The South American fruit fly, Anastrepha fraterculus (Wiedemann, 1830) s.l., is present in most countries of the Americas from the USA to Argentina (Hernández and Aluja 1993, Steck 1999, Zucchi 2007). Its centre of diversity is the South American subcontinent, where formerly it was thought to occurr in two, possibly unconnected bands: one along the western edge, including both highland and lowland areas of the Andean range, and the other along the east coast. However, recent data indicate its presence in parts of the Brazilian Amazon basin (Zucchi et al. 2011). It has been reported to infest about 110 host plants including major fruit crops (Norrbom and Kim 1988, Zucchi 2000, 2015, Norrbom 2004). The presence of this highly destructive pest results in quarantine restrictions for fruit export to many countries (Steck 1999).

The high levels of variability found among different populations throughout the geographical range of A. fraterculus led to the conclusion that it is a complex of cryptic species rather than a single biological entity (Stone 1942, Morgante et al. 1980, Malavasi and Morgante 1982, Solferini and Morgante 1987, Steck 1991, Steck and Sheppard 1993, Selivon 1996, Hernández-Ortiz et al. 2004). Differences have been reported based mainly on morphology, pest status and genetics (including karyotype, isozyme and molecular analyses); these are reviewed in Steck (1999) and some aspects discussed in subsequent studies (McPheron et al. 1999, Norrbom et al. 1999, Gomes Silva 2000, Smith-Caldas et al. 2001, Aluja et al. 2003, Hernández-Ortiz et al. 2004, Barr et al. 2005, Selivon et al. 2005, Goday et al. 2006, Selivon and Perondini 2007, Silva and Barr 2008, Prezotto 2008, Cáceres et al. 2009). However, in order to establish how much of this variation reflects population level variation, and how much reflects unrecognised cryptic species diversity, it is necessary to systematically correlate these genetic and morphological differences with the existence of reproductive isolation and other life history related traits (hosts, demography, etc).

Reproductive incompatibility has been reported both at pre- and post-zygotic levels between some A. fraterculus populations. At the pre-zygotic level, mating compatibility was evaluated among different populations from South America, involving lowland (Peru) and highland (Colombia) areas from the Andean region, and the south-eastern part of the continent (Brazil and Argentina). Most of the populations were shown to have some level of incompatibility with each other and thus appeared sexually isolated. Flies of different populations were often sexually active at different times of the day suggesting different sexual behaviour (Selivon 1996, Vera et al. 2006, Cáceres et al. 2009).

Post-zygotic studies between two populations from Brazil (Selivon et al. 1999) and between one Argentinean population and one Peruvian population (Cáceres et al. 2009) found partial hybrid inviability and sex ratio distortion confirming the existence of post-zygotic barriers. In the former case, cytological, isozyme and molecular studies revealed differences among groups (Malavasi and Morgante 1982, Selivon 1996, Selivon et al. 2005, Goday et al. 2006); while for the latter case, differences between groups were also found in terms of male sex pheromones and karyotypes (Cáceres et al. 2009).

The combined results of these studies suggested the existence of seven different biological entities referred to as: Anastrepha sp. 1 aff. fraterculus (Brazilian 1 morphotype) (Yamada and Selivon 2001), Anastrepha sp. 2 aff. fraterculus (Brazilian 2 morphotype) (Yamada and Selivon 2001), Anastrepha sp. 3 aff. fraterculus (Brazilian 3 morphotype) (Selivon et al. 2004), Anastrepha sp. 4 aff. fraterculus (Peruvian morphotype) (Selivon et al. 2004), A. fraterculus Mexican morphotype (Hernández-Ortiz et al. 2004), A. fraterculus Andean morphotype and A. fraterculus Venezuela coastal lowlands (Steck 1991).

Although previous studies provided strong evidence supporting the existence of several biological species, major knowledge gaps still existed in 2010. In particular, the described studies used different methodologies, did not use the same identified biological material and, most importantly, did not include all of the morphotypes. Therefore, in order to be able to formally describe and name these putative species, it was considered critical to apply a standardised, complete set of methodologies to all populations from the entire geographic distribution range in a comprehensive integrative taxonomy study. This would allow the characterization of each putative species and would provide sound diagnostic tools for addressing the related management and trade issues.

Definition of species limits and formal naming of these putative species will be relevant for plant protection authorities in determining which of them may or may not be quarantine pests. This would immediately allow some countries to gain access to international fresh fruit markets for those countries and commodities which can be determined to be outside the geographic and host range of correctly delimited A. fraterculus s.s. In addition, detailed studies on pest status, host range, economic impact and distribution would minimize any possible impact on trade between South American countries. Furthermore, knowing species boundaries and their levels of sexual compatibility within the complex would enable the implementation of the SIT.

Colonies from five A. fraterculus morphotypes (Mexican, Andean, Peruvian, Brazilian 1 and Brazilian 3) were established and used for behavioural, chemical, cytological, molecular and larval morphology studies. Linear and geometric morphometry were validated as tools for morphotype discrimination. Comprehensive morphometric studies supported the existence of eight morphotypes: the seven reported previously and a new Equadorian morphotype (Hernández-Ortiz et al. 2012, 2015).

Guidelines for performing mating compatibility field cage tests were developed. Reproductive compatibility studies were performed among the five morphotypes. Among the combinations studied, morphotypes were incompatible at the pre- and post-zygotic level (Rull et al. 2012, 2013, Devescovi et al. 2014). Male calling and courtship behaviour were recorded for four morphotypes. Sexual behaviour studies helped to identify behavioural characteristics that allowed distinct morphotype descriptions such as time of sexual activity, acoustical signals and sequence of courtship behaviour (Rull et al. 2012, 2013, Devescovi et al. 2014, Dias et al. in press). One post-copulatory study revealed differences in sperm storage and remating propensity between the Peruvian and Brazilian 1 morphotypes (Abraham et al. 2014). Hybrid females tended to mate with hybrid males and as a result Segura et al. (2011) suggested that hybridization is a possible speciation mechanism. In all combinations analysed, post-zygotic isolation was found to be weaker than pre-zygotic (Devescovi et al. 2014). Wolbachia was detected in several morphotypes (Cáceres et al. 2009, Lima 2015, Lima et al. unpublished data) and more in-depth studies aiming to characterize these strains are in progress.

The chemical profiles of the male pheromones and cuticular hydrocarbons of these five morphotypes were characterized as complex blends that were qualitatively and quantitatively unique for the different morphotypes (Břízová et al. 2013, Gonçalves et al. 2013, Vaníčková et al. 2012, 2015a,c,d, Milet-Pinheiro et al. 2015). The description of the mitotic karyotypes from the Mexican, Colombian and Equadorian morphotypes, from which information was absent or incomplete, allowed confirming that karyotypes are unique for each morphotype (Canal et al. unpublished data). The polytene chromosome map for one morphotype was constructed (Gariou-Papalexiou et al. unpublished data). Internal Transcriber Spacer 1 (ITS 1) was found to be a good molecular marker to identify different groups (Sutton et al. 2015). Microsatellites were developed (Lanzavecchia et al. 2014) and proved to be successful to discriminate among morphotypes (Lima et al. unpublished data, Manni et al. 2015). The phylogenetic relationships of Andean-Ecuadorian populations were determined with other molecular markers (Ludeña et al. 2011).

Based on all the collected evidence it is now possible to describe four morphotypes as new species with the exception of the three Brazilian morphotypes for which it is still necessary to solve problems with the unknown origin of the holotype male of Anastrepha fraterculus (Wiedemann) and the new Ecuadorian morphotype from which further studies are required for an integrative description (Hernández-Ortiz et al. 2015). A manuscript with the description of the new species is in preparation (V. Hernández-Ortiz personal communication) and improved diagnostic tools are now available based on morphology, molecular markers, chemical profiles, cytology and sexual behaviour.

The knowledge and gaps identified at the start of the CRP, as well as the progress made by the end of the CRP in addressing the gaps identified for the A. fraterculus complex are summarized in Table 1.

Across Asia and the Pacific the fruit fly subfamily Dacinae contains some 47 recognised pest species (Drew and Romig 2013). Of these, eight were recognized within the Bactrocera dorsalis complex, with some being the most economically damaging of all pest species within the subfamily (Drew and Hancock 1994, Clarke et al. 2005). Losses caused by B. dorsalis complex species include destruction of crops, restriction of international trade, and the establishment of a range of quarantine and regulatory activities carried out by various regional governments.

Background research on these flies has generated data on diagnostics, field surveillance, quarantine strategies, field pest control, and market access protocols (e.g. Tan and Nishida 1996, 1998, Muraji and Nakahara 2002, Naeole and Haymer 2003, Smith et al. 2003, Armstrong and Ball 2005). But the key knowledge gap of the B. dorsalis complex was a lack of consensus on species limits of the major pest species in the complex, particularly B. dorsalis s.s. (Hendel, 1912), B. papayae (Drew & Hancock, 1994), B. philippinensis (Drew & Hancock, 1994), B. carambolae (Drew & Hancock, 1994) and B. invadens (Drew, Tsuruta & White, 2005) (Clarke et al. 2005, Drew et al. 2005, Wee and Tan 2005, Ebina and Ohto 2006, Drew et al. 2008). Failure to resolve the taxonomic status of the members of this complex prevented further development towards SIT integration into AW-IPM programmes against these pest insects and limited international horticultural trade.

Background research on the taxonomy of the B. dorsalis complex has been unable to provide definitive identification of some species (Clarke et al. 2005). This has confounded collecting associated host plant records and defining geographic distributions. It was considered absolutely essential that the species be accurately identified to be able to apply AW-IPM field programmes that include a SIT component. Because the trade implications and response systems to detections and/or incursions are different for all members in the complex, “near-enough” identification is, unfortunately, not good-enough. Consequently countries have difficulty overcoming the phytosanitary barriers to export-trade to major importers such as Australia, Japan, Europe, New Zealand, South Africa, and the USA. Another severe problem would arise if one member of the complex is detected or becomes established in a country, but is unable to be differentiated from others in the complex. In this case that country would then be forced to admit that all members of the complex may in fact be present, which would result in extended trade embargoes.

Therefore a comprehensive integrative taxonomy approach involving biological, morphological, chemo-ecological and molecular studies of the various members of the B. dorsalis complex were needed to: (1) resolve species limits by seeking a consensus result from different tests; (2) examine congruence between data from the different approaches to either support taxonomic revision or retain existing species status; and (3) develop robust diagnostic tools for the identified species.

Quantified cross-species field-cage mating trials, as described in the FAO/IAEA/USDA Quality Control Manual (FAO/IAEA/USDA 2014), were completed for B. dorsalis s.l. populations from China, Kenya, Malaysia, Pakistan, the Philippines, Suriname and Thailand. Results presented in Schutze et al. (2013), Bo et al. (2014), and Chinvinijkul et al. (2015) demonstrated pre- and post-zygotic mating compatibility between all target taxa except for crosses involving B. carambolae, which always showed some level of sexual isolation from the other taxa. Post-zygotic isolation tests up to three generations were carried out for crosses involving B. dorsalis s.s. (China and Pakistan) and B. invadens; no evidence for hybrid incompatibility was detected (Bo et al. 2014).

Chemical components and ratios of sex pheromone stored in male rectal gland and emitted during courtship after feeding on methyl eugenol (ME) were determined qualitatively and quantitatively for B. dorsalis s.s., B. invadens, B. papayae and B. philippinensis. The four ccomplex members had identical volatile emission profiles and rectal pheromonal components consisting of 2-allyl-3,4, dimethoxyphenol (DMP) and E-coniferyl alcohol (E-CF) (Tan et al. 2011, 2013), and the ratios of DMP: E-CF were not significantly different between the different members. Probit analysis showed that the responsiveness of these four members to ME was similar as their ED₅₀ values (= dose at which 50 % of the population responded) were not significantly different (Hee et al. 2015a). However, differences were found for B. carambolae.

Wing shape variation analysed through geometric morphometrics was used for the first time in the B. dorsalis complex (Schutze et al. 2012a,b, Krosch et al. 2013). Variation in wing shape proved to be extremely informative in interpreting variation within the B. dorsalis complex.

Genetic, cytogenetic and molecular analyses of B. dorsalis complex specimens collected across the geographical range were carried out in participating laboratories in Asia, Australia and Europe:

Cytogenetics: One of the objectives was to identify and evaluate cytogenetic tools that could help to resolve the taxonomic status of the five taxa under study, focusing on chromosomal rearrangements, especially inversions. For this purpose, mitotic and polytene chromosomes were analysed from colonized specimen representing B. dorsalis s.s., B. papayae, B. philippinensis, B. invadens and B. carambolae. Analysis of mitotic karyotypes could not detect any differences among these five taxa (Yesmin and Clyde 2012, Yesmin 2013, Augustinos et al. 2014, Augustinos et al. 2015), showing that all had the typical B. dorsalis s.s. karyotype as previously described by Hunwattanakul and Baimai (1994). Polytene chromosome maps were developed for the first time of a member of the B. dorsalis complex, i.e. B. dorsalis s.s. (Zacharopoulou et al. 2011a). Subsequent analysis showed that the five members of the complex do not present any chromosomal rearrangements that could be used as diagnostic characters and therefore these taxa can be regarded as homosequential (Augustinos et al. 2014, 2015). Although B. carambolae presented the same mitotic and polytene chromosome karyotype as the other members of the complex, the presence of a high number of minor asynapses in F₁ hybrids of B. dorsalis s.s. × B. carambolae crosses may indicate the presence of small differences in the chromosomal organization among the parental entities. However, these observations cannot be regarded as diagnostic at the species level (Augustinos et al. 2014).

Microsatellites: Microsatellite DNA markers derived from B. dorsalis s.s. were tested on populations of B. dorsalis s.s. from Bangladesh, Cambodia, China (multiple populations), Hawaii (two populations), Laos, Malaysia, Taiwan, and Thailand (multiple populations) (Shi et al. 2012, Krosch et al. 2013, Aketarawong et al. 2011, 2014a). The same set of markers combined with microsatellite markers derived from B. papayae were used to compare populations of B. dorsalis and B. papayae from the Thai/Malay Peninsula (Krosch et al. 2013, Aketarawong et al. 2014b). No genetic isolation was found between the B. dorsalis and B. papaya populations, supporting the hypothesis that both are the same entity. On the other hand, microsatellite markers, which amplify for B. carambolae and B. dorsalis, showed different genetic clusters between these two species, although admixture populations were observed. Admixture is evidence that some gene flow (i.e. hybridisation) may occur in the field between these species (Aketarawong et al. 2015).

Haplotype analysis: CO1 haplotype networks showed that common haplotypes were shared between B. dorsalis, B. papayae, B. philippinensis and B. invadens, but not with B. carambolae (Schutze et al. 2012b, 2015a). This supports the hypothesis that the first four taxa are a single biological species, while B. carambolae is distinct.

Phylogenetic analysis: A phylogenetic study using six neutral genetic markers found that B. carambolae could be resolved as a monophyletic clade from the other four target species, which were mixed together as an unresolved comb (Boykin et al. 2014, Schutze et al. 2015a).

Based on genetic, cytogenetic, pheromonal, morphometric and behavioural data, which repeatedly showed no or only minor variation between B. dorsalis, B. invadens, B. papayae, and B. philippinensis, formal taxonomic name changes were made. B. philippinensis was made a junior synonym of B. papayae by Drew and Romig (2013). Subsequently, also B. papayae and B. invadens were synonymised with B. dorsalis (Schutze et al. 2015b, Hee et al. 2015b), while the status of B. carambolae has not been altered. This means that only B. dorsalis and B. carambolae remain as scientifically valid names. The name changes have been widely accepted by national and regional plant protection organizations around the world, the Secretariat of the International Plant Protection Convention and the FAO (http://www.fao.org/news/story/en/item/262972/icode/).

In the works of Drew and Romig (2013), San Jose et al. (2013) and Schutze et al. (2015b) new morphological descriptions of the target taxa are provided. However, the use of morphology alone is not sufficient for definitive diagnosis of B. carambolae and B. dorsalis, but molecular and pheromone markers are now available to distinguish them. Molecular protocols using neutral genetic markers to distinguish the two species from each other, and from other closely related taxa, are provided in Boykin et al. (2014). Microsatellite markers which amplify for both species and which are used in population genetic studies, are provided in Aketarawong et al. (2015). The Y-specific marker will also separate the two species. For adult flies, the presence of the ME metabolites DMP and E-CF in the male rectal gland following ME feeding can be used to discriminate B. dorsalis from B. carambolae (which produces only E-CF) (Tan and Nishida 1996, Wee et al. 2007, Wee and Tan 2007).

The knowledge and gaps identified at the start of the CRP, as well as the progress by the end of the CRP in addressing the gaps identified for the Bactrocera dorsalis complex are summarized in Table 2.

The Afro-tropical fruit flies Ceratitis fasciventris, C. anonae and C. rosa (i.e. the CeratitisFAR complex), together with C. capitata and C. cosyra, are considered major horticultural pests of that region (White and Elson-Harris 1994, De Meyer 200la). These species are of quarantine significance (EPPO/CABI 1997) as they are highly polyphagous and damage a wide range of unrelated wild and cultivated crops (De Meyer et al. 2002), resulting in enormous economic losses wherever they occur (Barnes 2000, De Meyer 2001b). They have different distribution patterns that partially overlap, resulting in sympatric occurrence in particular areas.

Ceratitis rosa, C. fasciventris and C. anonae were considered the three members of the CeratitisFAR species complex (Virgilio et al. 2007a, 2007b). Taxonomically, C. fasciventris was initially considered a variety of C. rosa (Bezzi 1920) but has recently been recognized as a different entity with species status (De Meyer 2001a).

Unlike C. capitata, which has over the last century spread from its home range in East Africa and attained an almost world-wide distribution (Fletcher 1989, White and Elson-Harris 1994, http://www-naweb.iaea.org/nafa/news/images/Distribution-Mediterranean-fruit-fly-Ceratitis-capitata.jpg), C. rosa, C. fasciventris and C. anonae have so far not been reported outside the African mainland (except for La Réunion and Mauritius) but are potentially invasive.

Due to the difficulty in distinguishing morphologically some members of the complex, especially females, a number of molecular approaches for species recognition were used (Baliraine et al. 2004, Barr et al. 2006, Virgilio et al. 2008). However, these diagnostic tools remained inadequate for quarantine purposes and much more robust molecular markers were needed.

Development of molecular diagnostics, using microsatellites (Delatte et al. 2013), revealed a more complex structure than the mere existence of three entities within the CeratitisFAR complex. Five genotypic groups were identified (Virgilio et al. 2013) and later confirmed by morphological differences of the males (De Meyer et al. 2015a). Morphological diagnostics for male specimens of the five entities, called R1, R2 (C. rosa type 1 and 2), F1, F2 (C. fasciventris type 1 and 2) and A (C. anonae) were developed. Morphometric diagnostics using wing landmarks were developed for both sexes to a certain extent (Van Cann et al. 2015). Microsatellites allowed distinction between the five entities. Cytological studies were restricted to one representative (F2) acting as a reference dataset (Drosopoulou, unpublished data).

Adult morphology and morphometry, pheromone, cuticular hydrocarbon and distributional data were collected that provide evidence for the specific status of all three formerly recognized taxonomic entities within the FAR complex (i.e. C. fasciventris, C. anonae, C. rosa) (Vaníčková et al. 2014, Břízová et al. 2015, De Meyer et al. 2015, Van Cann et al. 2015). More detailed studies were conducted for the two C. rosa types (R1, R2) (adult morphology and morphometry, pheromone, cuticular hydrocarbon, developmental physiology, behavioural, and ecological data), which provided evidence that the two C. rosa types represent two separate species (Tanga et al. 2015, Van Cann et al. 2015, Vaníčková et al. 2015b), of which one (currently referred to as R2) will be formally described. An altitudinal transect in Tanzania, where R1 and R2 occur in sympatry, confirmed that R1 is more tolerant to higher temperatures and R2 better adapted to colder environments (Mwatawala et al. 2015a). For the two C. fasciventris types (F1, F2) these additional studies could not be conducted as laboratory colonies of one of the two types could not be established, preventing experiments on developmental physiology and mating compatibility. Larval morphology did not provide evidence with regard to the specific status, except for C. fasciventris (F2) to some extent (Steck and Ekesi 2015). Moreover, as a spin-off of this research it was shown that characters previously considered diagnostic for differentiation between species and even between the genera Ceratitis and Bactrocera, proved to be variable.

The knowledge and gaps identified at the start of the CRP, as well as the progress made by the end of the CRP in addressing the gaps identified for the CeratitisFAR complex are summarized in Table 3.

The melon fly, Zeugodacus cucurbitae (initially referred to as Bactrocera (Zeugodacus) cucurbitae), is a major pest of cucurbit crops that has spread from its area of origin (South East Asia) across Africa, Hawaii, the Indian Ocean, Papua New Guinea and the Solomon Islands (Severin et al. 1914, Dhillon et al. 2005). In particular, it causes severe losses in food crops and restrictions to trade for some cucurbit crops.

Some populations were identified in Africa, islands in the Indian Ocean, Hawaii and South East Asia with different host use, which could indicate the existance of very closely related species. Although the SIT has been effectively applied against the melon fly in certain regions (Koyama et al. 2004), this issue needed to be resolved to enable the application of species-specific treatments such as the SIT against all populations in all regions.

In spite of earlier observations and indication of different host-use by B. cucurbitae in different geographic regions, genetic studies using mitochondrial and nuclear markers indicated very low intraspecific variability worldwide. Population genetic studies using microsatellites were able to distinguish five major groups worldwide: African mainland and Seychelles, Réunion and Mauritius, Central Asia, SE Asia, and Hawaii (Virgilio et al. 2010). However no phylogeographic patterns could be discerned using cytogenetics analyses (Zacharopoulou et al. 2011b, 2013) or mitochondrial and nuclear gene fragments (total of 2764 bp) (Virgilio, unpublished data). The invasion history for the species on the African mainland was also reconstructed (Delatte et al. unpublished data).

Cross-mating experiments were conducted at the start of the CRP between populations of Mauritius, Seychelles and a genetic sexing strain of Hawaii and these indicated no mating incompatibility or sexual isolation (Sookar et al. 2013). Given this fact and the genetic assessments, it was decided there was no need for additional cross-mating studies.

Further studies, including host preference and microsatellite markers, did not show any relation between genetic structure and host plants (Virgilio et al. 2010, Sookar et al. 2013). It was concluded that there is no evidence of the existence of host races or cryptic species within B. cucurbitae. However, as a spin-off of the conducted research, recent studies on the higher phylogeny of dacines have shown that the higher taxonomic classification under which B. cucurbitae is placed, is a paraphyletic grouping, requiring a taxonomic change in generic placement (Krosch et al. 2012, Virgilio et al. 2015). A nomenclatorial act has raised the subgenus Zeugodacus (as well as other subgenera belonging to the Zeugodacus group, sensu Drew 1989) to genus level. As a result, Bactrocera cucurbitae was put in a new generic combination: Zeugodacus cucurbitae, and should be referred to by this name from now onwards (Virgilio et al. 2015).

The knowledge and gaps identified at the start of the CRP, as well as the progress made by the end of the CRP in addressing the gaps identified for Z. cucurbitae are discussed in De Meyer et al. (2015b) and summarized in Table 4.