Phenology, seasonal abundance and stage-structure of spittlebug (Hemiptera: Aphrophoridae) populations in olive groves in Italy

Spittlebugs (Hemiptera: Aphrophoridae) are the dominant xylem-sap feeders in the Mediterranean area and the only proven vectors of Xylella fastidiosa ST53, the causal agent of the olive dieback epidemic in Apulia, Italy. We have investigated the structured population phenology, abundance and seasonal movement between crops and wild plant species of both the nymphal and adult stages of different spittlebug species in olive groves. Field surveys were conducted during the 2016–2018 period in four olive orchards located in coastal and inland areas in the Apulia and Liguria regions in Italy. The nymphal population in the herbaceous cover was estimated using quadrat samplings. Adults were collected through sweep nets on three different vegetational components: herbaceous cover, olive canopy and wild woody plants. Philaenus spumarius was the most abundant species; its nymphs were collected from early March and reached a peak around mid-April, when the 4th instar was prevalent. Spittlebug adults were collected from late April until late autumn. P. spumarius adults were abundant on the herbaceous cover and olive trees in late spring, and they then dispersed to wild woody hosts during the summer and returned to the olive groves in autumn when searching for oviposition sites in the herbaceous cover. A relatively high abundance of P. spumarius was observed on olive trees during summer in the Liguria Region. The present work provides a large amount of data on the life cycle of spittlebugs within an olive agroecosystem that can be used to design effective control programmes against these vectors in infected areas and to assess the risk of the establishment and spread of X. fastidiosa to Xylella-free areas.

ploughed in Southern Italy in June/July (thus not affecting the development of spittlebug nymphs) to avoid fire hazards during summer. One of the selected olive groves in each region was located in a coastal area and the other one in an inland area, and this choice was made in order to investigate spittlebug populations under different climatic conditions and agroecosystems within the Mediterranean basin. The sites in Apulia and Liguria are about 1,000 km apart and fall within the warm-summer Mediterranean climate area 45 . In 2017, soil tilling and the sowing of broad bean were carried out in the inland olive grove (Locorotondo municipality), making it no longer suitable for monitoring. For this reason, a new site, about 5 km from the previous one, was identified and monitored in 2017. Surveys in the Apulian olive groves only were thus prolonged to 2018. The principal characteristics of the olive groves surveys and their agroecosystems are reported in Supplementary Table S1.
A homogeneous area of 1 ha (primary sampling unit -PSU) was surveyed inside each olive grove. The nymphal populations of Aphrophoridae were monitored using the quadrat sampling method, which is frequently used to quantify the density of spittlebug nymphs 22,38,44 . At each sampling date, 30 quadrats of 0.25 m 2 (secondary sampling unit of nymphs -SSU n ;100 × 25 cm each) were randomly positioned on the ground cover inside each PSU. The vegetation and soil surface inside the quadrat were inspected carefully for spittlebug nymphs, which were counted, and their instars were determined directly in the field using the identification key of Vilbaste 46 . The spittlebug samplings were conservative, and the nymphs were therefore not collected in order to avoid interference with the population dynamics throughout the season. Only a few nymphs were reared to adults in the laboratory to confirm their species. The sampling of the nymphal stages was carried out weekly from early March until late-May, or until no nymphs were found in any SSU on two consecutive sampling dates. The field samplings in Apulia in 2016 were started in late March, because of unfavourable weather conditions, thus the first phases of nymphal development were not observed. 2 nd -3 rd and 4 th -5 th instars were aggregated because of the initial difficulties in the instar identification under field conditions. Spittlebug adults were sampled inside the PSU of each olive grove using a sweep net (38 cm in diameter) on three different types of vegetation: herbaceous cover, olive trees and wild woody plants (shrubs or trees). The sweep net has proven to be the most reliable sampling method for spittlebug adults 43,47 . Cicadas (Hemiptera: Cicadoidea), an important group of xylem-sap feeders in Mediterranean olive groves, were not sampled because they require different sampling methods and their role in Xf transmission has never been confirmed, although it has been suggested in a few studies 48,49 . Samplings were carried out at randomly selected points on each sampling date to avoid repeated disturbance of the same points throughout the season. The spittlebugs in the ground vegetation were sampled in 30 randomly distributed secondary sampling units (SSU h ). The sampling consisted of 4 sweeps performed along a 2.8-meter transect (four 0.7 m steps each). Therefore, the area effectively sampled by means of sweeping in an SSU h was estimated to be about 1.0 m 2 (0.7 m long × 0.38 m wide × 4 sweeps). One hundred and twenty sweeps were performed in the ground vegetation per site per sampling date. The insects in the olive canopies were sampled using a sweep net with a 2-meter long stick on 20 randomly distributed olive trees (an olive tree is an SSU o ). Ten sweeps were performed on each olive tree around the entire canopy. Sampling on wild woody plants was carried out with the same methodology as that used for the olive trees, but on 10 different, randomly chosen plants (a plant corresponds to an SSU s ). Owing to the presence of dense Mediterranean shrubland (garrigue), it was not always possible to perform the sweeps all around a single plant. The wild woody plant species present in each site were identified at a species level 50 . Sampling of spittlebug adults was carried out from the appearance of the first early adults in the quadrat samplings until October/November, or when no spittlebugs were found in any SSU on two consecutive dates. The adult spittlebug samplings were conservative, thus the collected insects were released immediately in the field after being identified. Any P. spumarius individuals were also sexed to determine the sex ratio trend during the season.
The air temperature and relative humidity were monitored in each olive grove using data loggers (HOBO U23-002; Onset Computer, Bourne, MA, USA). The data loggers were installed in the centre of the PSU, close to an olive trunk, 40 cm above the soil level, and they were covered by a solar radiation shield. Such a location permitted us to obtain similar weather data to those experienced by the nymphs and adults present on the ground cover. Whenever the data loggers failed to record weather data, due to technical issues, the information gaps were filled with data from the closest weather station. In order to study the phenology of P. spumarius, hourly air temperature data, expressed in °C, were used to calculate the degree-days (DD) according to the following formula: where n is the number of hours in one year, from 1 st January to 31 st December; T i is the air temperature at hour i and T 0 represents the lower developmental threshold temperature for P. spumarius. Compared to chronological observations, phenology models, based on an accumulated heat unit (degree-days), help to describe the temperature-dependent development of insects in a more realistic way. In this study, the lowest developmental threshold temperature was set at 8 °C on the basis of the preliminary results of experiments conducted at controlled temperatures (authors' unpublished data). Previous studies conducted by Chmiel and Wilson 32 , West and Lees 51 and Zajac et al. 33 also reported P. spumarius temperature thresholds well below 10 °C.
Statistical analyses. Statistical analyses were conducted and plots were drawn up using R statistical software, version 3.4.4 (R Core Team 2017). Descriptive statistics were calculated for the nymphal and adult stages sampled during the surveys (tidyverse packages). The effect of the chronological time (days after first sampling), the olive grove, the sampling region, and their interactions with the adult sex ratio of P. spumarius were tested using GLM with a quasi-binomial error structure (due to overdispersion) (glm function). The best fitting model was selected using a step-wise procedure based on qAIC index (Mumln package).

Results and Discussion
Species composition and relative abundance. The samplings of the spittlebug populations in four olive orchards located in two distant Italian regions, Apulia and Liguria, indicated the presence of three predominant species of Aphrophoridae: P. spumarius, N. campestris and Aphrophora alni, the latter only being found in the Ligurian olive groves. Philaenus spumarius showed the highest mean densities among the spittlebugs for both the nymphal and adult stages (Tables 1 and 2). Philaenus spumarius was the predominant species in both regions, contributing by 77.9% and 83.1% to all the spittlebug nymphs and by 79.4% and 73.8% to the total spittlebug adults sampled during the 2016-2018 surveys in Apulia and Liguria, respectively. The Ligurian olive groves usually showed higher densities of meadow spittlebugs than the Apulian ones. Neophilaenus campestris accounted for 20.6% and 14.7% of the spittlebug adults in Apulia and in Liguria, respectively. The N. campestris densities were usually lower than the ones registered for P. spumarius, with notable differences between the sampling years, olive groves, and, for the adults, between the vegetation compartments (Tables 1 and 2). Aphrophora alni accounted for 3.8% of the total adults in the Ligurian olive groves. Another three Cercopoidea species [Cercopis vulnerata (Rossi), C. sanguinolenta (Scopoli) (Hemiptera: Cercopidae) and Lepyronia coleoptrata (L.) (Hemiptera: Aphrophoridae)] were occasionally sampled as adults on the herbaceous cover in the Liguria (C. vulnerata and L. coleoptrata) and Apulia (C. sanguinolenta) olive groves.
philaenus spumarius. Nymphal instars. Philaenus spumarius nymphs were detected, starting from the second week of March, with no significant chronological differences between the sampling years or olive groves monitored in Apulia and Liguria ( Fig. 2 and Table 3). The physiological time of the first instar appearance varied from the lowest value of 78 DD in the Apulian inland site (Locorotondo) in 2017 to the highest value of 198 DD in the Apulian coastal site (Valenzano) in 2018 (Fig. 3). Less variability in the physiological time of the first instar appearance was recorded in Liguria, where it ranged from a minimum of 94 DD in the coastal site (Finale) in 2017 to a maximum of 100 DD in the inland site (Arnasco) in 2017. All the nymphal stages followed a similar pattern of appearance in the different olive groves to the one recorded for the 1 st instar for both the physiological (Fig. 3) and chronological timings (Fig. 2). The seasonal peak of nymphal populations was registered in mid-April in Liguria and slightly earlier in Apulia ( Fig. 2 and Table 3). In Apulia, the lowest DD recorded for the nymphal population peak was 180 DD on 7 th April in Locorotondo in 2017 and the highest value was 369 DD on 7 th April in Valenzano in 2017. The data collected in Apulia during the 2016 surveys did not permit us to detect a clear population peak since the samplings started late, and the DD for the population peaks were not considered in 2016. In Liguria, the seasonal peak of the nymphal population ranged from a minimum of 267 DD on 11 th April in Finale in 2016 to a maximum of 294 DD on 13 th April in Arnasco in 2017. The nymphal densities declined gradually throughout late April and May, with the 5th nymphal stage being observed until the first half of May (2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19) in Apulia and until the second half of May (17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31) in Liguria (Table 3). In Apulia, the lowest DD for which the last nymphal instars were found was 434 DD, which was recorded on 19 th May in Locorotondo in 2017, whereas the highest DD in Valenzano was 689 DD and was recorded on 12 th May in 2017. In Liguria, the last nymphal instars were recorded at a minimum of 517 DD on 17 th May in Arnasco in 2017 to a maximum of 664 DD on 19 th May in Finale in 2017. Overall, in Liguria, the DD values for the first instar appearance, population peak and last nymphal instars, all fell within the Apulian DD range. The time span of the nymphal stages, i.e. from the appearance of the first nymphs to the appearance of the first adults, ranged from 30-50 days (195-350 DD) in Apulia to 50-70 days (269-417 DD) in Liguria ( Table 3). The period of presence of each nymphal instar varied between 32 and 56-63 days in Liguria and between 14 and 36 days in Apulia, except for the first instar, which was collected over a limited time span of 2-3 weeks. The duration of each instar in Apulia was shorter for both the chronological and physiological times than in Liguria (Table 3 and Fig. 3). Overall, the early instar stages (1 st -2 nd -3 rd ) developed over higher cumulated DD in Liguria than in Apulia (Fig. 3). www.nature.com/scientificreports www.nature.com/scientificreports/ The mean densities of the nymphal stages were higher in Liguria, with a range of population peaks of 31-68 nymphs m −2 , compared to the Apulia peaks, which ranged from 7 to 39 nymphs m −2 (Fig. 2). The progression of juvenile stages of P. spumarius occurred as a series of highly overlapping distributions that extended approximately over three months during spring, both in Apulia and Liguria. Indeed, several different nymphal instars were found on each sampling date, and the 1 st instar only disappeared when the 4 th instar was present. The       www.nature.com/scientificreports www.nature.com/scientificreports/ maximum nymphal population densities were relatively constant for the different instars within each olive grove, but they varied between the olive groves, regions and years. For example, the overall densities of P. spumarius registered in Apulia during the 2018 surveys were much lower than those found in 2017 in the same sites (the maximum density in Locorotondo was 39.1 nymphs m −2 in 2017 and 9.3 nymphs m −2 in 2018) (Fig. 2). The first instar was an exception, because its densities were always low, probably because of the difficulty detecting these small nymphs under field conditions and its shorter development time than that of the other nymphal stages.
The temporal occurrence of P. spumarius nymphs followed a well-defined pattern in all of the monitored olive groves, that is, emerging in early March, peaking in April, and emerging to the adult stage from late April until the end of May. The temperatures registered in late winter and spring were quite similar in the investigated olive groves in both regions, although lower temperatures were recorded in the inland olive grove in Apulia than in the coastal one (see Supplementary Fig. S2). The slight difference in the phenology of spittlebugs in the two Mediterranean regions was therefore not completely unexpected, since the hatching and development of P. spumarius nymphs are influenced by the temperature and photoperiod, although some contrasting results have been reported 23,41,52,53 . The differences in DD needed the for nymphal development of P. spumarius observed in the surveyed olive groves were related to: i) the non-linear responses of the development rate at high temperatures and a rather low value for the upper development threshold, (see Chmiel and Wilson 32 , who estimated a nymphal upper threshold of 26.7 °C), ii) the gradual egg hatching, possibly influenced by differences in cold temperatures experienced throughout winter 23 . Further experiments focused on disentangling the effect of the temperature on juvenile development are currently ongoing in our laboratory under controlled conditions to control for any confounding effects, such as host plant communities or population differences.
Our results on the P. spumarius nymphal stage phenology are consistent with some that have recently been reported for other Mediterranean olive groves 9,43,44 . In previous studies carried out in the USA, a similar development timing was reported for California, whose climatic conditions are similar to the Mediterranean ones, whereas a delay of almost one month was registered in other States 29,32,54,55 . Some studies conducted in Europe, in the UK 22,56 , Germany 23 and the Black Sea region of Turkey 57 , investigated the phenology of the meadow spittlebug in northern regions and found nymphal stages between April and June. According to Halkka et al. 21 , the chronological delay in nymph development in Finland, compared to the Mediterranean basin, exceeds two months.
Adult stage. Newly emerged adults were detected from late April to early May. Adult emergence occurred without any relevant differences in chronological time between the olive groves and regions. In 2016, in the inland grove of Liguria, a single freshly emerged adult was already found on 11 th April (278 DD), 3 weeks earlier than the first ones observed in the coastal grove on 4 th May (426 DD). However, adults only started to emerge consistently on 20 th May, 71 days after the first observation of nymphs at 547 DD. In 2017, in Locorotondo, the first adults emerged at 273 DD on 28 th April, which is the lowest DD recorded at either site, while the highest DD recorded for first adult emergence was 626 DD on 28 th April in Valenzano in 2016 (Fig. 3).
Following emergence, Philaenus spumarius adults were found in high number on herbaceous cover in late May-mid June, in both the Apulian and Ligurian olive groves (Figs. 4, 5). The peak of adult abundance in spring in Apulia ranged from 434 DD in Locorotondo in 2017 to 1213 DD in Valenzano in 2018 (Table 4). In Liguria, the peak of adult population abundance in spring ranged from 571 DD in Finale in 2017 to 822 DD in Finale in 2016. Up to 1.55 ± 0.22 adults per olive tree were collected on olive canopies during late spring, about 1-3 weeks later than the abundance peak found on the herbaceous cover (Figs. 4 and 5).
The density of P. spumarius adults decreased in the entire olive agroecosystem during summer, although some differences were observed between Apulia and Liguria. In the Apulian olive groves, the number of P. spumarius decreased abruptly on the herbaceous cover from late spring, with no spittlebugs being sampled from mid-June until mid-September (Fig. 4), in part as a consequence of soil tilling. The spittlebug populations on woody plants (both olives and wild trees) decreased more slowly than on the herbaceous cover, and a limited presence of individuals was registered throughout the summer. In Liguria, P. spumarius did not disappear from the herbaceous cover during summer, and adults were present with relatively high densities on olive canopies (mean range 0.5-1 adult/tree) and, especially in coastal olive grove, on wild woody hosts (up to 1.6 adult/tree) (Fig. 5). The abundance and the pattern of P. spumarius registered on wild woody hosts differed greatly between the olive groves. These differences could probably be explained by considering the diverse availability of alternative host plants for the spittlebugs and the heterogeneous spatial structure of the considered olive groves. For instance, the densities of P. spumarius in the coastal olive grove of Liguria were high on shrubs and trees, and were similar to those registered on the herbaceous cover, while only a few meadow spittlebugs were collected on wild woody hosts throughout the entire sampling season in the inland olive grove. As far as the preference of P. spumarius for woody plants is concerned, the highest abundance was reported on Quercus spp., Pistacia lentiscus L. and P. terebinthus L. in Liguria and on Quercus spp. and Myrtus communis L. in Apulia.
In late summer, the number of adults on the herbaceous cover tended to increase again, probably due to the return of spittlebugs from aestivation sites (e.g. wild woody hosts inside and outside the agroecosystems). This return was considerable, particularly in the Liguria olive groves, where similar spittlebug densities to those observed in spring were recorded in autumn (mean density of up to 1.47 adults/SSU h ). The densities of P. spumarius in Apulia were lower (mean value up to 0.33 adults/SSU h ), probably because of a higher adult mortality rate during summer (Figs. 4 and 5). The overall number of adults sampled in the olive grove agroecosystem decreased again in late October or early November, with very few individuals being recorded in late autumn. The last adults in Liguria were sampled in late November or December. In Apulia, the samplings had often ended by late October, even though a few spittlebugs were still present in the olive groves. Indeed, some P. spumarius adults were collected in the same olive groves in late winter, thus proving that a few adults may overwinter in South Italy (Dr. V. Cavalieri, unpublished data). www.nature.com/scientificreports www.nature.com/scientificreports/ The population peaks of the adult stage were quite similar in the Apulia and Liguria regions, although there were important differences between the olive groves. The differences in P. spumarius density between the two regions might be due to several reasons, amongst which: i) intrinsic differences in suitability for spittlebugs of specific olive groves, considering the surrounding habitats, i.e. landscape composition, which can provide alternative habitats for spittlebug adults, thus contributing towards maintaining high population levels of vectors in olive groves through seasonal mass movements; ii) different levels of insect disturbance, as a result of agronomic  www.nature.com/scientificreports www.nature.com/scientificreports/ measures, such as soil tillage -which is usually carried out in summer in Apulia -can have an impact on the adult population, thereby determining mass movements to olive trees and other woody hosts, and may influence the oviposition suitability of certain sites, thus impacting the following year's juvenile populations. Given the different ecological situation of the sites inspected in the Liguria and Apulia regions, we hypothesize that the P. spumarius population abundance was higher in Liguria mainly because of the landscape structure (much more fragmented in Liguria) and the richness of non-cultivated refuge areas, which can represent a more suitable environment for spittlebugs 58 . Again in this case, the more intensive agricultural exploitation of the Bari province area, both at a landscape level and because of the use of more intense agronomic measures, could account for the lower population level of the spittlebug registered in our study compared to that found in previous studies performed in the Lecce Province (South Apulia) 44,59 .
Sex ratio. The sex ratio of P. spumarius adults, determined over the whole season, was slightly male biased in the Apulian olive groves (male proportion = 0.54), while it was female biased in the Ligurian sites (male proportion = 0.41). A significantly greater proportion of males was recorded during the adult emergence period, especially in Apulia. The proportion of males tended to decrease during the adult season in all the olive groves, until a parity of male / female proportions was reached at the end of the emergence period (mid-May) (Fig. 6). Subsequently, the proportion of males continued to decrease throughout the year, although it fluctuated to a great extent during summer, with a few or no P. spumarius males being collected on the last sampling dates. No differences were registered among the vegetation compartments. Therefore, the sex ratio varied slightly, although it was statistically significant, between the regions, olive groves and sampling dates, but with no significant interactions (Table 5). Males appeared earlier in spring, probably because their preimaginal development was faster  Table 4. First, last and peak dates and degree-days (DD) of the Philaenus spumarius adult stage (peak dates = max no. of adults sampled).  Table 5). The horizontal dash-dotted line shows sex ratio equality (0.5).  Table 7. Dates of the first, last and peak occurrences of the different instars of Aphrophora alni (peak dates = max no. of insects sampled).  www.nature.com/scientificreports www.nature.com/scientificreports/ than that of the females, as already reported in previous studies 21,60 . The sex ratio became female biased throughout the summer season, probably because of a shorter longevity of the males 29,61,62 . Overall, the male proportion was significantly higher in Apulia than in Liguria. A different prevalence of the sex-manipulating proteobacteria Wolbachia in the spittlebug populations of the two Regions might account for the local sex ratio differences 63 . other spittlebug species. The nymphal populations of N. campestris followed a similar trend as P. spumarius, although lower abundance levels were observed. The mean peak densities ranged between 1 and 16 nymphs m −2 . The highest nymphal density was registered in 2017 in the inland olive groves in both Apulia and Liguria. The nymphal stages spanned from mid-March to May, and peaked during April ( Fig. 7 and Table 6). Adults started to emerge earlier in Apulia, that is, mid to Late-April, compared to Liguria, where they started to emerge mid-May (Table 6). Aphrophora alni juveniles were collected from mid-March to late May on the herbaceous ground cover, albeit only in the Ligurian olive groves ( Table 7). The population densities were lower than 2 individuals m −2 throughout spring, except for a 5 th instar peak in late May 2016 in the inland olive grove (Arnasco), which reached 7.61 ± 2.40 nymphs m −2 . The mean density of the A. alni nymphs in the two Ligurian olive groves was higher in 2016 (1.05 ± 0.16 nymphs m −2 ) than in the 2017 survey (0.17 ± 0.04 nymphs m −2 ) (Fig. 8).
The adult populations of N. campestris followed a different pattern throughout the year from P. spumarius. Adults were mainly collected in the herbaceous cover, especially where grasses (i.e. Poaceae) were dominant. The highest abundances of N. campestris adults were registered on the herbaceous cover in Apulia in late May, soon after the end of the nymphal stages, whereas the highest population levels in Liguria were registered during autumn. N. campestris adults were not usually found during summer in any of the vegetation compartments of the olive agroecosystem in either Apulia or Liguria (Figs. 9 and 10). Such an absence of N. campestris in the olive groves is probably due to the movement of spittlebugs to aestivation host plants, that is, usually conifer trees 64 . Aphrophora alni adults were only sampled in Liguria, mainly on woody plants, on both olive trees and wild broadleaved trees (e.g. Quercus spp.). The population abundance levels were higher in 2016 than in 2017 in both olive groves (the mean number of individuals collected per sampling date in 2016 was 14.4 ± 3.70, while it was 3.85 ± 1.30 in 2017). The highest densities on olive trees were registered in June (up to 0.85 ± 0.25 adults/olive) and they then decreased during summer. The populations on wild woody plants peaked later, in July/August, with up to 1.4 ± 0.50 adults/tree. In 2016, the A. alni density on wild trees remained quite constant until late summer. A few individuals were occasionally sampled in autumn (Fig. 11). www.nature.com/scientificreports www.nature.com/scientificreports/ Neophilaenus campestris showed lower population levels than P. spumarius in both regions for both nymphs and adults, as already reported for other olive groves in Apulia 44,59 and in Greece 41,42 . Conversely, N. campestris has been found to be the predominant spittlebug species in olive groves in Spain, although dramatic fluctuations of population levels have been observed between olive orchards 43 . The chronological time of nymphal instars or adult appearance of N. campestris was quite similar to those of P. spumarius, although a certain time lag of the appearance of N. campestris nymphs was registered (Fig. 7). Such a delay could have partly been caused by the lower abundance of the former, and thus by the lower probability of sampling its nymphs at the very beginning of nymphal development 65 . Moreover, the onset of nymphal development of A. alni was unclear, given the low densities that were observed, but adults only started to emerge in late May, later than P. spumarius. To the best of our knowledge, no other information is available on the nymphal instar phenology of N. campestris and A. alni. Only a few investigations have been carried out on some Neophilaenus species in the UK 34,53,65 and in Germany 38 , in which nymphal development between late April and mid-June has been reported, that is, about one month later than for N. campestris under Mediterranean conditions. conclusions. A three-year study has been conducted under field conditions in order to provide a detailed description of the population dynamics, abundance and phenology of spittlebugs in Mediterranean olive groves. The study contributes to fill the knowledge gap on the ecology and biology of this group of insects and to pave the way towards a better understanding of the epidemiology of Xf-associated diseases in the Mediterranean basin, particularly in olive groves. Different xylem feeder species, all belonging to the Cercopoidea superfamily (spittlebugs) were collected during the three-year field survey. No sharpshooters (Hemiptera: Cicadellidae Cicadellinae) were found in our sampling campaigns. This result confirms what has already been reported for other olive groves in the Mediterranean region 41,43,59 . However, the presence of sharpshooter species in some olive agroecosystems cannot be ruled out. Indeed, a few Cicadella viridis L. adults have been collected in olive groves in the Italian Abruzzo 66 and Veneto Regions (Nicola Mori, Personal communication), and in Turkey 67 . In our study, two xylem-sap feeder species were predominant, P. spumarius and N. campestris. A third spittlebug species, A. alni, was only present in the Liguria olive groves. Three other spittlebug species, L. coleoptrata, C. vulnerata and C. sanguinolenta, were also recorded, albeit all in low numbers and for a restricted time span. Although olive groves represent a rather biodiverse agroecosystem, as the olive is a perennial crop and the grove has a ground cover of herbaceous plants, the richness of spittlebug species in olive groves seems to be limited. Data from other surveys carried out in olive groves [40][41][42]59,68 and in vineyards 69,70 also indicate the presence of a limited biodiversity of this taxon under cropping systems, compared to biodiversity registered in more natural grassland habitats in Northern Europe 37,71,72 . Our results on the phenology and abundance of the nymphal and adult stages of spittlebugs provide useful indications for vector control. The description of stage-structured populations of nymphs and adults allows to www.nature.com/scientificreports www.nature.com/scientificreports/ identify the best time to apply control. According to our observations, the newly hatched nymphs (1 st instar) always disappeared just before the peak of the 4 th instar nymphs, and the first adults were only sampled after this peak. Therefore, any control measure applied at the 4 th instar peak could potentially target the whole nymphal population before the onset of the adults, thus achieving the maximum efficacy. A first abundance peak was observed for the adults soon after emergence in all the olive groves on both the herbaceous cover and on the olive trees. These considerations, which are similar to those of Cornara et al. 9 , suggest that the period immediately after adult emergence is the crucial moment for both X. fastidiosa acquisition and transmission to olives by insect vectors. It is worth pointing out that, once infected, vectors are persistently infectious. Insecticides targeting the adult stage should be applied timely to the olive canopy, mainly in this period, in order to prevent the spread of the disease during the year 11,73,74 . Hence, transmission may occur all over the summer and at the beginning of autumn, even though a limited presence of adults was recorded on olive canopies later in the season. In Liguria − a Xf-free area − P. spumarius adults were collected in relatively high numbers on olives from their emergence to the end of September, as has also been noted in other regions in Northern Italy (Pavan 27 and authors' unpublished data). Indeed, where the summer conditions are less severe (e.g. Northern Europe), spittlebug adults can be collected over the whole summer under cropping and non-cropping systems 21,65,75 . The extended period of the presence of P. spumarius on olive trees in Liguria, together with the high populations of recorded spittlebugs, suggests that the eventual introduction of a strain of X. fastidiosa infecting olive plants in this area could potentially lead to a dramatic outbreak, although vectors are only one of the actors involved in epidemics. Furthermore, the difference in composition of spittlebug communities in Mediterranean olive groves could influence the epidemiology of X. fastidiosa. Mediterranean regions with lower population levels of P. spumarius than those of Neophilaenus spp, e.g. Spain 43,64 , could present less inoculum pressure on olive trees and a subsequent epidemic spread of the bacterium, since Neophilaenus spp. are rarely collected on olive canopies, as demonstrated in this study and others 10,42 .
The summer colonisation of wild shrubs/trees inside and around the olive agroecosystem by spittlebugs, and by P. spumarius and A. alni in particular, highlights the role of these plants as reservoirs of potential vector insects. Some of these plants, e.g. Quercus ilex, Rhamnus alaternus, Laurus nobilis, Phyllirea latifolia and Myrtus communis, can also act as X. fastidiosa ST53 reservoirs 10 . The colonisation of these alternative host plants by spittlebugs is possible as they are highly polyphagous, and wild plants might be more suitable (e.g. less water stressed) during the dry summer months than olive trees. The aestivation period and/or local migration towards natural or semi-natural patches of these wild host plants could occur, as already noticed in other hot Mediterranean   www.nature.com/scientificreports www.nature.com/scientificreports/ areas 18,41,43 . In autumn, the adult populations increased on the herbaceous ground cover, which represents the oviposition site, thus confirming that most of the spittlebugs did not die during summer. The spatial distribution of spittlebugs throughout the season, inside and outside the olive groves, needs dedicated studies, since knowledge on this topic could contribute towards the understanding of the spread of epidemics and help in the designing of an effective management of vector populations 76 . Similarly, the investigation of the role of the landscape composition may help predict P. spumarius population levels in specific areas 58 . Such a prediction is needed to assess the risk to new plantings of Xf-susceptible crops in infected areas, as well as to evaluate the probability of the establishment and spread of the bacterium in the case of its introduction into new areas.
The findings of our study underline the great variability of population levels of potential X. fastidiosa vectors in olive groves between and within different Mediterranean regions. There is still a general lack of information on both the abiotic and biotic factors that influence the composition of xylem-sap feeder communities in olive groves and on species abundance. Further studies on vectors in olive, almond and other agroecosystems potentially at risk to the spread of X. fastidiosa in Europe are urgently needed to improve control efforts and to contribute towards limiting the spread of X. fastidiosa epidemics.