As in the Tsuji (2006) analysis, Synapsida, Captorhinidae and Mesosauridae were assigned as outgoups when running the parsimony analysis in PAUP.
In addition, the scoring of Mesosauridae, Millerettidae and Captorhinidae is corrected to '0' as the maxilla and quadratojugal are clearly separated by the jugal in these taxa (Gow, 1972; Modesto, 2006; Clark and Carroll, 1973) and to '0&1' for synapsids as these bones are connected in caseids, eothyrids and some varanopids but separated in other varanopids and ophiacodontids (Berman et al., 1995; Brinkman, 1988; Langston, 1965; Laurin and Reisz, 1995; Romer and Price, 1940).
This state was correctly scored for Synapsida and Eunotosaurus but Mesosauridae were scored incorrectly, as the parasphenoid of Mesosaurus is slightly longer that it is broad (Modesto, 2006 fig.
Also, contact between paraoccipital process and the dermatocranium, which was previously scored as unknown, was scored as absent (67:4) for Mesosauridae, Eunotosaurus, and Belebey, and changed from '1' (to squamosal and supratemporal) to '4' (absent) for Millerettidae.
Firstly, Mesosauridae, which was one of the three outgroups of Tsuji (2006), is recovered within the ingroup Parareptilia because it is positioned nesting deeper in the topology than the Millerettidae-Eunotosaurus clade (B), which is called Millerosauria sensu Modesto and Reisz (2008).