Torymus sinensis Kamijo, a biocontrol agent against the invasive chestnut gall wasp Dryocosmus kuriphilus Yasumatsu in Spain: its natural dispersal from France and first data on establishment after experimental releases

Aim of study: The globally invasive gall wasp, Dryocosmus kuriphilus Yasumatsu, 1951 (Cynipidae: Cynipini), reached Spain seven years ago and is already regarded as an important pest of chestnuts (Castanea spp.) in this country as well as worldwide. In this paper, we present comprehensive data on the establishment in Spain of Torymus sinensis Kamijo, 1982 (Chalcidoidea: Torymidae), an effective non-native natural enemy of this pest, as a result of both natural dispersal and settlement after controlled releases since 2015. Area of study: Sites of the Spanish autonomous communities of Galicia, Asturias, Basque Country, Catalonia, Andalusia and Madrid where D. kuriphilus is present. Material and methods: To study the natural dispersal of T. sinensis from France, we selected two sampling sites in Catalonia, six in the Basque Country and two in Navarra known for their heavy Asian chestnut gall wasp (ACGW) infestation; to study T. sinensis establishment after authorized controlled releases by the concerned authorities, the field samplings were done mainly in Galicia (35 sites) and Andalusia (8 sites). Additionally an experimental release study was made in Madrid. Main Results: Our results showed that T. sinensis has spread throughout Spain by natural dispersal across the French border and now occurs in Catalonia (two sites), the Basque Country (three sites) and Navarra (one site) but not in the neighbouring region of Cantabria. The percentage of parasitism by T. sinensis on D. kuriphilus is higher in the Basque Country sites, which are close to the French border, thus indicating that its establishment in these localities is not recent. After controlled releases, T. sinensis has been successfully established in five release sites in Andalusia (Valle del Genal and Sierra Blanca, Málaga Province) and one release site in Madrid. However, in the region of Galicia (NW Spain), where the number of authorized releases has been higher, the establishment of T. sinensis still appears to be very low. Research highlights: Established populations of T. sinensis may exert a positive buffer against D. kuriphilus-driven chestnut infestation in Spain, similar to what is observed in other invaded European countries. Additional keywords: Controlled releases, Torymidae, invasive species, Cynipidae, natural spread, biological control. Abbreviations used: Asian chestnut gall wasp (ACGW); Torymus sinensis parasitism rate (TsPR); mean of ACGW larval chambers per gall (MACGW). Authors ́ contributions: Conceived and designed the work: JLNA, DGT, JRB, MJL. Performed the experiments and analyzed the data: JLNA, DGT, JRB, ONG, MJL, MEW. Contributed materials/analysis tools: JLNA, DGT, ONG, JRB, MJL, DB, CRC, MPRR, JMV, MEW. Wrote the paper: JLNA, DGT, MJL, CP. Citation: Nieves-Aldrey, J-L., Gil-Tapetado, D., Gavira, O.N., Boyero, J.R., Polidori, C., Lombardero, M.J., Blanco, D., Rey del Castillo, C., Rodríguez-Rojo, M.P., Vela, J.M., and Wong, M.E. (2019). Torymus sinensis Kamijo, a biocontrol agent against the invasive chestnut gall wasp Dryocosmus kuriphilus Yasumatsu in Spain: its natural dispersal from France and first data on establishment after experimental releases. Forest Systems, Volume 28, Issue 1, e001. https://doi.org/10.5424/fs/2019281-14361 Received: 03 Dec 2018. Accepted: 05 Mar 2019. Copyright © 2019 INIA. This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 International (CC-by 4.0) License. Funding: This study was financed by an Encomienda de Gestión from MAPAMA to Agencia Estatal CSIC, 16MNES003 awarded to JLNA, DGT, MJL and CP, and by project AGL2016-76262-R (AEI/FEDER, UE) awarded to MJL and JLNA. JRB, JMV, MEW and OG were supported by project PP.PEI.IDF201601.4 de Demanda Institucional from CMAOT to the public research Institution IFAPA. CP was funded by a post-doctoral contract from the Universidad de Castilla-La Mancha. Competing interests: The authors have declared that no competing interests exist. Correspondence should be addressed to José Luis Nieves-Aldrey: aldrey@mncn.csic.es Forest Systems 28 (1), e001, 11 pages (2019) eISSN: 2171-9845 https://doi.org/10.5424/fs/2019281-14361 Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria O. A., M. P. (INIA) José-Luis Nieves-Aldrey, Diego Gil-Tapetado, Oscar N. Gavira, Juan R. Boyero, Carlo Polidori, et al. Forest Systems April 2019 • Volume 28 • Issue 1 • e001 2

ACGW is one of the few oak gall wasps that are harmful to their Fagaceae host plants and thus have economic importance (Kinsey, 1935;Bailey & Stange, 1966;Scutareanu & Roques, 1993;Baldassari & Baronio, 1996;Nieves-Aldrey, 2001). The peculiar life cycle of ACGW (univoltine, with only parthenogenetic females known, whose small first instar larvae are present in overwintering chestnut buds and are thus visually hardly detectable) has undoubtedly contributed to the rapid expansion and settlement of this exotic invasive species, helped by the human-assisted dispersal of infested chestnut from nurseries (Hulme, 2009). The negative effects of ACGW on chestnuts include a reduction in the number of fruits, malformation of the branches and general weakening of the tree (Battisti et al., 2014;Gehring et al., 2018;Avtzis et al., 2019), with consequent chestnut production losses and considerable economic damage to the chestnut tree sector (Brussino et al., 2002;Zhang, 2009;EFSA, 2010).
After the first report of its presence as a pest in 1941 in Japan (Murakami et al., 1980), ACGW became established in several countries, from Korea and United States to Nepal and Canada (Cho & Lee, 1963;Payne et al., 1975;Abe et al., 2007;Huber & Read, 2012). The species arrived in Europe in 2002 (Brussino et al., 2002) and since then spread rapidly throughout neighbouring countries, from Slovenia in 2005 to the Czech Republic in 2012 (Borowiec et al., 2014) and Romania in 2015(Rádócz et al., 2016. The first published reports of the establishment of ACGW in the Iberian Peninsula (IP) were from Catalonia (northeastern Spain) (DOGC, 2012;Pujade-Villar et al., 2013), although an earlier detection in 2010 is mentioned in Borowiec et al. (2014) (as a personal communication from J. H. Delader), and data on heavy chestnut attacks in the Montseny Natural Park (Barcelona and Girona) in 2011 were reported by Cristina Castro Torres (in Rubio, 2014).
Since this first Spanish introduction, ACGW has spread throughout the Cantabrian coast (north of the IP) and was reported from the Basque Country and Cantabria in 2013 and from Asturias in 2014. In May 2014, the species was detected in Galicia (northwestern Spain) and northwestern Portugal (Pérez-Otero & Mansilla, 2014). The species was also recorded at the same time (2014) in the Andalusian provinces of Málaga and Granada (southern Spain) Wong et al., 2015, while important chestnut areas in the central and western IP now seem to be unaffected (as of May 2018) (Gil-Tapetado et al., 2018).
In territories where ACGW has been introduced and represents an important forestry pest, the wasps' protection inside the galls, in addition to potential environmental impacts, makes chemical control ineffective, and efficient natural enemies at the first stages of invasion are lacking. Thus, classical biological control with a parasitoid of D. kuriphilus in its native geographic area has been widely used in many invaded countries. This biological control agent is Torymus sinensis Kamijo, 1982 (Hymenoptera: Chalcidoidea: Torymidae), a specialist parasitoid with a univoltine life cycle that is synchronised with that of ACGW, which is native to East Asia (Murakami et al., 1980(Murakami et al., , 2001Moriya et al., 1989). It should be noted however, that in some invader areas a prolonged life-cycle diapause of T. sinensis has been reported as well as the species has shown to be able to adapt to the attack of native cynipid species associated to Quercus (Fagaceae) (Ferracini et al., 2015a(Ferracini et al., , 2017. Biological control assays with this species were first set up in Japan (Moriya et al., 1989(Moriya et al., , 2003 and continued in the United States (Cooper & Rieske, 2007) and Europe, where it has been introduced in Italy ( Quacchia et al., 2008;Gibbs et al., 2011), France (Borowiec et al., 2014(Borowiec et al., , 2018 and Croatia (Matošević et al., 2014), among other countries (Avtzis et al., 2019). The results of the introduction and effectiveness of ACGW biocontrol were mostly positive (Moriya et al., 2003;Quacchia et al., 2008;Matošević et al., 2014Matošević et al., , 2017Paparella et al., 2016), so that the demand to introduce T. sinensis into new areas is growing among the affected chestnut communities (orchards, timber production and natural forests management).
However, the introduction of T. sinensis into nonnative areas may also represent a possible threat to the native fauna (Gibbs et al., 2011;, and studies on balancing the benefits and risks have been carried out (e.g., in Switzerland: Aebi et al., 2011). Accordingly, the two main threats associated with T. sinensis introduction are attacks on non-target cynipid species (Italy: Quacchia et al., 2008;Ferracini et al., 2015bFerracini et al., , 2017 and the risk of hybridization and ecological competition with native torymid species (Japan: Moriya et al., 1989Moriya et al., , 2003Yara et al., 2007;Italia: Ferracini et al., 2018;Pogolotti et al., 2018). These two risks are currently being evaluated also in Spain by authors of this paper and will be published elsewhere.
After the introduction and spread of D. kuriphilus into the main chestnut production areas of Spain, concerned authorities in the regions (autonomous communities) of Galicia and Andalusia began controlled release assays of T. sinensis, authorised by the Spanish Ministry of Agriculture and Fisheries and Food (MAPA). Releases of T. sinensis in the affected areas of the Province of Málaga were authorized in 2015, 2016 and 2017, and in the same years, releases were undertaken in large areas of the Autonomous Community of Galicia (Pérez-Otero et al., 2017) and also in Asturias (north of the IP). However, the presence of T. sinensis in France since 2011 posed the possibility of natural dispersal of the species into Spain across the French border. The dispersal ability of T. sinensis has been discussed by Colombari & Battisti (2016) for the case of Italy suggesting a dispersal rate up to 70 km per year. In the case of Japan, the first country in which T. sinensis was introduced as an ACGW control, the estimation of dispersal rates was of less than 1 km/year during the first few years, followed for more rapid and gradual spread in the next years (equal to 12 km) and finally reaching a constant rate of ca 60 km per year (Moriya et al., 1989(Moriya et al., , 2003. Among the transboundary dispersal events of T. sinensis that have been published, we can mention the passage of the species from Italy to Switzerland , from Italy or Switzerland to France (Borowiec et al., 2014), and from Croatia to Bosnia (Matošević et al., 2014). In the case of Spain, we recently found T. sinensis in Catalonia (NE of the IP) (unpublished data) and it has also been independently detected by other authors (Jara-Chiquito & Pujade-Villar, 2018). Given that no T. sinensis releases were authorised in Catalonia, its presence in the region is more likely due to dispersal from France. It is thus necessary to see if this parasitoid also occurs in Navarra and the Basque Country (both next to France), where no releases of T. sinensis have occurred.
The aim of this paper is twofold: (i) first, we report the dispersal of T. sinensis from France into Spain throughout the Basque Country and Navarra and confirm the same path across the eastern border into Catalonia; (ii) second, we provide the first data on the establishment of T. sinensis in Andalusia, Galicia and central Spain after experimental releases.

Study sites
Natural dispersion Sampling and field work were conducted within the framework of three scientific projects focused on ACGW in Spain. The main project, developed at the level of the entire state, had the twofold aim of studying the recruitment of native parasitoids and assessing the environmental risks associated with the introduction of T. sinensis. At the same time, similar projects were conducted in the autonomous communities of Andalusia and Galicia.
The chestnut forests of the northern IP form a "continuum" from SW France along the Cantabrian Coast to Galicia. To study the natural dispersal of T. sinensis from France, we selected two sampling sites in Catalonia, six in the Basque Country and two in Navarra known for their heavy ACGW infestation, at the east and west corners, respectively, of the border with France, where the Pyrenees are not an effective barrier and both form biological corridors (Table 1, Fig. 1). These areas are isolated and very distant from the Spanish territories where the T. sinensis 2015-2017 releases were performed. An additional locality was also sampled in the community of Cantabria, close to the Basque Country, but further away (approximately 200 km) from the French border and at about the same distance from Asturias, where controlled releases were carried out in 2017 (Table 1, Fig. 1).

Establishment after authorized controlled releases
To study T. sinensis establishment after controlled releases by the concerned authorities, the field samplings were done mainly in Galicia (NW of the IP) and Andalusia, the Spanish regions were the first experimental releases of T. sinensis were authorized and performed. In Andalusia, 8 sites were repeatedly sampled from 2016 to 2018 while in Galicia we sampled 35 sites in 2018. In addition, 2 sites of the Autonomous Community of Asturias were also sampled for this purpose (Table 1).

Experimental release in Madrid
An experimental release of T. sinensis was performed in 2017 on infested chestnuts of the Royal Botani cal Garden Alfonso XIII of the Universidad Complutense de Madrid (Madrid, Central Spain).

Field samplings: collecting galls and rearing of parasitoids
With the aim of rearing T. sinensis (as well as native parasitoids with a similar life-cycle), whose emergence occurs from dry galls after winter, ACGW galls were preferentially collected during the winter and then transferred to indoor laboratory conditions until emergence of the insects. This usually occurred one or two months before the natural outdoor emergence because of the forced conditions of the laboratory. Collected galls were located into cardboard emergence boxes equipped with light extractors and kept in normal indoor laboratory conditions. For an accurate estimation of the real emergence date of T. sinensis in the field, an additional experiment was performed in Juanar (Málaga, Andalusia). Three infested chestnuts were selected, and the galled branches bagged. A total of 30 bags were placed (10 per tree), encaging an average of 8, 6 and 7 galls per bag. The emergence of T. sinensis was checked weekly.

Taxonomic identification
A careful morphological examination of the reared specimens attributed to T. sinensis is essential to prevent any possible misidentification with other native torymid parasitoid species. Confusion is particularly possible with three species that are

Estimation of parasitism
All the galls collected during the study together with their collection dates, the number of T. sinensis individuals recovered per locality, their emergence dates, and parasitism percentages (TsPR) appear in Table 1. To estimate parasitism rate of ACGW by T. sinensis we used the following equation where N Tsinensis is the number of T. sinensis indivi duals per sample; N galls is the number of ACGW galls per sample, and MACGW is the mean of ACGW larval chambers per gall in each locality (see Table 1). This num ber was calculated after dissecting at least ten galls for each sampling locality in Andalusia and Galicia. For the samples collected in the Basque Country, Navarra, Madrid and Catalonia, where galls were not dissected, 5.2 chambers per gall was used as an estimated mean number for the remaining Spanish localities (Table 1).

Navarra, Basque Country and Cantabria (northern IP)
From the samples collected in Navarra and the Basque Country, we recovered 147 females and 73 males of T. sinensis. The species was present in one of the sampled sites from Navarra (Bera de Bidasoa) and in three of the sampling sites in Guipúzcoa Province (Basque Country): Hondarribia, Jaizkíbel and Lezo. Compared with the relatively low figures of parasitism to date, from the Spanish sites where were T. sinensis was artificially released, the parasitism percentages in the Basque Country and Navarra were variable but were generally moderate to high (from 0.99% to 8.29%), the relatively high rates at Lezo (4.88%) and especially at Jaizkíbel (nearly 8.5%) being remarkable. These data unequivocally show the successful establishment of T. sinensis in these two autonomous communities, likely as a result of its natural dispersal from France. All the localities with T. sinensis are very close to the border with France, Lezo being the farthest away (approximately 14 km). In contrast, in the sam ples collected in three sites in Vizcaya Province (Bas que Country) (Orozco, Berango and Urdúliz), which are substantially farther from France (approximately 130 km), T. sinensis was not present. In addition, samples of ACGW galls collected in S. Roque de Riomiera, in the neighbouring Autonomous Community of Cantabria (225 km from France), did not yield any T. sinensis individuals.

Catalonia (northeast IP)
Twelve T. sinensis females were reared from ACGW galls collected during the summer in two sites, Sant Marçal and S. Hilari Sacalm, in the Natural Park of Montseny (Barcelona and Girona, Catalonia). The calculated parasitism rate was 0.96% at the first site and 0.19% at the second (Table 1). The specimens successfully emerged in the laboratory from fresh galls collected eight months before (see Table 1), despite the galls being collected in summer and not subsequently enduring a 'cold' period for larval diapause before pupation.

Establishment of T. sinensis after releases
Andalusia (southern IP) A total of 275 individuals of T. sinensis (154 females and 121 males) were recovered from 12,218 galls collected in eight sampling sites of Málaga Province, where experimental releases of T. sinensis were made in 2015-2017 (Table 1; Fig. 1). The settlement of T. sinensis was especially successful in Juanar, where 230 individuals were recovered, with an estimated mean parasitism rate of 0.51% (range 0.10-0.86, n = 5), despite the fact that experimental releases were made only in 2015 at this site, with 720 individuals (480 females and 240 males) being released. In addition, 34 T. sinensis individuals were recovered from another two sites, Igualeja and Júzcar, in the Valle del Genal, while we also recovered 3 specimens from Puerto de Ojén, which is not too close to Juanar, both being in the Sierra Blanca, and 8 from Yunquera in the Sierra de las Nieves.
To have available accurate data on the phenology of T. sinensis in the areas where it has been introduced is important to determine the best release date for a successful biological control programme with this species. Accordingly, in the experiment performed in 2018 at the site of Juanar, the emergence of T. sinensis under field conditions began in mid-March and ended in mid-April, with an emergence peak in late March (Fig.  2). Therefore, this date may be the most appropriate for releasing T. sinensis in Andalusia region.

Galicia (northweast IP)
Given the extent and economic importance of the chestnut in Galicia, authorized controlled releases of T. sinensis have been important in this region. Up to 220,395 individuals were released in three campaigns (2015-2017) according to Pérez-Otero et al. (2017), and even higher numbers were released in the 2018 campaign (704,270 individuals from data of the Xunta de Galicia). However, only 12 individuals have been recovered according to published data, all from release sites in Ourense Province (Pérez-Otero et al., 2017). Our own data even showed null settlement values at different localities. Indeed, from 4,871 galls collected in the winter of 2017 and early spring of 2018 from 33 sites in Galicia, only one male T. sinensis was recovered at one site (Riós, Ourense) where controlled release began in 2017. In addition, the species was not recovered from galls collected in winter (2017) from several sites of the Ribeira Sacra region (Central Galicia).

Madrid (central IP)
The experimental release of 120 individual T. sinensis on ACGW galls in Madrid returned 25 emerged T. sinensis individuals (13 females and 12 males ). The success of the experimental infection was high, with a resulting parasitism rate of 3.69%.

Natural spread from France
In this work, we report the presence of T. sinensis in northern Spain in three autonomous communities at the eastern and western ends of the border with France: the Basque Country, Navarra and Catalonia. Authorized releases of this biological control agent have been made in recent years in Galicia, Asturias and Andalusia, and thus in geographically distant regions from those mentioned above, but no releases have been made to date in communities close to the border with France. Discounting the possibility that the presence of T. sinensis in these northern autonomous communities is a result of many illegal or uncontrolled releases, the most likely explanation is that T. sinensis has naturally dispersed to these territories from France across the southern border with Spain. Published data of T. sinensis releases in France (Borowiec et al., 2014(Borowiec et al., , 2018 show that the closest release sites to Spain are Gironde, Lot, and Hautes-Pyrénées, 172-237 km away from the northern Spanish sampling sites. Taking into account that T. sinensis was released in France in 2013 and that the estimated dispersal capacity of this parasitoid is approximately 70 km per year (Colombari & Battisti, 2016), we suggest that T. sinensis may have reached the ACGW-infested Spanish areas in 2016. Fast natural dispersal of T. sinensis was observed also in Hungary .
The hypothesis of an early arrival of T. sinensis in northern Spain by natural dispersal from France is reinforced by the relatively high parasitism rates shown in Lezo and Jaizkíbel (4.8% and 8.2%, respectively). It is largely accepted that T. sinensis needs 5-7 years after the first release and or settlement to reach such high rates of parasitism (Quacchia et al., 2008;Borowiec et al., 2014;Matošević et al., 2014Matošević et al., , 2017. However, despite this early arrival in the Basque Country and Navarra, our data indicate that the progression into Spain throughout the Cantabrian coast was slow, given the apparent ab sen ce of T. sinensis in the sampled sites of Vizcaya Province and the neighbouring community of Cantabria. The spread of T. sinensis from France towards Spain takes advantage of the existence of two broad ecological corridors connecting both sides of the eastern and western Pyrenees. As a comparison between the two zones, the dispersal of T. sinensis across the western corridor, from France across the Basque Country and Navarra, would be easier than along the eastern corridor connecting France and Catalonia, since the first route has a lower altitude and a milder climate. With regard to Catalonia, our data confirm the recent publication of the presence of T. sinensis in this autonomous region (Jara-Chiquito & Pujade-Villar, 2018). These authors reported a total of 45 individuals (15 males and 30 females) from three localities in Girona Province, close to the sites sampled by us. However, they do not provide data on the parasitism levels.

Establishment after controlled releases
The release programmes with T. sinensis carried out from 2015 to 2017 in two large areas of Spain, the whole Galicia region and areas of the Andalusian Province of Málaga yielded different results about its settlement. Whereas in Andalusia, the first successful results came after three years of experimental releases (275 T. sinensis were recovered out of 41,420 released, present results), in Galicia (where releases were even more numerous), only 12 out of 220,395 released individuals were recovered in post-release controls   Pérez-Otero et al., 2017), and only one in our study. Up to now, all available data have shown a low rate of T. sinensis implantation in this autonomous community. It is expected however, given the massive releases of T. sinensis made by the Xunta of Galicia in 2018 that the situation could significantly change after update the recovery data of 2019. We had not still access to the unpublished data of recoveries from the Xunta of Galicia, but our own data have started to show emergences of T. sinensis from galls collected in additional samplings in 2018-2019 (3 females and 2 males recovered to date (February 2019) from galls collected at the site of Merouzo pequeno (Ourense)).
The settlement seems to be particularly successful in Juanar, the first site in Málaga Province where T. sinensis was experimentally released in 2015. Despite the lack of further releases in the area since then, more than 200 individuals were recovered from test-release controls in 2018, with an estimated parasitism rate of 0.51%, thus suggesting successful establishment. It is important emphasize however, that the release area in Galicia is much wider than the release area in Andalusia, and the relative sampling effort was much lower in the first one, which may be one of the reasons of the low establishment success in that region.
It is interesting to compare the parasitism rate recorded from the different sampling sites. Values were null or low (≈1%) at the release points (Table 1), while they were high at sites in the Basque Country (8.2%), i.e., where T. sinensis populations occur because of natural spread from France. The reasons behind this difference is difficult to explain using collected data. It seems not probable that settlement success depends on the time from the releases, since these were done starting back in 2015, possibly even before the arrival of the parasitoid in the northern regions by natural dispersal from France. On the other hand, we cannot exclude an effect of some unknown environmental and climatic conditions (which are extremely variable across the IP, and particularly between the northern and southern coastal areas) on the settlement success (Gil-Tapetado et al., 2018). Additional studies are necessary to understand whether the release of T. sinensis is more likely to fail under certain conditions and thus possibly not be useful against ACGW.
The data from the experimental release in Madrid showed successful establishment of the insect in the field, on ACGW galls, which is a promising starting point towards future controlled releases in this and nearby regions where C. sativa is present but infection with ACGW have been not detected yet.
We predict that T. sinensis will continue its spread over the northern part of the IP, following C. sativa forests infested by ACGW, possibly reaching in a few years many more areas than where it actually occurs today.
Established populations of T. sinensis may exert a positive buffer against D. kuriphilus-driven chestnut infestation in Spain, similar to what is observed in other invaded European countries.