24 HYATT ON THE TERTIARY SPECIES 



growing through a series of stages in the young, which repeat more or less fully the adult 

 characteristics inherited from their ancestors, and by producing series of new forms, more 

 or less parallel with those of other congeneric series, which are also new, or in j)art the 

 direct result of inheritance from the parental type form. 



Thus, although it may be said that the environment, which here consisted of an unoc- 

 cupied field, or one which may at least be inferred to have been uniiivorable for the 

 growth of other competing shells, was favorable to the fullest expansion of the tyj^e, 

 nevertheless the precision with which the series were evolved, and their concordances 

 cannot be explained by any hypothesis based solely upon these influences. Again, if this 

 be doubted, and the uniform action of the uniform environment be assumed as possible, 

 then how account for the manifest differences arising in the primary PI. levis, and 

 becoming hereditary in the series. These characteristics are not representative, but 

 essentially diflferential, in fact the only ones which in the lower forms indicate affinity, 

 and enable us to begin to thread our way through the complicated labryinth of represen- 

 tative characteristics. Thus, PI. 't"S"% PI- "'iZT, PI- '"'"S"', and PI. «"»/^«!;™«*, are all 

 more or less distinct and possess certain differences which are subsequently increased in 

 each of the progressive and in some, but not all of the retrogressive series. 



Thus, PI. Pf^^^' has the sub-acute outer side of the whorl less perceptible than in PI. 

 parvus, but it is a distinction constant in the series. Reversions to PI. levis might 

 of coiirse, and prol)ably would, occur, if we could know the facts by the actual 

 experiment of hatching broods, but these would only add strength to the ^jresent prop- 

 osition, since they would only be adults retaining the form of the young, which until a 

 late stage, even in PI. crescens, is identical with PI. levis. PI. "''"tlvi""'^ lias the narrow 

 umbilicus and gibbous whorls in the lower side of some varieties of PI. levis, and we 

 have only to consult the plates to see how persistently this is maintained in PI. supremus. 

 PI. •s''"'j^'j™™«'« has the foi-m of whorl of the more angular and flattened varieties of PI. 

 levis, with a tendency to increase this to an excessive degree in PI. discoideus. 



But here a curious interference is observable. An apparently retrogressive or revers- 

 ionary characteristic, the gibbous or rounded and smooth character of the under side of 

 the whorls described in many of the PI. discoideus, a^apears, and is increased in the succes- 

 sive forms causing so close a resemblance to PI. supremus, that a derivation of that form 

 from these was actually traced out by Hilgendorf. I have shown, however, that this is 

 probably erroneous, and that they are in distinct genetic series. 



All of the changes of form in the First Series as previously pointed out, are repre- 

 sentative, so that the only characteristics which can be compared to the class now 

 under discussion are the prominent costa?, and the small size of the whorls when considered 

 as one series. 



Having now this peculiar category of characteristics before us, it is easy to see that 

 they represent the essential differences of the series, and are quite distinct in their nature. 

 They are permanent when once introduced, and essentially progressive, or rather cumu- 

 lative, in each series, except where they are modified by the introduction of some pre- 

 potent representative or reversionary characteristic, as in PI. (rochiformis.^ 



1 Though even here they are present in the early stages of all forms with one e.xcepliou (the variety with rotundatus 

 like young). 



