Recently, Baker et al. (2005) published a Bayesian tree based on 658 base-pair control-region mitochondrial DNA sequences from 125 individual moas from widespread localities in New Zealand, a sample that included all then-accepted taxa (Worthy and Holdaway 2002, Bunce et al. 2003). They recognized 14 clades (interpreted as shown in Table 1) but avoided equating these clades with species by using only geographic identifiers. More recently, however, Baker (2007:18), in a paper whose title claims “advances in the study of geographic variation and speciation,” has equated each of the 14 clades with the status of species (Table 1). It is not my intention here to assess the validity of Baker's conclusions, nor is it to assess their relevance to taxonomy, but rather to point out that he applied incorrect nomenclature to 5 of the 14 clades while overlooking an important paper on moa systematics. The result has only further confused the taxonomy and nomenclature of moas.
People often confuse the three components to “labeling” taxa. First, taxonomy is the science of circumscription of a taxon, for example, at the species, generic, familial, or other levels, by diagnosis. Such taxa are based on a type or type series and delimit a group of individuals and their relationships to other such groups. Second, nomenclature is a technique for the naming of such taxa, and it is governed by rules (e.g., International Code of Zoological Nomenclature). Nomenclatural activity includes formulating new names for taxa and determining the correct name to be applied to existing taxa. The issues discussed below involve the determination of lineages leading to either a suggested split of taxa and the attribution of names to those taxa or synonymy of taxa, and thus are of a nomenclatural nature. Thirdly, the assignment of a specimen to a given taxon, or its identification, employs a name but has no bearing on nomenclature or taxonomy.
Clades 1 and 2.—
Baker et al. (2005) labeled these clades as Pachyornis mappini from eastern and western North Island, respectively. Baker (2007) modified this by calling clade 1 P. mappini and clade 2 Pachyornis, n. sp. A, which is erroneous in two regards. First, the type specimen of P. mappini is from the western North Island (Archey 1941), so that if only a single clade takes the name P. mappini, it would have to be the western population, not the eastern one. Secondly, as I have established elsewhere (Worthy 2005), the taxon formerly known as P. mappini Archey now takes the name P. geranoides (Owen), a nomenclatural change that also affects the genus Euryapteryx (see below).
Clade 6.—
Baker et al. (2005:8259) referred all specimens of Euryapteryx to E. geranoides, including birds from far north of North Island that “were previously assigned to Euryapteryx curtus,” which was followed by Baker (2007), though with the spelling of geranoides incorrect. The nomenclatural issue here is that E. curtus (Owen, 18466) has priority over E. geranoides (Owen, 1848), as is obvious by the dating of the two names and as shown in check-lists (e.g., Turbott 1990) and systematic accounts (e.g., Archey 1941, Worthy and Holdaway 2002). Baker (2007), however, also overlooked the nomenclatural issues raised by Worthy (2005), who transferred E. geranoides (Owen, 1848) to Pachyornis, where it becomes a senior synonym of P. mappini Archey. As a result, specimens referred to E. geranoides, as distinct from E. curtus, now take the next available name, E. gravis (Owen, 1870). However, if only one species name is applied to a clade including all individuals of Euryapteryx, then E. curtus has priority.
Clades 13 and 14.—
Baker et al. (2005) recognized two clades based on two individuals of Megalapteryx. They referred the northern specimen to M. didinus (Owen, 1883) and the southern one to M. benhami Archey, 1941, and questioned the synonymy of M. benhami with M. didinus (Worthy 1988). First, the type of M. didinus comes from Otago in southern South Island, and the type of M. benhami from northwest Nelson in northern South Island (Archey 1941), so the application of the names by Baker et al. (2005) and Baker (2007) to the two clades is wrong regarding the geographic origin of the types for the taxa used. Secondly, M. benhami was based on a very large specimen that was considered by Archey (1941) to be outside the size range of M. didinus. Baker et al. (2005) did not sample any specimens that could be referred to M. benhami on the basis of size. The genomic data presented in that paper therefore have no bearing on the taxonomic status of M. benhami.
Similar geographic issues of the origin of types may also affect the application of names to other clades in these studies. For example, the type of D. robustus Owen, 18466 is from “South Island” (Archey 1941), which can be restricted to “Waikawaiti” (sic = Waikouaiti) in Otago, because Owen (1846a:313, 319, 321) made it clear that the bones he named D. robustus Owen (1846b:48) came from this locality. Thus, if clades 9, 10, and 11 are to be elevated to the status of species, it is not clear which would take the name robustus, given that both clades 9 and 11 were found in Otago.
In conclusion, the phylogenomic results of Baker et al. (2005) and Baker (2007) may have merit in the recognition of unsuspected clades within moa populations, but Baker's (2007) application of existing nomenclature to these clades as though they were species, in disregard of previous literature, rather than advancing the systematics of moas, has only added confusion to an already complex state of affairs. It is lamentable that such profound taxonomic results should have been published as an aside in a review, the more so as such little understanding of nomenclatural issues was shown.
Literature Cited
Table 1.
Comparison of the nomenclature applied to clades of moas in Baker et al. (2005) and Baker (2007). Details in square brackets identify the geographic origin of the clade as shown in figure 1 in Baker et al. (2005).
