Skip to main content
NCBI home page
Journal of Insect Science logoLink to Journal of Insect Science
. 2020 Jul 17;20(4):7. doi: 10.1093/jisesa/ieaa056

Do Sharpshooters From Around the World Produce the Same EPG Waveforms? Comparison of Waveform Libraries From Xylella fastidiosa (Xanthomonadales: Xanthomonadaceae) Vectors Kolla paulula (Hemiptera: Cicadellidae) From Taiwan and Graphocephala atropunctata From California

Elaine A Backus 1,✉, Hsien-Tzung Shih 2
Editor: Phyllis Weintraub
PMCID: PMC7366069  PMID: 32677683

Abstract

When an exotic invasive species is a vector-borne plant pathogen, vector feeding behavior must be studied to identify potential host plant range and performance of specialized pathogen transmission behaviors. The most rigorous tool for assessing vector feeding behavior is electropenetrography (EPG). Xylella fastidiosa Wells et al. is a gram-negative bacterium native to the Americas, where it is the causal agent of lethal scorch-type diseases such as Pierce’s disease (PD) of grapevines. In 2002, a PD strain of X. fastidiosa invaded Asia for the first time, as confirmed from grape vineyards in Taiwan. Kolla paulula (Wallker), a native Asian species of sharpshooter leafhopper, was found to be the primary vector in Taiwanese vineyards. This study used an AC-DC electropenetrograph to record stylet probing behaviors of K. paulula on healthy grapevines. The main objective was to create an EPG waveform library for K. paulula. Waveform description, characterization of R versus emf components (electrical origins), and proposed biological meanings of K. paulula waveforms are reported. In addition, comparison of K. paulula waveforms with those from the most efficient, native vector of X. fastidiosa in California vineyards, Graphocephala atropunctata, is also reported. Overall, both species of sharpshooters had similar-appearing waveforms. Five new findings were identified, especially that the previously described but rare waveform subtype, B1p, was extensively produced in K. paulula recordings. Sharpshooter waveforms from species worldwide share a high degree of similarity. Thus, EPG methods can be rapidly applied to potential vectors where X. fastidiosa is newly introduced.

Keywords: applied signal, stylet probing, stylet penetration, electrical penetration graph, xylem ingestion

A nation’s agricultural biosecurity in the age of global commerce depends upon using all tools available to rapidly assess threats from exotic, invasive pests. When the invader is a vector-borne plant pathogen, it is important to assess both the feeding behavior of its vector and the degree to which specialized pathogen transmission behaviors are performed on those plants (Sandanayaka et al. 2017). For hemipteran vectors, the most rigorous tool for observing and quantifying their otherwise invisible feeding behaviors is electropenetrography (EPG) (McLean and Kinsey 1964, Tjallingii 1978, Backus and Bennett 2009). EPG has become vital to assess both the range of acceptable host plants (Sandanayaka et al. 2007, 2012, 2017; Sandanayaka and Backus 2008) and transmission behaviors (Fereres 2007, Backus 2016) of invasive vectors and/or pathogens. However, the key to rapidly applying EPG to invasive vectors and pathogens is to understand how to interpret the vector’s EPG waveforms.

One of the most destructive invasive plant pathogens in the world is Xylella fastidiosa Wells et al., the causative agent of bacterial scorch diseases such as Pierce’s disease (PD) of grapevines in North America and citrus variegated chlorosis (CVC) in South America (Purcell 1997). Xylella fastidiosa can be transmitted without strain specificity by many vector species, primarily sharpshooter leafhoppers (Cicadellidae: Cicadellinae) and spittlebugs (Aphrophroidae) (Almeida 2016b). Xylella fastidiosa can arrive inside invasive vectors like the glassy-winged sharpshooter, Homalodisca vitripennis (Germar), as occurred in Hawaii and French Polynesia (Grandgirard et al. 2006, Mizell et al. 2008). More commonly, however, X. fastidiosa invades via contaminated host plant material; subsequently, the bacterium co-opts competent native vectors in the invasion area. The latter scenario recently was recognized in Italy, where X. fastidiosa subsp. pauca str. ST53 was introduced sometime before 2013 in plant(s) from Costa Rica (Almeida 2016a). The bacterium was acquired by local meadow spittlebugs, Philaenus spumarius L., which then inoculated it to olive trees (Almeida 2016a, Cornara et al. 2017, Giampetruzzi et al. 2017). The resulting epidemic of olive quick decline disease has now killed millions of olive trees in Italy; X. fastidiosa also is spreading to other areas of Europe and other crops (Almeida 2016a).

Before both the Pacific and Italian introductions, the first invasion of X. fastidiosa into Asia occurred in Taiwan. Symptoms of PD were first seen in grape vineyards of Nantou County in 2002, and later also in vineyards in Taichung City and Miaoli County (Su et al. 2013a, b). Subsequently, X. fastidiosa subsp. multiplex str. Temecula (a PD strain in California) was isolated and identified (Su et al. 2013b). Two native sharpshooter species, Kolla paulula (Walker) and Bothrogonia ferruginea (F.) were identified as vectors (Tuan et al. 2016), among which K. paulula is the dominant species found in PD infection areas (Shih et al. 2013). These endemic Taiwanese sharpshooters transmit X. fastidiosa when they disperse from local riparian plants to grapevines in nearby vineyards (Shih et al. 2009, 2013), ecologically similar to the native California vector, blue-green sharpshooter, Graphocephala atropunctata (Say) (Purcell 1997, Redak et al. 2004). Because grapes are a high-value, luxury commodity in Taiwan, economic impact of PD has been significant for the island’s economy. The average annual production of grapes in Taiwan has reached 93,000 metric tons with an average price of US$2.4 per kg (AG Statistics Yearbook 2018, Agriculture and Food Agency, COA, Executive Yuan, Taiwan) over the past decade. During the period 2003–2012, 1,367 grape vines infected with PD of grape were destroyed in Taiwan (Su et al. 2013a), with a total economic loss of at least US$2.88 million.

There are no records of host plants for K. paulula in historical scientific papers. Therefore, feeding and survivorship tests of host plants of K. paulula were performed based on the definitions made by Oman (1949) (H.-T. Shih, unpublished data); however, such a verification process is time-consuming. To shorten the host verification time and understand the correlation between feeding behavior and pathogen transmission efficiency on different grape varieties, AC-DC EPG (Backus and Bennett 2009) was used to establish the feeding behaviors of K. paulula on different host plants. As explained in our recent review article (Backus and Shih 2020), the AC-DC electropenetrograph is an improvement in EPG technology with flexible settings to tailor the instrument to any type of arthropod regardless of size or type of feeding. It is recommended that this review paper (Backus and Shih 2020) be read before the present paper, for necessary background on EPG history and principles, especially electrical origins of waveforms, the R and emf components, and also background on sharpshooter waveforms.

As X. fastidiosa spreads around the world, there is a need to know whether local, noncosmopolitan sharpshooter species (especially those in Asia) exhibit the same feeding behaviors, waveforms, and likely X. fastidiosa inoculation behaviors as do the better-studied New World sharpshooters (Miranda et al. 2009, Cervantes and Backus 2018). Such knowledge would allow EPG to immediately be used to assess stylet probing and bacterial transmission behaviors, as well as search for novel management methods for the introduction are, such as choice of insecticides and development of host plant resistance.

As explained in Backus and Shih (2020), when the flexible amplifier sensitivities of the AC-DC electropenetrograph are used to compare electrical origins of waveforms, a library of waveform appearances can be prepared to speed up the process of identifying biological meanings of waveforms. The present project builds upon and expands the G. atropunctata project of Cervantes and Backus (2018) to produce the first waveform library for any Asian hemipteran species. Accordingly, the objectives of this project were to: 1) record EPG waveforms for the first time for K. paulula using an AC-DC electropenetrograph; 2) develop a waveform library, to determine degree of similarity with other sharpshooter EPG waveforms; 3) display sharpshooter waveforms from the gold wire loop + silver glue wiring method at Ri levels 106 and 108 Ohms (not previously done for Cervantes and Backus 2018); 4) identify any new or modified waveforms seen with K. paulula and hypothesize their biological meanings based on their electrical origins; and 5) identify the best electropenetrograph settings for future, quantitative comparison studies of K. paulula feeding on diverse host plants in a biosecurity and host plant resistance context.

It is not the intent of this paper to repeat all the detailed information found in the waveform library of G. atropunctata (Cervantes and Backus 2018), but instead to provide illustrations of the K. paulula waveform library and describe novel EPG findings with this new X. fastidiosa vector species, by comparison with G. atropunctata. Results showed that the overall waveform appearances for K. paulula are very similar to those of G. atropunctata, and can be interpreted using the existing sharpshooter waveform names. Nonetheless, five new findings were identified for K. paulula compared with G. atropunctata.

Materials and Methods

Plant and Insect Rearing

Kolla paulula is commonly found in low-altitude areas of Taiwan. Sharpshooters were originally field-collected from host plants Commelina communis Linn. and Bidens pilosa L. var. radiata in Douliu City, Yunlin County, Taiwan. They were brought to the Taiwan Agricultural Research Institute (TARI) in Wufeng District, Taichung, Taiwan, where they were tested using the methods of Su et al. (2013b) to ensure they were Xylella-free. A reproducing colony was developed and caged in BugDorm-44545 Insect Cages (45 × 45 × 45 cm) (MegaView, Taichung, Taiwan) on C. communis in an indoor rearing room under artificial lights. Like G. atropunctata and other sharpshooters, K. paulula does not reproduce on grapevines, but prefers wild, riparian plants like B. pilosa for oviposition as well as feeding (Redak et al. 2004, Krugner et al. 2019). A photoperiod of 13:11 (L:D) h and temperature ~25°C were maintained year-round. All insect colony plants were grown from seed in a greenhouse at TARI under natural light supplemented with artificial lights to maintain a 13:11 (L:D) h photoperiod.

Insect Wiring

Adult K. paulula were removed directly from colony rearing cages with no conditioning period on grapevines, similar to the protocol of Cervantes and Backus (2018). Very long recordings (18–24 h) allowed researchers to capture some of the behavioral adaptation period from native to naïve plant, in order to maximize the types of waveforms that might be produced and include them in the waveform library. Insects were enclosed in plastic vials and then put on ice for 1 min before wiring. Cold insects were immobilized further by low vacuum under a dissecting microscope (Wild MSA, 20497, Heerbrugg, Switzerland) for placement of the gold wire, which was 18 µm in diameter (EPG Systems, Inc., Wageningen, The Netherlands). A tiny loop was formed at the tip of the wire before it was dipped in silver glue and applied to the dorsum of each sharpshooter (see Cervantes and Backus 2018 for glue recipe and detailed methods). Note that, unlike for Cervantes and Backus (2018), the same (best) adhesive (silver glue) was used for all recorded insects in the present study.

EPG Methods

A four-channel, AC-DC electropenetrograph (EPG Technologies, Inc., Gainesville, FL; to purchase an AC-DC electropenetrograph, contact the CEO of EPG Technologies, Andrew Dowell, at andygator3@gmail.com) was used to record feeding waveforms. This instrument is similar to but simpler than the AC-DC correlation monitor (Backus and Bennett 2009); it differs by requiring choice of AC or DC applied voltage (not both, mixed) and simultaneous recording of four insects (instead of one). Backus and Shih (2020) have more details on design, as well as a block diagram of signal processing for the four-channel instrument.

Two grape plants (Vitis vinifera cv. ‘Kyoho’) each were placed into two Faraday cages, and a second or third true leaf from each plant was presented to each of four insects by attaching each leaf to a horizontal Plexiglas plate with Parafilm as described in Cervantes and Backus (2018). Only AC applied voltage was used for this experiment because we were primarily interested in waveform appearance, not effects of AC versus DC on the insects. We preferred to use AC because DC applied signals can have a deleterious effect on sharpshooter stylet probing behavior compared with AC applied signals (Cervantes and Backus 2018), similar to other large insects (Backus et al. 2018). Previous studies (Backus and Bennett 2009, Pearson et al. 2014, Cervantes and Backus 2018) have shown that the waveforms generated by the AC-DC electropenetrograph are virtually identical (as the instrument was designed to do) for either AC or DC applied signal when the same input resistor (Ri) level (a.k.a. amplifier sensitivity) and applied voltage level are used. Indeed, occasional switches of applied voltage from AC to DC showed no difference in waveform appearance or fine structure. Recordings were acquired and digitized using WinDaq Pro+ software and a DI-720 analog-to-digital board and a sample rate of 100 Hz per channel. Waveforms were displayed using WinDaq Waveform Browser (all from DATAQ Instruments, Akron, OH). The instrument gain for all recordings was 4,000× and WinDaq gains ranged from 2× to 64×.

In all recordings, pre- and post-rectification signals were simultaneously recorded to check for signal inversion caused by the rectifier. The offset knob was adjusted as needed, to ensure inverted signals were reverted. Thus, baseline voltages were determined via the pre-rectification output, but figures were derived from post-rectification output signals. Further explanation of the offset function is available in Backus and Shih (2020).

Experimental Design and Statistical Analysis

Four wired insects were recorded each day on Kyoho grape. One of two experimental designs was used for each daily recording. In the first design, either two or four channels were set with one of the four Ri levels compared for this study (i.e., 106, 107, 108, or 109 Ohms), and insects were recorded for 20 h each day. For the second design, each channel was set with a different Ri level so that all four chosen Ri levels were represented within one experiment. Insects were recorded for 18–24 h/day.

Applied signal voltage was deliberately kept low (serendipitously 15, 20, or 25 mV), to eliminate potential negative voltage effects (which usually occur only above 250 mV AC) (Backus et al. 2018, Cervantes and Backus 2018). Total number of insects at each Ri and applied voltage level was: nine insects each at Ri’s 106 and 107, eight insects each at Ri’s 108 and 109. Despite these relatively small sample sizes, our long-duration recordings produced sufficient numbers of probes for visual analysis of waveform appearances and statistical analysis of counts. Logistic regression was used to compare the proportion of insects whose probes were biphasic (dependent variable) in relation to applied voltage level and Ri level using JMP (SAS interactive statistical discovery software, Cary, NC). Means were considered significantly different at α = 0.05.

Waveform Naming Convention

The Backus convention for sharpshooter waveform names was exclusively used herein. This convention is explained in detail in Backus and Shih (2020) and Cervantes and Backus (2018).

Results

General Overview of K. paulula Waveforms

EPG waveforms for K. paulula were very similar to those seen for G. atropunctata (Cervantes and Backus 2018, especially table 1 therein) within the same Ri levels. However, because the same wiring method (gold loop + silver glue) was used for all insects in the present study (and not in Cervantes and Backus 2018), less adhesive-caused variability was seen among recordings (especially at high Ri levels) so that differences could be better-attributed to Ri level and/or applied voltage effects. At very low Ri levels, quality of wiring also made some difference in appearances. Nonetheless, more detailed comparisons can be made among Ri levels for K. paulula than for G. atropunctata.

Waveform Polarity and Voltages

In all EPG studies, overall waveforms within a probe can be: 1) biphasic, with most peaks being positively oriented and waveforms occurring both above and below baseline level, 2) monophasic positive, waveforms located above baseline with positively oriented peaks, or 3) monophasic negative, waveforms located below baseline with negatively or positively oriented peaks (Backus and Shih 2020).

Cervantes and Backus (2018) found that silver glue caused most G. atropunctata probes to be biphasic, probably because it is more electrically conductive. However, the present study with K. paulula showed that probes could be either biphasic or monophasic positive when attachment was standardized using silver glue, depending primarily upon applied voltage level and secondarily on Ri level. All probes recorded for each insect would have one or the other polarity. Using 15 mV AC applied voltage, 0% of insects produced biphasic probes, regardless of Ri level. When 20 mV AC was used, Ri levels of 106 through 108 Ohms had a total of 33% (4/12) of insects producing biphasic probes, whereas neither insect at 109 Ohms made biphasic probes. Percentage of insects making biphasic probes increased dramatically at 25 mV, when 80% (8/10) produced biphasic probes at Ri 106 through 108 Ohms, but one out of two was biphasic at 109 Ohms. Despite the low sample size, logistic regression showed that the proportion of insects with biphasic probes was dependent on applied voltage level (df = 3; χ 2 = 13.16; P = 0.0014); Ri level and interaction effects were not significant. Nonetheless, there was a trend that higher applied voltage stimulated biphasic polarity as Ri level was increased. Other than these polarity effects, no other changes in waveform appearance were due to voltage level. Most changes were caused by differences in proportion of R to emf, from Ri levels.

Very few probes were monophasic negative. When probes were biphasic, waveforms were always above the baseline in the beginning, then dived below the baseline level during pathway phase (described below), only to rise abruptly or gradually through baseline level into positive level during X wave phase (see below) (Fig. 1), stabilizing at a flat voltage level above baseline during sustained ingestion (data not shown). Most waveform figures presented herein show monophasic positive probes because biphasic probes had such steeply falling voltage levels that detailed views could not be contained in a single, page-sized figure. Only low-amplitude, compressed, coarse-structure views could be thusly contained (Fig. 1). Even with monophasic positive probes, there was still enough voltage level variation that some figure sections had to be deamplified to fit in the space allowed (see below).

Fig. 1.

Fig. 1.

Open in a new tab

Highly compressed examples of pathway and X wave phases of biphasic probes by K. paulula at the respective Ri levels tested. Applied signals were 25 mV AC for all four input resistor (Ri) levels. WinDaq gain and time scale are shown in the first part, for 106 Ohms; the same gains and scales also were used for the other three parts. Dashed gray line, baseline.

Relative voltage levels varied with Ri level, as with G. atropunctata. Thus, at low Ri levels, early peaks (especially family A waveforms, see below) were irregularly high to very high, while at 109 Ohms, family A waveforms were reduced in height; overall voltage level of the entire probe at high Ri levels was stereotypically quite flat.

Variability in Waveform Appearances

Waveform appearances were strikingly variable from insect to insect at low Ri levels (106 and 107 Ohms) despite exclusive use of silver glue. Therefore, the extremes in a continuum of appearances are displayed in two figures for each low Ri level. In contrast, waveform appearances were much less variable for 108 Ohms, and highly stereotypical for 109 Ohms. Therefore, each has only one figure. All figures are cited and described under Pathway Phase, below. This topic will be revisited (below) after individual waveform appearances are summarized, next.

Pathway Phase

Pathway extends from initial stylet insertion until stylet withdrawal (as with test probes, below) or X wave phase, whichever comes first. In a complete (or exploratory or ingestion) probe, Pathway is composed of two families, A and B, each with two types and several subtypes in K. paulula, the same as G. atropunctata and other sharpshooters.

New waveform family, T

Like all sharpshooter leafhoppers, K. paulula made test probes of very short duration (a few seconds) upon first arrival on a plant it had not previously experienced; these were performed before the insect began an exploratory probe, progressing into an ingestion probe. In the past, test probes have been called a special circumstance of waveform family A. However, recordings at different Ri levels with K. paulula showed strong variation in the appearance of test probes with Ri level (Fig. 2). At the lowest Ri levels, 106 and 107 Ohms, most test probes resembled a spikier version of A1 and A2 (Fig. 2, first two panels). However, this detail was usually partially (108 Ohms) to wholly (109 Ohms) lost at higher Ri levels, so that test probes became quite nondescript in K. paulula recordings. This observation warranted a new waveform name, family T, so that test probes could be distinguished from longer exploratory/ingestion probes during measurements for quantitative EPG studies.

Fig. 2.

Fig. 2.

Open in a new tab

Representative appearances of test probes (one per panel) by K. paulula at the respective Ri levels tested. Applied signals were 20 mV AC for all four input resistor (Ri) levels. WinDaq gain and time scale are shown in the first part, for 106 Ohms; the same gains and scales also were used for the other three parts. Dashed gray line, baseline.

Family A. Type A1

Kolla paulula A1 was very similar to that of G. atropunctata because it comprised one or two variable-width peaks or plateaus at the beginning of a probe, often composed of two vertical sections (Cervantes and Backus 2018). At 106 Ohms, A1 can range from moderately tall compared with short C2 (Fig. 3a), to very tall compared with very short C2 (Fig. 4a). At higher Ri levels, A1 can be so short that it is very difficult to distinguish from A2 (Fig. 5b), or be quite indistinct (Fig. 7b), or commonly even disappear altogether (Figs. 6b and 8b), especially compared with tiny C2 amplitudes. Thus, A1 has a strong R component.

Fig. 3.

Fig. 3.

Open in a new tab

Representative waveforms for part of the recordings at Ri = 106 Ohms, using applied signal of 15 mV AC. a. Overview of pathway and X wave phases. Phase-level names are along the top of the label bar, while family-level names are along the bottom of the label bar. b–g. Enlargements of boxes b–g in part a. Family-type level names are along the top of the label bar, while type- or subtype-level names are along the bottom of the label bar, respectively. Subtype names may also be near waveform images with vertical lines separating waveforms. Downward-pointing arrowheads denote vd’s. WinDaq gains and time scales shown in most parts. Gains and scales for parts c and d same as part b; likewise, for parts e and g same as part f.

Fig. 4.

Fig. 4.

Open in a new tab

Representative waveforms for part of the recordings at Ri = 106 Ohms, using applied signal of 25 mV AC. a. Overview of pathway and X wave phases. Phase-level names are along the top of the label bar, while family-level names are along the bottom of the label bar. b–d. Enlargements of boxes b–d in part a. Family-type level names are along the top of the label bar, while type- or subtype-level names are along the bottom of the label bar, respectively. Subtype names may also be near waveform images with vertical lines separating waveforms. e. Waveforms from later in X wave phase, not shown in part a. Upward-pointing arrowheads denote broad A2 peaks. WinDaq gains and time scales shown in most parts. Gains and scales for parts c–e same as part b.

Fig. 5.

Fig. 5.

Open in a new tab

Representative waveforms for part of the recordings at Ri = 107 Ohms, using applied signal of 20 mV AC. a. Overview of pathway and X wave phases. Phase-level names are along the top of the label bar, while family-level names are along the bottom of the label bar. b–e. Enlargements of boxes b–e in part a. Family-type level names are along the top of the label bar, while type- or subtype-level names are along the bottom of the label bar, respectively. Subtype names may also be near waveform images with vertical lines separating waveforms. f. Waveforms from later in X wave phase, not shown in part a. Downward-pointing arrowheads denote vd’s. WinDaq gains and time scales shown in most parts. Gains and scales for parts c–f same as part b.

Fig. 7.

Fig. 7.

Open in a new tab

Representative waveforms for part of the recordings at Ri = 108 Ohms, using applied signal of 20 mV AC. a. Overview of pathway and X wave phases. Phase-level names are along the top of the label bar, while family-level names are along the bottom of the label bar. b–g. Enlargements of boxes b–g in part a. Family-type level names are along the top of the label bar, while type- or subtype-level names are along the bottom of the label bar, respectively. Subtype names may also be near waveform images with vertical lines separating waveforms. Downward-pointing arrowheads denote vd’s. WinDaq gains and times scales shown in most parts. Gains and scales for parts c–g same as part b.

Fig. 6.

Fig. 6.

Open in a new tab

Representative waveforms for Ri = 107 Ohms, using applied signal of 20 mV AC. a. Overview of pathway, X wave, and return to pathway phases. Phase-level names are along the top of the label bar, while family-level names are along the bottom of the label bar. b–d. Enlargements of boxes b–d in part a. Phase- or family-type level names are along the top of the label bar, while type- or subtype-level names are along the bottom of the label bar, respectively. Subtype names may also be near waveform images with vertical lines separating waveforms. WinDaq gains and time scales shown in most parts. Gains and scales for parts c and d same as part b.

Fig. 8.

Fig. 8.

Open in a new tab

Representative waveforms for part of the recordings at Ri = 109 Ohms, using applied signal of 15 mV AC. a. Overview of pathway and X wave phases. Phase-level names are along the top of the label bar, while family-level names are along the bottom of the label bar. b–g. Enlargements of boxes b–g in part a. Family-type level names are along the top of the label bar, while type- or subtype-level names are along the bottom of the label bar, respectively. Subtype names may also be near waveform images with vertical lines separating waveforms. B1w/s means alternating B1w and B1s. h–j. XN segments from later in X wave phase, not shown in part a. WinDaq gains and times scales shown in most parts. Gains and scales for parts c–j same as part b.

Family A. Type A2

Kolla paulula A2 was a shorter-amplitude series of peaks directly following A1, similar to that of G. atropunctata. However, K. paulula A2 more frequently had brief voltage drops (vd’s), sometimes many (Figs. 3b and 5b, downward-pointing arrowheads) compared with G. atropunctata. Sometimes the vd’s were so regularly spaced that the waveform resembled rounded, tightly assembled plateaus with little rise in relative amplitude (Fig. 5b). The largest number of vd’s occurred at lower Ri levels, with diminishing numbers at higher Ri levels, especially 109 Ohms. This supports that vd’s in K. paulula are R-dominated. Some vd’s also could occur during B1 (below), but much less frequently than during A2.

Family B. Type B1

Like with G. atropunctata, B1 always followed family A in K. paulula. B1 in G. atropunctata almost always was composed of only two subtypes: B1w (flat wavelets) and B1s (spikelet bursts) long trains of alternating B1w/B1s comprised the bulk of pathway phase in that species and most other sharpshooters including K. paulula (best seen herein in Fig. 8b and c). Because they are visible at all Ri levels, B1s/B1w are composed of mixed R and emf components in G. atropunctata (Cervantes and Backus 2018), and the same is true in K. paulula.

Graphocephala atropunctata very rarely produced a third subtype, B1p, i.e., a few, tiny peaks about the same amplitude as B1w (and usually using silver paint as the adhesive at higher Ri levels). However, in striking contrast, B1p was a major waveform in K. paulula, often representing about 25–50% with sometimes as much as 70% of B1. At low Ri levels, B1p peaks were clearly distinct from and interspersed with B1s spikelet bursts, with B1w in between the peaks and bursts (Figs. 3c, d and 6b, c). At higher Ri levels, B1s and B1w sometimes merged and became less distinguishable (although not always, as shown in Fig. 8c and d; if B1s/B1w were indistinguishable, we often used the type name B1, rather than dividing it into the subtypes). In those cases, B1p became more prominent, with peaks increasing in amplitude, frequency, and duration (Figs. 5b–d and 6d). At 109 Ohms, B1p became so dominant that long trains of peaks of increasing amplitude would take up much of pathway phase leading right up to the start of X wave phase (Fig. 9), and often it was difficult to distinguish between B1p and XC1 (see below). Accordingly, we concluded that, in K. paulula recordings, detecting alternating B1s/B1w is dependent on R being present (thus, Ri lower than 109 Ohms), while detecting B1p was highly dependent on emf (109 Ohms).

Fig. 9.

Fig. 9.

Open in a new tab

Example of a long pathway phase at Ri = 109 Ohms, using applied signal of 25 mV AC, showing extreme predominance of B1p with gradually increasing amplitude of peaks appearing to evolve into XC1. Waveforms in each part follow directly from those in the part above, as indicated by bold horizontal arrows. *Amorphous B1 best labeled at that type level, not divided into subtypes. WinDaq gains and time scales shown in the top part also apply to other parts.

Family B. Type B2

This waveform type was a stereotypical series of short, triangular peaks that formed a distinct crescent, easily distinguished from B1 (Fig. 5c), and nearly identical to the B2 of G. atropunctata. B2 was seen at all Ri levels in K. paulula recordings, like with G. atropunctata. B2 was not as common during pathway with K. paulula as with G. atropunctata, although when seen, it was slightly higher in amplitude at lower Ri levels than at high levels (data not shown). At 109 Ohms, B2, while still distinguishable, was shorter in amplitude than B1p (Fig. 9). Thus, B2 is composed of both R and emf, with a slight emphasis on R.

X Wave Phase

The sharpshooter X wave is a striking transition between pathway phase and sustained ingestion phase, comprised of waveforms similar to and evolving in appearance from pathway to sustained ingestion. XC was composed of alternating waveform types XC1 and XC2 in K. paulula, similar to G. atropunctata and all other sharpshooters recorded.

Family XC. Type XC1

This distinctive waveform type was composed of high- to very high-amplitude peaks (1.5–2 times higher than XC2) (see below). Graphocephala atropunctata XC1 was usually narrower than XC2 (about 25–50% the width of XC2) (Fig. 7f). XC1 was recorded at all Ri levels except 106 Ohms and had similar appearance in each; at 106, we were unable to distinguish XC1 from XC2 (see below). Therefore, electrical origin of XC1 is dominated by emf. One to five XC1 peaks always preceded XC2 rounded plateaus. Because XC1 was difficult to separate from XB1p (uniquely in K. paulula), we used the following criteria to separate them. XB1p peaks were narrower (less than 25% the width of an XC2 plateau, therefore narrower than XC1) and shorter than or the same amplitude as an XC2 plateau. XB1p could also occur throughout XN preceding XC, but XC1 always occurred at the beginning of XC before XC2 (e.g., Figs. 7f, g and 8g–I). Thus, XC1 always occurred between XN (see below) and XC2.

Family XC. Type XC2

By definition, XC2 (‘trial ingestion’) events in all sharpshooters last no longer than 300 s (Backus et al. 2005), and the same was the case for K. paulula. Otherwise, XC2 for K. paulula was virtually identical to that for G. atropunctata. For example, XC2 appearance was highly variable during the X wave (300 s in duration), evolving over time (Fig. 5d–f) until it eventually looked the same as C2 (see below) Variability in XC2 appearance was a marker for the X wave.

Family XN

Similar to XN in G. atropunctata, XN in K. paulula showed a distinct change in appearance between Ri 106 Ohms and all other Ri levels. At 106, XN was exclusively positive-oriented (Fig. 4a) or mixed positive-negative-oriented (Fig. 3a). At all other Ri levels, XN was always negative-oriented, regardless of the polarity of the probe. Thus, XN as a whole was composed of mixed R and emf, but emf-dominated.

In G. atropunctata, XN is composed entirely of one type, XB1 (a B1-like waveform but occurring during the X wave). There are two subtypes in XB1: fB1w (‘fuzzy’ B1w, i.e., B1w with a high-frequency component superimposed on top), and XB1s (similar to B1s but during the X wave). In K. paulula, fB1w was almost never seen except at higher Ri levels such as 108 Ohms (Fig. 7f and g) and 109 Ohms (data not shown). Unlike B1s, XB1s was more variable, amorphous, and difficult to clearly distinguish, for both species. When amorphous, it was often just termed XB1. As with pathway B1p, XB1p (during the X wave) did not occur at low Ri levels, but was highly visible and sometimes dominant in K. paulula X waves at 108 and 109 Ohms (Figs. 7f, g and 8f–h). Therefore, XB1p is emf-dominated. XB1p often was so tall at higher Ri levels that it was very difficult to separate from XC1 (e.g., Fig. 7g, see below).

Sustained Ingestion Phase

Family C. Type C2

Similar to G. atropunctata, C2 for K. paulula had a highly stereotypical appearance that no longer evolved over time as was the case with XC2. Also, like XC2 (shown in all figures), C2 amplitude increased with increasing Ri level (not shown in most figures; see more C2 in Backus and Shih 2020). (Compare the difference in apparent amplitude between A1 and XC2 in Figs. 4a and 8a for a representation similar to C2). Therefore, C2, like XC2, was emf-dominated but also had a little R because a much lower amplitude version of it was visible even at 106 Ohms. With K. paulula, again similar to G. atropunctata, when the Ri was low the small C2 plateaus sometimes had a tiny, positive-going peaklet at the beginning or end of each rectangular plateau (Fig. 4e inset box). This tiny peaklet completely disappeared at higher Ri levels, where the plateau became rounded at the top (Fig. 8a inset box). This complete disappearance did not occur with G. atropunctata recordings at 109 Ohms, especially if silver paint was used as the adhesive. Thus, C2 was composed of R (causing the rectangular shape with peaklet) and emf (causing a rounded rectangle) mixed together. Otherwise, the differences between XC2 and C2 were that: 1) the amplitude and shape of XC2 evolved over time while that of C2 was stable, and 2) unlike XC2, C2 events were always longer than 300 s, often lasting one to several hours.

Family C. Type G

G was a common waveform for K. paulula, even more so than for G. atropunctata. That said, its appearance for K. paulula was identical to that of G. atropunctata (data not shown; see Cervantes and Backus 2018). G resembled C2 plateaus with several small spikelets on top of the otherwise rounded plateau. G occurred at all Ri levels, thus was composed of both R and emf. G always occurred directly after C2 at the same voltage level without falling to baseline; thus, it is thought to occur in the same xylem cell. G had a very slow repetition rate, with long to very long stretches of flat line between plateaus. G often occurred for long events in the last few hours of recordings.

Other, Nonphase Waveforms

Several other waveforms that do not correspond to the phases above were described for G. atropunctata and were again seen in K. paulula recordings. The most common ‘other’ waveform was R (data not shown; see Cervantes and Backus 2018). R was essentially a flat or slightly wavering line at one of two voltage levels: either at the same level as preceding C2 or G, or very slightly above baseline. Another unusual waveform was SR (also not shown; see Cervantes and Backus 2018), which comprised rapid, near-vertical declines and rises in voltage from very high to very low, just above baseline. These sudden peaks and valleys were repeated many times in a short duration. While SR was rarely seen in K. paulula recordings, it was very common with G. atropunctata recordings that used DC applied signal (Cervantes and Backus 2018), but rare in recordings that used AC applied signal (similar to K. paulula recordings herein).

Figure 10 shows a final ‘other’ K. paulula waveform type that is similar to but slightly modified from that of G. atropunctata. The waveform type is D, and with G. atropunctata it was composed of D1 and D2 (Cervantes and Backus 2018). With K. paulula, we never saw exactly a D1 waveform, but instead saw a D1-related waveform that seemed like a blend of C2 and D; so, we termed it ‘D+C’ (Fig. 10a–c). The amplitude of D+C was similar to that of C2, although events of this waveform were so rare that we could not compare it across all Ri levels to determine whether its amplitude changed with Ri. D2 for K. paulula was nearly identical to that of G. atropunctata, being a very large waveform of slow, upward-sloping line followed by an abrupt drop (Fig. 10a and c). Again, D2 was relatively rare with K. paulula using AC applied signal.

Fig. 10.

Fig. 10.

Open in a new tab

Example of D-type waveforms at Ri = 109 Ohms, using applied signal of 15 mV AC. Enlargements of boxes b–g in part a. a. Overview of waveforms. Type-level names are along the top of the label bar while subtype-level names are near waveform images with vertical lines separating waveforms. Downward-pointing arrowheads denote unusual potential drops. WinDaq gains and time scales shown in most parts. Gains and scales for part c same as part b.

Variability in Waveform Appearances, Revisited

At one end of the continuum in waveform appearances for 106 Ohms (Fig. 3), an insect’s probes slightly emphasized emf, because its family A waveforms were not very tall in relation to trial ingestion plateaus shown (the smallest waveform, e.g., XC2); also, the plateaus lacked R-component peaklets (Fig. 3a and g). B1p was present at this end of the continuum for 106 Ohms, but only tiny and sporadic (Fig. 3c–g). This version of 106 waveforms occurred when the applied voltage was low (15 mV). At the other end of the continuum for 106 Ohms (Fig. 4), an insect’s probes slightly emphasized R, mostly because the applied voltage was higher (20–25 mV). Family A waveforms were very tall compared with XC2 plateaus (compare Figs. 3a and 4a); also, plateaus displayed some R-component peaklets (Fig. 4a and e inset box) but almost no B1p. Thus, at 106 Ohms, R dominated overall, but could also be slightly to moderately influenced by emf, in inverse proportion to the amount of applied voltage. At 107 Ohms, much more emf was detected, e.g., B1p was more obviously present. Variability in waveform appearances reached its highest level (compare Figs. 5 and 6), with R- and emf-component waveforms blended together in different proportions with individual insects. There was no obvious relationship between waveform appearance and applied voltage level. Therefore, at 107 Ohms, R and emf were well-balanced overall, and variation was mostly likely caused by wiring quality. That said, reducing one cause of variability in wiring quality/electrical conductivity by standardizing the wiring method allowed greater resolution of differences in waveform polarity, caused by even small changes in applied voltage level.

At higher Ri levels, emf became more important, regardless of applied voltage level. At 108 Ohms, the A-family waveforms were shorter with greater prevalence of vd’s. B1p, B1w, and B1s were all present, but trains of B1p began to appear, with higher amplitudes (Fig. 7). At 109 Ohms, emf strongly predominated. A-family waveforms were almost nonexistent, while the amplitude of ingestion waveforms (XC2 and C2, both emf-dominated) was almost higher than the start of the probe. B1p was present in all probes, sometimes in very long trains (Fig. 9), and sometimes rivaling XC1 in amplitude. Relatively little, if any, R component was present at 109 Ohms.

Discussion

A summary of findings about K. paulula from this study compared with G. atropunctata from Cervantes and Backus (2018) is found in Table 1, including estimation of electrical origins for all waveforms cited therein. In addition, five features of K. paulula recordings are noteworthy and suggest different or better understanding of biological meanings compared with G. atropunctata stylet penetration (Cervantes and Backus 2018).

Table 1.

Summary of the most important sharpshooter and spittlebug waveforms, categories, and results from combined findings of this paper and Cervantes and Backus (2018)

Waveform categories and names Findings
Phase  Family Type Subtypes Electrical origins Comparison of K. paulula (Kp) vs G. atropunctata (Ga)
 T Ra More common with Kp than Ga, despite naive insects being recorded on new plants for both.
Pathway
 A A1 Ra Similar in appearance with both species.
A2 R(emf)b Similar in appearance with both species, although vd’s more common in Kp than Ga.
 B B1 B1s/B1w R+emfc Similar in appearance with both species; however, adhesive made a difference. More R seen with paint; emf with glue.
B1p emf(R)d Much taller and more common with Kp than Ga; only present in Ga at high Ri with silver paint.
B2 emf(R)d Similar in appearance with both species, although less common in Kp than Ga.
X wave
 XC XC1 emfa Similar in appearance with both species.
XC2 R+emfc Similar in appearance with both species.
 XN XB1e XB1s emf(R)d Similar in appearance with both species.
fB1w emfa Similar in appearance with both species.
XB1p emf(R)d Much larger and more common with Kp than Ga.
Sustained ingestion
 C C2 emf (plateaus), R (peaklets) Similar in appearance with both species; however, adhesive made a difference. More R seen with paint; more emf with glue.
 G R+emf Similar in appearance with both species.
Interruption
 N emf(R)d Similar in appearance with both species
Others
 D D1, D2 ? Variant D+C seen with Kp, not Ga
 R R+emfc Similar in appearance with both species
 SR ? Similar in appearance with both species
 W R+emfc Similar in appearance with both species
Open in a new tab

Naming convention is that of Backus (Backus and Shih 2020).

aR means that the waveform is strongly R-dominated and disappears at high Ri; similarly, emf means that the waveform is strongly emf-dominated and disappears at low Ri, especially 106 Ohms.

bR(emf) means that, while the waveform is detectable at all Ri levels, its amplitude is highest at low Ri; therefore, R dominates but some emf is also detectable.

cR+emf means R and emf are balanced because this waveform is seen at all Ri levels and is similar in appearance.

demf(R) means that, while the waveform is detectable at all Ri levels, its amplitude is highest at low Ri; therefore, emf dominates but some R is also detectable.

eThis waveform type was mistakenly referred to as B1 in Cervantes and Backus (2018). We correct that error herein.

Variability in Waveform Appearances and the emf/R Responsiveness Curve

Because we used the same wiring method (gold wire loop + silver glue) for all Ri levels, and only varied the applied voltages slightly, we were able to see the full effects of the emf/R responsiveness curve (Backus et al. 2019) for K. paulula waveforms. At 106 Ohms, R-component waveforms were highly dominant, although (as discussed above) their fine-structure appearance varied along an R-emf gradient dependent on applied voltage. At 107 Ohms, there was a balance of R and emf components despite some variability in waveform appearances also due to differences in applied voltage. Consequently, 107 Ohms probably represents the 50:50 level for R:emf ratio, thus the inherent resistance (Ra) of K. paulula. At 108 Ohms, there were still a few R components (e.g., part of waveform A) but much less than at lower Ri levels and no longer so dependent on applied voltage. At 109 Ohms, emf was highly predominant. Almost all of the variation in waveform appearance among insects disappeared, with little effect from applied voltage, as would be expected when waveforms depend upon inherently generated biopotentials such as streaming potentials. Waveforms at 108 Ohms were, in most ways, intermediate between 107 and 109 Ohms. In terms of emf/R responsiveness and inherent resistance, these results are very similar to those from other large hemipterans, such as heteropterans (Cervantes et al. 2016, Lucini et al. 2016, Lucini and Panizzi 2018), and support the concept of the R/emf responsiveness curves (Backus et al. 2019).

Waveform Polarity

For K. paulula, biphasic probes uniformly began at a positive voltage level, then fell into a negative level either once or twice (after rising first). The waveforms gradually rose above baseline at the start of X wave phase until they were entirely above baseline during sustained ingestion phase. We hypothesize that the one or two rise-and-fall changes in voltage level during pathway represent stylet depth, as was proven histologically with H. liturata (Backus et al. 2009). Thus, K. paulula probably made one or two branches of the salivary sheath in each probe. Biphasic polarity occurred at all Ri levels with surprisingly small changes in applied voltage level. This observation has not been made for any other EPG recordings of any insect, to our knowledge. Based on this finding and other work (Backus et al. 2016), we recommend that all EPG users, both DC and AC-DC, pay attention to and carefully standardize their applied voltage levels.

Positive followed by negative voltage levels of pathway waveforms during biphasic probes was clearly an R component because higher-amplitude positive or negative voltage levels occurred with higher applied voltage (an R determinant). Yet, it was always the ingestion waveforms that rose into positive voltage level, gradually during X wave then strongly and more stably during sustained ingestion. We thus further hypothesize that, while the R component is predominant for biphasic probes, ingestion (XC2 or C2) is so emf-dominant (due to powerful streaming potentials during ingestion of xylem sap; Backus and Shih 2020) that the mechanism shifts when an ingestion waveform begins. Understanding these mechanisms greatly aids in interpretation of K. paulula waveforms, especially making it possible to define number of salivary sheath branches and ingestion from xylem based on the B2 waveform and voltage level shifts, without histological correlation.

Biological Meanings of New Waveforms of K. paulula

Family T

Variation in test probe appearance with Ri levels provides important clues about fluid flows and composition occurring during these short probes. Test probes are thought to be primarily sensory in function, to provide an initial taste/feel of a new plant. The insect uses this sensory information to determine whether the plant is acceptable to initiate deeper stylet penetration to search for an acceptable xylem cell (Backus and McLean 1983, Backus 1988). In smaller leafhoppers, no chemosensilla have been found on the tips of the stylets, so chemical constituents of the plant interior are probably brought up the food canal in the stylets into the precibarium for tasting by the precibarial chemosensilla (Backus 1988). Interestingly, larger leafhoppers like sharpshooters have larger stylet tips that can apparently accommodate additional sensory structures. Homalodisca vitripennis has uniporous pegs/sensilla placodea (likely chemosensory based on their external morphology) on the tips of the mandibular stylets (Leopold et al. 2003) that probably can taste internal plant constituents in the shallow plant tissues shortly after initial stylet penetration. The mandibular stylets are not penetrated into the deeper tissues of the plant (because sharpshooters use the maxillary-stylets ahead approach) (Backus 1988) but are used during shallow test probes. Kolla paulula and G. atropunctata (both about the same size, but smaller than H. vitripennis) may have similar stylet tip sensilla.

A-like waveforms (seen at low Ri levels in test probes) have been visually correlated with salivation to create the salivary flange on the exterior of the plant just prior to stylet penetration, as well as the trunk of the salivary sheath internal to the plant (Backus et al. 2005). Strong electrical conductivity of saliva is a primary mechanism of the R component (Walker 2000), and A-family waveforms during waveform T are R-dominated. In contrast, B1s and B1w are a mixture of R and emf caused by opening/closing of precibarial valves/cibarial pump and streaming potentials, respectively (Backus and Shih 2020). Streaming potentials indicate movement and directionality of fluids into and out of the stylets (Walker 2000, Dugravot et al. 2008). B1 during pathway is highly visible at high Ri levels like 109 Ohms; thus, its absence during test probes is obvious. Accordingly, we propose that the very spiky A-type waveforms of low-Ri test probes (waveform T) represent salivation combined with tasting of chemicals (solubilized by and dissolved in the saliva) by the stylet tip sensilla located shallowly in the plant. We also propose that uptake of fluids into the precibarium does not occur during sharpshooter test probes, because the B1 waveform is not visible at high Ri levels like 109 Ohms. Instead, a nearly flat line usually occurs because there is virtually no emf component in test probes. Thus, stylet penetration in sharpshooters probably progresses from stylet tip chemosensing during test probe salivation to precibarial chemosensing during exploratory probe salivation.

Family B. Subtype B1p

B1p was seen rarely seen in G. atropunctata and merely named in Cervantes and Backus (2018); therefore, very little could be inferred about its biological meaning. Because of its unique (to date among sharpshooter species) predominance during B waveforms (both in pathway and X waves) in K. paulula, we can infer much more about its biological meaning.

The main electrical origin of B1p is emf, so much so that it can dominate pathway phase at 109 Ohms. The clear resemblance of B1p peaks to XC1 peaks (even, their progressive evolution into XC1 over time, in some probes) strongly suggests that the same mechanism underlies each type of peak. Relatively tall, peak- and plateau-like, emf waveforms such as XC1, XC2, and C2 are caused by streaming potentials generated by rapid fluid movement through the narrow, capillary-like food canal in the stylets, as further explained in Backus and Shih (2020). Repeated plateaus like XC2 and C2 probably occur when the cibarial diaphragm is gradually lifted (after the rise during the flat top of the plateau) (Dugravot et al. 2008) as the cibarium fills; then, as the diaphragm is released and falls asymmetrically, fluid is swallowed into the pharynx (Dugravot et al. 2008, Backus 2016).

Repeated, straight-walled and abrupt peaks like B1p and XC1 probably occur when the cibarial diaphragm is lifted but then immediately dropped, propelling fluid out of the cibarium without filling. Thus, B1p and XC1 rapidly bring fluids from the plant into the full length of the buccal cavity (i.e., the food canal, precibarium, and cibarium), and then rapidly propel them back out the same route. Trains of B1p peaks imply repetitive, rapid in-out movements (uptake and spitting out) of fluids, probably for tasting by the precibarial chemosensilla (Backus 1988, 2016). Because amplitude of the peak is proportional to the degree of uplift of the cibarial diaphragm (Dugravot et al. 2008), we hypothesize that the much taller XC1 peaks represent the first uptake and spitting out of fluids from a mature xylem cell because the strongly negative pressure potential of xylem sap requires very strong suction by the cibarial diaphragm. Uptake and spitting during XC1 is thought to be the mechanism of discharge egestion (Backus and Morgan 2011, Backus 2016). During the less-tall B1p peaks, similar discharge egestion is probably occurring in a variety of cells. These could include mesophyll/parenchyma cells or immature (not yet translocating) protoxylem cells along the pathway, which would have positive or neutral pressure potentials, thus require less cibarial diaphragm lifting. Alternatively, when B1p peaks occur right before XC1, they could represent the first mechanical testing of a mature xylem cell. Accordingly, the finding of this new, common variant of a previously rare waveform, B1p, indirectly adds support to the salivation-egestion mechanism of X. fastidiosa inoculation (Backus 2016) because it provides a new waveform in addition to XC1 for discharge egestion, thought to be the most significant behavior for inoculation.

It is also interesting to compare the biological meaning of B1p and other subtypes of B1, especially B1s. At present, evidence from streaming potentials supports that B1s represents precibarial valve turbulence (Backus and Shih 2020) and/or cibarial diaphragm quivering (tiny up-down movements of the diaphragm), or both (Backus 2016, E.A.B., personal observations). Thus, B1s represents uptake of very small amounts of fluid into just the precibarium, fluid movement around the precibarium, and gentle release of those small amounts of fluid out the stylet tips (termed uptake, swishing, and dribbling, respectively) (Backus 2016, Backus and Shih 2020). Accordingly, B1s dribbling, a slow and gentle expulsion of fluids (that might contain X. fastidiosa bacteria) from primarily below the precibarial valve, termed rinsing egestion (Backus and Morgan 2011), occurs almost continuously during pathway and X wave searching/testing of xylem cells by G. atropunctata and other sharpshooters. In contrast, the ubiquity of B1p during K. paulula stylet penetration suggests the more forceful discharge egestion (probably cleaning out the entire length of the precibarium) occurs throughout stylet penetration of that species, in all cell types. We hypothesize that B1p might, therefore, make K. paulula a more efficient vector than G. atropunctata. Of course, direct research such as classical transmission studies would be needed to confirm this hypothesis.

Other Waveforms

Family D. Subtype D+C

The biological meaning of family D is completely unknown, but we can say that D waveforms are emf-dominated. The new D+C subtype is interesting because part of it strongly resembles XC2 or C2 but with small, upward-sloping D in between C2 plateaus. The immediate temporal juxtaposition of D and C2 in this waveform suggests that the D-like portion is also caused by cibarial diaphragm uplift, like C2. If so, then perhaps the short-amplitude D in D+C also is similar (but with higher diaphragm uplift?) to D2. Consequently, we hypothesize that D-family waveforms represent some type of ingestion in K. paulula, and thus possibly other sharpshooters also; however, it is not presently known from what cell type the insect might be ingesting. Future histological research will be necessary to fairly identify the stylet location for waveform family D.

Conclusions

Herein we presented the first EPG recordings of an Asian sharpshooter leafhopper, K. paulula, indigenous to Taiwan and southern Asia and now a vector of introduced X. fastidiosa in Taiwan (Tuan et al. 2016). We developed a waveform library for K. paulula and compared it to the recently published waveform library for the North American species G. atropunctata (Say) (Cervantes and Backus 2018). Overall, the waveforms of K. paulula and G. atropunctata were similar, and we found no difficulty in using the Backus waveform naming convention for K. paulula, as established for G. atropunctata and other sharpshooters from the Americas. That said, we also found some interesting differences in waveform coarse and fine structure between the two species. For coarse structure, waveform polarity was related to small shifts in applied voltage level (5–10 mV) and Ri level, for the first time for any species. This finding continues to support that voltage levels should be carefully standardized in EPG recordings. For fine structure appearance, the B1p waveform was seen more commonly in K. paulula than another other sharpshooters studied to date.

For future quantitative studies of K. paulula, we recommend using the AC-DC electropenetrograph (Backus and Bennett 2009), so that a researcher can tailor the settings to the insect. As a result of this study, we propose that the K. paulula inherent resistance level is 107 Ohms, at which setting a researcher would find the best balance of R and emf components. However, because the proposed X. fastidiosa inoculation waveforms during X wave phase are important for quantitative studies and are emf-dominated, we recommend using 108 Ohms as the Ri level for future vector research. Use of an applied voltage level of 25–30 mV AC will retain some R components such as biphasic waveform polarity. Also, 108 Ohms is advantageous because it will allow the researcher to distinguish between B1p and XC1 at the end of pathway phase. We do not recommend use of DC, especially at Ri 109 Ohms, because of possible negative effects on stylet probing (Backus et al. 2018, Cervantes and Backus 2018). Also, it is very difficult to distinguish between B1p and XC1 at that Ri level. However, if DC must be used (especially at 109 Ohms as with the DC system), then it is crucial that the applied voltage level be kept below 20 mV via checking the output voltage from the monitor with a high-quality volt-ohm-meter.

Relative consistency of waveform appearances across several species of sharpshooter leafhoppers (Cicadellidae: Cicadellinae) as well as spittlebugs (Aphrophoridae) supports that biological meanings of sharpshooter waveforms can be applied to all species within the group. Thus, the potential of EPG technology for X. fastidiosa vectors now has been thoroughly demonstrated. It will be a powerful tool for the biosecurity of countries in the potential invasion zone of X. fastidiosa.

Acknowledgments

We thank Mei-Rong Tzeng (TARI) for technical help collecting and rearing insects and plants. We also thank Dr. Chiou-Chu Su for helping to identify whether the wild population of K. paulula has X. fastidiosa pathogens. The editorial suggestions of Dr. James Throne (ARS Parlier) and three anonymous reviewers were also greatly appreciated. This work was supported in part by the Ministry of Science and Technology Foundation (grant number MOST 108-2321-B-005-007) and in part by the Council of Agriculture Foundation (grant number 107AS-4.1.1-CI-C1) of the Republic of China (to H.-T.S.), and in-house funds from USDA-ARS (2034-22000-010-00D) (to E.A.B.).

Footnotes

Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. USDA is an equal opportunity provider and employer.

This article was prepared by a U.S. Department of Agriculture employee as part of his/her official duties. Copyright protection under U.S. Copyright Law Title 17 U.S.C. § 105 is not available for such works. Accordingly, there is no copyright to transfer. The fact that the private publication in which the article appears is itself copyrighted does not affect the material of the U.S. Government, which can be freely reproduced by the public. Articles and other publications prepared as part of a Federal employee’s official duties are property of the U.S. Government.

References Cited

  1. Almeida R. P. P. 2016a. Can Apulia’s olive trees be saved? Science. 353: 346–348. [DOI] [PubMed] [Google Scholar]
  2. Almeida R. P. P. 2016b. Xylella fastidiosa vector transmission biology, pp. 165–173. InBrown J. K. (ed.), Vector-mediated transmission of plant pathogen. American Phytopathological Society Press, St. Paul, MN. [Google Scholar]
  3. Backus E. A. 1988. Sensory systems and behaviours which mediate hemipteran plant-feeding: a taxonomic overview. J. Insect Physiol. 34: 151–165. [Google Scholar]
  4. Backus E. A. 2016. Sharpshooter feeding behavior in relation to transmission of Xylella fastidiosa: a model for foregut-borne transmission mechanisms, pp. 173–194. InBrown J. K. (ed.), Vector-mediated transmission of plant pathogen. American Phytopathological Society Press, St. Paul, MN. [Google Scholar]
  5. Backus E. A., and Bennett W. H.. . 2009. The AC-DC correlation monitor: new EPG design with flexible input resistors to detect both R and emf components for any piercing-sucking hemipteran. J. Insect Physiol. 55: 869–884. [DOI] [PubMed] [Google Scholar]
  6. Backus E. A., and McLean D. L.. . 1983. Behavioral evidence that the precibarial sensilla of leafhoppers are chemosensory and function in host discrimination. Entomol. Exp. Appl. 37: 219–228. [Google Scholar]
  7. Backus E. A., and Morgan D. J.. . 2011. Spatiotemporal colonization of Xylella fastidiosa in its vector supports the role of egestion in the inoculation mechanism of foregut-borne plant pathogens. Phytopathology. 101: 912–922. [DOI] [PubMed] [Google Scholar]
  8. Backus E. A. and Shih H. T.. . 2020. Review of the EPG waveforms of sharpshooters and spittlebugs including their biological meanings in relation to transmission of Xylella fastidiosa. J. Insect Sci. In press. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Backus E. A., Habibi J., Yan F., and Ellersieck M.. . 2005. Stylet penetration by adult Homalodisca coagulata on grape: electrical penetration graph waveform characterization, tissue correlation, and possible implications for transmission of Xylella fastidiosa. Ann. Entomol. Soc. Am. 98: 787–813. [Google Scholar]
  10. Backus E. A., Holmes W. J., Schreiber F., Reardon B. J., and Walker G. P.. . 2009. Sharpshooter X wave: correlation of an electrical penetration graph waveform with xylem penetration supports a hypothesized mechanism for Xylella fastidiosa inoculation. Ann. Entomol. Soc. Am. 102: 847–867. [Google Scholar]
  11. Backus E. A., Cervantes F. A., Godfrey L., Akbar W., Clark T. L., and Rojas M. G.. . 2018. Certain applied electrical signals during EPG cause negative effects on stylet probing behaviors by adult Lygus lineolaris (Hemiptera: Miridae). J. Insect Physiol. 105: 64–75. [DOI] [PubMed] [Google Scholar]
  12. Backus E. A., Cervantes F. A., Narciso Guedes R., Li A. Y., and Wayadande A. C.. . 2019. AC-DC electropenetrography (EPG) for in-depth studies of oviposition and piercing-sucking feeding behaviors. Ann. Entomol. Soc. Am. 112: 236–248. [Google Scholar]
  13. Cervantes F. A., and Backus E. A.. . 2018. EPG waveform library for Graphocephala atropunctata (Hemiptera: Cicadellidae): effect of adhesive, input resistor, and voltage levels on waveform appearance and stylet probing behaviors. J. Insect Physiol. 109: 21–40. [DOI] [PubMed] [Google Scholar]
  14. Cervantes F. A., Backus E. A., Godfrey L., Akbar W., and Clark T. A.. . 2016. Characterization of an EPG waveform library for pre-reproductive adult Lygus lineolaris and L. hesperus feeding on cotton squares. Ann. Entomol. Soc. Am. 109: 684–697. [Google Scholar]
  15. Cornara D., Saponari M., Zeilinger A. R., de Stradis A., Boscia D., Loconsole G., Bosco D., Martelli G. P., Almeida R. P. P., and Porcelli F.. . 2017. Spittlebugs as vectors of Xylella fastidiosa in olive orchards in Italy. J. Pest Sci. (2004). 90: 521–530. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Dugravot S., Backus E. A., Reardon B. J., and Miller T. A.. . 2008. Correlations of cibarial muscle activities of Homalodisca spp. sharpshooters (Hemiptera: Cicadellidae) with EPG ingestion waveform and excretion. J. Insect Physiol. 54: 1467–1478. [DOI] [PubMed] [Google Scholar]
  17. Fereres A. 2007. The role of aphid salivation in the transmission of plant viruses. Phytoparasitica. 35: 3–7. [Google Scholar]
  18. Giampetruzzi A., Saponari M., Loconsole G., Boscia D., Savino V. N., Almeida R. P. P., Zicca S., Landa B. B., Chacón-Diaz C., and Saldarelli P.. . 2017. Genome-wide analysis provides evidence on the genetic relatedness of the emergent Xylella fastidiosa genotype in Italy to isolates from Central America. Phytopathology. 107: 816–827. [DOI] [PubMed] [Google Scholar]
  19. Grandgirard J., Hoddle M. S., Roderick G. K., Petit J., Percy D., Putoa R., Garnier C., and Davies N.. . 2006. Invasion of French Polynesia by the glassy-winged sharpshooter, Homalodisca coagulata (Hemiptera: Cicadellidae): a new threat to the South Pacific. Pac. Sci. 60: 429–438. [Google Scholar]
  20. Krugner R., Sisterson M. S., Backus E. A., Burbank L. P., and Redak R. A.. . 2019. Sharpshooters: a review of what moves Xylella fastidiosa. Austral Entomol. 58: 248–267. [Google Scholar]
  21. Leopold R. A., Freeman T. P., Buckner J. S., and Nelson D. R.. . 2003. Mouthpart morphology and stylet penetration of host plants by the glassy-winged sharpshooter, Homalodisca coagulata (Homoptera: Cicadellidae). Arthropod Struct. Dev. 32: 189–199. [DOI] [PubMed] [Google Scholar]
  22. Lucini T., and Panizzi A. R.. . 2018. Electropenetrography (EPG): a breakthrough tool unveiling stink bug (Pentatomidae) feeding on plants. Neotrop. Entomol. 47: 6–18. [DOI] [PubMed] [Google Scholar]
  23. Lucini T., Panizzi A. R., and Backus E. A.. . 2016. EPG waveform characterization for redbanded stink bug, Piezodorus guildinii (Westwood) (Hemiptera: Pentatomidae), on soybean plants. Ann. Entomol. Soc. Am. 109: 198–210. [Google Scholar]
  24. McLean D. L., and Kinsey M. G.. . 1964. A technique for electronically recording aphid feeding and salivation. Nature. 202: 1358–1359. [Google Scholar]
  25. Miranda M. P., Fereres A., Appezzato-Da-Gloria B., and Lopes J. R. S.. . 2009. Characterization of electrical penetration graphs of Bucephalogonia xanthophis, a vector of Xylella fastidiosa in citrus. Entomol. Exp. Appl. 130: 35–46. [Google Scholar]
  26. Mizell R. F., Tipping C., Andersen P. C., Brodbeck B. V., Hunter W. B., and Northfield T.. . 2008. Behavioral model for Homalodisca vitripennis (Hemiptera: Cicadellidae): optimization of host plant utilization and management implications. Environ. Entomol. 37: 1049–1062. [DOI] [PubMed] [Google Scholar]
  27. Oman P. W. 1949. The Nearctic leafhoppers (Homoptera: Cicedellidae). A generic classification and check list. Mem. Entomol. Soc. Wash. 3: 1–253. [Google Scholar]
  28. Pearson C. C., Backus E. A., Shugart H. J., and Munyaneza J. E.. . 2014. Characterization and correlation of EPG waveforms of Bactericera cockerelli (Hemiptera: Triozidae): variability in waveform appearance in relation to applied signal. Ann. Entomol. Soc. Am. 107: 650–666. [Google Scholar]
  29. Purcell A. H. 1997. Xylella fastidiosa, a regional problem or global threat? J. Plant Pathol. 79: 99–105. [Google Scholar]
  30. Redak R. A., Purcell A. H., Lopes J. R., Blua M. J., Mizell R. F., and Andersen P. C.. . 2004. The biology of xylem fluid-feeding insect vectors of Xylella fastidiosa and their relation to disease epidemiology. Ann. Rev. Entomol. 49: 243–270. [DOI] [PubMed] [Google Scholar]
  31. Sandanayaka W. R. M., and Backus E. A.. . 2008. Quantitative comparison of stylet penetration behaviors of glassy-winged sharpshooter on selected hosts. J. Econ. Entomol. 101: 1183–1197. [DOI] [PubMed] [Google Scholar]
  32. Sandanayaka W. R. M., Chhagan A., and Ramankutty P.. . 2007. Host plant testing of the spittle bug Carystoterpa fingens by stylet penetration behaviour. N. Z. Plant Prot. 60: 78–84. [Google Scholar]
  33. Sandanayaka W. R. M., Jia Y., and Charles J. G.. . 2012. EPG technique as a tool to reveal host plant acceptance by xylem sap-feeding insects. J. Appl. Entomol. 137: 519–529. [Google Scholar]
  34. Sandanayaka W. R. M., Charles J. G., and Froud K. J.. . 2017. Potential use of electrical penetration graph (EPG) technology for biosecurity incursion response decision making. N. Z. Plant Prot. 70: 1–15. [Google Scholar]
  35. Shih H. T., Su C. C., Feng C. Y., Fanjiang C. C., Hung W. F., and Hung L. Y.. . 2009. Studies on the morphology, ecology, and host range for Kolla paulula (Walker, 1858) Hemiptera: Membracoidea: Cicadellidae: Cicadellinae. Formosan Entomol. 29: 353. [Google Scholar]
  36. Shih H. T., Wen Y. D., Fanjian C. C., Chang C. J., Chang C. M., Lee C. Y., Yao M. H., Chang S. C., Jan F. J., and Su C. C.. . 2013. Potential vectors of Pierce’s disease in Taiwan: ecology and integrated management, pp. 163–175. InChang C. J., Lee C. Y., and Shih H. T. (eds.), Proceedings of the 2013 international symposium on insect vectors and insect-borne diseases. Special Publication of TARI, Taiwan, ROC. No. 173. [Google Scholar]
  37. Su C. C., Chang C. M., Chang C. J., Su W. Y. Deng W. L., and Shih H. T.. . 2013a. Occurrence of Pierce’s disease of grapevines and its control strategies in Taiwan. Plant Pathol. Bull. 22: 245–258. [Google Scholar]
  38. Su C. C., Chang C. J., Chang C. M., Shih H. T., Tzeng K. C., Jan F. J., Kao C. W., and Deng W. L.. . 2013b. Pierce’s disease of grapevines in Taiwan: isolation, cultivation and pathogenicity of Xylella fastidiosa. J. Phytopathol. 161: 389–396. [Google Scholar]
  39. Tjallingii W. F. 1978. Electronic recording of penetration behaviour by aphids. Entomol. Exp. Appl. 24: 721–730. [Google Scholar]
  40. Tuan S. J., Hu F. T., Chang H. Y., Chang P. W., Chen Y. H., and Huang T. P.. . 2016. Xylella fastidiosa transmission and life history of two Cicadellinae sharpshooters, Kolla paulula and Bothrogonia ferruginea (Hemiptera: Cicadellidae), in Taiwan. J. Econ. Entomol. 109: 1034–1040. [DOI] [PubMed] [Google Scholar]
  41. Walker G. P. 2000. Beginner’s guide to electronic monitoring, pp. 14–40. InWalker G. P. and Backus E. A. (eds.), Principles and applications of electronic monitoring and other techniques in the study of homopteran feeding behavior. Thomas Say Publications in Entomology: proceedings. Entomological Society of America, Lanham, MD. [Google Scholar]

Articles from Journal of Insect Science are provided here courtesy of University of Wisconsin Libraries

RESOURCES