Birch (Betula spp.) leaves adsorb and re-release volatiles specific to neighbouring plants – a mechanism for associational herbivore resistance?


  • Sari J. Himanen,

    1. Department of Environmental Science, University of Eastern Finland, Kuopio Campus, PO Box 1627, FI-70211 Kuopio, Finland
    2. MTT Agrifood Research Finland, Plant Production Research, Lönnrotinkatu 5, FI-50100 Mikkeli, Finland
    Search for more papers by this author
  • James D. Blande,

    1. Department of Environmental Science, University of Eastern Finland, Kuopio Campus, PO Box 1627, FI-70211 Kuopio, Finland
    Search for more papers by this author
  • Tero Klemola,

    1. Section of Ecology, Department of Biology, University of Turku, FI-20014 Turku, Finland
    Search for more papers by this author
  • Juha Pulkkinen,

    1. Department of Biosciences, University of Eastern Finland, Kuopio Campus, PO Box 1627, FI-70211 Kuopio, Finland
    Search for more papers by this author
  • Juha Heijari,

    1. Department of Environmental Science, University of Eastern Finland, Kuopio Campus, PO Box 1627, FI-70211 Kuopio, Finland
    2. Kotka Maritime Research Centre, Mussalontie 428 B, FI-48310 Kotka, Finland
    Search for more papers by this author
  • Jarmo K. Holopainen

    1. Department of Environmental Science, University of Eastern Finland, Kuopio Campus, PO Box 1627, FI-70211 Kuopio, Finland
    Search for more papers by this author

Author for correspondence:
Sari J. Himanen
Tel. +358 40 738 9873


  • Plant-emitted semi-volatile compounds have low vaporization rates at 20–25°C and may therefore persist on surfaces such as plant foliage. The passive adsorption of arthropod-repellent semi-volatiles to neighbouring foliage could convey associational resistance, whereby a plant’s neighbours reduce damage caused by herbivores.
  • We found that birch (Betula spp.) leaves adsorb and re-release the specific arthropod-repelling C15 semi-volatiles ledene, ledol and palustrol produced by Rhododendron tomentosum when grown in mixed association in a field setup. In a natural habitat, a higher concentration of ledene was released from birches neighbouring R. tomentosum than from birches situated > 5 m from R. tomentosum. Emission of α-humulene, a sesquiterpene synthesized by both Betula pendula and R. tomentosum, was also increased in R. tomentosum-neighbouring B. pendula.
  • In assessments for associational resistance, we found that the polyphagous green leaf weevils (Polydrusus flavipes) and autumnal moth (Epirrita autumnata) larvae both preferred B. pendula to R. tomentosum. P. flavipes also preferred birch leaves not exposed to R. tomentosum to leaves from mixed associations. In the field, a reduction in Euceraphis betulae aphid density occurred in mixed associations.
  • Our results suggest that plant/tree species may be protected by semi-volatile compounds emitted by a more herbivore-resistant heterospecific neighbour.


Plants and trees live surrounded by individuals of the same species (conspecifics) and different species (heterospecifics). Co-occurring species may share their preferences for soil, light, nutrient and moisture conditions and benefit from the presence of each other through interactions such as mycorrhizal symbiosis (Urcelay et al., 2003). However, associational resistance, where a species experiences reduced herbivore attack through coexistence with other plant species (Karban, 2007), might be an important mediator of species abundance. Recent research has revealed that plants can sense their neighbours without direct physical contact; mechanisms for this include detection of a reduced ratio of red to far-red in light reflected from neighbouring foliage (Tao et al., 2008), or detection of plant-emitted volatile compounds (Baldwin et al., 2006; Heil & Karban, 2010). Plant volatiles have various ecological roles, including attraction and deterrence of herbivores (Baker et al., 1998; Bruce et al., 2005), while herbivore-induced emissions may attract natural enemies to suppress herbivore pressure in the process of indirect defence (Turlings et al., 1990; Dicke, 2009). In natural environments, it is evident that the volatile bouquets of neighbouring species become mixed. However, existing evidence that volatiles affect herbivore abundance on neighbouring plants (Karban, 2007) and behaviour of insects (De Moraes et al., 1998) in nature suggest that this does not result in an infochemical chaos, but that individual organisms can benefit from volatile cues.

Previous research on plant–plant signalling via airborne plant-emitted volatiles has revealed that induced volatiles are able to activate volatile compound release (Choh et al., 2004), defence gene expression (Arimura et al., 2000) and synthesis of phenolics, proteinase inhibitors (Tscharntke et al., 2001) and extrafloral nectar (Heil & Silva Bueno, 2007) in signal-receiving plants. It has recently been shown that exposure to volatiles can prime plants to have faster defence responses upon subsequent herbivore attack (Kessler et al., 2006; Heil & Silva Bueno, 2007). However, passive adsorption of volatiles emitted by neighbouring plants might also be an important, but understudied, way for volatiles to act in natural environments. Choh et al. (2004) found that volatiles can be adsorbed to and re-released from conspecific plants. To show this, they treated Lima bean (Phaseolus lunatus) leaves with a protein-synthesis inhibitor to suppress production of herbivore-induced volatiles. These leaves were then exposed to spider mite (Tetranychus urticae)-infested Lima bean plants, and were subsequently shown to emit volatiles associated with herbivore damage. As the receiver leaves were unable to synthesize these volatiles, a passive adsorption and release process was concluded. In the laboratory, exogenous monoterpene fumigation resulted in nonemitting oak (Quercus suber) leaves subsequently releasing monoterpenes for more than 12 h (Delfine et al., 2000). However, passive adsorption and re-release of volatiles has not been demonstrated interspecifically under field conditions.

Neighbour-emitted compounds could be ecologically beneficial to the plant to which they adsorb, if they either deter herbivores (Karban, 2007) or attract herbivore natural enemies. For the emitting plant, this would probably present a neutral situation: it will emit the same volatiles despite its neighbour. However, it might also have an impact on competitive dynamics, for example, if there were phytotoxic effects on receiver plant foliage (Barney et al., 2005). Improving the performance of a neighbour’s herbivores by attracting them in higher numbers or by attenuating the foraging efficiency of natural enemies of a neighbour’s herbivores might lead to negative effects of neighbouring. The fitness of the emitting plant may even be altered, for example, by reduced root and light competition. Benefits or costs of coexistence mediated by volatile communication are multifaceted (Heil & Karban, 2010). The importance of volatile-mediated plant–herbivore interactions is ultimately dependent on the amount of herbivore pressure and whether it is significant enough in the ecosystem to affect plant fitness (Karban, 2007).

Rhododendron tomentosum (Stokes) H. Harmaja (formerly known as Ledum palustre L.) is an abundant perennial evergreen shrub distributed throughout boreal ecosystems (Urcelay et al., 2003). R. tomentosum has a characteristic aromatic profile as a result of its high volatile terpenoid content (Butkiene et al., 2008). These compounds have both medicinal (Harbilas et al., 2009) and arthropod-repellent properties (Jaenson et al., 2005, 2006). The specific volatiles emitted by R. tomentosum are the sesquiterpene alcohols palustrol (C15H26O) and ledol (C15H26O) and the sesquiterpene ledene (C15H24) (Jaenson et al., 2005). In addition, it emits the monoterpene β-myrcene (C10H16) as a major compound (Jaenson et al., 2005, 2006; Butkiene et al., 2008). Betula spp. (Betulaceae) trees commonly coexist with R. tomentosum in natural ecosystems (Urcelay et al., 2003). Reanalysis of chromatograms from our earlier study of mountain birch (Betula pubescens ssp. czerepanovii (Orlova) Hämet-Ahti) leaf volatiles (Mäntyläet al., 2008) indicated traces of ledol and palustrol, when R. tomentosum had been growing nearby. The emission of these specific ‘sticky’ volatiles makes R. tomentosum an excellent species for studying how passive adsorption of volatiles might contribute to associational resistance and neighbouring plant interactions. Oxygenated sesquiterpenoids are particularly resistant to degradation by ozone in the atmosphere (Pinto et al., 2007), have semi-volatile characteristics (boiling point at 240–400°C and low vaporization rate to gaseous phase at 20–25°C (Hübner & Röpcke, 2009)) and could therefore persist on surfaces such as leaf foliage (Helmig et al., 2004).

We established a field experiment with potted silver birch (Betula pendula Roth) growing either alone or in identical plots with R. tomentosum as neighbours, that is, exposed to airborne volatiles but isolated below ground. We first tested whether the semi-volatile sesquiterpenoids or other volatiles released from R. tomentosum were adhered to and subsequently re-released from B. pendula leaves. Ecologically, this would mean a change in the volatile profile experienced by herbivores attacking B. pendula when it coexists with R. tomentosum. Additionally, we collected volatiles in a natural peatland ecosystem, where downy birch (Betula pubescens Ehrh.) and R. tomentosum coexisted.

Betula species host numerous herbivores from different feeding guilds (Koponen, 1983; Atkinson, 1992). To test for associational resistance in B. pendula neighbouring R. tomentosum, we screened for the attractiveness of R. tomentosum-neighbouring birch leaves to three species of generalist herbivores, the green leaf weevil Polydrusus flavipes Degeer (Coleoptera: Curculionidae), the autumnal moth Epirrita autumnata (Borkhausen) (Lepidoptera: Geometridae) and the birch aphid Euceraphis betulae Koch. (Homoptera: Aphididae). E. autumnata larvae are major pests of birch trees in boreal and subarctic ecosystems and may appear in extensive outbreaks that can result in total destruction of birch foliage (Haukioja et al., 1988). Leaf weevils can be a serious problem for birch forest plantations. In the field experiment, we also screened for season-long natural abundance of other herbivores in R. tomentosum associated and control B. pendula plots.

We report passive adsorption and re-release of specific R. tomentosum-emitted volatiles by birch leaves in field and laboratory experiments, as well as in observations in a natural habitat. Our results show a volatile-mediated ecologically meaningful plant–tree interaction, which might benefit the receiver without its own input contribution. Associational resistance in Betula spp. through coexistence with R. tomentosum is suggested based on certain herbivore-repellent effects observed for R. tomentosum and R. tomentosum volatile-exposed birches.

Materials and Methods

Our study consists of a field experiment to assess whether R. tomentosum neighbouring alters the volatile emissions from B. pendula and the abundance of its natural herbivores. In a natural habitat, we aimed to verify whether adsorbance of R. tomentosum-emitted volatiles occurs in nature and the extent of its distance dependency in B. pubescens. In a laboratory setup, we tested the attractiveness or repellency of R. tomentosum-emitted and R. tomentosum-originating foliage-adsorbed volatiles for selected birch herbivores.

Site descriptions

Field experiment Rhododendron tomentosum plantlets were collected from a natural habitat (Kuopio 62°56′48′′N, 27°23′32′′ E) between 8 and 12 May 2009, transplanted into 5 l pots filled with peat soil and kept outdoors in half-shade before transferral to experiment plots in the field. Four independent replicate plots (established on wooden platforms 3.0 × 1.2 m) were located in the same field site used for earlier ultraviolet (UV) radiation exposures (Turtola et al., 2006; Blande et al., 2009) at Ruohoniemi, Kuopio. Four platforms were used for controls and four platforms for R. tomentosum-exposure treatments (a total of eight plots in use with four separate platforms used for volatile collection). The experiment was established on 25 May 2009; in each plot, six potted 2-yr-old B. pendula seedlings (seed origin, Hausjärvi, Finland, 60°48′N, 24°01′E) were placed in two rows (three plants in each row), with an additional seedling placed at each end of the rows to ensure equal shading. In R. tomentosum mixed association plots, six potted R. tomentosum plants were placed next to the six experimental birch seedlings, c. 15–20 cm away from the birches. In control plots, an equal number of B. pendula seedlings were placed at identical positions to form conspecific association plots with an identical shading and growth environment. The Ruohoniemi experimental field site was a meadow established on a former field in 1999 (Karnosky et al., 2007). The vegetation surrounding the site was dominated by Graminaceae species, but various herbaceous plants and sedges were also present.

Natural habitat  The natural habitat site, located near Kuopio, Finland (62°47′07′′N, 27°35′44′′E), was a ditched Sphagnum peat bog site with young Salix spp. and B. pubescens saplings. The understorey vegetation of the site was dominated by Vaccinium vitis-idaea L., V. oxycoccos L. and Chamaedaphne calyculata (L.) Moench with aggregations of R. tomentosum.

Volatile collection and analysis

Field experiment collections  A portable volatile collection system (Mäntyläet al., 2008; Blande et al., 2009) with 35 × 43 cm PET (polyethylene) bags (Look Isopussi, Euracon Oy, Finland) for plant enclosure (heated at 120°C before use) was used. Volatiles were collected in the field experiment on three consecutive days, 8–10 June 2009. Before collection, B. pendula seedlings were transferred to a collection table located at the field site, and the first collection was started after c. 10 min. Two collections (the first in the morning 10 min after removing the seedling from the mixed or conspecific association and the second in the afternoon, c. 4 h later) were made from each seedling. Collections from control and R. tomentosum-exposure plots were made simultaneously with a total of eight seedlings sampled daily (one seedling per plot on each day). This resulted in a total of three collections per plot (i.e. 12 collections per treatment), plus an equal number of 4 h postexposure collections. Volatile emissions from R. tomentosum in the mixed associations were collected on each of the tree days (one plant per plot, total 12 collections).

Natural habitat collections  In the natural habitat, collections were made on 22 June 2009. Six experimental plots (a minimum of 5 m apart), where B. pubescens and R. tomentosum were growing closer than 0.5 m from each other, were randomly selected. Volatiles were collected from one B. pubescens branch and one R. tomentosum branch per site, separately. Control samples were collected from six B. pubescens plots where the distance to the closest R. tomentosum was more than 5 m.

Volatile analysis  Volatiles were collected at 0.5 l min−1 for 30 min from the Betula spp. and for 15 min from R. tomentosum, and run using a GC-MS instrument as in Blande et al. (2009). Emitted compounds were identified and quantified using synthetic standard compounds (Blande et al., 2009; Himanen et al., 2009) and verified with the Wiley mass spectra library. Palustrol and ledene were not available, so a ledol standard (ChemDiv Inc., Del Mar, CA, USA) was used to calculate their emissions. Emissions were expressed in ng g–1 leaf DW h–1 (leaves dried at 60°C for a minimum of 2 d) to allow comparison with most earlier publications on ecological interactions of volatiles using this unit. However, to allow chemical reaction comparisons, emissions from receiver foliage are shown as pmol per seedling h−1 (B. pendula, field experiment) or pmol per branch min−1 (B. pubescens, natural habitat) in Supporting Information, Figs S1 and S2.

Herbivore assays

Epirrita autumnata larvae were obtained from the University of Turku as eggs and kept at +4°C before use. The original parents had been collected as larvae from the field during the previous summer. The eggs were hatched at room temperature (20–25°C) and in natural light, and maintained on young B. pendula leaflets. Second instar larvae (5–8 mm) were used in the experiments. Green leaf weevils (P. flavipes) and birch aphids (E. betulae) were collected from birch trees in the Ruohoniemi field site and in the Kuopio Campus area, and maintained on B. pendula leaves at 4°C before testing.

Separate sets of choice tests were conducted with the herbivores to compare the attractiveness of control vs R. tomentosum-exposed B. pendula leaves. Leaves were collected from 4-yr-old B. pendula saplings of mixed clones grown for the last 2 yr in 7 l pots at the Ruohoniemi field site. Ten 1.5 l glass containers were used as exposure systems in which each had two 30 ml water-filled glass vials with three B. pendula twigs (c. 10 cm long) in one vial and either three identical B. pendula twigs (control) or three field-collected R. tomentosum twigs (identical in size) in the other vial. Vials were placed c. 3 cm apart and without physical contact between leaves. Exposure systems had no air flow, but two air outlets (7 mm diameter) were present at the top of each container. Systems were placed in a fume hood at room temperature for 24 h before herbivore assays. For choice tests, one leaf from each twig was detached with forceps and the petioles of leaves were inserted in water-filled 1.5 ml Eppendorf tubes through a hole in the cap. Two leaves (control and R. tomentosum-exposed) were then placed in 13-cm-diameter plastic Petri dishes lined with filter paper (30 replicates). Herbivores (one P. flavipes adult, five E. autumnata larvae or five E. betulae adults) were immediately released in the middle of each dish. The location of the herbivores on either leaf was recorded at 30 min, 1 and 2 h from release. In addition, the ability of E. autumnata to choose between B. pendula or R. tomentosum was tested (performed as the other choice tests, but using five R. tomentosum leaves and one B. pendula leaf).

The ability of P. flavipes to use volatiles in host plant finding was tested in a Y-tube olfactometer system (described in Himanen et al., 2009). P. flavipes were offered the choice of arms containing three B. pendula twigs or a blank. The repellency of R. tomentosum was also tested by comparing responses to an arm containing three R. tomentosum twigs vs a blank. Thirty-six P. flavipes adults were tested for each combination with twigs changed after every six herbivores tested.

Natural herbivore abundance assessments in the field experiment

The natural abundance of herbivores was monitored by calculating the number of leaves infested by different species (different aphid species, sawflies, mined leaves, leafrolls) in the field plots (observed from three seedlings per plot) on the following days: 23 July (t1), 4 August (t2), 18 August (t3) and 5 September 2009 (t4). At the same time the total number of leaves per seedling and the number of yellow leaves (an indicator of accelerated leaf senescence from potential phytotoxicity of a neighbour’s volatiles (Barney et al., 2005)) and herbivore-damaged leaves were calculated.

Statistical analysis

Before analysis, normality and equality of error variances were tested and some variables log(x + 1)-transformed for normality. Percentage values were also arcsin(x + 1)-transformed before analysis. Volatile emission, herbivore infestation and total number of leaves and yellowing leaves recorded in the field experiment were tested with linear mixed model analysis, with treatment (conspecific or R. tomentosum mixed association) as a fixed factor and experimental plot (and measurement day for volatiles) as random factors. For volatile compound emissions below the detection limit, the nonparametric Mann–Whitney U-test was used instead of a parametric test. Natural habitat volatile emission results were tested with one-way analysis of variance. Herbivore choice tests were analysed with the χ2-test. All tests were two-tailed and performed with SPSS 14.0 for Windows (SPSS Inc., Chicago, IL, USA).


Volatile emissions

The monoterpene β-myrcene and the sesquiterpene alcohols palustrol and ledol were the main volatile compounds emitted by R. tomentosum plants in both a natural habitat and in the field experiment (Table 1). Control B. pendula emitted various mono- and sesquiterpenes (Fig. 1). When grown in a mixed community with R. tomentosum, B. pendula scent included palustrol, ledol and ledene, released in significantly increased amounts (Mann–Whitney = 0, <0.001; = 12.0, < 0.001; and = 18.0, = 0.001, respectively) when sampled 10 min after removing the plants from treatments (Fig. 1a). Emissions of palustrol and ledene from these seedlings were still detectable 4 h after removal from a mixed community at concentrations significantly above (Mann–Whitney = 0, P < 0.001 and U = 36.0, = 0.039, respectively) control seedlings in which these compounds were absent (Fig. 1b). Emissions of palustrol in R. tomentosum-neighbouring B. pendula seedlings were reduced 4 h postexposure to 25.4%, in ledol to 16.4% and in ledene to 11.6% from the concentrations detected 10 min postexposure. Higher (linear mixed model, F1,22 = 5.3, = 0.031) amounts of α-humulene, a sesquiterpene synthesized by both B. pendula and R. tomentosum, were collected from R. tomentosum-neighbouring B. pendula seedlings than control seedlings at 10 min postexposure (Fig. 1a). The dominating monoterpene of R. tomentosum, β-myrcene, was not increased in R. tomentosum-neighbouring B. pendula emissions.

Table 1.   olatile compounds emitted by Rhododendron tomentosum genotypes in natural habitat and as potted seedlings in the field experiment Thumbnail image of
Figure 1.

 Volatile emissions from Betula pendula seedlings grown in conspecifics association (control, open bars) and Rhododendron tomentosum mixed association (R. tomentosum-exposed, closed bars) in the experimental field plots. Collections (30 min) were started 10 min (a) and 4 h (b) after removing the plants from the treatments (n = 12). Means ± SE and statistically significant differences (*, P < 0.05; ***, P < 0.001, linear mixed model or Mann–Whitney U-test) are shown.

Numerous mono- and sesquiterpenes were detected from B. pubescens growing in the natural habitat (Fig. 2). Ledol, palustrol and ledene were found in emissions of B. pubescens with R. tomentosum as a neighbour, and also in low concentrations in control samples (> 5 m from closest R. tomentosum). Ledene was the only compound emitted in significantly different amounts by control and R. tomentosum-exposed plants (analysis of variance, F1,10 = 5.3, = 0.044).

Figure 2.

 Volatile emissions of major (a) and minor (b) compounds from Betula pubescens growing in a natural habitat, with Rhododendron tomentosum growing at < 0.5 m (R. tomentosum as neighbour, closed bars) or > 5 m (control, open bars) from the target plant (n = 6). Means ± SE and statistically significant differences (*, P < 0.05, ANOVA) are shown.

Herbivore assays

We first checked that the volatile adsorption ability of detached birch leaves, which had to be used in the laboratory based herbivore assays, resembled that of intact seedlings. Emissions of B. pendula branches collected from the field site and exposed to R. tomentosum in the same way as the twigs used in the herbivore assays, and B. pubescens branches collected from the natural habitat (R. tomentosum as neighbours) were measured in the laboratory. The emission rates (mean ± SEM, as ng g−1 DW h−1, n = 3 for B. pendula and n = 5 for B. pubescens) were 178.5 ± 23.1 (B. pendula) and 80.2 ± 21.0 (B. pubescens) of palustrol, 56.2 ± 7.4 (B. pendula) and 25.2 ± 3.7 (B. pubescens) of ledol and below detection limit (B. pendula) and 39.3 ± 8.9 (B. pubescens) of ledene. None of these compounds was detected in emissions 48 h later (results not shown).

In the Y-tube olfactometer tests, 81.8% of P. flavipes adults preferred the blank arm to volatiles from R. tomentosum (χ2 = 8.91, = 0.004), whereas 81.3% chose the B. pendula arm over the blank arm (χ2 = 6.25, = 0.021; Fig. 3). In the choice test with R. tomentosum-exposed B. pendula vs control B. pendula, 86.7% of P. flavipes adults were positioned on control leaves 2 h after release (χ2 = 8.07, = 0.007; Fig. 4a), with a similar trend observed in recordings after 30 min (χ2 = 2.25, = 0.210) and 1 h (χ2 = 1.5, = 0.307) (Fig. 4a). No preference for control or R. tomentosum-exposed B. pendula leaves was observed for E. autumnata larvae (χ2 < 2.29 and > 0.18; Fig. 4b), but the larvae did prefer B. pendula over R. tomentosum leaves from 30 min after release until 2 h (χ2 = 25.8 at 30 min, χ2 = 37.2 at 1 h and χ2 = 52.6 at 2 h, P < 0.001 for all time points; Fig. 4c). No preference for R. tomentosum-exposed vs control leaves was observed for winged E. betulae aphid adults at any time point (χ2 < 0.78, > 0.46 for all time points; Fig. 4d).

Figure 3.

 Green leaf weevil (Polydrusus flavipes) attraction to Betula pendula and Rhododendron tomentosum volatiles vs blank arm in Y-tube olfactometer. Treatment, closed bars; blank, open bars. Choosing rates (% of total herbivores making a selection) and statistically significant differences between choices based on χ2 test (*, P < 0.05; **, P < 0.01) are shown.

Figure 4.

 Herbivore orientation to Rhododendron tomentosum-exposed (Rt-exposed) vs control Betula pendula leaves (a, b, d) or R. tomentosum vs B. pendula (c) leaves at 30 min, 1 and 2 h after release. Tested herbivores were Polydrusus flavipes adults (a), Epirrita autumnata larvae (b, c) and Euceraphis betulae adults (d). Choosing rates (% of total herbivores making a selection) and statistically significant differences between choices based on χ2 test (P < 0.01**, P < 0.001***) are shown.

Herbivore field observations

Observations for natural herbivore densities in the field were conducted at four time points for control and R. tomentosum mixed association plots. Three aphid genera, earlier reported from the field site (Holopainen et al., 2009), were found on the birch seedlings: Euceraphis, Callipterinella and Betulaphis sp. A higher percentage of control birch leaves than R. tomentosum-exposed birch leaves were infested with Euceraphis sp. on 18 August (linear mixed model, F1,22 = 5.9, = 0.023). Densities of other aphid species did not differ between treatments (Fig. 5a). Numbers of mined leaves, rolled leaves and sawflies did not differ between control and R. tomentosum-exposed birches (results not shown). The total number of leaves and percentages of herbivore-damaged and yellowing leaves were not statistically different between the treatments (Fig. 5b).

Figure 5.

 (a) Aphid incidence (Euceraphis, Callipterinella, Betulaphis species and total aphids, expressed as a percentage of leaves infested); (b) total number of leaves (left axis) and percentage of damaged and yellow leaves (right axis) in conspecifics association (control, open bars) and in Rhododendron tomentosum mixed association (R. tomentosum-exposed, closed bars) Betula pendula in the experimental field plots on 23 July (t1), 4 August (t2), 18 August (t3) and 5 September 2009 (t4). Means ± SE and statistically significant differences (*, P < 0.05, linear mixed model) are shown.


This study demonstrates that volatiles emitted by intact plants can be adsorbed to neighbouring heterospecific foliage, with their subsequent release (for more than 4 h after exposure) affecting the attractiveness of receiver leaves to herbivores. R. tomentosum-specific semi-volatile compounds were found in emissions of neighbouring birch leaves both in nature and in field and laboratory experimental setups. Certain birch-feeding herbivores (E. autumnata and P. flavipes) avoided R. tomentosum leaves and their volatiles. R. tomentosum-exposed birch seedlings were less attractive to E. betulae aphids than control seedlings in the field, and P. flavipes preferred nonexposed birch leaves to R. tomentosum-exposed leaves. Thus, semi-volatiles, in particular, when adsorbed to neighbouring species foliage, could lead to associational resistance and help plants protect themselves as a result of coexistence. This may support the idea that local plant/tree communities are coevolved with adaptation to the presence of their ‘chemical neighbour’ (Grøndahl & Ehlers, 2008, Orians & Björkman, 2009).

Persistence of volatiles – how variable is it?

Leaf properties such as the thickness of the cuticular wax layer, as well as volatility and polarity (Riederer et al., 2002) of the volatile compound in question, and environmental conditions such as wind, light, humidity and temperature (Niinemets et al., 2004) affect the adsorption of volatiles from surrounding air to leaves. The volatility and lipophilicity of plant-emitted volatile compounds vary substantially (Noe et al., 2006, 2008). Sesquiterpenes of a semi-volatile nature are highly temporally variable in their concentrations in the environment, since they are easily adsorbed to vegetation and soil and may form stationary envelopes on leaf surfaces (Helmig et al., 2004), but they are lost rapidly in atmospheric reactions and aerosol formation (Pinto et al., 2007). Two of the specific compounds emitted by R. tomentosum (palustrol and ledol) are oxygenated sesquiterpenes that typically have high polarity and thus lower volatility (Helmig et al., 2004). Oxygenated terpenes are ecologically long-lived (available to volatile receivers for a long time) in comparison to other classes of plant-emitted compounds, such as green leaf volatiles (Pinto et al., 2007). However, it seems that oxygenation is not required for adsorption to occur, since ledene and α-humulene, both nonoxygenated C15 sesquiterpenes, were also increased in R. tomentosum-exposed birch leaves. The monoterpene β-myrcene, which was emitted in very high concentrations by R. tomentosum, did not significantly increase in neighbouring birch emissions. This suggests that either the atmospheric stability of this monoterpene or its adsorbance to receiver foliage was significantly lower than that of the specific sesquiterpenes.

Effective distance for airborne signals

The distance between neighbouring plants is a key factor needed to predict the ecological effects of volatile-based plant–plant signalling. In a field study on wild tobacco (Nicotiana attenuate) and sagebrush (Artemisia tridentata), the effective distance for airborne signalling to increase herbivore resistance of tobacco plants was 0.15 m (Karban, 2001). Clipping sagebrush induced defence of neighbouring conspecifics at a distance of 0.6 m (Karban et al., 2006). Tscharntke et al. (2001) described how defoliating black alder (Alnus glutinosa) resulted in reduced herbivore incidence on neighbouring conspecifics (at 1.3 m distance) compared with more distant neighbours (10.6 m from the defoliated individual). Karban (2007) reported that the degree of leaf area loss to herbivory in Wyethia mollis could be predicted by the distance to a sagebrush neighbour. According to these and our current results, the presence of sticky semi-volatiles could mean that all the vegetation within at least c. 0.5 m might affect a plant’s ecological fitness by altered success of the plant to attract natural enemies or to resist herbivory. It is, however, noteworthy that in the natural habitat of our study, the birch emissions of control plots, located more than 5 m from R. tomentosum, also had small concentrations of R. tomentosum-originating sticky volatiles. This suggests that, at least in certain weather conditions, even greater distances might be relevant for volatile communication by specific persistent volatiles carried by wind in natural habitats. Barney et al. (2009) used a distance of > 10 m to separate invasive plant-emitted volatiles from background emissions in a monoculture and found similar reduction in, but no total disappearance of, the specific volatiles studied.

For insects, the role of volatiles emitted by the neighbouring plants of their hosts, which may interfere with or confuse signals, has not been addressed well in any studies to date (but see Stenberg et al., 2007). The specificity of volatile-mediated plant–herbivore and tritrophic interactions depends on how vulnerable the interaction in question is to environmental interference. Volatile-degrading gases such as O3 (Pinto et al., 2007), but also neighbour-emitted volatiles, could be considered as interferers that might attract or repel herbivores or mislead parasitoids and predators. Ecologically, sticky semi-volatiles might be difficult for insects to use as volatile cues, since they are not restricted to the site of emission but spread on neighbouring plants. However, some of them, such as ledol and palustrol, are specific in the way that they are only emitted by a few plant species (Butkiene et al., 2008) and are scarce anywhere else. So they are reliable indicators of the presence of these certain plants in a habitat and could thereby act as attractants for specialist herbivores but as repellents for generalists and specialist herbivores of other plant species. However, their signalling value then becomes misleading when the compounds are adsorbed to neighbouring interspecific foliage. Some other nonoxygenated semi-volatiles, such as the sesquiterpene (E,E)-α-farnesene and the homoterpene DMNT ((E)-4,8-dimethyl-1,3,7-nonatriene) are typically induced by herbivore feeding in many plant species. These compounds act as a signal for natural enemies of herbivores (Dicke, 2009; Himanen et al., 2009). However, as a result of their high reactivity in the atmosphere (Pinto et al., 2007), they are probably not adsorbed to neighbouring plants and would thus keep their signalling value for natural enemies of herbivores better.

Previous studies have demonstrated a role of volatiles, emitted by damaged intraspecific neighbours (Choh et al., 2004; Karban et al., 2006), damaged interspecific neighbours (Kessler et al., 2006) and intact plants that naturally emit defence-inducing compounds (Khan et al., 1997), in defence activation of intact plants. Thus, it was also interesting to see whether any induction of self-emitted volatiles could be found for R. tomentosum-neighbouring birches. The only increased self-emitted terpenoid from R. tomentosum-neighbouring birch was the sesquiterpene α-humulene, which does not belong to the major compounds believed to be induced by herbivory (Vuorinen et al., 2007) and was emitted in minor amounts. Since R. tomentosum also emitted α-humulene, adsorption seemed to occur for this sesquiterpene in the field experiment. To effectively assess any activation of defence in birches, analysis of within-plant molecular responses or a follow-up stable isotope labelling study would be beneficial.

Does adsorption of volatiles emitted by neighbouring plants confer associational resistance?

Associative resistance was defined by Atsatt & O′Dowd (1976) as neighbours hosting or attracting more natural enemies of herbivores of a particular plant individual, or reducing herbivore attack through luring herbivores away or preventing detection of the plant individual. Volatile compounds are an excellent mediator of associational resistance, since herbivores use plant-released volatile cues in the location of hosts, avoidance of unsuitable hosts and in pheromone communication (Bruce et al., 2005). Our results suggest that a high concentration of volatiles on birch foliage that is not self-emitted may change host selection behaviour by birch herbivores. Semi-volatiles have an additional advantage for conveying associational resistance, since they can be easily adsorbed during cool night temperatures and re-released in the warmer day temperatures (Schaub et al., 2010), when herbivorous insects are most active. This may be important regarding their infochemical value in synchrony with higher trophic level insect activity or abiotic interactions in nature (Heil & Karban, 2010).

The effects of neighbouring plant species with specific volatile profiles on herbivores can be more easily observed in single species-dominated ecosystems (Khan et al., 1997), but Karban (2007) succeeded in showing that an interspecific plant–plant interaction, affecting the degree of herbivore damage, can exist in nature. We showed that P. flavipes avoided R. tomentosum volatiles and verified that after exposure to R. tomentosum, birch leaves were less attractive to P. flavipes than control leaves. Additionally, Euceraphis spp. adults showed temporally reduced incidence in R. tomentosum-neighbouring birches in the field. Future work should include observing herbivore incidence, leaf area removal and fitness consequences in natural habitats of R. tomentosum and Betula sp. Support for potential associational resistance for R. tomentosum-neighbouring species comes from earlier studies. R. tomentosum extracts, with myrcene and palustrol as main compounds, had repellent effects on ticks (Ixodes ricinus) (Jaenson et al., 2005) and mosquitoes (Jaenson et al., 2006), yet isolated ledol had no effect on feeding of the Colorado beetle (Leptinotarsa decemlineata Say) (Szafranek et al., 2006). Instead, ledol induced an electroantennographic response in mahogany shoot borers (Hypsipyla grandella, Lepidoptera: Pyralidae) and was suggested partly to account for their orientation to Guarea macrophylla trees (Lago et al., 2006).

The benefits of plant–plant signalling for emitters and receivers are multifaceted (Heil & Karban, 2010). Sagebrush is an example of efficient plant–plant volatile signalling-based associational resistance in a natural setup, as it has reduced the leaf area that is removed by herbivores for wild tobacco (Karban, 2001) and W. mollis (Karban, 2007). In these cases, the receiver has clearly benefited from neighbouring. However, plant species may also use odours to compete for shared natural enemies of their own and their neighbour’s herbivores, thereby reducing the herbivore impact on themselves and conspecifics and increasing herbivory on competing neighbouring species (Stenberg et al., 2007). The passive process of volatile adsorption that we demonstrated here further complicates this discussion. In the case of sticky volatiles, the receiver could benefit from increased repellency towards herbivores, but ‘helping’ a neighbour might cost the emitter through increased competition.

Concluding remarks

Our study revealed a new mechanism for plant–tree interactions in a boreal ecosystem: R. tomentosum-emitted aromatic volatiles are passively adsorbed to Betula spp. foliage. When re-released, these volatiles were repellent to birch herbivores, which suggest the presence of associational resistance. This could be ecologically important, affecting plant fitness and even species distribution. Future work should address whether these neighbour-emitted volatiles affect the volatile-based orientation of natural enemies of herbivores for the adsorbing foliage, if defence processes are activated by adsorption of neighbour-emitted persistent compounds, and if these sticky volatiles are able to condense from the leaves of neighbouring plants to have allelopathic effects (Barney et al., 2009). Our findings also suggest an important methodological issue that when volatiles are collected in natural habitats, care should be taken in describing the nature of the growth environment. The presence or absence of certain plant species as neighbours of the focal species may increase or decrease the volatile emissions detected. Finally, our results show that interactions mediated by species-specific semi-volatile compounds are a good example of the complicated ecological effects that need to be considered at the level of a plant community rather than individual plants.


This work was funded by the Academy of Finland (grant no. 111543). We thank Maria Saastamoinen and Eposi Grace Evele for assistance in field work.