca microstachys Nutt., and Stipa pulchra Hitchc.) under an air control versus under the presence of volatile compounds from fresh leaves of E. globulus (treatment of interest), S. apiana (positive control; Muller and Muller, 1964; Muller et al., 1964; Muller, 1966), or Q. agrifolia (native tree treatment). Fresh leaf material was cut into 1.25 to 2.5-cm-wide pieces, and approximately 2 g of fresh, cut leaf material was placed in a small glass dish. The small dish was then placed in the center of a 90-mm-diameter glass petri dish (large dish) lined with filter paper. The small petri dish functions to isolate the torn leaf material from interacting with the germinating seeds except through the effect of volatile compounds. Twenty seeds of one species were placed on the filter paper, watered with 7 mL of deionized water, sealed with parafilm, and left for two weeks, at room temperature in the dark, before we recorded germination for each seed. Ten replicate dishes were tested for each of the four treatments (E. globulus, S. apiana, Q. agrifolia, and no leaves) for each test species. In all germination tests, seeds were considered germinated if the seed coat was broken and a hypocotyl or radicle was visible.

Germination in the presence of water-soluble compounds - We quantified germination rates of native species (A. glaber, E. californica, L succulentus, F. microstachys, and S. pulchra) in a water control versus in the presence of volatile compounds from fresh leaves of E. globulus, S. apiana (positive control), or Q. agrifolia (native tree treatment). Water-soluble compounds were extracted by macerating 10 g of fresh leaf material with 250 mL of deionized water for, 20 seconds in a blender. The filtrate was strained through four layers of fine-mesh cheesecloth and diluted with enough deionized water to bring each solution up to 500 mL. Twenty seeds of one test species were placed on filter paper in 90-mm-diameter glass petri dishes and watered with 7 mL of the prepared filtrate in the trial or with deionized water. We used five replicate dishes of each treatment (E. globulus, S. apiana, Q. agrifolia, and water) for each of the five native test species. All dishes were left at room temperature in the dark for two weeks before we recorded germination.     

Germination & growth in field-collected soil - We conducted two greenhouse experiments using field-collected soil to document the capacity for native species to germinate and establish in the presence of naturally-occurring concentrations of potentially allelopathic chemicals from E. globulus. We sampled soil from two plantations of E. globulus as well as from two adjacent coastal scrub communities (scrub soil). We collected soil in two different seasons, in April, 2015 (spring collection/experiment) and December, 2015 (winter collection/experiment) to capture any potential seasonal differences in accumulation of chemical extracts (del Moral and Muller, 1970). During each soil collection, duff and leaf litter were brushed off the soil surface, and soil was collected from the surface to, 20 to 25 cm depth. In the spring soil experiment, we germinated seeds in both soil types from each site and measured seed germination and total biomass at the end of the experiment. Ten seeds of each of three native species were sown together in 12 large (38cm2 x 25 cm deep) plastic flats, with three replicate flats for each of the four soil collection locations. We tested seed from species common to the native coastal scrub and grassland plant communities that are most likely impacted by the establishment of E. globulus plantations throughout California. We used seed from Lupinus succulentus Douglas ex K. Koch, Stipa pulchra Hitchc., and Festuca microstachys Nutt.. Two other species, Baccharis pilularis DC and Salvia mellifera Greene seed were also sown but failed to germinate across all treatments, so were excluded from the study. Seeds were obtained from S&S Seeds (Carpinteria, California) and came from locally-sourced stock from the Central Coast of California. The seeds of each native species were planted in a randomly selected grid location within each 38-cm flat. Each flat received enough water to maintain saturated conditions, between, 200 and 600 mL once or twice daily for the duration of the experiment. We recorded germination weekly and measured above-ground biomass two months after planting. The above-ground biomass of germinated plants was harvested, placed in individual paper bags, dried for 72 hours at 49 C°, and weighed.     In the winter soil experiment, our design and species selection were slightly different to increase our statistical power.  We used seeds of A. glaber, E. californica, Lupinus succulentus, and Stipa pulchra. Festuca microstachys was not used in the winter experiment due to insufficient seed availability. Ten seeds of a single native species were planted in 10-cm pots in a replicated block design, with each block containing each native species and each soil type (n=10). Each pot received the same amount of water at each watering, approximately 18 to 25 mL every two to three days. We scored germination on all seeds over a 60-day germination period and calculated total germination over the study period.     Data analyses - To determine whether water soluble or volatile compounds of E. globulus inhibited germination, we used factorial generalized linear models (GLMs, logit link function, binomial distribution) in the R base stats package (R Core Team, 2018).  Two separate models were performed, one for the volatile experiment and one for the water-soluble experiment.  For each model, the proportion of seeds germinated out of total seeds per dish was the response variable, while treatment, species, and the interaction of treatment and species were fixed-effect predictors. Tukey tests were carried out to compare germination across all pairwise treatment comparisons (glht function in the R package multcomp, Hothorn et al., 2008).  Given a significant interaction of treatment and species, pre-planned independent contrasts were used to determine all pairwise treatment comparisons for each species separately for a total of 30 independent contrasts (glht function in the R package multcomp, Hothorn et al., 2008). These contrasts compare the effect of each treatment/control to all other treatments/controls within a species, allowing us to compare any observed inhibition in E. globulus to that of the control (air/water), Q. agrifolia (native tree), and S. apiana (positive control). Due to the possibility of false discovery from multiple comparisons, we employed a Benjamini–Hochberg correction when determining significance of pre-planned contrasts (Benjamini & Hochberg, 1995).     To determine whether soil collected from under E. globulus inhibits germination, we again used a factorial generalized linear model with a logit link function, and binomial distribution.  The proportion of germinated out of total seed per pot was the response variable, while soil type (coastal scrub, Eucalyptus), season (winter, spring), species, site, and all two and three way interactions between the soil, season, and species were included as fixed-effect predictors. Given a significant interaction of soil type and species, pre-planned independent contrasts were again used to compare germination by soil type for each species separately for a total of five independent contrasts.     Finally, to determine whether soil collected from under E. globulus reduces above ground biomass following germination, we used a factorial generalized linear model with a natural log link function, gamma distribution. A gamma distribution is suitable for left-skewed response variables such as plant biomass. The total biomass for each pot was the response variable, while soil (coastal scrub, Eucalyptus), species, site, and the interaction between soil and species were included as fixed-effect predictors.

See R Script for annotations.