Introduction

International trade in fruit and vegetables continues to increase to feed an increasing world population (Bebber et al. 2014). Additionally, tourism is essential to the world’s economy. However, both trade and tourism are implicated in the dispersal of exotic species including fruit flies (Early et al. 2016; Robinson and McNeill 2022). Incursions by exotic fruit flies result in restrictions on trade and loss of productivity (Drew 1997; Ormsby 2021).

More than 300 species of fruit fly occur in Australia although only a few cause any economic impact (Plant Health Australia 2018). For Australia, the main exotic fruit fly threat comes from Asia because New Zealand and Antarctica are fruit fly free (MacLellan et al. 2021). Therefore, Australian surveillance is conducted primarily for exotic Asian tephritids. Fortunately, most endemic and exotic male tephritids are easily monitored by male-lure traps (Suckling et al. 2016).

Frequently, tephritid invasions follow a jump-diffusion model (Sadler et al. 2011) where tourism or trade provide the long-distance jump, followed by diffusion or dispersal into the local environment. Bactrocera dorsalis (Hendel) (Oriental fruit fly (OFF)) is a significant concern because it has a long history of successful invasions. In 1935, B. dorsalis invaded the Mariana Islands but was deemed eradicated by 1965 (Leblanc et al. 2013a). Bactrocera dorsalis invaded Hawaii (1946), Nauru (1992 but eradicated in 1999), and French Polynesia (1996) where it outcompeted Bactrocera tryoni (Froggatt) (Queensland fruit fly (Qfly) and became the dominant pest (Leblanc et al. 2013a). Recently, Bactrocera papayae (Drew and Hancock) and B. philippinensis (Drew and Hancock) were synonymised to B. dorsalis (Schutze et al. 2015) and this synonymisation added to the invasion history of B. dorsalis.

Bactrocera papayae was detected in Papua New Guinea in 1993 and Palau in 1996 (Leblanc et al. 2013a). In October 1995, the exotic B. papayae was detected in papaya near Cairns in northern Queensland (Fay et al. 1997; Gillespie 2003). Shortly afterwards in November 1997, B. philippinensis was detected in Northern Territory, but both Australian incursions were eradicated (Hancock 2013). Given the invasion history of B. dorsalis, another incursion is the primary driver for continued surveillance in all first ports of entry (Gillespie 2003; Wylie et al. 2008; Dominiak 2021a).

The potential for an invasion by B. dorsalis into northern Australia is high due to its proximity to Asia. Therefore, many exotic fruit fly traps are on islands in the Torres Strait, some of which are only 4–6 km from Papua New Guinea. These traps act as an early warning system and exotic fruit flies are eradicated annually (Huxham 2002; Murphy 2022). Australia’s preparedness of exotic incursion, particularly B. dorsalis, was reviewed by Hoskins et al. (2023).

For eastern Australia, another threat is the Mediterranean fruit fly (Medfly), Ceratitis capitata (Wiedemann), from Western Australia. One advantage of B. tryoni being endemic is that there is a very low likelihood that Medfly will establish in a B. tryoni endemic area (Dominiak and Mapson 2017), as Medfly was previously displaced from eastern Australia in the 1940s (Dominiak and Mapson 2017). There is a perception that Qfly will outcompete Medfly and that Medfly will not establish where there is a resident Qfly population (Dominiak and Mapson 2017). The HRN comparison between Qfly and Medfly will be the subject of another review.

Similar to C. capitata, if B. dorsalis did enter and attempt to establish again on mainland Australia, it would have to compete with the established fruit flies, primarily B. tryoni. In Queensland, other tephritids include B. aquilonis (May), B. frauenfeldi (Schiner), B. jarvisi (Tryon), B. musae (Tryon), B. neohumeralis (Hardy), and Zeugodacus cucumis (French) (Hancock et al. 2000). However, different hosts are known to variously support the full life cycle of different tephritid species (Woods et al. 2005; Lloyd et al. 2013; Dominiak et al. 2020; Follett et al. 2021). These papers report on the number of adults that can emerge from one kg of fruit. Subsequently, this metric was termed the Host Reproduction Number (HRN) (Dominiak 2022). The Host Suitability Index (HSI) (Follett et al. 2021) is a way of placing HRN into six major categories.

Regarding exotic invasions, Ekesi et al. (2009) noted that tephritids with a higher reproductive capacity (or HRN) were likely to eventually displace species with lower reproductive capacity. I wanted to test this HRN theory in northern Australia where B. dorsalis may make an incursion into areas where B. tryoni is the main tephritid. Theoretically based on HRN alone, B. dorsalis is more likely to outcompete B. tryoni in hosts that are poor hosts for B. tryoni. Conversely, B. dorsalis is less likely to establish in hosts that support large numbers of B. tryoni. Here, I wanted to use HRN theory to determine whether B. dorsalis and B. tryoni hosts more or less likely to support an incursion by B. dorsalis. The competitive advantage of B. dorsalis is compared to host records from its previous establishment in the Cairns region of north Queensland.

Materials and methods

I followed the methodology described by Dominiak (2022). In brief, I used Google Scholar as a search engine because it yields more results than other databases (Pozsgai et al. 2021). In the first search, I used “Bactrocera dorsalis” as the main search term. Additional terms were added, such as “host”, “suitability”, “susceptibility” or “fruit” and similar terms in successive searches. I examined the results and chose references if they provided useful data for HRN. Usually, useful references were contained in the first four pages of each search. The third search word was changed, based on the results of earlier searches. I repeated these procedures using “Bactrocera tryoni” as the main search term.

To formulate Table 1 containing hosts, HRN and HSI, I examined each reference for data of each host regarding the reproductive capacity to support adult fruit flies. Where possible, all fruit infestation results are standardized on an individual’s HRN. Most data were based on field sampling. Unfortunately, not all papers reported the information required. Some papers reported infestations per fruit and provided insufficient information to calculate the adults/kg metric. Data in some papers reported infestation rates in graphs or figures and accurate interpretation was too difficult to obtain exact figures. Other papers reported infestations of larvae or pupae per kg and this information is not reported here. Different papers report a range of HRN figures. In plant biosecurity, the worst-case scenario is always assumed to be the case. Therefore, I did not include any references that reported a HRN lower than the eventual highest HRN in Table 1. Within Table 1, the HRN is followed by the equivalent index (HSI) proposed by Follett et al. (2021).

Table 1 Bactrocera hosts with the adults/kg numbers (Host Reproduction Number (HRN)) and Host Suitability Index (HSI) in brackets. The potential reproductive advantage of Bactrocera dorsalis over Bactrocera tryoni or visa-versa is calculated by dividing the larger HRN by the smaller HRN
Full size table

I reviewed the papers for B. dorsalis and B. tryoni that contained HRN data. Hosts are listed in alphabetical order in Table 1. The reproductive advantage of B. dorsalis over B. tryoni is calculated by dividing the higher HRN of fruit for B. dorsalis by the lower HRN for B. tryoni. The reverse calculation was done where the HRN for B. tryoni was larger than the HRN for B. dorsalis. In Table 2, these calculated figures are compared to the frequency of fruit samples infested with B. dorsalis from November 1995 to December 1997 (Anon 1998). There were 229 detections of B. dorsalis across 33 hosts with the frequency ranging from one (ten different hosts) to 68 (mangoes).

Table 2 Comparison between the frequency of infested fruit (B. dorsalis) samples in 1995–1997 incursion and reproductive advantage
Full size table

In Table 2, HRN and therefore reproductive advantage was not available for all B. dorsalis hosts noted by Fay et al. (1997). Additionally, the HSI (Follett et al. 2021) for both species is provided as a further comparison. These HSI were very poor (< 0.1 adults per kg), poor (0.1–1.0), moderately good (1.0–10.0), good (10–100) and very good (> 100). Table 3 provides QDAF records of fruit collections from 1995 to 1997 in north Queensland and averages the number of fruit fly specimens per collection.

Table 3 Records from QDAF Insect Collection for specimens of B. dorsalis (papayae) and B. tryoni collected in far north Queensland during the OFF campaign 1995–1997
Full size table

Results

I found data for 39 comparisons between B. dorsalis and B. tryoni. In the reproductive advantage assessment for “B. dorsalis over B. tryoni” column (Table 1), 20 hosts scored a reproductive advantage for B. dorsalis over B. tryoni but 13 hosts with an advantage > 2. There were three hosts with HRN > 30 and no host with a HRN of < 30 and > 10. The advantage scores were quite polarised. There were 14 hosts in the “B. tryoni over B. dorsalis” column which shows the reproductive advantage for B. tryoni with 12 hosts with a score of > 2.

Using the calculated reproductive advantage, B. dorsalis is likely to have dominance over B. tryoni in soursop (45.4), mango (33.7), capsicum (31.1), purple passion fruit (9.9), Tahitian chestnut (8.7), papaya (8.4), cashew (6.7), lemon (6.4), rambutan (6.2), banana (5.9), lychee (5.2), grapefruit (2.4), Tahitian lime (2.1) with eight other fruits with < 2 advantage. Soursop, mango and capsicum appear superficially to be better candidate fruits for B. dorsalis surveillance.

Conversely, B. tryoni dominated B. dorsalis where the reproductive advantage favoured B. tryoni, for example, in kumquat (33.9), avocado (18.2), mandarin (15.9), oranges (12.9), yellow passion fruit (10.8), breadfruit (10), giant passion fruit (9.5), custard apple (8.3), star apple (5.1), pomelo (4.6), star fruit (2.5), Surinam cherry (2.3), with five other fruits with < 2 advantage. I predict that kumquats and avocados would be the poorest candidates for B. dorsalis surveillance.

Table 2 contains calculated reproductive advantage from Table 1 and the frequency of fruit infested with B. dorsalis detected in the previous Queensland B. dorsalis eradication program. Additionally, the HSI comparison is provided for both tephritids. Mangoes and capsicums are in the top four rankings. Table 3 provides the number of collections, the total number of specimens recorded and the average number of specimens per collection. Capsicums were ranked higher than mangoes in Table 3.

Discussion

In Tahiti, Bactrocera kirki (Froggatt) invaded in 1928 and became established (Leblanc et al. 2013a). Subsequently, B. tryoni was introduced in 1970 and displaced B. kirki in many areas. In 1996, B. dorsalis was detected and established, becoming the dominant fruit fly, displacing both B. kirki and B. tryoni (Leblanc et al. 2013a). The competitive displacement was attributed to several factors including a high reproductive rate (HRN) and increased longevity under mild tropical temperatures (Vargas et al. 1997; Leblanc et al. 2013a). Adult B. dorsalis has aggressive behaviour and aggressive larval competition (Ekesi et al. 2009; Leblanc et al. 2013a) and very broad host range (Follett et al. 2021; Dominiak 2022). Additionally, B. dorsalis had a developmental advantage in preferred hosts such as P. guajava (Vargas et al. 2012; Leblanc et al. 2013b). Another competitive advantage is that B. dorsalis attacks fruit at an earlier stage than many other tephritids.

For biosecurity, the understanding of host suitability and reproductive capacity is fundamental information to underpin emergency response plans and effective eradication programs for exotic tephritids. This information facilitates the targeting of the most reproductive hosts, as was achieved in Nauru (Allwood et al. 2002) and the Cook Islands (Allwood 2002). Similarly, these data will assist importing countries in the risk assessment associated with the potential importation of infested produce (Leblanc et al. 2013a).

In my review, some HRNs were not available for all fruits in this assessment. However, there are some consistencies between this HRN assessment and the frequency of fruit samples infested with B. dorsalis in the 1995–1997 data (Table 2). Firstly, there are ten hosts listed in Table 2. Of the five fruit most infested by B. dorsalis (mango, tropical almond, guava, capsicum, papaya in that order in Table 2), all five hosts had a calculated reproductive advantage for B. dorsalis. Of the five hosts less frequently infested by B. dorsalis, three hosts (tomato, grapefruit, star fruit in Table 2) had a calculated competitive advantage favouring B. tryoni.

Mangoes were the highest competitive ranking (33.7) and were most frequently infested (68). Tropical almond was the second most frequently infested host (28). Jose et al. (2013) considered that tropical almond was an important reservoir for B. dorsalis in Africa. Guava was the third most frequently infested host (20) and is a highly suitable fruit for many tephritids (Dominiak 2021b). Guava and tropical almond were both “very good” hosts (Follett et al. 2019) and the competitive advantage was < 2.5 for both hosts. Perhaps, both tephritid species could coexist in these hosts given the low reproductive advantage. Capsicums were the fourth most frequently infested host (13) and had a competitive advantage of 31.1 favouring B. dorsalis. These data need to be balanced with host availability in Cairns. Mangoes, guava and tropical almonds were widely available and provided a huge resource for both species. This availability may have negated any advantage for both tephritids. Overall, capsicums were well ranked in all three tables and offer a good candidate for B. dorsalis surveillance, followed by mangoes. Because of their size, fruiting capsicum plants could be used as sentinel plants in high risk areas (such as first ports of entry) to optimise early detection of any incursion in most parts of Australia. This is because fresh fruiting capsicum plants could be provided on a regular schedule unlike fruiting mango plants which cannot be put in pots due to their size and weight. Additionally, moving fruiting mango plants in pots offer larger challenges due to their size and weight.

Based on initial data easily available, I infer that reproductive advantage scores of 5 or less indicate a minimal competitive advantage for both tephritid species. If this was the case, I propose that pomelo (4.6), star fruit (2.5), grapefruit (2.4), Surinam cherry (2.3), Tahitian lime (2.1), rose-apple (1.9), Pacific lychee (1.8), strawberry guava (1.8), tropical almond (1.7), Jew plum (1.7), guava (1.6), abiu (1.6), Indian jujube (1.6), tomato (1.2), loquat (1), and Malay apple (1) were about equal for both B. dorsalis and B. tryoni and these tephritids may co-exist in these hosts. Additionally, any host with a MG ranking (kumquat), P ranking (pomegranate) or NH (finger lime) is unlikely to support an incursion by B. dorsalis. Some of these hosts are unlikely to be grown in northern Queensland.

There may be some support for the role of endemic tephritids in minimizing the chances of successful establishment by B. dorsalis. Fay (2008) reported that OFF may have arrived 12 months before detection. Furthermore, Cantrell et al. (2002) suggested introduction occurred 2.5 years before detection. It might be inferred that the endemic tephritids caused challenges for the establishment of OFF. Kay (2023) found that B. tryoni and B. neohumeralis exhibited similar aggressive behaviour of B. dorsalis on fruit. These behaviours (including crabbing, wing supination, and pushing (Kay 2023)) may partially explain the delay between incursion and detection at Cairns.

In the 1995 incursion, B. dorsalis was originally detected in papaya and the local grower, John Crawford, observed OFF attacking papaya earlier than local tephritids (Fay et al. 1997). Follett et al. (2019) reported a field HRN (142) of papaya for B. dorsalis, while a HRN of 500 can be reached in screen cage tests (Follett et al. 2019). The HRN for B. tryoni in papaya was 17 so perhaps the establishment of B. dorsalis in papaya was to be expected. The reproductive advantage for papaya was comparatively low at 8.4 favouring B. dorsalis.

Reitz and Trumble (2002) noted that B. dorsalis has a higher net reproductive capacity on certain hosts. However, this competition is host-mediated and the less competitive tephritids persisted in hosts where they were better adapted. In my assessment, there are eight examples where the competitive advantage was > 10 for either species. It is possible that the two tephritids may coexist, but one may evolve superior competitive abilities and eventually displace the less competitive tephritid species over time (Reitz and Trumble 2002).

Another reason for inconsistencies may be the altitude where data was collected. Bactrocera dorsalis is reputed to be a lowland fly (Ekesi et al. 2009; Reitz and Trumble 2002). My HRN comparison ignores aspects such as earlier utilisation associated with B. dorsalis. In my literature review, papers provided HRN data but rarely gave altitude data and, therefore, HRN figures used here may not all be on the same basis (altitude). Also, pesticide use could skew the eradication data. For example, the use of pesticides and sampling of some hosts such as Terminalia spp. and feral guavas as reported by Fay et al. (1997) could have skewed Australian invasion and eradication data. During the B. dorsalis eradication, bait spraying was targeted at specific hosts such as mangoes and guavas, and then coffee where later B. dorsalis was found breeding. Additionally, usual pesticide applications in commercial crops influence infestation success by the endemic and invader tephritid. Some crops are known to be very suitable to B. tryoni and be subject to pesticide programs which might equally disadvantage B. dorsalis. Conversely, coffee has few insect problems and receives few pre-harvest insecticides (Fay et al. 1997). There were seven examples of infested Coffea arabica but no HRN recorded. In Africa, Coffea canephora has a HRN of 370 for B. dorsalis (Mwatawala et al. 2009).

Ekesi et al. (2009) noted several mechanisms for competitive displacement between tephritids. Rapid population growth (based on HRN) resulted in pre-emptive use of resources, particularly during egg laying and larval feeding. Bactrocera dorsalis is very mobile and has a high dispersive power compared to Ceratitis, however, this advantage may not be the case when compared to another Bactrocera such as B. tryoni. Therefore, rapid displacement of other species, as occurred in Africa, may not occur in Australia where many Bactrocera spp. are resident (Hancock et al. 2000). Bactrocera tryoni and B. neohumeralis exhibit similar aggressive behaviour on fruit compared to B. dorsalis (Kay 2023) with seven identified behaviours in competitive contests. The range of B. tryoni extends along the entire eastern seaboard of Australia, including three capital cities and likely first ports of entry (Wylie et al. 2008; Dominiak and Mapson 2017). The sister species, B. aquilonis (May) is distributed across northern parts of Western Australia, Northern Territory and Queensland (Cameron et al. 2010). The range of B. neohumeralis and B. bryoniae extends from northern Queensland down to Sydney, the largest city in Australia (Dominiak and Worsley 2016; Dominiak and Millynn 2022). From the results presented here, Capsicum annum, which is a preferred host of B. bryoniae (Dominiak and Millynn 2022) and also is a preferred host of B. dorsalis (this paper).

Dominiak and Worsley (2017) reported the presence of B. jarvisi in northern NSW while Z. cucumis was found just in the south of the Queensland border (Fay et al. 2022). Some, if not all, of these species may exhibit the aggressive behaviour to each other (Kay 2023), and to B. dorsalis. Therefore, it is likely that these behaviours, and HRN, are part of the competition that will contribute to any possible establishment success of B. dorsalis in Australia (Clarke and Measham 2022).

This comparison has some limitations. It is a comparison between two species based only on HRN. There are other factors including host availability, ripeness or other physical factors which may influence competition between the fruit flies. In biosecurity, the worst-case scenario is always assumed to be the case, irrespective of where the HRN originated. I used the highest HRN figures from all papers. However, some values originate from different continents and the comparison may not be fair. The issues of altitude and climate may influence a fair comparison. Latitudes may be important as a component of climate, or proximity to the equator and this could be an indicator of tolerance to heat or cold. Hassani et al. (2022) found B. dorsalis had a distinct preference for lowland (0-300 m ASL) and gradually decreased up to 800 m ASL. Therefore, it is likely that reproductive advantage might be higher at lower altitudes and disfavour B. tryoni. Additionally, soils may influence survival. Bactrocera tryoni are known to survive relatively well in wet or saturated soils (Hulthen and Clarke 2006; Dominiak et al. 2011) whereas B. dorsalis has poor survival in wet soils (Hou et al. 2006). It is likely that B. tryoni may dominate B. dorsalis at the bottom of slopes in higher rainfall areas, irrespective of HRN.

Therefore, the direct HRN comparison and the calculated competitive advantage may be the worst-case scenario and not warranted in all cases. For instance, for B. tryoni in Valencia oranges, Lloyd et al. (2013) found a HRN of 10 in central Queensland whereas Dominiak et al. (2020) reported a HRN of 142 in southern New South Wales, 1120 km further south. This difference may reflect different climate or Qfly population pressure and this difference demonstrates that one HRN does not fit all scenarios. However, for a B. dorsalis incursion into Queensland, Lloyd et al. (2013) might be more locally appropriate; if so, the reproductive advantage would change from 12.9 to B. tryoni to about one or neutral for both species. Follett et al. (2021) discussed some of the issues influencing host suitability. The use of a wide range of data from different studies and localities is recognised and problematic. However, there is no option when attempting to quantify the invasion threat of an exotic species, given the currently availability of data.

Notwithstanding, this was an assessment to demonstrate how HRN and reproductive advantage could be used to better target vulnerable industries and to identify where surveillance might be best placed. The threat of incursion exists for all countries due to increased trade in fruit and vegetables (Bebber et al. 2014) which are further compounded by international movements of tourists (Robinson and McNeill 2022). For instance, there were 15 documented incursions of fruit flies in New Zealand since the early 20th century (Kean 2017), two of which were of B. dorsalis (Hancock 2013). Although they were eradicated, these incursions may occur in future. Therefore, trap surveillance for exotic fruit flies such as B. dorsalis is key for their early detection in fruit fly-free countries. Hosts with higher competitive advantage may be better options for surveillance sites.

The first ports of entry of fruit fly host materials may signal an obvious early warning of fruit fly incursions (Gillespie 2003; Wylie et al. 2008; Augustin et al. 2012). There was no trapping surveillance at Cairns at the time of the B. dorsalis incursion in 1997 (Fay et al. 1997; Gillespie 2003), however the detection in fruit prompted an increased fruit fly trapping program in Australia. In Australia, B. dorsalis was not detected in these ports and therefore, a second tier of surveillance in vulnerable crops may be a valuable support to the first ports of entry. HRN could be used in ports, market access, and urban areas, which mostly offer favourable sites for many tephritids because of the increased temperature and moisture regimes (Dominiak et al. 2006; Raghu et al. 2000) and a continuum of hosts. In these areas, hosts are grown opportunistically and are usually poorly managed, which may result in secondary spread of tephritids into nearby commercial orchards (Dominiak et al. 1998).

Furthermore, HRN should be integrated in the System Approach of pest management (Dominiak 2019; van Klinken et al. 2020) and advising the appropriate disinfestation treatment required (Balagawi et al. 2021).

Regarding the possible introduction of infested fruits by air passengers, HRN could be used to inform targeted surveillance and education at airports. Fruit fly host materials are commonly carried by international tourists. Of the domestic flights in Papua New Guinea during February and March 2000, 38.9% of passengers carried fresh commodities, including 34 commodities that were known fruit fly hosts. About 20% of crops are infested in Papua New Guinea (Putulan et al. 2004). Hosts were mainly bananas, oranges, cucumbers, mangoes, passion fruits and guavas (Putulan et al. 2004). Similarly, in Japan from 1979 to 1998, authorities confiscated 1,093 items of OFF hosts, mainly mangoes, Syzygium sp., longans, guavas, star apples and capsicums (Iwaizumi 2004).

What are the future steps for HRN and its use in Australia? Unfortunately and regarding potential incursions, there is HRN data only for B. tryoni and not the other endemic tephritids. Therefore, the possible influence of these other tephritids is unknown but may impact any calculated reproductive advantage between B. dorsalis and B. tryoni. Thus, more data is required for HRN for hosts, and particularly for other Australian fruit fly species. Additional data on issues such as climate, altitude, latitude and host ripeness would facilitate a more meaningful analysis of invasive potential. Additionally, more research is required to evaluate the role of HRN in fruit fly management and trade. This additional data would provide the basis for a more informed view of incursion risk, and the establishment of any fruit fly in any circumstance or country.

However, based on my assessment, capsicums may be well-suited as sentinel plants for the early detection of B. dorsalis in many parts of Australia. Capsicums are grown throughout the year (particularly in protected cropping) in most parts of Australia. These capsicums are sold mainly for domestic consumption although some capsicums are imported. Capsicums are ranked third in the B. dorsalis reproductive advantage in Table 1. This review should not be seen as definitive because of much missing data. However, it does demonstrate how HRN can be used to better inform surveillance, management and trade.