464 BARON FRANCIS NOPCSA ON THE 



111 1922 Abal (2) dropped the idea tiiat birds and Dinosaurs 

 originated from quadrupedal arboreal forms, and suggested that 

 birds might be the offspring of a bipedal Tarsius-shsijyed reptile. 

 Heilman tried (11) to reconcile my views with those held by 

 Hay and Abel, etc. 



It is apparent from this review, that anatomical arguments 

 against the hypothesis of a running Proavis are mainly based 

 on the structure of the manus and the pes, and partly on the 

 occurrence of certain quills. None of the authors cited has 

 considered either the shoulder-girdle or the pelvis of birds, in 

 spite of these parts being correlated with the feet. 



Since Hay and Abel's argument from the occurrence of a 

 functional hallux in different birds is of great anatomical weight, 

 it has to be dealt with first. The functional hallux in many 

 birds must either be an old structure inherited from an arboreal 

 Proavis, or it must be a newly-acquired character that was 

 missing in the running Proavis. 



Although we know comparatively little about the Triassic 

 bipedal reptiles among which the running Proavis might be 

 looked for, nevertheless by the study of fossil footprints some 

 points of the structure of the feet of bipedal Triassic reptiles can 

 be elucidated (13). In my argument, four types of Triassic 

 footprints are of special interest, for they throw some light on 

 the evolution of the hallux in these otherwise nearly unknown 

 types. These four types can be grouped ai-ound the footprints 

 called Plectoterna, Amonopits, Anchisauripus, and Gigandipus. 

 Believing that I have proved all Dinosaurs to have developed 

 from bipedal lightly-built animals (15), I consider these tracks 

 to be Dinosaurian. 



Plectoterna has long and slender toes, and a very long metatarsus 

 resting always on the ground and indicating a plantigrade animal 

 like Hallopus (text-fig. 1). Aonmopus has also slender toes, but 

 it shows, instead of the metatarsal impression only, the impres- 

 sion of a rounded pad, proving that this animal was digitigrade, 

 as is Procompsognathus (text-fig. 2). Anchisauo'ipus has thicker 

 toes than either of these two types, and was frequently planti- 

 grade. It is the track of Anchisaurus (10) (text-fig. 3). Gigan- 

 dipus is of still heavier build, but always digitigrade, and the 

 track of some unknown Megalosaurian. 



Around each of these four types several other Triassic foot- 

 prints can be grouped that differ mainly in the development of 

 the hallux. To the Plectoterna type belong Palamopus, Uxocanijje, 

 Herpedactylus. Xipliopeza^ and Gorvipes ; to the Avimopus type 

 are allied EupaUmiopus, Polemarchus, and Platypterna ; the 

 AncMsauripus type is suggested by Anomoepus, Apatichnus, and 

 Grallator ; and the Gigandipus type is approached by Sauropus 

 and Hyphepus. 



Considering the hallux of these different tracks, it may be 

 remarked that in some the hallux articulates very low down and 

 is not rotated {Palamopus, Exooampe, Anomoipus) ; in others the 



