Introduction

Tetranychus kanzawai Kishida is one of the important pests of tea [Camellia sinensis (L.) O. Kuntze] fields; when the damage is noticeable, it results in fallen leaves (Fukunaga and Ishiyama 2001). Acaricides are the main control strategy against T. kanzawai in tea fields. Because this mite inhabits the lower (abaxial) surface of tea leaves (Ehara 1993), application of higher volumes of pesticides is needed to ensure that sufficient acaricide reaches these lower surfaces. However, repeated application of pyrethroids (Osakabe 1986; Ikeda and Yamamoto 1989) and continuous application of the same pesticidal component (Osakabe 1973) has resulted in adaptation of the mite by selection for pesticide resistance populations. As a consequence, it is now considered to be difficult to control T. kanzawai in tea fields. On the other hand, in recent years, there have been increasing instances where the low densities of T. kanzawai have led to the curtailment of acaricide applications. There have been suggestions that the decrease in T. kanzawai frequency could be due, at least in part, to changes in the type of pesticide used, whereas it has also been attributed to biological control of the spider mite by its natural enemies (Ozawa et al. 2006). Masui et al. (2018) found that the species composition of natural enemies was affected by insecticide application in a citrus field. However, Mochizuki (1990) demonstrated that a strain of Neoseiulus womersleyi (Schicha), which is resistant to synthetic pyrethroids, controls T. kanzawai in a tea field. Therefore, if phytoseiid mites could survive in tea fields that had been sprayed with synthetic pyrethroids, they may reduce the abundance of T. kanzawai. Consequently, attention has been drawn to the phytoseiid predatory mites, which are the predominant natural enemies of T. kanzawai.

In the tea fields of Japan, pesticides to control pests other than spider mites—e.g., Empoasca onukii Matsuda (Hemiptera, Cicadellidae), Scirtothrips dorsalis Hood (Thysanoptera, Thripidae), Caloptilia theivora (Walsingham) (Lepidoptera, Gracillariidae), and Homona magnanima Diakonoff (Lepidoptera, Tortricidae)—are generally applied several times a year. There is no doubt that synthetic pyrethroid pesticides adversely affect natural enemies against the spider mite (Takafuji and Inoue 1993), especially the phytoseiid mite fauna in tea fields (Ozawa et al. 2006). Ullah et al. (2016) reported that two neonicotinoids (acetamiprid and imidacloprid) had severe effects on phytoseiid predators and, thus, could account for the observed change in predator species composition in Japanese orchards. Moreover, it is known that the composition of phytoseiid mites in Japanese pear orchards differs depending on the types of pesticides that are used by farmers (Kishimoto 2002). Therefore, it seems likely that the differences in pesticide susceptibility among phytoseiid mites in tea fields also lead to changes in their composition.

Conversely, Kakoki et al. (2018) found that when small amounts of a nonselective pesticide were sprayed onto the plucking surface of the tea plant using the ‘spotting spray method’, the leaf layer functioned as a protective shelter to spiders because not all of the leaf layer was covered with pesticide, and it will maintain some safety space for spiders and other natural enemies of tea pests in part of leaf layer. Furthermore, Hamamura (1986) showed that N. womersleyi was able to escape the effects of low concentrations of 12 synthetic pyrethroids, including fenpropathrin, cypermethrin, and permethrin. From these observations, we hypothesized that a partial spray method that does not completely cover the leaf layer with pesticide helps the predatory phytoseiid mites to escape damage from pesticides, and that the decline in the occurrence of T. kanzawai is affected, at least in part, by increasing numbers and diversity of phytoseiid mite populations.

To test this hypothesis, we examined whether the method of pesticide spraying used against many species of pest in tea fields and the difference in coverage area by the pesticide in different leaf layers affects the diversity and abundance of phytoseiid mite species. As part of this study, we investigated behavioral avoidance of three species of predatory phytoseiid mites, namely Amblyseius eharai Amitai and Swirski, Phytoseiulus persimilis Athias-Henriot, and Euseius sojaensis (Ehara), in response to the synthetic pyrethroids fenpropathrin, permethrin, bifenthrin, and cypermethrin.

Materials and methods

Nonselective pesticides with high activity against natural enemies of pests (Ozawa 2013) were selected as test compounds, and tests were conducted under circumstances (method and amount of spray applied) which maximized the likelihood of detecting any effects (Table 1). In addition, acaricide spraying for control of T. kanzawai was not carried out in order to confirm any suppressive effect of the phytoseiid mites, against T. kanzawai. Phytoseiid mites were identified as described previously (Toyoshima et al. 2015a).

Table 1 Timing, rates, and concentrations of pesticide applications
Full size table

Temporal variation in the abundance of Tetranychus kanzawai and Phytoseiidae in a tea field (field trial)

Study field and pesticide application

We conducted this study in a tea field of the Kagoshima Prefectural Institute for Agricultural Development, Tea Division (Minamikyushu City, Kagoshima Prefecture, Japan) in Japan (31°37N, 130°45E). The trial area in the tea field (cultivar ‘Kuritawase’) was 240 m2 (14-m rows), with three treatments (including the control), and three replicates per treatment, each replicate plot being 7 × 1.7 m2. In addition, a no-pesticide buffer row of tea plants was provided between rows in a treatment. The treatments were as follows: (1) an application rate of 40 L/1000 m2 (subsequently referred to as ‘40-L treatment’ or ‘partial spraying’), using a lower-volume sprayer (Kagoshima-style MCS-KAGO3-2; Matsumoto Kiko, Kagoshima, Japan; spray pressure, 1.5 MPa; spray speed, 0.7 m/s); (2) an application rate of 200 L/1000 m2 (subsequently referred to as ‘200-L treatment’ or ‘conventional spraying’), using a conventional sprayer (MCS8A; Matsumoto Kiko; spray pressure, 1.0 MPa; spray speed, 0.5 m/s); and (3) no spraying (control treatment). Pesticide spraying against E. onukii, S. dorsalis, C. theivora, H. magnanima, and Adoxophyes honmai Yasuda (Lepidoptera, Tortricidae) was carried out, and the effects of these sprayings on the populations of T. kanzawai and Phytoseiidae were monitored.

Monitoring of Tetranychus kanzawai and Phytoseiidae

The sampling of T. kanzawai was conducted from August 2015 (sampling of phytoseiid mites was started in May 2015) to October 2017, with survey intervals of about 7 days from May to October, and about 15–30 days from November to April. Leaves 10 cm from the plucking surface were collected. We randomly sampled 100 leaves from each treatment and put them into a polyethylene self-sealing bag to transport them. We counted the leaves that were infested by adult and immature stages of T. kanzawai and phytoseiid mites (not including the egg stage) under a stereoscopic microscope (SZ-61; Olympus, Tokyo, Japan) in the laboratory. Phytoseiid mites were preserved in 70% (vol/vol) ethanol and were prepared as specimens for identification. We identified phytoseiid mite species under a phase-contrast microscope (DM750; Leica, Wetzlar, Germany) according to Toyoshima et al. (2015a) and Ehara (2007).

Behavioral avoidance of phytoseiid mite species against synthetic pyrethroid insecticide (laboratory trial)

We chose A. eharai, P. persimilis, and E. sojaensis, which had been confirmed as present in the tea field under study, as the test species of phytoseiid mites for this investigation. In 2017, we captured A. eharai in the study tea field from October to December, and E. sojaensis from the tea field of the Kagoshima University campus (31°57′N, 130°54′E) in November. Amblyseius eharai and E. sojaensis individuals (fed with tea pollen) were kept in a 24-well multiwell plate (BM Equipment, Tokyo, Japan) in an incubator (MLR-352H; Panasonic, Tokyo, Japan) at 25 ± 0.5 °C (16 h/8 h light/dark) for tests. In 2017, P. persimilis individuals were collected from the study tea field from May to December and they were raised together with T. kanzawai in an incubator under conditions of 25 °C and 16 h/8 h light/dark on kidney bean (Phaseolus vulgaris L. cv. ‘Taisho-Kintoki’) leaf disks (2 × 2 cm). Tetranychus kanzawai that is food for P. persimilis were also collected from the Kagoshima Prefectural Institute for Agricultural Development (Tea Division) from May to October 2017. We reared T. kanzawai that had been multiplied on kidney bean seedlings at 25 °C and under a 16 h/8 h light/dark photoperiod (Shinkaji 1991).

We selected fenpropathrin (100 ppm), permethrin (100 ppm), bifenthrin (20 ppm), and cypermethrin (60 ppm) as the test pesticides for use in the study tea field trial, with distilled water being used as the control. The test apparatus was prepared following the method of van den Boom et al. (2003), and each test was carried out by the following procedure: (1) a primary leaf of kidney bean was cut into a 2 × 2 cm square, and one side was immersed in one of the pesticide solutions and allowed to dry in air; (2) two leaf pieces (one treated with pesticide and one without) were connected via a strip of parafilm cut to a T shape (3.5 × 3 cm) with the treated leaf surface facing upwards. In a Petri dish (9 cm diameter), absorbent cotton was laid in the base, the leaf pieces and parafilm were placed on top, and the dish base was filled with water (Fig. 1); (3) one adult female of the phytoseiid mite species under investigation was placed on the parafilm, equidistant from the treated and control leaf pieces, using a fine face brush; after 60 min, the location of the phytoseiid mite was noted and the condition of the phytoseiid mite was recorded. The condition was classified as being healthy, distressed, or escaped. In 60 min after of test, the mites, showing the normal behavior as that before the test, were classified as ‘healthy’ and those showing the status of phytoseiid poisoning with trembling legs were classified as ‘distressed’. Each test was terminated when submergence of the phytoseiid mite was confirmed. Submergence was classified as an avoiding behavior (‘escape’), and the test was terminated at that moment; therefore, the survival or death of the phytoseiid mite could not be confirmed following avoidance. Immediately after submergence was confirmed, we moved the individual from the Petri dish to a Munger cell (4 × 5 cm, 2.5 cm diameter) (Munger 1942), provided tea plant pollen as feed, and observed the condition of the phytoseiid mite for further 24 h following submergence. These classifications of condition (i.e., healthy, distressed, and escaped) have previously been used (e.g., Wakou and Odagiri 1985; Hamamura 1986; Mochizuki 1990; Hashimoto et al. 2000). Twenty-five individuals were tested for each pyrethroid–mite species combination.

Fig. 1
figure 1

The setup of the pesticide-avoidance (evasion) laboratory test with of phytoseiid mites

Full size image

Statistical analysis

In the avoidance test, the results were converted to ratios, then a χ2 test was carried out on the results from each pyrethroid/control comparison; if a significant difference was observed (α = 0.05), the Ryan multiple comparison test was performed. If zero was present in any of the cells (e.g., with distilled water, no distress should be observed), the χ2 test would be inappropriate, so the Fisher’s exact probability test was used instead after dividing the result into avoidance and healthiness, followed by the Ryan multiple comparison test was used if a significant difference was detected. The statistical analysis of the field trial data used a repeated-measures analysis of variance, performed with log10(N + 0.5)-transformed data for mite abundance per leaf in each treatment. If a significant difference was observed, multiple comparisons were carried out using Tukey’s honestly significant difference (HSD) test. JMP v.7 software (SAS Institute 2007) was used to carry out all statistical analyses.

Results

Temporal variation in the abundance of Tetranychus kanzawai and Phytoseiidae in a tea field (field trial)

Throughout the study periods, the number of T. kanzawai did not differ significantly among the treatments (F2,79 = 2.33, p = 0.10). In 2015, the number of T. kanzawai in the 40-L treatment was significantly higher than in the 200-L treatment or the control, whereas there were no significant differences in numbers among treatments in 2016 and 2017. The maximum value of the ratio of leaf on which T. kanzawai fed was 17.3% in the area subjected to the 200-L treatment, 20% in the area subjected to the 40-L treatment, and 10% in the untreated (control) area (Fig. 2).

Fig. 2
figure 2

Temporal variation in the percentage of leaves infested by Tetranychus kanzawai and abundance of T. kanzawai and Phytoseiidae. Treatments within a tea season followed by the same letter are not significantly different (ANOVA followed by Tukey’s HSD test: p > 0.05; n.s. not significant). Arrows in the figure indicate the dates of pesticide application

Full size image

From May to December 2015, the total number of phytoseiid mites was not significantly different between the 40-L treatment and the control, nor between the 40-L and 200-L treatments, but that in the 200-L treatment was significantly lower than in the control (F2,24 = 4.04, p = 0.024). There was no significant difference in the number of phytoseiid mites among the different pesticide treatments (January to December 2016: F2,33 = 0.40, p = 0.67; January to October 2017: F2,32 = 1.71, p = 0.19) (Figs. 2, 3). The species of phytoseiid mite collected after pesticide spraying were A. eharai, P. persimilis, Amblyseius obtuserellus Wainstein and Begljarov, E. sojaensis, and Typhlodromus vulgaris Ehara. Euseius sojaensis, which was a dominant species before the pesticide application, was barely detectable after, and the composition of the phytoseiid mite population had markedly changed as a result of the pesticide treatment (Fig. 3). Throughout the study period, the number of A. obtuserellus was significantly higher (F2,91 = 3.20, p = 0.043) in the control than in the 200-L treatment, whereas the number in the 40-L treatment was similar to that in the control (Table 2). However, there was no significant difference in the number of A. eharai (F2,91 = 1.07, p = 0.35), P. persimilis (F2,91 = 1.22, p = 0.30), E. sojaensis (F2,91 = 1.35, p = 0.26), or T. vulgaris (F2,91 = 1.00, p = 0.37) among treatments (Table 2). There were also no significant differences among the treatments in the total numbers of the five species of phytoseiid mites that were confirmed in the study tea field (F2,91 = 1.14, p = 0.32; Table 2).

Fig. 3
figure 3

Differences in the composition of the Phytoseiidae. Study periods were May to December 2015, January to December 2016, and January to October 2017. Means within a tea season capped with different letters are significantly different (ANOVA followed by Tukey’s HSD test: p < 0.05). Italicized letters denote significant differences for separate phytoseiid mite species

Full size image
Table 2 Mean (± SE) total numbers of Phytoseiidae collected during the study period (May 2015–October 2017)
Full size table

Figure 3 shows the species composition of the mite population in each year. In 2015, the numbers of Phytoseiidae (F2,24 = 4.04, p = 0.024) and A. eharai (F2,24 = 4.49, p = 0.016) were similar in the 40-L and control treatments, but the numbers in the 200-L treatment were significantly lower than those in the control treatment. There were no significant differences in the number of the other phytoseiid mites among the treatments. In 2016, the number of A. obtuserellus (F2,33 = 3.08, p = 0.053) was similar between the 40-L treatment and the control, whereas the number in the 200-L treatment was significantly lower than that in the control. Apart from A. obtuserellus, no significant differences in the number of phytoseiid mites were observed among the treatments. In 2017, the numbers of four (T. vulgaris were not detectable) species of phytoseiid mites were not significantly different among the treatments. After the spraying of pesticides, P. persimilis and A. eharai became dominant.

Behavioral avoidance of phytoseiid mite species against synthetic pyrethroid insecticide (laboratory trial)

For P. persimilis and E. sojaensis, the avoidance scores were not significantly different between the control and each of the pyrethroid treatments, but A. eharai exhibited significantly more escape and avoidance in response to permethrin or bifenthrin compared to the control (Fig. 4). The number of healthy individuals of A. eharai was significantly lower in all the pyrethroid treatments than in the control plot (Fig. 5). Phytoseiulus persimilis acted normally in response to most of the synthetic pyrethroids, and no significant differences compared with the control were observed (Fig. 5). By contrast, there were few healthy individuals of E. sojaensis remaining after the pyrethroid treatments, and many distressed individuals were also observed. Table 3 shows the destiny of the A. eharai individuals 24 h after escaping from the platform. The A. eharai individuals were transferred to individual rearing containers (Munger cells) immediately after their submergence (avoidance). In the permethrin and bifenthrin treatments, most A. eharai individuals that escaped from the platform and became submerged in the water were acting normally. On the other hand, one out of nine individuals in the fenpropathrin treatment and one out of eight in the cypermethrin treatment showed abnormal behavior, with one A. eharai individual from the cypermethrin treatment (n = 8) being dead within 24 h.

Fig. 4
figure 4

Results of the avoidance test of three phytoseiid mite species, testing leaf discs treated with four pyrethroid pesticides (n = 25). Bar sections labelled with different letters indicate significant differences among pyrethroids (significant χ2 tests between pyrethroids vs. control plot, followed by Ryan multiple comparison procedure: p < 0.05)

Full size image
Fig. 5
figure 5

Results of the avoidance test (classified into healthy, distressed, or escaped) of three phytoseiid species (n = 25; 60 min after application). Column sections labelled with different letters indicate significant differences among categories (Fisher’s exact probability test followed by Ryan multiple comparison procedure: p < 0.05)

Full size image
Table 3 Fate of Amblyseius eharai individuals 24 h after being rescued from water
Full size table

Discussion

Ikeda and Yamamoto (1989) reported that if synthetic pyrethroids were sprayed in the summer, there would be a very high probability that a resurgence of T. kanzawai would occur in the autumn and continue to the second tea season of the following year, as these phenomena were mainly caused by a pyrethroid-based decrease in the numbers of N. womersleyi, a natural enemy of T. kanzawai. In the tea field of the current study, the temporal abundance growth of T. kanzawai in 40-L and 200-L treatments in August 2016 were confirmed; however, the densities of T. kanzawai tended to decrease as time passed. In addition, N. womersleyi was not detected, and A. eharai, which also plays a role in preventing increases in spider mite population size (McMurtry 1992), and P. persimilis, which can suppress an outbreak of spider mites (Takafuji and Chant 1976; Takahashi et al. 1998), were the dominant species in the field trial of the current study.

Nagatomo et al. (1991) reported that the presence of N. womersleyi only was confirmed in a tea field adjacent to our study tea field in the 1980s, whereas in 2015–2017, A. eharai, P. persimilis, A. obtuserellus, E. sojaensis, and T. vulgaris were all confirmed, and the species diversity of phytoseiid mites had become quite rich. Ullah et al. (2016) indicated that ‘displacement in predator complexes of fruit-tree orchards could be due to different degrees of pesticide susceptibility’. Moreover, it is known that the composition of the phytoseiid mite population differs depending on the management of the tea fields (Santoso et al. 2004). As a general trend in the Japanese tea fields, the species diversity of phytoseiid mites has changed drastically since the 1960s (Toyoshima et al. 2015b), such that acaricide spraying to control T. kanzawai is occasionally curtailed (Ozawa et al. 2006). Thus, the phytoseiid species richness may contribute to the relatively rare occurrence of T. kanzawai, and, even then, at low densities (Toyoshima 2014).

In the indoor trials, most of the P. persimilis individuals continued their normal behavior even after contact with leaves that were saturated with fenpropathrin, permethrin, bifenthrin, or cypermethrin. In addition, about 40% of the submerged (escaped) P. persimilis individuals were also found in the cypermethrin treatment. According to previous studies (Wakou and Odagiri 1985; Hamamura 1986), escape from the platform and subsequent submergence of phytoseiid mites was regarded as a behavioral avoidance from pesticides on the platform, so we believe that P. persimilis in our indoor experiments also showed behavioral avoidance from the pesticide. In addition, Hamamura and Isobe (1998) discovered that P. persimilis has an extremely high resistance to fenpropathrin, permethrin, and cypermethrin at practical (used in fields) concentrations. Moreover, in our indoor experiments, 96% of P. persimilis individuals exposed to bifenthrin remained healthy. This result supported the finding that the number of individuals in the pesticide spraying area was not significantly different from that in the control treatment in the field experiments. Therefore, it was considered that P. persimilis tends to survive due to resistance and behavioral avoidance when fenpropathrin, permethrin, bifenthrin, or cypermethrin were sprayed on tea fields under conditions similar to those used in this test. There are also similar findings from previous reports on tea fields (Ozawa et al. 2014) and pear gardens (Yamazaki and Itoyama 2015).

Kishimoto et al. (2018) reported that synthetic pyrethroid insecticides have a strong influence on A. eharai. In the indoor experiments of the current study, however, the mortality rate of A. eharai was lower than that reported by Kishimoto et al. (2018), though that discrepancy may have been caused by shorter pesticide contact time and also to the existence of refugia without pesticide. In the laboratory tests, many of the A. eharai individuals submerged, following which most of the rescued individuals recovered, showing a normal behavior after 24 h. In addition, in the tea field of this current study, no significant differences were observed in the number of A. eharai in the sprayed treatments. When pesticides are sprayed (200 L/1000 m2) on the plucking surface, only about 20% of pesticide would reach inside the leaf layer (15 cm below the plucking surface) (Kawai et al. 1999), so that the number of N. womersleyi individuals in the sprayed and the control areas did not differ (Kawai 2001). However, when the spraying volume was high enough to reach inside the leaf layer (400 L/1000 m2), the population size of the predatory mite decreased (Kawai 2001). Consequently, in the spray treatment of our study (200 L/1000 m2), the ability to resist and be repelled by the pesticide might enable A. eharai to survive; alternatively, in the leaf layer, not enough of the pesticide might cover to achieve the lethal dosage for A. eharai. In the current study, control of T. kanzawai, which is conventionally achieved by spraying acaricide into the inside of the tea bushes (400 L/1000 m2), had not been carried out, but this may also have led to the maintenance of high predatory phytoseiid mite populations. In the 200-L treatment, the A. eharai incidence was significantly lower than that in the control plots in 2015, but not in 2016 and 2017. It seemed either that the growth of the tea bush provided refugia for A. eharai to shelter from pesticides and/or that the sensitivity of A. eharai to pesticides had changed. To distinguish between these phenomena, further research is needed.

In contrast, E. sojaensis, which was a dominant species before pesticide spraying, proved to be particularly susceptible to pyrethroids. The moment E. sojaensis individuals came into contact with pesticide-treated leaves, they began to act distressed and were unable to walk normally. This finding is in agreement with a previous report (Shibao et al. 2006), which showed that the mortality rate of E. sojaensis in response to permethrin at concentrations of 200, 100, and 50 ppm were 100%, whereas at 25 ppm the mortality was 63%. Euseius sojaensis showed marked susceptibility to the nonselective pesticides. Also, in the fruit orchard fields, where nonselective pesticides are frequently used, E. sojaensis is only rarely observed (Kishimoto et al. 2018). In the current study, E. sojaensis was not detected in the 200-L treatment until the end of the study, about 2.5 years after the pesticide was applied, whereas in the 40-L and the control areas only a few individuals were confirmed. When the volume of pesticide applied was 200 or 40 L/1000 m2, the pesticide coverage at 5 cm below the plucking surface was estimated to be > 90% or > 40%, respectively, whereas at 25 cm below the plucking surface it was estimated to be > 20% or > 0.1%, respectively (Kakoki et al. 2018). Thus, the pesticide coverage in the leaf layer was significantly higher in the 200-L treatment (Kakoki et al. 2015). In addition, the pesticide coverage on the lower surface of the leaf (5 cm below the plucking surface) was estimated to be 0.2% and 11.3% in the 40-L and 200-L treatments, respectively (Kakoki et al. 2015, 2018). It is possible that the side effect on E. sojaensis of the 40-L treatment was alleviated to an extent in a manner dependent on the degree of pesticide coverage in the leaf layer. Incidentally, the number of individuals of A. obtuserellus in the control and 40-L treatments tended to be more than that in the 200-L treatment.

In conclusion, by continuing the method of pesticide spraying that leaves refugia in the leaf layer free (in part) from pesticides, our research suggests that adequate populations of various species of phytoseiid mites can be maintained. In addition, it is known that a combination of phytoseiid mite species [Neoseiulus californicus (McGregor) and P. persimilis] can effectively control T. urticae (Shibao and Ioku 2016). Conserving the diversity and abundance of Phytoseiidae in this way may contribute to the stabilization of T. kanzawai at low densities.