Neoseiulus californicus (McGregor, 1954) preying in different life stages of Tetranychus urticae Koch, 1836 (Acari: Phytoseiidae, Tetranychidae)

The predaceous mite Neoseiulus californicus is one of the major biological control agents of tetranychids in greenhouses of several countries. The two-spotted spider mite Tetranychus urticae is one of the main pests affecting rose (Rosa spp.) cultures in Brazil. Chemical methods are used for its control, causing a significant environmental impact. Thus, this work aimed to study the predatory potential of N. californicus as an agent of biological control of T. urticae on roses. For the predatory capacity studies, 40 T. urticae mites /arenas of Jack bean leaves (Canavalia ensiformis) were offered to one specimen of each life stage of N. californicus. The adult females were the most efficient in preying upon immature stages, followed by nymphs. For the functional and numerical responses, adult females of N. californicus were confined to arenas made of Jack-bean leaves, and offered immature stages of T. urticae at the following densities: 0.14, 0.28, 0.70, 1.4, 2.8, 4.2, 4.9, 6.3, 7.7, 9.8, 14.1, 17.6 and 28.2 / cm2. The number of killed prey (functional response) and the number of eggs laid by the predator (numerical response) were assessed every 24 hours for 8 days. A type II functional response was inferred through a regression analysis. N. californicus preyed on a maximum of 60 pecimens of T. urticae per adult female per day.


INTRODUCTION
The two-spotted spider mite Tetranychus urticae Koch (Acari: Tetranychidae) is a major pest in greenhouses. The stability conditions of temperature and humidity maintained for plant growth also generally favour the rapid development of this pest (Zhang 2003).
The two-spotted spider mite usually causes chlorotic or discoloured spots on the leaves, which later turn into bronze and get dry. Ultimately, defoliation, sometimes total, occurs (Cruz 2000). Other damages caused to cultures are: plant growth reduction and web formation and depreciating the final product (Barbosa 2003). The production and consumption of ornamental flowers and plants in Brazil have followed the expanding world market trend (Junqueira and Peetz 2002). Roses, Rosa spp., are the major cut flower products in Minas Gerais (Landgraf and Paiva 2005). In Brazil, the rose-growing area is of 426 ha (Junqueira and Peetz 2002), with Minas Gerais accounting for 35.38 % of the production (Landgraf and Paiva 2009). Pest control is one of the challenges faced by ornamental flower and plant growers, with any foliage and flower injury caused by insects and other arthropods being unacceptable for marketing (Carvalho et al. 2009). In spite of often being the only strategy adopted by producers, chemical control of T. urticae has become increasingly difficult due to the rapid development of resistance to pesticides and decrease in the number of products registered for use (Zhang 2003;Sato et al. 2004Sato et al. , 2005. As a result, producers may increase pesticide doses, causing health risks to labour workers, environment and consumers (Kim et al., 2006). Predatory mites of the family Phytoseiidae are considered the most effective natural enemies for biological control of pest mites (Moraes et al. 2004).
The first consistent and well-documented attempts to use predatory mites as a means of controlling agricultural pests date from the second half of the twentieth century. However, pest mite control projects using predatory mites, as well as ecological studies under field conditions are still scarce in Brazil (Moraes 2002). The species Neoseiulus californicus (McGregor) (Phytoseiidae) (Reis et al. 2005) is among the main predatory mite used in greenhouses. Phytoseiid mites are the subject of many taxonomic, biological and ecological studies for being mainly predators, and of great importance in the control of phytophagous mites, especially those belonging to the family Tetranychidae (McMurtry and Croft 1997).
Assessment of the functional response and predatory capacity of Phytoseiidae when offered T. urticae is a critical step in determining their ability to regulate preys.
The functional response concept first described by Holling (1959) has been widely utilized to evaluate effectiveness of predacious insects and mites (Laing and Osborn 1974;Everson 1980;Sabelis 1985;Trexler et al. 1988;De Clercq et al. 2000;Badii et al. 2004;Reis et al. 2003Reis et al. , 2007Timms et al. 2008). The functional response of the predator to the prey can follow one of three following mathematical models: Type I (linear), Type II (convex) or Type III (Sigmoid). In Type I model, the proportion of prey consumed increases linearly with prey availability up to a maximum. In Type II model, the proportion of prey consumed declines monotonically with prey density. Type III model depicts a sigmoid relationship in which the proportion of prey consumed is positively density-dependent over some ranges of prey density (Holling, 1959(Holling, , 1961Trexler et al. 1988;De Clercq et al. 2000;Timms et al. 2008).
Thus, the objective of this work was to study the predatory capacity, functional and numerical responses of the predatory mite N. californicus as a contribution to its use as an agent of biological control of T. urticae in rose greenhouses.

Two-spotted spider mite rearing
Adults of T. urticae were collected from rose (Rosa alba L.) leaves under greenhouse conditions at the "Fazenda Experimental da Empresa da Pesquisa Agropecuária de Minas Gerais -EPAMIG", in Lavras, MG, Brazil, without pesticide application. Rearing was conducted at the Acarology Laboratory of the "Centro de Pesquisa em Manejo Ecologico de Pragas e Doenças de Plantas -EcoCentro, EPAMIG" , at the "Universidade Federal de Lavras -UFLA", Lavras, MG, at 25 ± 2°C, 70 ± 10 % RH and 14 hours photophase, on Jack-bean [Canavalia ensiformis (L.) DC.] leaves placed on a disk of foam (1 cm thick) covering the bottom of a Petri dish 15 cm in diameter, without cover, and constantly moistened with distilled water. Hydrophilic cotton was placed around the leaves as a physical barrier to prevent mite escape, according to the technique of McMurtry and Scriven (1965).

Predatory mite rearing
Adults of N. californicus were collected from Jackbean leaves in a greenhouse without pesticide application, at the "Instituto Federal de Educação, Ciência e Tecnologia do Sul de Minas Gerais", Inconfidentes Campus. The mites were placed on flexible black plastic PVC arenas (3 cm in diameter), floating in distilled water in Petri dishes (15 cm in diameter), without cover, according to the methodology of Reis and Alves (1997). These arenas were remained in the Acarology Laboratory, EPAMIG/ EcoCentro, at a temperature of 25 ± 2°C, 70 ± 10 % RH and 14 hours photophase. The predatory mites were fed with castor bean (Ricinus communis L.) pollen and T. urticae from the stock rearing colony.

Studies of predator abilities
The predatory capacity potential of N. californicus was evaluated through bioassays in arenas (3 cm diameter), constituted of leaves of Jack-bean plant floating in distilled water in Petri dishes (15 cm diameter), without cover, according to methodology adapted from Reis et al. (1998).
The possible combinations between each stage of development of the predatory mite N. californicus [larva, nymphs (protonymph and deutonymph), adult female and male] and each phase of development of the pest-mite T. urticae [egg, larvae, nymphs (protonymph and deutonymph) and adult] were studied. Four bioassays, (one for each phase of development of T. urticae), were carried out in a completely randomized design. In each arena, for each phase of the predator, 40 pest-mites from the rearing stock were supplied in the phase intended for the study, with 10 replications.
The evaluations were made after 24 hours, counting the number of T. urticae consumed by the predatory mite. The mean daily mite consumed were compared with variance analyses and subse-quent mean comparison test (Tukey test). The data were transformed in √ x + 0.5 (Ferreira 2008).
For the functional response (number of preys killed) and numerical response (number of eggs laid), adult females of N. californicus were confined into arenas as previously described. Immature stages of T. urticae were offered to one N. californicus adult female. Larvae and nymphs were used based on the prey's preference in immature stages for predation by phytoseiids (Gravena et al., 1994;Reis et al., 2000), at the following densities: 0.14, 0.28, 0.70, 1.4, 2.8, 4.2 and 4.9 (seven replicates), 6.3 (four replicates), 7.7 (three replicates), 9.8, 14.1, 17.6 and 28.2 / cm 2 (two replicates) according to Reis et al. (2003Reis et al. ( , 2007 methodology. The number of replicates decreased when T. urticae densities increased due to the difficulty in handling a large number of mites. The number of preys consumed (functional response) and the number of eggs laid by the predator (oviposition rate) were evaluated every 24 hours for 8 days. These data (consumed preys and number of eggs) were submitted to regression analyses (Ferreira 2008). The number of preys was daily replaced up to the limit of the initial number in each prey density.

Predatory capacity potential
All the life stages of N. californicus consumed more larvae of T. urticae than other stages. The adult females were the most efficient consuming 86.3 %, followed by the adult males with 54.8 %, the nymphs  Figure 1). Similar results were observed by Forero et al. (2008) in studies with N. californicus in rose cultivation in Sabana de Bogotá. They also reported a higher consumption of larvae and nymphs, thus confirming the predator's preference for immature forms of the phytophagous mite. Several studies with the predatory mite species: Phytoseiulus persimilis Athias-Henriot, Galendromus occidentalis (Nesbitt) and N. californicus also showed, that these predators preferred the nymphal stage than the eggs Panonychus citri (McGregor) in citrus, when offered simultaneously (Xiao and Fadamiro 2010).
According to Blackwood et al. (2001), spider mite larvae may be more profitable than eggs for generalist phytoseiids, in regards to both nutritional benefit and handling time. In their study, these latter authors observed that females of several generalist species often attempted to pierce eggs without success before attacking larvae. One possible explanation for this observation would be that generalist species might have mouthparts not as effective at piercing the T. urticae egg chorion as mouthparts of more specialized species. Flechtmann and Mc-Murtry (1992) found that cheliceral characteristics differ between Phytoseiidae species primarily feeding on spider mites and pollen.
As in the present work, Cedola et al. (2001) evaluated the predatory capacity of N. californicus at the nymph stages (protonymph and deutonymph) and observed that these immature stages have a lower predatory capacity than the adult females. Such a result can be explained by the greater energy requirements for oviposition compared with the immature phases, although the larger size of the adult females also must be taken into account. All N. californicus phases showed a low consumption of T. urticae adults, compared to the other offered phases (Table 1), probably due to their greater mobility and size.
The N. californicus larvae were less efficient predators, likely due to their reduced size and the short duration of this stage, as also observed by Reis and Alves (1997) and Reis et al. (1998).

Functional Responses and Oviposition Rate
Both the functional response (preys consumed) and oviposition rate of N. californicus were found to increase with increasing prey density, presenting a positive and significant interaction with the coefficient of determination (r 2 ) of 0.94 and 0.83, respectively (Figures 2 and 3). The maximum oviposition of 1.5 eggs / day was found at the density 6.3 preys / cm 2 (Table 2). The regression analysis results showed that the functional response is of type II (convex). The num-ber of attacked preys by the predator's female increases asymptotically with the increasing number of offered preys. Then it gradually decelerates until reaching a level at which the predatory capacity rate becomes more or less constant, regardless of prey density. Such response, according to Hassel (1978) is supposedly typical of arthropod predators. Although predators which exhibit the Type III functional response are commonly regarded as efficient biological control agents (Fernández-Arhex and Corley 2003;Pervez and Omkar 2005), many of the predators that have been successfully released as bio-logical control agents have been shown to exhibit a Type II functional response (Xiao and Fadamiro, 2010). According to Castagoli et al. (2002), less clear features were evidenced for N. californicus that is generally considered as a specialist or selective type II predator, but it seems that an intermediate behaviour between specialist and generalist type seems more appropriate.
At a density of 14.1 preys / cm 2 , the percentage of predation decreases fewer than 50 %, suggesting a predator's satiation, or that there may have been interference in its ability to prey due to increased prey density (Sandness and McMurtry 1970;Reis et al. 2003), indicating that N. californicus will be more efficient at densities between 10 and 20 preys / cm 2 than for higher ones (Table 2). Koehler (1999) corroborate the observation that mite control by predators is very efficient, particularly at low prey densities, including the mite of the family Phytoseiidae. Blümel et al. (2002) concluded that N. californicus is an efficient predator for the control of mites on rose and vegetable cultivations, also surviving for long periods without the presence of preys, and by feeding on pollen and other alternative food sources. A similar result was found  For effective employment of N. californicus as a biological control agent, its predation activity must be better understood. The functional responses of N. californicus can differ depending upon a number of factors: strains with different nutritional histories (Castagnoli et al. 1999), environmental temperature (Gotoh et al. 2004) and the use of pesticides (Poletti et al. 2007). Ahn et al. (2010) studied the functional response of N. californicus on T. urticae in strawberries and concluded that the functional response of adult female was not influenced by non-glandular trichomes on abaxial leaves but was affected by temperature. Castagnoli et al. (1999) stated that it is important to know the host and the prey history of the predator strain used for successful biological control. However, this study has also shown that N. californicus has an excellent predatory capacity on T. urticae (60 pest-mites per adult female per day).

CONCLUSION
Although obtained in laboratory conditions, the results of this work allow to conclude that N. californicus can be considered an efficient predator of immature stages of T. urticae, at different densities of the prey. The results suggest that the testing should be done in a greenhouse for consolidation of biological control with N. californicus in integrated management systems. ACKNOWLEDGEMENT To Fundação de Amparo à Pesquisa do Estado de Minas Gerais -FAPEMIG, for financial support; to Conselho Nacional de Desenvolvimento Científico e Tecnológico -CNPq and, Coordenação de Aperfeiçoamento de Pessoal de Nível Superior -CAPES, for the scholarships granted.