Distribution and biological features of Typhlodromus (Anthoseius) recki (Acari: Phytoseiidae) on Tetranychus urticae, T. evansi (Acari: Tetranychidae) and Aculops lycopersici (Acari: Eriophyidae)

The mite family Phytoseiidae contains predatory species used in biological control. However, among the 2,521 valid species, only a few are presently considered for pest management. The study focuses on a species, frequently observed in agrosystems, but whose biology is almost totally unknown: Typhlodromus (Anthoseius) recki. The objectives were to assess its distribution and to determine under laboratory conditions: (i) the development of three populations fed on three tomato pests (Tetranychus urticae, T. evansi, Aculops lycopersici), (ii) the dispersion along the tomato stem, and (iii) the dispersion from mint to infested tomato leaf discs. Based on its distribution, this species seems to be adapted to Mediterranean climate, and is preferentially found on Boraginaceae, Solanaceae, Asteraceae and Lamiaceae. Results show that T. (A.) recki is a generalist predator, with good performance regarding the number of prey consumed. Differences were observed between the populations considered; one population out of the three tested shows even a higher mean consumption of T. urticae than that recorded for Neoseiulus californicus and N. cucumeris. It is attracted by food source and is able to walk along the tomato stem. These results are encouraging enough to consider T. (A.) recki as a good natural enemy candidate for biological control (especially based on biodiversity conservation). Further experiments should be developed to test the predatory ability of this species at larger scale (entire tomato plants and tomato fields or greenhouses).


Introduction
Mites of the family Phytoseiidae are predators, and some species are used for controlling pest mites and small insects, in various crops all over the world (Gerson et al. 2003; McMurtry et al. 2013. This family contains 2,521 valid species (Demite et al. 2020) but only a few are presently used in biological control, probably because of little knowledge on the biological features of most species. We thus hypothesise that much more species than the 2030 species presently known for their predatory efficiency could be exploited for biological control issues (Van Lenteren 2012; Van Lenteren et al. 2017. Phytoseiidae mites are naturally occurring in crops and uncultivated areas ). Most Phytoseiidae species are generalist predators (McMutry & Croft 1997; McMurtry et al. 2013); they feed on prey and also on pollen, plant exudates, sometimes plant tissues, fungi, etc. Their occurrence and diversity are strongly affected by plant phylloplane features (Karban et al. 1995; Kreiter et al. 2002; Schmidt 2004; Tixier 2018. Tomato plants have specific morphological and chemical characteristics that can act as a selective filter on Phytoseiidae communities. Trichomes on tomato leaves and stems hinder mite dispersal and development, Phytoseiidae get stuck on the sticky trichomes and die (Kennedy 2003; Simmons and Gurr 2005; Sato et al. 2011; Van Houten et al. 2013a. Alkaloids contained in tomato and ingested by prey adapted to this plant, can also be toxic for nonadapted predators (Koller et al. 2007; Ferrero et al. 2014a. However, some studies report the presence of Phytoseiidae on tomatoes (i.e. Moraes et al. 1986; Tixier et al. 2020. Studies on Phytoseiulus longipes Evans show that strains of this species collected on Solanaceae in Argentina and Brazil are adapted to tomato and can control both Tetranychus urticae (Koch) and T. evansi Baker & Pritchard on this crop (Furtado et al. 2007; Ferrero et al. 2014a. However, as P. longipes is endemic to the Neotropical region, the introduction of this species in Europe is confronted with regulatory and environmental problems. In a recent study, we showed that Typhlodromus (Anthoseius) recki Wainstein is one of the species most frequently reported on Solanaceae in Europe . It is also observed on plants with high trichome density, or presenting spiny and sticky trichomes (especially on plants of the families Lamiaceae and Boraginacae) (Moraes et al. 1986; Swirski and Amitai 1997; Tixier et al. 2020. Little is known on the biology of T. (A.) recki; it has been reported on tomatoes, feeding on mites and thrips, in Turkey and Iran (Hajizadeh et al. 2006; Kumral andCobanoglu 2015a, b). In the present study, a series of experiments was carried out to characterise the feeding and dispersal habits of T. (A.) recki. Furthermore, its current distribution was also analysed.

Distribution of Typhlodromus (Anthoseius) recki
Information on the geographic distribution and plant occurrence was retrieved from 59 publications (from 1958 to 2018) reporting T. (A.) recki occurrence (database available on request).

Plant Nomenclature
We followed the Angiosperm Phylogeny Group III's nomenclature (2009) for family, genus and species names.

Feeding habits of Typhlodromus (Anthoseius) recki in the laboratory
The aim was to characterise the ability of T. (A.) recki to feed on three prey species. Three populations of T. (A.) recki collected in France on Solanaceae and Lamiaceae were considered (Table 1). Two "control" species, commonly used in biological control (Knapp et al. 2018), both collected on Solanaceae in France (Table 1), were also studied: Neoseiulus californicus (McGregor) and N. cucumeris (Oudemans). The predatory mites were maintained in rearing units, as described by McMurtry and Scriven (1965). Pollen of Typha sp. was added daily as food source.
Feeding ability was assessed on three prey species: (i) T. evansi collected on tomato, Solanum lycopersicum L., in a greenhouse at SaintJeannet (AlpesMaritimes, SouthEast Three treatments were tested: (i) T. urticae on bean leaf discs, (ii) T. evansi on S. nigrum leaf discs and (iii) A. lycopersici on S. nigrum leaf discs. As T. urticae does not develop well on S. nigrum, it was impossible to carry out the tests on this latter plant. Three tetranychid females (T. evansi or T. urticae) were placed on a leaf disc (2.5 cm in diameter) of S. nigrum or P. vulgaris var. Contender, disposed on humid cotton in a Petri dish (10 cm in diameter) to avoid mite escape. After two days, the Tetranychidae females were removed, the number of eggs laid was counted and a predatory female (young gravid female issued from the rearing units) introduced in the experimental arena. The experimental conditions were 25°C, 70% relative humidity and 16D: 8N photoperiod. The number of prey and predator eggs was counted each day during four days as a rapid test to screen the predator efficiency.
For A. lycopersici experiments, S. nigrum leaf discs (2.5 cm in diameter), highly infested with this pest, were placed on humid cotton in a Petri dish (10 cm in diameter). One predatory female (young gravid female issued from the rearing units) was introduced and the number of eggs laid by the predator was counted daily during four days. The experimental conditions were the same as for the Tetranychidae experiments. For each prey and predator population / species treatment, 15 replicates were performed.
Analysis of variance (ANOVA) followed by mean comparison tests (NewmanKeuls) were carried out to compare the mean daily fecundity and mean daily egg consumption of the populations / species considered for each food treatment. The analyses were performed using Statistica (Statsoft France, 2010).

Dispersal abilities of Typhlodromus (Anthoseius) recki in the laboratory
The experiment aims at determining "in vitro" conditions how T. (A.) recki disperses along tomato stems. The population of T. (A.) recki used was that collected in SaintGuilhemle Désert (Hérault, South of France) on Mentha suaveolens Ehrhart (Table 1). This population was chosen because of its higher fecundity compared to the other strains, when fed on A. lycopersici. The experimental design consisted of a detached tomato stem (var. Marmande) (10 cm in length) placed in a vertical position within a box filled with agar ( Figure 1). Agar was maintained moistened (wet cotton strips placed around the agar layer) to avoid mite escape and keep the leaf discs and the stem turgescent. A leaf tomato disc with pollen or A. lycopersici was placed at one extremity of the stem, whereas a tomato leaf disc without food was placed at the other extremity. Shelters (wool yarns and cardboard) were disposed on the leaf discs, as nests where Phytoseiidae use to lay eggs. The boxes were turned each day to avoid phototropism effects. Fifteen replicates were carried out for each treatment (pollen or A. lycopersici). For each replication, one young gravid female of T. (A.) recki (issued from rearing units) was placed in the middle of the stem. The position of the female, the number of eggs and immatures on each leaf disc was counted for seven days. Student ttests were carried out to compare the mean number of eggs laid on leaf discs with and without food.   recki. This experiment aims at determining under laboratory conditions if T. (A.) recki would disperse from mint to tomato. The population used was the same as that considered for the stem dispersal experiment (Saint GuilhemLeDésert). The mint leaves were collected at Saint GuilhemLeDésert. Five females of T. evansi were deposited for three days on a tomato leaf disc (var. Marmande, 2.5 cm in diameter). The tomato leaf disc was cut into two equal parts. Each tomato halfdisc was placed adjacently to a halfleaf disc of mint (same dimension as the tomato halfdisc) containing two females of T. (A.) recki. The mint leaf discs were checked to avoid prey presence. The number of T. evansi (females, eggs and immatures) and the number of Phytoseiidae (females, eggs and immatures) on each halfdisc were daily counted during 7 days and 15 replications were performed. Student ttests were carried out to compare the number of T. evansi and T. (A.) recki on the two half discs through the time. A Chi2 test was carried out to compare the percentages of females, eggs and eggs+immatures on each halfdisc, through the time. The analyses were performed using Statistica (Statsoft France 2010).

Results and discussion
Distribution of Typhlodromus (Anthoseius) recki Typhlodromus (A.) recki is reported from 22 countries, mainly in the Southern WestPalearctic region ( Figure 2) under Mediterranean climate. It is reported 421 times on 213 plant species, belonging to 138 genera and 51 families. It is recorded less than 5 times on 37 plant families, and the highest record rates were observed on six plant families (56% of reports): Asteraceae (61 reports), Lamiaceae (50 reports), Rosaceae (40 reports), Boraginaceae (36 reports), Fabaceae (26 reports) and Solanaceae (25 reports). This might suggest close relationships between this species and those plants. However, because of no close phylogenetic relationship among these plant families, such occurrence seems not related to evolutionary aspects, as also stated by Tixier (2018) and Tixier et al. (2020). Typhlodromus (A.) recki is reported less than 5 times on 123 plant genera. The highest number of reports (19% of total reports) is observed on species of the genera Solanum (19 reports), Echium (17 reports), Vitis (14 reports), Prunus (12 reports) and Salvia (11 reports). Again, these genera are not phylogenetically close, and T. (A.) recki occurrence seems not related to plant support evolution. Those data provide elements on the "probability" to retrieve T. (A.) recki on some plant species, although we are aware that the recorded occurrences do not exactly reflect the real distribution.

Tetranychus urticae and bean
The average number of eggs laid over 4 days by females of T. (A.) recki, N. cucumeris and N. californicus was significantly different (P = 0.0008) ( Table 2). This was due to the significantly higher fecundity of N. californicus (1.4 eggs/ day/ female) as no difference was observed between the three populations of T. (A.) recki and N. cucumeris. Those differences were not significant at days 1 (P=0.07) and 2 (P=0.06) but were significant at days 3 (P<0.001) and 4 (P=0.001) (Table 4, Figure 3a). For days 3 and 4, the fecundity of N. californicus was significantly higher than that of N. cucumeris and T. (A.) recki. Differences in fecundity through the time were only observed for N. californicus, a higher fecundity being observed at days 3 and 4 than at days 1 and 2 (Table 4). Even if not significant, we can note that among the T. (A.) recki populations, the highest daily fecundity was observed for the population collected on Datura stramonium L. (0.95 eggs/ day/ female) and the lowest for the population collected on S. lycopersicum L. (0.57 eggs/ day/ female) ( Table 2).
The number of T. urticae eggs consumed by the populations/species was significantly different (P = 0.02) ( Table 3) eggs consumed). At day 3 and 4, the highest values were observed for N. californicus (Table  4). Egg consumption significantly decreased from day 1 to day 4, except for N. californicus for which they increased (Table 4). The mean number of eggs consumed during 4 days was the highest (

Tetranychus evansi on Solanum nigrum
The number of eggs laid by the three populations of T. (A.) recki, N. cucumeris and N. californicus was significantly different (P = 0.016) ( Table 2). The highest fecundity was observed for T. (A.) recki from D. stramonium L. (0.3 eggs / female/ day) and the lowest for    (Table 4). For all species, the daily consumption significantly decreased through the time (Table 4). The cumulated number of eggs consumed during four days ranged from 6.4 (N. californicus) to 18.1 (T. [A] recki from D. stramonium) (Figure 3d). These results, as in Escudero and Ferragut (2005), Moraes & McMurtry (1985) and Koller et al. (2007), show the low mean fecundity of N. californicus when fed on T. evansi, even lower than the values obtained in these publications (0.79 eggs/ female/ day, 0.5 eggs/ female/ day and 0.56 eggs/female/day, respectively). The low mean fecundity observed for N. cucumeris and the three populations of T. (A.) recki are in accordance with the results obtained for eight Phytoseiidae species by Moraes and McMurtry (1985). For T. evansi egg consumption, data in literature are scarce.  reported a lower mean consumption by Phytoseiulus persimilis AthiasHenriot (1.4 1.7 eggs/ female/day) than the one here obtained for N. cucumeris and T. (A.) recki (3.11 4.52 eggs/female/day).

Aculops lycopersici on Solanum nigrum
The mean fecundity of the populations / species was not significantly different (P = 0.055) ( Considering each predator separately, T. evansi was the less favourable food source. However, no distinction between the populations/species was observed when the prey was T. urticae or A. lycopersici. Only, a higher fecundity was observed for N. californicus fed on T. urticae than on A. lycopersici (Table 2).

Dispersal abilities of Typhlodromus (Anthoseius) recki along tomato stem in laboratory studies
Females dispersed along the stem bridge whatever the food source (pollen or A. lycopersici). A low proportion was observed on the stem. They essentially moved during the first day ( Figure  4). Afterwards, their movement was less frequent. The females were preferentially found on the discs with food, both for pollen and A. lycopersici treatments ( Table 5). The number of eggs Table 4 Mean fecundity and mean number of eggs consumed each day (± standard error) for the populations T. (A.) recki (D: collected on Datura stramonium, S: collected on Solanum lycopersicum, M: collected on Mentha suaveolens), N. californicus and N. cucumeris when the females are feeding on eggs of T. urticae (on bean leaf discs), of T. evansi (on S. nigrum leaf discs) and all stages of A. lycopersicum (on S. nigrum leaf discs).  laid on the leaf disc with food was significantly higher than those laid on the leaf disc without food (P < 0.0001 pollen disc : 1.50 eggs/female/day, disc without pollen: 0.25 eggs/ female/ day; P = 0.03 disc with A. lycopersici : 1.09 eggs/ female/ day, disc without A. lycopersici: 0.40 eggs/ female/ day). No significant difference in fecundity was observed for leaf discs with A. lycopersici and pollen (P = 0.13). For a same replicate, eggs were sometimes observed on both leaf discs suggesting that the female was able to recross the stem bridge. Even if those results are encouraging, cut tomato stem have different physical and chemical properties than notcut ones. Thus, further experiments on entire plants are needed to conclude on the ability T.  (Table 6). Overall, the number of females and the percentage of occupation of each halfdisc (mint vs. tomato) were not significantly different. However, at days 3 and 6, the number of females of the tomato halfdisc was higher than on the mint halfdisc and the percentage of females on tomato halfdisc was higher than on mint at days 2,   The percentage of eggs+immatures was significantly higher on mint halfdiscs at day 1 and 2, whereas it was significantly higher on tomato halfdiscs at day 7 ( Figure 5). The number of T. evansi females decreased significantly through time (from 5 females / leaf disc at day 0 to 1.72 at day 7). Several hypotheses can explain this observation: (i) female escaped (but this is unlikely as they would be observed on moistened cotton), (ii) female were predated even if predation usually occur on eggs and immature stages and/or (iii) female died (but cadavers were not retrieved). The number of eggs and immatures of T. evansi increased until the day 3 and then decreased but no significant difference was observed through the time (P=0.32) ( Table 6).

Conclusion
This study presents first data on the biology of T. (A.) recki. It confirms that this species is a generalist predator, with biological features close to those of other efficient generalist predators as K. aberrans and T. (T.) pyri. Its mean consumption of T. urticae is similar to N. californicus, which is considered as an efficient predator of spider mites (i.e. Knapp et al. 2018). Furthermore, it seems to perform better on T. evansi than the generalist species already tested. It is similarly attracted by prey and pollen, confirming its generalist statute. It is the first time that this species is tested on A. lycopersici, one of the most important pests in tomato crops (TrottinCaudal et al. 2003; Duso et al. 2009). These preliminary results are encouraging considering both biological features and dispersal abilities. However, as laboratory observations do not entirely reflect real situations, trials on entire plants are required. Furthermore, because of the biological characteristics of this generalist predator, such as for K. aberrans and T. (T.) pyri, biodiversity conservation biological control should be favoured, in line with the banker plant strategy already tried in tomato greenhouses to control white flies (i.e. Frank 2010). Several characteristics of T. (A.) recki (endemic species of Europe, current distribution, its occurrence on Solanaceae and the fact that it is attracted by prey) constitute elements to further investigate its predation ability at plant and crop level.