JAMES PERRIN SMITH 25 



has a keeled lineage extending back to the Devonian, while the keeled 

 Ceratites extend no further back than Middle Triassic. But the tendency 

 to form a keel sometimes crops out even in the ancestry of Ceratites, since 

 at least one species of Lecanites has shown this character. And both 

 stocks appear to have come from the same Devonian genus, Gepliyroceras, 

 Ceratites from the main group, and Eutomoceras from the keeled sub- 

 genus Timanites, as shown in tliis series : 



Gephyroceras-Lecanites-Meekoceras-Ceratites; 



Timanites-Dalmatites-Hungarites-Eutomoceras; 



Timanites-Aspenites-Hedenstroemia-Pinacoceratidae. 



It w^ould seem that there may have been in the descendants of the 

 Gephyroceratida3 a strong tendency to form keels. This was already 

 present in Timanites, a subgenus and contemporary of Gepliyroceras, and 

 is continuous in the Hungaritid^e and Hedenstrcemin^e, which branched 

 out from Timanites, as shown in Longobardites, PI. IX, figs. 14-16. The 

 same character appears belated in the keeled Ceratites, certainly not in- 

 herited from the collateral Timanites branch, and not known to have been 

 present in the ancestor of the two stocks. 



Equally difficult to explain is the apparent genesis of the polyphy- 

 letic genus Trackyceras from the two lines, one from Meekoceras-Cera- 

 tites, the other from Tirolites. To state that both lines had a strong 

 tendency to develop rough shells, a median furrow, and complex septa 

 does not explain the phenomenon. Nor yet does it explain the strong re- 

 semblance of mature Sagenites of the Tropitidte to Trackyceras, so strong, 

 in fact, that careful paleontologists have confused them, although their 

 ontogeny separates them at once. 



The term orthogenesis is a statement of a fact, rather than an 

 explanation. Ammonites have developed constantly in certain directions, 

 in form and ornamentation of the shell, and increasing complexity of 

 septation, in parallel series coming from the same or nearly related 

 ancestors, as well as in series coming from different ancestors. In neither 

 case are the characters hereditary, though in both cases the tendency 

 to develop those characters seems to have been hereditary. Genera 

 derived from nearly related ancestors have frequently become more alike 

 with the lapse of time, and this has also occurred often with genera 

 whose ancestry was wholly different. This has made the study of 

 Ammonite-pliylogeny exceedingly difficult; in it fact and fancy have 

 been so mixed that it has sometimes been called the "happy hunting 

 ground" of theorists. But it has also been the happy hunting ground 



