22 TYPE AMMONITES— VI June 



This would make the ParapatoceratidzE an aberrant off-shoot of primitive 

 Kosmocerataceae. A similar theory would make the Stepheoceratan- 

 Parkinsonian uncoiled forms, Spiroceras and others, aberrant develop- 

 ments of primitive Parkinsonidae. 



Such a theory receives some support from the existence in the 

 Kosmoceratan of a small unnamed Kosmoceratid with polygyral serpenti- 

 cone whorls barely in contact. Such a form is, presumably, a rather 

 primitive Kosmoceratid, which, with a very little disturbance, might 

 easily become an uncoiled form. A similar, little-in-contact form in 

 the Proplanulitan Age, congruous with young Toricellites, may be 

 imagined as ancestral to the Parapatoceratidae. 



Uncoiled forms of widely-separated dates are obviously polyphyletic. 

 Uncoiled forms of approximately similar date are not necessarily mono- 

 phyletic. So the family Parapatoceratida; may not be systematically 

 correct : it may be only a convenience. Aberrantism may be looked 

 upon as somewhat of a disease, which may have attacked more than 

 one of the genera of Kosmocerataceje to produce forms like Crioconites 

 and Parapatoceras. 



Kosmoccratidae. Most of these genera are remarkable for length 

 of lateral and ventral tuberculation, truly spinosity, and for the number 

 of ribs which may be collected to join to one ventral tubercle (spine). 

 In the order of declining collection of ribs the genera seem to stand 

 as follows : — Kosmoceras , Spinikosmokeras, Hoplikosmokeras, Lobo- 

 kosmokeras, Kalakosmokeras, Bikosmokeras, and the supposed decadent 

 genera, Zngokosmokeras and Kuklokosmokeras. Kosmoceras and Spini- 

 kosmokeras stand at what may be called the acme of rib-collection to 

 one ventral tubercle, 3 or 4 or even more ribs flowing into one ventral 

 tubercle, while the genera placed after them in the foregoing sentence 

 show decline, in steps, down to unity, with final loss of ventral tubercles. 



The Kosmoccratidae are distinct from the Gulielmiceratidas in 

 attaining a rounded venter without losing costate ornament. 



The genera and families of the Kosmocerataceae show the following 

 phenomena : — i, Faunal repetition : they come in as independent waves 

 not only during two Ages, but during the successive hemera; of those 

 Ages, then they disappear for an Age — Reineckeian — not only from 

 this country, where their absence might be explained by loss of 

 Reineckeian deposits, but, so far as is known, from other places where 

 Reineckeian deposits are well-developed. Then in a later Age, Kosmo- 

 ceratan, they re-appear in abundance, show very flourishing stages of 

 high development — heterogeneity — gradually fading away to less hetero- 

 geneity of characters, but not attaining the simplification — loss of 

 ornament and other features — of even the Gulielmiceratidse. Recog- 

 nisable Kosmoceratids continue into early Vertumniccratan (Professor 

 A. Morley Davies's finds), and then perhaps all die out ; but there is 

 a possibility that genera like Kuklokosmokeras have given rise to forms 

 whose Kosmoceratid affinity would not be readily suspected. 



The phenomena shown by genera of the Kosmocerataceae in regard 

 to evolution of characters and the times at which they appear seem 

 to be congruous with those noted for the Oppelaceae (T.A. V, pp. 7 

 et seqq.). It is the later forms which show the greater amount of acmic 

 characters or high elaboration ; while it is the earlier forms which reveal 

 the larger amount of paracmic or declining characters — none of the 



