OSTEOLOGY OF AMERICAN PERMIAN VERTEBRATES 369 



believe that the temporal region was closed in. I spent many 

 hours in the endeavor to reach an understanding of the temporal 

 structure, which appeared to be so inconsistent with that of the 

 remainder of the skeleton. That there was only an upper temporal 

 vacuity did not at this time occur to me, since we know of no 

 other paleozoic reptile having such a structure. I had almost 

 despaired of the solution of the riddle when an additional skull 

 was found which shows on the left side the temporal region in an 

 almost perfect condition. With this specimen as a key, the struc- 

 ture of the skull was readily and completely unlocked. 



The drawings of the skull herewith given (Fig. 3, A, B) are 

 composites derived from all the different specimens, though 

 chiefly based on one of about two-thirds adult size. Nearly 

 every character shown or described has been corroborated by 

 several specimens; in the palate only, less corroborative material 

 is available. 



In only one specimen is there anything of the premaxillae pre- 

 served. In the reconstruction of the general form of the skull 

 I believe that the figures are tolerably accurate, though I cannot 

 be sure that the skull may not have been a little wider or narrower 

 or higher than I have shown it. It is difficult to make exact 

 measurements of such small bones. The preparation of the speci- 

 mens has required very patient work with a fine needle under a 

 dissecting microscope, for the most part, but the bones, when care- 

 fully prepared, come out clean and smooth. 



The parietals are longer than broad. In one excellent speci- 

 men found later, the distinguishing suture in front is clearly shown. 

 The parietal foramen is large, as shown in three skulls; it is situated 

 toward the front part of the parietals, just a little behind the 

 posterior margins of the orbits. Posteriorly, the parietals are 

 produced into an acute angle in the middle, projecting slightly 

 over the supraoccipitals. I can find in no specimen the shghtest 

 indication of a distinct dermosupraoccipital, on the upper surface 

 of the skull at least. On the posterior outer angle each parietal 

 is produced into a slender process, as in lizards. In two specimens, 

 the articulating tabulare was cleanly removed, showing a striated 

 sutural surface near the angle. The arch thus formed, of parietal 



