The Old World genera Erpornis Hodgson, 1844, and Pteruthius Swainson, 1832, were long considered babblers Timaliidae (Deignan 1964, Sibley & Monroe 1990, Howard & Moore 1991). This was based on gross morphology (in Erpornis, similarity to Yuhina; in Pteruthius, bold plumage pattern reminiscent of, e.g., the laughingthrush genus Garrulax) and biogeography (presence alongside many ‘other’ species of babblers in the Oriental region), rather than phylogenetic analysis.

In the first molecular phylogenetic analysis of babblers, Cibois et al. (2002) found that the species known as ‘Yuhina’ zantholeuca was not closely related to other species of Yuhina but was placed in an unresolved position near the crows Corvus and cuckooshrikes Coracina. As a consequence, the species was removed from Yuhina and placed in the monotypic genus Erpornis (Cibois et al. 2002). Erpornis was subsequently found to be closely related to the New World vireos Vireonidae (Barker et al. 2004). Another study showed that Pteruthius does not belong with the babblers but is sister to the Erpornis–Vireonidae clade (Reddy & Cracraft 2007). These findings were corroborated by Jønsson et al. (2016), Oliveros et al. (2019) and Stervander et al. (2020).

Following the discovery that Erpornis and Pteruthius represent successive sister groups of the vireos, they were lumped within an expanded Vireonidae in several taxonomies (Dickinson & Christidis 2014, Winkler et al. 2015, Fjeldså et al. 2020b). Placing Erpornis and Pteruthius in Vireonidae was probably necessitated by the lack of available family-group names for these two taxa. However, a good reason to refrain from placing Erpornis and Pteruthius into Vireonidae is that it changes the meaning of Vireonidae and affects several aspects of the systematics and biogeography of the family. For example, the time to their most recent common ancestor (MRCA) is shifted at least ten million years backwards (Stervander et al. 2020: fig. A2.1), the geographic origin of the MRCA shifts from the Americas to Asia, and the synapomorphies and diagnostic character states of ‘Vireonidae’ are modified. One can argue that these changes were unnecessary and potentially confusing. Given the long stability of the meaning of the name Vireonidae (e.g. Sharpe 1903, Wetmore 1930, Mayr & Amadon 1951, Deignan 1964, Wolters 1980, Sibley & Monroe 1990, Dickinson 2003), the alternative taxonomic solution, to exclude Erpornis and Pteruthius and place these in new family taxa, should be proposed. For example, Slager et al. (2014), who were aware of the close relationship of Erpornis and Pteruthius to the vireos, used the name Vireonidae in its traditional sense, i.e. for the New World taxa alone, excluding Pteruthius and Erpornis.

Placing Erpornis, Pteruthius and Vireonidae in separate family taxa is in line with the current classification of Passeriformes, as reviewed in Fjeldså et al. (2020a). Fjeldså et al. (2020b) followed the idea, within Passeriformes, of temporal banding (Avise & Johns 1999), in which the classification at higher ranks (i.e. families and genera) should correspond to groups of similar age, based on calibrated phylogenies. They did not, however, establish a strict cut-off line (as had Jønsson et al. 2016 for Corvides), knowing that confidence intervals of nodes are often wide and can vary among analyses (e.g. Oliveros et al. 2019, Stervanderet al. 2020). Most taxa treated as families by Fjeldså et al. (2020b), however, diverged around 20 Mya, a date similar to estimates for the Erpornis lineage and the Pteruthius clade (Table 1). In the chronogram of Oliveros et al. (2019), 38 taxa treated as families (sensu Fjeldså et al. 2020b) are of equal or younger age than Pteruthius and six of these are also of equal or younger age than Erpornis (Table 2). Similarly, in the unconstrained chronogram of Stervander et al. (2020), 56 taxa treated as families (sensu Fjeldså et al. 2020b) are of equal or younger age than Pteruthius and 28 of these are also of equal or younger age than Erpornis (Table 2).

TABLE 1

Divergence times of Pteruthius and Erpornis from Vireonidae calculated in various dated phylogenies.

TABLE 2

Taxa treated as families (sensu Fjeldså et al. 2020b) that were inferred to be of equal or younger age than Pteruthius. Taxa marked with an asterisk were inferred to be younger than both Pteruthius and Erpornis.

Because no family-group names are available for Erpornis and Pteruthius, we propose:

Discussion

An alternative solution would be to treat Erpornithidae as a subfamily of Vireonidae, based on (i) the view that placement of the Asian Erpornis with the New World vireos conveys useful biogeographic information because it would emphasise the Asian origin of Vireonidae, (ii) the fact that in the phylogeny of Oliveros et al. (2019) only a small number of families are younger than Erpornithidae, and these are all in the nine-primaried New World oscines, and (iii) the age of Erpornithidae is younger than a large number of nodes within families in the phylogeny of Oliveros et al. (2019), which means that if temporal banding is applied to these groups, many new families should be named.

Ranking is obviously subjective, and therefore disagreements about the ‘appropriate’ rank cannot be solved empirically. In this case, ranking Erpornis as a subfamily and placing it in Vireonidae has two negative consequences: it results in an unnecessary change of the name of the vireos (‘Vireoninae’) and attaches the name ‘Vireonidae’ to a different, larger, older and geographically much more widespread group. We are not convinced by the arguments in favour of merging Erpornis with the vireos in a single family. First, merging the Asian Erpornis with the New World vireos in itself would not convey any information about the geographic origin of the vireos. Rather, the origin of a particular group, i.e. the direction of past geographic movement, is inferred from mapping the geographic range of multiple groups on a phylogeny. Second, whereas in the phylogeny of Oliveros et al. (2019) only six families were younger than Erpornithidae, and these were all in the nine-primaried New World oscines, in the phylogeny of Stervander et al. (2020) no fewer than 28 were of equal or younger age than Erpornis (Table 2), and these were not restricted to the nine-primaried New World oscines. Several nine-primaried New World oscine taxa were ranked as families by Barker et al. (2013) to keep the names of several major previously recognised groups stable (e.g. Thraupidae, Cardinalidae, Icteridae, Parulidae). We agree with this approach because it serves nomenclatural stability, even if it results in relatively ‘young’ family-level taxa. We have applied this approach to Erpornithidae and Pteruthiidae, so that Vireonidae can be maintained for the vireos. Third, the fact that several lineages in other families are older than Erpornithidae does not imply that these should also be ranked as families. Only when temporal banding is enforced with a strict threshold will there be a need to change the rank of these taxa, and we believe this is not necessary. Ranking Erpornis as a family is in line with the current classification of Passeriformes, as reviewed in Fjeldså et al. (2020a), which is only loosely based on temporal banding.