An orchard treated in isolation is recolonised from its surroundings. Knipling (1979) formalised the argument that suppression applied uniformly to an entire population achieves more than the same effort applied intensively to parts of it, and the approach has since been developed for many pests, fruit flies prominent among them (Dyck et al., 2005). For fruit flies the area that matters includes commercial orchards, village and garden trees, roadside and wild hosts, markets and packhouses. A coastal district adds year-round host availability, since mango is followed by guava, citrus and other fruits, and tropical almond, Terminalia catappa, a favoured wild host, is common along shorelines.

Males of B. dorsalis are strongly attracted to methyl eugenol, a plant compound they feed on and use in courtship (Tan and Nishida, 2012). The attraction serves both for detection, in baited traps, and for control by male annihilation, in which the lure is combined with an insecticide and distributed densely enough to remove most males before they mate. The method eradicated the species from the island of Rota in the 1960s (Steiner et al., 1965). Females are targeted with protein bait sprays applied as spots to foliage. Orchard sanitation removes the larvae in fallen and unharvested fruit. Parasitoids, notably the egg-attacking braconid Fopius arisanus, and the release of sterile males can be added where programmes have the means (Vargas et al., 2015). Experience with these tools in African mango systems is collected in Ekesi et al. (2016).

The International Plant Protection Convention sets the terms on which official action is recognised by trading partners. ISPM 9 describes how an eradication programme is to be evaluated, organised and verified (IPPC, 1998). ISPM 26 covers the establishment and maintenance of areas free of fruit flies, including surveillance, corrective action after a detection and the suspension and reinstatement of status (IPPC, 2006). Detailed guidance on trap types, lures, densities and servicing is published jointly by the Food and Agriculture Organization and the International Atomic Energy Agency (FAO/IAEA, 2018). The framework below does not restate these documents. It arranges their requirements, together with the control tools, into a sequence that a district can follow.

A district at risk should run a standing network of methyl eugenol traps before any fly is found, concentrated where entry is most likely: the port, fruit markets, bus and lorry terminals, and orchards near them. Trap placement and servicing follow the published guidance (FAO/IAEA, 2018). A suspect specimen is sent at once to a competent taxonomist, since species of the B. dorsalis complex are hard to separate and molecular confirmation may be needed. The plant protection authority is notified as soon as identity is confirmed, because official recognition of all later steps depends on it.

Confirmation triggers a delimiting survey. Trap density is raised sharply around each detection and decreases outward in zones, and ripe fruit of mango and other hosts is collected and held to rear out larvae, which shows whether the fly is breeding or only arriving. Movement of host fruit out of the infested zone is restricted while the survey proceeds. Within a few weeks the task force should have a map of positive traps and infested fruit samples from which to judge the extent of the population.

The delimiting survey supplies the evidence for the most consequential decision of the response. Eradication is realistic when the infestation is small, recently established and bounded, and when reinvasion can be prevented. A review of tephritid eradication programmes worldwide found that many have succeeded, most often against incipient populations detected early (Suckling et al., 2016). A coastal district with continuous hosts, a neighbouring infested region and unregulated fruit movement may meet none of these conditions, and the honest objective is then sustained suppression to a level that allows production and trade. The task force should record which objective it has adopted and why, in the terms set out in ISPM 9 (IPPC, 1998), and should revisit the decision at fixed intervals.

Suppression combines three measures applied to every host tree in the area, including those in homesteads. Male annihilation stations are distributed on a grid and renewed as the lure weathers. Protein bait is applied as spot sprays at regular intervals through the fruiting season, which uses far less insecticide than cover spraying and spares natural enemies. Fallen and over-ripe fruit is collected at least weekly and destroyed by deep burial, bagging in the sun, or placement in tent-like structures that retain emerging flies while letting parasitoids escape. Sanitation is the measure most dependent on household cooperation, and it is where community organisers and schools are of most use.

Growers need to sell fruit while the campaign runs. Mango cultivars differ in the damage they suffer, and late-maturing ones are typically worst affected because fly numbers build through the season (Vayssières et al., 2009). Harvesting at the mature green stage and bagging fruit on the tree reduce infestation. For consignments leaving the district, post-harvest disinfestation by hot water immersion or vapour heat, to schedules accepted by the importing authority, provides the assurance that inspection alone cannot (Heather and Hallman, 2008).

Trap catches, expressed as flies per trap per day, and the proportion of sampled fruit infested are compiled weekly and mapped. Under an eradication objective, freedom is declared only after trapping at delimiting intensity has yielded no detection for a period covering several generations of the fly, as provided for in ISPM 26 (IPPC, 2006). Under a suppression objective, the review compares catches and fruit losses with the targets set at the outset. In either case the sentinel network of Phase 1 remains in place.