THE SUDDEN ORIGIN OF NEW TYPES 413 



ceeded the apertures became curved and were often produced 

 in a spine, whilst in the Silurian period they attained even 

 greater complexity. Moreover, the specialisation of the latter 

 forms with regard to the pelagic habit is exhibited not merely 

 by the Upper Silurian Diplograptus, with its pneumatophores, 

 but by Retiolites with its netted periderm. The latter structure 

 is an adaptive device in the direction of lightness combined 

 with strength, which is met with again, for instance, in the 

 pelagically specialised swimming Crinoid Saccocoma of the 

 Jurassic. 



The continuous evolution of one generic type from another 

 is particularly well shown by graptolites and is therefore of 

 great value in its bearing upon my theory. It is clear alsQ 

 that the expression of a Graptolite genus differs widely and 

 essentially from the ordinary conception of the term. Thus, 

 since the same type of hydrotheca is found in the three suc- 

 cessive forms Bjyogmptiis reflextis, Tetragraptus denticnlatits 

 and Didymograptiisfasciculatus, it is considered that they stand 

 to each other in direct genetic succession and exemplify a 

 progressive specialisation from the original much-branched 

 form. A similar conclusion is reached in the case of the 

 successive series B)'yograptus Callavci, Tetragraptus Hicksi and 

 Didymograptus affiiiis, which all agree in possessing in common 

 another type of hydrotheca. Furthermore, since the species 

 of the genus Monograptus, as at present constituted, exhibit 

 a great variety of hydrothecae, it is evident that this genus is 

 polyphyletic and is merely a convenient conception for sig- 

 nifying that divergent branches in the genealogical tree of 

 the graptolites reached a similar stage of evolution at the 

 same period. 



Among Echinoderms the general principle discussed in this 

 paper can also be detected, although the origin of the class 

 as a whole must have taken place in pre-Cambrian times and 

 is consequently beyond our ken. Stalked Cystoidea belonging 

 to the primitive bilateral Carpoidea (showing an absence of 

 the characteristic pores of Cystids) are already known from 

 Cambrian rocks, ^.^. Trochocystitcs boJieniicus Barr. (Mid-Cambrian 

 of Bohemia) ; hence the sessile Aristocystis of the Ordovician 

 of Bohemia, which is usually regarded as a very primitive form, 

 can only dimly indicate to us the ancestral forms (the Ampho- 



