Electrostatic detection and electric signalling in plants: do flowers act as antennas?

Flying insect pollinators are electrically charged. As bumblebees (Bombus terrestris) and honeybees (Apis mellifera) are almost always positively charged, they present a static electric field that is modulated by the harmonic motion of their wings. Previous research has demonstrated that as a bee approaches a flower, there is a change in the stem potential of the plant, even before the bee lands, suggesting a capacity for flowers to sense the approaching bee through its electric field. Using a combination of laboratory and field studies, we explore the potential for flowers to act as receiving antennas for electrical signals, and to transmit these signals to neighbouring plants. Results show that flowers can detect non-contact electric stimuli, presumably by charge induction, and appear not to be tuned to a specific frequency. In the field, non-contact electric stimuli can also be detected in neighbouring flowers that did not receive the aerial stimulus. This evidence demonstrates the ability of flowers to capture local, aerial electrical signals, and the plant to then transmit signals through the soil to neighbouring plants. This work highlights the significance of environmental electric fields in pollination biology.


Introduction
Flying insect pollinators, such as bumblebees (Bombus terrestris) and honeybees (Apis mellifera) generally accumulate a positive charge during flight [1,2].Flowers have a static electric field, which is negative relative to the electric field of the flying bee.The charge density of the floral electric field varies at different parts of the flower head depending on flower geometry [1].When a positively charged pollinator approaches a flower, these charges interact through Coulomb forces which can facilitate pollen transfer [3] and potentially provide foraging cues to the pollinator [1,4,5].Reciprocally, the electric fields of the pollinator could also provide a cue to the flower that a pollinator is approaching, even before it lands.In effect, the stem potential of Petunia flowers changes during bumblebee visitation prior to landing [3] and, in fact, repeated stimulation with the electric field of a bumblebee causes an increase in the emission of floral volatile organic compounds [6].
Electrical communication in plants has been considered in the context of action or variation potentials relaying signals between cells in the same organism [7], and there has been a suggestion that fungal mycelia can mediate plant-plant electrophysiological communication [8].Direct communication of an electrical stimulus from one plant to another has been given only little attention [9,10], and the stimulus in these studies has been delivered by conductive contact with the plant.Consideration has been given to the potential for the static electric field of pollinators to act as a non-contact electrical stimulus for flowers [1], but it has been understudied.
Nectar is essential to most plant-pollinator interactions, but it is energetically costly to produce [11].If plants could increase nectar production only whilst detecting the presence of pollinators, it would conserve metabolic energy.It is therefore conceivable that electric fields in the environment, natural or anthropogenic, may play a role in mediating plant-pollinator interactions, thus impacting global food production.Here, we explore the potential for flowers to act as receiving antennas for non-contact electrical signals that are further transmitted to neighbouring plants.

Laboratory experiments
Daffodils (Narcissus Spp.) were maintained in a laboratory at the University of Bristol's Life Sciences Building.Daffodil stems were cut to 220 mm long and placed in a conical centrifuge tube containing 35 ml of water in a 0.8 m tall parallel plate capacitor.The top of the daffodil was always set to 0.36 m in height.Electrophysiological measurements were made using a sharp stainless steel measurement electrode (0.67 mm diameter) inserted into the stem of the daffodil, shielded by a Faraday cage (Fig. 1).The  Experimental electric fields were generated using a parallel plate capacitor.A 14 Vpp, 3.2 s, 0.1-1000 Hz frequency sweep was generated in MATLAB (Version 2018b) and output through an analogue output module (NI-9263 housed in cDAQ-9174, Austin, Texas, United States).The signal was amplified by a factor of 40 using a custom high-voltage amplifier and passed to the top capacitor plate while the bottom plate was earthed, resulting in a field strength of 350 V/m across the plates.The sweep was played five times low-high frequency and five times high-low frequency with a six-second off period between the change in frequency direction and electrophysiological recordings were made throughout.After each full sweep playback, a floral component was removed to determine its role in the antennal properties of the flower and another playback and recording was made.The removal order of the sex organs, trumpet, and petals of the daffodil (Fig. 1) was randomised.After all three were removed, the rest of the daffodil was incrementally sliced to remove its remaining parts, the head, half the stem, and the rest of the stem.Finally, the electrode was removed from the stem of the flower and a recording was taken to ensure the signal was being detected through the daffodil and not directly detected by the electrode.
Recorded signals were analysed in Spike2 (Cambridge Electronic Design Ltd, Cambridge, UK) using a Fast Fourier Transform (FFT; Hanning window and size 1024) to both assess the spectral sensitivity of daffodil and calculate the power of the signal recorded at each stage of floral component removal.Differences in power between removal stages were analysed using an analysis of variance (ANOVA; one-way, two-tailed) and a Tukey HSD post-hoc test in R (version 4.1.1,R Core Team, 2021, Vienna, Austria) within the RStudio interface (2022.12.0 Build 353, Posit Software, PBC).

Field experiments.
To investigate the potential for plants to receive and transmit aerial electrical signals, the response was also measured in adjacent plants set in the same soil medium, yet electrically shielded from the aerial stimulus.A 183 Hz 10 Vpp sine wave (approximate pollinator wingbeat frequency [12]) was produced by a portable signal generator (FG-100 DDS Function Generator) and delivered for 10 seconds through a 1.5 mm diameter silver ball (spherical) electrode 10 mm from the flower head.The signal was measured with a stainless-steel electrode inserted into the stem of the flower, amplified through a DC (Direct Current) coupled amplifier (DAM50, World Precision Instruments, Hitchin, UK) and captured with a national instruments DAQ (NI USB-6009, Austin, Texas, United States) using a custom MATLAB (Version 2018b) script.Because the signal being measured was also transmitted through the ground, the equipment could not be traditionally earthed through a connection with the soil.Instead, the amplifier, DAQ, and signal generator were connected on the same floating ground to ensure they were maintained at the same resting potential.The shielded flower was inside a Faraday cage to ensure the signal could not reach it aerially (Fig. 2).Three control measurements were taken to ensure the signal detected was not a result of crosstalk: 1. Recordings of the stem potentials after the flower's connection to the rest of the plant was severed, 2. Recordings with the electrode floating in the Faraday cage, and 3. Recordings with the electrode in surface contact with the stem below the cut of the flower (Fig. 4).The recordings from the plants were analysed using the pspectrum function in MATLAB (Version 2018b) to ensure the 183 Hz stimulus was detected.The peak-to-peak voltage of the 183 Hz sine wave recorded in the flower was analysed for the effect of distance from the stimulus flower using a general linear model (GLM) and compared between conditions using ANOVA (one-way, two-tailed) with Tukey HSD post-hoc test in R (version 4.1.1)within the RStudio interface (2022.12.0 Build 353).

Laboratory experiments
The frequency sweep stimulus was detected and measured in the daffodils, but spectral analysis demonstrated a broadband frequency response with no sensitivity peaks.228 Hz was chosen as the arbitrary value to analyse the stimulus power due to the broadband frequency response of the plant.The power of the signal recorded in the daffodil reduced as components of the flower were removed (ANOVA, F11=343.5, P<0.001; Fig. 3).Results indicate that the petals provide the largest influence on the power of the signal recorded, likely due to their relatively large surface area.Further analyses of the geometry and dielectric properties of these components are needed to understand how flowers function as antennas.

Field experiments
Field tests with hogweed plants (Heracleum sphondylium) in the gardens of the University of Bristol Veterinary School, Langford, Bristol, demonstrated an electrical connection between plants.The signal measured in the plant that received the stimulus was 17.04 mVpp, and in a different, Faraday cage shielded hogweed plant 1.59 m away (distance from stem base to stem base), a 15.79 mVpp signal was detected (Fig. 4).When the stem of the shielded hogweed was cut below the position of the measurement electrode, a signal of 3.21 mVpp was detected, this represents no increase in signal above the background noise and was not specific to 183 Hz.These findings support the results of Volkov et al. [9,10], that electrical signals can be propagated between plants, and are the first demonstration of transmission of a non-contact electrical stimulus between plants in the field.To test the connection between plants further, a series of measurements were taken using buttercups (Ranunculus repens) in Royal Fort Gardens, University of Bristol.The same methodology applied to the hogweed plants was used to stimulate and record the buttercups.However, an additional control was conducted; after the flower's connection to ground was cut, to ensure it did not desiccate, the stem was placed in water electrically isolated from ground.Two buttercups were used as stimulus flowers, and the response was measured in eight different buttercups, shielded by a Faraday cage.The distance of the shielded buttercups from the plant receiving the stimulus ranged from 0.33 m to 4.25 m and did not correlate with the signal strength measured (GLM; F1=0.82,R 2 =-0.09,P=0.391); increasing the distance between flowers did not diminish the strength of the transmitted signal.No difference was found between the signal measured in the stimulus flowers and the shielded flowers, but the signal was higher in both the stimulus and shielded flowers than in the control measurements (ANOVA, F4 = 138.1,P <0.001; Fig. 5).A voltage drop over distance may be expected as this is true for traditional electrical circuits; however, it is important to note that many factors may be at play here including the soil geochemistry and the biology of the plants involved.Interestingly, measurements show no signal above the noise when the connection to the soil is severed, clearly demonstrating that the electrical signal is transmitted through the soil by some means, potentially through fungal mycorrhizal networks [8] or direct soil conduction.

Conclusion
Here, we demonstrate that non-contact electrical stimuli can be detected within flowers, likely through induction of mirror charges.We also provide the first evidence of the transmission of aerial electrical signals between neighbouring plants.The complex geometry of daffodil flowers contributes towards their antennal properties.Removing structural components of the flower alters its ability to act as a receiving antenna.The difference in signal strength measured between components removed might be a result of their difference in size, or dielectric properties; further analysis is required to ascertain this relationship.As demonstrated in hogweed and buttercups, some flowers can transmit a non-contact electrical signal to neighbouring plants, and no voltage drop was detected at the distances measured.These findings broaden our understanding of the electrical interaction between plants and pollinators and stimulate new thinking about plant behaviour.Nectar is extremely energetically costly for flowers to produce, requiring up to 37% of daily photosynthate allocation [11], so if production could be triggered by a stimulus informing the flower of a pollinator's presence, it would ensure the energy is not wasted.Further work will focus on potential metabolic changes in the plant as a result of electrical stimulation, such as increased nectar production.Enticingly, the observed electrical connection may be a result of transmission through fungal mycorrhizal networks, as proposed by Thomas and Cooper [8], and/or direct, low-resistance conduction through wet soil.These findings diversify our view of insectplant interactions, and opens intriguing avenues for future studies into the ability of plants to respond to aerial electric cues and for the encoded information to be communicated with neighbours.

Figure 1 .
Figure 1.The parallel plate capacitor set-up for the experiments removing floral components of daffodils.Each daffodil was placed in 35 ml of water to prevent desiccation.Flower icon for graphical illustration only and not to scale.

Figure 2 :
Figure 2: The setup for the field plant-plant connection experiments, created with BioRender.com.Flower icons for graphical illustration only and not to scale.

Figure 3 .
Figure 3.The response strength recorded in the daffodil (Narcissus Spp.) stem at different stages of floral component removal.Numbers on the top row denote the sample size of experimental replicates.Letters denote significance between conditions; conditions that share the same letter are not significantly different at threshold P<0.05.ANOVA performed in R studio (F11=343.5,P<0.001).

Figure 4 .
Figure 4.The strength of the signal recorded in hogweed (Heracleum sphondylium) after delivery of a 10 second, 10 Vpp, 183 Hz sine wave from a 1.5 mm ball electrode 10 mm away from the head of the stimulus flower.Grey bars denote recordings in the plant that received the stimulus, black bars denote recording in an electrically shielded hogweed plant 1.59 m away (distance from stem base to stem base).Electrode placement is denoted by the schematic at the top, created with BioRender.com.Flower icons for graphical illustration only and not to scale.

Figure 5 .
Figure 5.The response measured in buttercups (Ranunculus repens) after delivery of a 10 Vpp 183 Hz sine wave from a 1.5 mm ball electrode 10 mm away from the stimulus flowers head.Multiple electrically shielded flowers were recorded from the same stimulus flower; Stimulus flower n=2, shielded flower n=8, floating n=1, cut n=7, watered n=5.Electrode placement is denoted by the schematic above, created in BioRender.com.Letters denote significance between conditions; conditions that share the same letter are not significantly different at threshold P<0.001 (ANOVA, F4 = 138.1,P <0.001).