The filter papers containing ovisacs were pinned to the underside of the aforementioned infected source plants, which were then kept in a growth chamber until the first instar mealybug crawlers hatched. Approximately 72 h after hatching on the infected source plants, mealybugs were transferred to mature vines in the vineyard and to uninfected vines in the laboratory, for a 48 h inoculation access period. The timing of hatching led us to perform field inoculations on 18 July 2011, which coincided with the emergence of the new Ps. maritimus generation in Napa Valley. Twenty replicate source vines were propagated and used, with one to five recipient test vines inoculated per source plant in each inoculation experiment . All recipient test vines were treated with an insecticide upon completion of the inoculation access period.The experimental field inoculations were located in three rows of a vineyard block of V. vinifera cv. Cabernet Franc clone 01 grafted to 110R rootstock, obtained from Duarte Nursery and planted in Oakville, Napa Valley, CA in 1994. No vines in the experimental area were symptomatic for grapevine leafroll disease prior to our experimental inoculations. To confirm initial GLRaV-3-free status prior to inoculations, three petioles were collected from each experimental vine in July 2011 before inoculations were performed, for diagnostic testing . The block consisted of 8315 vines planted at 588 vines per hectare. Row spacing was 1.8 m, and vine spacing was 1.5 m, with a vertical shoot positioning trellis system and bilateral pruning. Row direction was northwest-southeast. Drip irrigation was provided using one 3.8 – L·h–1 emitter every 1.5 m. A minimum of five buffer vines were left untreated at each end of the rows. Experimental vines were spaced every third vine, and treatments were fully randomized. The three treatments included inoculations with no leaf cages, plastic grow pots inoculations using mesh leaf cages, and negative controls for which no experimental manipulation was performed. Each treatment included 30 replicate vines, for a total of 90 experimental vines.

The experiment comprised an area including 360 total vines, including the 90 experimental vines plus the spacer vines. The spacer vines were monitored periodically throughout the study for symptoms of grapevine leafroll disease. A survey for any signs of mealybugs was performed in October 2012. On 11 October 2012, 15 months postinoculation, a commercial testing service collected and analyzed material from some vines that were symptomatic for grapevine leafroll disease in the experiment and tested for a broad panel of known grape pathogens: GLRaV-1, GLRaV-2, GLRaV-2 strain Red Globe, GLRaV-3, GLRaV-4, GLRaV-4 strain 5, GLRaV-4 strain 6, GLRaV-4 strain 9, GLRaV-7, Syrah virus 1, Grapevine virus A, GVB, Grapevine virus D, Grapevine fanleaf virus, Xylella fastidiosa, GFkV, Rupestris stem pitting-associated virus, Rupestris stem pitting-associated virus strain Syrah, and Grapevine red blotch-associated virus. For inoculations, ten Ps. maritimus first instar insects were gently moved with a paintbrush from leaves of infected source plants onto the underside of one fully expanded mid-height leaf, located on a vertical cane growing from a middle spur on the south cordon of each grapevine. For the caged treatment, a cloth mesh cage was placed over the inoculated leaf and secured at the petiole using a twist tie. For the uncaged treatment, no covering was used on the inoculated vine. The experimental area was commercially treated with spirotetramat insecticide on 20 July 2011, after a 48 h inoculation access period. After inoculations the experimental area was managed following standard commercial practices.Three months after inoculations, the petiole of the inoculated leaf was collected on 14 October 2011 for diagnostic testing. In the instance where that petiole had fallen off the vine or could not be found, a petiole near the inoculated leaf was collected; inoculated petioles were missing from 9 of 60 inoculated vines.

Immediately following the first appearance of symptoms in 2012 and 2013, petioles were collected from each experimental vine and tested for presence of GLRaV-3. Petioles were collected from each experimental vine in September 2014, and tested for the presence of GLRaV-3, GVB, and GFkV. On each sampling date, three petioles were collected from each vine and pooled for diagnostic testing. If a vine had symptomatic leaves at the time of sample collection, symptomatic leaves were preferentially collected over asymptomatic leaves. During each growing season in 2011 through 2014 , experimental vines were surveyed regularly for visible leaf roll disease symptoms, beginning immediately after inoculations. On each survey date vines were marked as either asymptomatic or symptomatic, with surveys beginning in May and continuing through October. Shortly after symptoms first emerged in 2012, a detailed symptom survey of each symptomatic vine was performed to determine possible variation in disease symptom severity among vines and if there was an association between location of inoculation and initial appearance of symptoms within vines. For this survey, the position of each spur and the number of symptomatic and asymptomatic leaves on each spur were recorded. In Year Two, berry quality of all vines was measured three times during the weeks immediately preceding commercial harvest. Degrees Brix , pH, and titratable acidity were measured on 31 August, 21 September, and 3 October 2012, and harvest was 4 October 2012. In Year Three, berry quality of a randomly selected subset of 30 vines was measured on 28 August and 14 September, and harvest was 14 September 2013. The 30 vines were evenly divided between uninfected negative controls, uninfected and infected vines from the caged inoculation treatment, and uninfected and infected vines from the uncaged inoculated treatment. For berry quality analysis, on each sampling date approximately 200 berries were collected from each vine to minimize variance in measurements .

Within each grapevine, berries were collected from the top, middle, and bottom of each harvestable cluster of grapes and pooled for laboratory analysis. All samples were processed by Constellation Laboratories in California, USA. Total soluble solids as °Brix were measured using an Atago refractometer, and pH was measured using an Orion pH meter. Titratable acidity of the juice was measured via direct titration with 0.1 N NaOH, using phenolphthalein as an indicator.To test whether the newly infected field vines could be a source of GLRaV-3 one season after mealybug inoculations, a transmission experiment was performed in the laboratory from cuttings of these newly infected fieldvines. Ps. maritimus were not used because of the above mentioned difficulty in obtaining virus-free first in stars for transmission experiments. Instead we used first instars of Planococcus ficus, which are easily maintained in colonies and therefore can be ready for use in transmission studies at any time. Furthermore, Pl. ficus is a known vector of GLRaV-3 . Field cuttings were collected on 4 October 2012 and the stem bases were placed in flasks of water. First instar Pl. ficus were allowed a 24 h acquisition access period on the field cuttings, then transferred to the underside of a leaf of virus- free V. vinifera cv. Pinot noir recipient test vines; ten insects per recipient test vine were confined using a leaf cage for a 24 h inoculation access period. Following inoculations, plants were treated with a contact insecticide and then kept in a greenhouse for four months until petiole sample collection for diagnostic detection of GLRaV-3. For this experiment, a randomly selected subset of experimental field vines of each treatment was tested as a potential GLRaV-3 source. In total, nine symptomatic vines were tested; five from the caged inoculation treatment and four from the open inoculation treatment, and seven recipient test vines were inoculated in the laboratory from each symptomatic field vine. One of these 63 recipient test vines died before petiole sample collection to test for infection with GLRaV-3. Eleven total asymptomatic field vines were tested as a negative control: three from the caged inoculation treatment, three from the open inoculation treatment, big plastic pots and five uninoculated negative control vines. There were no symptomatic negative control vines in the field experiment. For each asymptomatic field vine, three replicate recipient test plants were inoculated, for a total of 33 recipient test vines from asymptomatic field vines. Additionally twenty uninoculated test vines were included with the recipient test vines in the experiment as negative controls, for a total of 116 experimental and control test plants.For each field and laboratory experiment, proportions of resulting successful inoculations from replicate source plants were compared using a Pearson chi-square test; proportions of successful inoculations did not differ, and therefore infected source plants were pooled for further analyses. A chi-square test revealed that caged and uncaged treatments did not differ in the field or laboratory studies, and data from caged and uncaged treatments were therefore pooled for all analyses. For each transmission experiment, proportions of recipient test plants that became infected with GLRaV-3 in each treatment were compared using chi-square tests. We calculated the estimated probability of transmission bya single insect following Swallow . The Swallow estimator can be used to estimate the probability that one insect will transmit a pathogen based on the number of insects used per recipient test plant, the number of recipient plants tested, and the proportion of recipient test plants that become infected. For the detailed symptom survey in Year Two on symptomatic vines only, we tested for a difference in the proportion of leaves that were symptomatic among spurs, using a generalized linear model with a Gaussian distribution; proportion data were arcsine-transformed prior to analysis to better meet the assumptions of the model. All above analyses were conducted using R Version 3.2.0. To assess the effects of GLRaV-3 infection on berry quality, °Brix, pH, and titratable acidity of symptomatic and asymptomatic vines were compared using a repeated measures ANOVA, using SPSS Version 23.

We found no effect of GVB infection on any of the variables measured in our field experiment; therefore the four vines that became infected with both GVB and GLRaV-3 were included with GLRaV-3-infected vines in our analyses.Our vineyard inoculations provide the first mealybugborne GLRaV-3 transmission study under realistic commercial vineyard conditions, providing corroboration that other laboratory transmission studies of GLRaV-3 are predictive of mealybug-borne transmission in commercial vineyards. In the field study, three months after vector inoculation, GLRaV-3 infections were detected in the petiole of the inoculated leaf of approximately two thirds of all vines that ultimately became infected, indicating that early localized infections in commercial vineyards can be detected using diagnostics well before the appearance of disease symptoms. Grapevine leafroll disease symptoms first appeared early in the year of the growing season following mealybug-mediated inoculations, and were present in all infected vines within a two week time frame. Appearance of disease symptoms was more consistent and narrow in timing than was diagnostic detection, which increased for two years following inoculations. Symptoms first appeared without localization to the point of inoculation, indicating that systemic infection had established before the first expression of symptoms. Furthermore, newly infected field vines were effective sources for mealybug-borne transmission one year after inoculation, providing additional evidence of rapid establishment of systemic infection. Berry quality was also affected one year after inoculations, indicating that infection had an effect on vine physiology as early as one growing season following inoculations. Only vines that were infected with GLRaV-3 also tested positive for GVB, indicating that GVB may have some dependence on GLRaV-3 during transmission or establishment in a new host. There were much fewer infections with GVB than with GLRaV-3. There was no evidence that GVB affected disease symptoms or progression compared with vines that were infected only with GLRaV-3. Results of laboratory-based transmission studies can differ from realistic field conditions , and there is considerable variation in estimates of transmission efficiency of GLRaV-3 among laboratory studies . The laboratory and field studies were consistent with each other in that there was no effect of caging the insect vectors on the recipient test vines on virus transmission. There was higher transmission efficiency based on our laboratory experiment compared with our field study. This may have been due in part to the controlled conditions indoors compared with outdoors, and the improved ability of first instar mealybugs to settle and feed on recipient test vines in the laboratory.