Abstract

A study was conducted to characterize the species composition of fruit flies in South Wollo and North Wollo administrative zones of northeastern Ethiopia. Fruit flies were collected using two methods: rearing from infested mango, guava, and sweet orange fruits and trapping the adults with male lures and food-based attractants. Three fruit fly species were recovered from the fruits collected: Bactrocera dorsalis, Ceratitis cosyra, and Ceratitis capitata; in addition to these species, Ceratitis fasciventris was trapped. Bactrocera dorsalis was the most abundant species and accounted for 97.9 and 78.89% of the total trapped and emerged adults, respectively. The number of B. dorsalis and C. capitata showed a significant difference among trapping locations and attractants. The highest number of B. dorsalis was trapped with methyl eugenol at Kalu (722.2 flies per trap per week). Among the indigenous fruit fly species, C. capitata was higher at Habru which was guava and coffee-dominated habitat, whereas C. cosyra was relatively higher at Kobo which is mango dominated. More adults were trapped in male lure traps (97.03%) than in food-based attractants (2.97%). Hence, fruit fly management should focus on guava and mango crops which host all the fruit fly species, including the dominant B. dorsalis.

1. Introduction

Fruit flies (Diptera: Tephritidae) are one of the most diverse groups of insects, comprising over 4000 species in 481 genera [1]. They are considered by far the most important group of horticultural pests worldwide. Each continent is plagued by several fruit fly pests, both indigenous and invasive ones, causing tremendous economic losses (De Meyer et al. 2015). In addition to the direct losses through damage, they can negatively impact commodity trade through restrictions to market access [2].

Tephritidae are distributed worldwide in temperate, tropical, and subtropical regions [3]. The genus Anastrepha is restricted to the Western Hemisphere, with various species, while most of the Bactrocera are native to tropical South and Southeast Asia [4]). The genus Ceratitis is native to tropical Africa [5]. In most African countries, many indigenous fruit fly species belong to Ceratitis. However, Bactrocera spp. (mainly Asian origin) appear to dominate over Ceratitis populations [6]. A complex of fruit fly species commonly coexists in the fragmented fruit and vegetable production systems in Africa [7]. An important aspect of fruit fly management is accurate information on the species and their host spectrum [8].

Fruit flies, especially alien invasive species, constitute a major threat to horticulture in Africa [9]. Since the arrival of Bactrocera dorsalis (Hendel) on the African continent [10], direct yield losses in mango, depending on local conditions, cultivars, and seasons, have raised up to 90%, which is significantly higher than earlier losses attributed just to native fruit fly species [11].

In Ethiopia, different species of fruit flies have been reported to infest the berry and stone fruits in considerable extents [1215]. However, most of the producers have not recognized the fruit flies as a problem and did not associate fruit rotting with this pest [15]. Recently, the invasive species B. dorsalis is being reported as the most devastating mango fruit fly in eastern, southwestern, and central Ethiopia [14, 16, 17] even causing up to 100% loss on guava in central rift valley region after its entry [17], where Ceratitis capitata (Wiedemann) and Ceratitis fasciventris (Bezzi) were previously reported as a dominant species [18]. Recently, in the study area, up to 78% fruit infestation on guava and 28% on mango have been reported [19].

In northern Ethiopia especially northeastern part, having wide coverage of fruit crop production, the information about the species composition causing infestation on the major fruit crops of the area is not well studied, though this information primarily required to plan management options for the loss being caused. Hence, this research was conducted to characterize the species compositions in major fruit crop-producing areas in northeastern Ethiopia.

2. Materials and Methods

2.1. Study Areas
2.1.1. Fruit Collection Sites

The study was conducted in the South Wollo administrative zone (Kalu district), and North Wollo administrative zone (Habru and Kobo districts) of eastern Amhara, Ethiopia. A total of nine collection sites were selected in consultation with zone and district agricultural office experts, to represent the major mango, guava, and citrus fruit production areas of the respective districts. Three sites from each district were selected for the study. However, one fruit collection site from Kalu district (Degan) was omitted for guava fruit since there was no guava in the area during the study period. Table 1 presents the major areas surveyed and their approximate georeferenced positions.

Fruit crops are widely grown in the study area, being the most important mango, guava, orange, banana, and coffee. The districts were selected based on their fruit crop production potential purposively; Kalu and Kobo districts are well-organized fruit production areas, and Habru is a major guava production area besides mango, orange, and coffee in eastern Amhara.

2.1.2. Trap Sites

Three orchards were selected for adult trapping; the orchards were selected based on the availability of mango, guava, and citrus fruits, having an area of more than two hectares to accommodate the replicated attractant devices. The details of trapping sites are presented in Table 2.

2.2. Sampling Techniques

Two methods of sampling were employed to collect the species of fruit flies, rearing of larvae to the adult stage from infested mango, guava, and sweet orange fruits and trapping of adult fruit flies using food-based attractants and male lures.

2.3. Collection of Fruits

Samples of infested ripe fruit of 75.4 kg of mango, 212.6 kg of guava, and 8.27 kg of sweet orange were collected weekly for four consecutive weeks. Single samples ranged from 0.5 to 3.6 kg, 2.1 to 7.6 kg, and 0.01 to 1 kg for mango, guava, and sweet orange, respectively. Fruit collection was conducted during peak maturing periods of fruit crops, between July and December 2018: July and August for mango and October-November for guava and sweet orange. The sample size varies among crops and study areas due to the area coverage differences in the study area.

2.4. Rearing of Fruit Flies from Infested Fruits

The matured fruits collected purposively having confirmed oviposition puncture were weighed, counted, and placed in plastic containers with a net lid and sterilized sand at the bottom and then incubated for about six weeks. Fruits were placed in containers in groups of 5–20 depending on the fruit and container size.

The fruits in the containers were checked in the three-day interval for adult flies. Emerged adults were provided with honey into the roof of the cage and water (cotton wool soaked with water) on the floor of the cage for feeding. Emerged adults were left in the cage for 3–5 days for growth and full development of morphological development. The adult fruit flies that emerged from different fruits was counted and preserved in 70% alcohol for identification [7]. Identification was performed at the Laboratory of Entomology at the Sirinka Agricultural Research Center with the help of a guide book [7], and for confirmation, voucher specimens were sent to the International Center of Insect Physiology and Ecology (ICIPE), Addis Ababa, Ethiopia.

2.5. Collection of Fruit Flies Using Different Attractants

Three different male lures, namely, methyl eugenol (ME), trimedlure (TML), and terpinyl acetate (TA), and two food-based attractants, torula yeast and protein hydrolase, have been used to attract fruit flies, since different attractants can be combined to reach the possible higher number of fruit flies [20]. Finally, the specialized male lures were evaluated with food-based attractants. ME for B. dorsalis and TML and TA for Ceratitis species were evaluated along with torula yeast and protein hydrolase. For further details of the attractant’s longevity and specificity, see Table 3.

Empty water bottles of one-liter capacity were used for making a modified trap by cutting the bottle at the 2/3 level; the cutoff neck served as an entry funnel into the rest of the bottle; then, the modified trap was hanged with a binding wire by penetrating at the bottom of the bottle. Baits with polymeric plug formulation were smeared in the inner side of the opened bottle, whereas liquid-formulated attractants were placed in the bottle using a piece of cotton wick. Locking from the opened side of the bottle with the binding wire by dipping in a mixture of carbaryl (Sevin 85% WP) in a 1 to 4 ratio with attractants was used to kill the fly in the bottle. The trap bottles with the male lures and food-based attractants were suspended with a binding wire on mango trees at a height of 1.5–2 meters from the ground for eight consecutive weeks (1 July to 26 August 2018). Each lure was replicated three times in each trap area with a minimum spacing of 20 meters among and 50 meters with attractant types; there were three traps without any attractant with carbaryl (Sevin 85% WP) as control at each trap sites (totally, 18 traps per site including the control treatments); the trap sites were 60–180 kilometers apart from each other.

New water bottle traps were used to avoid contamination of the outer surface of the bottles with the baits, which may keep the flies to settle to the outer side of the bottle instead of getting in. The trap-holding wires were smeared with grease to prevent the entry of ants. Renewing of the attractants was employed in a weekly interval for food-based attractants (torula yeast and protein hydrolase) and monthly for male lures (ME, TA, and TML). After each inspection, there was a clockwise rotation of traps; inspection was employed in a 7 days interval [20]. The trapped flies were preserved in vials containing ethanol 70% for further identification. The samples were identified at the Laboratory of Entomology at the Sirinka Agricultural Research Center with the help of color print guide books [7, 22]; for confirmation, voucher specimens were sent to the International Center of Insect Physiology and Ecology (ICIPE), Addis Ababa, Ethiopia.

Meteorological parameters like maximum and minimum temperature, solar radiation, and rainfall data were taken from the MARKSIM DSSAT weather file generator.

The relative abundance of each fruit fly species was estimated as the proportion of the total adult of each fruit fly species and the total adult collected of all fruit fly species:where Isi is the total number of individual species and ∑ Nsi is the total number of the species population.

Flies per trap per week was estimated with the following daily formula by contextualizing to the week basis:where F is the total number of fruit flies, T is the number of inspected traps, and D is the average number of days traps were exposed in the field [21].

2.6. Data Analysis

Data for all insect counts were converted to the number of flies per trap per week and were subjected to ANOVA using the generalized linear model (PROC GLM, SAS Institute). ANOVAs were performed on catch data for all species, considering locality and attractants as main factors. The data were transformed using procedure Log10 () for B. dorsalis and for C. capitata and C. cosyra species. Means were separated by the Student–Newman–Keuls Test when ANOVA was significant at .

3. Results

3.1. Fruit Fly Species and Their Abundance

A total of 7064 adult fruit flies have emerged from the collected three fruit hosts’ species. Three fruit fly species were recovered, namely, B. dorsalis, C. cosyra, and C. capitata. Bactrocera dorsalis was the most abundant species and accounted for 78.89% of the total emerged adults; C. capitata contributed 17.19% of the emerged adult fruit flies, whereas C. cosyra had the lowest proportion of emerged adults from the infested fruits collected (3.92%) (Figure 1).

Bactrocera dorsalis was also the dominant species from the trapped adults that accounted for 97.9%; Ceratitis fasciventris was observed only from the trap with the lowest proportion of trapped adults (0.03%) (Figure 1).

All of the observed fruit fly species emerged from guava and mango; one fruit fly species (B. dorsalis) emerged from sweet orange. Only B. dorsalis emerged from all observed host fruits, indicating its polyphagous nature. Ceratitis fasciventris was not reared from the collected host fruits, but it was found only from the trap with the lowest proportion of trapped adults. However, this species was reported as dominant on mango in eastern Ethiopia [14].

Bactrocera dorsalis, C. cosyra, and C. capitata represented the 71.95%, 56.68%, and 83.36%, respectively, of emerged adults from guava fruits and 27.9, 43.32, and 16.64%, respectively, from mango fruits (Figure 2). On the other hand, sweet orange contributed 0.14% adult emergence only for B. dorsalis. The three evaluated hosts were observed to have a simultaneous occurrence of B. dorsalis and Ceratitis species. As shown in Figure 2, B. dorsalis shared mango and guava with C. capitata and C. cosyra; on sweet orange, B. dorsalis was the only species reared.

The invasive species Bactrocera dorsalis dominated the indigenous species of fruit flies in the current study which is supported by the earlier reports in different parts of Ethiopia, and it has established in some parts of eastern Ethiopia [14]. Bactrocera dorsalis is the dominant species reared from mango fruits, followed by C. fasciventris and C. cosyra in western Ethiopia [15]; also in southwestern Ethiopia, it was the predominant species, with 96% of captures and the only fruit fly species emerging from mango fruits incubated [16]. In other African countries, B. dorsalis was reported as the most prevalent species (98%) in Uganda on mango [23]. In Kenya, it displaced C. cosyra on mango within 4 years from invasion, constituting 98 and 88% of the total population in traps and mango fruit at Nguruman, respectively [24].

The polyphagous nature of the species had been reported in different countries out of its origin. In Tanzania, suitable hosts are available throughout the year for B. dorsalis. Mango appears to be the principal host in the period October to January, while guava is the principal host in the period February to August [6]. However, mango is an important host for C. cosyra during the Mango fruiting season [6]. In Kenya, before the arrival of B. dorsalis, the indigenous fruit fly species C. cosyra was the predominant fruit fly pest of mango [24]. Bactrocera dorsalis is also present in indigenous coastal forests in Kenya and is capable of reproducing in wild fruits; hence, a sufficient reproductive base exists even when mangoes are not fruiting [25]. From the survey conducted in Peninsular Malaysia to determine the infestation levels of fruit flies, B. dorsalis was recorded as the most abundant species (91%), compared to Bactrocera carambolae Drew and Hancock (9%) [26].

In this study, C. capitata was mostly (83.36%) reared from guava fruit, and a significantly small amount (16.64%) was reared from mango fruit (Figure 2). In addition, this species contributed a very small amount of the trapped adult fruit flies (1.2%) (Figure 1), whereas significantly more specimens were reared from fruits (17.19%) (Figure 2). This is probably due to the time of occurrence and trapping time difference, since the trap was left in the field during July and August, which is before the maturity of guava, whereas it is the time for the peak maturing of mango in the study areas, especially late-maturing improved varieties. It may be due to displacement of the indigenous species on mango, as Vargas et al. [27] reported that, when B. dorsalis became established in Hawaii, it displaced C. capitata in many hosts and lowland habitats.

A significantly higher proportion of C. cosyra (43.32%) were reared on mango fruit than C. capitata and B. dorsalis (Figure 2). This result is supported by Ekesi et al. [24], who stated that C. cosyra has not been completely displaced in mango orchards at Nguruman. There are probably some advantages this species allows for some level of coexistence with B. dorsalis. One advantage may be its more specialized host-searching abilities on mangoes have been linked more closely to this host plant over a long period in Africa. Also, the same author reported that B. dorsalis frequently shared the same fruits with the indigenous fruit fly species C. cosyra but often occurred at higher numbers than C. cosyra [28]. Ceratitis cosyra dominated in most wild fruits [8].

The economic importance of C. cosyra has been growing since the widespread commercialization of mango and the introduction of exotic mango varieties. Late ripening mango varieties are the ones that suffer the most, due to C. cosyra infestation. Ceratitis cosyra is widespread in sub-Saharan Africa [2] and the dominant one in Abugubeiha area, Sudan [29]. This probably is the other reason for the rearing of a proportionally higher number of C. cosyra on mango fruits; i.e., the study covered the peak harvesting period of late-maturing improved mango varieties.

In the current study area, from sweet orange, the only emerged fly was B. dorsalis, meaning that it completely displaced other fruit flies, as previously reported in Tanzania, in the some host C. sinensis [6].

3.2. Fruit fly Species Collected through Traps

Overall, 27641 (27060 B. dorsalis, 231 C. cosyra, 342 C. capitata, and 8 C. fasciventris) specimens were collected. The total number of specimens was higher in male lure traps (97.03%) than in food-based attractants (2.97%) (Table 4). Methyl eugenol caught 97.5% B. dorsalis, and the remaining 1.54 and 0.91% of specimens were collected with protein hydrolase and torula yeast, respectively (Table 4). TML captured 78.35% of C. cosyra, followed by TA with 12.56%, and the remaining 5.63 and 3.46% were collected by torula yeast and protein hydrolase, respectively.

3.2.1. Bactrocera dorsalis

Analysis of variance showed that there is a significant difference between treatments (attractants) (F = 265.04; DF = 3; ) for B. dorsalis (Table 5). The highest number of B. dorsalis per trap per week was trapped at Kalu with methyl eugenol (722.2 flies trapped per trap per week (FTW)) followed by the same attractant at Kobo (376.5) and Habru (90) (Table 5). Both food-based attractants, torula yeast and protein hydrolase, trapped much lower B. dorsalis than the male lure (ME), statistically in par each other (Table 5). According to the earlier reports, torula yeast would be a better monitoring tool than other attractants for B. dorsalis like biolure [30], even more effective than hydrolyzed proteins over time [21].

For B. dorsalis, the effects of attractants were consistent between locations, with ME attracting a significantly higher number of adults. Methyl eugenol is a natural phenylpropanoid highly attractive to B. dorsalis males [31]. Males of the B. dorsalis are strongly attracted to ME [32]. This research also confirmed its high attractiveness.

The number of B. dorsalis per trap per week was shown to be significantly different among trapping locations (F = 59.27; DF = 2; ), where the overall mean of adult B. dorsalis per trap per week was significantly higher at Kalu (184.4) which is eight times as high as at Habru (23.6) and two times as high as at Kobo (96.2) (Table 5); this is probably due to the year round availability of diverse reservoir alternative host fruit species and varieties at and around the trap site at Kalu (Table 2). Also, since the Kalu trap site is a governmental nursery and semifruit production site, fruits set by mother plants of different varieties escaping from the purposive abortion of flowers for the preparation of scion may be left on the field; also, it is the most probable entry point of the exotic species (B. dorsalis) into the region, since they use seed sources for rootstock preparation from southern Ethiopia where the species was reported earlier [15, 16].

3.2.2. Ceratitis Species

There were significant differences in the number of C. capitata between attractants (F = 53.62; DF = 4; ); the highest number of C. capitata was trapped by TML at Habru (4.83 adult flies per trap per week); this lure also trapped a significantly higher number of flies at Kalu (2.79) and Kobo (1.88), while the other male lure TA recorded par with the food-based attractants except at Habru (1.29) (Table 6). This result is analogous with some previous reports as trimedlure is an effective lure for surveying and monitoring activities of C. capitata males [33]. Also, other Ceratitis species are known to be attracted to it [34].

There was also a significant difference for C. capitata among the trap sites (F = 15.04; DF = 2; ). The highest number of C. capitata was trapped at Habru (1.45) followed by Kalu (0.66) and Kobo (0.42) (Table 6). The study was conducted in the area where C. capitata and C. fasciventris were previously reported as dominant species [18]. The current result showed the displacement of the indigenous Ceratitis species by the invasive ones (B. dorsalis), especially from the lowland habitats, since the Habru site was relatively at higher altitude. This finding is in line with analogous results reported by Ekesi et al. [28], who reported a significant inverse relationship between numbers of flies and elevation. On the other hand, C. capitata persists from the displacement by the invasive species (B. dorsalis) in lowlands on coffee [24]. This is probably the other reason for the relatively higher C. capitata catches at Habru, which is a known coffee production area in the region, and may explain the coexistence of C. capitata with the dominant species B. dorsalis.

The number of C. fasciventris is shown to be nonsignificant among trapping sites (F = 0.46; DF = 2; ) and attractants (F = 3.69; DF = 4; ) (Table 6). Even the overall number of C. fasciventris is shown to be negligible, from 0.05 flies per trap per week at Habru to 0.13 flies at Kalu and Kobo. The limited number of C. fasciventris was trapped with torula yeast and protein hydrolase-baited traps in all sites. Unlike other fruit fly species trapped during this study, C. fasciventris was not found in collected fruits; however, this species was reported as a dominant species especially on citrus (Fardu et al. 2009).

The weekly trapped C. cosyra showed a significant difference for both attractants (F = 7.90; DF = 4; ) and trapping sites (F = 5.71; DF = 2; ) (Table 6). The highest number of C. cosyra was trapped by protein hydrolase at Kobo (7.25 flies per trap per week), followed by the same attractant at Habru (0.54) and torula yeast at Kalu (0.38). However, most of the attractants were not statistically significant even with the control traps, except protein hydrolase at Kobo (Table 5). Among the trap sites, C. cosyra was higher at Kobo (1.49). This site was mango dominated, as the whole Kobo district, following the Kobo-Girana valley irrigation scheme, is a major mango producer. Ceratitis cosyra, widespread in sub-Saharan Africa [2], has been linked more closely to mango over a long period in Africa and has more specialized host-searching abilities on mangoes [24].

Food lures are generic by nature, and besides the target fruit fly species, traps tend to catch a wide range of other tephritid and nontephritid flies [21]. Both PH and TY had trapped the species detected in all trapping sites through male lures in a comparative amount for Ceratitis species, and this may also be considered as an additional advantage for the survey and monitoring purposes.

4. Conclusion

Monitoring of Tephritidae fruit flies in eastern Amhara revealed the existence of four fruit fly species under the genera Ceratitis and Bactrocera. These are C. capitata, C. cosyra, C. fasciventris, and B. dorsalis. The last one is the dominant species from the trapped and emerged adults in the study area. From the indigenous species, C. capitata was commonly trapped by TML-baited traps at Habru, which is a guava and coffee dominated and relatively higher altitude trap site. The higher proportion of C. cosyra was reared on mango and trapped at Kobo, which is a mango-dominated site, mainly with protein hydrolase-baited traps. Ceratitis fasciventris was observed only from the traps with the lowest proportion. All the observed fruit fly species emerged from guava and mango, from sweet orange B. dorsalis only. More adults were trapped in male lure traps (97.03%) than food-based attractants (2.97%); hence, fruit fly management should focus on guava and mango crops which host all the fruit fly species, including the dominant B. dorsalis.

Data Availability

The data for this paper are available from the corresponding author upon request.

Disclosure

This paper is part of MSc thesis presented at Bahir Dar University.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Authors’ Contributions

A.M. and Y.W. were responsible for conceptualization. A.M., M.A., and Y.W. were responsible for methodology. A.M. was responsible for formal analysis. A.M. and M.A. were responsible for investigation. A.M. was responsible for writing the original draft. A.M., M.A., and Y.W. were responsible for writing, reviewing, and editing.

Acknowledgments

The authors would like to thank the Amhara Region Agricultural Research Institute (ARARI) for offering the first author the opportunity to pursue this study and covered the expenses incurred for the research. Grateful acknowledgements are extended to ICIPE Ethiopia staff Dr. Tadele Tefera, Dr. Shifa Ballo, Dr. Girma Demssie, and Mr. Alemayehu for helping us for species identification. This research was financially supported by the Amhara Agricultural Research Institute (ARARI).