Aedini phylogenetic classification

The traditional classification of tribe Aedini (Knight & Stone, 1977) included nine genera, the largest of which was Aedes Meigen. The first revision of aedine mosquitoes was conducted by Edwards (1932), who recognised an Aedes group of seven genera – Aedes Meigen (the largest with 29 subgenera, many of which had previously been described as genera), Armigeres Theobald, Eretmapodites Theobald, Haemagogus Williston, Heizmannia Ludlow, Opifex Hutton and Psorophora Robineau-Desvoidy. Aedini was defined by Belkin (1962) to consist of the Aedes group of Edwards and genera Ayurakitia Thurman, Udaya Thurman and Zeugnomyia Leicester. Ayurakitia was later reduced to subgeneric status in Aedes by Mattingly (1971); hence, the nine genera recognised by Knight & Stone (1977). Belkin recognised that some subgenera of Aedes were heterogeneous groups and noted that the genus would likely have to be revised into smaller more natural groups of species.

Radical changes to the traditional classification of genus Aedes were introduced in a series of studies which began with the restoration of Verrallina Theobald and Ayurakitia to generic status (Reinert, 1999, 2000a respectively) and were followed by the landmark study of Reinert (2000b), who conducted a comprehensive and systematic analysis of over 65% of the species of genus Aedes that were recognised at the time. Based principally on morphological features of the male and female genitalia, as well as anatomical characters of fourth-instar larvae and pupae, Reinert split the composite genus Aedes into genera Aedes and Ochlerotatus Lynch Arribálzaga (originally proposed as a genus), with 22 of the 43 recognised subgenera placed in genus Aedes and 21 placed in genus Ochlerotatus. This was followed by a series of four major studies based on cladistic analyses of morphological characters (Reinert et al., 2004, 2006, 2008, 2009), which were complemented by many new treatments of species, generic-level taxa and character descriptions. In the first of these, Reinert et al. (2004) carried out an extensive study of higher-level relationships within Aedini based on 172 morphological characters of eggs, fourth-instar larvae, pupae and adults of 119 species representing the 12 genera and 56 subgenera of the tribe that were recognised following the taxonomic changes of Reinert (1999, 2000a, 2000b). Aedini and all existing genera except Ochlerotatus and Aedes were recovered as monophyletic. Ochlerotatus formed a polyphyletic assemblage that included Haemagogus, Opifex, Ochlerotatus subgenus Nothoskusea Dumbleton and Psorophora, and Aedes was polyphyletic relative to Armigeres, Ayurakitia, Eretmapodites, Heizmannia, Udaya, Verrallina and Zeugnomyia. Based on these findings, the authors proposed the elevation of 32 subgenera of Aedes and Ochlerotatus to generic status. In addition, Downsiomyia Vargas was resurrected from synonymy with Finlaya Theobald as a genus for the Kochi Group of species and a new genus Tanakaius Reinert et al. was established for two other species previously included in Aedes (Finlaya), resulting in the formal recognition of 46 genera within tribe Aedini. The decision to afford generic rank to all of these taxa was based on the principal of equivalent rank and the perceived integrity of monophyletic groups.

Reinert et al. (2006) analysed the relationships of a wider range of taxa, with a focus on Finlaya and related taxa. Relationships were explored using 232 characters from eggs, fourth-instar larvae, pupae, adults and immature habitat coded for 116 exemplar species. The ingroup comprised 74 species, including 66 species that were classified as Aedes (Finlaya) prior to Zavortink (1972) and Reinert et al. (2004), and eight related species. The outgroup comprised the non-aedine species used in the previous study, and representatives of subgeneric taxa of the polyphyletic Aedes and Ochlerotatus. The results corroborated the monophyly of the 13 genera proposed in 2004 that were included in the study, even though different taxa and additional morphological characters were used in the two analyses. Four subgenera of ‘Ochlerotatussensu auctorum (Bruceharrisonius Reinert, Macleaya Theobald, Molpemyia Theobald and Pseudoskusea Theobald) were raised to generic status, two nominal generic-level taxa (Hulecoeteomyia Theobald and Phagomyia Theobald) were resurrected as genera from synonymy with ‘Finlayasensu auctorum and 11 new genera were proposed based on the strength of character data and the principle of equivalent rank, bringing the total to 63 genera in tribe Aedini.

Reinert et al. (2008) reported the results of a third study that focused on Ochlerotatus and related taxa of uncertain taxonomic position, based on 297 characters from eggs, fourth-instar larvae, pupae, adults and immature habitat coded for 158 exemplar species. The ingroup comprised 54 species and the outgroup included four non-aedine species and 100 aedine species, 21 of which were previously classified as incertae sedis (Reinert et al., 2006). The analyses supported the generic status of 30 taxa included in the study that were previously recognised as genera (Reinert et al., 2004, 2006). Geoskusea Edwards, Levua Stone & Bohart, Ochlerotatus, ‘Ochlerotatussensu auctorum, Pseudoskusea and Rhinoskusea Edwards comprised a clade that was treated as a genus Ochlerotatus. The latter four taxa were restored to their previous subgeneric rank within the genus. Five nominal generic-level taxa (Acartomyia Theobald, Chrysoconops Goeldi, Culicelsa Felt, Gilesia Theobald and Protoculex Felt) were resurrected from synonymy with ‘Ochlerotatussensu auctorum for generic groups, and three additional genera, Danielsia Theobald (raised from synonymy), Luius Reinert et al. (newly proposed) and Hopkinsius Reinert et al. (newly proposed), were established for species previously included in ‘Finlayasensu auctorum, resulting in the recognition of 62 genera in tribe Aedini.

Reinert et al. (2009) reported the results of a comprehensive phylogenetic analysis of the tribe based on 336 morphological characters of eggs, larvae, pupae, adults and the habitat of immature stages coded for 270 exemplar species, including an outgroup of four species from different non-aedine genera. As in their previous studies, unordered multistate characters were analysed under implied weights but the analyses were performed with TNT (Goloboff et al., 2003) and clade support was assessed using the parsimony jackknife (Farris et al., 1996) and symmetric resampling (Goloboff et al., 2003). The results validated the monophyly and status of the genera recognised in the previous studies and, with only two exceptions, all the groups represented by two or more species were again recovered as monophyletic clades. The taxon Tewarius Reinert, newly added, was confirmed to be monophyletic. Subgenera Mattinglyia Lien of genus Heizmannia and the nominotypical subgenus of genus Verrallina were found to be paraphyletic with respect to subgenera Heizmannia and Neomacleaya Theobald, respectively. Genus Stegomyia Theobald was shown to be polyphyletic, and seven subgenera (and many unplaced species) were established for the genus (Reinert et al., 2009). Thus, based on the series of morphological and phylogenetic studies of Reinert and his colleagues (Reinert, 1999, 2000a, 2000b; Reinert et al., 2004, 2006, 2008, 2009) and the addition of genera Borichinda Harbach & Rattanarithikul (in Harbach et al., 2007) and Nyctomyia Harbach, 2013 (replacement name for Nyx Harbach & Linton, 2013, in Harbach et al., 2013), the internal classification of tribe Aedini was radically reorganised and now formally comprises 81 genera and 48 subgenera compared to 12 genera and 56 subgenera that were recognised prior to these studies.

The nomenclatural changes of Reinert (2000b) and Reinert et al. (2004) provoked a heated debate and much controversy. Given the major changes to the classification of Aedes and the implications for those working in applied taxonomy, this was perhaps not surprising. Some critics were opposed to nomenclatural changes based on an exemplar approach and direct translation of the results of cladistic analyses into a reclassification of the tribe [though one wonders why there has never been any objection to the elevation of Verrallina and Ayurakitia to generic status by Reinert, 1999, 2000a respectively]. However, despite the criticism and controversy, Reinert et al. (2004) went on, as noted above, to publish the results of three further detailed studies, which appear to have been largely ignored by mosquito workers, due at least in part to the earlier criticism. A consequence of the heated debate was a deep impasse between researchers and applied taxonomists who supported the reclassification advanced by Reinert et al. (2009) and those who ignored or summarily rejected it. In defence of Reinert et al., it must be emphasised that the application of objective phylogenetic analyses of explicitly coded character data produces repeatable results, and to undertake such analyses but not translate the results into a classification is an abnegation of responsibility. The MTI recognises that more data and phylogenetic analyses are needed to corroborate and further promote the re-classification of Aedini, but to completely deny it based purely on operational functionality is an impediment to further research and the realisation of a natural classification of the tribe. Furthermore, more accurate delineation of interspecific and intraspecific relationships is imperative for a deeper understanding of aedine phylogeny and taxonomy.

A thorough review of molecular studies that include taxa of Aedini can be found in Reinert et al. (2009). One of the first of these studies was conducted by Besansky & Fahey (1997), who examined the relationships of three Aedes (traditional sense) and one Haemagogus species among 13 mosquito species based on sequences for the nuclear protein-coding white gene. Another taxon-limited study by Kumar et al. (1998) examined the relationships of Aedes (in the traditional sense), a species of Haemagogus, a species of Armigeres and 15 culicine species based on the ribosomal RNA cistron. Cook et al. (2005) were one of the first research groups to use the cytochrome oxidase c subunits COI and COII of mitochondrial DNA. They studied species representing subgenera traditionally recognised as subgenera of Aedes. The results of these studies revealed overall support for the taxonomic groupings of Reinert (2000) and Reinert et al. (2004).

Three publications on mosquito barcoding, Cywinska et al. (2006), Kumar et al. (2007) and Wang et al. (2012), included more representatives of aedine genera than previous molecular studies, but, as Reinert et al. (2009) pointed out, their results are based on neighbor-joining, a phenetic as opposed to a phylogenetic method, and as such their results are diminished in interpretative utility.

Reidenbach et al. (2009) are the only researchers who have performed a phylogenetic analysis of combined morphological and molecular data for Culicidae. Their study included mosquitoes from 25 genera and data for six nuclear genes and 80 morphological characters. In agreement with Reinert et al. (2009), genus Psorophora was recovered as the sister to representatives of six other aedine genera. The relationships of the six other genera included in the study of Reidenbach et al. could be compared with the results of Reinert et al. (2009), but the findings are unlikely to be particularly meaningful in view of the far greater number of taxa included in analysis of those authors.

In summary, the ranking of taxa as genera or subgenera is subjective, but essentially involves the criteria of monophyly and equivalent rank in a phylogenetic analysis. In the analyses of Reinert et al., taxa that were widely accepted as genera (i.e. Armigeres, Eretmapodites, Haemagogus, Heizmannia, Opifex, Psorophora, Udaya and Zeugnomyia) were taken as the reference clades for equivalence, monophyletic clades of equivalent rank to those taxa were afforded generic rank and monophyletic groups within those clades were afforded subgeneric rank. It is noted that intuitive interpretation is not without merit as the application of explicit methodology confirmed the monophyly of most subgenera and species groups that were diagnosed by unique combinations of characters. The principal problem, then, is not in recognising monophyletic groups, but in deciding which taxonomic ranks should be assigned to such taxa once their phylogenetic relationships are established. Bearing in mind that acceptance of broad genus-group concepts for convenience is phylogenetically insupportable, in cases where a taxon is found to be paraphyletic or polyphyletic it should be reclassified to ensure that taxonomic ranking reflects monophyly. On that basis, and because most phylogenetic systematists do not accept paraphyletic or polyphyletic classifications, Reinert et al. raised many of the subgenera and species groups of tribe Aedini to generic level. It should be noted that many of the taxa recognised as subgenera in the traditional classification of genus Aedes were originally described as genera. The phylogenetic and traditional classifications of Aedini are contrasted in the attached file.

File attachments: 
Scratchpads developed and conceived by (alphabetical): Ed Baker, Katherine Bouton Alice Heaton Dimitris Koureas, Laurence Livermore, Dave Roberts, Simon Rycroft, Ben Scott, Vince Smith