Interactions with soil biota shift from negative to positive when a tree species is moved outside its native range



  • Studies evaluating plant–soil biota interactions in both native and introduced plant ranges are rare, and thus far have lacked robust experimental designs to account for several potential confounding factors.
  • Here, we investigated the effects of soil biota on growth of Pinus contorta, which has been introduced from Canada to Sweden. Using Swedish and Canadian soils, we conducted two glasshouse experiments. The first experiment utilized unsterilized soil from each country, with a full-factorial cross of soil origin, tree provenance, and fertilizer addition. The second experiment utilized gamma-irradiated sterile soil from each country, with a full-factorial cross of soil origin, soil biota inoculation treatments, tree provenance, and fertilizer addition.
  • The first experiment showed higher seedling growth on Swedish soil relative to Canadian soil. The second experiment showed this effect was due to differences in soil biotic communities between the two countries, and occurred independently of all other experimental factors.
  • Our results provide strong evidence that plant interactions with soil biota can shift from negative to positive following introduction to a new region, and are relevant for understanding the success of some exotic forest plantations, and invasive and range-expanding native species.


Plant species are frequently moved from their native ranges into new regions, and sometimes they achieve higher growth rates, biomasses per individual, or higher densities, in their introduced ranges relative to in their native range (Shea & Chesson, 2002; Leger & Rice, 2003; Hawkes, 2007). Because introduced species can be both desirable (e.g. by being useful in forestry and agriculture) and undesirable (e.g. by operating as biological invaders), there is substantial interest in understanding underlying mechanisms that explain why their performance sometimes changes when they grow in their new range (Hierro et al., 2005; Gurevitch et al., 2011; Kueffer et al., 2013). Numerous factors have been proposed to explain enhanced plant performance in introduced ranges, such as improved climate or edaphic conditions, changes in the strength of biotic interactions, and genetic differences between the native and introduced populations (Hierro et al., 2005; Gurevitch et al., 2011; Kueffer et al., 2013). To date, experiments explicitly comparing plant performance in native vs introduced ranges (hereafter referred to as ‘home-away comparisons’) remain relatively rare, primarily because of logistical obstacles in making these types of comparisons. When home-away comparisons are made, substantial uncertainty frequently remains regarding the relative importance of drivers leading to enhanced growth in the plant's introduced range because multiple mechanisms are usually not considered simultaneously (Maron et al., 2013).

One mechanism that has gained substantial attention in the plant invasion ecology literature during the past decade is the role that interactions with soil biota may have in controlling the success of introduced plant species (Klironomos, 2002; Reinhart & Callaway, 2006; Pringle et al., 2009). Soils contain an immense diversity of organisms, some of which can adversely affect plant performance (e.g. pathogens and root herbivores), and some that enhance plant performance (e.g. mycorrhizas). Several studies have shown that the net effect of soil biota on some well-known invasive plant species are positive or less negative relative to co-existing native plant species, likely contributing to the success of those invaders (Agrawal et al., 2005; van der Putten et al., 2007; Engelkes et al., 2008; Kulmatiski et al., 2008). Research has also shown that the absence of key mutualisms from the native range of a plant species (e.g. mycorrhizal fungi or N2-fixing bacteria) can limit plant performance in introduced ranges (Nunez et al., 2009; Wandrag et al., 2013). A very limited number of studies have explicitly compared plant species interactions with soil biota in both their native and introduced ranges. The few studies making this comparison have suggested that differential interactions with soil biota in native and introduced ranges may contribute to the success or failure of some introduced species (Reinhart et al., 2003, 2010; Callaway et al., 2011a).

While these studies highlight the likely importance of soil biota in determining the success of introduced plant species, home-away comparisons that have focused on soil biota have rarely addressed a variety of additional factors that may also differentially influence plant growth in their native vs home ranges (te Beest et al., 2009). Notably, several hypotheses seeking to explain the success of plant invaders have focused on genetic differences in their introduced vs native range (e.g. founder effects and the ‘evolution of increased competitive ability’ hypothesis (EICA); Mayr, 1970; Blossey & Nötzold, 1995), which may contribute to altered plant performance in their new range (e.g. Leger & Rice, 2003). Additionally, several hypotheses have highlighted the roles that differences between native and introduced ranges in climate (Broennimann et al., 2007) or soil nutrient availability (and various factors that influence it) (Davis et al., 2000; Levine, 2000) may have on the performance of introduced plant species. These additional factors may not only confound the interpretation of home-away comparisons focused on soil biota, but could also interact with soil biota to control plant growth (Stamp, 2003; Johnson, 2010; Gurevitch et al., 2011; Gundale et al., in press). For instance, several studies have shown that for some species, different populations can interact differently with soil biota (Ramos-Zapata et al., 2010; Reinhart et al., 2011). Likewise, both pathogenic and mutualistic interactions with plants are known to vary in intensity as a result of plant nutrient status (Stamp, 2003; Johnson, 2010), suggesting that variability in soil fertility within or across native and introduced ranges may determine the net effect that soil biota has on plant performance. No experiment evaluating interactions between plants and soil biota across native and introduced ranges has yet explicitly controlled for differences in abiotic factors (i.e. climate or soil physical properties) and population genetic differences between native and introduced ranges. Further, no home-away comparison study to date has explicitly investigated whether effects of soil biota vary across different levels of soil fertility or for more than one introduced population of the same species.

Using a widely introduced and invasive tree species, Pinus contorta (McIntosh et al., 2012; Gundale et al., in press), we conducted two glasshouse experiments in a single location (i.e. where climate could be held constant) to understand how differences in soil abiotic and biotic properties in its native range (in Canada) and introduced range (in Sweden) influenced its growth; and further, whether the effects of soil biota on its growth differed among different introduced plant populations (i.e. different native range provenances) and across levels of soil nutrient availability. Because P. contorta is known to achieve higher growth rates in northern Sweden relative to in its native range (Elfving et al., 2001; McIntosh et al., 2012; Gundale et al., in press), we hypothesized that: seedlings would perform better in Swedish soil compared to Canadian soil; enhanced performance on Swedish soil would, in part, be due to differential effects of Swedish and Canadian soil biota; and these differential effects of soil biota across native and introduced ranges would vary among different tree provenances and across different levels of soil fertility. Testing these hypotheses in combination provided a robust home-away comparison of a plant species interaction with soil biota, while simultaneously enabling the evaluation of several potential interacting or confounding factors that have not been considered simultaneously in previous home-away comparison studies.

Materials and Methods

Field sites

The study was conducted using soils collected from 16 experimental Pinus contorta Douglas ex Loudon plantations in Sweden (between 62°04′ N and 65°50′N), and four P. contorta sites in British Columbia, Canada (between 57°40′N and 59°40′N). Each Canadian site was the seed source (i.e. provenance) for four of the Swedish plantations, which were established between 1970 and 1972.

In August 2012, all Swedish and all Canadian sites were visited in parallel by two separate field crews, who collected 10 cm depth soil cores from each site using a 110 diameter PVC coring tool. Nine cores were collected from each of the 16 Swedish stands, and 36 cores were collected from each of the four Canadian stands, from 5 m spaced sampling points within each stand (< 1 m from individual trees, stand size > 10 ha). The cores were placed in PVC tubes, wrapped with cellophane, and placed in identical coolers containing ice. After sample collection was complete, Canadian soils were express shipped to our laboratory facility in Umeå, Sweden. Swedish soils were delivered by car to the laboratory on the same day that the Canadian soils were shipped, and were then left at room temperature until the Canadian soils arrived. As such, Canadian and Swedish soils were handled similarly before the experiment started. Once all soils were present, each core was sieved (2 mm), and placed in a polyethylene bag. Swedish and Canadian soils were then paired by provenance, so that for each tree provenance eight of the nine samples from each of the four Swedish plantations were randomly paired with 32 of the 36 samples collected from their corresponding Canadian site (i.e. resulting in 32 paired Swedish and Canadian soil samples for each of the four tree provenances, 256 samples in total). Three quarters of these paired samples (i.e. 24 of 32 pairs for each provenance, or 192 samples in total) were randomly selected and gamma-irradiated (40 kGy; Synergy Health, UK) to sterilize the soil, while the remaining unsterilized samples were refrigerated (i.e. eight of 32 pairs for each provenance location, 64 samples in total). Gamma-irradiation effectively sterilizes soil, while causing minimal changes to soil physical properties (McNamara et al., 2003). The remaining unsterilized soil samples that were not paired (i.e. one of nine and four of 36 samples from each Swedish and Canadian site, respectively) were sieved and combined to create a Swedish and Canadian soil sample containing representative soil biota from each region, for use as inoculum, exactly as done in numerous other soil inoculation studies (e.g. Kardol et al., 2007; Ayres et al., 2009; Rodriguez-Echeverria et al., 2009). Nutrient availability (NH4+, NO3+, PO4) and pH was analyzed from a composite sample from each stand using standard protocols (Gundale et al., 2011); however only pH and NH4+ differed between countries, and thus are reported. We harvested seeds from each of the four Canadian tree provenances by collecting mature 2-yr old cones from south-facing exposed branches of five different trees in each of the Swedish plantations (i.e. 20 trees in total for each provenance) during late September, 2012. The cones were then opened with the use of heat, and the seeds were harvested.

Two glasshouse experiments were conducted utilizing the paired Canadian and Swedish soils. For both experiments, soil samples derived from individual soil cores served as the independent experimental unit, with each sample being placed in a single pot. Experiment one, which addressed our first hypothesis, was conducted using the unsterilized soil pairs, that is, 32 paired Swedish and Canadian soils; 64 pots in total. The experiment included three fully crossed factors, soil origin (Canada vs Sweden), tree provenance (seeds derived from the four different provenances introduced to Sweden), and fertilizer (unfertilized control or fertilized weekly with a cumulative total equivalent to 50 kg ha−1 of nitrogen and phosphorous, as NH4NO3 and Na2HPO4), and using four replicates of each (i.e. 2 × 4 × 2 × 4 replicates = 64 pots). For the Canadian soils, the replicates consisted of samples from each of the four Canadian stands, and for the Swedish soils the replicates consisted of randomly paired samples from one of four stands planted with the corresponding tree provenance. Experiment two, which addressed our second and third hypotheses, was conducted using gamma-irradiated sterile soil from both Sweden and Canada, that is, 96 paired Swedish and Canadian soils, or 24 pairs corresponding to each set of provenance locations; 192 pots in total), and involved a factorial cross of soil origin (Canada or Sweden), soil biota treatments (sterile control, or sterilized and inoculated with unsterile Swedish or Canadian soil), fertilizer (with or without fertilizer), and tree provenance (four different tree provenances), using four replicates (i.e. 2 × 3 × 2 × 4 × 4 replicates = 192 pots); stands served as replicates, as described in the first experiment.

For each experiment, each sample was combined with sand (50:50 mixture by volume), and an 800 ml total volume of this mixture was then placed in a 1 l pot. The sand was first autoclaved to prevent accidental introduction of soil biota. For the inoculation treatments, 10% of the soil volume was replaced with nonsterilized soil before mixture with sand, meaning 5% of the final pot volume consisted of unsterile soil (Kardol et al., 2007; Ayres et al., 2009). For both experiments, pots were randomized on glasshouse tables with an 18 h : 6 h, light:dark photoperiod, and were planted with one of the four tree provenances described earlier. The seeds were germinated in Petri plates, and two seeds were placed on the soil surface of each pot 2 d after germination. After 2 wk, the smallest seedling in each pot was removed. The glasshouse experiment ran for five months, during which pots were watered every 3 d, which prevented soils from becoming excessively dry. The glasshouse atmosphere was maintained at 75% relative humidity, and day and night time temperature was set at 25 and 15°C, respectively. At the end of the experiments, we observed 100% seedling survival. We destructively harvested the experiment by washing soils from roots, oven-drying the plants (65°C, 48 h) and recording aboveground, belowground, and total biomass of each seedling.

Statistical analysis

For experiment one, data were analyzed using a three-factor ANOVA, with soil origin, tree provenance, and fertilizer serving as fixed factors. For experiment two, data were also analyzed with a four-factor ANOVA, with soil origin, tree provenance, soil biota treatment, and fertilizer treatment serving as fixed factors. For all analyses, data were log transformed when parametric assumptions were not met. When significant differences occurred for factors with more than two levels, or for significant interactions, post hoc Student–Newman–Keuls tests were used to show pairwise differences at P = 0.05. All data analyses were done using SPSS version 19.0 (SPSS Inc., Armonk, NY, USA).

Results and Discussion

Our study compared the growth of Pinus contorta on soils of its native and introduced range, in a common glasshouse location where climate could be held constant. In doing so, our experiments provide strong evidence that growth of P. contorta was greater when grown on soil throughout its introduced range in Sweden compared to locations where it originated in northern BC, Canada, and that interactions with soil biota served as the main driver of this response.

In support of our first hypothesis, experiment one showed that when seedlings were grown on nonsterilized Swedish soil, they exhibited more aboveground relative to belowground biomass and gained c. 43% more total biomass, relative to when they were grown on nonsterilized Canadian soil (Table 1; Fig. 1). The differential response to soil origin was almost half as large as the growth response elicited by the fertilizer treatment, which caused a two-fold growth increase (Table 1; Supporting Information Fig. S1). Higher growth on Swedish soil occurred despite significantly lower mean soil pH and extractable NH4+ concentrations across the Swedish sites relative to the Canadian sites (Supporting Information Fig. S2), suggesting that differences in soil fertility cannot explain this response. Several studies have described that plants often perform better in their introduced ranges relative to their native ranges (Leger & Rice, 2003; Reinhart et al., 2003; Hawkes, 2007; Callaway et al., 2011b), and for Pinus contorta it suggests that plant–soil interactions likely contribute to its enhanced growth in its introduced ranges.

Table 1.  Analysis of variance comparing the main and interactive effects of soil origin (SO), tree provenance (TP), and fertilization (F) on the total, aboveground, belowground biomass, and the ratio of aboveground to belowground biomass of Pinus contorta seedlings grown in unsterilized soil
 DFaTotal massbAboveground massbBelowground massbAbove : Belowc
  1. The soil originated from either Canada or Sweden. Bold F- and P-values are significant at an alpha value of 0.05.

  2. a

    Numerator degrees of freedom. Error degrees of freedom were 48 for each factor.

  3. b

    Data were transformed (log (x + 1)) before analysis.

  4. c

    The ratio of aboveground to belowground biomass.

Soil origin (SO) 1 7.874 0.007 8.319 0.006 3.8390.056 11.899 0.001
Tree provenance (TP)30.6570.5830.4560.7150.7480.5290.4720.703
Fertilization (F) 1 21.465 <0.001 23.469 <0.001 14.716 <0.001 11.923 0.001
SO × TP30.6120.6100.5660.6400.8780.4590.2650.851
SO × F10.0010.9760.0020.9640.0010.9710.4210.520
TP × F30.0930.9640.0790.9710.1900.9030.1700.916
SO × TP × F30.4020.7520.4180.7410.5490.6511.8170.157
Figure 1.

The mean (± standard error (SE)) total biomass (a), aboveground biomass (b), belowground biomass (c), and the ratio of aboveground to belowground biomass (i.e. Above : Below) (d) of Pinus contorta grown on unsterilized Swedish or Canadian soils. Asterisks above each pair of bars indicate significant main effects as indicated by ANOVA in Table 1 (**, P < 0.01; ***, P < 0.001).

Our second hypothesis, which predicted that soil biota effects would underlie differences in seedling growth between Swedish and Canadian soil, was strongly supported by experiment two. When soils were sterilized, soil origin (i.e. Canada or Sweden) was not a significant factor explaining seedling growth (Table 2). This result contrasts sharply with the highly significant effect of soil origin in the first glasshouse experiment, where unsterilized soil was used (Table 1). Additionally, the reciprocal inoculation component of the second glasshouse experiment provided compelling evidence that soil biota were responsible for the higher growth in Swedish soil relative to Canadian soil revealed in experiment one. When sterilized Canadian or Swedish soils were inoculated with Swedish soil biota, the ratio of aboveground to belowground mass was 25% greater and total biomass was 61% greater compared to when these soils were inoculated with Canadian soil biota (Table 2; Fig. 2). These results show that soil biota have a net negative effect on Pinus contorta growth in its native Canadian soils compared to a net positive interaction in its introduced Swedish range, independent of any differences of soil abiotic properties (e.g. nutrient availability or mineralization rates) that may have differed between the two region.

Table 2. Analysis of variance comparing the main and interactive effects of soil origin (SO), inoculum origin (IO), tree provenance (TP) and fertilization (F) on the total, aboveground, belowground biomass, and the ratio of aboveground to belowground biomass of Pinus contorta seedlings grown in sterilized soil
DFa Total massb Aboveground massb Belowground massb Above:Belowc
F-value P-value F-value P-value F-value P-value F-value P-value
  1. The soil originated from either Canada or Sweden and was sterilized via gamma irradiation. Inoculation origin consisted of three treatments, that is, a sterile control soil, or sterilized soil inoculated with either Canadian or Swedish soil biota. Bold F- and P-values are significant at an alpha value of 0.05.

  2. a

    Numerator degrees of freedom. Error degrees are 144 for all factors.

  3. b

    Data were transformed (log (x + 1)) before statistical analysis.

  4. c

    The ratio of aboveground to belowground biomass.

Soil origin (SO)10.8330.3630.3660.5460.8870.3480.0240.878
Inoculum origin (IO)2 12.080 <0.000 16.688 <0.001 6.431 0.002 20.354 <0.001
Tree provenance (TP)30.5200.6690.3970.7551.1050.3491.3250.269
Fertilization (F)1 18.902 <0.001 33.490 <0.001 3.1040.080 57.013 <0.001
SO × IO20.0150.9850.1470.8640.0170.9830.0370.963
SO × TP30.3040.8220.1990.8970.4950.6860.4160.742
SO × F10.4560.5010.0970.7551.2280.2700.2360.628
IO × TP60.1460.9900.1230.9930.4400.8510.9960.430
IO × F20.3920.6761.2510.2890.0940.910 4.427 0.014
TP × F30.2240.8800.4140.7430.0900.9660.0560.982
SO × IO × TP30.7900.5800.8210.5550.6430.6950.8490.534
SO × IO × F60.2400.7870.1000.9050.5660.5690.4880.615
SO × TP × F20.7050.5511.1520.3300.4450.7212.2030.090
IO × TP × F31.1930.3141.1650.3281.0730.3810.6140.719
SO × IO × TP × F60.5380.7790.4990.8080.6860.6610.3980.879
Figure 2.

The mean (± SE) total biomass (a), aboveground biomass (b), belowground biomass (c), and the ratio of aboveground to belowground mass (i.e. Above : Below) (d) of Pinus contorta grown on sterilized soil, and sterilized soil re-inoculated with Swedish or Canadian soil biota. Half of the soils within each treatment originated from Canada, and the other half from Sweden. Different letters (a, b or c) above bars indicate significant post hoc differences (Student–Newman–Keuls test) at P = 0.05 following ANOVA (Table 2).

Several mechanisms have been proposed to explain how soil biota can enhance plant growth in species' introduced ranges. Reinhart & Callaway (2006) proposed that improved growth of introduced species could emerge due to their escape from soil pathogens, or due to the establishment of novel positive associations in their new range. The results of experiment two provide some additional insight into the potential mechanisms underlying the observed higher growth of Pinus contorta seedlings in Swedish soil. Seedlings growing on soil inoculated with Canadian biota resulted in 21% lower total biomass relative to the noninoculated control soil (Fig. 2a), indicating a net negative effect of Canadian soil biota on seedling performance. The most plausible mechanism for this net negative effect is an over-riding influence of soil pathogens, which are frequently problematic in Canadian P. contorta stands (e.g. the parasitic fungi in the Armillaria genus; Krebill, 1973). In contrast to Canadian soil biota, inoculation with Swedish soil biota resulted in c. 27% higher growth relative to the noninoculated control soil. This pattern may be driven by the establishment of stronger or more positive mycorrhizal associations in Swedish soil (Reinhart & Callaway, 2006). In Sweden, P. contorta has been shown to associate with many of the same ectomycorrhizal fungal genera as the widespread native species P. sylvestris (Kardell et al., 1987; McIntosh et al., 2012). It is known that ectomycorrhizal fungi can have variable effects on host plants, ranging from strongly positive to parasitic, which can depend on the host species identity as well as the environmental context of where the association occurs (Bethlenfalvay et al., 1982; Johnson et al., 1997; Näsholm et al., 2013). It is therefore possible that P. contorta seedlings encounter ectomycorrhizal species or genotypes in Swedish soil that provide them with greater resources per unit of carbon exchanged compared to their native range (Johnson et al., 1997).

Our experiments provided partial support for our third hypothesis. The first part of this hypothesis predicted that responses to soil treatments would vary among the four introduced tree provenances, which was not supported in either the first or second glasshouse experiments (Tables 1, 2). In both of these experiments, significant main or interactive effects of tree provenance were never detected. The second part of our third hypothesis, which predicted that effects of soil biota on plant growth would vary across different levels of soil fertility, was weakly supported by our data. Experiment two showed that the strong effect of soil biota treatments on total plant biomass occurred independently of fertilization treatments (i.e. no interactive effect), with the exception that the aboveground to belowground biomass ratio increased in response to both inoculation treatments but only in the absence of fertilizers (Table 2; Supporting Information Fig. S3). These results indicate that seedling growth allocation between root and photosynthetic structures can be highly sensitive to the combination of soil fertility and soil biota, while total biomass only responds to these two factors separately. An additional insight gained from experiment two was that the positive effect of Swedish biota (i.e. 61% greater total biomass relative to the Canadian soil biota, Fig. 2a) was of much greater magnitude than the growth increase caused by the fertilization treatment (i.e. 34%; Fig. 2, Supporting Information Fig. S4), highlighting the importance that biogeographical differences in soil biota may have on Pinus contorta performance relative to soil fertility.

A key novelty of our study is the inclusion of native and home range soils, multiple plant provenances, and independent manipulation of soil biota and soil fertility in a common glasshouse environment where climate was held constant. No previous home-away comparison studies focused on soil biota have accounted for all of these potential factors at the same time, each of which can differ between native and introduced ranges (e.g. Reinhart et al., 2003; Callaway et al., 2011a; Maron et al., 2013; Wandrag et al., 2013; Yang et al., 2013). Our experiments unambiguously demonstrated that the net effect of soil biota on the performance of a plant species shifted from negative to positive when grown on soils in its introduced range, regardless of climate, the plant population considered, or differences in soil abiotic properties. Differential effects of native and home range soil biota on plant growth, as shown in our experiments, are potentially important in explaining several increasingly recognized biogeographical patterns. For example, tree species introduced for production forestry purposes sometimes achieve higher growth rates in the regions to which they are introduced (Richardson & Rejmanek, 2011; Yetti et al., 2011). Likewise, nonnative or range-expanding native invaders often exhibit higher growth rates or densities within communities relative to their historical ranges (Hawkes, 2007; Lamarque et al., 2012). While numerous factors may influence the success of nonnative or range expanding native species, our data provides convincing evidence that biogeographical differences in soil biotic communities may contribute to these patterns.


The project was supported by Swedish TC4F to M.J.G. The authors thank M. Sandström and A. Väppling for their assistance with the glasshouse experiment, and K. Gundale and Thomas Hörnlund for assistance with fieldwork. The authors also thank E. Walfridsson at Skogforsk for extracting and germinating seeds.