was placed in a matrix of 31 characters and 14 taxa to test its relationships among archosauromorphs. Included taxa (Appendix 1) are from a variety of basal archosauromorphs known from relatively complete material and included in many past phylogenetic studies (e.g.,
Gower and Sennikov 1997;
Senter 2004). Characters were culled from several of these past studies (Appendix 2).
The matrix (Appendix 3) was analyzed in PAUP* version 4.0b10 for 32-bit Microsoft Windows (Swofford 2002). The characters were unordered and of equal weight. One character (number 3) was parsimony-uninformative but was retained for use in future expansion of the matrix.
Broom was set as the out-group taxon, and the tree was rooted with the out-group as paraphyletic. The default settings in PAUP* for a branch and bound search were used and resulted in 15 most parsimonious trees (MPTs) with a length of 60 steps, a consistency index of .5833, a retention index of .7283, and a rescaled consistency index of .4248. Decay indices were determined using the 'keep' command in PAUP* and compiling consensus trees longer than the original consensus tree until all nodes collapsed. Bootstrap values are based on 1,000 replications.
In the strict concensus tree (Figure 11),
nests within Archosauriformes in a polytomy with
, and Archosauria (sensu
Gauthier 1986) and is, thus according to this analysis, closer to Archosauria than is
. However, this idea is weakly supported (Bootstrap value of 40% for 1000 replicates), and only a single step (Tree length = 61) forms an unresolved polytomy between
, and Archosauria. Better supported is the more advanced placement of
among basal archosauriformes as two additional steps (Tree length = 62) are necessary for
to clade (unresolved) with
. The clade Archosauriformes collapses at 63 steps.
Unambiguous synapomorphies supporting the inclusion of
into Archosauriformes include the lack of post-axial intercentra, a femur with a medially inflected head, a sigmoidal femoral shaft, the lack of a distinct intertrochanteric fossa on the proximoventral surface of the femur, and the presence of osteoderms.
Long and Murry (1995), unsure of its taxonomic position, originally assigned
but suggested that it might represent an archosauromorph, whereas
Small and Downs (2002)
postulated that the Ghost Ranch form possessed archosauriform affinities. The current study tentatively supports these hypotheses; however, the inclusion of skull, braincase, and tarsal characters into the matrix based on the Ghost Ranch material is needed to confirm these taxonomic affinities.