Diversity and abundance of Phytoseiidae (Acari: Mesostigmata) in three crop management strategies of citrus orchards in Tunisia

Surveys of Phytoseiidae were carried out in three Tunisian citrus orchards; the first was conducted according to organic farming practices; the second was conducted according to integrated pest management; the third was conventionally conducted with an intensive use of pesticides. The aim of the present study was to assess Phytoseiidae diversity and densities in these orchards both on trees and weeds. The highest Phytoseiidae diversity on citrus and weeds was observed in the organic citrus orchard (eight species on citrus and eight species on weeds). In the conventionally managed orchard, only three and four species of Phytoseiidae were found on weeds and trees, respectively. The same species, Euseius stipulatus, was dominant on citrus in the three orchards, whereas it was observed only on some weeds. The prevalent Phytoseiidae species were different according to the orchards (Typhlodromus [Typhlodromus] phialatus in organic farmed plot, Neoseiulus californicus in integrated managed plot and Phytoseiulus persimilis in the conventional one). Prevalence of such species was due to their high abundance on some plants (i.e. Phaseolus vulgaris for P. persimilis). No clear impact of weed management on citrus Phytoseiidae fauna was observed. However, pesticide use seems to affect densities and diversity of Phytoseiidae. Even if no clear correlation between E. stipulatus and Tetranychus sp. was observed, it seems that E. stipulatus could feed on Tetranychus sp. but cannot quickly and efficiently control high densities of these pests.

Species diversity and abundance of Phytoseiidae are mainly affected by climate conditions, habitat stability and food resources (McMurtry and Croft, 1997). Most Phytoseiidae are able to survive and develop when preys are absent because of their polyphagous diet (McMurtry and Croft, 1997;Nyrop et al., 1998). This generalist feeding habits (pollen, fungi, and other mites) explain in part the perennial presence of Typhlodromus (Typhlodromus) pyri Scheuten and Kampimodromus aberrans (Oudemans) in European vineyards and their ability to control mite pests in such agrosystems (Duso, 1992;Lorenzon et al., 2012). Crop management could also affect Phytoseiidae densities and diversity. Many studies related the negative effects of pesticides on Phytoseiidae (i.e. Kreiter et al., 1998;Childers et al., 2001;Chen et al., 2003;Hardman et al., 2006Hardman et al., , 2007Bonafos et al., 2008;Meyer et al., 2009;Peverieri et al., 2009). In surveys comparing treated and untreated apple orchards in North Carolina, Farrier et al. (1980) showed that there were twofold more species on untreated trees compared to treated ones. Surveys carried out by Fitzgerald and Solomon (2001) showed that Christmas tree plantations chemically treated had lower Phytoseiidae densities than the untreated ones. Finally, some authors report the positive impact of agrosystem diversification on natural enemy communities (i.e. Altieri and Letourneau, 1982;Tscharntke et al., 2005). Weeds are major components of agrosystems; some authors report that they can affect occurrence of pests and beneficial insects (Altieri et al., 1977). However, only few studies have focused on the effect of weed management practices on Phytoseiidae dynamics. Some studies have shown that herbicides have directly detrimental effects on Phytoseiidae in apple orchards (Rock and Yeargan, 1973;Hislop and Prokopy, 1981), vineyards (Kreiter et al., 1993) and citrus orchards (Pereira et al., 2006;Aguilar-Fenollosa et al., 2008, 2011Mailloux et al., 2010). Some also shows that weeds can harbor Tetranychidae mites but also Phytoseiidae mites (i.e. Muma, 1975;Aceujo et al., 2003;Fenollosa et al., 2008Fenollosa et al., , 2011.
The aim of the present study is to characterize Phytoseiidae diversity and densities in three Tunisian citrus orchards managed with contrasted pesticide and weeding managements.

Studied orchards
Mite families Phytoseiidae and Tetranychidae were surveyed in three citrus orchards (two in the Cap Bon and one in Bizerte regions) from September 2009 to August 2011.
In the orchard (1) conducted with organic farming practices, only one insecticide (spinosad) was applied in April 2011 to control aphids. The ground cover was dominated by Poaceae and mechanical weeding was done once in April (Table 1). In the orchard (2) conducted according to integrated pest management (one insecticide: imidaclopride applied in April), ground vegetation was diversified and was ploughed once in April (weeds on the rows were not destroyed). In the orchard (3) conventionally conducted eleven pesticides were applied to control several pests and diseases (sulfur, abamectin, malathion, dimethoate, cyhexatin + tetradifon, methyl tiophanate, benomyl, copper, Bacillus thuringiensis). Herbicide (glyphosate) was applied two times (in November and March), and ground was ploughed several times. Phaseolus vulgaris L. was planted in April as inter-cropping on the rows under the trees ( Table 1).
Populations of whiteflies, thrips and scales were present in the three orchards but only few individuals were seen during countings and the densities remained very low during the whole study.
The climate in the Cap Bon and Bizerte regions is semi-arid marked by irregular precipitations and temperate, respectively. During the survey, the Mechanical weeding once in highest temperatures were observed in July-August (ranging between 30 to 38°C) and the rainfalls were marked from October to May with maximum rainfall observed in November with 15 mm for orchards (1) and (2) and 44 mm for orchard (3).

Mite survey
From September 2009 to August 2010, samplings were conducted at least one time a month. At each sampling date, 30 citrus leaves were randomly taken in each plot. To characterize mite fauna in ground vegetation, two liters of weeds were randomly collected. Then, each plant (weeds) and citrus leaves were transported in freezing boxes to the laboratory for mite extraction. Mites were extracted from citrus leaves and weeds using the 'soakingchecking-washing-filtering method' (Boller 1984). Then all Tetranychidae and Phytoseiidae found were counted and Phytoseiidae were identified at species level. The generic classification of Chant and McMurtry (2007) and specific literature (Ferragut et al., 2009;Papadoulis et al., 2009) were used for species identification. Specimens of Tetranychidae all belong to the genus Tetranychus, but no further identification at species level was carried out.

Data analyses
The number of Phytoseiidae species "species richness" and the Simpson diversity index (1 -D) (Simpson, 1949) were calculated to compare diversity on citrus trees and weeds in the three orchards considered. Simpson's diversity index ranges between 0 and 1; a value of 1 represents an infinite diversity and a value of 0, no diversity. Species similarity between the three orchards was estimated us-ing the Jaccard index. This index corresponds to the number of species shared by two orchards divided by the total number of species (Jaccard, 1912). This index ranges from 0 (no common species) to 1 (all species in common).

Time variation of mite density and diversity
Orchard 1. Eight species of Phytoseiidae were found on citrus (Fig. 1a). The most abundant species was E. stipulatus (92 %). The highest densities were observed in December (3 individuals / leaf) and May (2 individuals / leaf). Damages of Tetranychus sp. were observed, the highest densities (3 individuals / leaf) being observed two weeks before the Phytoseiidae peak. Then, densities of both Phytoseiidae and Tetranychidae decreased until March (Fig. 2). Two smaller peaks of Tetranychidae were observed in March and July (1 individual / leaf).
Phytoseiidae appeared on weeds in March and the highest densities were observed in June. Eight species were found ( Fig. 1b; Table 2). Sixty-three percent of Phytoseiidae found were collected on Poaceae, which were dominant in this orchard. Five species are found on citrus and weeds: E. stipulatus, Graminaseius graminis (Chant), P. persimilis, T. in very low densities (Table 2). Tetranychus sp. were present only at three dates on the same plant (S. nigrum).
Orchard 2. Four species of Phytoseiidae were identified on citrus, E. stipulatus being dominant (98 %). The highest densities were observed in April (2 individuals / leaf) (Fig. 3a). They then declined progressively to reach low numbers in summer. Few specimens of Tetranychus sp. were observed on citrus (4 females/ 30 leaves founded in July).
Eight Phytoseiidae species were found on weeds. The highest density of Phytoseiidae was observed in March (Fig. 3b). Four species are found both on citrus and weeds (E. stipulatus, N. californicus, T. (A.) rhenanoides and T. (T.) phialatus) (Jaccard index = 0.5). However, as in the orchard (1) the dominant species are different. The prevalent species on weeds was N. californicus. This species was mainly collected on Mercurialis annua L., and Malva sp., the two most abundant plants in the orchard in spring. Euseius stipulatus, prevalent on citrus, was observed on three plants in the inter-rows (Malva sp., Convolvulus arvensis and S. nigrum). The highest Tetranychidae densities were also observed in March (129 individuals/ sample) collected mainly on M. annua. After weeding, no more Tetranychidae was collected until July.
Orchard 3. Five species of Phytoseiidae were collected on citrus; E. stipulatus being the dominant species (79 %). No Tetranychidae was found. The Phytoseiidae densities were very low (always less than 0.25 individual / leaf), the highest numbers being observed between March and May (Fig. 4a).
Seven species of Phytoseiidae were observed on weeds. Four species are found both on citrus and weeds (E. stipulatus, N. californicus, P. persimilis and T. (A.) rhenanoides) (Jaccard index = 0.5). However, the dominant species were different. Phytoseiulus persimilis prevailed on weeds (78 %) and was mainly observed on the planted species P. vulgaris. The highest densities were observed in July (Table  2). Euseius stipulatus, dominant on citrus, was only found on Malva sp. Tetranychidae were observed on this plant since June and the highest densities were found in July.

Orchard comparison
Phytoseiidae on Citrus trees. Even if no statistical analyses can be carried out, some tendency on Phytoseiidae densities can be drawn. The highest densities of Phytoseiidae were observed in the orchard (1) and the lowest in the orchard (3). The highest number of Phytoseiidae species was also observed in the orchard (1). However, in the three orchards, the same species E. stipulatus prevailed. Euseius stipulatus, N. californicus and T. (A.) rhenanoides were present in the three orchards. Typhlodromus (T.) phialatus was found in the orchards (1) and (2), P. persimilis in the orchards (1) and (3) whereas Graminaseius graminis (Chant), Typhlodromus (Anthoseius) yasminae Faraji and Proprioseiopsis bordjelaini (Athias-Henriot) were found only in the orchard (1) and Paraseiulus talbii (Athias-Henriot) only in the orchard (3). Jaccard index between orchards (1) and (2) is the highest (0.6) whereas between orchards (1) and (3) and orchards (2) and (3) this index is of 0.4. Even if the species richness was higher in the orchard (1), Simpson indices are low in the three orchards (Table 3), because of the great prevalence of E. stipulatus.

Phytoseiidae on weeds. Phytoseiidae diversity
(1 -D) was higher in orchard (1) than in the two other ones whereas species richness is equivalent (Table 3). Furthermore, in this orchard, Phytoseiidae richness was similar on citrus trees and weeds, whereas in the two other ones the number of Phytoseiidae species was higher on weeds than on citrus (  (1) and (2), orchards (2) and (3) and of 0.44 between orchards (1) and (3).
Tetranychidae occurrence. Variation in time of Tetranychidae was different in the three orchards, on citrus trees and weeds. Tetranychidae were abundant on citrus in orchard (1) and nearly absent in the two others. In the ground cover of orchard (1), Tetranychidae were sporadically present but always on the same plant (S. nigrum). Tetranychidae were abundant in the ground cover of the orchard (3) on P. vulgaris and Amaranthus retroflexus L. In orchard (2), Tetranychidae were mainly observed in March especially from M. annua.

Phytoseiidae species found and their biological control efficiency
Among the thirteen Phytoseiidae species collected in this survey, five were found in the three orchards: E. stipulatus, P. persimilis, N. barkeri, N. californicus and T. (A.) rhenanoides. The majority of these species are, according to McMurtry and Croft's classification (1997), generalist predators that feed on a great variety of food sources including mites, insects, pollen ... This may explain their high abundance in absence of pest preys.
Euseius stipulatus was the dominant species on citrus trees in the three orchards. This species also prevails in other citrus orchards in Tunisia (Sahraoui et al., 2012), Spain (Pereira et al., 2006;Abad-Moyano et al., 2009, 2010 and in the Mediterranean citrus orchards in general (McMurtry, 1977). Euseius stipulatus was found even when Tetranychidae were absent. This species is considered according to McMurtry and Croft (1997) as specialized pollen feeder (Type 4). Several authors reported that its development rate is higher fed on pollen than on phytophagous mites (i.e. Ferragut et al., 1987, Zhimo andMcMurtry, 1990). Furthermore, some studies have shown that its occurrence could be related to pollen abundance (Villanueva and Childers, 2004). The abundance of E. stipulatus in spring (in the three orchards) could thus be due to the presence of pollen. However, in orchard (1), this species was also abundant in December whereas pollen quantity was low. This species is also known to feed on Panonychus citri (McGregor) (Ferragut et al., 1988(Ferragut et al., , 1992, T. urticae (Abad-Moyano et al., 2009) and eriophyid mites (Ferragut et al., 1987). Its abundance in December could thus be due to the occurrence in orchard (1) of specimens of Tetranychus sp., of eriophyid mites, or tydeid mites (which are abundant during this period) suggesting that this species could feed on these preys in citrus orchards. This species was present throughout the year during the surveys except in summer (July-August) when the temperature exceeds 30 -35 ºC. Ferragut et al. (1987) showed that this species stops laying eggs at 32°C.
Among the other Phytoseiidae species present on citrus trees, N. californicus and P. persimilis can be considered as good candidates for biological control of phytophagous citrus. Phytoseiulus persimilis is known as a specialist predator, especially efficient to control T. urticae in greenhouses all over the world (McMurtry and Croft, 1997).
Typhlodromus (A.) rhenanoides is a generalist species, but also reported to reproduce and develop on T. urticae and the red mite P. citri (Tsolakis et al., 2012), two pest mites of citrus in Tunisia (Grissa and Khoufi, 2012).
Even if the three species P. persimilis, N. californicus and T. (A.) rhenanoides are known to feed on some citrus pests, their densities in the present surveys were too low to play a key role in biological control.

Relations between mite fauna on weeds and on trees
Some authors suggested that the ground cover plants may serve as overwintering plant hosts and provide alternative food for predacious mites (Childers, 1994;Fadamiro et al., 2009)

Considerations on weed management effects
Phytoseiidae species richness was equivalent on weeds of the three orchards whereas diversity was higher in the orchard (1). This is certainly due to favorable plants present in the inter-rows. Indeed, in the orchards (2) and (3), only one or two plants harbored great quantities of N. californicus and P. persimilis, respectively. In the orchard (1), the prevalent species T. (T.) phialatus was found on various plants. The Phytoseiidae densities were much lower in weeds of orchard (1) than in the two other orchards. However, this abundance is due to the high densities of N. californicus on M. annua and Malva sp. in March and that observed in orchard (3) is due to the high densities of P. persimilis on P. vulgaris in July and August. Weeding management can certainly affect Phytoseiidae density and diversity by affecting the plant diversity in interrows. Indeed, the ground cover of orchard (1)

Considerations on orchard management effects
In the organic managed orchard (1), the species richness both on citrus and weeds was higher than in the two others. This agrees with other studies showing that arthropod diversity is higher in organic farming systems (i.e. Altieri and Nicholls, 2004;Bengtsson et al., 2005;Hole et al., 2005). However, it is also the unique orchards where despite Phytoseiidae occurrence damages of Tetranychus sp. have been observed.
The very low numbers of Phytoseiidae on citrus in the orchard (3) with extensive use of pesticides is certainly due to negative effects of pesticides. Pratt and Croft (2000) showed for instance, that insecticides were highly toxic for Phytoseiidae. Pyrethroids, which are known to be highly toxic to predacious mites (Hardmann et al., 2007;Bostanian et al., 2012), were used in the orchard (3). Moreover, applications of some fungicides including benomyl and sulfur could negatively impact Phytoseiidae densities (Childers and Enns, 1975). In addition, the lower densities in orchard (2) in regards to orchard (1) could be due to the use of imidaclopride (in orchard [2]), known to be toxic for Phytoseiidae (Bostanian et al., 2010).

CONCLUSION
The objective of this study was to assess the total abundance, species richness, and composition of Phytoseiidae and Tetranychidae in different farming systems and to examine the contributions of the vegetation present on Phytoseiidae abundance and composition. Our results provided a detailed picture of the mite community of Tunisian orchards and increase the knowledge of acarofauna associated with this crop. Low pesticide applications were correlated to high diversity and densities of Phytoseiidae on citrus. However, these high densities were not sufficient to limit Tetranychus damages. On the other hand, in treated orchards, densities and diversities of predators were lower but no damage was observed. Other factors than pesticide could explain Tetranychidae outbreaks in orchards colonized by a non-specialist predator of these mites: citrus variety (Grissa and Khoufi, 2012), rootstock (Bruessow et al., 2010) and nutritive stress of plants (Aucejo-Romero et al., 2004). Yet, little is known of such effects that it would be interesting to better characterize. Phytoseiidae were present in weeds, suggesting that they can constitute reservoirs for these predators. However, E. stipulatus prevailing on citrus were poorly present on weeds, wondering the impact of such a reservoir on pest regulation on trees. More than the impact of weeding management, the most important point seems to determine what would the best plant species in the inter-row to ensure high densities of the Phytoseiidae species also found on trees. According to the present results, it seems that Malva sp., Convolvulus arvensis and S. nigrum could be interesting. However, many questions remain to better manage weed to improve biological control on associated trees. For example, is there Phytoseiidae migration between weeds and trees? How do weed management affect this migration? Further studies are thus planned to answer these questions.