Effectiveness of inundative releases of Anthocoris nemoralis (Hemiptera: Anthocoridae) in controlling the olive psyllid Euphyllura olivina (Hemiptera: Psyllidae)

This study investigated the effectiveness of fi eld releases of nymphs of Anthocoris nemoralis (F.) (Hemiptera: Anthocoridae) in controlling the olive psyllid, Euphyllura olivina Costa (Hemiptera: Psyllidae). Field trials were conducted in two successive years (2014 and 2015) in two organic olive orchards located in the region of Sfax (Tunisia) using two treatments: low (release of 10 A. nemoralis nymphs/tree) and high (release of 40 A. nemoralis nymphs/tree) applied two times, the fi rst on March 3 and second on March 17. In both olive orchards, the high treatment was the most effective in controlling the increase of E. olivina in the spring. The A. nemoralis population grew gradually and reached a single peak towards the end of April. In control and low treatment plots, despite the signifi cant increase in predator populations, psylla abundance was not controlled. After the second release, however, in high treatment plots a reduction in psyllid density was recorded. An expected effect of the A. nemoralis releases was a reduced parasitic activity of Psyllaephagus euphyllurae (Masi) (Hymenoptera: Encyrtidae).


INTRODUCTION
In Tunisia, olive (Olea europaea L. subsp. europaea) is one of the most important agricultural crops. Olive plantations of about 66 million trees cover about one third of the arable land (ca. 1.7 million hectares). In several arid areas, olives are the only crop and are of great economic and social importance . Olives are attacked by several insect pests (Jarraya, 2003). Among these pests, the olive psyllid, Euphyllura olivina Costa (Hemiptera: Psyllidae) is the most serious (Ksantini, 2003). During spring, high psyllid populations can affect the vegetative development, fertility and reproduction of olive trees (Prophetou & Tzanakis, 1976;Arambourg, 1984;Jardak et al., 1985;Saeb et al., 2001;Ksantini et al., 2002;Jardak et al., 2004;Tzanakakis, 2006), causing economic losses of up to 60% of total production in some Mediterranean Basin countries (Jardak et al., 1985;Tzanakakis, 2006). Outbreaks of olive psyllid populations are associated with new shoot production (Pereira et al., 2013), weather conditions (Chermiti, 1989) or decrease in natural enemy populations due to the use of insecticides (Chermiti, 1992;Ksantini et al., 2002). Deltamethrin and dimethoate are the most frequently used to control the olive psyllid, olive fruit fl y and olive moth (Ksantini, 2003).
The aim of the fi eld trials was to assess the effect of the immature stages of A. nemoralis (mixture of nymphs from 2 nd to 4 th instar) on the abundance of E. olivina. As done previously with ladybird beetles (Iperti, 1999) we released anthocorid nymphs instead of anthocorid adults in order to reduce the risk of emigration from release sites.
The immature stages are active but unable to fl y, and are unlike adults that have tendency to leave the olive trees on which they were released and move to other trees. The time between the two releases was two weeks, which is the average time required for the second instar of A. nemoralis to reach adulthood (12 to 15 days at 22°C) (Gharbi et al., 2011).
In order to simulate an augmentative release of predators, two treatments were used: low and high, the fi rst consisting of two releases of 10 A. nemoralis nymphs per tree and second of two releases of 40 A. nemoralis nymphs per tree. Control plots where no predators were released allowed us to follow the natural development in the numbers of E. olivina.
In 2014 and 2015, there were two releases of anthocorid nymphs in the orchard plots the fi rst on March 3 and second on March 17. Sigsgaard et al. (2006) and Yanik & Unlu (2015) suggest that more than one release is necessary in order to maintain high population levels of predators in orchards.
Small Plexiglas boxes (15 × 10 cm, and 3 cm high), without tops, were used for releases. These boxes were placed 1.5 m above the ground in the foliage of each tree, so that the nymphs could freely disperse in all directions. These boxes were removed before the second release two weeks later, which like the fi rst released either 0, 20 or 80 nymphs per tree. Each tree received the same treatment in both years. One of the objectives of this study is to determine if any of A. nemoralis individuals released in the fi rst year survived and reinforced those released in the second year or dispersed following the decrease in the pest population in the fi rst year.
The design of the experiments was a Latin square. In each orchard there were nine plots, with three plots per treatment and sixteen trees in each plot. The distance between different plots was 14 m and 24 m in the Taous and Chaal orchards, respectively.

Field sampling
In 2014 and 2015, E. olivina and A. nemoralis populations were monitored weekly from the beginning of February until the end of May. To estimate population levels of A. nemoralis and E. olivina, all trees were sampled within a plot, fi ve twigs (10-15 cm in length) were randomly collected from each tree. These were taken to the laboratory in cool boxes and immediately examined Ksantini et al., 2002;Jardak et al., 2004;Tzanakakis, 2006). Non-insecticide alternatives are needed.
The most abundant predator of psyllids in Tunisian olive orchards is Anthocoris nemoralis (F.) (Hemiptera: Anthocoridae), which makes up 49% of the total number of the natural enemies (Gharbi et al., 2012). Considered to be a specifi c predator of psyllids (Anderson, 1962;Solomon et al., 2000), this predator is characterized by its ability to detect psyllid infestations by means of volatiles (Drukker et al., 1995;Scutareanu et al., 1997), high search effi ciency (Brunner & Burts, 1975) and strong numerical response (Trapman & Blommers, 1992;Gharbi et al., 2011). There are many reports describing its role in the biological control of E. olivina populations (Arambourg & Chermiti, 1986;Ksantini, 2003). According to Chermiti (1989), in 1985 in the region of Mahdia, A. nemoralis reduced the population of the second generation of E. olivina to such a level that it was not necessary to use pesticides.
Anthocoris nemoralis may provide an alternative to the chemical control of E. olivina as is indicated by the great success of A. nemoralis mass releases in pear and pistachio orchards (Rieux et al., 1994;Faivre-D'Arcier et al., 2001;Sigsgaard et al., 2006;Yanik & Unlu, 2015).
This paper describes results of fi eld trials carried out to quantify the effectiveness of augmentative releases of A. nemoralis nymphs in controlling the abundance of E. olivina. This study also provides some information on the temporal variation in the abundance of the pest after augmentative releases of A. nemoralis. This knowledge on the interaction between these species should help improve the management and result in a better control of the olive psyllid.

Orchards
The areas studied were two olive orchards near Sfax, referred as Taous (35°02´23˝N, 10°28´34˝E) and Chaal (34°40´10˝N, 10°20´23˝E). "Chemlali" was the most prevalent cultivar. Both olive orchards have been managed without using pesticides against pests and diseases, and the soil was ploughed superfi cially with a scarifi er three or four times a year to control weeds. The planting of the olive trees in the Taous orchard was on a 7 m × 7 m grid and the trees were about 10 years old (average height 3.35 m, tree canopy diameter 2.33 m) and artifi cially irrigated. In the Chaal orchard planting depended on natural resources and the trees were about 70 years old (average height 4.83 m, tree canopy under stereomicroscope and the different stages (eggs, nymphs and adults) of E. olivina and A. nemoralis counted.
To evaluate the effect of the releases on the abundance of other natural enemies, we recorded the total number of different species of predators present, Chrysoperla carnea (Stephens) (Neuroptera: Chrysopidae) and syrphids, as well as the mummies of E. olivina parasitized by P. euphyllurae in the periods after the releases.

Data analysis
Psyllid numbers were expressed in terms of the numbers of psyllid nymphs (Chermiti, 1989(Chermiti, , 1992Ksantini, 2003) on twigs collected in terms of the number of all stages/linear length in meters (lm). The same approach was adopted for A. nemoralis.
These variables were analysed using an ANOVA (SPSS Inc., 2012). The fi eld data was analysed as repeated measures, where individual samples from a plot were the random variable. Data were transformed using Log-transformation in order to normalize the distribution and homogenize the variance.
The Student t-test was used to compare pairs of treatments. Orchard was treated as a factor.

RESULTS
In early February, in both orchards, the psyllid populations increased rapidly and reached very high densities, with the fi rst peak around the middle of March. The highest peak was recorded in the control plots in the Taous orchard on March 17, 2015 with a value of 208.41 individuals/lm (Fig. 1B). Then, the population of the pest declined slightly towards the end of March with the minimum recorded between March 31 and April 7. Subsequently, it increased with a second smaller peak between April 14 and April 28.
After the second peak the pest population declined, which was associated with a change in climatic conditions (increase in temperature and decrease in humidity) and lignifi cation of the twigs.
Before the releases of A. nemoralis, the linear densities of the olive psyllid E. olivina were 95.44 and 159.88 individuals/lm in the Taous orchard, in 2014 and 2015, respectively ( Fig. 1A and B)   During the fi rst two weeks after the fi rst release, the psylla linear densities continued to increase and were not signifi cantly affected differently in the treatments. One week after the second release an appreciable reduction was recorded in the psylla linear density only in the high treatment, whereas in the control and low treatment plots the densities followed almost similar trends, with slightly lower numbers in the low treatment.
The trends were the same in both the Chaal and Taous orchards, where psyllid linear density in high treatment plots was signifi cantly lower than in low treatment and control plots (Taous 2014: F = 24.265; df = 2,76; P < 0.001; Taous 2015: F = 41.678; df = 2,79; P < 0.001; Chaal 2014: F = 48.332; df = 2,77; P < 0.001 and Chaal 2015: F = 31.527; df = 2,77; P < 0.001). However, there was no difference in the E. olivina linear densities recorded in the control and in low treatment plots. Late in the season there was a marked decline in psyllid densities in all the plots (Fig. 1A and D).
In two orchards, A. nemoralis was almost absent until the beginning of March, even in the second year of the experiment. In the control plots, they gradually increased and peaked in abundance towards the end of April, which is when as usual the pest population began to decline.
Before the fi rst release, A. nemoralis was present at low densities of between 0 and 1 individuals/lm. Then, they gradually increased in abundance as the psylla population increased throughout the season.
After the releases, the same scenario was recorded in all orchards, where there was a rapid increase in predator den- sity in high treatment plots, which peaked around March 31 or April 7. In the low and control treatment plots, predator populations increased slowly and peaked two to four weeks later.
Despite the signifi cant increase in predator densities it was insuffi cient to control psylla numbers below damaging levels in control and low treatment plots ( Fig. 2A-D).
In 2015, the psyllid population developed in a similar way as in 2014 with the A. nemoralis population beginning from zero. Hence the releases made during the year 2014 have no effect on the predator population the following year. Table 1 shows the total numbers of the most abundant of E. olivina natural enemies. It is noteworthy that the total number of A. nemoralis collected was highest in the plots in which there were releases. The increase in abundance of A. nemoralis had no effect on the numbers of lacewings and syrphids. Their increase in abundance, however, was associated with a reduction in mummifi ed larvae of psyllids and subsequently of the parasitoid P. euphyllurae.
The decrease in the number of psyllid nymphs parasitized by P. euphyllurae with increase in the number of A. nemoralis, seems to indicate an inverse relationship between these two natural enemies. The presence of a large number of generalist predators appears to have adversely affected parasitoid fi tness and the target herbivore populations. When A. nemoralis was present along with P. euphyllurae there was a great reduction in the cumulative number of hosts parasitized.

DISCUSSION
In both of the years studied the E. olivina infestations were higher at Taous than Chaal. It is likely that the climatic conditions (near the sea) and irrigation at Taous were more favourable for the psyllid, which confi rms the results of Chermiti (1989) and Ksantini (2003).
Every year, there were two peaks in the populations of psyllids: the fi rst and more important, peak was in the middle of March and the second peak at the end of April.
Yayla (1983) and Gharbi et al. (2012) report that A. nemoralis is the most voracious predator of E. olivina. At the beginning of March, this predator was very rare on olive trees and only started to increase gradually from mid-March, peaking at the end of April, slightly later than that of E. olivina. The population of A. nemoralis is the major part of the overall predator population found on olive tree in the spring (Gharbi et al., 2012).
Anthocoris nemoralis arrives late and in such low numbers that it is unable to control olive psyllid populations before they damage the crop (Chermiti, 1989(Chermiti, , 1992. This was also the case in the second year of the study. The failure of A. nemoralis to become established is attributed, partly to its annual migration to orchards from hedgerows, where it builds up its population in spring by feeding on psyllids and other arthropods (Scutareanu et al., 1999). Nevertheless, because of their late migration to crops, their numerical response is often insuffi cient to prevent damage by psyllids (Sarasua et al., 1994;Vilajeliu et al., 1998;Scutareanu et al., 1999;Erler, 2004).
Reports of releasing anthocorids in fi elds to control psyllids are very scarce. In Europe, Anthocoris has been successfully used against pear psyllids (Sigsgaard et al., 2006). In olive orchards, this study showed that the release of a low number of A. nemoralis (20 nymphs/tree) was insuffi cient to control the olive psyllid, but release of a high number (80 nymphs/tree) was followed by an increase in the number of A. nemoralis offspring and a signifi cant suppression of the target herbivore population.
Field releases of anthocorids is an effi cient method of controlling the spring build-up of the E. olivina population and keeping it at an acceptable level. Early releases increased the overall effect of the natural enemies at a crucial time (fl owering period), which is a pre-requisite for successful biological control in olive orchards.
The effect of anthocorids was reinforced by the activity of the other natural enemies present in the orchards. Every year, P. euphyllurae was present especially between mid April and mid May after which it markedly decreased in abundance from beginning of May when no psyllid nymphs are present (Gharbi et al., 2012). The number of A. nemoralis was highest in treated plots, whereas the number of psylla nymphs parasitized by P. euphyllurae was highest in control plots. This inverse relationship between A. nemoralis and P. euphyllurae could be due to the fact that A. nemoralis feeds on nymphs parasitized by P. euphyllurae, which indicates that generalist predators can reduce the effectiveness of parasitoids, as previously reported by Yanik & Unlu (2015). This interaction between a generalist predator and a specialist parasitoid is an example of a non-additive effect of natural enemies on psyllid populations and is similar to such interactions reported in earlier studies (Colfer & Rosenheim, 1995Ferguson & Stiling, 1996;Lucas et al., 1998;Snyder & Ives, 2001;Erbilgin et al., 2004). It is likely that the rate of encounter of A. nemoralis with parasitized hosts is very high because they are very mobile (Agarwala et a l., 2003). In fact, the high abundance of A. nemoralis on the psyllid-infested olive trees may indicate that the competitive ability of A. nemoralis in this guild is very high. Rapid maturation of adults, short pre-oviposition period and nymphal development time, and non-discriminatory mating behaviour (Anderson, 196 2;Horton et al ., 2000) ensure that A. nemoralis can successfully compete with P. euphyllurae for the same resource.
Augmentative releases of the immature stage of A. nemoralis in orchards resulted in signifi cant reductions in psyllid densities. The release of high numbers of adult A. nemoralis is currently widely used and can result in good control. But their effect depends on proper timing (availability of prey) and no application of insecticides according to Faivre-D'Arcier et al. (2001). The release of immature stages rather than adults is more effective because nymph dispersal is limited and a quicker effect can be achieved than would result from using anthocorid eggs (Sigsgaard et al., 2006). Indeed, nymphs are preferred over eggs' (Rieux et al., 1994) because the high temperatures in South Eastern part of Tunisia can desiccate anthocorid eggs.
Overall, our study presents a basis for future studies on biological control in olive groves. Further studies on interand intra-guild interactions should be carried out before deciding to release generalist predators to control psyllid infestations in olive orchards. In addition, factors affecting tree health and yield of olive trees should be elucidated