Proposal (338) to South American Classification Committee

Change linear sequence of species in the genus Turdus

Effect on SACC: This would modify our current sequence to reflect the phylogenetic hypothesis in Voelker et al. (2007).

Background: Our current sequence is as follows. Other than placing presumed sisters adjacent, I am not sure that there is any published rationale for the sequence of species groups.

Turdus flavipes
Turdus leucops
Turdus fuscater
Turdus chiguanco
Turdus serranus
Turdus nigriceps
Turdus reevei
Turdus olivater
Turdus maranonicus
Turdus fulviventris
Turdus rufiventris
Turdus falcklandii
Turdus leucomelas
Turdus amaurochalinus
Turdus ignobilis
Turdus lawrencii
Turdus fumigatus
Turdus obsoletus
Turdus hauxwelli
Turdus haplochrous
Turdus grayi
Turdus nudigenis
Turdus assimilis
Turdus albicollis

New data: Voelker et al. (2007) sequenced mtDNA (1000 bp of cyt-b, 333 bp of ND3, and 1035 bp of ND2) for 55 Turdus spp. and smaller samples for another 5 species to produce a phylogeny for 60 of 65 species (!). Their tree has a high number of strongly supported nodes. South American species all fall into one clade that includes Nesocichla from Tristan da Cunha, a few species from Middle America, and two from Africa. "Platycichla" falls within this clade - we already merged Platycichla into Turdus based on earlier papers.

Translating all that to a linear sequence, using the usual conventions ("basal" taxa first; for sister taxa, NW-most taxon listed first; for polytomies, stay as close to traditional sequence as possible), the result is:

leucops (no close relatives; unsupported branch links it to Old World pelios; support for overall placement is weak but falls outside a well-supported group that includes all the rest of SA Turdus)

falcklandii (support for overall placement is weak but falls outside a well-supported group that includes all the rest of SA Turdus, which is biogeographically surprising)

reevei (no close relatives but comes out somewhere near base of SA species)

flavipes (ex-Platycichla; same comment as for reevei)

leucomelas (sister to fumigatus + hauxwelli)

fumigatus (sister to hauxwelli)

hauxwelli (sister to fumigatus)

obsoletus (support for exact placement in SA species is weak; possibly sister to following species group)

rufiventris ("basal" position in strongly supported group that includes next 5 species)

maculirostris (this and next 3 form a strongly supported group, but internal nodes have essentially no support)

grayi (this with the next two form a polytomy - listed N > S)

nudigenis (note: maculirostris currently treated as a subspecies of this, but they are not particularly close genetically; proposal needed)

haplochrous

lawrencii (sister to a strongly supported group that includes next 3)

amaurochalinus (sister to ignobilis + maranonicus)

ignobilis (sister to maranonicus)

maranonicus (sister to ignobilis)

fulviventris (part of a group that includes all species below, but placement uncertain)

olivater (sister to a group consisting of next 4)

nigriceps (sister to a group consisting of next 3)

fuscater (sister to serranus + chiguanco)

chiguanco (sister to serranus; somewhat surprising)

serranus (sister to chiguanco; somewhat surprising)

assimilis (sister to albicollis)

albicollis (sister to assimilis)

Recommendation: Regardless of any minor problems that might arise with this new sequence, it is backed by phylogenetic data, in contrast to the traditional sequence, which is maintained solely by historical momentum; therefore, I recommend YES.

Lit Cit:

VOELKER, G., S. ROHWER, R. C. K. BOWIE & D. C. OUTLAW. 2007. Molecular systematics of a speciose, cosmopolitan songbird genus: defining the limits of, and relationships among, the Turdus thrushes. Molecular Phylogenetics and Evolution 42: 422-434.

Van Remsen (in consultation with Gary Voelker), March 2008

________________________________________________________________________________________________________

Comments from Stiles: "YES. It is clearly an improvement over the traditional sequence, which has no consistent basis and is at odds with numerous points of the now quite well-resolved phylogeny."

Comments from Nores: "YES. Esta secuencia resulta mucho más coherente que la anterior, pero de todos modos hay algunas cosas para mí no muy convincentes. Por ejemplo, la relación entre las especies sudamericanas con Turdus pelios; la separación tan marcada de las especies del ex-Platycichla; y sobre todo la inclusión de Nesocichla en Turdus."

Comments from Jaramillo: "YES - much improved sequence, and based on good data. This makes much more sense that the traditional arrangement."