The position of the New Guinea Flightless Rail, Megacrex inepta, remains unclear. Kirchman (2012) found it basal to both Rallus and Gallirallus, while Trewick's (1997) results would put it in Gallirallus, which I follow here.
This is an interesting case, and indeed not addressed at all by my previous trees, hence...:
Trewick (1997 [pdf
]), and Kirchman (2012 [pdf
]) each produced a partial cytochrome b and a partial 12S rRNA sequence of Megacrex
; to date, these four sequences are the only available genetic data for this taxon. Within each data set, the signal given by the two genes looks compatible, but the data sets conflict with one another.
I have attached a tree based on cytb+12S sequences, resulting largely from a merging of these two data sets. Trewick's sequences have accession numbers U77###, Kirchman's sequences have accession numbers JQ34#### and JQ36####--so they can easily be identified in the tree.
forms the sister group of a clade made of Amaurornis olivacea
, A. phoenicurus
, and Gallicrex cinerea
. Trewick's Megacrex
ends up in a completely different place, within the Rallus
clade, and actually appearing nearly identical to Kirchman's Gallirallus lafresnayanus
. It seems obvious from these sequences that the samples that Trewick and Kirchman each called Megacrex inepta
cannot have been derived from birds belonging to the same species. Nor even to the same genus, actually.
It's probably not possible to be 100% certain without a third study. However:
- It seems rather improbable to me that Gallirallus lafresnayanus
and Megacrex inepta
, two completely flightless rail species, one restricted to New Caledonia, the other to New Guinea, would appear genetically as close as Trewick's Megacrex
- However, externally, these two species are both large dull flightless rails with thick bill and legs, hence I presume that they could conceivably be confused for one another in a museum tray.
- Livezey (1998 [pdf
]), based on morphological characters, placed Gallirallus lafresnayanus
within the Gallirallus
group (Fig.3), and Megacrex inepta
(Fig.4). Although support was far from high, (1) this result runs nevertheless counter to them being nearly conspecific, and (2) this result is nevertheless strikingly more congruent with the relationships suggested by Kirchman's sequences, than with those suggested by Trewick's sequences.
Thus, if facing a choice, I think that I would tend to accept Kirchman's sequences, while presuming that Trewick's Megacrex
might have been a misidentified Gallirallus lafresnayanus
(I'd rather not bet my hand on this type of thing if I can avoid it, though.