We evaluated 3 species of predacious mites (Acari: Phytoseiidae), Galendromus occidentalis (Nesbitt), Phytoseiulus persimilis Athias-Henriot and Neoseiulus californiens (McGregor), as biological control agents for citrus red mite, Panonychus citri (McGregor) (Trombidiformes: Tetranychidae), on citrus in southern Alabama. Three separate experiments were carried out during 2008 and 2011 to evaluate various factors (i.e. release rate, release frequency and initial prey density) that may impact the performance of the predacious mites. In the first experiment conducted in 2008 on trees with moderate initial prey densities (i.e. < 4 P. citri motiles per leaf), one single release of P. persimilis or G. occidentalis at a rate of 100 or 200 per tree effectively prevented the prey from exceeding the economic threshold (5 motiles/leaf) for the entire duration (35 d) of the experiment. The result of the second experiment in 2008 on trees with high initial prey densities (i.e. ≥ 5 motiles per leaf) showed that 2 releases of P. persimilis or G. occidentalis at a rate of 100 or 200 per tree per release could not provide adequate suppression of P. citri below the economic threshold. In both experiments, P. citri densities were significantly lower in most predacious mite treatments compared to the control (no release). Also, lower P. citri densities were recorded at the higher release rate (200 per tree) compared to the lower rate, but this was only significant in a few cases. The third experiment conducted in 2011 in large plots on trees with low initial P. citri densities (i.e. < 1 motile per leaf) showed that 2 releases of N. californiens or P. persimilis at a rate of 200 per tree per release effectively maintained P. citri at low densities (< 1.5 motiles per leaf) throughout the duration (56 d) of the experiment. Limited observations in spring 2012 confirmed the establishment of the predacious mites released in the 2011 study. These results showed that all 3 phytoseiid species were effective in reducing P. citri densities on citrus. However, initial prey density may be an important factor influencing their performance.
The citrus red mite, Panonychus citri (McGregor) (Trombidiformes: Tetranychidae) is a key pest of citrus (Sapindales: Rutaceae) in many parts of the world (Gotoh & Kubota 1997; Jamieson et al. 2005; Childers et al. 2007). Both immatures and adults feed on citrus leaves. Severe leaf infestation may result in grey or silvery spots known as stippling injury, reduced photosynthesis, premature shoot dieback and decreased plant vigor (Kranz et al. 1977). High infestations may also result in fruit feeding and injury (Childers et al. 2007).
Panonychus citri is also an important pest of satsuma mandarin (Citrus unshiu Marcovitch) in Alabama (English & Turnipseed 1940, Fadamiro et al. 2007, 2008). Although this citrus variety has been grown for over a century in Alabama and other parts of the Gulf Coast region of the United States (English & Turnipseed 1940), satsuma mandarin production is a growing industry in the region partly because of strong industry, state support and development of new markets (Campbell et al. 2004). In Alabama, P. citri is typically a spring pest (Fadamiro et al. 2008), and population densities of the pest during this period are usually above the economic threshold of 5 motiles per leaf, proposed by Childers (1994).
Historically, P. citri has been controlled in the Gulf Coast region and other parts of the world through applications of conventional acaricides (Childers 1994; Jamieson et al. 2005; Childers et al. 2007). However, use of broad spectrum acaricides may speed up phytophagous mite resistance (Shanks et al. 1992; Omoto et al. 1995; Bergh et al. 1999), disrupt predators (Thistlewood 1991; Welty 1995; Antonelli et al. 1997; Jamieson et al. 2005; Urbaneja et al. 2008), and intensify food safety concerns. These concerns have resulted in renewed interests in the use of predacious mites for biological control of spider mites (McMurtry 1983; McMurtry & Croft 1997; Jamieson et al. 2005; Rhodes et al. 2006).
Predacious mites, in particular those belonging to the families Phytoseiidae and Stigmaeidae, have been widely evaluated in the laboratory and field for biological control of phytophagous mites in many crop systems (Childers & Enns 1975; McMurtry 1983; Childers 1994; Wood et al. 1994; Takano-Lee & Hoddle 2001, 2002; Pratt et al. 2002; Colfer et al. 2004; Opit et al. 2004; Jamieson et al. 2005; Rhodes et al. 2006; Fraulo & Liburd 2007; Arthurs et al. 2009; Cakmak et al. 2009). The key species evaluated in many systems include 3 commercially available phytoseiids, Phytoseiulus persimilis Athias-Henriot, Galendromus occidentalis (Nesbitt), and Neoseiulus californicus (McGregor). Some indigenous species have also been evaluated (Muma & Denmark 1970; Childers & Enns 1975; Childers 1994). A recent survey of the predacious mite fauna of satsuma mandarin in Alabama identified 29 species from 9 families, with the dominant species being Typhlodromalus peregrinus Muma and Proprioseiopsis mexicanus (Garman) (Phytoseiidae), and Agistemus floridanus Gonzalez (Stigmaeidae) (Fadamiro et al. 2009). Many of these species (e.g., T. peregrinus and P. mexicanus) were found in association with P. citri (Fadamiro et al. 2008, 2009). However, abundance of predacious mites in local orchards was generally too low to provide significant suppression of phytophagous mites (Fadamiro et al. 2009). Attempts to mass rear the key indigenous predacious mites for augmentative releases against P. citri in Alabama orchards have not been successful (X. Y. & H. Y. F, unpublished data), and thus our interest in evaluating commercially available phytoseiids as potential biological control agents for P. citri.
In a recent laboratory study, we evaluated the predation potential of 3 commercially available phytoseiids, P. persimilis, G. occidentalis and N. californicus, on P. citri (Xiao & Fadamiro 2010). Phytoseiulus persimilis is a specialist predator of Tetranychus spp., while G. occidentalis and N. californicus are selective predators of spider mites (McMurtry & Croft 1997; Blackwood et al. 2001; Fraulo & Liburd 2007). The laboratory study included a series of experiments to assess the numerical and functional responses, and prey-stage preferences of the phytoseiids when offered P. citri as prey. The results showed that all 3 species were effective in reducing P. citri density, preferred P. citri nymphs over eggs, and showed a functional type II (convex) response in which the number of prey consumed increased linearly with prey availability up to a maximum (Xiao & Fadamiro 2010). These initial findings coupled with their favorable life history traits (McMurtry & Croft 1997; Pratt & Croft 2000), aided our selection of these commercially available phytoseiids for further evaluation in the field. In the present study, field experiments were conducted in 2008 and 2011 to evaluate the effectiveness of releases of the phytoseiid species at different rates and frequencies for suppression of P. citri in satsuma orchards in southern Alabama.
MATERIAL AND METHODS
Study Sites
Three separate field experiments were conducted at 2 locations in Baldwin county, southern Alabama. The first 2 experiments were conducted at Coker orchard in 2008. This orchard had ∼ 250 citrus trees comprised primarily of satsuma mandarin (‘Owari’ variety). The orchard has had a history of high infestations of P. citri and is typically managed using conventional practices including routine applications of pesticides. However, no pesticides were applied in the orchard during this study. Mean temperatures at this location during the experiments in spring (Mar-May) 2008 were 15.3–24.0 °C (min: 9.5 °C, max: 28.6 °C). Relative humidity was in the range of 65–75%. The third experiment was conducted in 2011 in a citrus orchard block at the Gulf Coast Research and Extension Center (GCREC), Fairhope, Alabama. This orchard block had ∼ 176 citrus trees comprised mainly of satsuma mandarin (‘Owari’ variety), and was not conventionally sprayed during the trials. Mean temperatures at this location during the experiments in spring (Mar-May) 2011 were 10.1–21.3 °C (min: 5 °C, max: 26.3 °C). Relative humidity was in the range of 50–90%.
Predacious Mites
The phytoseiids, P. persimilis, G. occidentalis and N. californicus, were purchased from Biocontrol Network, Inc. (Brentwood, Tennessee) and kept separately in a cooler (4 °C) for 1–2 days, if the weather was unsuitable for immediate release. Only P. persimilis and G. occidentalis were tested in the 2008 experiments. However, G. occidentalis was not commercially available for the 2011 experiment and was, therefore, replaced with N. californicus.
General Procedure for Predacious Mite Release and Sampling
The study evaluated the effectiveness of releases of the phytoseiids at different rates and frequency in suppressing P. citri in 3 experiments. First, satsuma trees in each orchard were sampled for P. citri, using a protocol described by Fadamiro et al. (2008) to determine initial population densities. Three experiments were conducted on trees with different initial P. citri densities relative to the economic threshold of 5 motiles per leaf proposed by Childers et al. (2007). Experiment 1 was conducted on trees with initial P. citri densities slightly lower than the economic threshold (i.e. 3–4 motiles per leaf, abbreviated as moderate initial prey densities). Experiment 2 was conducted on trees with initial P. citri densities higher than the economic threshold (i.e. ≥ 5 motiles per leaf, abbreviated as high initial prey densities). Experiment 3 was conducted on trees with initial P. citri densities much lower than the economic threshold (i.e. < 1 motile per leaf, abbreviated as low initial prey densities). All selected trees were of similar size (canopy of 2 m high, 1.5 m diam) and had not previously been treated with pesticides during the season. In the first 2 experiments, treatments were applied in single-tree plots. Two test trees were separated by at least one “buffer” tree to minimize wind-aided dispersal of mites between test trees. The third experiment was conducted in larger plots each consisting of 12 trees. Each experiment was a randomized complete block design and trees were randomly assigned to the treatments.
Shipments of predacious mites from the supplier arrived in plastic vials (6 cm ht × 5 cm diam), each consisting of 500–1000 individuals on a carrier. The average number of predacious mites per carrier was estimated by rolling the vial to evenly disperse mites within the carrier, placing samples into a Petri dish, and counting the number of predacious mites under a stereo dissecting microscope (20x). Prior to release, the predacious mites were checked to confirm identity and viability by observing subset samples of ∼ 20 individuals in a Petri-dish for about 10 min. Viability was always > 95% for all species. The release rates evaluated (100 or 200 predatory mites per tree) were selected based on the results of a preliminary trial in 2007 that showed minimal efficacy at lower release rates. The tested release rates are also consistent with the supplier recommendation (i.e. ∼ 2000/ acre for field releases), and published rates (e.g., ∼5000/ha in cotton, Colfer et al. 2004).
For each phytoseiid species, individuals, at the appropriate release rate, were evenly distributed into four 3.5 ml plastic containers (each with holes to allow predacious mites to exit and disperse), which were used as release devices. The 4 plastic containers were then hung on branches located on 4 sides (one container per side) of a test tree at ∼ 1.5 m above the ground. Predacious mite releases were performed around 11 am to 2 pm when temperature was around 24–27 °C, at RH 65–70%.
Experiment 1: Single Release of G. occidentalis or P. persimilis in Single-tree Plots with Moderate Initial Prey Densities
The first experiment was conducted in a block of Coker orchard in 2008 to evaluate single releases of G. occidentalis or P. persimilis at 2 release rates (100 or 200 per tree) on satsuma trees infested with P. citri at moderate initial densities (i.e 3–4 motiles per leaf). The following 5 treatments were evaluated on single trees: 1) one single release of G. occidentalis at a rate of 100 per tree (G100-1), 2) one single release of G. occidentalis at a rate of 200 per tree (G200-1), 3) one single release of P. persimilis at a rate of 100 per tree (P100-1), 4) one single release of P. persimilis at a rate of 200 per tree (P200-1), and 5) no-release control. Predacious mites were released at the same time in all treatments (with the exception of the no-release control) in late March. Each treatment was replicated 3 times. To evaluate P. citri density, 24 (6 per tree quadrant) randomly selected leaves were taken from each test tree at 0, 7, 21 and 35 d after predacious mite release. The leaves were collected in properly-labeled paper bags, held in a cooler and transported to the laboratory where they were examined under a dissecting microscope at 20 × magnification. The numbers of P. citri eggs and motiles (nymphs + adults) per leaf were counted and recorded. In all experiments, predacious mites were rarely observed on the leaves and thus were not recorded.
Experiment 2: Two Releases of G. occidentalis or P. persimilis in Single-tree Plots with High Initial Prey Densities
A second experiment was conducted in another block of Coker orchard in 2008 to evaluate 2 releases of G. occidentalis or P. persimilis at 2 release rates (100 or 200 per tree) on satsuma trees infested with P. citri at high initial densities (i.e. > 5 motiles per leaf). The following 5 treatments were evaluated on single trees: 1) 2 releases of G. occidentalis each at a rate of 100 per tree (G100-2), 2) 2 releases of G. occidentalis each at a rate of 200 per tree (G200-2), 3) 2 releases of P. persimilis each at a rate of 100 per tree (P100-2), 4) 2 releases of P. persimilis each at a rate of 200 per tree (P200-2), and 5) no-release control. Predacious mites were released at the same time in all treatments (with the exception of the norelease control), in early March (first release) and late March (second release), respectively. Each treatment was replicated 3 times. To evaluate P. citri density, 24 (6 per tree quadrant) randomly selected leaves were collected from each test tree at 0, 7, 14, 28, 35, 49, and 63 d after the first predacious mite release. The leaves were collected, handled and analyzed as described above for experiment 1.
Experiment 3: Two Releases of P. persimilis or N. californicus in Large Plots on Trees with Low Initial Prey Densities
A third experiment was conducted at the GREC, Fairhope, AL in 2011 in relatively larger orchard plots each consisting of 12 satsuma trees. The spacing between the trees and tree rows were 4.5 m and 6 m, respectively. The experiment was arranged as a 3 × 3 randomized complete block design (3 treatments, 3 replicates). A buffer of 1 tree row was left between the block and 2 tree rows between plots in each block. The test trees were infested with P. citri at low initial densities (i.e. < 1 motile per leaf). The following 3 treatments were evaluated: 1) 2 releases of P. persimilis at a rate of 200 per release per tree (P200-2), 2) 2 releases of N. californicus at a rate of 200 per release per tree (N200-2), and 3) no predacious mite release control. The procedures were similar to those described for the previous experiments but with some modifications. In each plot, 6 inner trees were selected and tagged for predacious mite release. Predacious mites were released at the same time in all treatments (with the exception of the no-release control), in mid February (first release) and mid March (second release), respectively. Panonychus citri density was evaluated in each plot by collecting 24 randomly selected leaves (6 per tree quadrant) from each of the 6 test trees (for a total of 144 leaves per plot) at 0 and 7 d after the first predacious mite release. To ensure timely processing of the leaf samples before they dried up, the sample size was subsequently reduced by half. Thus, 12 randomly selected leaves (3 per tree quadrant) were taken from each of the 6 test trees (for a total of 72 leaves per plot) at 14, 21, 28, 35, 42, 49, and 56 d after the first predacious mite release. The leaves were collected, handled and analyzed as described above for experiment 1. To control for the different sample sizes, data were analyzed and presented as number of P. citri per leaf.
Statistical Analyses
For all experiments, mean numbers of P. citri eggs, and motiles per leaf per sampling date were computed for each treatment and used for statistical analyses. Data were not normally distributed and thus were transformed using √ X + 0.5 and then analyzed by repeated measures multivariate analysis of variance (MANOVA) on the 2 main factors (sampling date and treatment), with time as the repeated measures factor (Ott & Longnecker 2001; Norman & Streiner 2008; Frank et al. 2011). Repeated measures MANOVA was used to account for the possibility of pseudoreplication in the experimental design (Lazic 2010; Frank et al. 2011). The assumption of equal variances and correlations across time in the response variable and appropriateness of using the unadjusted univariate F-test values and the sphericity square root transformation (√) test were based on the sphericity of the model. Where the sphericity test, which is part of the within-subject analysis, was significant, we reported the adjusted F-test values including adjusted degrees of freedom from the Wilks’ λ test. The Wilks’ λ is a test statistic used in multivariate analysis of variance (MANOVA) to test whether there are differences between the means of identified groups of subjects on a combination of dependent variables. When the results of repeated measures MANOVA showed significant effects of sampling date, treatment and a significant sampling date* treatment interaction, the data was further analyzed by using one-way analysis of variance (ANOVA) followed by the Tukey-Kramer HSD comparison test to determine significant differences among the treatments on each sampling date (P < 0.05; JMP® 7.0.1, SAS Institute 2007).
RESULTS
Experiment 1: Single Release of G. occidentalis or P. persimilis in Single-tree Plots with Moderate Initial Prey Densities
Repeated measures MANOVA showed significant effects of sampling date (Wilks’ λ = 0.144, Adj. df = 6, P < 0.0001), treatment (Wilks’ λ = 0.149, Adj. df = 8, P < 0.0001), and sampling date* treatment interaction (Wilks’ λ = 0.084, Adj. df = 24, P < 0.0001), on the numbers of P. citri eggs and motiles. Thus, the data was analyzed by sampling date using one-way ANOVA. No significant differences were recorded among the treatments in the numbers of P. citri eggs or motiles at 0 (prerelease) and 7 days after release (DAR). However, significant differences were recorded among the treatments at 21 and 35 DAR (Table 1). Numbers of P. citri eggs and motiles were significantly higher in the control (no release) than in the predacious mite treatments at 21 DAR (Table 1, Figs. 1A and 1B). Similarly, numbers of P. citri eggs were significantly higher in the control than in the predacious mite treatments at 35 DAR, with the exception of the treatment in which G. occidentalis was released at a rate of 100 per tree (G100-1). The sharp decline in the population density of P. citri in the control at the end of the experiment (35 DAR) was likely due to a general population crash induced by rainfall. In general, lower densities of the prey were recorded at the higher release rate (200 per tree) compared to the lower rate, although this was only significant at 21 DAR for eggs (Fig. 1A). No significant differences were recorded in the ability of both phytoseiid species to suppress the prey.
TABLE 1.
ONE-WAY ANOVA VALUES FOR EXPERIMENT 1: DENSITIES OF P. CITRI IN SINGLE-TREE PLOTS WITH MODERATE INITIAL PREY DENSITIES TESTING SINGLE RELEASE OF EITHER G. OCCIDENTALIS OR P. PERSIMILIS VERSUS THE CONTROL (NO RELEASE).
Experiment 2: Two Releases of G. occidentalis or P. persimilis in Single-tree Plots with High Initial Prey Densities
Repeated measures MANOVA showed significant effects of sampling date (Wilks’ λ = 0.062, Adj. df = 12, P < 0.0001), treatment (Wilks’ λ = 0.295, Adj. df = 8, P < 0.0001), and sampling date * treatment interaction (Wilks’ = 0.122, Adj. df = 48, P < 0.0001) on the numbers of P. citri eggs and motiles. One-way ANOVA showed no significant differences among the treatments in the pre-release densities of P. citri. However, significant differences were recorded among the treatments in the numbers of P. citri eggs at 7, 14, 28, 35, and 63 DAR (Table 2, Fig. 2A). Similarly, the numbers of P. citri motiles were significantly different among the treatments at 7, 35, 49 and 63 DAR (Table 2, Fig. 2B). In general, lower numbers of P. citri eggs and motiles were recorded in the predacious mite treatments compared to the control (no release) on most sampling dates. As observed in the first experiment, the steep decline in the population density of P. citri in the control at the end of the experiment (35 DAR) was likely induced by rainfall. Among the predacious mite treatments, numerically lower numbers of P. citri eggs were recorded at the higher release rate (200 per tree) compared to lower release rate (100 per tree), but a significant effect of release rate was recorded only at 7 DAR (P. persimilis) and 63 DAR (both species) (Fig. 2A). No significant differences were recorded in the performance of the phytoseiid species.
Experiment 3: Two Releases of P. persimilis or N. californicus in Large Plots on Trees with Low Initial Prey Densities
Significant effects of sampling date (Wilks’ λ = 0.935, Adj. df = 16, P < 0.0001), treatment (Wilks’ λ = 0.989, Adj. df = 4, P < 0.0001), and sampling date * treatment interaction (Wilks’ λ = 0.972, Adj. df = 32, P < 0.0001) were recorded. Further analysis of the data using one-way ANOVA showed no significant differences in the number of P. citri eggs among the treatments at 0 (pre-release), 35, and 42 DAR. However, significant treatment effects were recorded at 7, 14, 21, 28, 49 and 56 DAR (Table 3, Fig. 3A). For P. citri motiles, significant treatment effects were recorded at 7, 14, 21, 28, 49, and 56 DAR (Table 3, Fig. 3B). In general, P. citri densities were significantly higher in the control than in the predacious mite treatments on most of the sampling dates. The only significant difference between the species was recorded at 56 DAR when a significantly lower numbers of P. citri eggs were recorded in the P. persimilis treatment compared to N. californicus (Fig. 3A).
Fig. 1.
Mean (± SE) number of P. citri eggs. (A) and motiles; (B) recorded in the test involving single release of G. occidentalis or P. persimilis in single-tree plots with moderate initial prey densities (Experiment 1). G. occidentalis - 100: one release of G. occidentalis at the rate of 100/tree; G. occidentalis - 200: one release of G. occidentalis at the rate of 200/tree; P. persimilis - 100: one release of P. persimilis at the rate of 100/tree; P. persimilis - 200: one release of P. persimilis at the rate of 200/tree; Control: no release. Arrows indicate date (29 Mar 2009) when predacious mites were released. Means having no letter in common are significantly different (P < 0.05).
DISCUSSION
The results of the 3 experiments showed that the phytoseiids, P. persimilis, G. occidentalis, and N. californicus, were effective in reducing P. citri densities on citrus. However, their efficacy appeared to be influenced by release rate and initial prey density. In the first experiment, conducted on trees with moderate initial prey densities (i.e., < 4 P. citri motiles per leaf), one single release of P. persimilis or G. occidentalis at a rate of 100 or 200 per tree effectively prevented P. citri from exceeding the economic threshold (5 motiles/leaf) for the entire duration (35 d) of the experiment. The results were similar for both release rates, although numerically lower prey densities were achieved at the higher release rate. The results of the second experiment, conducted on trees with high initial prey densities (i.e. ≥ 5 motiles per leaf) of P. citri, were not as promising. Although P. citri densities were significantly lower in most predacious mite treatments compared to the control (no release), the data showed that 2 releases of P. persimilis or G. occidentalis at a rate of 100 or 200 per tree per release could not provide adequate suppression of P. citri below the economic threshold. As in the first experiment, lower prey densities were recorded at the higher release rate compared to the lower rate, but this was only significant in a few cases. Together, these results suggest that initial prey density may be an important factor affecting the ability of the phytoseiids to effectively control P. citri. The results of the large plot experiment conducted on trees with low initial P. citri densities (i.e. < 1 motile per leaf) showed that 2 releases of P. persimilis or N. californicus at a rate of 200 per tree per release effectively maintained P. citri at low densities (< 1.5 motiles per leaf) throughout the duration (56 d) of the experiment. It is worth noting that P. citri densities in the control plots also remained below the economic threshold, suggesting that treatment of plots with initial mite densities of < 1 motile per leaf using miticides or biological control agents may not be necessary.
TABLE 2.
ONE-WAY ANOVA VALUES FOR EXPERIMENT 2: DENSITIES OF P. CITRI IN SINGLE-TREE PLOTS WITH HIGH INITIAL PREY DENSITIES TESTING TWO RELEASES OF EITHER G. OCCIDENTALIS OR P. PERSIMILIS VERSUS THE CONTROL (NO RELEASE).
Fig. 2.
Mean (± SE) number of P. citri eggs (A) and motiles (B) recorded in the test involving 2 releases of G. occidentalis or P. persimilis in single-tree plots with high initial prey densities (Experiment 2). G. occidentalis - 100: 2 releases of G. occidentalis at the rate of 100/tree; G. occidentalis - 200: 2 releases of G. occidentalis at the rate of 200/tree; P. persimilis -100: 2 releases of P. persimilis at the rate of 100/tree; P. persimilis - 200: 2 releases of P. persimilis at the rate of 200/tree; Control: no release. Arrows indicate dates (1st release: 6 Mar 2009; 2nd release: 29 Mar 2009) when predacious mites were released. Means having no letter in common are significantly different (P < 0.05).
TABLE 3.
ONE-WAY ANOVA VALUES FOR EXPERIMENT 3: DENSITIES OF P. CITRI IN LARGE PLOTS ON TREES WITH LOW INITIAL PREY DENSITIES TESTING TWO RELEASES OF EITHER N. CALIFORNICUS OR P. PERSIMILIS VERSUS THE CONTROL (NO RELEASE).
Katayama et al. (2006) identified N. californicus as a major predator of P. citri in satsuma citrus groves in central Japan but we are not aware of previously published systematic field studies that evaluated P. persimilis, G. occidentalis, or N. californicus against P. citri in citrus orchards/groves. However, studies have reported that the phytoseiids were effective in suppressing Panonychus ulmi in apples (Monetti & Fernandez 1995), and Tetranychus urticae in citrus (Grafton-Cardwell et al. 1997; Abad-Moyano et al. 2010). Several authors have also reported their efficacy in suppressing Tetranychus spp. in many other crop systems, including strawberries (McMurtry 1982, 1991; Van de Vrie & Price 1994; Rhodes et al. 2006; Fraulo & Liburd 2007; Cakmak et al. 2009), avocado (Hoddle et al. 2000; Takano-Lee and Hoddle 2001), hops (Strong & Croft 1995, 1996), ivy geranium (Opit et al. 2004), greenhouse vegetable crops (McMurtry 1991; Arthurs et al. 2009), and ornamental plants (Hamlen & Lindquist 1981; Pratt & Croft 1998). Many of the above studies demonstrated that release rate/frequency and initial prey density are critical factors that may impact successful biological control of phytophagous mites by the phytoseiids (Hamlen & Lindquist 1981; Hoddle et al. 2000; Pratt & Croft 2000; Opit et al. 2004; Fraulo & Liburd 2007). For instance, Fraulo & Liburd (2007) reported that release rate and frequency had a great impact on the ability of N. californicus to provide effective and season-long suppression of spider mites on strawberry.
These results demonstrated that one or 2 releases of P. persimilis, G. occidentalis, or N. californicus at release rates of 100 or more per tree could provide effective suppression of P. citri on citrus, in particular at low-moderate initial prey densities. Both the eggs and nymphs of P. citri were suppressed by the phytoseiids in this study. However, the data showed relatively higher suppression of nymphs, consistent with a recent laboratory finding that the 3 phytoseiid species prefer nymphs to eggs of P. citri (Xiao & Fadamiro 2010). Previous studies with these phytoseiids suggest that they can survive periods of starvation in the laboratory (Xiao & Fadamiro 2010) and tolerate high field temperatures (McMurtry & Croft 1997). Moreover, they are highly active with high prey searching efficiency (Pratt & Croft 2000, Blackwood et al. 2001), and adapted to disturbed habitats, such as intensively-managed orchards (McMurtry & Croft 1997). These factors suggest that all 3 phytoseiid species may do well in the severe southern Alabama climate. This is supported by our limited observations in 2012 at the GREC (Fairhope) location which confirmed the establishment of the predacious mites (P. persimilis and N. californicus) previously released in 2011. In a small sample collected in spring 2012 at this location, the predacious mites were detected both on the trees on which they had been released in 2011 as well as on control trees, suggesting their spread or dispersal throughout this orchard. Further field evaluations including cost analysis are necessary to determine the economic feasibility of large-scale biological control of P. citri with these predacious mites.
Fig. 3.
Mean (± SE) number of P. citri eggs (A) and motiles (B) recorded in the test involving 2 releases of N. californicus or P. persimilis in large tree plots with low initial prey densities (Experiment 3). N. californicus - 200: 2 releases of N. californicus at the rate of 200/tree; P. persimilis - 200: 2 releases of P. persimilis at the rate of 200/ tree; Control: no release. Arrows indicate dates (1st release: 28 Feb 2011; 2nd release: 14 Mar 2011) when predacious mites were released. * indicates marginal significant difference (P = 0.054). Means having no letter in common are significantly different (P < 0.05).
In this study, we evaluated releases of single predator species rather than multiple predator species. Studies that compared the effectiveness of single versus multiple predator species have produced varying results, ranging from negative (Rosenheim et al. 1995; Schausberger & Walzer 2001) to neutral (Denoth et al. 2002; Chow et al. 2008; Cakmak et al. 2009) effect. Future studies will determine the efficacy of combined releases of 2 or more phytoseiid species, as well as integration of predacious mites with petroleum oils (e.g., FC 435-66 oil) and other effective reduced-risk acaricides (Fadamiro et al. 2005) for managing P. citri in Alabama citrus orchards.
ACKNOWLEDGMENTS
We thank Monte Nesbitt and Alan Burnie for helping with the field trials, and the satsuma grower cooperator (Coker orchard). Funding for this study was provided through grants by the Alabama Agricultural Experiment Station and Auburn University Horticulture Line Item grants program to HYF
