Biological control is defined as the activity of parasitoids, predators, or diseases in maintaining the other organism’s population density at a lower average than would occur in their absence (DeBach, 1964). Classical biological control involves the introduction of a biocontrol agent to a new habitat to control pests that have spread beyond their original habitat. According to (Caltagirone, 1981), classical biological control is the regulation of exotic pest populations (insects, mites, mammals, weeds, pathogens) by exotic natural enemies due to the lack of local natural enemies that can suppress the exotic pests, thus creating a favorable situation for invasive species to spread. It works on the premise that old associations between the host (pests) and its biocontrol agent have already been established (co-evolved), and thus should be an effective control agent (Hokkanen & Pimentel, 1989)

The most famous classical biological control was the introduction of the vedalia beetle predator Rodolia cardinalis (Coleoptera: Coccinellidae) to control the cottony cushion scale Icerya purchasi (Hemiptera: Margarodidae) in California(Source Title, 1978)(Caltagirone & Doutt, 1989) The introduction of R. cardinalis is regarded as the beginning of classical biological control. The use of biological agents for weed management started in 1795 when the cochineal insect Dactylopius ceylonicus from Brazil was introduced to India to control the invasive weed Opuntia monacantha (Willd.) Haw. (Cactaceae) (Winston et al., 2014) By 2012, a recorded number of 1,555 cases of biological control of weeds were reported in 90 countries in which 468 biocontrol agents were introduced to manage 175 species of weeds in 48 plant groups (Schwarzlander et al., 2018) However, only 982 (63.2%) introductions demonstrated stability, with a total of 332 (70.9%) biological agents used (Schwarzlander et al., 2018)

In Indonesia, Cecidochares connexa (M.) (Diptera: Tephritidae) is a gall fly that has been used as a biocontrol agent to suppress the population of the invasive species Chromolaena odorata (L) R.M. King and H. ob (Asteraceae). C. odorata is a fast-growing invasive species that entered Indonesia in 1934 at the Lubuk Pakan tobacco plantation in North Sumatra and spread rapidly to other major Indonesian islands(Tjitrosoedirdjo, 2005)(Setyawati et al., 2015)(Padmanaba et al., 2017) C. connexa has been established in several countries, including the Philippines, India, Guam, and Papua New Guinea(Aterrado & Bachiller, 2002)(Bhumannavar & Ramani, 2006)(Reddy et al., 2010)(Day et al., 2013) The success of its establishment was one of the reasons for its introduction to Indonesia. (McFadyen et al., 2003) reported that C. connexa causes galls on the terminal and axillary vegetative meristems of C. odorata, and can decrease seed production by 50% and reduce C. odorata population by up to 37,2% within two years (Tjitrosemito, 1999)

C. connexa was first introduced in Indonesia in 1993 from Colombia, USA (Tjitrosemito, 1999) In 1995-2001, releases were carried out in several areas, including eastern Indonesia, one of which was South Sulawesi. The first release of C. connexa in Sulawesi was conducted in the Bantimurung sub-district (60 km northeast of Makassar), South Sulawesi, in March 1999, with a total of 240 galls. A second release was carried out in February 2000 at two different locations, namely the Bantimurung sub-district (60 km northeast of Makassar) with a total of 300 galls and Camba sub-district (25 km northeast of Bantimurung) with 232 galls (Wilson & Widayanto, 2004) In April 2002, the establishment of C. connexa was evaluated at release sites. C. odorata infested by C. connexa was found approximately 10 km from the release point (Wilson & Widayanto, 2004)

Since 2002, no studies have evaluated the effectiveness C. connexa in Bantimurung, South Sulawesi, nor have any studies been conducted to see the changes in the food web interaction due to the introduction. As mentioned by (Pearson & R.M, 2005), biocontrol agents introduced in new areas could indirectly change the interactions in the ecosystem and attack non- targets through food web subsidies or host change. This interaction can change the overall food web and interactions of many species in an ecosystem. One example of research related to food web subsidies is the introduction of Urophora affinis and U. quadrifasciata to control Centaurea maculosa and C. diffusa in western North America (Winston et al., 2014) Their research showed that the two gall flies that were introduced became an additional food source for populations of generalist predators, deer mouse Peromyscus maniculatus(Pearson et al., 2000)(Ortega et al., 2004)(Winston et al., 2014) Another study on ecosystem changes was conducted by (Jaya, 2006)and (Lukvitasari et al., 2021) who showed that C. connexa could be attacked by local parasitoids and predators. Several types of parasitoids have been found to attack C. connexa including those from the Families Ormyridae, Eupelmidae, Bethylidae, Braconidae, Ichneumonidae, Encyrtidae, Eucoilidae, Eulophidae, Eurytomidae, Pteromalidae, and Torymidae(McFadyen et al., 2003)(Safi’i, 2006)(Buchori et al., 2020)(Lukvitasari et al., 2021) These findings show that there is a new association between C. connexa and local parasitoids, which invokes the old hypothesis of (Hokkanen & Pimentel, 1989) regarding new associations that can be formed between biological control agents and local natural enemies. Thus, biological agents introduced in new areas can potentially become new hosts for natural enemies.

The aim of this research is to study community interactions and possible new associations between C. connexa and local parasitoids in Polewali Mandar, West Sulawesi. Does the establishment of the bioagent C. connexa in West Sulawesi cause an array of new interactions with local parasitoid communities? Does the complex of host parasite interactions similar as was found in the Western part of Indonesia? These questions are important since Sulawesi is located on the Eastern side of the Wallace line, thus may have different biodiversity compared to the western part of Indonesia. These are among the questions raised to understand the impact of C connexa released in the field at Polewali Mandar, West Sulawesi.

Our result showed that C. connexa are already established in West Sulawesi and can be found in open fields and cocoa plantations. This study is the first to report that C. connexa are able to adapt and become established in Polewali Mandar, West Sulawesi. This finding is in line with (Tjitrosemito, 2002)which stated that C. connexa has been established on most of the larger Indonesian islands and that C. connexa can spread widely over short periods of time. According to (Harjaka & Mangoendihardjo, 2010) C. connexa can spread over a radius of more than 200 km from the release site in Gunungkidul, Yogyakarta, to several areas in East Java within 10 years. (Day et al., 2013) also reported that C. connexa had spread more than 100 km from most of the release points in Papua New Guinea within seven years. It is likely that C. connexa can also be found in areas more than 250 km from the release point. However, the introduction of C. connexa did not seem to reduce the population of C. odorata, as indicated by its abundant population at the study site. (Kenis et al., 2019) reported that the success of biological control is classically characterized by a decrease of the target population. However, our observation found C. odorata to be abundant in the field, hence raising the question of C. connexa effectiveness in as biological control agent. Field observations showed that the attacked branches do not instantly die but continue to develop and it even create new branches. Thus attack of C. odorata seems to increase the branch production.

Our observation showed that C. connexa could lay more than one egg on each shoot of C. odorata. In addition, we also found more than one C. connexa larvae in one gall that can emerge as imago. (McFadyen et al., 2003) reported that C. connexa could lay up to 16 eggs per shoot on the shoots of C. odorata. However, the number of larvae that can survive to become imago is between 2-4 individuals.

Compared to open fields, cocoa plantations have fewer galls and C. connexa. This is because fewer C. odorata plants may be found in cocoa plantations, which leads to fewer galls being discovered. Pruning and spraying of herbicides routinely on C. odorata around cocoa plantations may affect to the population of C. odorata, resulting in a decrease in the number of galls formed. The presence of shade is also believed to be a factor that slows the growth of C. odorata in addition to pruning and spraying. This has an impact on C. connexa which can parasitize parasitoids.

Overall we found six species of parasitoids associated with C. connexa with Aprostocetus sp 1 being the most abundant . The parasitoid complex is different than the community of parasiotids found in West Java(Buchori et al., 2020) Quadrastichus sp was the most abundant parasitoid found in Bogor and Lampung. The difference may seem be due to the different diversity that are present in the island of Java, Sumatra and Sulawesi. This study however, strengthens the findings of(Buchori et al., 2020) (Lukvitasari et al., 2021) (Jaya, 2006) and (Pahlevi, 2006) that new associations have been formed from exotic bioagents introduced to new habitats. The fact that C. connexa from different islands are attacked by different parasitoid complex has shown that adaptation of local parasitoids to invasive/exotic pests can happen. (McFadyen et al., 2003) also reported the attack of C. connexa by parasitoid insects five years after release and the first report of a new association between C. connexa and a parasitoid. (Boughton et al., 2012) also reported that Neomusotima conspurcatalis Warren (Lepidoptera: Crambidae) was introduced to Florida in 2008 to control Lygodium microphyllum (Schizaeales: Lygodiaceae) and are then attacked by local parasitoids in just a few years. These findinga are interesting because it reveals the local adapatation of local parasitoids with exotic herbivore species, thus strengthening (Hokkanen & Pimentel, 1989) argument to use local parasitoids as bioontrol agent, since new associates can happen quite rapdily in the field. Our findings raise other questions e.g. does the presence of C. connexa change the interaction dynamics of the parasitoid with its original hosts? Are there shifts in the preference of the parasitoid from its original hosts, and if there is a shift, what are the implications for the original hosts? Will there be a natural enemy-free space that results in the explosion of new pests? Further investigation is required to corroborate some of the assertions made above because there is a shortage of knowledge on the subject.

One important question that can be raised from this study is whether C. connexa only functions as an additional host for parasitoids or actually functions as a subsidy that can increase the population of parasitoid insects and can indirectly increase the role of these parasitoids to attack other hosts. The parasitoids found attacking C. connexa in this study are most likely generalist parasitoids that attack a variety of gall-forming insects in several habitats in the tropics. Thus, when C. connexa attacks C. odorata and form galls, the local parasitoids can shift and atttack C. connexa as a new host. Parasitoids of the Genus Aprostocetus were the most common and had the highest parasitization rate. This is different from what was previously reported by (Buchori et al., 2020) and (Lukvitasari et al., 2021)that the dominant parasitoid found was Quadrastichus sp. Aprostocetus parasitoids are likely common parasitoids that commonly attack other insects such as Orseolia javanica (Diptera: Cecidomyiidae) on rice in Bogor, West Java (Maqsalina, 2021) Ophelimus eucalypti (Hymenoptera: Eulophidae) on eucalyptus plants Medan, North Sumatra (Anisa et al., 2023)

Parasitoids are generally considered beneficial for pest control as they parasitize insect pests. In this case, they have a detrimental impact, as they attack the biological control agent of C. connexa which is used to control C. odorata populations. Overall, the parasitoid discovered in the cocoa plantations was also present in the open fields. The parasitoid observed to attack C. connexa in this study is most likely a generalist parasitoid that targets different forming-gall insects in several habitats in tropics. The introduced C. connexa which was established in release areas to control C. odorata weed, is a new host for local parasitoids. According to (Pearson & R.M, 2005) biological agents introduced into new areas to control invasive pests have the potential to become hosts for local natural enemies. If they are unable to control the population of invasive weeds, it results in an abundance of weed populations.

One reason for the ineffectiveness of C. connexa in suppressing C. odorata populations is the presence of parasitoids that parasitize C. connexa. Although the level of parasitization is still relatively low, the presence of parasitoids on C. odorata can reduce attacks from C. connexa. (Abdala-Roberts et al., 2019) stated that one of the indirect effects of the presence of natural enemies is that they can reduce herbivore populations and indirectly affect the plants that are attacked.

Classical biological control using C. connexa to suppress populations of the invasive weed C. odorata has been unsuccessful. Although C. connexa is widespread and established, its presence is insufficient to control C. odorata because it is attacked by a variety of parasitoids in a variety of habitats, including industrial plantations, forests, open fields in Bogor(Jaya, 2006)(Safi’i, 2006) palm oil plantations and open fields in Bogor and Lampung (Lukvitasari et al., 2021)cocoa plantations, and open fields in Polewali Mandar. Consequently, its effectiveness is reduced. In addition, C. odorata endured C. connexa attacks by producing additional branches in response to the attack.

The gall fly C. connexa was found in open fields and cocoa plantations in Polewali Mandar, West Sulawesi, more than 250 km from the release point. This is the first report of C. connexa in Polewali Mandar, West Sulawesi. The abundance of C. connexa in C. odorata was influenced by habitat type. There is a new association between C. connexa and local parasitoids with the discovery of six types of parasitoid morpho-species from four different families. Aprostocetus sp.1 was the dominant parasitoid, with a parasitization percentage of 13.57% in open land and 10.63% in cocoa plantations.