Corrigendum: A Short-Snouted, Middle Triassic Phytosaur and its Implications for the Morphological Evolution and Biogeography of Phytosauria

This corrects the article DOI: 10.1038/srep46028.

220. This character was scored as (1) by Li et al. 8 . Here we interpret the acromion process of the scapula of Diandongosuchus as flat and in the same plane as the ventral edge of the scapula. The state for Diandongosuchus is identical to that of phytosaurs. Therefore, we scored this character as state (0). 223. This character was scored as unknown (?) by Li et al. 8 , but we changed it to state (0), short postglenoid process.
226. This character was scored as state (0) by Li et al. 8 . Here we changed the score to state (1).
Diandongosuchus, like all other phytosaurs, has a distinctly hooked anterior margin of the coracoid. This was reconstructed incorrectly by Li et al. 8 when it appeared that the scapula covered the coracoid notch and formed a coracoid foramen; reinterpretation of the articular surfaces puts the scapula more posterior on the coracoid and reveals the distinct anterior hook of the coracoid. Additionally, Diandongosuchus and other phytosaurs also lack a coracoid foramen.
228. This character was scored as unknown (?) by Li et al. 8 . The posterodorsal edge of the coracoid lacks a distinct groove so it was changed to state (0).
237. This character was scored as unknown (?) by Li et al. 8 , but we changed it to state (0). The lateral surface of the right ulna is exposed in Diandongosuchus, and it appears that the surface is 6 rounded without a distinct lateral tuber as in taxa scored as state (0). The morphology of the ulna is nearly identical to that of other phytosaurs (e.g., Smilosuchus gregorii, USNM 18313; 1 : figure   31E).
238. This character was scored as state (1) by Li et al. 8 , but we changed it to state (0). The lateral surface of the right ulna is exposed in Diandongosuchus and the distal surface is rounded similar to that of Smilosuchus gregorii (USNM 18313; 1 :figure 30A) and lacks the distinct squared-off morphology of loricatans.
239. This character was scored as unknown (?) by Li et al. 8 . Here, we scored the character as state (0) because the distal end of the ulna is compressed like that of other phytosaurs and stem archosaurs.
242. This character was scored as state (0) by Li et al. 8 , but is rescored as unknown (?) because the proximal carpals are not preserved.
243. This character was scored as state (0) by Li et al. 8 , but is rescored as unknown (?) because the proximal carpals are not preserved.
246. This character was scored as unknown (?) by Li et al. 8 . Here, we rescored the character as state (0). The proximal ends of metacarpals I-III of the left manus of Diandongosuchus remain in articulation and have the plesiomorphic character state of having a simple overlap with one another. 7 247. This character was scored as unknown (?) by Li et al. 8 . We took measurements of the forelimb elements and concluded that the manus accounts for less than 0.3 of the total length of the humerus + radius. Hence, we scored this character as state (0). 250. This character was scored as state (0) by Li et al. 8 , but we now think that this character cannot be scored because of poor preservation of the manus. Therefore, we scored this character as unknown (?).
256. This character was scored as state (0) by Li et al. 8 , but we now think that this character cannot be scored because of poor preservation. Therefore, we scored this character as unknown (?).
257. This character was scored as state (1) by Li et al. 8 . Here we revised this score as state (0) because the one manual ungual that is preserved in Diandongosuchus, although it has a sharp tip, is rather short and blunt, and it is not recurved like that of theropod dinosaurs.
259. This character was scored as state (0) by Li et al. 8 , but we now think that this character cannot be scored because of poor preservation. Therefore, we scored this character as unknown (?).
260. This character was scored as state (1) by Li et al. 8 , but we now think that this character cannot be scored because of poor preservation. Therefore, we scored this character as unknown (?). 8 261. This character was scored as state (0) by Li et al. 8 , but we now think that this character cannot be scored because of poor preservation. Therefore, we scored this character as unknown (?).
269. This character was scored as unknown (?) by Li et al. 8 . The preacetabular process of the ilium is visible in medial view, and it is clear that the preacetabular process is present and short. Therefore, we scored the character as state (0). 275. This character was scored as state (0) by Li et al. 8 . However, the dorsal margin of the ilium is covered by the femur. Therefore, we scored this character as unknown (?).  (2), obturator foramen modified into a notch that opens medioventrally. A similar notch is also present in Smilosuchus gregorii (USNM 18313), and this character was scored as small (0) for Parasuchus based on the description by Chatterjee 10 . 283. This character was scored as state (1) by Li et al. 8 . However, The pubic shaft of Diandongosuchus is similar throughout its length and does not expand posteriorly into a boot at its distolateral margin as in Arizonasaurus babbitti (MSM 4590) and Postosuchus alisonae (UNC 15575). Therefore, we scored this character as state (0). 9 284. This character was scored as state (0) by Li et al. 8 . But, because character 283 is now scored as state (0), character 284 is scored as inapplicable (-).
285. This character was scored as state (1) by Li et al. 8 . But, because character 283 is now scored as state (0), character 285 is scored as inapplicable (-).
286. This character was scored as state (1) by Li et al. 8 . However, this portion of the articular surfaces of the pubis are hidden. Therefore, we score this character as unknown (?).
287. This character was scored as state (1) by Li et al. 8 . Although the pelvis of Diandongosuchus is disarticulated, the anteroventral edge of the right ischium indicates that the contact between the ischium and pubis extended ventrally like that of Euparkeria capensis. Here, we rescored the character as state (0). 288. This character was scored as state (1) by Li et al. 8 . The proximal articulation surface between the pubes is not relatively thickened relative to the rest of the pubis in Diandongosuchus in comparison with poposauroids (e.g., Shuvosaurus inexpectatus). Therefore, we scored this character as state (0). (1) by Li et al. 8 . The ischia of Diandongosuchus meet only on the ventromedial edge and do not have extensive contact like those of paracrocodylomorphs 1 and, hence, we scored this character as state (0). 14 376. This character was scored as state (2) by Li et al. 8 . The calcaneal tuber is not just short of twice the mediolateral width of the fibular facet; it is more likely about as tall as broad.

This character was scored as state
Therefore, we rescored this character as state (1). 377. This character was scored as state (2) by Li et al. 8 . The orientation of the calcaneum with respect to the astragalus of Diandongosuchus cannot be determined because the ankle region is slightly disarticulated and compressed. Therefore, we rescored this character as unknown (?).
379. This character was scored as unknown (?) by Li et al. 8 . The calcaneum of Diandongosuchus retains the short processes and is rescored as state (0).  (1) straight. (Fig. S3) The proximal half of the lateral margin of the humerus is distinctly concave in Batrachotomus (SMNS 80275; Fig. 3a), with the proximal quarter nearest the deltopectoral crest angled outward (state 0). Diandongosuchus (Fig. 3c) and phytosaurs are scored as state (1), where the proximal quarter of the lateral margin does not expand outward but keeps the proximal half of the lateral margin in a straight plane.

IV. Support for Phytosauria outside of Archosaurs
As recovered by Nesbitt 1 , we continue to find Phytosauria outside of Archosauria (Fig. S4, though this is contrary to our results using the matrix by Ezcurra 12 ; Fig. S5). In his analysis, Nesbitt 1 provided the following character support for this topology that we review here with respect to Diandongosuchus.
Character 32: Palatal processes of maxillae do not meet on midline. This cannot currently be assessed for Diandongosuchus because of the compressed preservation of the holotype skull.