GREGORY, PRESENT STATUS OF ORIGIN OF TETRAPODA 329 



lies a foramen opening into a median tube (Pander/ I860, Tab. o, Figs. 

 1-5) which has been identified by authors {e. fj., Xewberr}-, Dean, Jaekel) 

 as the pineal foramen. In Dipterus also the frontals are short, but in the 

 Palseoniscidse they are longer. The postfrontals in Osteolepidag and 

 Ehizodontid^e are elongate narrow elements extending from the postero- 

 superior margin of the orbits to the pterotics ("squamosals"). 



The parietals in the Ehipidistia are longer than the frontals, from 

 which they are separated by sharp transverse sutures. In Dipierus, on 

 the other hand, the parietals were short, in Pateoniscid^ they were short 

 and wide. 



Behind the parietals in Osteolepis microlepidotus (Pander, 18-60, Tab. 

 I) was a transverse series of small quadrangular elements, closely ap- 

 pressed to the parietals and ''pterotics," which at first sight suggest the 

 transverse occipital series of Stegocephali. The innermost or median 

 pair of this series suggests the paired dermosupraoccipitals or postparietals 

 of Stegocephali; next to these comes a second or middle pair; the third 

 pair at first suggest the tabularia, and the outermost pair in the species 

 under consideration suggest the '^epiotic cornua" of such Stegocephali as 

 Archegosaurus. 



Nevertheless, in spite of these attractive resemblances with the trans- 

 verse postparietal series in Stegocephali, I am finally constrained to treat 

 them as analogies only. First, this transverse postparietal series is best 

 developed in Osteolepis microlepidotus, is imperfectly developed in 0. 

 inacrolepidotus, Gyroptychius, Tristichopterus, and is entirely absent in 

 Rhizodopsis, MegalicMliys, Glyptopornus and Holoptychius. But struc- 

 tures that are developed only in a few forms, rather than in whole groups, 

 offer very unsafe guides for homologization with similar elements in other 

 widely separated groups, a principle too often neglected by comparative 

 anatomists. Secondly, this transverse occipital series in Osteolepis does 

 not bear a transverse line of sensory pits, as it should do if it were 

 liomologous with the transverse postparietal series of Stegocephali (cf., 

 Moodie, 1908). Thirdly, there is another row of transverse plates consti- 

 tuting the nuchal or "supratemporal series," lying behind the occipital 

 suture, which bear a transverse sensory line and are homologized bv all 

 authorities with the postparietal row in Stegocephali. This nuchal series 

 may terminate dorsally either in a single median plate, as in Dipnoi, all 

 Ehipidistia, many Actinopter3^gii ( ? supraoccipital), or in double or 

 paired median plates, as in PoJypterus, Amia, Lepidosteus and all Stego- 

 cephali. 



Conce:fning the nuchal series in Tristichopterus alaius, one of the 

 EhizodontidcT. Traquair (1875, p. 386) wrote as follows: 



