Influence of host plant species on the development, fecundity and population density of pest Tetranychus urticae Koch (Acari: Tetranychidae) and predator Neoseiulus pseudolongispinosus (Xin, Liang and Ke) (Acari: Phytoseiidae)

Abstract

The two-spotted spider mite, Tetranychus urticae Koch (Tetranychidae), is a pest of agricultural crops that could potentially be controlled by the predatory mite Neoseiulus pseudolongispinosus (Xin, Liang and Ke) (Phytoseiidae). This study investigated the development, fecundity and population density of these two mite species on three different species of bean (Phaseolus lunatus L., Lablab purpureus [L.] and Phaseolus vulgaris L. [Papilionaceae: Leguminosae]). The morphological characteristics of the host plants, including leaf area, thickness and hairiness, main stem diameter and plant height affected development rate, fecundity and population density of T. urticae and also the searching success and abundance of the predatory species, N. pseudolongispinosus. L. purpureus was found to be a superior host plant for both predator and prey species. These findings emphasize the importance of host plant characteristics on the performance of species used for biological control.

Keywords:

• host species
• Tetranychus
• Neoseiulus
• development
• fecundity
• biological control

Introduction

The two-spotted spider mite, Tetranychus urticae Koch (Acari: Tetranychidae), is an extremely polyphagous pest that has been reported on more than 900 host plant species and has been described as a notorious pest of at least 30 economically important crops (Helle & Sabelis 1985). T. urticae is distributed across the globe and has been found to be one of the major pests of greenhouse crops. Traditionally, this mite has been controlled using acaricidal sprays (Heinz et al. 2004); however, it quickly developed resistance to chemicals used for its control (Kropcznska & Tomczyk 1966). Furthermore, these mites usually live on the undersides of leaves, where they feed and produce webbings that cover certain parts of the leaf, reducing the effectiveness of chemical control (Gerson & Weintraub 2007). Consequently, biological control using predatory mites of the genus Neoseiulus (Phytoseiidae) is increasingly being adopted to control spider mites and other soft-bodied insect pests on crops (Pratt et al. 2002; Sarwar et al. 2009, 2010, 2012).

The scientific name of the predatory mite Amblyseius pseudolongispinosus (Xin, Liang and Ke) (Phytoseiidae) is a synonym of Neoseiulus pseudolongispinosus (Xin, Liang and Ke) (Acari: Phytoseiidae). This species has been collected from a wide range of plants in China (Xin et al. 1981). Zou et al. (1986) found N. pseudolongispinosus foraging on 16 tetranychid mites belonging to seven genera. It has been used as a biological control agent against soft-bodied insect pests of greenhouse crops (Sarwar & Kongming et al. 2011; Sarwar & Xuenong et al. 2011) and has been considered as a potential biological control agent for various spider mites, including Tetranychus cinnabarinus (Boisduval) (Tetranychidae). Its key attributes are its ability to establish soon after release and spread to subpopulations of the pest species, survive chemical sprays, persist on crops and provide good control of the target pest (Sarwar 2013a,b). However, the effectiveness of N. pseudolongispinosus in controlling spider mite populations has been found to vary between crop species (Zou et al. 1986; Sarwar & Kongming et al. 2011; Sarwar & Xuenong et al. 2011).

The morphological attributes of host plants, such as plant structure or architecture, leaf hairiness and the type and density of trichomes, have been found to affect the searching success of natural enemies through the interaction between the pest and its predator (Krips 2000; Raghu et al. 2004); the microstructure of plants also affects the abundance of small arthropods inhabiting the plants (Larsson et al. 1997). Roda et al. (2000) concluded that a complex leaf topography (trichomes) and spider mite webbing protected predatory mite eggs from intra-guild predation, and Cedola et al. (2001) emphasized the need to consider plant attributes as an essential and interactive component of biological control practices. Similarly, Al-Zyoud et al. (2005) showed that the effectiveness of natural enemies against arthropod herbivores varied with the characteristics of the plant on which they were found. The effectiveness of generalist phytoseiid mites as biocontrol agents against spider mites might be directly influenced by canopy structure or indirectly affected by differences in plant microclimate (Prischmann et al. 2006).

The objective of this study was to investigate the three-way interaction between the host plants and the herbivore T. urticae, and finally the predatory mite N. pseudolongispinosus once it had been introduced. Morphological variations in the leaf structure of three host bean plant species were identified to determine their impacts on the biology of T. urticae, and on the development and population density of N. pseudolongispinosus, both directly and indirectly. It was intended that these findings could then be used to improve the mass rearing technique for the predatory mite.

Materials and methods

The mite rearing experiments were conducted in the Laboratory of Insect Natural Enemies, Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Beijing, China. The two mite species (T. urticae and N. pseudolongispinosus) were initially collected from the field (crops, trees and shrubs) and reared on adzuki bean plants (Phaseolus angularis Wight). They were then individually picked and reared on three host plant species (hyacinth bean Lablab purpureus [L.], sieva bean Phaseolus lunatus L. and common bean Phaseolus vulgaris L. [Papilionaceae: Leguminosae]) to produce stock colonies. The seedlings used had between two and four leaves and were individually kept in saran plastic screen rearing cages (0.5×0.5×1 m); seedlings were replaced as necessary, e.g. when a quarter or half of the leaves of the potted plant were senescent and/or deteriorated. The plants were kept at a temperature of 25±2 °C, 60±5% RH and 16:8 h (L:D) photoperiod. Mites from the colonies were used for the trials and the data were analysed as a complete randomized design.

Cultivation of host plants and determination of their characteristics

The morphological characteristics of the three host bean plants (family Papilionaceae [Leguminosae]) used in the study were: (1) L. purpureus: aromatic, smooth textured, veined, alternate pinnate and hairy tri-leaflet; (2) P. lunatus: dense hairs along the mid-rib on the undersurface of opposite tri-leaflet; and (3) P. vulgaris: small alternate leaves and dense hairs. All three plant species possess a main stem with no branches. The suitability of each plant species for feeding and reproduction of T. urticae and N. pseudolongispinosus was initially confirmed after the preliminary tests. One seed of each species was sown separately in a 10-cm-diameter pot and grown for 40 days to obtain 15-cm-tall bean plants. Uniform practices (e.g. watering and fertilization [farm yard manure]) were adopted for growing the potted plants. The size, thickness and hairiness of leaves (seven leaves per plant), main stem diameter, and plant height of the 40-day-old potted plants (about 15-cm-tall plant per replicate) were assessed, and the characteristics of seven replicates of each species were statistically analysed.

The 15-cm-tall healthy and actively growing potted bean plants were also used to inoculate and rear the spider mite stock colonies. Mite larvae of a similar age (4–6 hours post-hatching) were collected from infested plants and gently deposited individually on plants using a wet brush, to ensure a uniform infestation. A total of seven bean plants from each of the three different species were used as hosts for the mite larvae (one plant per pot) for the experiments. The potted bean plants of each species were replaced with new ones twice a week. Three weeks later, seven plants of each species infested with spider mites were individually inoculated with N. pseudolongispinosus larvae, which were again collected individually with a wet brush. Following inoculation with predatory mite larvae, all of the plants were placed in a growth chamber at 25±2 °C temperature, 60±5% RH and 16:8 h (L:D) photoperiod. They were checked daily using a hand lens (20×) and any intruding mites, insects or predators other than T. urticae and N. pseudolongispinosus were removed using a brush.

Influence of host plant species on the life history of T. urticae and N. pseudolongispinosus

The influence of host bean species on either T. urticae or the predatory mite N. pseudolongispinosus was evaluated in terms of the duration of their immature stage (egg to adult) and fecundity. The spider mites were selected from 3-week-old colonies. For the developmental trials of T. urticae, seven newly laid (1–2-hour-old) eggs were picked and individually transferred on to a leaf, which was placed over a piece of water-filled sponge in a 10-cm-diameter Petri dish; this was replenished daily to keep the leaf fresh for the mites. The stages and numbers of eggs, larvae, nymphs and adults were determined individually and recorded every 6 hours. As soon as the immature or adult spider mites ecdysed, they were transferred individually to a new leaf, which was placed in a Petri dish covered by a 0.1 mm mesh lid. A total of seven replicates comprising seven sets of Petri dishes containing a leaf and a pair of T. urticae were conducted for each of the three host plant species. The newly emerged adults from each T. urticae pair (after 24 hours) were kept on host plant leaves in a Petri dish that contained a sponge block with filter and polythene papers at the top; these were then used for the longevity trials. The female and male mites were provided with fresh leaves in the Petri dish as and when required (every 2 days). The number of eggs laid on a leaf by each female within 7 days of egg laying were counted daily and recorded to evaluate the fecundity of T. urticae. To prevent mites from escaping during the experiments, each dish was sealed with wet cotton wool. If mortality other than natural mortality occurred, mite species were replaced and the replicate rerun. All leaves chosen for the fecundity trials were of the same quality (young 2-week-old, dark green, primary leaves) and were about 5 cm wide at the widest point near the base; and eggs were transferred individually on to each leaf. The mean development period of the mites was recorded as the total time from the egg hatching until adult emergence.

Population densities of T. urticae and N. pseudolongispinosus

The population growth rate of T. urticae was estimated 21 days after releasing, by assessing the density of offspring that had been produced by 10 pairs of T. urticae per plant, feeding and reproducing on each of the tested host species (30 pairs in total released on to three replicates of a plant species). The numbers of T. urticae per leaf (eggs, nymphs, adults) were inspected on site using a hand lens (20×). Three primary, young and dark green leaves were randomly sampled from three replicates of each species. 21 days after releasing 10 pairs of T. urticae on to each of the potted host plant of each species, 10 pairs of N. pseudolongispinosus per host plant species were released on to the leaves (30 pairs on all three replicates of a plant species). The densities of these predators were then assessed 3 weeks later on three replicates of each plant species via an on-site inspection using a hand lens. The specific-stage (larvae) of spider mites or predatory mites were selected and used for plant inoculation for the trials (each mite used for plant invasion in the experiment was in the larval stage). To assess the density of predatory mites (eggs, nymphs, adults), three leaves of uniform size from the canopy of three plants of each species were sampled at random (selecting a total of nine leaves from three replicates of each host species). The escape of mites or intrusion by any other type of organism was prevented by placing the plants in saran plastic screen cages.

Host suitability trials for N. pseudolongispinosus

This experiment was conducted to determine whether factors other than morphological characteristics of the host plants could affect the foraging behaviour of N. pseudolongispinosus, such as the production of allelochemicals or appearance of webbings as a result of foraging by T. urticae. Seven plants from each of the three species were selected and a leaf containing five to 15 prey mites was cut using scissors. Each leaf was then placed upside down on a layer of water-saturated cotton wool in a Petri dish (2 cm diameter×1 cm deep) and surrounded by wet cotton wool to prevent the mites from escaping. In addition to the five to 15 mites, eggs or nymphs, adults and their additional spider mites were supplied daily to secure a food supply for the predators and facilitate their development. A pair of newly emerged N. pseudolongispinosus was positioned with a brush on each Petri dish leaf for egg laying. The eggs laid on each leaf were checked four times a day and transferred to another Petri dish (3 cm diameter×1.5 cm deep). These eggs were then used to conduct the life-history trials. The Petri dish was covered with a 0.1-mm-mesh lid for ventilation purposes. A total of seven replicates were conducted, each of which consisted of a 2-week-old leaf in a Petri dish. After 2 days, the leaf was replaced with a new, similarly treated leaf of the same plant and mites were transferred from the old to the new leaf using a brush without injuring the mites. For each replicate, the number of eggs laid on the Petri dish leaf and numbers of larvae and nymphs (mites in the immature stage) were assessed and recorded daily. Also the duration of the immature stages and fecundity of the predator were evaluated on plant hosts. The Petri dish leaf technique was also used to determine the longevity of both species of mite by checking survival on the leaves every 6 hours. Additional observations that were recorded daily included: (1) the number of eggs laid per female over the first week following the introduction of a freshly mated pair of mites on to a Petri dish leaf; (2) the survival of both sexes of mite until natural death; and (3) the length of the egg and immature stages. Mite cultures were maintained in a controlled climatic chamber set at 25±2 °C temperature, 65±10% RH with 16:8 h (L:D) photoperiod.

Data analysis

The data on duration of the immature stages, number of eggs produced, and longevity of spider mites and predatory mites were analysed by analysis of variance (ANOVA) using SPSS (2005) software. The effects of the morphological characteristics of the host bean species on the mean development time (egg to adult) and the population density or abundance of T. urticae and N. pseudolongispinosus after 21 days were determined using a least significant difference (LSD) test.

Results

Variation in the morphological characteristics of the three host bean species, and the mean duration of the immature stages and fecundity of T. urticae and N. pseudolongispinosus feeding and reproducing on each of the bean hosts are shown in Tables 1, 2 and 3. These findings highlight the key differences in the life histories of the predator and prey mites when reared on different host species.

Table 1 Morphological characteristics of different host bean plants tested.

	Leaf (mean±SD)
Host bean plant species	Area (cm^2)	Thickness (mm)	Trichomes/ 0.25 cm^2	Main stem diameter (mm) (mean±SD)	Plant height (cm) (mean±SD)
Phaseolus Lunatus	76.60±1.72 b	0.354±0.01 b	48.20±2.43 b	3.20±0.12 a	82.20±1.28 b
Lablab purpureus	82.20±5.34 b	0.424±0.02 c	12.40±2.44 a	3.60±0.18 a	81.20±1.39 b
Phaseolus vulgaris	31.60±1.07 a	0.272±0.01 a	83.40±4.41 c	4.20±0.12 b	71.00±1.18 a

Data are average of 7 replicates. Means in the same row followed by different letters are significantly different by LSD tests at P<0.05.

Table 2 Influences of different host bean plants on life history parameters of Tetranychus urticae.

	Duration, fecundity, longevity and density of pest mite on different host bean plants (mean±SD)
Parameter	Phaseolus lunatus	Lablab purpureus	Phaseolus vulgaris
Egg	5.32±0.20 a	5.14±0.18 a	5.57±0.21 a
Larva	1.78±0.06 a	1.75±0.07 a	1.89±0.09 a
Protonymph	1.85±0.16 a	1.82±0.17 a	2.64±0.07 b
Deutonymph	2.72±0.13 a	2.68±0.13 a	3.25±0.19 b
Mean development period	11.67±0.45 a	11.39±0.46 a	13.35±0.37 b
Daily numbers of eggs laid/female	4.42±0.42 b	5.85±0.40 c	3.00±0.57 a
Longevity (days)
Female	23.57±0.71 b	25.00±0.69 b	19.28±1.89 a
Male	13.85±0.96 b	15.28±0.77 b	10.71±1.18 a
Number of mites/leaf (increased from 10 pairs for a duration of 21 days)	13.33±1.76 b	15.66±2.84 b	5.33±1.85 a

Data are average of 7 replicates. Means in the same row followed by different letters are significantly different by LSD tests at the P<0.05. Numbers of mites/leaf were the population density increased from 10 pairs of mites for duration of 21 days.

Table 3 Influences of different host bean plants on life history parameters of Neoseiulus pseudolongispinosus.

	Duration, fecundity, longevity and density of predatory mite on different host bean plants (mean±SD)
Parameter	Phaseolus lunatus	Lablab purpureus	Phaseolus vulgaris
Egg	3.46±0.10 a	3.28±0.06 a	3.53±0.11 a
Larva	1.82±0.07 a	1.78±0.08 a	1.96±0.06 a
Protonymph	2.92±0.07 a	2.89±0.09 a	3.15±0.12 a
Deutonymph	3.82±0.07 a	3.80±0.08 a	4.00±0.09 a
Mean development period	12.02±0.12 a	11.75±0.18 a	12.64±0.24 b
Daily numbers of eggs laid/female	2.42±0.52 b	3.57±0.36 c	1.14±0.14 a
Longevity (days)
Female	24.00±1.11 b	27.71±1.20 c	20.00±0.81 a
Male	16.28±1.01 b	19.57±0.64 c	13.42±0.86 a
Number of mites/leaf (density increased from 10 pairs for a duration of 21 days)	8.66±1.45 a	9.66±2.84 a	2.00±0.57 b

Data are average of 7 replicates. Means in the same row followed by different letters are significantly different by LSD tests at the P<0.05. Numbers of mites/leaf were the population density increased from 10 pairs of mites for duration of 21 days.

Characteristics of host plants and influence on life history traits of mites

There were significant differences in the morphological characteristics of the three host bean species tested. Plants of P. lunatus and L. purpureus were taller (F =23.141; P=0.001), and had larger (F=70.679; P=0.001) and thicker (F=21.309; P=0.001) leaves than P. vulgaris; however, their leaves were less hairy (F=120.563; P=0.001) and they had a smaller main stem diameter (F=11.692; P=0.002) than P. vulgaris. These morphological differences, in turn, led to differences in the biological parameters of the mites inhabiting them. Both mite species exhibited significantly higher development rates on P. lunatus and L. purpureus than on P. vulgaris. The host plants P. lunatus and L. purpureus significantly differed from each other in leaf thickness and hairiness (Table 1). The increased thickness and reduced hairiness of P. lunatus and L. purpureus leaves enabled the motile stages of both mites to move freely and swiftly on the leaf or petiole, and on to the main stem of the plants. Thus, the mite species on both these hosts could be more likely to find a microhabitat with more nutrients or a place with a sufficient food supply.

Influence of host plants on T. urticae

There were significant differences in the duration of development between T. urticae feeding on the three different host species (F=5.993; P=0.01), with durations of 11.39 days and 11.67 days for the immature stage of the spider mites feeding on L. purpureus and P. lunatus, respectively, compared with 13.35 days on P. vulgaris; and the durations of the protonymph and deutonymph stages varied in a similar way (Table 2). There was also a significant effect of host plant species on the fecundity of T. urticae (F=9.000; P=0.002), with females producing 5.8, 4.4 and 3.0 eggs per day when they fed on L. purpureus, P. lunatus and P. vulgaris, respectively. T. urticae that fed on P. vulgaris also had shorter longevities for both females (F=5.769; P=0.012) and males (F=5.57; P=0.013) than mites that fed on L. purpureus and P. lunatus (Table 2). There was no significant difference in the incubation period of eggs and the duration of the larval period between the three bean species (egg: F=1.105; P=0.353; larva: F=0.886; P=0.429), but there was a significant difference in the duration of the protonymph (F=10.636; P=0.001) and deutonymph (F=4.108; P=0.034) stages (Table 2).

Population densities of T. urticae and N. pseudolongispinosus on the host plants

The population density of T. urticae was significantly higher on L. purpureus (15.66 mites/leaf) than on P. lunatus (13.33 mites/leaf), and the same pattern was seen for N. pseudolongispinosus (9.66 mites/leaf versus 8.66 mites/leaf, respectively). By contrast, both species had significantly lower densities on P. vulgaris (5.33 mites/leaf versus 2.00 mites/leaf) for T. urticae and N. pseudolongispinosus, respectively (F=6.008; P=0.037 and F=4.937; P=0.054) (Tables 2, 3). Host plants with larger and thicker leaves had higher densities of T. urticae mites due to increased levels of reproduction over a 21-day period (Tables 1, 2, 3). Consequently, the bean species with larger, thicker and less hairy leaves, a smaller main stem diameter, and greater plant height tend to have higher densities of T. urticae and N. pseudolongispinosus.

Suitability of host plants for N. pseudolongispinosus

The predatory mites N. pseudolongispinosus, which feed on T. urticae, were able to survive, develop and reproduce successfully on all three of the tested bean species. This species exhibited the highest densities while living and reproducing on L. purpureus, followed by P. lunatus and P. vulgaris; however, the duration of the immature stages was not significantly affected by host species (egg: F=1.800; P=0.194; larvae: F=1.615; P=0.226; protonymph: F=1.954; P=0.171; deutonymph: F=1.860; P=0.184). The longevity of both males and females varied greatly with host plant species (female: F=13.261; P=0.001; male: F=12.822; P=0.001). The external characteristics of the bean species also significantly influenced the densities of N. pseudolongispinosus that were produced over a 21-day period (F=5.619; P=0.013) and the immature development rates (mean generation time), with faster rates on L. purpureus (11.7 days) and P. lunatus (12.0 days) than on P. vulgaris (12.6 days) (Table 3).

Discussion

The external characteristics of host bean plants: leaf area, thickness and hairiness, main stem diameter, and plant height were found to vary significantly between the species tested. The stem diameter of P. vulgaris and plant height of P. lunatus had no effect on the abundance of T. urticae or N. pseudolongispinosus over a 21-day period. However, the size or area, thickness and hairiness of leaves significantly affected the abundance of both species of mite. Both T. urticae and N. pseudolongispinosus developed faster, lived longer (higher survival rates) and produced more offspring on L. purpureus and P. lunatus than on P. vulgaris, leading to higher densities after 21 days of reproduction. From this, it can be inferred that L. purpureus and P. lunatus are better host species than P. vulgaris for rearing the prey and predatory mites in the laboratory. However, more information is needed to interpret the influences of host characteristics on the development, fecundity and population density of the predatory mites. The host plants L. purpureus and P. lunatus have smoother leaf surfaces with less hairiness along the mid-rib (mid-vein), whereas P. vulgaris has heavy leaves with dense hairs, which hampered the movement of N. pseudolongispinosus, leading to a lower foraging efficiency of this predator. Skirvin & Fenlon (2001) similarly reported that the chrysanthemum's morphology was the only factor that influenced the foraging behaviour of Phytoseiulus persimilis Athias-Henriot in terms of its functional responses, with the predatory mite foraging fewer eggs on plants with smooth and waxy leaves.

The development rates of the predator also varied with host structure. According to our daily observations during the experiment, nymphs of N. pseudolongispinosus had more difficulty in catching their prey and struck T. urticae many more times before handling them successfully on P. vulgaris than on L. purpureus and P. lunatus. The predatory mites also frequently fell off the leaf and took longer to locate the prey colony on P. vulgaris. Thus, the hairy leaves of P. vulgaris impeded the predatory mites from searching and handling their prey, causing them to expend more energy in developing into adults and reproducing. Similarly, Krips et al. (1999) and Krips (2000) found that the predator P. persimilis had the highest walking speed and population growth along with the highest searching efficiency and predation rates on Gerbera species that had the lowest leaf hair density. Thus, Krips et al. (1999) demonstrated how the architecture of the plant canopy and surface texture of the host leaf influenced the searching behaviour of predatory mites, and Carter et al. (1984) suggested that overly smooth leaf surfaces might have had negative effects on the searching efficiency of coccinellid larvae.

The fine trichomes on the leaves L. purpureus and P. lunatus may provide additional benefits for T. urticae in terms of helping them to retain their grip on the plant, which, in turn, benefits the reproduction of T. urticae and increases the foraging efficiency of N. pseudolongispinosus. Walter & O'Dowd (1992) concluded that leaf domatia or trichomes hampered the predatory mites’ efficiency to search and handle the prey in terms of its abundance; and Price et al. (1980) and Van Haren et al. (1987) demonstrated that the mechanical hindering and sticky exudates from trichomes had detrimental effects on the searching ability or dispersal success of predators. Stavrinides & Skirvin (2003) reported that the number of prey consumed by a predator was inversely correlated to trichome density.

The predatory mite N. pseudolongispinosus, which fed on T. urticae, took 11.7 to 12.6 days to complete its life cycle. It has previously been shown that the development rate, fecundity and generation time of Neoseiulus not only varies between host plants, but may also be affected by geographical and climatic factors (Sarwar 2013b). Zhou & Chang (1989) studied the biology of A. pseudolongispinosus at five different temperature regimes (18–34 °C) and showed that the mean generation time varied from 32.6 to 12.6 days. Zhang (1995) reported that mean oviposition and development rates of N. pseudolongispinosus were 2.8 eggs female⁻¹ day⁻¹ and 0.15 day⁻¹, respectively. Variations in the rates of population increase of predatory mites observed in previous studies are due to the predatory potential of the species studied, which affects the accumulation of energy and consequently the number of eggs produced, but may also be affected by climatic conditions.

In this study, thicker leaves were found to be harbour to higher densities of T. urticae and N. pseudolongispinosus. T. urticae fed more efficiently on L. purpureus and P. lunatus than on P. vulgaris, leading to higher population densities on these two host species. Wermelinger et al. (1991) reported that the nutrition level of host plants influenced the development rate, egg production and longevity of T. urticae. Prey that fed on thicker host leaves have a higher rate of nutrient uptake and thus greater nutrient storage in the eggs and immature stages, which, in turn, supports a faster population increase of N. pseudolongispinosus; this was demonstrated by the finding that the predator reached a high density within 21 days in this study. Gillespie & Quiring (1994) also determined that exudates from glandular hairs on leaves were toxic to predatory mites; therefore, in addition to the morphological characteristics of the host plant, the herbivore's performance is also likely to be affected by the overall physiological state of the plant, which may lead to the production of chemicals that interfered with all three trophic levels (predator, prey and plant host). The size of leaves has also been shown to affect the efficiency of the searching success of N. pseudolongispinosus, for a given density of prey in a given space (Shih & Wang 2001; Wang & Shih 2001). However, the findings of the present study suggest that the predatory mites’ searching success was more likely to be affected by the morphological features of the plant, which supports the findings of Malison (1996), who noted that the number of phytoseiids per leaf was positively correlated with leaf area; and of Prischmann et al. (2006), who reported that canopy size did not significantly affect the density of the pest or predatory mites, but architecture did have an impact. These findings indicate that it is the morphological characteristics of the host plant leaves rather than stem diameter that affect the risk of predation of spider mites.

Marquis et al. (2006) suggested that plant quality and predation risk of herbivores could be influenced by plant productivity, structural complexity, vigour and size. Price et al. (1980) showed that the host plant characteristics and prey densities had the greatest impacts on natural enemies by influencing their searching success, the quality of diet of resources (prey and host plant), and consequently their biology. From these findings, three inferences can be made about the influence of host plant species on the abundance of N. pseudolongispinosus feeding on T. urticae: (1) the prey might be impaired by host plant exudates, leading to reduced rates of nutrient uptake and thus lower nutritional value of the prey being offered to the predators; (2) the morphological structures and characteristics of the host plants, including exudates from the trichomes, may impair the activity, searching and handling success of the predators; and (3) variations in prey densities along with the size of the leaves, affect the predators’ searching and handling behaviour.

Conclusions

There were significant differences in the abundances of both T. urticae and N. pseudolongispinosus reared on three bean species, which, in turn, led to differences in their development rates and fecundities. All three of the bean plant species tested were suitable hosts for T. urticae, and N. pseudolongispinosus developed and reproduced effectively on T. urticae populations that fed on all of these species. However, the different characteristics of the host bean species affected: (1) the searching success and foraging capacity of the predators and population densities of the prey; (2) the availability of suitable food for both the prey and predator; and (3) the quality of food. Both T. urticae and N. pseudolongispinosus populations developed faster and produced larger numbers of offspring over a 21-day period when they fed on L. purpureus and P. lunatus than on P. vulgaris. The main host leaf attributes that impacted on the mite populations were the size, hairiness and thickness of leaves, as well as plant architecture such as stem width and canopy height. Consequently, L. purpureus and P. lunatus are more suitable hosts than P. vulgaris for the rearing of both T. urticae and N. pseudolongispinosus in the laboratory, as they result in shorter development durations, greater longevity and higher fecundity rates of the mites. It is thought that the thicker and larger leaves of these two bean species contain greater nutritional value or a higher concentration of nutrients, allowing higher densities of both the prey and predator to be maintained. In conclusion, L. purpureus and P. lunatus are better host bean species than P. vulgaris for the laboratory rearing of both T. urticae and N. pseudolongispinosus.

Acknowledgements

The author wishes to thank the Institute of Plant Protection staff, China, and Higher Education Commission, Islamabad, Pakistan, for technical and financial support.