XVIII PHYLOGENY AND ONTOGENY 5 i i 



examples of this type. In the second type the pedicle opening 

 is restricted to the ventral valve, and the direction of the pedicle 

 makes a right angle with the plane of the valves ; in the lower 

 forms the pedicle lies in a slit or sinus (^Trematidae)^ but by fur- 

 ther specialisation it becomes enclosed by shell growth so as to lie 

 within the periphery, and finally becomes subcentral in some 

 genera (^Discinidae) . The third type shows the pedicle opening 

 confined to the ventral valve and submarginal. A pseudo-delti- 

 dium may preserve the original opening (^Clitamhonites^ ; or this 

 shelly plate may become worn away or reabsorbed in the adult 

 so that the deltidial fissure through which the pedicle passes 

 remains quite open (^Orthidae). In the fourth type the incipient 

 stage marks a return to the simple conditions of the first type ; 

 but ultimatel}^ a pair of deltidial plates develop, and may com- 

 pletely limit the pedicle opening below. Examples of this type 

 are Spirifera and Rhynchonella. By means of these four types 

 the Brachiopods have been divided into four Orders : the Atre- 

 mata (type i.) ; the JSfeotremata (type ii.) ; the Protremata 

 (type iii.) ; and the Telotremata (type iv.). 



The Telotremata were the last to appear, but the four types of 

 pedicle-opening with the various forms of calcareous brachial 

 apparatus were in existence in the Bala period of the Ordovician. 



As Paterina is the most primitive form of all, we may place it 

 at the root of the phylogenetic tree. From it sprang the Atremata^ 

 which gave off the Neotremata ^nd Protremata ; the most primi- 

 tive Neotremata seem to be the Trematldae^ while the connecting 

 link between the Protremata and Atremata is furnished by the 

 Kutorgiiiidae. From the genus Oonchidium and its allies we 

 may see how the Rhynchonellidae ushered in the Telotremata as 

 an offshoot from the Protremata. The Telotremata subsequently 

 gave off two main branches, which became specialised with the 

 loop-bearing and spire-bearing forms respectively. 



The evolution and mutual i-elationships of genera have been 

 indicated with much probability by Hall, Clarke, and others. 

 The Obolelloid type may be connected with the Linguloid by 

 means of Lingulella and Iji7igidepis, while in Liiigida itself we 

 find the point of divergence for the ancestors of Trimerella, and 

 for a line of variation culminating in Digtiomia. The Palaeo- 

 zoic Rhynchonelloids branched off at an early period from the 

 same stock as Orthis, and are connecting links between this 



