I would definitely not base any taxonomic recommendation on this but, for the entertainment, I have attached an unbootstrapped tree including all available cox1 Ploceidae sequences. I used the same colour codes for the clades as in the [graphical abstract
] of the recent paper.
Many of these sequences are from Sonet et al. 2011 [pdf here
] (pp.117-131): these are short (~300bp) to very short (100bp) sequences obtained from old to very old museum specimens; some of them may have problems -- in fact for some of them I see a strong suggestion that they do
have problems. Almost all the sequences in the data set include the region corresponding to the 100bp of the shortest ones (the two that do not only lack a couple of bases at the start of this region), hence artefacts associated to some sequences having no overlapping DNA at all should at least be absent. Due to the shortness of many of the sequences, bootstrapping this data set would largely destroy the structure that can be seen in the unbootstrapped tree; this structure is not to be regarded as having received statistical support; yet, it does
reflect the divergence levels that are present between the sequences of the data set.
The length of each sequence is given in parentheses, in the tip label, before the GenBank/BOLD accession number.