ANAGENESIS. 73 



There is a similar succession in the evolution of the varietal and specific 

 characteristics of the series. Thus, in Plate II., Fig. 10, 11, and 15 repre- 

 sent the extreme varieties of Am. semicostatwn, the former with immature pilae 

 without channels, the second with well developed pilae without channels, and 

 the third with well developed pilae and distinct channels. 1 The young of Am. 

 tar deer escens, Fig. 19, shows that the derivation of this species was prohably 

 from the unchannelled forms of Am. semicostatwn similar to Fig. 11. The 

 same is true of Am. Bodlcyi with reference to Hartmanni, as is shown by com- 

 paring the young of the former, Plate II. Fig. 23, with the adult of Hartmanni, 

 Fig. 17, which had very slight channels even in the adult, and this is still more 

 apparent in the involute flattened form of Fig. 24, in which the channels 

 were earlier developed. The close connection of all its characters, both of 

 young and full grown, with its immediate ancestor, forbids us imagining an 

 independent descent from a variety of any other species than Hartmanni, and 

 the evidence is strong that it had no descendants beyond its own species. 

 The channelless variety of Am. falcaries, Fig. 26, gives similar evidence of its 

 genetic connection with the channelless varieties of Am. Hartmanni, or if this is 

 doubted, a more direct connection with Am. semicostatum may be claimed ; but 

 certainly there is no evidence for any connection with the channelled varieties 

 of semicostatum or Hartmanni. 



The pilae began with psiloceran-like immature folds, after the keel appeared 

 in the arnioceran varieties of Cor. hridion, and before the keel in the more acceler- 

 ated development of other varieties of the same species. In some forms the 

 pilae were completed and became tuberculated before the channels appeared. 2 

 This same confusion with regard to the time of the appearance of characteristics 

 with relation to each other is a peculiarity of highly accelerated species, as 

 before noted in Ver. ophioides. 



Such observations are of importance, since they enable us to understand that 

 characteristics do not necessarily develop with invariable regularity. The usual 

 order of their succession may be in a measure changed, or even reversed, when 

 acceleration takes effect upon one character more than another. So far we have 

 found this occurred only in species where all the principal characters of the series 

 were undergoing exceptional acceleration. 



The common form of the younger stages of all the Ammonitinae during the 

 goniatitinula stage is shown in the plates. 3 The depressed goniatitic helmet shape 

 was succeeded in Psiloceras by a laterally flattened helmet shape. In Caloceras 

 the same form was succeeded for a very prolonged period, in species with flattened 

 abdomens, like Cat. carusensef by a stage in which the abdomen became broader 

 and the sides slightly divergent. In Am. ceras, as in Cat. carnsense, 5 this often 

 occurred at a much earlier period, replacing entirely the psiloceran helmet 



1 By accident these specimens were all of different sizes. Thus they give false impressions. Although 

 fig. 15 is older and larger than the others here figured, my observations were made on specimens of similar 

 size and age. 



2 PI. iii. fig. 9, 20. 



3 PI. i. fig. 4 a, for Psil. planorbe ; pl. vi. fig. 7, for Cor. Sauzeanum : also in Embryology of Cephalopods. 



4 PI. ii. fig. 1, 2. 6 PI. i. fig. 15. 



10 



