The phylogeny of 18 taxa in the Lanius excubitor complex, and the related species L. sphenocercus, L. ludovicianus and L. somalicus, was estimated based on the mitochondrial cytochrome b gene and the non-coding D-loop (in total 1.3 kb). According to the mitochondrial gene tree, Lanius excubitor s.l. is non-monophyletic, with some of its subspecies being more closely related to L. sphenocercus, L. ludovicianus, and L. somalicus. Also the division of the L. excubitor complex into a northern (L. excubitor) and a southern (L. meridionalis) species, as has been proposed based on morphological and ecological similarity and geographical distributions, is not compatible with the mitochondrial tree. Overall, genetic divergences among the ingroup taxa are small, indicating a recent radiation. A tree based on the nuclear ornithine decarboxylase (ODC) introns 6–7 is unresolved with respect to the ingroup, but provides strong support for a clade containing the Lanius excubitor complex, L. sphenocercus, L. ludovicianus and L. somalicus. We discuss the incongruence between the current taxonomy and the mitochondrial gene tree, and conclude that based on the latter the Lanius excubitor complex may be treated as at least six species, L. borealis, L. elegans, L. excubitor, L. lahtora, L. meridionalis, and L. uncinatus, but that other taxonomic treatments are also possible. However, uncertainty regarding to which extent the mitochondrial gene tree reflects the species phylogeny prevents us from recommending taxonomic change without further investigation. This study highlights the possible danger of relying on a single molecular marker, such as mitochondrial DNA, in taxonomic revisions and phylogenetic inference.

Keywords