Spondias mombin as a reservoir of fruit fly parasitoid populations in the Eastern Amazon: an undervalued ecosystem service

Characterization of the study area

The state of Amapá (143,453.70 km²), located in the far north of Brazil (Fig. 1) is considered the most preserved state in Brazil with approximately 72% of its territory under some form of environmental protection (Conservation International Brazil, 2007). It thus constitutes an excellent case study for investigating natural relationships between native fruits, tephritid species and associated parasitoids.

The state is bordered to the south and west by the State of Pará, to the east by the Atlantic Ocean, to the north by French Guiana and to the northwest by Suriname (Porto, 2007). The climate, according to the Koëppen-Geiger classification, is of the Aw (tropical savanna) and Am (tropical monsoon) type, with an average annual temperature of 26 °C and an average annual precipitation between 2,300 and 2,400 mm (Peel, Finlayson & McMahon, 2007). The rainy season occurs from January to June and a characteristically dry period is more frequent from September to November (IBGE - Instituto Brasileiro de Geografia e Estatística, 2021). The soils are of the type Yellow Latosol, Red-Yellow Latosol, Red-Yellow Argisol and Gleissolos (Alves, Alves & Mochiutti, 1992). The Amapá ecosystems, due to their super humid equatorial climate, present themselves in the form of forests, very rich within their plant variety, and is divided into: upland forest (not affected by floods), floodplain forest (flooded only during the flooding of rivers), and fields (flooded fields, closed fields and clean fields) (Morais, 2000; Zoneamento Ecológico Econômico, 2008).

The study species Spondias mombin L. (Anacardiaceae)

The genus Spondias (Anacardiaceae) includes 18 species distributed in the Neotropics, Asia and Oceania (Mitchell & Daly, 1995). Spondias mombin (the ‘yellow mombin’ or ‘hog plum tree’), popularly known in Brazil as the “taperebá” or “cajá” tree, is found in the both the Atlantic Forest and Amazon in upland and floodplain forest environments, and is also present in inhabited areas, albeit in a sub-spontaneous state (Cavalcante, 2010). It is an economically important species in the North and Northeast states of Brazil where it is cultivated at a small scale. Its fruit can be consumed in natura or processed into pulps, juices, jams, nectars and ice creams of excellent quality and high commercial value. The species has only recently been commercially exploited and, as yet, there is insufficient technical knowledge for the development of large-scale enterprises (Sacramento & Souza, 2009).

Spondias mombin is a large tree, able to grow to heights of over 25 m which mainly fruits during the rainy season (Queiroz, 2000). In Amapá, the flowering peak of S. mombin occurs from September to October and fruiting lasts approximately eight months, peaking from November to March (Freitas, Santos & Oliveira, 2010). The fruits are globose or elliptical drupes, ranging in color from yellow to light orange, with a thin, smooth skin, juicy pulp and tangy-sweet flavor. They are consumed both in natura or processed into juices and ice creams that are sold in farmer’s markets in the North and Northeast states of Brazil (Santos-Serejo et al., 2009; Cavalcante, 2010).

In the Amazon, the fruits are extracted directly from the forest and, to date, there have been very few attempts to develop commercial orchards. The small production volume means that companies that process this fruit are completely dependent on supply from small-scale extractivism, which is seasonal and often insufficient for the viable operation of processing facilities (Souza, 2005). One potential income-generating option for local rural workers would be to set up domestic orchards (Deus, Sousa & Adaime, 2016). Spondias mombin is also a potentially good option as a component of Agro-Forest Systems, including those developed for restoring degraded areas (Bezerra, Barros Neto & Silva, 2010).

Fruit fly infestation of S. mombin in Amapá

Studies on fruit flies and parasitoids in Amapá have resulted in 14 publications based on 103 samples. Data from another 119 samples are available in unpublished form (Table 1). For the present analysis we have collated data from 252 samples, representing 30,418 fruits (338.60 kg) from S. mombin, of which 223 (88.5%) were infested by fruit flies (Table 1). In the most intensive work carried out, the vast majority of samples were collected from January to May, the rainy season and the greatest abundance of ripe fruits in the field (R Adaime, 2007, unpublished data; Jesus-Barros et al., 2012). However, it is possible to find fruits in the field in the driest months, then some samples were collected in September, November and December (R Adaime, 2007, unpublished data; Sousa et al., 2014). Considering that there is little fruiting during this period, it is necessary to carry out studies to evaluate the role of these fruits in maintaining populations of fruit flies and their parasitoids in native and agricultural areas.

From the collected fruits we have obtained 16,592 puparia of fruit flies with infestation rates generally lower than 100 puparia/kg of fruit; the highest reported being 454.5 puparia/kg (Table 1). These data were collected following two typical methodologies: (i) in most studies (76.9%), a grouped fruits methodology was adopted, whereby the fruits are packaged together, with several fruits from the same plant constituting a single sample; (ii) Cunha et al. (2011), Sousa et al. (2014) and Lemos et al. (2017) used an individualized fruit methodology, where each fruit represents a subsample. When we take together the amount of fruits from an individualized sample, they are equivalent to a grouped sample. Therefore, for the purpose of discussion we consider everything as a “grouped sample”.

Five species of Anastrepha have been reported from S. mombin fruits in the state of Amapá: Anastrepha antunesi Lima, Anastrepha fraterculus (Wiedemann), Anastrepha obliqua (Macquart), Anastrepha sororcula Zucchi, and Anastrepha striata Schiner (Table 1). Infestation by the invasive Bactrocera carambolae Drew & Hancock (Brasil, 2018) has also been recorded (Silva et al., 2011a; Lemos et al., 2014). Of these species, A. obliqua is the most commonly found on S. mombin fruits (Silva, Lemos & Zucchi, 2011; Deus, Sousa & Adaime, 2016), although A. antunesi is also frequently obtained from samples, albeit at a lower abundance than A. obliqua.

In Brazil, 121 species of Anastrepha have been recorded (Zucchi & Moraes, 2021). However, only a few of these species are economically important, including A. fraterculus (the South American fruit fly), A. obliqua (the West Indian fruit fly), and A. striata (the guava fruit fly) (Leonardo et al., 2017). These species were obtained from several samples of S. mombin collected in Amapá (Table 1). Exactly 116 host plants of A. fraterculus, 51 of A. obliqua and 31 of A. striata have already been reported in Brazil (Zucchi & Moraes, 2021; Adaime et al., 2014). Anastrepha antunesi and Anastrepha sororcula have no economic importance (Uramoto & Zucchi, 2009), the former being frequently associated with S. mombin and the latter with Psidium guajava (Adaime, Sousa & Pereira, 2016). Bactrocera carambolae (the carambola fruit fly), native to Asia, is an invading species in South America (Castilho et al., 2019a). Its first detection in Brazil was in 1996, in Oiapoque county, state of Amapá (Silva et al., 2004). This species is currently classified in Brazil as a quarantine pest and is under official control by the Ministry of Agriculture and Food Supply (Brasil, 2018). In addition to S. mombin, it has been recorded on 25 other host plants in Brazil, all of which occur in Amapá (Belo et al., 2020).

Although it is possible to find fruits of S. mombin with significant infestation (up to 8 pupae), the simultaneous occurrence of A. obliqua and A. antunesi is uncommon (only 0.5 to 1.3% of the fruits), which may mean low interspecific competition (Cunha et al., 2011; Sousa et al., 2014; Nascimento et al., 2015). Another factor that may explain the rare sharing of the same fruit by A. obliqua and A. antunesi is the significant abundance of this plant species in the sampling areas (Nascimento et al., 2015) - in addition to the presence of fruits from other hosts, since they are polyphagous (Zucchi & Moraes, 2021). This highlights the importance of studies based on individual fruits, since these allow researchers to better study the tritrophic relationship between the host fruit, the larvae of flies that infest it, and the parasitoids associated with them (Silva et al., 2011c).

Fruit fly parasitoids from S. mombin in Amapá

The vast majority of fruit samples of S. mombin contained parasitoids of fruit flies, especially those collected in the rainy season (Table 1). Four species of Braconidae have been recorded in association with S. mombin in the state of Amapá: Doryctobracon areolatus (Szépligeti), Opius bellus Gahan, Asobara anastrephae (Muesebeck), and Utetes anastrephae (Viereck). The eucoiline parasitoid Aganaspis pelleranoi (Brèthes) Figitidae has also been reported on S. mombin. Doryctobracon areolatus and O. bellus are the most widely distributed species in Brazil and the most common in fruits from the Amazon region (Silva, Lemos & Zucchi, 2011; Sousa, Adaime & Pereira, 2016). Asobara anastrephae, U. anastrephae, and A. pelleranoi are also abundant in the region, but are generally represented by a few specimens in the samples (Deus & Adaime, 2013; Sousa, Adaime & Pereira, 2016).

Braconidae parasitoids are the most important natural enemies of fruit flies at the global level (Wharton, 1996; Ovruski et al., 2000), besides being widely distributed. They are frequently obtained from fruits infested by fruit flies in the Brazilian Amazon and in Brazil as a whole (Canal & Zucchi, 2000; Sousa, Adaime & Pereira, 2016). The Figitidae occur throughout Brazil, although they are poorly studied (Guimarães & Zucchi, 2011). In the Brazilian Amazon, there are records in the states of Amazonas, Amapá, Pará and Roraima (Sousa, Adaime & Pereira, 2016).

Members of the Braconidae and Figitidae are mainly koinobiont endoparasitoids that oviposit in the larval stage, developing in the living host and killing it at the end of the cycle when the adult wasp emerges from the host’s puparium (Ovruski et al., 2000; Brodeur & Boivin, 2004; Marinho, Costa & Zucchi, 2018). Most native Braconid parasitoids are generalists, being able to parasitize Anastrepha species on a wide variety of plant species (López, Aluja & Sivinski, 1999). They are the most studied fruit fly parasitoids worldwide (Wharton, Marsh & Sharky, 1997), being preferred in biological control programs due to host specificity by the Tephritidae family (Clausen, 1940). Of the four species of Figitidae that occur in the Brazilian Amazon, A. pelleranoi is the most generalist, being frequently associated with Tephritidae puparia (Guimarães & Zucchi, 2011).

Parasitism rates varied between studies, reaching as high as ∼50% of puparia (Table 1). Specifically, Cunha et al. (2011) recorded a parasitism rate of 46.6%, almost equally split between D. areolatus (57.1% of specimens) and O. bellus (42.9%). Nascimento et al. (2015) reported a somewhat lower parasitism rate of 27.8% for the same species of parasitoids [D. areolatus (55.4%) and O. bellus (44.6%)]. Overall, the braconid D. areolatus is the most abundant species in samples of S. mombin in Amapá, including the capital city Macapá (Silva, Jesus & Silva, 2006a) and Santana (Silva et al., 2007a). However, in some samples O. bellus was the most abundant species Deus et al., (2013). In the Brazilian Amazon, D. areolatus and O. bellus have been associated with 16 and nine species of Anastrepha, respectively (Marinho, Silva & Zucchi, 2011). The association of a given parasitoid species with a fruit fly species can only be considered when they are found exclusively (one species of parasitoid and one fruit fly species) in a sample of a single fruit species (Leonel Junior, Zucchi & Wharton, 1995). Under these criteria, the parasitoids D. areolatus and O. bellus are associated with A. obliqua (Sousa et al., 2014) and A. antunesi (Nascimento et al., 2015) in S. mombin fruit.

Although braconid parasitoids prefer certain species of host flies, the physical (weight, color and size) and chemical (chemical volatiles) characteristics of fruit are also important factors in parasitoid attraction (Marinho, 2009). For example, parasitism is more difficult in large fruits, as the larvae have to burrow deep into the pulp (Sivinski, Aluja & López, 1997). Consequently, fruit weight and size are usually inversely proportional to the rate of parasitism, i.e., the heavier and larger the fruit, the less likely it is to be parasitized (Hernández-Ortíz, Pérez-Alonso & Wharton, 1994; López, Aluja & Sivinski, 1999). Fruits of S. mombin are quite variable in size (Cunha et al., 2011): mean smaller diameter 26.5 ± 0.24 mm (ranging from 21.4 to 33.7), mean larger diameter 36.2 ± 0.30 mm (ranging from 29.3 to 43.5), weight mean 12.7 ± 0.31 g (ranging from 6.6 to 22.7) and pulp thickness mean 9.5 ± 0.14 mm (ranging from 6.7 to 12.2). Cunha et al. (2011) found no significant correlation between percentage of parasitism and fruit weight, with no correlation between parasitism and pulp thickness. This suggests that the range thickness of S. mombin fruits does not strongly influence parasitism or the efficiency of parasitoids in finding larvae of their host flies. Significant differences in the chemical composition of fruits of S. mombin were observed by Bezerra, Barros Neto & Silva (2010). The authors collected only two samples from the municipality of Mazagão, state of Amapá, and observed different patterns for almost every qualitative parameter that was analyzed, highlighting the enormous diversity of genetic material in the region. To date, no studies have been published showing the relationship between parasitism and chemical characteristics of S. mombin fruits.

Ovipositor length also varies among species (D. areolatus −3.15 to 4.59 mm; O. bellus −0.93 to 1.66 mm; U. anastrephae −0.85 to 1.87 mm; and A. anastrephae −2.49 to 3.28 mm (Marinho, 2009) and is a determining factor in parasitism. The shorter ovipositor of U. anastrephae may be an adaptation for foraging on smaller fruits, potentially competing with D. areolatus, which has a longer ovipositor (Sivinski, Aluja & López, 1997).

Potential for biological control of fruit flies using parasitoids

In Amapá, some forest inventories have already been completed in “várzea” floodplain forest ecosystems, providing phytosociological data on S. mombin (Table 2). The highest concentrations of this species are located along the Jari River basin, with densities of 69 to 130 individuals per hectare (Carim et al., 2017).

Considering fruit infestation rates by fruit flies and fruit fly parasitism rates by parasitoids, we can make a crude estimate of the number of parasitoids ‘produced’ per 1 kg of S. mombin fruits. Specifically, our studies suggest that floodplain forest ecosystems in the state of Amapá can ‘produce’ up to 66.7 specimens of parasitoids of Anastrepha from 1 kg of fruits of S. mombin (Table 1). If we consider that 10 to 130 S. mombin trees can be found per hectare in a floodplain forest (Table 2), and that each tree can potentially produce up to 10,000 (=100 kg) fruits (Janzen, 1985), then these forests conceivably host over 900,000 parasitoids per hectare.

When estimating the production of parasitoids per hectare, we must take into account the natural mortality factors of these insects. In field conditions, the risk of mortality is associated, for example, with the search for a source of food, which, if not found before the insect runs out of energy, will cause its death by starvation (Bernstein & Jervis, 2008). There is also the incidence of extrinsic factors, such as climate (Weisser, Völkl & Hassell, 1997), predation (Jervis, 1990; Heimpel, Rosenheim & Mangel, 1997; Rosenheim, 1998) and desiccation (Jervis, Ferns & Heimpel, 2003). In addition, we must consider that the rates of parasitism in fruits that are collected in the field and placed under laboratory conditions do not accurately represent reality. These fruits, when removed from the natural environment, may contain larvae of first and second instars that are not yet susceptible to being parasitized Adaime et al., (2018). Thus, when the fruits containing the immatures are removed from the field, it is no longer possible for parasitism to occur (Uchôa-Fernandes et al., 2003). Thus, the actual rate of parasitism may be even higher than predicted, as observed by Adaime et al., (2018) in larvae of A. coronilli that infest fruits of B. grossularioides. Parasitism may also be underestimated in studies using fallen fruit because, when the infested fruits fall, larvae abandon them and penetrate the soil to begin the pupa phase Adaime et al., (2018). This is a limitation for most studies on large canopy trees because in the cases fruit samples are usually collected from the soil. In any case, our estimated amount of parasitoids per hectare of foodplain forest (900,000) is much higher than that estimated by Sousa et al. (2021) for a fragment of upland forest in Amapá, with the occurrence of G. argenteum, natural host of Anastrepha atrigona Hendel (2,500 parasitoids per ha). Sousa et al. (2021) use a specific plot-based methodology and a robust statistical approach, the application of which to S. mombin in Amapá floodplain forests would make it possible to measure the contribution of this plant species to maintaining fruit fly parasitoid populations.

To date, besides Amapá, S. mombin has been sampled in other four of the nine states that make up the Brazilian Amazon. The highest percentages of parasitism recorded in each state were 29.5% in Acre (Thomazini & Albuquerque, 2009), 30.0% in Amazonas (Dutra et al., 2013), 31.1% in Roraima (Marsaro Júnior et al., 2010), 46.9% in Amapá and 70.8% in Pará (Brandão et al., 2019) (Table 3). The density of parasitoids may be even higher than estimated for Amapá (66.7 parasitoids/kg of fruit): Marsaro Júnior et al. (2010) obtained 165 parasitoids/kg of fruit from Roraima (also in the Brazilian Amazon), while López, Aluja & Sivinski (1999) obtained an impressive 207 parasitoids/kg of S. mombin fruit in Mexico. Also in Mexico, Murillo et al. (2015), found that 88% of the A. obliqua larvae obtained from S. mombin were parasitized. Due to the great difference in sampling effort, it is not possible to directly compare the data collected in Amapá during almost 20 years (Table 1) with the point data produced in other locations in the Brazilian Amazon (Table 3) and in Mexico. Therefore, it is necessary that studies be carried out on a larger spatial and temporal scale to identify which factors influence the percentage of parasitism and the number of parasitoids obtained per kilo of fruit.

We should also take into account that S. mombin does not occur exclusively in floodplain forests, but can also be found in upland forest and in secondary vegetation areas where it regenerates spontaneously from seeds or from stakes and roots (Bosco et al., 2000). Amapá has a total floodplain forest area of approximately 543,348 hectares and 10,362,374 hectares of upland forest (Zoneamento Ecológico Econômico, 2008), in addition to highly conserved upland forests. Another factor to be considered is that the majority of Amapá floodplain forests are difficult to access. This undoubtedly contributes to conservation of the region, but also hampers data collection leading to incomplete knowledge about many aspects of biodiversity (Gaston & Rodrigues, 2003; Mace, 2004; Ladle & Hortal, 2013).

We have shown that the floodplain forests and upland ecosystems of Amapá have enormous potential as areas for the multiplication of parasitoids of fruit fly species. In these ecosystems (where S. mombin is present) millions of parasitoids from at least five species (O. bellus, D. areolatus, A. anastrephae, U. anastrephae and A. pelleranoi) are produced every year at zero economic cost. The results of almost 20 years of research indicate that S. mombin can potentially be used to maintain and multiply parasitoids which, in turn, act as agents of natural biological control of agricultural pest species in the genus Anastrepha. In the Brazilian Amazon, parasitoids from S. mombin can regulate fruit fly populations in fruits of socioeconomic importance, for example Anastrepha striata in guava (Sousa et al., 2019). In addition to its role as a host for biological control agents, S. mombin could contribute both economically and environmentally to projects involving the restoration of legal reserves. These reserves are created by large landowners as part of their obligations under federal law for preserving a proportion of natural habitat. The main goals of legal reserves are to ensure an economical and sustainable use of natural resources in rural properties, to help conserve and rehabilitate ecological processes, and to promote the conservation of biodiversity (Brasil, 2012).

Finally, although Amapá is considered the best-preserved state in Brazil (Conservation International Brazil, 2007), large swathes of other states in the Brazilian Amazon are degraded or in the process of degradation. Given this situation, recovering these areas using native plant species with commercial potential and that can also provide valuable ecosystem services is an interesting alternative. As such, S. mombin has enormous potential for use in restoration and agroforestry projects.