Variation in plant leaf traits affects transmission and detectability of herbivore vibrational cues

  • Estefania Velilla (Contributor)
  • Jernej Polajnar (Contributor)
  • Meta Virant-Doberlet (Contributor)
  • Daniel Commandeur (Contributor)
  • Ralph Simon (Contributor)
  • Jacintha Ellers (Contributor)
  • Wouter Halfwerk (Contributor)

Dataset

Description

Many insects use plant-borne vibrations to obtain important information about their environment, such as where to find a mate or a prey, or when to avoid a predator. Plant species can differ in the way they vibrate, possibly affecting the reliability of information, and ultimately the decisions that are made by animals based on this information. We examined whether the production, transmission and possible perception of plant-borne vibrational cues is affected by variation in leaf traits. We recorded vibrations of 69 Spodoptera exigua caterpillars foraging on four plant species that differed widely in their leaf-traits (cabbage, beetroot, sunflower and corn). We carried out a transmission and an airborne noise absorption experiment to assess whether leaf traits influence amplitude and frequency characteristics, and background noise levels of vibrational-chewing cues. Our results reveal that species-specific leaf traits can influence transmission and potentially perception of herbivore-induced chewing vibrations. Experimentally-induced vibrations attenuated stronger on plants with thicker leaves. Amplitude and frequency characteristics of chewing vibrations measured near a chewing caterpillar were, however, not affected by leaf traits. Furthermore, we found a significant effect of leaf area, water content and leaf thickness - important plant traits against herbivory, on the vibrations induced by airborne noise. On larger leaves higher amplitude vibrations were induced, whereas on thicker leaves containing more water airborne noise induced higher peak frequencies. Our findings indicate that variation in leaf traits can be important for the transmission and possibly detection of vibrational cues.,Animal rearing The beet armyworm S. exigua (Hübner; Lepidoptera, Noctuidae) is a polyphagous insect pest with a worldwide distribution, and is considered a serious pest of vegetables, field, and flower crops. S. exigua was reared at 26°C ± 1°C and 80% relative humidity on a 12:12 h (L:D) photoperiod. Larvae were fed with a corn-based artificial diet, and adults were given 10% sucrose solution. Mating was facilitated by placing a male and a female moth in a plastic round container with a mesh cloth sealing the top. The mesh cloth was used as a surface on which eggs were laid. Eggs were collected by removing the cloth and cutting and placing the sections of the cloth containing eggs on diet-filled petri dishes. As eggs hatched, larvae developed on these diet-filled petri dishes. The life cycle of S. exigua under our rearing conditions was completed between 25-30 days and included 5 instars. Each instar transition took between three and five days. Plant rearing and measuring of plant leaf traits Four plant species were used for this experiment: sunflower H. annuus, cabbage B.oleracea var. Capitata, beetroot B. vulgaris, and corn Z. mays. Seeds were bought from commercial companies: Seedo and 123Zaden. Plants were reared in growing chambers at 20˚C - 25˚C with an 8:16 h (L:D) photoperiod and with 70% relative humidity. Quartz sand was used instead of potting soil to control for any variation in soil quality that could influence plant traits. A fixed amount of 50% Hoagland’s solution was provided every second day. The amount changed gradually with the growth of the plants (~10-100ml). Plants that were between six- and eight-weeks post germination were used in the experiments. We measured four leaf traits in the lab: Leaf area (cm2), leaf mass (g), leaf thickness (mm), and punch force (N) following the handbook for standardized measurement of plant functional traits (Pérez-Harguindeguy et al., 2013). Leaf area was measured by scanning the leaves, and subsequently measuring the area with the program ImageJ. The dry weight was calculated by first drying the harvested leaves for at least 48 hours in a 70 °C oven, and then weighing them. Dry weight and leaf area were used to calculate the specific leaf area SLA (cm2/g) (one-sided area of a fresh leaf, divided by its dry weight). Leaf thickness was measured with a manual calliper on nine locations on the leaf, seven for corn (Fig. 1). Punch force was measured as the maximum (i.e. pulse) force needed for a 1 mm diameter blunt needle to puncture the leaf, which was clamped tightly on either side of the puncturing spot using a Mecmesin Ultratest Newton meter with Force Gauge AFG 1000-N (Mecmesin, Broadbridge UK). These measurements were done on the same nine locations within eudicot leaves and in, the same seven locations for corn. The measuring points one, two, and three covered the main vein, and the points four, five, six, seven, eight, nine the soft tissue (Fig. 1). The fresh weight and dry weight were used to calculate the fresh weight to dry weight ratio, which we interpreted as the water content. We measured leaf traits of two sets of plants. The first set (n= 101) consisted of plants paired up with the plants used in the recordings of caterpillar chewing vibrations (chewing experiment). This paired design allowed us to measure leaf traits in plants undamaged by chewing while being representative of the leaf traits of chewed plants, but relied on the assumption that paired plants used to measure traits represent the leaf trait variation of the plants on which chewing measurements were done. Although this assumption may not have been met for all pairs, it still seems reasonable compared to the alternative design which would involve measuring leaf traits and vibrations on the same leaf, given that chewing behaviour could be affected by prior measuring of the traits (e.g. damage by punch force tests could have elicited secondary metabolites), and trait measurement could have been affected by foraging (e.g. loss of mass or strength after herbivory attack). The second set of plants (n=40) corresponds to the transmission experiment, described in a section below. In this case, there was no confounding effect of herbivory, and we therefore used the same plants to measure vibrations and leaf traits. Recordings of caterpillar chewing vibrations The purpose of these recordings was to investigate the effect of variation in leaf traits on the production of caterpillar chewing vibrational cues. We first made recordings of caterpillars chewing on different plant species and we then related the amplitude and frequency characteristics of those recordings to variation in leaf traits. A custom-built wooden box (90cm x 60 cm x 60 cm) with a Plexiglas top was used to reduce airborne noise during recording of the chewing vibrations. The box was lined with noise absorbing foam and was placed on a vibration reduction marble table with passive suspenders to minimize substrate-borne noise of the building. The airborne noise amplitude inside the box at the location of the plant, was approximately 35 dB (A) measured with an Extech SDL600 sound level meter set to (set to fast and max). Caterpillar chewing was recorded using a Laser-Doppler Vibrometer (LDV; Polytec PDV-100, set to 5 or 20 mm/s/V, sampling rate 22 kHz). The output of the laser was acquired using a TASCAM DR-60D MKII audio recorder (44,1 kHz, 16-bit resolution). The recording level of the Tascam was set so that there would not be overload of the signal. A reference signal using the same recording level was made against the vibration-isolation table after each recording to allow calculation of absolute amplitude. A reference signal can be made by generating a sinusoidal test signal of 2.80 V (RMS) with a frequency of approximately 1 kHz by setting the laser service mode to Output = Full (Polytec PDV-100 user manual, section 5). We recorded this test signal on the table on which we placed our plants with caterpillars and we recorded it with the TASCAM with the exact same recording settings used per recording of caterpillar chewing vibrations. A single plant was placed inside the noise-isolation box, and a free-moving caterpillar was placed on a leaf of the plant. The caterpillar was always dropped on the middle of the leaf, but it foraged freely and therefore the foraging position changed per trial. A piece of reflective tape placed central on the measuring side of the leaf was used to enhance reflection of the laser beam. Recordings started always around the same time of day (~13:00). A recording started when a caterpillar started eating, and the recording lasted 30 seconds. The distance from feeding site to the measurement site (reflective tape) was noted. The distances ranged from 0.50 to 33.3 mm, with a mean ± sd distance of 12 ± 8.7 mm. Individual plants and caterpillars were used only once for each recording. Recordings were done throughout three larval stages of the animals (L2-L4/5). Before the recordings, caterpillars were weighed and allowed to acclimatize to the experimental chamber. We had 19 replicates for cabbage, 24 for beetroot, 18 for sunflower, and 16 for corn. However, eight recordings were too low in amplitude and the chewing vibrations were not distinguishable from the background noise levels. Therefore, our total sample size was reduced to 69 caterpillar – plant combinations. Transmission experiment To test which leaf-traits affected the transmission of caterpillar chewing cues among plant species, we conducted a vibrational playback experiment. The purpose of this experiment was to measure the transmission of a synthetic vibratory signal throughout the leaves of the different plant species we used during the chewing vibrations recordings. We then related these transmission measurements to the different leaf traits. Using a Brüel & Kjær mini-shaker Type 4810, we played back a two-minute frequency sweep starting at 20 Hz and ending at 2 kHz. The playback was corrected for the frequency response of the shaker to ensure constant velocity at all frequencies, and adjusted to the RMS amplitude level of chewing vibrations recorded in caterpillar trials. This adjustment was necessary because the frequency response of the mini shaker is highly non-linear in the low frequencies (< 100 Hz), biasing the playback to the frequencies > 100 Hz in a non-corrected format. Therefore, by correcting our playback file to the frequency response of the shaker, we made sure all frequencies were equally represented. We used two LDVs, one to register vibrations next to the source (Fig. 1) and the other on the nine different points on the leaf, seven for corn (Fig. 1). These points were the same points that were used for measuring leaf thickness and punch force (Fig. 1). Furthermore, we recorded the sweep on the adjacent leaf to test relative energy loss during transmission. Vibrations were transferred via a rod mounted on the shaker and attached with Blue Tack adhesive to the underside of the leaf on a point in the centre of the soft tissue, not touching the main vein, between points five and eight (Fig. 1). The distance between each point and the stimulus was noted. The distances ranged from 1 to 20 mm. We tested five plants per species. Airborne-noise exposure experiment To test whether amplitude of vibrations induced by airborne noise was affected by leaf traits, we did an acoustic noise playback experiment with white noise (0.1 – 20kHz) (Rebar et al. 2012). Using a Behringer MPA40BT speaker positioned 60 cm from the plant, we played back 10 seconds of white noise at 70 dB (A) measured at the position of the leaf with an Extech SDL600 sound level meter (set to fast and max) and recorded with the LDV on three points (points 1:3, Fig. 1). Using an LDV we recorded the vibrations induced by the airborne noise playback. We tested five plants per species. Analysis of vibratory measurements Chewing recordings were first filtered with a 100 Hz high pass filter in R version 3.3.1 (R Core Team 2016), run in the RStudio interface (RStudio Team 2015) with the function “fir” from the Seewave package version 2.1.4 (Sueur, 2008). Recordings were filtered to remove high amplitude, low-frequency background building noise. Using Raven Pro 1.5 software (Cornell Lab of Ornithology 2017), we selected ten chewing events per recording (see Fig. A1 in Appendix 1 for an example). We measured root mean square (RMS) amplitude from the waveform, and first and third quartiles. Peak frequency was taken from the spectrum (sampling frequency: 44100, window type: “Hanning”, window size: 1024, overlap: 50). All measurements were done on the filtered recordings. Reference recordings were filtered in the same way as chewing recordings. RMS measurements from the reference recordings were used to calculate absolute RMS amplitude (mm/s) of chewing events. To calculate absolute RMS amplitude, we used the following formula: RMS amp (mms)= RMSmeasurementRMS(reference)*2.80*LDV vel scaling setting The value of 2.80 represents the RMS amplitude (in Volts) output of the LDV, and the LDV velocity scaling setting was either 5 or 20 mm/s/V (Polytec PDV-100 user manual, section 5). Analysis of vibratory sweep and acoustic noise playbacks The sweep recordings were high-pass filtered in the same way as the chewing recordings. The main frequency range of chewing recordings was determined by plotting frequency spectrum of a representative recording per plant species (Fig. 2). We determined RMS amplitude of reference recordings in Raven Pro, which were made for every recorded point, including the ones on the adjacent leaves. Attenuation (dB) was calculated for every measurement point of every individual plant using the following formula: 20Log10=RMS(measurement)RMS(reference) We also calculated peak frequency from the spectrum (sampling frequency: 44100, window type: “Hanning”, window size: 1024, overlap: 50). Because we did not measure transmission on the stem, we decided to use the midvein points and the adjacent leaf as a proxy for transmission. Hence, our statistical analyses not only explore differences in the average amplitude change (dB) and mean peak frequency (Hz) measurements across all points on the same leaf, but also of the midvein points, as well as transmission via the stem to the nearest leaf. Noise recordings were also high-pass filtered and RMS amplitude (dB) and peak frequency (Hz) measurements obtained with Raven Pro. All recordings were normalized in R by dividing them by the maximum amplitude of the loudest recording.,In the dataset called "plant.traits_transmission experiment_2017" (Excel file) used during the transmission experiment, the variables LT1, LT2...LT9 (leaf thickness) are expressed in a different unit than in the datasets from the caterpillar chewing experiments, called "plant.traits" and "plant.traits.complete". In those datasets leaf thickness is expressed in mm. Therefore, in the R script "transmission.experiment", I divide the leaf trait values by 100 to convert leaf thickness to mm. In the dataset "PT_whole" (Excel file), the variables "dry.weight" and "fresh.weight" are also in the wrong units (mg). Therefore, in the R script "substrate_effects_on_chewing_cues" I divide them by 1000 to properly calculate water content and SLA. The dataset "plant.traits.complete" (Excel file) contains the corrected units of measurement for all variables.,
Date made available1 Jan 2020
PublisherUnknown Publisher

Cite this