CLASSIFICATION BY DESCENT. 129 



Harpoceras in its restricted sense ; but c contains Am. aalensis, i. e. Grammoceras, 

 Am. opalinus, i. e. Lioceras, Am. Murchisonce, i. e. Ludwigia. 



Now, if we were to take Ludwigia away and place it in the other group, would 

 the remaining forms fall naturally into another separate group ? This must be 

 answered in the negative. Harpoceras and Lioceras are, in my opinion, separated 

 as widely from Grammoceras as they are from the Division I. We have therefore 

 three main divisions to deal with, and their relationship to one another and to 

 Arietites is as follows : 



Beanch A. — Eildoceras, Lillia, Ludwigia, &c. This is the branch which bears 

 the most similarity to the form of the parent stem, and whose changes are of recent 

 date so far as the Inferior Oolite is concerned. 



Branch B. 1 — Harpoceras, Lioceras, Pseudolioceras, &c. A branch from the 

 original stem which has undergone considerable modification prior to the Upper 

 Lias, but has since then been somewhat stationary. In the Upper Lias it is far 

 ahead of the species of Branch A in development ; but the latter have managed to 

 become close rivals to it in the Inferior Oolite, and hence we have convergence 

 between Ludwigia and Lioceras. 



Branch C. — Grammoceras. A branch which left the parent stem at an early 

 date. It has undergone considerable modification prior to the zone of Am. 

 Jamesoni ; but since then it has been somewhat stationary. Hence at the top of 

 the Upper Lias the less-developed members of Branch A come to resemble it, 

 leading to convergence between Hildoceras and Grammoceras ; while the higher 

 developed members of that branch have passed it. 



Perhaps the simplest form of classification would be to include all these genera 

 in one family, Hildoceratidte , 2 divisible into the Branches A, B, C. 3 To Harpoceras, 

 as used by Zittel, the above-constituted family, Hildoceratidce, is probably equal 

 in scope ; but in rank it should be equal to his Harjwceratidce, the only difference 

 being that certain genera, which belong more appropriately to other families, have 

 been eliminated. For instance, the Ammonites, which may be grouped around 

 Am. insignis, Zieten, and Am. fallax, Benecke (Hammatoceras pars), are descended 

 from some form allied to Deroceras, or Microderoceras (JEgoceras, "Waag. pars,), and 

 are therefore more closely allied to Stephanoceras ; the Ammonites which form the 



1 The groups of Am. trimarginatus and Am. canaliculatus may be descendants of the forms of 

 this branch. 



2 It is preferable to use Hyatt's name Hildoceratidse in this present case on account of its priority 

 and fitness. In this family he placed the genera Hildoceras, Grammoceras, and Lioceras, which are the 

 foundations of our present divisions. Thus the family Hildoceratidze closely represents the limits 

 which we wish to assign to the present family, besides having priority over every other in date. The 

 family Hildoceratidse should contain genera descended from Arietites, but no others. 



3 The Branches A, B, C are three perfectly distinct, but contemporaneous, genetic series having 

 nothing to do with each other except through their common Lower-Lias ancestor. 



17 



