Electronic Supplementary Material

Plants suppress their emission of volatiles when growing with conspecifics

Rose N. Kigathi, Wolfgang W. Weisser, Daniel Veit, Jonathan Gershenzon & Sybille B. Unsicker*

*Corresponding author: Sybille B. Unsicker, Max-Planck-Institute for Chemical Ecology, Hans-Knöll Str. 8, 07745 Jena, Germany, phone: +49(0)3641 571328, fax: +49(0)3641 571302,

Table S1 The effects of above ground, below ground and full contact interaction on the volatile emission of Trifolium pratense analyzed by ANOVA (for details see Materials and Methods). Each type of contact had 10 replicates, 5 “herbivory”, and 5 “no herbivory”, giving a total of 60 plants. Part A: Total volatiles emitted includes herbivore induced volatiles, which increased by more than 5-fold after caterpillar feeding, and constitutive volatiles. Part B shows individual results for the major herbivore induced compounds. Bold numbers indicate significant results. DMNT, (E)-4,8-dimethylnona-1,3,7-triene; MeSA, methyl salicylate.


Table S2 Emission (mean ± SE) of all major volatile compounds in ng (g dry wgt)-1 h-1 from Trifolium pratense. Plants grew either without neighbors, in intra-specific interaction with another T. pratense, or inter-specific interaction with a Dactylis glomerata plant under three different contact treatments: above ground, below ground or full contact. Volatile emission was measured from undamaged control plants (No Herbivory) and after feeding by two Spodoptera littoralis caterpillars (Herbivory).(See Table S6 for retention indices (RI) of the compounds)

†Compounds identified by comparison to authentic standards. Others were identified by comparison of mass spectra and retention times to those in the literature.

Table S3 The effects of above ground, below ground and full contact interaction on A: above ground biomass, root biomass, the root:shoot ratio , flower biomass of Trifolium pratense. B: Specific leaf area, leaf area eaten in cm2 and percent herbivory on T. pratense analyzed by ANOVA. Biomass data were analyzed only in the control plants (no herbivory) to reveal changes caused by neighbor identity alone. Each type of contact had five replicates, giving a total of 30 plants (for details see Materials and Methods). Bold numbers indicate significant results. (See Figs S2, S3b for graphs).

Table S4 The effects of above ground, below ground and full contact interaction on the volatile emission of Trifolium pratense analyzed by ANOVA (for details see Materials and Methods). Each type of contact had 10 replicates, 5 “herbivory”, and 5 “no herbivory”, giving a total of 60 plants. Here only data from the 5 “herbivory treatments are presented (30 plants) including the percent level of herbivory as a covariate. a: Total volatiles emitted include herbivore induced volatiles, which all increased more than 5-fold after caterpillar feeding, and constitutively emitted volatiles. b: Individual results for the major herbivore induced compounds. Bold numbers indicate significant results. DMNT, (E)-4,8-dimethylnona-1,3,7-triene; MeSA, methyl salicylate. c: Individual results for the major constitutive compounds.

Table S5 Emission (mean ± SE) of volatile compounds in ng (g dry wgt)-1 from Dactylis glomerata. Plants grew either without neighbors or inter-specific interaction with a Trifolium pratense plant under three different competition treatments: above ground, below ground or full contact. Volatile emission was measured from undamaged control plants (No Herbivory). Calculated RI is the retention index determined using alkane standards as references on the DB5 column.

†Compounds identified by comparison to authentic standards. Others were identified by comparison of mass spectra and retention times to those in the literature.

Table S6 Retention indices of compounds emitted by Trifolium pratense. Calculated RI are retention indices determined using alkane standards as references on the DB5 column. For comparison, retention indices from the Adams library (Adams, 2007) and MassFinder libraries are provided.

†Compounds identified by comparison to authentic standards. Others were identified by comparison of mass spectra and retention times to those in the literature.

Supplementary Figure Legends

FIG. S1 Sketch and photographs of the experimental set-up. A) After transplanting, above ground parts of all plants were placed in a cylinder of PET film (height 44 cm, Æ 13 cm) attached to a Plexiglas top that was supported with two threaded metal bars. The film was attached both on top and at the bottom of the bars with a clamp (b and c). In case of above ground contact, two PET cylinders were connected through an opening of 5 x 5 cm (a). Throughout the duration of the experiment, charcoal-filtered air was pumped into the cylinders at a flow of 2 l/min through a Teflon tube. The air exited the system through an opening in the Plexiglas lid. In all interaction treatments, two pots (350 ml, 7 x 7 x 8.5 cm) were connected with adhesive tape (c). b) Scheme of the volatile collection system. Clean air was pumped in through Teflon tubes (continuous arrows) and pumped out through a trap (dotted arrows). For details on the volatile collection see “Materials and Methods” section. c) When the volatiles were sampled, all plants were bagged individually and an additional Teflon tube with the Poropak filled trap (ARS, Gainesville, FL, USA) was inserted into the cylinder through the second opening in the Plexiglas lid.

FIG. S2 Effect of plant-plant interaction on Trifolium pratense growth. Graphs depict: a) above ground biomass, b) root biomass c) root:shoot ratio and d) flower biomass of plants growing under different types of interaction including: T. pratense plants growing with no neighbor (grey hatched bars), growing with a T. pratense neighbor (white bars), and growing with a Dactylis glomerata neighbor (grey bars). The types of contact are none, above ground contact, below ground contact, and both above ground and below ground contact. Bars represent mean ± SE. [Replace “competition” with “neighbor” in legend in table.]

FIG. S3 Effect of plant-plant interaction on herbivory. Graphs depict: a) total volatile emission for herbivore-damaged plants, b) percent of leaf area lost to herbivory and c) actual leaf area lost after herbivore feeding of plants growing under different types of plant interaction including: Trifolium pratense plants growing with no neighbor (grey hatched bars), growing with a conspecific neighbor (white bars), and growing with a heterospecific neighbor, a Dactylis glomerata plant (grey bars). The contact treatments are none, above ground, below ground, and both above ground and below ground. Bars represent mean ± SE.

FIG. S4 Comparison of constitutively emitted volatiles from Trifolium pratense (red dots) and Dactylis glomerata(green diamonds) by non-metric multidimensional scaling. The lack of much overlap between samples from the two species suggests that the two species have distinctively different volatile blends.

FIG. S5 Evidence of root competition in experimental treatments with below ground contact. Photographs show differences in root architecture due to the presence of competing species: (a) Trifolium pratense plants growing with below ground contact to Dactylis glomerata and (b) T. pratense plants growing without below ground contact. The T. pratense plants without below ground contact to neighbors have more lateral roots than those with contact to D. glomerata. Trifolium pratense plants growing with below ground contact to another T. pratense had roots similar to those in the no neighbor treatment. Design of pots for different treatments: Pots for below ground contact are shown in top (c) and side (d) views displaying the openings on the side (5 x 6 cm) which when connected allowed root interaction. (e) Pots without below ground contact viewed from top and side. In all interaction treatments, two pots (350 ml, 7 x 7 x 8.5 cm) were connected with adhesive tape (see Materials and Methods for details).

FIG. S1

FIG. S2

FIG. S3

FIG. S4

FIG. S5

8