Performance of a Leaf-Galling Phylloxera ( Daktulosphaira vitifoliae ) on Roots of Diverse Vitis spp. Rootstocks in North East Victoria, Australia

,


Introduction
Grapevine phylloxera, Daktulosphaira vitifoliae (Fitch), is a pest of major economic signifcance to viticulture worldwide. Phylloxera feed on roots and leaves of Vitis spp. [1,2]. Te phylloxera root-feeding form, called the radicicole is typical on ungrafted cultivated varieties of Vitis vinifera (L.) ssp. Sativa throughout the world. On young nonlignifed root tips, phylloxera root feeders induce hooked galls (called nodosities) and malformed swellings (called tuberosities) on older lignifed roots. Nodosities and tuberosities are a source of nutrition and sites for oviposition and larval development [3][4][5]. Individual insects attach at feedings sites and develop through four intermediate stages to adults that lay up to 400 eggs [6]. Nodosities impede root growth and tuberosities cause splits on the roots. Te splits lead to decay because of secondary fungal infections eventually causing vine death [7,8]. Te leaf-feeding form, called the gallicole, not only feeds on leaves of North American Vitis species [2,9] but can also attack leaves of V. vinifera [1,[10][11][12].
Te phylloxera life cycle is complex and varies in different viticultural environments. High genotypic diversity and reproduction that is predominantly asexual have been revealed in D. vitifoliae populations in Europe [3] and Australia [13][14][15]. In Europe, separate introductions from the North East coast of North America can be traced back to two genetic groups of the dominant American species V. riparia and V. labrusca [16]. A signifcant deviation from Hardy-Weinberg equilibrium suggests that parthenogenesis is the primary reproductive mode of reproduction for Californian D. vitifoliae populations [17]. Te asexual females feed on either roots (radicicole) or leaves (gallicole). During the growth season, gallicoles and radicicoles undergo several asexual generations where the parthenogenetic females lay several hundred eggs [3,18,19]. Te frst instars are mobile and move to the leaves and/or roots where they establish new feeding sites on growing shoots or young root tips. Radicicoles overwinter as frst instar nymphs under the bark of roots [18,20] and can give rise to winged adults (called alates) that emerge from the ground to disperse and fnd new feeding sites.
In Australia, management of phylloxera is achieved in part through quarantine [21]. For long term management, North American Vitis spp. that have coevolved with phylloxera are broadly used due to host responses that resist or tolerate infestations [2,21]. Lack of complete resistance and, in some cases, potential breakdown in resistance limits the uptake and durability of rootstocks. Studies in Europe and the USA indicate that biotype-C phylloxera is adapted to feeding on rootstocks with the riparia parentage, including the Teleki rootstocks, 101-14 and Schwarzmann [22]. As a result, vineyard productivity in some regions may be lower where biotype-C populations predominate. To date, several phylloxera resistant loci have been identifed in the Borner rootstock [23], V. cinerea C2-50 [24], Muscadine rotundifolia, and a complex hybrid [25]. DNA-markers linked to these phylloxera resistance loci can now be used to breed durable resistant rootstocks harbouring two or more resistant loci [26]. Aside from phylloxera resistance, rootstocks are also used to counteract other soilborne pests, such as nematodes, as well as maintain vine productivity in response to abiotic stress, soil pH, and porosity [27].
Phylloxera resistance or tolerance attributes of rootstocks is determined by insect genotype-plant interactions [28,29]. On tolerant rootstocks, phylloxera can feed and then reproduce and maintain a population for several years without causing yield loss or vine death [30]. Previous rootstock screening studies in Australia have focused on six root-galling phylloxera strains using a triphasic screening approach [28,31]. Tis encompasses a suite of protocols that integrate laboratory (excised roots) and greenhouse (potted vines) trials to screen new and existing rootstocks for phylloxera resistance [1,21] and to analyse the ability of phylloxera to survive and develop on rootstocks [1,32]. Small-scale feld trials in commercial vineyards are conducted where possible [22,33,34]. Together, this information is incorporated into management decisionmaking tools, such as the Grapevine Rootstock Selector, to provide information that enables growers to select rootstocks that are resistant or tolerant to both phylloxera and root-knot nematodes [35].
In Australia, phylloxera has been established in the North East Victoria Phylloxera Infested Zone (PIZ) since the early 1900s [13,14,36]. Most vineyards are infested by parthenogenetic lineages that feed on roots, and there is insufcient evidence to support the classical life cycle [37]. Gallicoles predominate on Vitis spp. in North America, Europe, and South America [3,19,38,39]. In Australia, gallicoles have only been observed infrequently in the North East Victoria Phylloxera Infested Zone (PIZ) [37,40]. Te winter egg has been recorded only once in Australia [41], and it is not clear how the leaf-galling populations are sustained from year to year. Corrie and Hofmann [37] found some genotypes of phylloxera on leaves and roots with evidence that those on roots arose from sexual reproduction in leaves. Eighty-three diverse phylloxera genetic strains were characterised in the early 2000s, of which 49 were identifed as root feeders, 23 leaf feeders, and 11 that feed on both leaves and roots [14,42]. Recent extensive surveys on vine roots conducted in the King Valley region of the North East Victoria PIZ, where only G4 phylloxera strain had been reported [13], identifed 32 newly characterised genetic strains [15].
Tis publication reports on a case study conducted over three consecutive years (2015-2018) in a commercial vineyard named BGM located 274 km north-east of Melbourne, Australia in the North East Victoria Phylloxera Infested Zone (PIZ) [36]. Tis site was of interest because of the detection of G38 phylloxera strain previously reported only as leaf-galling under feld conditions. Te G38 phylloxera was frst characterised from a leaf-galling sample that was collected near the Warby Ranges and Mount Glenrowan south east Australia on the AxR#1 rootstock [13,37,43]. Te results of that study implied that G38 could be the outcome of sexual reproduction between individuals of G2 and G3 phylloxera that were sampled from both leaf and root [43].
Te vines in this study site were planted in 1988 on uncultivated soil without a previous history of grapevine cultivation. Te vines used in this study were in four rows, planted in a randomised complete block trial in which the plots were V. vinifera on own roots and grafted to diferent rootstocks. In the summer of 2015, the vines were observed to have yellowing leaf symptoms comparable to phylloxera infestations for the frst time. Root inspection revealed nodosities and yellow clusters similar to phylloxera colonies on several rootstock cultivars and V. vinifera. Te insects from the roots were collected for morphological and molecular identifcation. Tey were confrmed as phylloxera and characterised as the G38 genetic strain using six nuclear DNA microsatellite markers [14,44]. No leaf galls were observed on the canopy of vines. Te detection of the G38 phylloxera strain on roots at a site with existing replicates of V. vinifera vines on own roots or on diferent rootstocks ofered a serendipitous opportunity to study the survival and development of the G38 phylloxera strain under feld conditions and subsequently under controlled environments using insects collected from the BGM vineyard.

Insect Stock Cultures.
In February 2016, adults, eggs, and frst instars of the G38 phylloxera strain were randomly collected from roots of grafted rootstocks and V. vinifera vines in the BGM vineyard described in the introduction.
Te insects were collected from four rows where the G38 phylloxera strain was frst detected. Several single adult lineages of the G38 phylloxera were cultured and established in the laboratory under controlled conditions (25 ± 2°C; 12 h L : D). Te genetic lineage of the insects was confrmed as the G38 phylloxera strain using six nuclear DNA microsatellite markers as per methods by Umina et al. [14] and Agarwal et al. [44] before the cultures were multiplied.
To produce a stock culture for laboratory and glasshouse experiments, insects were maintained under quarantine and mass reared on excised roots of V. vinifera cv. Chardonnay as per methods by Kingston [45].

Description of Vines in Four Rows at BGM Vineyard Used
for In-Field Assessments. Te four rows of vines used for infeld assessments (Table S1) were established in 1991 from callused rootstock cuttings of the varieties Schwarzmann, 101-14, 3309 Courderc, 125AA Kober, 1103 Paulsen, 5A Teleki, SO4, R99 and Sori and callused cuttings of own root Cabernet Sauvignon (clone LC10), were established through 600 mm wide polythene flm in 1991. Te rootstock cuttings were sourced from the Murray Valley Vine Improvement Association; and the Cabernet Sauvignon (clone LC10) were sourced from a local vineyard in North East Victoria. In the following season the rootstocks were subsequently feld grafted to Cabernet Sauvignon (clone LC10) and trained up to form a bilateral cordon. Te own rooted Cabernet Sauvignon vines were also trained up in the same season to the equivalent system. Both the Cabernet grafted rootstock vines and the own rooted Cabernet Sauvignon vines were reworked to Sangiovese (clone MAT 7) in the spring of 2005.

Plant Material Used for Laboratory and Glasshouse
Experiments. Vines used for controlled glasshouse and laboratory experiments were sourced as dormant cuttings from the Yalumba Nursery, Nuriootpa, South Australia which is in a Phylloxera Exclusion Zone and free from phylloxera. Te cuttings were planted in 4.5L plastic black pots using heat sterilised 80% general purpose potting mix (Spotswood Potting Mixes and Fertilisers, Yarra Glen, Victoria) and 20% perlite (Peards Nursery, Albury, NSW) and fertilised with  Table 1). Once planted, vines were kept in a shade house and drip irrigated for two min daily over 12 months to allow optimal root development before inoculating with G38 phylloxera eggs from stock cultures.  [48]. To install the traps, a container was placed on levelled ground at approximately 10 cm from the base of trunk on each of the experimental vine and secured with metal tent pegs.

In-Field
Trapping commenced over the vegetation period from mid-Spring until mid-Autumn (November-April) of 2015/ 2016 following the frst detection of phylloxera on roots of vines in the vineyard and continued for three subsequent summer seasons (2015/2016, 2016/2017, and 2017/2018) during phylloxera peak activity [49]. Vines on which traps were placed were randomly selected from four rows of a block of mixed rootstock and own rooted vines (V. vinifera) ( Figure 1). One trap was placed per vine. Eight to ten replicate vines were selected for each rootstock cultivar and V. vinifera. Ten Vitis varieties were selected for the study (Table S1).
Emergence trap samples were collected once every month over the three seasons. Sample collection involved rinsing the traps condensate with 80% ethanol and pouring contents into sampling containers. Te traps were thereafter rinsed with tap water and replaced. Trap samples were transported to a quarantine laboratory at Agriculture Victoria, Rutherglen. Te samples were observed under a low power dissecting microscope and frst instars and alate adults found in each trap were quantifed (Table S2). Total number of frst instars and alates in traps over the three years were summed up for each Vitis spp. cultivar.

Root Inspection-Confrming Presence/Absence of Phylloxera on Vines
Roots. Insects may have been caught in traps placed on resistant vines through the movement of frst instars along the roots from nearby susceptible vines. Roots were, therefore, visually inspected to record the presence or absence of phylloxera life stages from a sample of vines where traps were placed for each Vitis spp. Visual detection of adults with eggs and 2 nd -4 th instars intermediate stages confrmed the successful development and completion of an asexual generation (Table S3). To conduct the visual assessment, a quantitative destructive technique was used. A dozen lignifed and nonlignifed root sections (5-30 cm in length) per vine were dug up to a depth of 5−20 cm and roots and soil samples were collected. Roots of eight replicate vines per rootstock type and V. vinifera on own roots were visually inspected. Te samples were transported to a quarantine laboratory at Agriculture Victoria, Rutherglen where they were inspected for presence of phylloxera life stages with the aid of a low power dissecting microscope. Visual leaf symptoms noted as chlorosis on leaves were observed at veraison on experimental vines during the third year of study ( Figure S1).

Performance of G38 Phylloxera in Potted Vines.
Te development of G38 phylloxera on diverse Vitis spp. was studied using potted vines. Te experiment used a completely randomised block design with eight replicate vines Australian Journal of Grape and Wine Research   Table 1). Sori and Richter 99 were not available for the potted trials. Te vines were obtained from Yalumba nursery, Nuriootpa in South Australia. Vines were infested with the G38 phylloxera as per methods by Forneck et al. [1] and Korosi et al. [31]. Briefy, the vines were removed from pots and a lignifed root (5 mm in diameter) and surrounding fbrous roots enclosed in a muslin cloth. Te roots were inoculated with 20 eggs of G38 phylloxera, and a small amount of potting mix immediately placed gently around the roots. Te ends of the cloth were carefully tied with a cable tie to enclose the eggs, roots, and soil into a "root pocket." A sticky gumbased insect barrier, Tanglefoot ™ , was applied along the edges of the cloth to stop phylloxera from escaping and the vine repotted by topping up with a potting mix and perlite mixture. Te root pockets were repotted with the rest of the roots and the vines thereafter fertilised with 3.5 g Osmocote ™ and 500 ml Trive ™ per potted vine. Te vines were maintained in the glasshouses at 22°± 2°C and a relative humidity of 60-70% and drip irrigated for 2 min daily.
Te experiment ran for eight weeks. Vines were removed from pots and the root pocket snipped to assess phylloxera development and damage on roots. Assessments were done under a low power dissecting microscope. Te following parameters were calculated: (i) total insects summed up as the phylloxera developmental stages (eggs, frst instars, intermediate instars, adults and alate); (ii) count of nodosities; and (iii) counts of tuberosities (Table S4).

Performance of G38 Phylloxera on Excised Roots.
Roots of 10 Vitis varieties (V. vinifera, Ramsey, Schwarzmann, 101-14, 3309 Courderc, 5BB Kober, 110 Richter, 1103 Paulsen, and 140 Ruggeri and Börner) were cut into 4-5 cm lengths and placed in plastic Petri dishes (90 × 25 mm). Ten 1-4-day old eggs of G38 phylloxera were placed on a root piece using an artists' paint brush. A Petri dish with one excised root and 10 eggs was considered a replicate. Each treatment (rootstocks and V. vinifera) consisted of fve replicates. Petri dishes were sealed with cling flm and then aluminium foil to restrict light fltering through and maintained in the growth room at 25°C ± 2°C. Over the entire experiment, cotton wool on the end of the roots was saturated with 2 mL ultrapure water once every week to keep the roots hydrated and viable.
Te number of eggs that hatched was recorded at day 8. First, instars that established feeding sites and subsequently survived to adults were recorded at days 18, 25 and 32. Adults were distinguishable from intermediates if they had eggs in proximity. Eggs that were laid on days 18 and 25 were counted and removed from the roots to discount for overlapping generations. Parameters were calculated as follows: (i) survival was calculated as the number of adults that developed from the initial eggs inoculated on roots; (ii) average fecundity was calculated as the number of eggs produced per adult per day after the frst adult detection; (iii) gross fecundity was calculated at day 32 as the total eggs laid over the lifetime of all adults per root. Tis was the sum of Australian Journal of Grape and Wine Research the number of adults in each root for each of days 18, 25 and 32; (iv) proportion of inoculated eggs reaching the adult stage was calculated as adults at day 32 out of total eggs inoculated on roots (n � 10) (Table S5).

Statistical Analysis. Generalised Linear Models (GLMs)
with Poisson distributions were used to compare the total number of insects caught in traps over the three years across the Vitis spp. A GLM with Poisson distribution was used to compare trap catches over three years with the trapping year and rootstock ftted I the model. A GLM was also used to compare the total number of nodosities, and total number of insects (eggs, intermediates, and adults) recorded among Vitis spp varieties in pots, net and gross fecundity among excised root varieties. Te proportion of eggs that reached the adult stage among excised root varieties was analysed using a GLM and model was ftted with a binomial distribution. All models were built in R using the "MASS" package and evaluated using loglikelihood tests in the "epiDisplay" package.
Tukey's post hoc tests with corrections were used to compare diferent treatments using the "lsmeans" package where our GLM model was signifcantly diferent to a null model as determined by the loglikelihood tests. All residual values were examined graphically to ensure normality and homogeneity of variances. Observations with standardised residuals greater than 3.0 were excluded from analyses.
For observations from roots, the number of phylloxera caught in traps over the three years was modelled using logistic regression, and the odds ratio was calculated. From the odds ratio, the probability of phylloxera reaching adulthood for each rootstock was calculated using the formula, p � (e z /1 + e z ) where p � probability and z is the relevant odds ratio [50]. Te analysis was performed in GenStat 18 th Edition [51].
For each trial type, phylloxera survival from eggs to adults on the vines was used to defne rootstock resistance and susceptibility ratings. Rootstocks were rated resistant if neither insect nor nodosities and tuberosities were present on roots, tolerant if nodosities, frst instars at low numbers (<5) and intermediate stages without adults were present on roots and susceptible if nodosities and tuberosities and all phylloxera life stages were present on roots.

In-Field Assessments of G38 Phylloxera
3.1.1. Emergence Traps. Phylloxera was captured in emergence traps across all rootstock treatments during the study ( Table 2). Te GLM revealed that the total number of insects caught in traps was signifcantly diferent across the vine cultivars (χ 2 � 44503; d.f � 9; p < 0.0001). Traps placed at the 101-14 vines consistently caught the highest numbers of phylloxera over the three years (Table 2). In comparison, less than 10 frst instars were caught in traps collected from the 1103 Paulsen and Richter 99 rootstock vines (Table 2). Tere were no statistical diferences between trap catches for Sori, Schwarzmann, 5A Teleki, 3309 Couderc, SO4, 125AA Kober, and V. vinifera (Table 2). Alates were caught in trap samples from V. vinifera and all rootstock cultivars except for 1103 Paulsen and Richter 99 (data not shown).
Te number of insects captured in traps increased signifcantly over the three seasons following the initial detection of the G38 phylloxera and diferences were attributable to vine cultivar (χ 2 � 3973; d.f � 29; p < 0.001; Figure 1). Traps placed on vines with the 101-14 rootstock captured highest number of insects during the third year of the study, followed by Schwarzmann, 3309 Courderc, Sori, Teleki 5A, SO4, and 125AA Kober (Figure 1). Collectively across all the rootstocks and V. vinifera, phylloxera catches in traps were highest in December and April (data not shown), indicative of multiple generations per season.

Root Inspection-Confrming Presence/Absence of G38
Phylloxera Stages. We confrmed successful development from the egg to the adult stage of G38 on each rootstock cultivar. Te G38 phylloxera successfully developed and reproduced on all rootstock cultivars except for Richter 99 ( Table 3). Scions of V. vinifera, Schwarzmann, 101-14, Sori, 3309 Courderc, 125AA Kober, and SO4 showed yellowing symptoms consistent of those of phylloxera infestations by the third year of study (Table 3; Figure S1).

Susceptibility
Rating. G38 phylloxera reproduced freely on 101-14 under feld conditions, consistent with these vines being classifed as susceptible to this strain (Table 1). Te rootstocks Sori, Schwarzmann, 5A Teleki, 3309 Couderc, SO4, and 125AA Kober appeared to tolerant the G38 phylloxera strain. Traps placed on vines with rootstocks 1103 Paulsen and Richter 99 caught the least number of insects and thus appear to be resistant to the G38 phylloxera strain (Table 1). Furthermore, the probability of G38 phylloxera completing an asexual generation on 1103 Paulsen and Richter 99 rootstocks was negligible.
Neither nodosities nor reproductive adults were observed on potted vines of rootstocks 110 Richter and 1103 which further supports classifcation of these rootstocks as being resistant to G38 phylloxera (Table 1). Survival and completion of an asexual generation by the G38 phylloxera accompanied by nodosities and/or tuberosities upon root inspection were recorded for 3309 Courderc, Schwarzmann, Sori, and 101-14, with means being higher than those recorded for V. vinifera on own roots. Potted vines of these rootstocks are, therefore, likely to tolerate G38 phylloxera. Survival and reproduction of the G38 phylloxera on excised roots was greatest for rootstock 101-14 compared to the other nine rootstocks tested, consistent with this rootstock being classifed as susceptible to strain G38.

Discussion
Tis study reports that G38 phylloxera, which is known to be a leaf-galling strain also occurs as root galling on diverse rootstocks and V. vinifera. Tis fnding presented an opportunity to examine the performance of a phylloxera strain that feeds on both roots and leaves on diverse rootstocks, employing the triphasic screening approach, which utilises feld trials together with more controlled laboratory and greenhouse trials. No leaf galls were observed on vines in the block under study during the three years. An important outcome from the feld trial was that the rootstock cultivars, 101-14, 3309 Courderc, Sori, 5A Teleki, and Schwarzmann, Table 4: Mean (±standard errors) of nodosities, tuberosities and total count of phylloxera (eggs, intermediates and adults) on roots of eleven Vitis spp (n � 8). Te roots were contained in muslin cloth pockets and cocultivated with the G38 phylloxera for 8 weeks.
Results from the three types of assays used in this study demonstrated that the rootstock 101-14 is susceptible to G38 phylloxera, and this phylloxera strain may be selected under diverse rootstock plantings [58][59][60]. Using excised root bioassays, the gross fecundity of G38 phylloxera was highest on 101-14 rootstock compared to the other nine Vitis varieties tested. Tough excised root bioassays do not indicate a complete plant response, they are considered to be a suitable measure for phylloxera biotyping [55]. Results suggest that the G38 phylloxera possibly belongs to the biotype-C group, which is a classifcation that includes strains that show superior performance on nodosities on roots of rootstocks derived from American Vitis spp. such as 5C Teleki and 101-14 and 3309 Courderc and reduced ability to establish on V. vinifera roots [22,61,62]. Rootstocks that had very low populations, such as was observed in 1103 Paulsen, 140 Ruggeri, Ramsey, 5BB Kober, and 110 Richter using potted vines assays, may use hypersensitivity as a resistance mechanism [63]. Te root response to phylloxera expressed as tissue browning, indicative that the oxidation of phenolic compounds and nodosities are unsuitable as feeding sites for developing nymphs [64].
Rootstocks with parentages that are susceptible to certain phylloxera types have caused enormous economic losses.
For instance, the rootstock AXR#1 (V. vinifera × V. rupestris) succumbed to infestations of virulent phylloxera biotypes in California, causing loss in production of between US$1 billion and $6 billion [65,66]. Te failure of AXR#1 led to caution concerning the use of rootstocks with insufcient tolerance to phylloxera, especially those with parentages that are highly susceptible. In this study, several rootstock cultivars had the riparia parentage, but it appears that only those with V. riparia × V. rupestris were susceptible or tolerant to the G38 phylloxera. Te 1103 Paulsen and Richter 99 both with V. berlandieri × V. rupestris parentage had low insect populations consistently across the three experimental methods and are potentially tolerant to the G38 phylloxera. Our fndings corroborate those in California that have shown phylloxera biotypes adapted to feeding on rootstocks with V. riparia parentage, such as 101-14 and Schwarzmann and some Teleki hybrids [22,67,68]. Tere is evidence to suggest that two genes are involved in nodosity formation based on fndings from a study of traits related to phylloxera susceptibility from the V. vinifera and V. rupestris hybrid AXR1 [69]. Tere is also a report of a hybrid of V. riparia and V. cinerea Arnold that shows high resistance to phylloxera [23].
Te emergence of biotype-C genetic strains that are adapted to feeding on V. riparia rootstocks has been highlighted as a concern when selecting rootstocks due to the potential for resistance breakdown [24]. Te ability of phylloxera strains to utilise rootstocks as a food source and their interactions with hybrids such as V. riparia in conferring resistance needs to be studied further to enable access to suitable traits for rootstock breeding programs. Our understanding of phylloxera genetic diversity in Australia is limited by lack of extensive surveys, especially in regions where phylloxera has existed since the early 1900s, which could explain why G38 phylloxera may not have been found on roots before. Recent surveys in the King Valley region, however, found a higher phylloxera genetic diversity than was previously thought [15].
In the current study, Börner rootstock, a hybrid from V. riparia and V. cinerea "Gm183," was found to tolerate infestations by G38 phylloxera in potted vines and excised roots assays. Generally, Börner is considered to resist phylloxera infestations [23,70,71] and high level resistance has been demonstrated against some phylloxera genetic strains [21,28,72,73]. Tis result is consistent with the fndings of Powell [21] and Powell and Krstic [28] who found Börner to be tolerant to the root-galling G19 phylloxera strain under feld assessments and tolerant to the G30 phylloxera strain in vitro. In the study of Powell and Krstic [28], Börner had higher numbers of G19 phylloxera in traps compared to other rootstocks and V. vinifera under feld conditions. Since Börner is considered an attractive genetic resource for rootstock breeding due to its good grafting characteristics and adaptation to varied soil types [24,72,74], further studies are needed to confrm the contribution of V. riparia to incomplete resistance of plant accessions to some phylloxera strains such as the G38.
Screening for resistance of rootstocks to diferent phylloxera strains forms an essential component of rootstock selection and provides planting recommendations to the grape and wine industry. Te three screening methods employed each have their disadvantages and advantages, such as ease of replication and/or applicability to the feld, and diferences in overall fndings have been observed in the past when comparing the three screening methods [15,28,61]. In this study, rootstocks such as 1103 Paulsen and 110 Richter led to diferent fndings when the results of the three trials were compared, implying that environmental conditions in the feld could afect the response of these rootstocks to phylloxera. Potted vines always had optimal water content and relative humidity was high compared to likely conditions in the feld. We did, however, observe necrotic nodosities on roots within the pockets, and this could contribute to restrict feeding due to deteriorating nutrition leading to insect mortality. We, therefore, recommend inoculating eggs on the roots of potted vines without the pocket enclosure. Furthermore, the G38 phylloxera was cultured on Chardonnay. Te original feeding host could have infuenced the ability of the G38 phylloxera to feed on diferent Vitis spp. varieties [75].
Te reliance of the Australian grape and wine industry on a few popular rootstock varieties is of concern, as it is currently the only recommended management approach for phylloxera. For instance, growers favouring 101-14 risk having their rootstocks succumb to infestations by phylloxera genetic strains such as G38 phylloxera, which show relatively high virulence and preference for particular rootstocks. With the uptake of new rootstocks and growers preferring certain rootstocks over others, phylloxera tolerant vines-if widely planted-may inadvertently promote the spread of phylloxera feeding forms such as G38, which could spread quickly across certain rootstocks. Phylloxera strains such as G38, which exist as both root and leaf galling forms and show high levels of galling and fecundity on rootstocks, ought to be included as tests strains in screening trials that evaluate rootstock resistance. Future studies could explore whether situations exist in vineyards that might promote adaptive changes in feeding and phylloxera ftness. An improved understanding on these parameters would shape the grape rootstock breeding programs with the development of new rootstocks that are better adapted to situations in Australian vineyards. Selection of rootstocks with broad resistance to multiple phylloxera strains will also reduce the risk of phylloxera dispersal and hence strengthen quarantine.

Conclusion
Te relatively high numbers of G38 phylloxera on rootstocks indicate that host preference is a survival mechanism with potential to increase the risk of phylloxera spreading to uninfested vineyards in Australia under poor containment practices. Te introduction of unknown phylloxera lineages that feed on both leaves and roots thus presents a potential additional risk factor in the management of phylloxera. Management eforts in Australia should not only focus solely on the radicicole phylloxera forms but also the gallicole forms. Te results from this study should initiate further investigations into the persistence of phylloxera strains that exist as both root and leaf because their impact on rootstocks is likely to be underestimated due to limited research. Human assisted and potentially natural dispersal of leafgalling phylloxera forms between regions is a high risk for the industry and their management needs to be considered in existing quarantine protocols.

Data Availability
Te data used to support the fndings of this study are included within the supplementary information fles.

Conflicts of Interest
Te authors declare that there are no conficts of interest regarding the publication of this paper.

Supplementary Materials
Supplementary 1. Table S1: Four rows of grafted rootstock and V. vinifera vines at the BGM vineyard located in the North East Victoria Phylloxera Infested Zone where feld assessments were conducted. Emergence traps were placed at base of trunk of selected vines marked in bold font. Supplementary 2. Table S2: Counts of frst instars and winged adults (alates) collected in emergence traps over three summer seasons (in 2015, 2016 and 2017). Supplementary 3. Table S3: Visual inspection of roots for phylloxera life stages on selected rootstocks and V. vinifera. Visual inspections were performed by digging and assessing roots for presence of eggs, frst instars, and intermediate stages and adults (winged and wingless) in the 3rd year of study to confrm that the G38 phylloxera completed an asexual generation cycle on the vine at which traps were placed. Supplementary 4. Table S4: Counts of phylloxera eggs, frst instars, intermediates and adults, nodosities and tuberosities eight weeks after infestation with the G38 phylloxera genetic strain using potted vine experiments. Supplementary 5. Table S5: G38 phylloxera living on excised roots of various rootstock varieties and V. vinifera on days 18, 25, and 32. Supplementary 6. Figure S1: Visual symptoms of rootstock vines that were monitored for G38 phylloxera infestations using bucket traps. Images were taken at veraison during the third year of study. (Supplementary Materials)