Sure Proterosuchus (Fig. 2) nests at the base of the Archosauriformes, but it had already strongly evolved into the long, low, hook-jawed morphology of other proterosuchids following the path set out by younginids like this one.
Garjainia is just a giant Youngina.
And size makes all the difference. With that large, tall skull, small pectoral girdle, separate pubis and ischium and relatively short tail, Garjainia shares several traits with Youngina (Fig. 1) that are not present in Proterosuchus (Fig. 2), a more basal archosauriform that took a different, more amphibious path.
In Garjainia the skull is relatively larger. The naris and rostral profile are raised, the teeth are larger but fewer, the maxilla is deeper and convex ventrally, the orbit is smaller, the antorbital fenestra included a fossa dorsal to the fenestra. The lateral temporal fenestra was invaded posteriorly by the squamosal and quadratojugal. Whereas sweet gentle Youngina was a sweet little insect-eater that occasionally chomped on small vertebrates, Garjainia was a fearless giant (for its time) killer and everything was on its menu.
Neck, torso and tail
To carry such a large skull the cervicals were larger and taller. The chest was deeper and the neural spines were taller to anchor larger back muscles. The tail was relatively shorter with shorter, taller vertebrae. It was also more heavily muscled.
Pectoral girdle and fore limbs
The pectoral girdle was little changed but the humerus and forearm were much more robust in Garjainia, again to carry the weight and move rapidly when necessary.
Pelvic girdle and hind limbs
The ilium and ischium both extended posteriorly. The pubis lost virtually all connection to the ischium and became transversely wider rather than anteroposteriorly long, creating the basic archosauriform pubis that so many of them retained. The hind limbs were more robust, of course.
Youngina is NOT ancestral to Lepidosauromorpha!
Wikipedia promotes a completely misguided (yet traditional) family tree of the Reptilia that nests Youngina and Thadeosaurus then Claudiosaurus as basal to Archosauromorpha + Lepidosauromorpha. Nothing could be further from the truth! This comes from a misguided and traditional hypothesis that basal lepidosauromorphs had a complete lower temporal bar, as in Sphenodon. This is wrong, as we learned earlier. This mistake also comes from not having a large gamut reptile tree that documents where every one of 344 taxa nest in complete resolution, rather than cherry picking a few favorite taxa to test.
Nesbitt 2011, for reasons unknown, put the proto-rhynchosaur Mesosuchus beyond the base of the Archosauriformes, rather than Youngina (taxon exclusion). He employed the very derived Erythrosuchus in his tree, rather than the more primitive Garjainia. For good measure he threw in the thalattosaur, Vancleavea. We talked about that earlier here, here and here.
Sorry that traditional paleontologists have messed things up so badly. Trying to fix it one step at a time.