Priority host plants of the Queensland fruit fly, Bactrocera tryoni (Froggatt), based on the host reproduction number for tephritid management, surveillance and trade

The Host Reproduction Number (HRN) is a measure of the number of adult fruit flies that can emerge from one kilogram of fruit. HRN is a useful tool in surveillance, management and trade. I reviewed the literature for Queensland fruit fly (Qfly) and found data on 297 hosts. There were 81 Qfly hosts with HRN data and 216 Qfly hosts with no HRN data. The HRN will help to inform and triage hosts for target surveillance and management programs. Additionally, HRN will inform disinfestation activities pertaining to incursion response management, risk mitigation and trade options. There is a need for scientists to report HRN in a consistent manner so that Qfly management programs and trade protocols might be optimised.


Introduction
Many countries rely on trade and tourism for their economies.However, increasing international trade, including of fruit and vegetables, increases the chances of inadvertent introduction of exotic pests such as tephritids (Bebber et al. 2014;Early et al. 2016;Robinson and McNeill 2022).Additionally, tourists using air travel carry host fruit, potentially infested with tephritids (Cantrell et al. 2002;Putulan et al. 2004).Similarly, land-based tourists in vehicles are known to carry fruit, some of which is infested with fruit flies (Dominiak et al. 2000;Cantrell et al. 2002, Dominiak andCoombes 2009).After the initial long-distance jump assisted by trade or travellers, local dispersal results in an incursion, establishment and dispersal into a new environment (Meats et al. 2003;Sadler et al. 2011;Dominiak and Fanson 2020).
Tephritid fruit flies are one of the most important invasive and destructive crop pests, causing fruit degradation capitata Wiedemann) was reviewed but there was no data on hosts with no HRN (Dominiak and Taylor-Hukins 2022).
Here, I reviewed the Qfly literature and found HRN data on 81 hosts and I revised the list of 216 hosts without HRN data.

Methods and results
I followed the methodology described by Dominiak (2022) and Dominiak and Taylor-Hukins (2022).In essence, I used Google Scholar as a search engine because it yields more results than other databases (Pozsgai et al. 2021).The main research term used was "Bactrocera tryoni".I added additional terms such as "host" or "suitability" or "susceptibility" or "fruit" and similar terms in successive searches.The search results were examined and references chosen if they provided useful data for HRN.Usually, useful references were contained in the first four pages of each search.I changed the third search word based on the results in earlier searches.
In the formulation of Table 1, each reference for data of each host was examined regarding the host status or reproductive capacity to support adult fruit flies.I used field infestation results where possible to standardize an individual host HRN.Unfortunately, not all papers reported the information required (e.g.Bateman et al. 1966).Some papers reported infestations per fruit and provided insufficient information to calculate the HRN metric.Other papers reported infestation rates in graphs or figures and interpretation was too difficult to obtain accurate 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 worstcase 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 HSI proposed by Follett et al. (2021).The reasons for different HRN in different hosts was reviewed in references such as Lloyd et al. (2013) and Follett et al. (2021) and will not be reviewed in detail further here.
Variations in my comparison could be due to many factors.Dominiak et al. (2020) reported the maximum HRN while Lloyd et al. (2013) reported mean HRN and hence I may have intrinsically underestimated the Queensland HRN in this comparison.NSW is at the southernmost range of Qfly compared to Queensland and the Pacific Islands and Qfly may interact with hosts differently.Therefore, this current HRN ranking should be regarded as the initial step in understanding the fly-host relationship.Some reported high HRNs may not apply to all circumstances.For instance, altitude seems to play a role in Africa (see review in Dominiak 2022).There were greater infestations of Anastrepha sp. in the dry season while other species were more prevalent in the wet season (Birke 2023).Fruit infestation by Bactrocera spp. was positively correlated to fruit weight and vitamin C but negatively correlated with fruit firmness, acidity and total phenolic content (Aarti et al. 2023).Larger fruit were more prone to co-infestation than smaller fruit (Birke 2023).Higher sugar content and lower levels of phenols and tannins were highly correlated to infestation and emergence (Orono et al. 2019).Therefore, variations in HRN may be argued down from the worst-case scenario based on local circumstances.
The current knowledge of Qfly HRN is valuable.There was additional data available however not in the HRN format or in a way that HRN could be calculated.Therefore, based on the list of 216 hosts with no recorded HRN, there is considerable research yet to be done to fully understand how hosts influence management at a grower level.Particularly, the role of native hosts growing near commercial production sites remains poorly understood.
HRN has three main work areas in fruit flies.Firstly, the use of HRN can inform and triage surveillance targets for early detection or pest population monitoring.Early detection will increase the chance of corrective action before establishment occurs in fruit fly free zones and Pest Free Production Sites (Fay et al. 1997;Cantrell et al. 2002;Dominiak and Mapson 2017;Leblanc et al. 2013b).Secondly, in local or regional management, fruit fly control measures are more likely to be successful in shorter times if all the risk hosts in the area are known.For example, programs are less likely to be effective if regulatory staff walk past and ignore the roses (HRN = 354) in the front garden to treat only the peach (HRN = 179) or orange (HRN = 142) in the backyard (Dominiak et al. 2020).Regarding fruit carriage by land and air travellers and possible incursions, low HRN fruit are less likely to result in establishment compared to high HRN fruit (for the same weight of fruit carried) (Meats et al. 2003;Dominiak and Fanson 2023).This HRN knowledge could Syzygium tierneyanum, Syzygium xerampelimum, Terminalia arenicola, Terminalia aridicola, Terminalia ferdinandiana, Terminalia melanocarpa, Terminalia muelleri, Terminalia platyphylla, Terminalia seriocarpa, Terminalia subacroptera, Trichosanthes anguina, Trichosanthes cucumerina var.anguinea, Vitis labrusca and Ziziphus jujuba.Additionally, Gossypium hirsutum and Maclura pomifera were added by Khan et al. (2012) and Reynolds et al. (2015) respectively.

Discussion
Of the reported 297 hosts, the 81 hosts with a recorded HRN are provided in Table 1.However, there are 216 ( 72.7%) other known hosts with little or no recorded HRN.Previously in an Australian review, Hancock et al. (2000) identified 243 hosts; 58 hosts now have HRN.There were 40 "major hosts" identified by Hancock et al. (2000) but I could find HRN data on only 21 of these hosts.Within the data available, the Hancock's "major host" had HRN that ranged from 6.7 to 4,065.Perhaps part of the "major host" categorisation was based on the proportion of hosts infested.For example, in Western Australia, they examined 62 hosts but found Qfly in only eight hosts with loquat most commonly infested (44 trees) compared to eight fig plants, five apricots and one guava (Sproul and Froudist 1992).The HRN for these hosts are 324 (loquat), 10 (fig), 63 (apricot) and 4,065 (guava) respectively so HRN alone does not explain the rate of infestation.Lloyd et al. (2010) reported a wide range of the proportion of hosts infested.I suspect that the classification of "major host" was linked to the proportion of host infested, rather than HRN.
There are some inconsistencies with HRN metrics in Table 1.Kumquats in southern New South Wales (NSW) were recorded with HRN = 234 but kumquats had a HRN of 25 and 30 in Queensland and the Pacific Islands respectively.In Queensland, only two of 24 samples were infested (Lloyd et al. 2010).It seems that kumquats are a more important Qfly host in southern NSW.Similarly, the HRN for sweet oranges ranged from 142 in NSW to < 30 in the other two locations.In Queensland, 26 of 227 samples were infested (Lloyd et al. 2010).Again, oranges seem to have a different risk profile in southern Australian climates.The HRN for avocadoes ranged from 129 (Pacific Islands) to 22 (Queensland).Roses ranged from 354 (Queensland) to 40 (NSW).Guava has a high HRN for many fruit flies (Dominiak 2021) but the HRN for guava ranged from 92 (NSW) to 4,065 (Pacific Islands).However, the high 4,065 was the highest HRN in a nine-year program with a HRN < 100 in seven of the other eight years.Vargas et al. (2007) suggested the same disinfestation protocol as a host such as oranges (HRN = 142) where the biosecurity risk is considerably higher.Should oranges be treated the same as blueberries (HRN = 869)?In a systems approach, low HRN may be combined with winter window of non-breeding in cooler months or in the southern most edge of the Qfly range (Clarke et al. 2022;Follett et al. 2022).This window may be part of the seasonal weight changes in Qfly: Dominiak et al. (2021) reported lower weights in autumn which were possibly linked to longevity rather than dispersal.Perhaps this lower weight also may trigger a non-breeding phase to survive the challenging conditions of winter.
Given the variable nature of HRN reporting, there is a need for a more consistent approach to HRN to optimise surveillance, management and trade.In particular, there is a need for all scientists to report the maximum and seasonal HRN for biosecurity purposes.The collection of HRN data would lead to the opportunity to develop host specific trade protocols for treatment of poor hosts because low hosts inherently provide a lower risk of transporting target fruit fly into new areas (Follett and McQuate 2001).
be used to better target the education of travellers (Dominiak and Coombes 2009).
Finally, market access is reliant of the demonstration of minimum risk to the importing country or state.Some commercial hosts have no recorded HRN and this may limit market access opportunities using the Systems Approach (Jang 2006;Dominiak 2019;Van Klinken et al. 2020).A Systems Approach integrates two or more independent phytosanitary measures to cumulatively provide an appropriate biosecurity confidence (Follett and Neven 2006;Sequeira and Griffin 2014).The Systems Approach offers a more robust system to provide a more nuanced approach of using a system or combination of risk mitigation measures that reach a level of efficacy previously provided by an end point treatment but are not vulnerable to a possible failure of that end-point treatment (Quinland et al. 2020).The concept of poor host status, along with other conditions, was used as part of a systems approach to export 'Sharwil' avocadoes from Hawaii to continental United States (Follett and Vargas 2010).Recently, Follett et al. (2022) proposed a systems approach for 'Malama' avocadoes using poor host status, low pest prevalence and limited harvest period for export from Hawaii to mainland USA.
For trade, a probit 9 or 8.7 treatment response level is required by many importing countries.For many countries, a 99.99% mortality level is acceptable for quarantine treatment efficacy, treating only about 30,000 individuals (Follett and Neven 2006).However, this level of treatment may be overly demanding for poor hosts that are rarely infested or infested at very low levels.In low HRN cases, a less stringent treatment protocol may provide the equivalent level of quarantine security such as probit 9 (Follett and McQuate 2001).A low HRN status may become a particular category for trade, just as a "hard green" status identifies a particular susceptibility to tephritids such as Qfly.
An alternate approach is to use probability to demonstrate that an appropriate level of protection or risk mitigation has been reached, equal to an end-point treatment (Quinlan et al. 2020).A Bayesian network was used to model the Qfly populations to persist or survive in an area-wide management of habitat (van Klinken et al. 2019.Similarly, the scenario tree analysis was used to statistically demonstrate that pest populations had been eradicated or were non-existent (Dominiak et al. 2011;Magarey et al. 2019;Charlton et al. 2022).A similar approach could be used to statistically demonstrate the components of a systems approach can reach a status equivalent to probit 9. HRN would be a useful component to such an analysis.These statistical techniques have been used in pest and landscape (area) analysis but are yet to be demonstrated in a pest and commodity (volume) setting.
Regarding current end-point treatments, hosts with a low HRN (e.g.lemons HRN < 4) may not need to be treated with

Table 1
(continued) (Diptera: Tephritidae) into a fruit fly free area.Aust J Exper Agric 40:763-771 Dominiak BC, Gott K, McIver D, Grant T, Gillespie PS, Worsley P, Clift A, Sergeant ESG (2011) Scenario tree analysis of zero detections and the eradication of yellow crazy ant (Anoplepis gracilipes (Smith)), in New South Wales, Australia.Australian Plant Prot Quart 26:124-129 Dominiak BC, Mellberg A, Galvin T, Gillespie PS (2018) Rose hips, The economic and social impact of the Bactrocera papaya Drew and Hancock (Asian Papaya fruit fly) outbreak in Australia.Pp 205-207.In Management of fruit flies in the Pacific.A.J. Allwood and R.A.I.Drew (eds).ACIAR Proc Early R, Bradley B, Dukes J, Lawler JJ, Olden JD, Blumenthal DM, Gonzalez P, Groholz ED, Ibanez I, Miller LP, Sorte CJB, Tatem AJ (2016) Global threats from invasive alien species in the twenty-first century and national response capacities.Nat Commun 7:e12485 Fay HAC, Drew RAI, Lloyd AC (1997) The eradication program for papaya fruit fly (Bactrocera papayae Drew and Hancock) in north Queensland.Pp. 259-261.In Management of fruit flies in the Pacific.A.J. Allwood and R.A.I.Drew (eds).ACIAR Proc Follett PA, McQuate GT (2001) Accelerated quarantine treatment development for insects on poor hosts.J Econ Entomol 94:1005-1011 Follett PA, Neven LG (2006) Current trends in quarantine entomology.Ann Rev Entomol 51:359-385 Follett PA, Vargas RI (2010) A systems approach to mitigate oriental fruit fly risk in 'Sharwil' avocados exported from Hawaii.Acta Hortic 880:439-445 Follett PA, Haynes FEM, Dominiak BC (2021) Host suitability index for polyphagous tephritid fruit flies.J Econ Entomol 114:1021-1034 Follett PA, Asmus G, Hamilton L, Stockton D (2022) Susceptibility of 'Malama' avocadoes to Bactrocera dorsalis (Diptera: Tephritidae) in Hawai'I and a proposed systems approach for market access.J Econ Entomol 115:1894-1900 Global Invasive Species Database (2022) http://www.iucngisd.org/gisd.Accessed 1 August 2022 Hancock DL, Hamacek EL, Lloyd AC, Elson-Harris MM (2000) The distribution and host plants of fruit flies (Diptera: Tephritidae) in Australia.Queensland Department of Primary Industries, Brisbane Jang EB (2006) Systems approaches for managing the phytosanitary risk of trading in commodities that are hosts of fruit fly.In Ekesi, S., Mohamed, S., De Meyer, M. (eds) Fruit fly research and development in Africa -towards a sustainable management strategy to improve horticulture.Springer.pp 737-751