Parasuchians (phytosaurs) are those very croc-like Triassic swamp giants with a nostril rising on a bony volcano almost between their eyes (Fig. 1).
Figure 1. A selection of phytosaurs (parasuchians).
Parasuchians all have a similar appearance.
The question is, where did they come from? Which taxa are their closest ancestors?
Nesbitt (2011)
nested parasuchians between Euparkeria and Archosauria (Ornithodira (including pterosaurs) + pseudosuchia. Apparently it didn’t matter to Nesbitt’s study that his parasuchians didn’t resemble the most closely nested taxa.
Brusatte et al. (2010)
nested parasuchians between Proterochaampsidae and Aetosauria + the rest of the Pseudosuchia; or (when pterosaurs were removed) between Revueltosaurus and Aetosauria + the rest of the Pseudosuchia. In both cases the Avemetatarsalia (pterosaurs + dinosaurs and kin) were considered closely related. So again, not many taxa here display a gradual accumulation of parasuchian traits.
According to the large reptile tree
everything becomes much more clear and a gradual accumulation of parasuchian traits is clearly visible in ancestral taxa (Fig 2).
Figure 2. The origin of the Parasuchia (Phytosauria) with Diandongosuchus, Mesorhinosuchus and related taxa. This series demonstrates a gradual accumulation of parasuchian traits. It would be nice to find one with the nostrils midway on the snout.
Taxon inclusion is key to this understanding. Using specimens rather than suprageneric taxa is also important.
References
Brusatte SL , Benton MJ , Desojo JB and Langer MC 2010. The higher-level phylogeny of Archosauria (Tetrapoda: Diapsida), Journal of Systematic Palaeontology, 8:1, 3-47.
Nesbitt SJ 2011. The early evolution of archosaurs: relationships and the origin of major clades. Bulletin of the American Museum of Natural History 352: 292 pp.
The Pterosaur Heresies 
Phytosaurs phorever!
So you link phytosaurs to Proterosuchus? Interesting, I thought that there were numerous characteristics in contrast to this interpretation. The presence of osteoderms, the absence of intercentra, the loss of the ectepicondylar flange of the humerus, and the overall specialization of the ankle, braincase, etc. It seems like there is significant evidence that phytosaurs were significantly more active reptiles similar to something like Euparkeria rather than Proterosuchus.
Knowing that taxa often present convergent traits, I rely on the LRT and PAUP software to provide results. It is important to add that several taxa intervene between phytosaurs and Proterosuchus, but the latter is the deep ancestor of the former according to the LRT.
True, but the accuracy of the LRT depends on its character selection. If I were to make a phylogenetic analysis as broad as yours, I would include every well-defined character that has ever been published. Can you assure me that you included important features such as those I’ve listed above? Because if not, I worry that you may be going down the wrong path.
Set your worries aside. Character selection is the least important part of the LRT, as proven by how terrible the character selection is (according to many assessments). Another set of characters would result in the same tree given the same taxon inclusion list. It’s all about the taxon list. BTW, you are welcome to produce two taxa from the LRT that should not be sisters (or sister clades), so long as you re-nest one of them with another included taxon that you think would make a better sister. Confirm or refute. That’s what I do. That’s what scientists do.