Lomasuchinae is regarded here as an inactive taxon because of the lack of effective support for the monophyly of Lomasuchus and Mahajangasuchus or, for that matter, Peirosauridae and Mahajangasuchus. Support for the former, on which the taxxon was erected (Carvalho et al. 2004), is based on a single unambiguous character (lateral contour of the dentary sigmoidal in dorsal view), which in Lomasuchus was not described as such by Gasparini et al. (1991) or figured and which in Mahajangasuchus is not apparent (Buckley and Brochu, 1999:fig. 3A). Support for the latter (Peirosauridae + Mahajangasuchus) is also based on a single, but different, character with similar lack of clarity—the thickness of the splenial at the symphysis (Buckley and Brochu, 1999:char 110). This character is preserved in neither Lomasuchus nor Mahajangasuchus, in which “damage precludes assessment of whether or not they [the splenials] participated in the mandibular symphysis (Buckley and Brochu, 1999:154).
In the analysis of Carvalho et al. (2004), they adapted the matrix of Ortega et al. (2000), but removed some taxa and added others. The result is that, of the 183 characters that they included, 53 are uninformative, 30 of which are constant. They first ran the matrix and then rewqeighted the characters to reduce the number of trees. Exactly how this was done is not clear. An unweighted analysis of the same data yields nine, rather than the reported three, minimum-length trees of 374 steps, with consistency and retention indices of 0.44 and 0.53, respectively, rather than the reported values of 0.68 and 0.83. At one additional step in length (375 steps), all nodes break down except three, and there is no resloution regarding the relationships of Mahajangasuchus and Lomasuchus. There is little support for several of the taxa erected by Carvalho et al. (2004).
Furthermore, Peirosaurus and Uberabasuchus are not most closely related, altough the latter comes from the same locality as the holotype of the former and appears to be the same taxon. The line of data for “Peirosaurus” in the matrix, apparently, is based on the specimen from Argentina that Gasparini et al. (1991) referred to P. torminni, the basis of which has not been clarified. The anterior portion of the premaxilla, the most diagnostic portion of the holotype of Peirosaurus torminni, is not preserved in the Argentine specimen.