Kelly's citrus thrips, Pezothrips kellyanus (Bagnall) (Thysanoptera: Thripidae), is a new pest of citrus (Stevens et al. 1998; Webster et al. 2006; Vassiliou 2007; Navarro et al. 2008). It became a pest in New Zealand (Blank & Gill 1997) and southern Australia (Mound & Jackman 1998) during the 1990s. In the Mediterranean Region, the first damage caused by P. kellyanus was recorded a few years later, and now this thrips is considered a pest in Greece (Varikou et al. 2010), Cyprus (Vassiliou 2007), Sicily (Italy) (Marullo 1998; Conti et al. 2003), and Spain (Navarro-Campos et al. 2012a). Pezothrips kellyanus nymphs feed on the surface of young citrus fruits for 5 to 6 wk starting at petal fall (Navarro-Campos et al. 2013). This feeding habit causes patches or rings of scarred tissue around the fruit apex that enlarge as the fruit grows. This damage is particularly severe on navel orange, lemon (Conti et al. 2003), and grapefruit (Mound & Jackman 1998; Baker et al. 2004; Vassiliou 2007, 2010). Although feeding damage does not affect the internal quality of the fruit, this damage leads to economic losses due to reduced market value of the affected fruits. The percentage of citrus fruits with a complete ring scar may reach 70% per orchard (Vassiliou 2010).

Despite the worldwide distribution and economic importance of P. kellyanus, its biological control is still under development (Baker et al. 2011; Navarro-Campos et al. 2012a). Therefore, chemical control is currently the only practical alternative for growers. However, its implementation, results, and side effects are poorly known. First of all, the number of treatments necessary to reduce thrips populations is unclear. From 1 to 3 insecticide applications are directed to these pests (Conti et al. 2004; Vassiliou 2007). Second, the efficacy of insecticides on P. kellyanus nymphs, the stage that produces the damage, has never been determined. The efficacy of the treatments has been determined based on fruit damage, but whether these applications reduce either the 1st or the 2nd generations of thrips, or both, remains unknown. Third, the side effects of these treatments on the natural enemies of other important citrus pests have not been studied. Importantly, insecticides are sprayed in spring, when the populations of key natural enemies are increasing after winter (Martínez-Ferrer 2007; Tena et al. 2008; Urbaneja et al. 2008, 2009; Sorribas & Garcia-Marí 2010). These natural enemies are responsible for the excellent biological control of many occasional and secondary citrus pests on orange cultivars in Spain (Jacas & Urbaneja 2009). Finally, P. kellyanus may develop resistance to insecticides if its chemical control relies on only a single class of insecticides (Baker et al. 2004).

Therefore, it is important to determine the efficacy against P. kellyanus of insecticides with different modes of action. Chlorpyrifos, an organophosphate insecticide, is one of the most-widely used insecticides for pest control in citrus against hemipterans (scales and aphids) and thrips (Morse & Grafton-Cardwell 2012a; Navarro-Campos et al. 2012b; Planes et al. 2013). It is used against the latter because of its fast-acting effect. However, its persistence against more than 1 generation of P. kellyanus is unknown. Spinosad, a mixture of tetracyclic-macrolide compounds, has been identified as potential candidate for integrated pest management (IPM) programs in citrus because of its fast action (insects dying of exhaustion within 1–2 d) and its low persistence (Thompson et al. 2000; Cisneros et al. 2002). Its residues on the leaf surface are degraded by sunlight within a few days (Salgado 1998). Because of these characteristics, spinosad is recommended in citrus against Scirtothrips citri (Moulton) (Thysanoptera: Thripidae) in California (Immaraju et al. 1989; Khan & Morse 2006; Morse & Grafton-Cardwell 2012b). Spirotetramat is a new systemic and persistent foliar insecticide. It is a tetramic acid derivative with a novel mode of action that interferes with lipid biosynthesis, leading to the death of immature stages of the target insect 2 to 10 d after application (IRAC 2014). Spirotetramat is active against a wide spectrum of sucking insects, including aphids, scales (soft and armored), mealybugs, whiteflies, psyllids, and selected thrips species (Grafton-Cardwell et al. 2007). Therefore, it could be used against Aonidiella aurantii (Maskell) (Hemiptera: Diaspididae) and P. kellyanus with a single application at the end of spring. Moreover, its long persistence could make it active against a possible 2nd generation of P. kellyanus.

In this study, we determined: i) the efficacy of 3 insecticides with different modes of action (chlorpyrifos, spinosad, and spirotetramat) against nymphs and adults of P. kellyanus in the field; ii) their persistence against subsequent generations of this thrips; and iii) their effectiveness in decreasing the percentage of damaged fruit. These results would allow us to make an educated recommendation about the number of treatments necessary when these insecticides are used. Finally, iv) we also determined the side effects of these treatments on phytoseiid mite predators, one of the key group of natural enemies in citrus.

Table 1.

Insecticides used in the assays.

Materials and Methods

FIELD ASSAYS

Alzira Orchard

This assay was conducted in a 16-yr-old navel orange ‘Lane-late’ orchard (Citrus sinensis Blanco var. Navel Lane-Late grafted on Citrange ‘Carrizo’ [Citrus sinensis L. Osbeck × Poncirus trifoliata Blanco]; Sapindales: Rutaceae) located near the town Alzira (39°08′59″N, 0°25′59″W) (Valencia, Spain) in 2010. The orchard had 1.4 ha and the planting pattern was 6 × 5 m. It was drip irrigated and the naturally occurring cover crop was mowed annually at the beginning of spring. The population density of nymphs was above the economic threshold at petal fall (25 May). On this day, 35 trees were sampled and selected according to their similar infestation level by P. kellyanus nymphs (15–25% occupied fruits). On the next day, insecticides were applied with a hand gun, using outside coverage with a volume of about 4.5 L per tree (approx. 1,500 L/ha). Ten, 8, and 8 trees (replicates) were sprayed with chlorpyrifos, spinosad, and spirotetramat, respectively, and the remaining 9 trees were not treated and served as controls. To avoid possible interferences, the 8 trees surrounding every treated tree received the same treatment.

To monitor thrips populations and determine the efficacy of the insecticides, we sampled 32 fruits (8 per orientation) per tree 1 d prior to the spray, 2 d later, and then weekly until the end of the study. On each fruit, we determined the presence of P. kellyanus nymphs. Insecticide efficacy was calculated using Abbott's formula (Abbott 1925). The percentage of damaged fruit was determined on 22 Nov in the same trees. We sampled 40 fruits per tree for damage assessment and distinguished between slightly and severely damaged fruits. We considered severely damaged fruits to be those that had complete ring-like scars and slightly damaged ones to be those with incomplete ring-like scars.

Tavernes Orchard

This assay was conducted in a 10-yr-old navel orange ‘Lane-late’ orchard (Citrus sinensis Blanco var. Navel Lane-Late grafted on Cleopatra mandarin [Citrus reshni Hort. ex. Tan.]; Sapindales: Rutaceae) located near the town Tavernes de la Valldigna (39°4′20″N, 0°15′57″W) (Valencia, Spain) in 2010. The orchard had 3.5 ha and the planting pattern was 6 × 4 m. It was drip irrigated and the naturally occurring cover crop was mowed annually at the beginning of spring. The experimental design was a randomized block with 4 replicates of 4 treatments. Each replicate contained 3 rows of 16 to 30 trees. The population density of nymphs exceeded the economic threshold 2 wk after petal fall (31 May). On this day, 10 trees from each central row were labeled and sampled. One day later, 1,500 L/ha were applied with an air blast sprayer at 30 atm of pressure (Fede mod. Select dynamic; Fede S. L.; Cheste, Spain) to achieve outside tree coverage as is normal for citrus aphids-thrips treatments (Chueca et al. 2009). To follow thrips populations and determine the efficacy of the insecticides, we sampled 32 fruits per tree (8 per orientation) the day prior to the spray, 2 d later, and then weekly until the end of the study. We determined the presence of P. kellyanus nymphs on each fruit. On 23 Nov, the percentage of damaged fruit was determined as above.

To determine the population trends of P. kellyanus adults and natural enemies of citrus pests under the different insecticide treatments, a portable, engine-powered suction device was used to collect all arthropods (Tena et al. 2008). The device was constructed by modifying a commercial vacuum-blower (Husqvarna Zenoah Co., model HBZ2601, Japan) adapted to collect insects from the foliage. We modified it by adding a cylindrical plastic pipe 50 cm long with a 30 cm diameter opening. The sampling was standardized by placing the opening of the cylindrical pipe 4 times, for 5 to 8 s each time, on the foliage of 10 citrus trees per date and tree (40 times in total). We sampled 10 trees from the central row in each replicate (4 replicates per treatment). The material collected was bagged and transported to the laboratory, where it was held at -20 °C to kill all insects. Adult thrips and natural enemies were counted and identified up to genus or species level under a binocular microscope. Insecticide efficacy on P. kellyanus adults was calculated using Abbott's formula (Abbott 1925).

We also determined the side effects of the selected insecticides (Table 1) on phytoseiid abundance. We counted the number of live phytoseiids on the underside of 5 interior and mature leaves per tree. Leaves were randomly selected in the canopies of the same trees sampled for P. kellyanus. The mean number of phytoseiids per leaf was determined for each block and treatment on each date sampled. Cumulative phytoseiid-days per leaf were calculated as an index of phytoseiid population for each replicate as:

Where Σ is summation over all sampling dates from the 1st evaluated day, on 31 May, to the last one, on 13 Jul; I_t is the interval between two successive sampling dates; and x_i and x_j are phytoseiid densities on those dates (Hardman et al. 2006; Kahn & Morse 2006).

Results

EFFICACY AGAINST P. KELLYANUS NYMPHS

The percentage of fruits occupied by P. kellyanus nymphs exceeded the economic thresholds (7%) at petal fall at the orchard in Alzira, and 1 wk after the petal fall at Tavernes. This percentage was similar in all the treatments in both orchards (Table 2).

Table 2.

Occupancy and insecticide efficacy (when occupancy was significantly different from the control) of fruits by Pezothrips kellyanus in 2 orchards located in Alzira and Tavernes.

In Alzira, the percentage of occupied fruits was significantly higher in the control trees than in the treated trees 2 d after the treatments (Table 2). The efficacy of spinosad and chlorpyrifos was significantly higher than that of spirotetramat (Table 2). Seven and 14 d after the treatments, the percentage of occupied fruits was low, and there were no significant differences among treatments; therefore, efficacy could not be calculated.

In Tavernes, the percentage of occupied fruits was significantly higher in control and spirotetramat plots than in spinosad and chlorpyrifos plots 2 and 7 d after the treatment (Table 2). The efficacy of spinosad and chlorpyrifos was high, and there were no significant differences between them on both sampling days. Fourteen days after the treatment, the percentage of occupied fruits decreased, and there were no significant differences among treatments. Twenty-one days after the treatments, the percentage of occupied fruits increased again and remained close to the economic thresholds for the following weeks in all the treatments. On 8 Jul, the orchard was treated with chlorpyrifos, and P. kellyanus populations decreased.

DAMAGE

In Alzira, the percentage of severely damaged fruits was significantly lower in the treated trees than in control trees (Fig. 2A), and it was significantly lower in trees treated with chlorpyrifos and spinosad than with spirotetramat (F_{3, 34} = 13.85; P < 0.001). The efficacy of chlorpyrifos and spinosad was significantly higher than that of spirotetramat (F_{2, 25} = 5.53; P = 0.01). The percentage of slightly damaged fruits was significantly lower for the trees treated with chlorpyrifos and spinosad than control and spirotetramat trees (F_{3, 34} = 5.72; P = 0.003). There were no significant differences between the efficacy of chlorpyrifos and spinosad (F_{1, 16} = 1.81; P = 0.22).

Fig. 1.

Number of Pezothrips kellyanus adults collected with a vacuum device in a navel orchard located in Tavernes (mean ± SE). Trees were treated with chlorpyrifos, spinosad, or spirotetramat.

Fig. 2.

Percentage (mean ± SE) of fruits slightly and severely damaged by Pezothrips kellyanus nymphs in every insecticide plot in orchards of Alzira (A) and Tavernes (B). Trees were treated with chlorpyrifos, spinosad, or spirotetramat. Different letters indicate significant differences (P < 0.05) between treatments (1-way ANOVA followed by Tukey post hoc tests).

In Tavernes, however, the percentages of slightly and severely damaged fruits were high, and there were no significant differences among the 3 treatments and the control (slightly damaged: F_{3, 15} = 0.33; P = 0.09; severely damaged: F_{3, 15} = 0.53; P = 0.67) (Fig. 2B).

SIDE EFFECTS

The numbers of phytoseiids per leaf were similar in all plots the day before treatments in Tavernes (Table 3). Their densities did not differ significantly the following days. However, the accumulated phytoseiidday values, used as an overall summary statistic, were significantly lower in the plots treated with spinosad and spirotetramat than in those untreated or treated with chlorpyrifos.

We captured and identified 1,740 natural enemies with the vacuum device (Table 4). Hymenopteran parasitoids were the most abundant, in total 927 were collected, followed by neuropteran predators (286) and arachnid predators (241). In general, the total number of natural enemies captured was higher in untreated plots (control) than in the treated plots in the following days. There were no significant differences among treatments on day -1 (F_{3, 15} = 0.85; P = 0.49), day 14 (F_{3, 15} = 0.50; P = 0.69), and day 21 (F_{3, 15} = 0.85; P = 0.49) in the total number of natural enemies captured. However, the total numbers of natural enemies captured in the plots treated with spinosad and chlorpyrifos were significantly smaller than in those untreated on day 2 (F_{3, 15} = 6.27; P = 0.0084). Similarly, the total numbers of natural enemies captured in the plots treated with the 3 insecticides were significantly smaller than in those untreated on day 7 (F_{3, 15} = 10.71; P = 0.001). We could not determine the side effects of the insecticides on the main natural enemies of citrus, namely the hymenopteran parasitoids Aphytis melinus DeBach (Aphelinidae), Cales noacki Howard (Aphelinidae), Citrostichus phyllocnistoides (Narayanan) (Eulophidae), and Metaphycus spp. (Encyrtidae) and the predators of the family Coccinellidae, because of the small numbers of specimens of these species collected (Table 4).

Table 3.

Side effects of insecticides used against Pezothrips kellyanus on phytoseiid populations.

Discussion

Our results, based on weekly monitoring of P. kellyanus immature populations on fruit, indicated that a single insecticide application of either chlorpyrifos or spinosad can suppress P. kellyanus nymphs when only 1 generation of P. kellyanus attacks the fruit. This was the case of the assay in Alzira. Both chlorpyrifos and spinosad displayed a knockdown effect against P. kellyanus nymphs and, 2 d after the treatment, reduced the percentage of occupied fruits below economic thresholds. Both pesticides reached efficacies higher than 90% mortality of the thrips populations. Afterwards, thrips populations remained low in both treated and untreated trees. Thus, only 1 generation attacked the fruit in this assay. As a consequence of this attack, the percentage of damaged fruit at harvest was less than 25% in the trees treated with chlorpyrifos or spinosad, whereas it reached almost 50% in untreated trees. Therefore, these insecticides were able to reduce the abundance of P. kellyanus and its damage with a single application when only 1 generation of this thrips attacked the fruit. Both pesticides had been tested previously against this pest with similar results (Benfatto et al. 2000; Purvis et al. 2002). Baker et al. (2004) considered spinosad to be a potential candidate for IPM of P. kellyanus in Australia, and Vassiliou (2007) identified chlorpyrifos as the most effective insecticide among 15 tested in his study. The results of these 2 studies were based on the observation of damaged fruit at harvest, though they did not monitor P. kellyanus populations before or after the treatments.

By contrast, our assay in Tavernes showed that a single application of chlorpyrifos or spinosad could not suppress a 2nd generation of P. kellyanus. As in Alzira, both insecticides reduced the percentage of occupied fruits to below economic thresholds 2 d after treatment. These percentages remained low for 21 d after the treatments, when a new generation of nymphs attacked the fruit in all blocks. Thus, the persistence of chlorpyrifos and spinosad applied against the 1st generation was not enough to control the 2nd one. Consequently, a 2nd treatment would have been necessary to suppress it. Although the percentage of occupied fruits was only one-third as large as the 1st generation, and fruits were larger, this 2nd application seems necessary because the percentage of damaged fruit was very high (above 50%) in this assay. Importantly, all insecticides applied against the 1st generation were able to reduce the percentage of damaged fruit when compared with the control. Vassiliou (2007) sprayed twice against P. kellyanus, but the percentage of damaged fruit was approximately 70%. Therefore, a 2nd application does not guarantee a reduction of damaged fruits. In his assay, Vassiliou (2007) did not monitor P. kellyanus populations, and it is not known whether the application timing was correct. Consequently, this 2nd application can be recommended only when thrips populations are monitored. Finally, if a 2nd treatment is necessary, the insecticides used should be different from those used against the 1st generation to help avoid development of resistance and to assure continued effectiveness of the available pesticides.

Some populations of citrus thrips have developed resistance to pesticides (Morse & Brawner 1986; Immaraju et al. 1989). In Californian citrus, S. citri developed resistance to a long list of insecticides (Morse & Brawner 1986; Immaraju et al. 1989; Khan & Morse 1998). Baker et al. (2004) found that some P. kellyanus populations in southern Australia had substantial levels of chlorpyrifos resistance. In Spain, chlorpyrifos has been used widely to control A. aurantii and other armored scales during the last 2 decades. The high efficacy obtained with chlorpyrifos in our assays suggests that Spanish populations of P. kellyanus have not yet developed resistance to this insecticide. To avoid development of resistance, citrus growers should avoid applying chlorpyrifos against both generations of P. kellyanus or against A. aurantii and P. kellyanus within the same year. The most obvious way of delaying the development of resistance to insecticides is to use them only when required, relying whenever possible on other methods of control that are included in IPM programs (Morse & Brawner 1986). In Spain, growers spray twice during this period to protect the fruit from P. kellyanus scarring. The 1st treatment is generally applied at petal fall and the 2nd one is usually applied 15 d later as a routine. However, insecticides should not be applied at petal fall as a habitual practice because the 1st generation of thrips may appear later, as occurred in the Tavernes assay. Furthermore, the existence of a 2nd generation may vary among years and locations. In fact, we did not observe a 2nd generation of nymphs in this orchard in 2012 (pers. observations). Consequently, population monitoring is critical to determine the optimum spray timing, assure the efficacy of treatments, and delay the appearance of resistant populations. In California, timing is considered vital to achieve adequate control of S. citri with a single application of a relatively short residual pesticide, so that destruction of beneficial organisms is minimized (Morse & Brawner 1986; Morse et al. 1988).

Table 4.

Side effects of insecticides used against Pezothrips kellyanus on natural enemies evaluated by total number of insects captured with a vacuum device per treatment and evaluation day.

Spirotetramat has been registered recently against A. aurantii, Panonychus citri (McGregor) (Prostigmata: Tetranychidae), and thrips (Grafton-Cardwell et al. 2007; Grafton-Cardwell & Scott 2008; Morse & Grafton-Cardwell 2009; MAGRAMA 2014). Spirotetramat might be an especially useful insecticide as it could control A. aurantii and P. kellyanus with a single application in spring. In our study, spirotetramat showed an efficacy around 60% and reduced both slight and severe scarring around 40% relative to control at Alzira in 2009. However, its efficacy was lower than that of chlorpyrifos and spinosad. Also, it did not display a knockdown effect, likely because its contact efficacy is rather limited (Nauen et al. 2008). Thus, its effect on P. kellyanus seems to be limited when compared with spinosad and chlorpyrifos. Moreover, despite its systemic and translaminar activity (Nauen et al. 2008), it did not prevent the attack of the 2nd generation in Tavernes.

To avoid the disruption of the excellent biological control of some important orange pests (Jacas & Urbaneja 2009), the insecticides selected for use against P. kellyanus should have relatively short residual effects, so their impact on beneficial organisms will be minimized (Morse & Brawner 1986; Morse et al. 1988). This is especially important because P. kellyanus is treated at the end of spring when most natural enemy populations are increasing in Spanish citrus (Martínez-Ferrer 2003; Tena et al. 2008; Urbaneja et al. 2008, 2009; Sorribas & Garcia-Marí 2010).

Spinosad and spirotetramat decreased the number of cumulative phytoseiid-days. Spinosad also reduced the number of cumulative phytoseiids-days in a similar study carried out in California (Kahn & Morse 2006). Although spinosad and spirotetramat showed low toxicity in our study, more detailed study is needed to clarify the side effects of spinosad on phytoseiids because it was highly effective against P. kellyanus and is therefore a candidate to be used within IPM programs. This is especially relevant if 2 treatments are necessary to control P. kellyanus.

In addition to counting phytoseiids, we also collected beneficial insects with a vacuum device. Our data showed that the 3 insecticides had a negative effect on the total number of beneficial insects, and this detrimental effect lasted 1 wk. However, we could not determine the side effects of the insecticides tested on the other 2 groups of natural enemies that are key to Spanish IPM programs in citrus, namely coccinellid predators and hymenopteran parasitoids (Urbaneja et al. 2008). Therefore, we would recommend determining the side effects on representative parasitoids and coccinellids of citrus IPM programs under laboratory conditions to ascertain the actual impact. Some of these studies have already demonstrated that chlorpyrifos is harmful to the parasitoid A. melinus (González-Zamora et al. 2013; Vanaclocha et al. 2013) and the coccinellid Cryptolaemus montrouzieri Mulsant (Coleoptera: Coccinellidae) (Planes et al. 2013).

In conclusion, our study shows that chlorpyrifos and spinosad display a knockdown effect and can control the 1st generation of P. kellyanus in citrus with a single application. However, their persistence is not enough to avoid a 2nd generation when it occurs. Therefore, an additional application might be necessary in those cases where this 2nd generation occurs. However, this 2nd application, as the 1st, is justified only when thrips populations are correctly monitored and exceed the potential damage threshold. Finally, because IPM programs on navel oranges in Spain are based on biological control of most of their pests, the development of alternative control strategies to avoid the disruption of the established biological control is urgently needed.