Developing a mealybug pheromone monitoring tool to enhance IPM practices in New Zealand vineyards

Mealybugs are phloem-feeding insects found on many crops worldwide. In New Zealand vineyards, they transmit the economically important Grapevine leafroll-associated virus 3 (GLRaV-3). For some mealybug species, synthetic sex pheromones have been commercialised, and are used as monitoring tools. The mealybugs Pseudococcus longispinus and Pseudococcus calceolariae are major pests in many New Zealand vineyards. We present work on the development of a combined P. longispinus and P. calceolariae pheromone lure. The optimal dose for monitoring P. longispinus was found to be 10 µg of the (S)-(+)-enantiomer, either alone or in the racemic mixture. Addition of the corresponding alcohol did not improve trap catch of P. longispinus. Both the P. longispinus and the P. calceolariae pheromone lures remained active in the field for 90 days. Combining the 2 species’ pheromones had no negative effects on male mealybug trap catch for either species. We conclude that the pheromone ester alone is the best lure for the male P. longispinus. Combining the two mealybug species’ pheromones into a single lure provides the New Zealand viticultural industry with an efficient monitoring tool. Late-vintage deployment of baited lures will provide information on mealybug abundance and local distribution that will inform the scope of future insecticide programmes, to target areas based on need rather than an area-wide application by default.


Introduction
Mealybugs are phloem-feeding pests on many crops worldwide, including fruit production crops (Ahlawat and Pant 2003;Grasswitz and James 2008;Charles et al. 2010;Flores et al. 2015;Firake et al. 2016), ornamental crops (Waterworth et al. 2011), and native plants (Aguirre et al. 2016). The economic losses mealybugs cause have elicited variable responses aimed at minimising their negative impacts (Daane et al. 2012). In vineyards, mealybugs are vectors of several economically important grapevine leafroll-associated Communicated by Donald Weber.
In New Zealand, two introduced mealybug species cause damage in vineyards and orchards: citrophilus mealybug (CMB) Pseudococcus calceolariae (Maskell) and longtailed mealybug (LTMB) Pseudococcus longispinus (Targioni-Tozzetti) (Hemiptera: Pseudococcidae). Both species transmit Grapevine Leafroll Disease, and of particular concern to New Zealand is the economically important Grapevine leafroll-associated virus 3 (GLRaV-3). GLRaV-3 causes wine grapes to have reduced carbohydrate contents, delayed ripening, an increase in titratable acids, and reduced yield by 30-50% (Maree et al. 2013). When found in large numbers in the vines, mealybug waste in the form of honeydew contributes to the growth of sooty mould. If grape bunches are affected with this contaminant at harvest, there is a risk of tainting the wine, meaning mould-affected crops can be rejected.
The viticultural industry in New Zealand has successfully adopted an integrated pest management (IPM) approach to controlling mealybug populations in vineyards. Central to this approach is the use of insecticides that are compatible with biological control (Charles et al. 2010). However, for IPM to be sustainable over time, a good knowledge of mealybug populations in the vines and their distribution throughout the vineyard is required. The current method used to monitor mealybugs in the vineyard relies on personnel undertaking visual leaf (or cluster) inspections. This process is time consuming, costly, and as a consequence, is rarely completed to a point where the data generated can inform future management decisions reliably.
Female sex pheromones for both mealybug species have been identified (Millar et al. 2009;El-Sayed et al. 2010a, b;Unelius et al. 2011;Ramesh et al. 2013), providing opportunities for the use of this technology to greatly benefit monitoring and possibly control in the future (Mansour et al. 2018;Lucchi et al. 2019;Ricciardi et al. 2019;Franco et al. 2021).
In horticulture, mealybugs were often historically managed by organophosphate insecticides and more recently, by neonicotinoids (Furness 1977;Daane et al. 2012). A number of these insecticides have been withdrawn from use in some jurisdictions, e.g. chlorpyrifos and chlorpyrifos-methyl in California and Europe, and imidacloprid in Europe. The withdrawal of insecticides can have significant impacts on the cost of grape production. Goodhue et al. (2022) estimate an increase in the cost of production of table, raisin, and wine grapes in California of $4.2-4.3 million USD per annum. Issues around insecticide resistance, stricter governmental policies on insecticide, including insecticide withdrawal together with negative public perception of such compounds means practitioners have had to evaluate new ways of managing mealybugs (Charles et al. 1993;Daane et al. 2008;Cocco et al. 2021). One of those tactics has been the use of synthetic sex pheromones. Sex pheromones deployed in very small amounts (often micrograms) offer an alternative targeted pest management tool that is more socially acceptable, environmentally safe, residue-free, and has been used successfully against several orders of insects (Suckling and Karg 2000;El-Sayed et al. 2006;Waterworth et al. 2011). Currently, 21 species of mealybug pheromones have been identified and synthesised worldwide (El-Sayed 2022).
In New Zealand, monitoring P. calceolariae in pheromone-baited traps has been successful in field trials conducted over a number of years (El-Sayed et al. 2010a, b, New Zealand patent number WO2011053168;Suckling et al. 2015). However, under New Zealand growing conditions, some questions remained about the efficacy of the California-sourced P. longispinus synthetic sex pheromone when used in field trials. Thus, before either mealybug pheromone is commercially released into the New Zealand market, it was deemed important that the efficacy of both compounds be demonstrated so that end-users could be confident in the results. Regional and temporal variations in the relative abundance of either species necessitate the use of a dual pheromone product so that false negatives are not interpreted from trapping efforts. Thus, for the short term at least, we expect growers will use this technology to guide decisions around whether or not to use mealybug insecticides, and where, based on trap catch data.
This research sought to develop a combined mealybug sex pheromone monitoring tool. We assessed the efficacy of the P. longispinus and P. calceolariae pheromones in New Zealand, and sought to understand properties of the lure for P. longispinus that might influence trap catch efficacy in New Zealand commercial vineyards. This included the effects of enantiomeric purity and alcohol additives (to investigate possible increased attraction), lure longevity, and optimal dose.

Pseudococcus calceolariae pheromone
The P. calceolariae pheromone used was a mixture of chrysanthemyl 2-acetoxy-3-methylbutanoate isomers, synthesised following the methods of Unelius et al. (2011). It contained 33% of the active (R,R,R) isomer (3) (Fig. 1), with inactive stereoisomers making up the remainder of the mix. These other stereoisomers have been shown not to significantly inhibit trap catch (Unelius et al. 2011;Flores et al. 2015).

Traps
Red rubber septa were loaded with the relevant chemicals in a hexane solvent. Individual rubber septum lures were placed on a corflute white sticky base (15 × 15 cm) and placed inside a corflute red delta trap (28 × 20 cm). The traps were tied to the grapevine cordon at c. 70-120 cm above ground.
All lures were tested at the Nelson site, while only 10 of the lures were tested at the Gisborne site.  row-column design for 15 treatments, generated with CycDesign 5.1 (VSN International Ltd 2013). Treatments initially numbered 14 and 15 in each block were then randomly allocated to one of the six extra traps (four replicates of 6 treatments) to ensure fairly even distribution across the three blocks. The Nelson site trial was laid out on 18 April 2018. Sticky bases were initially collected and replaced, with lures transferred to the new base on 24 April, then weekly for 3 weeks. The bases were subsequently replaced every 2-5 days (24 April,and 2,9,16,18,22,25,30 May), giving a trial period of 42 days. Numbers of male P. longispinus mealybugs per trap were recorded at each assessment. For the third and fourth (9 and 16 April) assessment dates, catches were mostly large, so counts were made only on a half (opposite diagonal quarters) or a quarter of the trap and then the total number extrapolated.
At the Gisborne site, four replicates of the 10 treatments were laid out in this trial, with each replicate comprising 5 blocks by 2, with replicates laid out in a 2 × 2 array. The treatments were allocated using a row and column Latinized row-column design, generated with CycDesign 5.1 (VSN International Ltd 2013).
The Gisborne site trial was set up on 23 April 2018, and then sticky bases were replaced approximately weekly, on 28 April, 3, 9, 4, 21 and 27 May, and 5 June, giving a trial period of 43 days. Lures were transferred to the new sticky base. Numbers of male P. longispinus mealybugs caught were recorded at each assessment.

lure longevity, dose response, and combined pheromone field trials
Two trials were carried out, one each at Marlborough and Waipara, New Zealand.
The main trial was conducted at the Marlborough site. Thirteen lures were used (Table 2), and were chosen to allow an assessment of lure longevity (L, six aged lures), dose-response (D, 6 lures of differing doses) and combined pheromones (C, four lures ± P. longispinus lure by ± P. calceolariae lure) ( Table 2). Lures that were aged were left outside in a delta trap at The New Zealand Institute for Plant and Food Research Limited Lincoln site (PFR; Lincoln) for the required number of days (0,30,60,90,120,150) prior to use. The lures inside the delta trap were in full sun and wind exposure, and were at a probable average temperature of 16-25 °C during the day, and 8-12 °C at night. Upon retrieval, the lures were sealed in a foil bag and deposited into a − 81 °C freezer pending deployment for the experiment.
Four replicates of each of 13 lures were used, with the trial laid out as a randomized block design with one full replicate in each of four rows of vines.
For the Waipara trial, only the four combined pheromone lures (C) were used (Table 2), with seven replicates of each. The trial was laid out as a 4 × 7 Latin rectangle, with a full replicate of four traps set up in each of seven rows of vines.
For both the Marlborough and Waipara trials, lures were left in the vineyard (without replacement of the lures) for 5 weeks, with trap bases replaced and lures transferred to the new base at each assessment.
Neighbouring traps were separated by 18 m at the Marlborough PFR site and 20 m at the Waipara site. Sticky bases were replaced every 6-9 days, at which point the numbers of male mealybugs were counted and recorded.

lure longevity and combined pheromone field trial
The trial was conducted in Gisborne in an established commercial vineyard with a known mealybug infestation. The 13 treatments used (Table 2) allowed an assessment of lure longevity of P. longispinus pheromone (10 µg), lure longevity of P. calceolariae pheromone (100 µg), dose response of P. longispinus pheromone (0, 10 and 20 µg), and the combined pheromones (10 µg LTMB + 100 µg CMB, and 20 µg LTMB + 100 µg CMB). Some lures belonged to more than one of these lure sets (Table 2). A potential commercial comparison lure was also tested and is not reported on in this manuscript. However, data for it were included in the statistical analyses.
The aged lures for the lure longevity sets were placed outside in a red delta trap at PFR (Lincoln) in full sun and wind (to replicate vineyard conditions) for the required number of days (0, 30, 60 or 90). The lures inside the delta trap were at a probable average temperature of 16-25 °C during the day, and 8-12 °C at night. Upon retrieval, the lures were sealed in foil bags and deposited into a -81 °C freezer until deployment for the experiment.
The lures were laid out in a 7 × 8 array, with seven traps per vineyard row, and a full set of traps in two adjacent rows. The trial was laid out using a 2-row Latinized row column design, generated with CycDesign 5.1 (VSN International Ltd 2013). Four replicates of treatments 1-12, and eight replicates of treatment 13 were used.
Traps were placed 20 m apart. Sticky bases were replaced 12×, every 2-6 days, and the numbers of male mealybugs on each sticky base were counted and recorded as described above. The trial was set up on 5 May 2020 and concluded on 5 June 2020.

Statistical analyses
Data for each date were examined graphically only.
The analysis approach was broadly similar for all trials except the Nelson trial in 2018. Methods for this trial varied, because data for two traps were missing at one date, and for some traps at some dates, insect counts were extrapolated from half or quarter counts of the trap base.
For all other trials, the total Pre-and Post-treatment catches were calculated for each trap. Initially, these counts were analysed using a negative binomial hierarchical generalized linear model (HGLM, Lee et al. 2006). This was fitted with Lures (and contrasts between lures) included as a fixed effect using a Poisson distribution and a logarithmic link. Trap (one level per trap) was fitted as a random effect with a gamma distribution and a logarithmic link, allowing the aggregation parameter of the negative binomial distribution to be estimated. Replicates, row and columns of traps (as in the layout) were included as spatial random effects also using a gamma distribution and logarithmic link: each was assessed for importance using a X 2 test of the change in deviance on dropping the term as implemented in Genstat's HGRTEST procedure (GenStat Committee 2015).
For Gisborne in 2018, replicate was found to be significant, and for the 2020 Gisborne trial, both row and column were found to be important. These terms were therefore retained as random terms (in addition to Trap) in the final analysis.
For the Marlborough and Waipara trials in 2019, none of the spatial factors was found to be important, so the data were analysed using a Poisson generalized linear model (McCullagh and Nelder 1989), with dispersion estimated. This is a simple special case of the negative binomial HGLM.
For the data from the Nelson 2018 trial, various methods were tried to allow for the missing data and sub-sampling. In the final analysis, counts for traps where only part of the trap was assessed were multiplied by 2 or 4 (where only half or a quarter was assessed). The data for each trap for each date were then analysed with a negative binomial HGLM similarly to the above, except that the aggregation parameter was assessed by fitting to Date × Trap.
For the HGLM analyses, assessment of treatment differences and contrasts between treatments were made similarly to assessment of random effects, using a X 2 test of the change in deviance on dropping the term as implemented in Genstat's HGFTEST procedure (GenStat Committee 2015). For the Poisson GLM analyses, treatment effects were assessed with F-tests done within the analysis of deviance.
For all analyses, results are presented as means and approximate associated 95% confidence limits: these were obtained on the link (logarithmic) scale, and back-transformed for presentation. For the 2018 Nelson trial these were then multiplied by the number of assessments to estimate the total Pre-and Post-treatment counts.
All analyses were carried out with Genstat (Payne et al. 2017).

Alcohol blend: 2018 trials
The P. longispinus male trap catch in Nelson varied with the pheromone:alcohol ratio (p < 0.001 for an overall test), with the catch higher for M100:0 (equivalent to (S)10) and lower for M0:100 than for the intermediate lures (Fig. 2a).
Trap catch also varied between the pheromone:alcohol lures in Gisborne (p < 0.001). The trap catch was again higher for M100:0 (equivalent to (S)10) and lower for M0:100 than for the intermediate lures (Fig. 2b).

Dose response: 2018 and 2019 trials
The dose rate effect for the rac (racemic) and (S) (positive enantiomer) P. longispinus lures in Nelson (2018) was not strong (p = 0.230 and p = 0.113 for rac and (S) overall rate effects, respectively). However, the changes with rate were similar for both (p = 0.747 for the rac vs (S) by rate interaction), and the catches on average were similar for rac and (S) (p = 0.617 for the rac vs (S) main effect). Averaged over rac and (S) there was a significant rate effect (p = 0.029), essentially because of the higher catch for the lowest rate (rac20 and (S)10) than for the higher rates, which had similar catches on average.
Trap catch varied between the racemic lures in Gisborne (2018) (p < 0.001) with catches increasing substantially from the lowest rates to the highest. The catch for M100:0 (equivalent to (S)10) was similar to that for the equivalent rac lure (p > 0.05; rac20).

Lure longevity: 2019 and 2020 trials
Catches of male P. longispinus mealybugs in Marlborough (2019) varied with the number of days the lures had aged in the field (p < 0.001), with an increase from 0 to 30 days, but then a decline as the numbers of days increased to the maximum 150 days tested (Fig. 3). Changes in catch over assessments followed a similar pattern for each number of days aged, with the catch initially quite low, increasing to the third assessment and then reducing (Online Resource 2).
In the Gisborne (2020) field trial, trap catch varied slightly with lure age (p = 0.011), with a higher daily catch for the 30-and 60-day aged lures (mean ~ 8) than for the unaged and 90-day aged lures (mean ~ 6) ( Fig. 3 and Online Resource 1). However, there was less variation over time for the LTMB 10 μg lures than for the lures with CMB 100 μg (p = 0.038 for the lure type x days aged interaction) (Fig. 3).
As the lures were in the field for five weeks, the total number of males caught was confounded by phenology. However, when the catch at each assessment is considered separately for each treatment (Online Resource 1 and 2), the data support the negative correlation with lure age, despite male mealybug phenology.

Combined lures: 2019 and 2020 trials
For the Marlborough (2019) field trial summarised across assessments, the total trap catch of P. longispinus for the combined lure was similar to the additive effect of the lures for each mealybug species on their own, with the presence of either lure increasing the catch (Fig. 4a).
The effect of the P. longispinus lure on trap catch of P. calceolariae was negligible (p = 0.651).

Discussion
The pheromone for P. longispinus was identified by Millar et al. (2009) and the pheromone for P. calceolariae was identified by El-Sayed et al. (2010a, b). The identification of these pheromones enables the use of pheromone monitoring traps, which, if available to the wine sector, could inform viticulturists of mealybug populations and their distribution within the vineyard. This information will help inform as to when and where to apply registered mealybug insecticides. Regular annual use of the pheromone technology will also allow them to track mealybug populations over multiple seasons. This will help to identify areas where mealybugs in the vines might overlap with planted areas known to be affected by leafroll virus. Hogg et al. (2021) showed traps catches using the Planococcus ficus (Signoret) (Hemiptera: Pseudococcidae) pheromone lure to be correlated with mealybug The corresponding alcohol is typically present in volatile collections from mealybugs whose pheromone is an ester (pers. comm. Jocelyn Millar) and are probably formed from hydrolysis of the ester pheromone. However, none of the corresponding alcohols has elicited any behavioural activity in the mealybug species tested to date. Our initial electrophysiology testing supported this, as there was no antennal response to the alcohol (unpublished data). However, owing to the variable nature of the pheromone's performance in previous trapping trials, we still considered it worth investigating the corresponding alcohol since the lack of antennal response may have been due to its low concentration in the headspace, as was found in other studies (e.g. Francke et al. 2002).
Our 2018 results, however, did not show equal or enhanced attraction of male P. longispinus to any of the pheromone alcohol (2) mixed lures, compared with the pheromone ester lure. The pheromone alcohol lures showed very similar catch patterns at the Nelson (high population) and Gisborne (low population) sites.
We found aging either the P. calceolariae pheromone lure or the P. longispinus pheromone lure for up to 90 days in field conditions before beginning trap catches had a negligible effect on male mealybug catch. Arguably, the ideal time for deploying mealybug-pheromone traps in New Zealand is between March and May (c. 60-90 days; Southern Hemisphere autumn). Therefore, based on the results of this study we propose that replacement of lures within a 90-day period is not required to obtain reliable mealybug trap catch data in New Zealand vineyards.
While the results of this study demonstrated that male P. longispinus mealybug catch increased with increasing P. longispinus pheromone dose, it is important to look towards a time when this technology is commercialised. Specifically, with increased dose is an expected increase CMB = citrophilus mealybug (P. calceolariae) lure; LTMB = longtailed mealybug (P. longispinus) lure. Pheromone amounts refer to the active ingredient. Note that the upper confidence limit for a value near 0 is difficult to obtain so is not shown in the production costs, which will translate into relatively minimal benefit to end-users. A 64 × increase in P. longispinus pheromone loading (from 10 to 640 µg) doubled the male catch. We consider 10 µg offers end-users an optimal pheromone loading to catch ratio, set at an economically sustainable price point. This was supported by Waterworth et al. (2011), who found no difference in trap catch between 3.2 and 100 µg loadings of P. longispinus pheromone. However, based on the results of our current research under New Zealand growing conditions, we will not be recommending a loading lower than 10 µg, owing to the probable loss of efficacy over a proposed lure longevity recommendation of 90 days.
A dose response of the P. calceolariae males to the female sex pheromone can be found in Unelius et al. (2011) (field work done in Hawke's Bay, New Zealand), showing a 1.5 × increase in catch between 100 and 1000 µg of P. calceolariae pheromone. Therefore, for monitoring, 100 µg was regarded as the sensible dose, considering both catch and monetary cost. A dose of 100 µg has been used in monitoring trials undertaken in New Zealand for several years.
No substantial interference was found in the male catch of either species when the combined lure was assessed under vineyard conditions. Hence, we recommend combining the 2 sex pheromones (P. calceolariae and P. longispinus) onto one lure. From a practical, logistical, and financial perspective, a single lure containing both mealybug compounds within a single delta trap is likely to be a more attractive option for growers.
Recently published work conducted in Italy (Cocco et al. 2018;Lucchi et al. 2019), California (Walton et al. 2006;Daane et al. 2020) and Israel (Sharon et al. 2016) suggest control of the vine mealybug (P. ficus) via synthetic pheromones may be possible using mating disruption. Ricciardi et al. (2022) used a combined Lobesia botrana Den. & Schiff. (Lepidoptera: Tortricidae) and P. ficus pheromone mating disruption device to successfully reduce trap catches of both species, reduce infested grape flowers, reduce the number of L. botrana nests and the number of P. ficus individuals in Italian vineyards. Ballesteros et al. (2021) successfully reduced the numbers of male P. calceolariae caught in sticky traps by applying the sex pheromone in apple and tangerine orchards in Chile at a rate of 6.32-9.45 g/ ha, indicating potential mating disruption. Ricciardi et al. (2019) conducted wind tunnel assessments of small-scale mating disruption trials of P. calceolariae at PFR Lincoln, New Zealand. These authors found that when a 4 × 4 array of 16 rubber septa lures loaded with 1 and 30 µg of the P. calceolariae pheromone was deployed, female location by males was reduced by 9.2 × compared with the control. The combined mealybug pheromone lure that we present in this study has potential to be used as a mating disruption tool. This would provide significant benefit to New Zealand viticulturists by providing residue-free, low carbon, sustainable pest control for what are significant viticultural pests. Pressure from local and international consumers for residuefree, sustainable products, including wine, reduces the pest management options available to New Zealand growers. Such signals are creating the impetus to identify and test new options for pest management tools to meet sustainability criteria.
There are several challenges to the use of mating disruption in P. calceolariae and P. longispinus, such as multiple mating (P. calceolariae males can fertilize up to 13 females, Ricciardi et al. 2019), and aerial dispersal of mealybug crawlers blowing in from upwind sources, repopulating vineyards with mealybugs. Advancing mating disruption technology in New Zealand or Australia would first require substantial research investment to determine efficacy against species like P. longispinus and P. calceolariae, as well as undergoing registration for the appropriate purpose.

Conclusions
P. calceolariae and P. longispinus lures remain viable in the field for up to 90 days, although to date, field use has typically been for no longer than 42 days.
The 10 µg pheromone was an effective dose to trap male P. longispinus. Importantly, this lower concentration offers end-users a dose loading likely to be more economically sustainable.
There was no substantial interference in the trap catch of male P. calceolariae with the addition of the P. longispinus pheromone, or in terms of trap catch of male P. longispinus with the addition of P. calceolariae pheromone.
The combined mealybug sex pheromone lure is valid for the purpose of monitoring P. longispinus and P. calceolariae in commercial vineyards in New Zealand, and may have potential for use in other jurisdictions as well.
Greg Pringle for his contribution to funding applications and coordination of research, and Tara Taylor for research support. We thank Asha Chhagan, Flore Mas, and Lloyd Stringer for their constructive comments on an earlier draft. This work was conducted under approval from the Environmental Protection Authority New Zealand; Plant and Food Research Experimental Plant Protection Compounds approval number HSC100112.
Funding Open Access funding enabled and organized by CAUL and its Member Institutions. Potential conflict of financial interest: One of the authors (CRU) is currently supplying the citrophilus mealybug pheromone to an English company that in turn sells it to UPL New Zealand for use in the combined mealybug pheromone lure product.

Conflict of interest
The other authors have no competing interests to declare that are relevant to the content of this article.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visithttp:// creat iveco mmons. org/ licen ses/ by/4. 0/