SITE CHARACTERISTICS

The study was conducted on Ilha de Santana, municipality of Santana, state of Amapá, Brazil (Fig. 1). The island occupies an area of approximately 2,005 ha and is situated on the banks of the Norte Canal, facing the city of Santana, between the geographic coordinates 00.0666667°S and 00.1000000°S, and 00.0666667°W and 51.2083333°W (Valente et al. 1998).

The predominant climate in the area is Amw' under the Köppen classification system, characterized as a tropical wet climate with a well-defined dry season, with mean temperatures never dropping below 18 °C and annual fluctuations seldom exceeding 5 °C. Annual precipitation ranges from 1,300 to 1,900 mm, with well-defined rainy and dry seasons (Dec to Mar and Aug to Nov, respectively). The predominant soils are Yellow Latosol and Haplic Gleysol. Yellow Acrisol, Fluvic Neosol, and Indiscriminate Hydromorphic Soils are also present (Valente et al. 1998).

The island is located 600 to 800 m from the Port of Santana. Access is obtained using small boats. The predominant activity on the island is small-scale family agriculture, with fruit growing as the most representative activity. Products are sold once a week, mainly in the form of fruit concentrate, in public farmers' markets in the municipalities of Macapá and Santana.

SAMPLING PROCEDURES

Monthly sampling of fleshy fruits from various plant species was performed from Jan to Jul 2015, a period in which there is high availability of fruiting plants in the region. To quantify the rate of infestation by fruit flies and the percentage of parasitism, we applied the grouped samples method described by Silva et al. (2011a). Each sample was determined by the availability of fruits. For each plant species with a high availability of fruits, a sample consisted of approximately 20 mediumsized fruits or 50 small fruits (partially or fully ripe fruits, collected directly from the tree or recently fallen to the ground). When there was not a sufficient amount of fruits, we collected whatever was available on the trees. The geographic coordinates of the sampling sites were recorded with a GPS device. In the field, the fruits were counted and stored in properly tagged plastic bottles that were wrapped in organza bags and closed with rubber bands. Later, the bottles were placed on plastic trays and transported by river to the Port of Santana, then by road to the Plant Protection Laboratory at Embrapa Amapá, in the city of Macapá, where the fruits were stored.

ACQUISITION OF PUPARIA AND ADULT INSECTS

In the laboratory, the fruits were weighed and transferred to plastic trays containing a thin layer of sterilized, moistened vermiculite. The trays were covered with organza fabric fastened in place with rubber bands. The material on the trays was examined every 3 d. Recovered puparia were removed and transferred to transparent plastic jars (8 cm diameter) containing a thin layer of moistened vermiculite. The bottles were covered with organza fastened in place with a vented lid, then placed in climate-controlled chambers under controlled temperature (27 ± 0.5 °C), relative humidity (70 ± 10%) and photoperiod (12:12 h L:D). They were checked daily for puparia. Adult insects that emerged (fruit flies and parasitoids) were killed and stored in glass vials containing 70% ethanol, duly tagged for subsequent identification.

IDENTIFICATION OF INSECTS

Specimens of Anastrepha were identified using the illustrated identification key published by Zucchi et al. (2011). Only females were used for identification, which was performed by examination of everted aculei, using a stereomicroscope and optical microscope (40×). Other characteristics were also observed, including wing pattern, mesonotum, mediotergite, and subscutellum. Confirmation of B. carambolae identity was based on the identification key published by Drew & Hancock (1994). To identify parasitoids (Hymenoptera: Braconidae), we used the work of Canal & Zucchi (2000) and Marinho et al. (2011). Neosilba specimens were identified according to McAlpine & Steyskal (1982) and Strikis (2011).

IDENTIFICATION OF BOTANICAL MATERIAL

To identify the forest plant species, we collected branches containing their reproductive structures (flowers and fruits), which were later processed into herbarium specimens using the mounting and preservation techniques described by Fidalgo & Bononi (1984). The plant species were identified using identification keys and specialized literature, as well as comparison with specimens available at the Herbário Amapaense (HAMAB), the herbarium at the Amapá Institute for Scientific and Technological Research (IEPA) in Macapá, Amapá, Brazil.

Fig. 1.

Sampling sites of fruits on Ilha de Santana, state of Amapá, Brazil (Jan to Jul 2015).

DATA ANALYSES

The following data were calculated: 1) infestation index = number of puparia obtained/weight (kg) of fruit collected; 2) emergence = (number of emerged flies + number of emerged parasitoids)/total number of puparia × 100; and 3) percentage of parasitism = (number of parasitoids emerged/number of puparia) × 100.

TEPHRITIDAE

The species of Tephritidae obtained were: A. striata, A. obliqua, A. fraterculus, Anastrepha antunesi Lima, and B. carambolae (Table 1). Anastrepha striata occurred on 2 hosts (M. emarginata and P. guajava) and A. obliqua occurred on 3 (S. mombin, M. emarginata, and S. cumini). Anastrepha antunesi and A. fraterculus occurred exclusively on S. mombin (Table 2). Bactrocera carambolae occurred on 8 of the 11 plant species infested by fruit flies: A. carambola, E. uniflora, Licania sp., M. emarginata, M. indica, P. guajava, S. mombin, and S. cumini (Tables 1, 2, and 3).

Table 1.

Rates of infestation of various plant species by fruit flies on Ilha de Santana, Amapá, Brazil (Jan to Jul 2015).

Continued

Anastrepha accounted for 75.2% of all fruit flies obtained in this study, followed by Bactrocera (23.7%) (Table 3). The highest abundance of fruit flies occurred on P. guajava, with 1,611 specimens, 90.2% of them consisting of A. striata, 9.6% of B. carambolae, and 0.2% of the lonchaeid Neosilba zadolicha McAlpine & Steyskal.

LONCHAEIDAE

The genus Neosilba accounted for only 1.1% of all fruit flies obtained in this study (Table 3). The species obtained were: N. zadolicha, Neosilba pendula (Bezzi), Neosilba pseudozadolicha Strikis, and Neosilba glaberrima (Wiedemann) (Tables 1 and 2). Five hosts were recorded for lonchaeids: C. baccatum, M. emarginata, I. edulis, Passiflora sp., and P. guajava (Tables 1 and 2). Malpighia emarginata was infested by 3 species (N. pendula, N. pseudozadolicha, and N. zadolicha) (Tables 1 and 2). Neosilba glaberrima and N. zadolicha infested 2 and 4 plant species, respectively. Infestation by Lonchaeidae was observed in at least 1 plant species in the months of Jan to Apr and Jul (Table 2).

PARASITOIDS

Three species of Braconidae (Hymenoptera) were obtained from fruits of S. mombin, Licania sp., M. emarginata, S. cumini, and P. guajava: D. areolatus, O. bellus, and Utetes anastrephae (Viereck). The most abundant species was D. areolatus, accounting for 89.0% of all parasitoids obtained (Table 1). The highest percentages of parasitism were observed on S. mombin (43.1%) and Licania sp. (42.6%) (Table 1).

HOST PLANT USE BY B. CARAMBOLAE

In all months of sampling, infestation by B. carambolae was observed on at least 2 plant species, peaking at 5 species in Apr. Eugenia uniflora and P. guajava were found to be infested in 5 of the 7 mo of sampling, and M. emarginata was infested in 6 mo (Table 2).

Bactrocera carambolae was the only species to infest A. carambola, E. uniflora, Licania sp., and M. indica (Tables 1, 2, and 3). On M. emarginata, it corresponded to 80.9% of all Tephritidae obtained (Table 3). On the other hand, on P. guajava, specimens of B. carambolae accounted for only 9.6% of all Tephritidae obtained, and the prevalent species was A. striata. On S. mombin, B. carambolae represented only 0.5% of Tephritidae obtained.

The sampling events were performed in 5 rural establishments (designated here as E₁ to E₅), as well as 3 isolated collection points (P₁ to P₃) located on the sides of secondary roads. Bactrocera carambolae was obtained from 48 samples collected in all of the sampled rural establishments, as well as 1 of the isolated collection points (Table 4).

TEPHRITIDAE

This research identified 3 new reports of fruit flies, in addition to those previously reported by Silva et al. (2007): A. fraterculus, A. antunesi, and B. carambolae (Table 1). Therefore, to date, 6 species of fruit flies are known to occur on Ilha de Santana: A. antunesi, A. fraterculus, A. leptozona, A. obliqua, A. striata, and B. carambolae.

Anastrepha striata was the predominant species. This species exhibits a marked preference for P. guajava, a plant species with high availability of fruits during the sampled period. All 34 samples of P. guajava collected in this study (396 fruits, 16.8 kg) were infested with A. striata (Table 3). This dominance of A. striata can therefore be explained by the abundance of fruits of P. guajava, its main host at many localities in South America, including the Brazilian Amazon (Aluja 1994; Silva et al. 2011b; Marsaro Júnior et al. 2013). In addition, A. striata is the most abundant and widely distributed tephritid species in the state of Amapá, and is also the most polyphagous (infesting 25 host plants in 16 botanical families) (Silva et al. 2011c; Jesus-Barros et al. 2012).

Table 2.

Fruit flies obtained from host plants sampled in each month of sampling on Ilha de Santana, Amapá, Brazil (Jan to Jul 2015).

Table 3.

Specimens of Anastrepha spp., Bactrocera carambolae, and Neosilba spp. obtained from plant species on Ilha de Santana, Amapá, Brazil (Jan to Jul 2015).

Anastrepha obliqua has been particularly associated with S. mombin. In surveys conducted in the state of Amapá, A. obliqua was predominant on this host (Silva et al. 2007, 2011d; Deus et al. 2009, 2013). In this work, the mean infestation index was 173.2 puparia per kg, lower than the highest index reported for this host in the state of Amapá, i.e., 385.1 puparia per kg in samples obtained at Serra do Navio (Deus et al. 2013). These results show the importance of A. obliqua as a pest of S. mombin in the state of Amapá (Silva & Ronchi-Teles 2000; Deus & Adaime 2013; Deus et al. 2016), especially as the fruit is very well-liked by the local population, who purchase it as concentrate to be made into juice.

This is the first report of A. obliqua associated with S. cumini in Brazil. On the other hand, A. obliqua has already been associated with M. emarginata in the states of Amapá, Pará, and Roraima (Ohashi et al. 1997; Amorim et al. 2004; Marsaro Júnior et al. 2011; Lemos 2014). Anastrepha antunesi and A. fraterculus had already been associated with S. mombin in the state of Amapá (Deus & Adaime 2013).

Bactrocera carambolae was obtained from 8 plant species, with E. uniflora, S. cumini, and Licania sp. representing new reports of host plants for the species in Brazil. The other plant species had already been reported to be hosts of B. carambolae in Amapá (Silva et al. 2004; Lemos et al. 2010, 2014).

LONCHAEIDAE

All Lonchaeidae obtained in this study are first reports for Ilha de Santana. Neosilba glaberrima and N. zadolicha infested the highest number of hosts (Tables 1 and 3), confirming the results obtained by Strikis et al. (2011) and Lemos et al. (2015), which indicated that these are the most polyphagous species of Lonchaeidae in the Amazon region. They are also among the species of highest economic importance for South America (Uchôa 2012).

In the Brazilian Amazon, N. pendula had already been reported on M. emarginata in the states of Pará and Roraima (Pereira & Adaime 2016). Araújo & Zucchi (2002) reported the species as an important primary invader of M. emarginata fruits in the state of Rio Grande do Norte.

No specimens of Lonchaeidae were obtained from S. mombin. This result was repeated during the sampling performed by Lemos et al. (2015) in 3 municipalities of Amapá (14 samples, containing 210 fruits in total).

Five samples of 4 plant species were infested exclusively by species of Neosilba: N. glaberrima on P. edulis (1 sample), N. pendula on M. emarginata (1 sample), N. zadolicha on I. edulis (1 sample), and N. zadolicha and N. glaberrima on C. baccatum (2 samples). This finding may indicate that these lonchaeid species are primary pests of the plant species in question. In addition, Uchôa (2012) mentioned that for some plant species of economic importance in South America, lonchaeids may be more abundant and important as pests than tephritids.

This work makes new associations between species of Neosilba and host plants in the Brazilian Amazon: N. glaberrima on Passiflora sp. and C. baccatum; N. pseudozadolicha on M. emarginata; and N. zadolicha on M. emarginata and C. baccatum (Table 2).

PARASITOIDS

This work adds 1 species of parasitoid (U. anastrephae) not reported by Silva et al. (2007) on Ilha de Santana. Therefore, to date, 4 species of parasitoids have been reported at this locality: A. anastrephae, D. areolatus, O. bellus, and U. anastrephae.

The higher abundance of D. areolatus found in this work is consistent with the notion that this is the most abundant and disseminated species of native parasitoid of Anastrepha in Latin America, including in Brazil (López et al. 1999; Ovruski et al. 2005; Marinho et al. 2011). The longer ovipositor of D. areolatus (3.8 mm) allows it to infest larvae of fruit flies in fruits of various sizes, enabling it to outperform other parasitoids (Aluja et al. 2013).

The highest percentages of parasitism were observed on S. mombin (43.1%) and Licania sp. (42.6%), which are native plants in the sampled region (Table 1). Overall, native plants in the wild harbor significantly more parasitoids per fruit than cultivated plants (López et al. 1999; Aluja et al. 2003).

All 3 species of parasitoids obtained in this work were present on S. mombin. This plant species has been reported as an important reservoir of parasitoids in the state of Amapá (Sousa 2015). Considering that 11 samples of S. mombin were collected (4.9 kg of fruits) and that 368 specimens of parasitoids were obtained from these samples, this means that 75.1 parasitoids were obtained per kilogram of fruit. The greatest number of parasitoids obtained to date from S. mombin in the Brazilian Amazon was 165 parasitoids per kg of fruit, in the state of Roraima (Marsaro Júnior et al. 2011). However, these are low values if compared with those reported by López et al. (1999) in Mexico (207 parasitoids per kg of fruit).

Table 4.

Occurrence of Bactrocera carambolae (samples collected/samples infested) on plant species sampled on Ilha de Santana, Amapá, Brazil (Jan to Jul 2015).

Two of the 3 samples of Licania sp. fruits were infested by fruit flies (presence of puparia). Four puparia were obtained from 1 sample, from which emerged 2 specimens of B. carambolae. Only parasitoids emerged from the other sample (2 specimens of D. areolatus and 1 unknown specimen). This finding merits special attention, as it suggests that at least 1 parasitoid species may be infesting larvae of B. carambolae. However, this cannot yet be positively affirmed. Should this be confirmed by an additional study, it will be the first report of a natural parasitoid of B. carambolae in Brazil, given that, up to the present time, no native parasitoid has been reported for this species (Adaime et al. 2014a). Lemos (2014) individually observed 1,262 puparia of B. carambolae in Amapá, originating from 9 plant species, without obtaining any parasitoid specimens. In Suriname and French Guyana, there is also no indication of native parasitoids specifically attacking larvae of B. carambolae (Sauers-Müller 2005; Vayssières et al. 2013). However, Vayssières et al. (2013) considered the hypothesis that parasitoids do attack immatures of Bactrocera, but do not successfully develop due to poor adaptation to this host or the immune response of its larvae. According to their observations, the only species of parasitoid to emerge from pupae of B. carambolae was Diachasmimorpha longicaudata (Ashmead) (Hymenoptera: Braconidae), which was introduced to the region in the year 2000.

We did not detect any parasitoids in Lonchaeidae in this work, although species of Eucoilinae (Figitidae) have already been reported in association with frugivorous larvae of Neosilba in Brazil (Uchôa 2012).

HOST PLANT USE BY B. CARAMBOLAE

Among the 8 plant species identified as hosts of B. carambolae in this study, 5 (S. mombin, M. indica, M. emarginata, P. guajava, and A. carambola) were also sampled on Ilha de Santana by Silva et al. (2007), without being infested by the pest. The authors collected 5 samples of A. carambola (416 fruits, 9.2 kg) and 9 samples of M. emarginata (2,741 fruits, 14.4 kg) but did not obtain any puparia. This suggests that the composition of fruit fly species at this locality has changed in the 10 yr interval between the 2 surveys.

Bactrocera carambolae infested 70.4% of the samples of M. emarginata (19 out of 27 samples), 66.7% of the samples of A. carambola (4 out of 6 samples), 55.5% of the samples of E. uniflora (5 out of 9 samples), and 41.2% of the samples of P. guajava (14 out of 34 samples) (Table 3).

Analyzing the percentages of occurrence of B. carambolae when compared with other fruit flies (Table 3), we can assume the pest to prefer hosts that are not infested by other species. This is clearly illustrated by noting that B. carambolae was the only species to infest A. carambola, E. uniflora, Licania sp., and M. indica in this study. In addition, when B. carambolae infests hosts that are strongly associated with a given species of fruit fly, such as S. mombin with A. obliqua, or P. guajava with A. striata, its percentage of occurrence is very low.

Averrhoa carambola, E. uniflora, M. emarginata, and P. guajava are responsible for sustaining the population of B. carambolae in the sampled areas (Tables 2 and 4). In the case of A. carambola, the availability of fruits and consequent infestation occurred in the initial months of the year. In the case of the other plant species mentioned, particularly M. emarginata, fruits were available throughout the sampling period.

Bactrocera carambolae was found at all 5 sampled rural establishments and at 1 of the 3 isolated collection points (Table 4). The percentage of samples infested by B. carambolae at each establishment or isolated collection point ranged from 11.8% (E₄) to 100% (E₅). The number of samples infested by the pest varied according to sampling month, peaking in Apr and May, with 10 and 11 samples infested, respectively (Table 4).

On M. indica (unidentified genetic material, not grafted), although 11 samples were collected (80 fruits, 14.4 kg), only 7 puparia were obtained, originating from a single sample collected in Apr in establishment E₃ (Table 4) and from which emerged only a single specimen of B. carambolae. Lemos (2014) also sampled fruits of M. indica (50 fruits, unidentified genetic material, not grafted) in 3 municipalities in Amapá without observing any infestation. On the other hand, Lemos et al. (2014) obtained 22 puparia and 19 adults of B. carambolae from a single fruit of M. indica (Tommy Atkins cultivar), reporting an infestation rate of 28.5 puparia per kg. This relationship merits further investigation, as the origins of the genetic materials cultivated in the region are unknown, as is their potential resistance to B. carambolae. The low rate of tephritid infestation on M. indica can be at least partly explained by the density of laticiferous ducts present on the epicarp and mesocarp of the fruit, which has a toxic effect on eggs and larvae, as shown by Joel (1980, 1981) for C. capitata and by Adaime et al. (2014b) for A. obliqua and Anastrepha ludens (Loew). These topics should therefore be studied further, especially considering that M. indica is widely used for urban landscaping in Macapá and Santana, municipalities that contain over 70% of the population of the state of Amapá. If their genetic materials are susceptible to infestation by the carambola fruit fly, these trees used in urban landscaping could be responsible for sustaining elevated populations of the pest.

Finally, it should be noted that B. carambolae seems to be adapting to infest native hosts in the Amazon, such as Licania sp. Lemos et al. (2014) highlighted this same fact, mentioning infestations on Eugenia stipitata McVaugh (Myrtaceae) and Pouteria macrophylla (Lam.) Eyma (Sapotaceae) in the state of Amapá. In both reports, although B. carambolae infested native plants, the fact that they were not located in completely unaltered environments should be taken into account.

In conclusion, the most important species of fruit flies on Ilha de Santana are: B. carambolae, for being a species of quarantine importance; and A. obliqua and A. striata, for infesting plant species of local socioeconomic importance (S. mombin and P. guajava, respectively). Species of Neosilba, though potential pests, are not abundant at the sampled locality. In addition, we can conclude that A. carambola, E. uniflora, M. emarginata, and P. guajava are responsible for sustaining the population of B. carambolae in the sampled areas.