replying to: H. Püschel et al; Scientific Reportshttps://doi.org/10.1038/s41598-024-78400-8 (2024).

Patagomaia chainko Chimento et al.1 is a fossil mammal based on a partial hindlimb and pelvis coming from the Late Cretaceous (Maastrichtian) of Santa Cruz province, Argentina. In spite of being described as belonging to the therian lineage by Chimento et al.1, Puschel et al.2 proposed that it may belong to the Gondwanatheria clade. It should be mentioned that most gondwanatherians are only poorly represented, and include taxa with a great degree of missing entries. Among the taxa assigned to Gondwanatheria, the Malagasy Adalatherium hui, is the only one that can be compared with Patagomaia, having clearly associated postcranial material3,4.

Puschel et al.2 conclude that all putatively therian characters of Patagomaia are either also present in Adalatherium, or cannot be assessed in the latter because of its poor preservation. They also change scorings in the phylogenetic analysis by Chimento et al.1 and argue that when these scores are changed, gondwanatherian affinities for Patagomaia are found to be more likely.

In spite that we consider that some Puschel et al.2 comments are valuable, we think that their point is far from being well supported. The postcranium of Adalatherium shows several features that are absent in Patagomaia, including proximal end of tibia strongly bowed and that in all probability exhibited asymmetrical proximal articular facets, with the medial one much wider than the lateral one4, and incomplete acetabular dorsal rim. In fact, in Adalatherium, a ridge marks the beginning of a dorsal acetabular emargination4, a condition shared with basal mammals, including the stem-therian Vincelestes, and different from Patagomaia and most therians5,6,7,8,9,10. Nevertheless, the distorted nature of Adalatherium and the incomplete available material known for Patagomaia, make more detailed comparisons difficult.

In spite that Chimento et al.1 performed three phylogenetic analyses, using three different matrices, Puschel et al.2 focused in only one of these datasets, in which they make some changes of the scorings. Nonetheless, when Puschel et al.2 include their changes in the matrix, Patagomaia results in an unresolved polytomy (together with gondwanatherians and therians), as they show in their Fig. 2a (see also Supplementary Information). They must further adjust even more the results and make a Bayesian analysis and show a 50% Majority Rule tree, with the aim to nest Patagomaia among gondwanatherians. It is noteworthy that in the three analyses carried out by Chimento et al.1 it was not necessary to modify the analysis to obtain resolved trees. Thus, Puschel et al.2 point is far from being well supported.

Puschel et al. argue that because the evidence for therian affinities of Patagomaia is not compelling (a conclusion that we disagree), and the presence of definitive gondwanatherians in Patagonia is well-documented, it is likely that it may represents a member of the later lineage, if not Magallanodon itself (a gondwanatherian found in roughly coeval beds to those Patagomaia)11,12. We will welcome the inclusion of Patagomaia within Magallanodon when postcranial materials of the later taxon become published. However, the presence of Patagomaia in layers contemporary with those of Magallanodon is only circumstantial evidence and is not sufficient reason to support this hypothesis, which in turn might be preventing us to recognize the presence of otherwise unregistered clades (i.e., Theria). It could be alternatively argued (with a comparable degree of uncertainty) that therian affinities for Patagomaia would be expected as this would help to explain the evolutionary origin of the rich therian record in Patagonia by the early Cenozoic.

In any case, Puschel et al.2 admit that therian characters recognized by Chimento et al.1 in Patagomaia are effectively present in the material (e.g., complete dorsal rim, presence of neck, position and orientation of the lesser trochanter, a well-defined and wide trochanteric fossa, robust greater trochanter on femur). However, because some of these features are present in Adalatherium, they consider that the many similarities present in pelvis and hindlimb structure between therians and gondwanatherians reflects locomotor-related homoplasies. An alternative hypothesis is that these common characters are synapomorphic features, and that gondwanatherians may be nested among therians, as advocated by some previous authors13,14,15. Therian features reported for Patagomaia are certain, thus its inclusion within Gondwanatheria may impact into the phylogenetic affinities of this mammalian clade.

In the absence of teeth (which are the most characteristic gondwanatherian elements) and with most postcranial features being also shared with therians, we conclude that there is no hard evidence supporting the inclusion of Patagomaia in Gondwanatheria. In sum, Puschel et al.2 do not provide new materials or conclusive phylogenetic evidence, and we consider their work to be inconclusive.