Using Nectar Cover Cropping in Vineyards for Sustainable Pest Management

Objective 1: Parasitoid survival and fecundity in the laboratory on nectar resources

Laboratory trials were proposed to investigate if buckwheat flowers and cahaba vetch extrafloral nectaries increase longevity and fecundity of three natural enemies, G. ashmeadi (glassy-winged sharpshooter parasitoid [GWSS]), Anagrus epos (grape leafhopper parasitoid) and Anagyrus pseudococci (vine mealybug parasitoid). We have completed these studies for A. pseudococci and G. ashmeadi and methodologies are outlined below. Trials with the leafhopper parasitoid, A. epos, could not be conducted. Efforts to establish A. epos and A. erythroneurae (variegated leafhopper parasitoid) colonies in 2008 and 2010 were unsuccessful. In 2009, we were unable to attempt to establish Anagrus sp. colonies since labor was concentrated on insect monitoring and maintenance of the 2009 field trial and processing 2008 sticky traps.

Maintenance of insect colonies and buckwheat plants

Colonies of G. ashmeadi and GWSS were maintained at the University of California, at Riverside, CA (UCR). G. ashmeadi colonies were reared on GWSS eggs laid on ‘Eureka’ lemon leaves, a preferred lemon variety for GWSS oviposition and parasitoid foraging (Irvin and Hoddle 2004). Maintenance of Citrus limon cv. ‘Eureka’ trees used for GWSS colonies and GWSS and G. ashmeadi colony maintenance is described in Irvin and Hoddle (2005). Petri dishes containing leaves with GWSS eggs previously exposed to G. ashmeadi were held at 26o ± 2oC and 30-40% RH under a L16:8D photoperiod and checked daily for parasitoid emergence.

Mummies containing unemerged A. pseudococci were provided by University of California, Berkeley, CA (UCB). Laboratory colonies of A. pseudococci at UCB were held at 24oC ± 2oC under a L16:8D and reared on the vine mealybug, Planococcus ficus, feeding on organic butternut squash, Cucurbita moschata L.. Maintenance of A. pseudococci and host colonies is described in Daane et al. (2004). On arrival, mummies were placed in Petri dishes (10 x 1.5 cm) and held at 26oC ± 2oC and 30-40% RH under a L14:10D photoperiod. Petri dishes were checked daily for parasitoid emergence.

P. ficus colonies were held at 26o ± 2oC and 30-40% RH under a L14:10D photoperiod with fluorescent lighting and maintained at UCR. One organic butternut squash heavily infested with P. ficus (provided by Kearney Agricultural Research and Extension Center, Palier, CA) was placed into a wooden cage (32 x 34 x 37 cm) painted white, with a glass top, mesh back for ventilation and hinged front door containing a material sleeve for access. Squash was stabilized by resting on a simple wooden stand constructed of two pieces of wood held parallel by two pieces of wood. One additional P. ficus colony was set up each week from this initial colony by brushing ovisacs laid by P. ficus in the initial colony gently with a paint brush and transferring them to a new cage containing one squash.

Plants of buckwheat (F. esculentum; obtained from Outsidepride, Salem, OR) and vetch (Vicia sativa L. cv. ‘Cahaba White’; obtained from Bailey Seed Company, Salem, OR) were grown from seed in a greenhouse at 26o ±5oC under natural L14: 10D light. Seeds were sown in 1 gal pots, placing 4-5 seeds per pot. Synchronous nectar production was ensured by performing staggered sowings at 7-10 day intervals. Plants were fertilized every three weeks with Miracle-Gro (20 ml/3.5 liters of water, Scotts Miracle-Gro Products Inc., Marysville, OH). Prophylactic applications of pyrethrin + canola oil (Garden Safe Brand Fruit & Vegetable Insect Spray, Schultz Company, Bridgeton, MO) were applied to vetch plants every 7-10 days to control greenhouse insect pests. Plants used for experiments were free of Pyrethroid applications for at least 14 days and were sprayed down with water and dried before use in experiments.

Experimental set up

For each parasitoid species, 10-16 replicates of three treatments (water only, buckwheat and vetch) were set up in the laboratory at 26o ± 2oC and 30-40% RH under a L14:10D photoperiod. Wooden cages (32 x 34 x 37 cm), as previously described, were used for treatment replicates and a white piece of cardboard was placed on the bottom of each cage. Water was provided in each cage via a 7.4 ml glass vial (2 dram Fisherbrand Glass Vial, Fisher Scientific, Pittsburgh, PA) with a 5 cm cotton wick which was placed on the bottom of each cage and topped up daily. Plant treatments consisted of one 1 gal potted buckwheat or vetch plant with the bottom and top wrapped in Parafilm (Parafilm‘M’ Laboratory Film, Pechiney Plastic Packaging, Chicago IL). Plants in cages were watered as needed with an 8 oz wash bottle (Nalagene, Thermo Fisher Scientific, Rochester, NY) inserted through a hole in the Parafilm. Tape was placed over the hole after each watering. Plants were removed and replaced every 4-5 days to ensure constant supply of nectar. One newly emerged (? 12 h old) naive male and female G. ashmeadi or A. pseudococci was released inside the cage. Parasitoid longevity was recorded daily until death for each sex, and males were not replaced once dead. Female longevity is reported here.

Hosts were provided to G. ashmeadi by placing the petioles of ‘Eureka’ lemon leaves (a preferred lemon variety for GWSS oviposition and parasitoid foraging; Irvin and Hoddle 2004) containing GWSS eggs (<24 h old) through holes drilled in the lid of a 130 ml plastic vial (40 dram plastic vial, Thornton Plastics, Salt Lake City, UT) filled with water. Leaves bearing hosts exposed to parasitoids were removed and replaced every three days until death of the female parasitoid. On the 1st, 4th, and 7th day, 80, 80 and 60 hosts were provided to G. ashmeadi, respectively. On subsequent host changing days, 40 GWSS eggs were provided. Host numbers and age were selected based on previous studies for G. ashmeadi (Irvin and Hoddle 2005, Pilkington and Hoddle 2006, Irvin and Hoddle 2007). Exposed leaves bearing host egg masses were placed into Petri dishes (9 x 1 cm, Becton Dickinson Labware, Becton Dickinson and Co., Franklin Lakes, NJ) lined with moist filter paper (9cm Whatman Ltd. International, Maidstone, England) and left at 26oC for three weeks to allow offspring to emerge. Leaves sometimes decayed, which often prevented successful insect emergence, therefore unemerged eggs were dissected and the numbers of unemerged parasitoids were recorded and included in progeny calculations.

Hosts were provided to A. pseudococci by placing one butternut squash infested with at least 50 3rd and 4th instar P. ficus on a wooden stand inside the cage. This was removed and replaced after six days to ensure that A. pseudococci were provided with life stages suitable for parasitism during their entire lifetime. Host numbers and age were selected based on previous studies for A. pseudococci (Daane et al. 2004). Small pieces of tissue paper were placed over large areas of honeydew excreta daily to prevent parasitoids from getting stuck and dying unnaturally. Exposed squash bearing P. ficus and plants removed from cages were placed into label wooden cages (previously described) for three weeks to allow offspring to emerge. The number of male and female A. pseudococci offspring was recorded for each cage replicate.

Statistical Analyses

Only parasitoids that died of natural causes were included in statistical analyses. Those females that did not mate (producing only male progeny) were excluded from the male progeny totals and sex comparison analyses. For each species, the effect of treatment on the total number of offspring and longevity was determined using ANOVA in SAS (1990). For A. pseudococci, total offspring and longevity data were square-root transformed prior to analyses. Tukey’s Studentized range test at the 0.05 level of significance was used to separate significant means. Logistic regression was used to determine the effect of treatment on logit offspring sex ratio (percentage female). Pair-wise contrast tests at the 0.05 level of significance were used to separate means. Means (± SEM) presented here were calculated from untransformed data.

Objective 2: Cover crop phenology

A tractor and cultivator were used to create a furrow surrounding each of 156 plots (twelve rows of thirteen plots) at the University of California, Agricultural Operations. Plots were 1 m squared and separated by 1.8 m. In the middle of each month from August 2007 until August 2008, five replicates of buckwheat and vetch were sown in 10 randomly selected plots following recommended agricultural sowing rates (buckwheat: 50 lb/acre = 5.62 g seed/m2; vetch: 60 lb/acre = 6.71 g seed/m2). Seed was sown in each plot, covered with approximately 2.5 cm of soil using a rake and watered with 2.5 gal of water from a watering can. Plots were irrigated via a 0.89 mm orange O-Jet 6000 Series Micro-Spray sprinkler head (Olson Irrigation Systems, Santee, CA) installed in the middle of each plot for 2 h. An adjustable pressure regulator was installed to deliver 15 PSI under which sprinklers emitted 7.2 GPH. Plots were irrigated every 2-4 days from August 2007 until July 2009 depending on time of year.

Irrigation was turned off during periods of rain. Weeding of plots was conducted as necessary. Vetch plots were susceptible to mites, aphids, thrips, the three-cornered alfalfa hopper (Spissistilus festinus [Say]) and the false chinch bug. Therefore, prophylactic applications of Ortho Systemic Insect Spray (8% acephate and 0.5% fenbutatin-oxide; The Ortho Group, Marysville, OH) were applied once a month in March 2008 and April 2008 following label directions, and then every 7-10 days from May 2008 until September 2008. From October 2008 until July 2009, vetch plots were sprayed with Ortho Systemic Insect Spray once a month. Plots were checked weekly and then every 2 days when plants were nearing nectar production. The number of days until at least one plant produced nectar (i.e., floral nectar from flowers of F. esculentum or extrafloral nectar from the stipples of vetch) was recorded per plot. At six weeks, the height of ten randomly selected plants per plot was measured with a ruler and average six week height calculated for each plot. Plants were monitored until nectar production ceased (i.e., when all flowers died for buckwheat or when vetch plants stopped producing extrafloral nectar). The length of the nectar producing period was recorded per plot.

Statistical analyses

Multiple regression was used to determine the effect of sowing date, plant species and sowing date * plant species interaction on average six week height (data logged transformed), days until nectar production (raw data) and length of nectar production (raw data). Tukey’s Studentized range test at the 0.05 level of significance was used to separate significant means. Means (± SEM) presented here were calculated from untransformed data.

Objective 3: Natural enemy enhancement and pest population suppression

In 2008, thirteen plots (28.7m x 6m [2 rows] and separated by at least 36 m) were selected in four blocks of Cabernet Sauvignon grapes at Bella Vista vineyard, Temecula, CA (GPS coordinates: 33o 33’26.18’’N x 117o 00’52.12’’W; elevation: 1,637 feet). One or two cover crop plots and control plots were randomly allocated per block, to total seven cover crop plots and six control plots for the entire study. In both years, six plots maintained under current vineyard practices, which included machine and hand cultivation between rows to remove unwanted weed vegetation, were used as controls. For the 2008 field trial, vetch was randomly allocated the north or south side of each cover crop plot and seeds were sown on February 19. On May 1, 2008, buckwheat was sown on the opposite side to vetch in each cover crop plot. Vetch did not establish, therefore, this side of the row was cultivated on June 11, 2008 and re-sown with buckwheat. In 2009, treatments were re-randomized using the same thirteen plots outlined above. Only buckwheat was sown in the 2009 field trial since vetch previously performed poorly in phenology trials, and there was no cahaba vetch seed available in 2009 because the only seed supplier in the U.S. lost their entire crop to mildew fungal disease in 2009. Buckwheat seed was initially sown for the 2009 field trial on April 1, 2009.

Seed was sown in cover crop plots at recommended agricultural sowing rates, which translated to 336 g of buckwheat seed or 404 g of vetch seed per 60 m2 plot. Seed was evenly sprinkled over the soil and covered with approximately 2.5 cm of soil using a rake. Sprinkler irrigation was installed on the existing grower’s grape irrigation in the cover crop plots. In 2008, irrigation consisted of five sprinklers (blue Micro Bird Spinner sprinkler heads per plot, 12 GPH, 360o x 12 ft diameter coverage; Temecula Valley Piping and Supply, Temecula, CA) each installed into 7 mm tubing attached to a 18 cm bamboo stick each side of the 60 m2 plot. Plots encompassed two rows so a total of 10 sprinklers per plot were installed. In 2009, irrigation consisted of 9 sprinklers (red 7.9 mm Micro sprinkler jet sprayer, 12-14 GPH, 360o x 12 ft diameter coverage; DIG Irrigation Products, Vista, CA) each installed into 7 mm tubing on a 32.5 cm plastic spike (DIG Irrigation Products) each side of the 60m2 plot (total of 18 sprinklers). Irrigation sprinkler number and type was changed in the second year of the trial to increase spray coverage and theoretically increase likelihood of buckwheat establishment. Buckwheat seed was re-sown in cover crop plots 2-3 times throughout the trial, and each cover crop plot was irrigated for 2 h the day after each sowing to ensure good germination, then approximately every 7-10 days for approximately 6 h. Additionally, irrigation was supplemented with 16 gal of water per plot, applied via a 16 gal NorthStar ATV Tree Sprayer (Northern Tool + Equipment, Burnsville, MN) and 4WD motorbike, approximately three times a week. Information on number of irrigation days and supplemental watering were recorded in order to estimate the amount of additional water required by the cover crop plots. Susceptible cover crop plots were treated with Rabbit Scram (Enviro Protection Industries Co., Kirkwood, NY) following label directions starting on July 25, 2008 and April 29, 2009.

Despite all these efforts, we encountered a number of problems with establishing and maintaining a cover crop during these trials. Four out of seven allocated cover crop plots were established in the 2008 field trial, and no replicated plots of buckwheat were established in the 2009 field trial. Poor establishment of cover crops in 2009 was due to a batch of poor quality seed with a low germination rate (just 10-33% in greenhouse studies) that was brought from suppliers, irrigation issues (including sprinkler head blockages and flooding), birds eating seeds before they germinated, rabbits eating large patches of germinated seedlings, extreme summer temperatures killing seeds and seedlings, and severe damage to cover crop plants from tractor and vineyard workers during routine vineyard maintenance. The 2009 field trial was discontinued in July 2009 after it was apparent that establishment of buckwheat replicates was unattainable. This report outlines additional methods and results from the 2008 field trial.

Experimental design

During the 2008 trial, four replicates of the cover crop treatment were established with buckwheat growing on one side of the cover crop plot. The other three allocated ‘cover crop plots’ had irrigation installed, but since buckwheat did not establish, these plots were reassigned as an ‘irrigated treatment’. Consequently, the three treatments for this study were: (1) buckwheat cover crop with irrigation; (2) irrigation with no buckwheat cover crop; and (3) a cultivated control with no buckwheat cover crop or irrigation. Including an ‘irrigated treatment’ will help determine whether effects of buckwheat cover crop on insect fauna, grape yield, fruit quality and vine vigor was due to the buckwheat cover crop or the irrigation required by the cover crop plants.

Details on insect monitoring

Sticky traps. Two transparent sticky traps (16.7cm x 13.2cm) mounted at a height of 1.45m and orientated parallel to the vines were placed on the north and south side of the middle row of each plot, 3.7 m apart. Traps were made from clear Perspex (Plaskolite Inc, Columbus OH) blotted on both sides with hot Tanglefoot glue (The Tanglefoot Company, Grand Rapids MI) using a 7.62cm wide paintbrush. Traps were collected and replaced weekly from June 10 through August 9, 2008. On collection, individual traps were placed between two labeled acetate sheets (21.5cm x 28 cm, C-line Products, Inc. Mount Prospect IL) indicating date trap was deployed, treatment, replicate, direction (north or south) and side of trap (‘open side’ facing out into the plot row or ‘foliage side’ positioned towards the grape foliage). Traps were stored in a -20oC freezer until counting of insects. Traps were viewed under a dissecting microscope, and each insect was identified to family or genus level. Counts of each insect were kept using a five-slot multiple-tally denominator (The Denominator Company Inc., S Woodbury CT). The number of pests and natural enemies were recorded separately for each side of the sticky trap to provide information on whether insects were flying towards or away from the grape canopy and buckwheat cover crop. Groups of beneficial insects that were counted on sticky traps and that are reported here were parasitic and predatory wasps, predatory thrips, and ‘other beneficials’ which included pirate bugs, ladybugs, lacewings, big eyed bugs, predatory mites, spiders, carabid beetles and scarab beetles. Groups of pest species that were counted on sticky traps were thrips, leafhoppers and ‘other pests’ which included miridae, sharpshooters, false chinch bugs, mites and aphids.

Visual counts. To obtain visual counts of leafhoppers and predators, a total of five leaves were visually examined per plot every two weeks between June 5 and August 2, 2008. Five vines on each of the north and south side were chosen at random in each cover crop and control plot. One first generation leaf (a large, mature leaf located three to four nodes up from the basal node of a cane) was examined with an OptiVisor (Donegan Optical Co., Lenexa KS) per vine and numbers of variegated leafhoppers, grape leafhoppers, lacewing eggs and predators were recorded.

Funnel beat sampling. Funnel beat sampling was carried out every two weeks between June 7 and July 9, 2008. The funnel was constructed using PVC pipes (JM Eagle, Livingston NJ) and consisted of four 60.96cm-long leg pieces and four 85.09cm-long frame pieces, secured with four 90? PVC joints (Dura Plastic Products Inc., Beaumont CA). The overall opening dimension of the frame was 0.86m by 0.86m. The frame was fitted with a static-resistant polyester material (Jo-Ann Fabric and Craft Stores, Hudson OH) which tapered to 0.1m by 0.1m at the bottom, creating an attachment point for collection bottles. The attachment point was constructed from 5 cm of the top of a 16 oz wash bottle (Fisher Scientific, Pittsburgh PA) glued into a hollowed out lid which had been glued to a second hollowed out lid. To conduct a funnel beat sample, the funnel was placed under a randomly selected vine near the middle of each plot, and a 16oz wash bottle was screwed into the attachment point. Using a small mallet, the main branch of the vine was struck forcefully and repeatedly, and foliage was agitated for 30 seconds. The sample was swept down from the funnel into the collection bottle, capped, labeled and placed into a large cooler for transport to the laboratory. Vines used for funnel trap sampling were labeled to ensure each vine was not sampled twice within a 28 day period. It was noted whether vines were wet from irrigation at time of sampling, and later these data were removed from statistical analyses. Samples were stored in 95% ethanol and the number of pests and natural enemies were counted as previously described for the sticky traps. Groups of beneficial insects that were counted in funnel beat samples and that are reported here were parasitic and predatory wasps, predatory thrips and ‘other beneficials’ which included pirate bugs, ladybugs, lacewings, big eyed bugs, predatory mites, spiders, nabilae and staphylinid beetles. Groups of pest species that were counted in funnel beat samples were thrips, leafhoppers and ‘other pests’ which included miridae, sharpshooters, false chinch bugs, mites, aphids, Hemiptera, ants and weevils.

Sweep net sampling. A sweep net was used to sample buckwheat and grape foliage in buckwheat and control plots, respectively, every two weeks between June 19, 2008 and August 14, 2008. This was conducted by vigorously shaking foliage into a sweep net (Bioquip, Rancho Dominguez, CA) for 1 min and placing contents in a labeled 1 gal Ziplock bag. Sweep netting was conducted between 7am and 12 pm. The time plots sampled were randomized, and the time the sweep net sample was conducted was recorded for each sample. Bags containing samples were placed into a large cooler for transport to the laboratory. Samples were stored in -20oC freezer until counting of insect groups. Groups of beneficial insects that were counted in sweep net samples and that are reported here were parasitic and predatory wasps, predatory thrips and ‘other beneficials’ which included pirate bugs, ladybugs, lacewings, big eyed bugs, spiders, nabilae and earwigs. Groups of pest species that were counted in sweep net samples were thrips, sharpshooter, leafhoppers, ants and ‘other pests’ which included miridae, false chinch bugs, mites, aphids, psyllidae, Hemiptera and grasshoppers.

Statistical analyses

Sticky trap data. Buckwheat only grew on one side of the cover crop plot, while sticky traps were deployed on both the north and south side of the row within the cover crop plot. Therefore, the sticky trap data in the cover crop plots was further reclassified as ‘buckwheat present in the plot, but not in the row’ and ‘buckwheat present in both the plot and row’. Consequently, for this data set there were four post-hoc treatments: (1) control; (2) irrigation treatment; (3) irrigation and buckwheat present in the plot, but not in the row; (4) irrigation and buckwheat present in both the plot and row.

The experimental design for the sticky trap data was a Double-Split Plot design, and since there is no obvious model for simultaneously analyzing these data across time, the data was analyzed on a per time point basis. The ten sampling dates associated with sticky trap data were averaged into five distinct bi-weekly time periods. Insect counts were log transformed using ln (x+1) prior to performing statistical analyses and presented as means estimated by the model. For each time period, the effect of treatment, row (north versus south side of the row within the plot) and trap side (open side versus foliage side of trap) on Total Pests, Total Beneficials, Thrips, Leafhoppers, Other Pests, Parasitic and Predatory Wasps, Predatory Thrips and Other Beneficials was determined using a linear mixed model. Tukey-Kramer at the 0.05 level of significance was used to separate means (Kramer, 1956). Means (± SEM) presented here were calculated from untransformed data.

Funnel beat data. Funnel beat samples were collected over four dates from the middle of each buckwheat plot. Sample positioning was not row specific, therefore, for this data set there were three treatments: (1) control; (2) irrigation treatment; and (3) irrigation and buckwheat treatment. As previously indicated, all data associated with vines that were wet from irrigation during the time of sampling were removed from the analyses to avoid confounding a wet vine effect with the treatment effects. Insect counts were log transformed using ln (x+1) prior to performing statistical analyses. The effect of treatment, date and date x treatment interaction on Total Pests, Total Beneficials, Thrips, Leafhoppers, Other Pests, Parasitic and Predatory Wasps, Predatory Thrips and Other Beneficials was determined using a linear mixed model. Where an interaction term was not significant, this term was removed and the model re-run. Where the date x treatment interaction term was significant, data was analyzed by date and treatment. Tukey-Kramer at the 0.05 level of significance was used to separate means. Means (± SEM) presented here were calculated from untransformed data.

Visual counts data. Visual counts were conducted on both the north and south side of the row within the cover crop plot. Therefore, for this data set there were four post-hoc treatments: (1) control; (2) irrigation treatment; (3) irrigation and buckwheat present in the plot, but not in the row; (4) irrigation and buckwheat present in both the plot and row. The five pseudo-replications (leaves) within each row were averaged into a single sample before performing statistical analyses. Additionally, the six sampling dates associated with the visual counts data were averaged into three distinct monthly dates (June, July and August). Total Leafhopper counts (variegated leafhoppers + grape leafhoppers), Predatory Insect counts and Lacewing Egg counts were log transformed using ln (x+1) prior to analyses. A linear mixed model was used to determine the effect of month, row, treatment and month x treatment interaction on Leafhopper, Predatory Insects and Lacewing Egg counts. Where an interaction term was not significant, this term was removed and the model re-run. Where the month x treatment interaction term was significant, data was analyzed by month and treatment. Tukey-Kramer at the 0.05 level of significance was used to separate means. Means (± SEM) presented here were calculated from untransformed data.

Sweep net data. Sweep net samples were conducted on the buckwheat foliage in the cover crop plots and grape foliage in the controls. Therefore, there were only two treatments: (1) buckwheat and (2) control. The time the sweep net sample was conducted was broken into two groups: before 9am and after 9am. The effect of treatment, time, treatment*time interaction and sample date on Total Pests (data log transformed), Total Beneficials (data square-root transformed), Total Leafhoppers (square-root transformed), Ants (log transformed), Other Beneficials (log transformed) and Other Pests (raw data) was determined using a linear mixed model. Where an interaction term was not significant, this term was removed and the model re-run. Where the date x treatment interaction term was significant, data was analyzed by date and treatment. Similarly, this was conducted when a significant time x treatment interaction term existed. For Thrips, Sharpshooters, Parasitic and Predatory Wasps and Predatory Thrips data (which did not fit a normal distribution), Friedman’s Chi-square was used on raw data to determine the effect of treatment and time. Means (± SEM) presented here were calculated from untransformed data.

Objective 4. Grape yield and quality

On September 18, 2008, the number of grape clusters present within a 3 m section of vine in the center of each plot was counted. Ten randomly selected clusters were harvested from each section (five from each of the north and south side of the row), placed into labeled Ziplock bags and transported to the laboratory in a cooler for grape yield and quality measurements. The weight of each cluster was recorded to within 0.01g and the number of berries per cluster counted. Each berry was inspected and categorized as ‘normal’, shriveled (berry shriveled due to dehydration), having broken skin (skin pierced and broken open by large insects such as wasps) or being split/squashed (as a possible result of handling).

Additionally, 25 berries were randomly selected from the ‘normal’ berry category and berry size (diameter) was measured for each berry using digital calipers (150 mm Absolute Mode Digital Caliper, Tresna, Guilin Guanglu Measuring Instrument Company, Guangxi Province, China) to within 0.01 mm. Each of the 25 berries was inspected for damage from thrips (scarring) and presence of black sooty mold. Finally, all 25 berries were placed into a 1 gal Ziplock bag and squashed with the palm of the hand to extract juice. A refractometer (Pocket Refractometer Pal-1, Atago, Itabashi-ku, Tokyo, Japan) was used to measure sugar content of the juice.

Statistical analyses

For overall cluster counts, sample positioning was not row specific. Therefore, for these data there were three treatments: (1) control; (2) irrigation treatment; and (3) irrigation and buckwheat treatment. The remaining grape yield and quality parameters were measured from both the north and south side of the row within the cover crop plot. Therefore, for these data there were four post-hoc treatments: (1) control; (2) irrigation treatment; (3) irrigation and buckwheat present in the plot, but not in the row; (4) irrigation and buckwheat present in both the plot and row. The effect of treatment on the overall number of clusters per 3 m row (data logged transformed) was determined using simple linear regression. The effect of treatment, direction and treatment*direction interaction on weight of clusters (square-root transformed), total number of berries per cluster (raw data), number of ‘normal’ berries (square-root transformed), Brix content (raw data) and berry size (raw data) was determine using linear mixed model. To separate means, pair-wise t-tests were performed and p-values were adjusted using Tukey’s method. A generalized linear mixed model was used to determine effect of treatment, direction and treatment*direction interaction on the number of scarred berries since these data were not normally distributed. To separate means, pair-wise t-tests were performed and p-values were adjusted using Tukey’s method. A Chi-square was used to determine the effect of treatment and direction on the number of broken and shriveled berries. A logistic model was used to conduct pair-wise comparisons for treatment. Means (± SEM) presented here were calculated from untransformed data.

Objective 5. Vine vigor

The influence of cover crops on vine vigor was investigated in October 2008 by measuring the weight of winter prunings from three randomly selected vines in the center of each treatment plot. For each vine, the number of canes growing from each arm was recorded. All canes were removed from the vine by cutting just above the basal node, and any remaining leaves and secondary shoots growing from the primary cane were removed. Canes from each vine were placed into 1 gal Ziplock bags and labeled with treatment and replicate. The contents of each bag were weighed to within 0.01 g. The average weight per cane was calculated for each vine by weighing the contents of each bag and dividing the weight by the number of canes.

Statistical analyses

Average cane weight was calculated for each plot, therefore, sample positioning was not row specific. Consequently, for these data there were three treatments: (1) control; (2) irrigation treatment; and (3) irrigation and buckwheat treatment. The effect of treatment on average cane weight (data logged transformed) was determined using one-way ANOVA. Tukey’s Studentized range test at the 0.05 level of significance was used to separate significant means. Means (± SEM) presented here were calculated from untransformed data.

Soil analyses

It was observed during the trial that soil type may have differed between plots used in the cover crop field trial. Soil type can have a marked effect on moisture retention which could have significantly effected the establishment and performance of cover crop and vine vigor, consequently effecting pest and natural enemy abundance. To aid interpretation of results from this field study, ten core soil samples were removed from each plot (four buckwheat plots, three irrigated plots and seven control plots) using a 0-20 cm soil auger. This was conducted by removing five sample from each side of the plot at 6 m intervals. All ten subsamples for each plot were placed into a labeled 1 gal ziplock bag transported to the laboratory. In the laboratory, soil within each bag was mixed and 500 g per was removed and sent to Analytical Laboratory, University of California, Davis, CA for soil testing. Tests included particle size analysis (sand, silt and clay content), organic matter content and moisture retention at both 0.3 and 15 ATM. At The Analytical Laboratory each of 13 soil samples was placed into individual paper bags and dried at 40°C for 12 h. Soil was crushed in a Bico-Braun soil pulverizer to pass through a 2 mm sieve.

A detailed description of methods used for particle size analysis is described in Sheldrick and Wang (1993). Particle size analysis was determine via the hydrometer method which is based on the dispersion of soil aggregates using a sodium hexametaphosphate solution and subsequent measurement based on changes in suspension density. The method has a detection limit of 1% sand, silt and clay (dry soil basis). A detailed description of methods used for moisture retention is described in Klute (1986). Moisture retention determines the soil moisture content under constant preset pressure potential. Soil was brought to near saturation and then was allowed to equilibrate under both 0.3 and 15 ATM. The method detection limit is 0.5%. Organic matter content was determined using the Walkley-Black method. This method quantifies the amount of oxidizable organic matter in which OM is oxidized with a known amount of chromate in the presence of sulfuric acid. The remaining chromate is determined spectrophotometrically at 600nm wavelength. The calculation of organic carbon is based on organic matter containing 58% carbon. The method has a detection limit of approximately 0.10%. A detailed description of the Walkley-Black method is described in Nelson and Sommers (1982).

Statistical analyses

A one-way ANOVA was used to determine the effect of treatment on the percentage of sand, silt and clay, organic matter content and water retention at 0.3 and 15 ATM. Tukey’s Studentized range test at the 0.05 level of significance was used to separate significant means.

Objective 6. Grape pathogens, pathogen vectors and grape pests

Approximately twenty five buckwheat and vetch plants were needle inoculated in the laboratory with X. fastidiosa and tested with ELISA kits after four weeks to determine whether these plants could act as a host for X. fastidiosa, the causative agent of Pierce’s Disease (PD) in grapes. Since both plants tested positive to X. fastidiosa, further testing was conducted to determine whether GWSS could acquire X. fastidiosa from buckwheat or vetch and successfully transmit the pathogen to grape vines. If GWSS can transfer the pathogen from the cover crop plants to grapes, then cover crop plants may act as a potential reservoir of X. fastidiosa and be detrimental to grape growers. This is an important question that needs addressing. Consequently, the ability of GWSS to transmit X. fastidiosa from the cover crop to grapes, and then from grapes to the cover crop was investigated. For this work, forty GWSS (to allow for mortality) were released into cages containing cover crop plants infected with X. fastidiosa. Insects were left for a 48-hour feeding and acquisition period, then insects were collected and five GWSS were placed into a sleeve cage on each of five grape plants. The insects were left to feed for 48 hours, after which the insects were collected into individually labeled 1.5mL microcentrifuge tubes and frozen at -80?C for processing. Following the 48 hr feeding period, the grape test plants were grown in a greenhouse and tested for X. fastidiosa infection 8, 12 and 16 weeks post-feeding using ELISA and plate culturing techniques. Using the same protocols, GWSS transmission from cover crop to cover crop, and grapevine to grapevine (controls) were tested.

Greenhouse transmission results for vetch were inconclusive. Vetch plants grown in the greenhouse were susceptible to pest problems and were extremely difficult to keep alive long enough to allow adequate testing, and the grape to grape controls resulted in no transmission, which may indicate there was a problem with the GWSS used for this cohort. Given these issues with greenhouse testing, field transmission tests were conducted. Trials that investigated natural inoculation of buckwheat and vetch under field conditions were conducted at Agricultural Operations, UCR where X. fastidiosa is known to occur. Buckwheat and vetch was sown in the field in August 2008 and after three weeks, 10 buckwheat and 10 vetch plants were randomly selected and individually covered with acetate cages. Acetate cages were 12” tall and 4” in diameter, with 2 x 4” ‘windows’ on opposites covered with nylon mesh organdy. The top was also covered with nylon mesh organdy. The seam of the acetate and the fabric were glued using a hot glue gun. Cages were secured in place over plants with a 3 foot long length of 1” diameter PVC pipe positioned in the ground directly east of the plant, and the cage was placed over the plant and fastened to the PVC using a size 32 rubberband (114g or 0.25lb) to prevent the cage from being blown over by afternoon winds.

One hundred and twenty-five adult GWSS were collected from the field and placed in a bug dorm with a potted grapevine (variety Redglobe). The grapevine had been needle-inoculated and infected with Xylella fastidious subspecies fastidiosa (Temecula strain of PD). Insects were left to feed for a 48-hour acquisition access period (AAP), then live GWSS were aspirated into plastic 40-dram vials (5 per vial). One vial was placed into each cage for each plant. Four potted grapevine controls (non-infected) were placed beside the buckwheat and vetch plants and fitted with nylon organdy sleeve cages. One vial of GWSS was released onto each potted grapevine. All GWSS were given a 96-hour inoculation access period (IAP), after which, insects were collected and plants labeled. Grapevine controls were returned to the greenhouse. Buckwheat plants started dying at 2-3 weeks post-IAP, so all plants were collected at three weeks post-IAP and tested for PD. Four plants were too dry for culture testing, so these were tested with ELISA only in case dead cells could be detected. The remaining six buckwheat plants were tested with ELISA and culture. The vetch plants were sampled at four-weeks post-IAP by collecting a small branch from the base of the plant. Lowest leaves from each branch of the grapevine controls were collected and the lowest 2cm of petiole tissue from each leaf was used.

Objective 7: Weed competitiveness

The ability of buckwheat and vetch to outcompete weeds was not investigated in the 2008 field trial because vetch was such a poor competitor that plots needed to be weeded by hand to ensure establishment of vetch for the trial. Vetch is a winter cover crop and did not perform well in hot weather. This objective could also not be examined in the 2009 field trial because replicated buckwheat plots could not be established and vetch was not investigated in 2009 field trial (see Objective 3).

Objective 8: Dispersal of natural enemies

We intended to investigate the dispersal of natural enemies from cover crop plots into the vineyard, spraying the plants that natural enemies were visiting and living on. This dispersal information will help determine how many rows of cover crops growers would require for adequate dispersal of biological control agents from resource-providing plants. It has been shown in the laboratory that a topical application of triple mark containing yellow fluorescent SARDI dye, casein from milk and albumin from chicken egg white marked up to 88% of the minute parasitoid G. ashmeadi (Irvin and Hoddle, unpublished data). On July 22 2008, insects in the four replicates of buckwheat plots were triple marked with a solution consisting of 20% chicken egg white (strained All Whites, Papetti Foods, Elizabeth, NJ) (containing ~5% albumin [Anon 1, 2011]), 78% milk (Ralphs 2% Reduced Fat Milk, Inter-American Products, Inc., Cincinnati, OH) (containing ~80% casein [Anon 2, 2011]) and 2% yellow SARDI fluorescent pigment (liquid dye) (Topline Paint, Pty Ltd, Aldelaide, SA, Australia). This was conducted by spraying buckwheat plants in a 30 m x 4.8 m plot with 5 liters of the triple marking solution via a 2-stroke 10 liter Stihl backpack sprayer (Andreas Stihl AG & Co., Virginia Beach VA). Four control plots were selected that were located in the same block as the buckwheat plots. Control plots were not treated with the triple mark to investigate the natural gradient of unmarked insects captured on sticky traps and to investigate the efficiency of buffer zones used to separate treatments by determining whether protein-marked insects are detected in the controls.

In additional to the sticky traps deployed as part of the insect monitoring study (see Objective 3; these traps were considered row 0 in this dispersal study), an additional four transparent sticky traps were placed on the 1st, 3rd, 6th and 10th row adjacent to the center of plot following the protocol previously outlined. Sticky traps were placed in each cardinal direction totally eight additional traps. Sticky traps were collected and replaced three and six days after marking to determine how long insects remained marked under prevailing field conditions. The number of pest and beneficial insects was counted on each trap in the same manner as was previously described for the sticky traps in Objective 3. Groups of beneficial insects that were counted included parasitic and predatory wasps, predatory thrips, pirate bugs, spiders and ‘other beneficials’ which included ladybugs, lacewings, big eyed bugs, predatory mites, carabid beetles, scarab beetles, anthicid beetles and staphylinid beetles. Groups of pest species that were counted on sticky traps were thrips, leafhoppers and ‘other pests’ which included miridae, psyllidae, sharpshooters, false chinch bugs, mites and aphids. Results for Leafhoppers, Other Pests and Thrips are not reported here because the main goal of this experiment was to determine the movement of beneficial insects from buckwheat plots.

To detect insects which were marked with yellow dye, traps were viewed under UV light for presence of yellow fluorescent dye. This was conducted by placing the parasitoid on a clean white 130 plastic vial lid under a dissecting microscope in a dark room. Two Croplands SARDI UV flashlights (SARDI, Urrbrae, SA, Australia) were held on either side of the dissecting microscope with the UV lights illuminating the parasitoids. Insects with yellow dye on their body were considered marked. Insects with yellow dye directly beside them on the sticky trap were also considered marked. Marked insects were circled with a yellow ultrafine-point Sharpie marker. Once the entire side of a trap was reviewed, marked beneficial insects were removed by peeling back a corner of the top acetate sheet and removing the insect with a toothpick (Greenbrier International Inc., Chesapeake VA). Toothpicks were broken in half and the tip containing the insect was inserted into a 200µL microcentrifuge tube (Eppindorf North America, Hauppauge, NY), sealed, and labeled. Microcentrifuge tubes were stored at -20oC until transport to James Hagler laboratory (USDA-ARS Phoenix Arizona) for ELISA testing to detect milk and egg proteins.

Three hundred and fourteen yellow dye marked leafhopper parasitoids (out of a total of 39,141 parasitic and predatory wasps) were removed from sticky traps across all treatments, replicates and sampling dates for ELISA analyses. All yellow dye marked and unmarked spiders (4 marked / total of 77), pirate bugs (3/70) and predatory thrips (5/343) were removed for ELISA testing. Additionally, a cohort of 8 leafhopper parasitoids per trap (totaling 2,349 parasitoids) that scored negative to yellow dye were removed for ELISA testing. All collected insects were packed and transported by car in a large cooler filled with dry ice from UCR to the USDA-ARS Arid-Land Agricultural Research Center in Maricopa, Arizona for detection of albumin and casein protein. Insects serving as negative controls were required as part of the ELISA methods where eight negative control insects are placed per plate of ELISA samples. Therefore, 272 parasitoids, 8 spiders, 8 pirate bugs and 24 predatory thrips were removed from sticky traps deployed on July 15, 2008 (before spraying of the triple mark) and sent to USDA-ARS to serve as negative controls. Insects scored positive for the presence of albumin or casein if the ELISA optical density reading exceeded the mean negative control reading by three standard deviations (Hagler 1997).

Statistical analyses

The effect of treatment (buckwheat or control), distance from middle of the plot (row 0, 1st row, 3rd row, 6th row and 10th row), row side (north or south side of row), presence of buckwheat (traps placed at row 0 on the side of buckwheat plots containing buckwheat were considered having buckwheat present while all other traps were considered having buckwheat absent), side of trap (open side versus foliage side) and treatment x distance interaction on the number of Total Pests, Total Beneficials, Parasitic and Predatory Wasps, Predatory Thrips, Pirate Bugs, Spiders and Other Beneficials counted on sticky traps was determined using Poisson Regression Model on log transformed data. Where a significant treatment x distance interaction existed, data was analyzed by treatment and distance. Pair-wise comparisons using a Bonferroni adjustment correlating to the number of comparisons were used to separate means for detecting significant differences between distances and treatments.

Logistic regression model was used to determine the effect of treatment, distance, row side, presence of buckwheat, side of the trap and treatment x distance interaction on the percentage of parasitoids marked with yellow dye, albumin, casein and a double mark. Non significant terms were removed and the model re-run. The odds ratio was used to determine the trend of significant row, buckwheat presence and trap side effects. The effect of distance from the middle of the plot on the percentage of parasitoids marked with yellow dye, albumin, casein and a double mark was determined for each treatment (buckwheat and control) using logistic regression. Distance was specified as a linear variable. The estimate was used to determine the trend of significant distance effect where a positive estimate equaled an increasing percentage of marked insects occurring with increasing distance and a negative estimate equaled a decreasing percentage of marked insects occurring with increasing distance. When data did not hold to linear fit assumptions, then data were specified as categorical data and Fishers Exact test was used to determine whether significant effect of distance existed.

The percentage of marked spiders, pirate bugs and predatory thrips was pooled together and called ‘other beneficials’. Logistic regression was used to determine the effect of treatment, distance, row side, presence of buckwheat, side of the trap, treatment x distance interaction and species of beneficial insect (spiders, pirate bug or predatory thrips) on the percentage of other beneficials marked with yellow dye, albumin, casein and a double mark. There was no significant effect of species of beneficial insects on the percentage of beneficial insects marked by yellow dye, albumin, casein and a double mark. Consequently, data are presented pooled over species. The effect of distance from the middle of the plot on the percentage of other beneficials marked with yellow dye, albumin, casein and a double mark was determined for each treatment using logistic regression in the same manor as described for parasitoids.