The cibarial pump of the xylem-feeding froghopper Philaenus spumarius produces negative pressures exceeding 1 MPa

Abstract

The xylem sap of vascular plants is an unlikely source of nutrition, being both nutrient poor and held under tensions (negative pressures) that can exceed 1 MPa. But some insects feed on xylem sap exclusively, extracting copious quantities using a muscular cibarial pump. However, neither the strength of the insect's suction, nor the direct energetic cost of xylem ingestion, have ever been quantified. Philaenus spumarius froghoppers were used to address these gaps in our knowledge. Micro-CT scans of its cibarium and measurements of cibarial muscle sarcomere length revealed that P. spumarius can generate a maximum tension of 1.3 ± 0.2 MPa within its cibarium. The energetic cost of xylem extraction was quantified using respirometry to measure the metabolic rate (MR) of P. spumarius while they fed on hydroponically grown legumes, while xylem sap excretion rate and cibarial pumping frequency were simultaneously recorded. Increasing the plants' xylem tensions up to 1.1 MPa by exposing their roots to polyethylene glycol did not reduce the insects’ rate of xylem excretion, but significantly increased both MR and pumping frequency. We conclude that P. spumarius can gain energy feeding on xylem sap containing previously reported energy densities and at xylem tensions up to their maximum suction capacity.

1. Introduction

Insects that exclusively feed on the xylem sap of vascular plants are found only within the Auchenorrhyncha, a suborder of the true bugs (Hemiptera) [1]. These insects are remarkable from a nutritional ecology standpoint as they survive entirely on a liquid that is not only nutritionally poor [2], but also difficult to extract due to the MPa-level of tension (i.e. negative pressure) in xylem vessels. To cope with their unusual diet, xylem-feeding bugs possess a large, muscular cibarial pump for extracting xylem sap as well as a specialized filter chamber in their gut [3] that enables them to extract more than 99% of the trace organic nutrients it contains [4]. However, xylem sap is so dilute that the insects must extract, filter and excrete between 100 and 1000 times their own weight in sap per day [5]. This unique challenge of extracting vast quantities of liquid against phenomenal pressure differences has led some to conclude that xylem-feeding insects are evidence that in vivo xylem tensions cannot be as great as generally accepted [6–8]. Previous studies have attempted to establish the energetic cost of ingesting xylem sap, either by calculating energy use from the difference in energy content between xylem sap and the insect's excreta [4] or from hydrodynamic principles [9]. Similarly, by extrapolation from the muscle mechanics and pump morphology of the blood-feeding bug Rhodnius prolixus, it was estimated that xylem-feeding insects could not generate tensions in excess of 0.3 to 0.4 MPa [10,11]. However, in the absence of direct measurements of these parameters, the accuracy of these estimates remains unknown. Addressing these gaps in knowledge by quantifying the cibarial mechanics and energetics of these insects would not only reveal how they survive on such an unusual diet, but also allow them to be used as biological pressure probes to substantiate the range of xylem tensions that exist within plants in situ. Crucially, given the damage that xylem-feeding insects cause as vectors of plant disease, this knowledge could provide valuable biological insights into how they interact with their plant hosts.

Xylem-feeding insects insert their stylets directly into the plant's mature xylem vessels [5,12]. As they are clearly capable of extracting xylem sap, they must generate a tension that exceeds that of the liquid within the xylem vessel. It is also expected that these insects must have the ability to extract xylem sap under most circumstances, as demonstrated by their rapid and constant rate of excretion while feeding at midday, when a plant's transpiration rate and xylem tension are both at their highest [9]. Their maximum suction capacity should therefore coincide with the upper limits of commonly encountered xylem tensions in their food plants. Several studies have investigated the suction capacity of these insects by determining whether their rate of xylem sap excretion decreases in response to increasing xylem tension, as measured using a Scholander–Hammel pressure bomb. One study showed that increasing xylem tension to 1.0 MPa had no effect on the excretion rate of feeding adult Philaenus spumarius [13], while the excretion rate of the glassy-winged sharpshooter Homalodisca vitripennis (formerly H. coagulata) began to drop precipitously once xylem tension exceeded 1.0 MPa, falling to zero at 2.1 MPa [9]. However, in the absence of additional evidence, these results can be interpreted as indicating not the existence of a large cibarial tension generated by the insect during ingestion, but an erroneous over-estimation of xylem tension by the pressure bomb technique [6,7]. The accuracy of this technique (and by extension the existence of the large xylem tensions it reveals) has been challenged on the grounds that (a) it lacks independent verification and (b) since it is performed on excised leaves rather than intact plants, the measurement itself introduces artefacts [6]. These controversial criticisms could be refuted if an alternative method for measuring xylem tension directly corroborated the range of tensions measured using the pressure bomb. Cell pressure probes—glass micropipettes implanted into xylem vessels—have partly achieved this goal, with pressure bomb and in vivo measurements agreeing across a range of 0 to 0.7 MPa [14]. But measurement of xylem tensions any greater than 1.0 MPa in vivo have yet to be achieved [15]. However, it is possible to calculate the maximum tension generated by a cibarial pump from its morphology, muscle properties and biomechanical principles alone, thereby revealing the highest xylem sap tensions that these insects can feed at in situ without reference to measurements of xylem tension as determined using the pressure bomb method.

For this study, we examined adult common meadow spittlebugs (P. spumarius), a cosmopolitan xylem-feeding Hemipteran, the adults of which are referred to as froghoppers. Micro-CT scans of the insect's head were used to reveal the morphology and muscle architecture of the cibarium, while the force-producing capacity of the cibarial dilator muscles was determined by measuring sarcomere length. These measurements were used to calculate maximum cibarial tension. Then, by measuring the metabolic rate (MR) of feeding froghoppers using flow-through respirometry, combined with video analysis of pumping frequency and xylem excretion, we quantified the metabolic cost of xylem extraction. Manipulating the xylem tension of the plants the insects were feeding on by exposing the roots to the osmolyte polyethylene glycol (PEG) revealed how the metabolic cost of xylem extraction changed with xylem tension. As P. spumarius feed exclusively on xylem sap, they must produce a net energy gain from this activity. Therefore, determining how the metabolic cost of xylem sap extraction varies with xylem tension, and then comparing this to the energy density of xylem sap, we could estimate net energy uptake or loss. Using these two separate approaches—biomechanical and metabolic—we were able to quantify the tensions that these insects can create with their cibarial anatomy, as well as demonstrate how costly it is to feed on xylem sap.

2. Results

(a) . Resting metabolic rate

Froghoppers had an average resting metabolic rate (RMR) of 48.7 ± 28.2 µW (0.15 µl O₂ h⁻¹; electronic supplementary material, table S1; n = 13).

(b) . The effect of xylem tension on feeding parameters

(i) . Pumping frequency

For PEG-manipulated trials, the average pre-PEG pumping frequency (f_pump) was 0.89 ± 0.26 Hz and the average post-PEG f_pump was 0.97 ± 0.24 Hz (electronic supplementary material, table S2; n = 10). Average f_pump ranged from 0.44 to 1.50 Hz and 0.43–1.52 Hz before and after the introduction of PEG, respectively (electronic supplementary material, table S2). The linear mixed-effects model indicated a statistically significant but slight increase overall (electronic supplementary material, table S5; figure 1a), giving the following regression:

$$f_{\mathrm{pump}}\hspace{0.17em}(\mathrm{Hz})=0.257\hspace{0.17em}[\mathrm{xylem}\hspace{0.17em}\mathrm{tension}\hspace{0.17em}(\mathrm{MPa})]+\mathrm{0.740.}$$ 2.1

Figure 1.

Relationship between xylem tension and feeding parameters. Effects of xylem tension on pumping frequency, f_pump (a), excretion rate, Q (b), the ratio of Q to f_pump (c) and cibarial metabolic rate, MR_cib (d). Solid circles indicate the means recorded from individuals, with each individual measured at two xylem tensions (pre- and post-PEG); circle colour corresponds to individual ID. Solid lines indicate linear regressions from linear mixed-effect models with individual froghopper included as a random effect. Grey bands indicate 95% confidence interval limits. (Online version in colour.)

(ii) . Xylem excretion rate, Q

For PEG-manipulated trials, the average pre- and post-PEG Q were both 0.57 ± 0.25 µl min⁻¹ (electronic supplementary material, table S3; n = 8). The linear mixed-effects model showed no statistically significant association between xylem tension and excretion rate (electronic supplementary material, table S5; figure 1b). Q values ranged from 0.16 to 1.10 µl min⁻¹ and 0.08–1.08 µl min⁻¹ before and after the introduction of PEG, respectively (electronic supplementary material, table S3).

(iii) . Pump stroke volume, Q/f_pump

To investigate the relationship between f_pump and volume of xylem extracted per pump, the average pump stroke volume (SV) was calculated as

$$\mathrm{SV}\hspace{0.17em}(\mu \mathrm{l}\hspace{0.17em}\mathrm{pum}{\mathrm{p}}^{-1})=\frac{Q\hspace{0.17em}(\mu \mathrm{l}\hspace{0.17em}\mathrm{mi}{\mathrm{n}}^{-1})}{\hspace{0.17em}{f}_{\mathrm{pump}}(\mathrm{pumps}\hspace{0.17em}\mathrm{mi}{\mathrm{n}}^{-1})}.$$ 2.2

The mean volume of xylem extracted per pump stroke, both before and after the addition of PEG, were identical at 0.011 ± 0.002 µl pump⁻¹ (n = 7). For PEG-manipulated trials, SV had no statistically significant association with the measured xylem tension, as shown by the linear mixed-effects model (electronic supplementary material, table S5; figure 1c).

(iv) . Cibarial metabolic rate

The introduction of PEG was associated with a significant increase in the cibarial metabolic rate (MR_cib, calculated as total feeding MR - RMR) of the feeding froghoppers, with the linear mixed-effects model indicating a significant association between measured xylem tension and MR_cib (electronic supplementary material, table S5; figure 1d). The regression for this relationship is

$${\mathrm{MR}}_{\mathrm{cib}}(\mu \mathrm{W})=27.186\hspace{0.17em}[\mathrm{xylem}\hspace{0.17em}\mathrm{tension}\hspace{0.17em}(\mathrm{MPa})]+\mathrm{11.004.}$$ 2.3

For PEG-manipulated trials, the average pre-PEG MR_cib was 25.5 ± 14.8 µW and the average post-PEG MR_cib was 39.3 ± 19.0 µW (electronic supplementary material, table S4; n = 8). MR_cib ranged from 3.2 to 65.5 µW and 5.4 to 79.1 μW before and after the introduction of PEG, respectively (electronic supplementary material, table S4).

(v) . Volumetric cost of xylem extraction, MR_cib/Q

For PEG-manipulated trials, the ratio of MR_cib to excretion rate, equivalent to the energy expended per unit of xylem extracted, had a statistically significant association with xylem tension as shown by the linear mixed-effects model (electronic supplementary material, table S5), with all of the individuals exhibiting an increase in MR_cib/Q (figure 2a). The regression for this relationship is

$$\frac{{\mathrm{MR}}_{\mathrm{cib}}(\mathrm{W})}{Q(\mu \mathrm{l}\hspace{0.17em}{\mathrm{s}}^{-1})}=0.002\hspace{0.17em}[\mathrm{xylem}\hspace{0.17em}\mathrm{tension}\hspace{0.17em}(\mathrm{MPa})]+\mathrm{0.001.}$$ 2.4

Figure 2.

A comparison between the volumetric cost of extracting xylem sap and the sap's energy density, in relation to the tensions within the cibarial pump of P. spumarius and the xylem sap of their host plants. (a) Xylem tensions measured with the pressure bomb plotted against the energy expended per unit of xylem sap volume extracted for PEG- and non-PEG-manipulated trials. Increasing xylem tension is associated with an increase in the cost of xylem extraction. Equation for OLS linear regression (grey line) is MR_cib/Q = 0.0041[xylem tension (MPa)] + 0.0002. Grey band indicates 95% confidence interval limits. Purple horizontal dashed lines show examples of mean xylem sap energy content (ε_xylem) for a preferred (P. salicina) and non-preferred (P. persica) host of xylem-feeding H. vitripennis. (b) Daytime field xylem tensions of eight species of P. spumarius's host plants as measured with a pressure bomb, by species and histogram (grey bars). Xylem tensions are colour-coded by ambient temperature recorded at the time of measurement. Mean maximum cibarial tension (1.29 MPa) indicated by blue vertical dashed line with the total range of calculated cibarial tensions (1.06 MPa and 1.57 MPa, respectively) given by the vertical blue bar. (Online version in colour.)

Combining these paired pre- and post-PEG measurements (n = 5) with those made using froghoppers feeding at a constant xylem tension, either on plants growing hydroponically without PEG manipulation (n = 4) or growing in soil (n = 2), gave the following linear regression:

$$\begin{array}{r}M{R}_{\mathrm{cib}}(\mathrm{W})/Q(\mu \mathrm{l}\hspace{0.17em}{\mathrm{s}}^{-1})=0.0041[\mathrm{xylem}\hspace{0.17em}\mathrm{tension}\hspace{0.17em}(\mathrm{MPa})]+\mathrm{0.0002.}\end{array}$$ 2.5

(c) . Morphological measurements

A summary of cibarial pump and cibarial dilator muscle (CDM) morphology is given in table 1. Mean sarcomere length in the CDM was 5.0 ± 0.3 µm (n = 2, 271 sarcomeres total), while the muscle itself contained an average of 23% mitochondria and 63% myofibrils by volume (electronic supplementary material, table S6; n = 2). Using the relationship between maximum specific muscle tension and sarcomere length established by Taylor [16], the specific tension of the CDM was calculated to be 297 kN m⁻²:

$$\mathrm{maximum}\hspace{0.17em}\mathrm{specific}\hspace{0.17em}\mathrm{muscle}\hspace{0.17em}\mathrm{tension}\hspace{0.17em}(\mathrm{kN}\hspace{0.17em}{\mathrm{m}}^{-2})=57.854\times \mathrm{sarcomere}\hspace{0.17em}\mathrm{length}+\mathrm{7.929.}$$ 2.6

Table 1.

Cibarial morphology data from micro-CT scans and calculations of muscle force and cibarial tension.

insect ID	body mass (mg)	cibarium volume (nl)	pennation angle between vectors (deg)	total muscle physiological cross-sectional area (mm2)	CSACibD (mm2)	FCDM (mN)	FCibD (mN)	maximum cibarial tension (MPa)
M1	8.57	25.346	30.830	0.67139	0.13526	199.54	171.34	1.2668
M2	7.72	11.821	38.532	0.49428	0.10820	146.90	114.91	1.0620
M3		19.337	40.989	0.90533	0.16055	269.06	203.10	1.2650
M4	10.08	6.900	30.512	0.59831	0.09755	177.82	153.19	1.5704
mean ± s.d.	8.79 ± 1.20	15.851 ± 8.138	35.216 ± 5.344	0.66733 ± 0.17452	0.12539 ± 0.02831	198.33 ± 51.86	160.64 ± 36.81	1.2911 ± 0.2096

The physiological cross-sectional area of the CDM was determined to be 0.67 mm² (n = 4), which, when multiplied by specific muscle tension, gives a total isometric contraction force parallel to muscle fibre orientation (F_CDM) of 198 mN. As the CDM fibres attach to the apodeme at a pennation angle (θ) of 35.2°, the component of the CDM's force contributing to pulling the apodeme forward (F_CibD) and lifting the cibarial diaphragm is given by

$$F_{\mathrm{CibD}}=F_{\mathrm{CDM}}\times \cos (\theta ),$$ 2.7

where F_CibD is the component of the CDM's contraction force (N) acting on the cibarial diaphragm. And as this force is exerted across the area of the cibarial diaphragm, this yields a maximum tension within the cibarium of 1.3 ± 0.2 MPa (table 1) according to the equation

$$p=\frac{F_{\mathrm{CibD}}}{CS{A}_{\mathrm{CibD}}}.$$ 2.8

3. Discussion

Plants use transpiration to generate high tensions within their xylem sap, harnessing the free water potential gradient between soil and air to drive an energetically favourable flow of liquid [17]. By contrast, froghoppers and other xylem-feeding Hemipterans must actively generate an even greater tension within their cibarium using ATP-powered muscle to extract the liquid mechanically. From our morphological and metabolic analysis of P. spumarius, we show that this insect possesses a cibarium that is eminently capable of generating tensions exceeding those routinely encountered in the xylem vessels of plants, but at a substantial energetic cost.

(a) . Muscle morphology

The bipennate CDM drives P. spumarius's cibarial pump, expanding the volume of the cibarium as it contracts. As low-frequency pumping can occur for days without pause [7], the muscle must have a substantial aerobic energy-producing capacity, as well as the ability to produce high force contractions. Mitochondria and myofibrils make up 23% and 63% of the muscle's volume, respectively, within the range seen in the leg muscles of the cockroach Periplaneta americana, which possess 35% mitochondria by volume in the coxal muscle to less than 10% in the tibial extensor [18]. From this, it may be concluded that CDM is much like other synchronous insect skeletal muscle. However, the sarcomeres in the CDM are longer than those found in the coxal muscles of the cockroach (5.0 µm compared to 3.7–4.2 µm), indicating a far higher capacity for force generation, which at 297 kN m⁻² is almost 50% higher than the approximately 200 kN m⁻² assumed for most muscle [19]. The CDM sarcomeres are certainly not the longest known for an insect, however, being only half the length of those in the adductor muscle associated with the jaws of the trap-jaw ant Myrmoteras (10.3 µm) [20]. These results indicate that the CDM of P. spumarius is adapted for slow contraction and high specific tension.

(b) . Cibarium morphology and maximum tension

The bipennate CDM has its origin on the inner surface of the postclypeus and inserts onto an apodeme that is oriented perpendicular to the cibarial diaphragm (figure 3; electronic supplementary material, figure S1). The CDM pennation angle (θ) of 35.2° is similar to the high pennation angles seen in ant mandible closer muscles specialized for force production (i.e. Atta sextens, 36.9°; Pogonomyrmex badius, 38.4°) [21]. During CDM contraction, the apodeme is pulled anteriorly, exerting a force on the cibarial diaphragm and generating the tension within the cibarial chamber necessary to draw in xylem sap. Based on a study by Ranieri et al. [22], the pressure drop across the stylets of an adult P. spumarius necessary to drive a flow of 12 cm s⁻¹ (the maximum flow velocity based on Q values measured in this study and the smallest P. spumarius stylet CS area from Ranieri et al. [22]) would be less than 0.001 MPa, making this pressure drop negligible compared with the tensions generated within the cibarium, which range from 1.06 to 1.57 MPa (table 1). These values indicate the absolute maximum suction limit of the cibarium, where muscle force on the diaphragm would balance xylem tension within the cibarial chamber. However, when a froghopper is ingesting xylem sap below this limit, each contraction of the CDM extracts an average volume of 0.011 ± 0.002 µl pump⁻¹, independent of xylem tension. This SV is 70% of the mean cibarium volume of 0.0158 µl measured from the micro-CT scans.

Figure 3.

Three-dimensional reconstructions of adult P. spumarius cibarial anatomy from micro-CT scans. (a) Anterior (i), lateral (ii) and dorsal (iii) views of the insect head. (b) Pathway for xylem flow. Xylem sap is drawn up through the stylets (1) held within the labium (2), into the precibarium (3) and then into the cibarium (4) before continuing to the pharynx (6) and oesophagus (7). The cibarial dilator muscles (CDM; 5) fan out from the apodeme located on the dorsal side of the cibarial pump (3). (c,d) Dorsal and lateral view, respectively, of the head with postclypeus anterior and mouthparts posterior. The apodeme movement vector parallel to the apodeme v and muscle contraction vector parallel to the muscle fibres u give the pennation angle θ. CDM, cibarial dilator muscles (purple); apo, apodeme (green); cib, cibarium (yellow). (Online version in colour.)

(c) . Physiological responses to changes in xylem tension

The power (P) of a pump with an efficiency of η is related to the pressure differential Δp across it and its volumetric flow rate Q as

$$P=\frac{\mathrm{\Delta }p\times Q}{\eta }.$$ 3.1

Applying this to the cibarial pump of P. spumarius, P is the net metabolic cost of feeding (W), Q is the xylem excretion rate (m³ s⁻¹) and Δp is equivalent to xylem tension (Pa). Cibarial pump efficiency (η) is a more complicated variable, including the inefficient conversion of metabolic energy (P, measured using respirometry) into mechanical work by the muscle, and the relationship between muscle's contraction velocity, frequency and pennation angle. Assuming that η is constant, as xylem tension increases Δp, there must be corresponding changes in P and/or Q. The froghoppers showed no significant change in Q with increasing xylem tension, as has been observed in previous experiments on this species [13]. The pre- and post-PEG averages of Q were 0.57 µl min⁻¹, although individual excretion rates reached upwards of 1.0 µl min⁻¹. These values agree well with those from studies on adult P. spumarius, with average rates varying from 0.43 µl min⁻¹ [23] up to 1.2 to 1.5 µl min⁻¹ [5,13]. Malone et al. [13] also observed that excretion rates were unchanged even when xylem tensions increased up to 1 MPa.

While the rate of xylem excretion did not change significantly with an increase in Δp, a small but significant increase in f_pump (approx. 10%) was observed after the introduction of PEG. The pre- and post-PEG average f_pump values were 0.89 Hz and 0.97 Hz, respectively, with the highest observed frequency being 1.3 Hz. Again, these values fall within the range of published f_pump values measured from P. spumarius (0.7 to 1.7 Hz [12,13]).

Although f_pump showed a significant increase with Δp, the ratio of Q to f_pump (i.e. SV) did not. This could indicate that the increase in f_pump, although significant, was not particularly large given the variability in excretion rate. Having no significant change in Q or the ratio of Q/f_pump, but a significant increase in f_pump suggests that the relationship between these variables is fairly weak. Given that Q did not change with Δp, according to equation (3.1), this can only leave a compensatory increase in P.

(d) . Metabolic cost of xylem extraction

Increasing xylem tension was associated with a significant increase in the froghoppers' cibarial metabolic rate (MR_cib). Feeding froghoppers exhibited an average pre-PEG MR_cib of 25.5 µW (4.25 µl O₂ h⁻¹) and an average post-PEG MR_cib of 39.3 µW (6.56 µl O₂ h⁻¹), representing a 50 to 85% increase in their total MR compared to their RMR. The energy expended per unit of xylem volume extracted (MR_cib/Q; J µl⁻¹) also showed a significant increase with xylem tension, increasing at a rate of 0.0041 J µl⁻¹ MPa⁻¹ (from the regression line in figure 2a, J µl⁻¹ = 0.0041 MPa + 0.0002).

(e) . Energetics of xylem feeding

Xylem sap is the least nutritional fluid present in plants [2] containing low concentrations of amino acids and some sugars [4,23–25]. There is substantial variation in the volumetric energy density of xylem sap (ε_xylem) as calculated assuming complete combustion of its organic molecules. Values of ε_xylem from trees vary from 3.89 ± 1.38 J cm⁻³ (peach Prunus persica) to 24.66 ± 5.11 J cm⁻³ (crape myrtle Lagerstroemia indica) [9], while the ε_xylem of herbaceous tomato plants (Lycopersicon sp.) ranges from 12.3 to 39.9 J cm⁻³ [23]. Without knowing the metabolic efficiency of a feeding froghopper, these values still demonstrate that for many plant species, more energy can be obtained per unit of extracted xylem than is spent on its extraction across the entire range of xylem tensions accessible to the insect's cibarial suction (figure 2a). For example, a froghopper with a Q of 0.01 µl s⁻¹ feeding on a tomato plant with an average ε_xylem of 26.1 J cm⁻³ would yield an energy extraction rate of 0.26 mW. Assuming the plant's xylem tension was 1.1 MPa, using the regression in figure 2a to determine energy expenditure per volume of xylem extracted (J ul⁻¹), and multiplying by Q, gives a xylem extraction cost of 0.05 mW. This, combined with an average RMR of 0.04 mW gives a total energy expenditure rate of 0.09 mW for the whole feeding insect, equivalent to 34% of the total energy present in the extracted xylem sap. Even with a metabolic efficiency of 40%, the extracted xylem would still satisfy 118% of the insect's total energy requirements. Lower xylem tensions would only improve the insect's net energy gain. By contrast, energy loss is predicted for insects feeding on P. persica at comparatively low xylem tensions (figure 2a). Plant species with low ε_xylem, including P. persica, are avoided by xylem-feeding H. vitripennis [26], demonstrating that some plants possess combinations of ε_xylem and xylem tension that are energetically unviable for xylem-feeding insects [27].

(f) . Cibarial pump efficiency

For a pump with a known P, Q and Δp, it is possible to calculate its efficiency η from equation (3.1) as

$$\eta =\frac{\mathrm{\Delta }p\times Q}{P}.$$ 3.2

The overall efficiency of a froghopper's cibarial pump is likely determined by the relative inefficiency of the muscle powering it. Applying equation (3.2) to simultaneous measurements of MR_cib, excretion rate and xylem tension gives an average efficiency of 33.2 ± 19.1% and 29.7 ± 15.2%, pre- and post-PEG, respectively (n = 6). With minimal change in CDM contraction velocity, as indicated by relatively constant f_pump and Q values, the small difference between these values is unsurprising. While the efficiency of synchronous and asynchronous insect flight muscle has been determined to lie between 10 to 16% [28,29], to our knowledge there are no published values for insect skeletal (non-flight) muscle efficiency. However, in vitro estimates of vertebrate locomotory muscle give maximum peak efficiencies of 25% without pre-stretching the muscle, and up to 50% if the muscle is pre-stretched [30]. Thus, calculated values here are biologically reasonable.

(g) . Biological pressure probe

Measuring xylem tension in situ using an implanted pressure probe is technically challenging, and this technique is limited to measuring tensions below 1 MPa due to cavitation issues [15]. Thus, the evidence for xylem tensions greater than 1 MPa relies on measurements made on excised plant tissue using a pressure bomb. But as the morphological data here show, P. spumarius has evolved the capacity to generate tensions potentially up to 1.6 MPa (table 1), indicating that they encounter and ingest xylem sap with tensions up to this level in situ. Furthermore, the calculated mean cibarial pump tension of 1.3 MPa exceeds greater than 97% of daytime xylem tensions that P. spumarius encounter while feeding, as measured from their host plants in the field using a pressure bomb (figure 2b). These xylem tensions, and their relationship to calculated cibarial tension, expand the range of confidence we can have in the accuracy of the pressure bomb method for measuring xylem tensions.

(h) . Implications for other xylem feeders

Adult P. spumarius were observed feeding on plants with xylem tensions up to 1.1 MPa, while tensions greater than 1.5 MPa appeared to hinder the froghopper's ability to ingest the xylem sap, as excretion stopped at this tension (E.A.B., E.L.G. & P.G.D.M. 2019, personal observation). However, xylem-feeding insects that have adapted to feed on trees or plants in arid zones likely possess an even greater capacity for generating tension to match the higher xylem tensions in their host plants. But the ability to feed successfully under these conditions will depend not only on the force-producing capacity of their CDM relative to the cross-sectional area of their cibarial diaphragm, but also on the ε_xylem of the host plant; if the tension of xylem sap increases without a concomitant increase in its energy density, the net energy balance of the feeding insect must eventually become negative.

4. Material and methods

(a) . Animals

(i) . Spittlebug nymphs and froghopper adults

Both spittlebug nymphs and adult froghoppers of P. spumarius were collected from around UBC Point Grey campus, from early May 2019 until early October. They were maintained in the laboratory on pea, alfalfa or fava bean plants within an aluminium mesh cage that received natural light through a window and was kept at ambient laboratory temperatures (20–25°C). Spittlebugs were kept until they moulted into adult froghoppers, at which point they were used for experiments.

(ii) . Body mass

All froghoppers were weighed to 0.01 mg on an electronic balance (XPE205 DeltaRange, Mettler Toledo, Greifensee, Switzerland) after being used in an experiment.

(b) . Plants

Plants used in experiments were grown and maintained according to protocols described in the electronic supplementary material.

(c) . Transmission electron microscopy

The CDM was fixed in situ, post-fixed with osmium tetroxide in sodium cacodylate, embedded in epoxy resin then sectioned using an ultra-microtome. These sections were post-stained with uranyl acetate and lead citrate, imaged using TEM and analysed. Full details are provided in the electronic supplementary material.

(d) . Micro-CT scans

P. spumarius froghoppers were decapitated and their heads placed in fixative before being transferred into 70% EtOH for storage. The fixed heads were then stained for 7 days in 1% phosphotungstic acid in 70% EtOH as a contrasting agent and scanned at a resolution of 3.0 µm using a three-dimensional X-ray microscope (Xradia Versa 520, Zeiss, Oberkochen, Germany). Scans were compiled and turned into three-dimensional reconstructions using Dragonfly image analysis software (v. 4.5.0.711, Object Research Systems Inc., Montreal, QC, Canada), before relevant morphological features were measured as described in the electronic supplementary material.

(e) . Respirometry trials

(i) . Measuring metabolic rate

Flow-through respirometry measuring VCO₂ was used to determine the MR of resting and feeding froghoppers. Froghoppers were recorded feeding on hydroponically grown plants before and after exposure to PEG to change the plant's xylem tension during feeding, as well as while feeding on hydroponically grown and soil-grown plants without xylem tension manipulation. A full description of the method, calculations and procedure is given in the electronic supplementary material.

(ii) . Measuring xylem tension

The xylem tensions of plants in the laboratory and the field were measured on excised leaves and stems using a Scholander–Hammel pressure bomb (Model 615, PMS Instrument Company, OR, USA) following standard procedures as described in the electronic supplementary material.

(f) . Measuring feeding parameters

Cibarial pumping frequency was determined by analysing video footage of actively feeding froghoppers enhanced using Eulerian video magnification software (electronic supplementary material, movie S2), while xylem excretion rate was determined from the diameter of excreted droplets measured from still frames extracted from this same video following the procedure given in the electronic supplementary material.

(g) . Statistical analyses

Data were analysed in R v. 3.5.1 [31] in RStudio v. 1.1.463 (RStudio, Inc., 2009–2018). Overall means for the pre-and post-PEG values of f_pump, excretion rate and MR_cib are all given ± s.d. Overall means for sarcomere length are also given ± s.d. The feeding parameters of f_pump, Q, the ratio of Q/f_pump, MR_cib and the ratio of MR_cib/Q were all analysed as response variables using linear mixed-effects models, with xylem tension as the predictor variable and the random effect corresponding to individual insects, using the nlme (v. 3.1-137) and lme4 (v. 1.1.21) packages [32,33]. The pre-PEG xylem tensions used in these models were the average values taken from non-manipulated plants of the same species, while post-PEG xylem tensions were those measured from the plant used in the feeding trial after PEG exposure.

Data accessibility

The data are provided in the electronic supplementary material [34].

Authors' Contributions

E.A.B.: conceptualization, data curation, formal analysis, investigation, methodology, validation, visualization, writing-original draft, writing-review and editing; E.L.G.: formal analysis and methodology; P.G.D.M.: conceptualization, funding acquisition, methodology, supervision, writing-original draft, writing-review and editing

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Competing interests

We declare we have no competing interests.

Funding

This work was supported by the Natural Sciences and Engineering Research Council of Canada: [Discovery grant nos: RGPIN-2014-05794 and RGPIN-2020-07089, Accelerator supplement: RGPAS 462242-2014 to P.G.D.M., Undergraduate Student Research Award to E.L.G.].