XX INTRODUCTION. 



in every region the series of strata contains only a very 

 discontinuous record of its successive faunas and floras. When 

 a region is a land area, as a rule, no deposit capable of preserva- 

 tion for long periods can accumulate, and the characters of its 

 life can only be inferred from fossils which have been entombed 

 in sediments apparently of the period in question elsewhere. 

 When the region happens to be covered with comparatively 

 deep water, the sediments will contain scarcely any but aquatic 

 organisms, rarely yielding a trace of the life on the nearest 

 land. 



Definite Direction of Evolution. Under such circumstances 

 it can be readily understood, that the time has not yet arrived 

 for deducing general laws from the data of Palaeontology. In 

 fact, our k'nowledge of the evolution even of the Vertebrata 

 is most casual and fragmentary in character. Nevertheless, 

 sufficient is known to indicate that changes in the vertebrate 

 skeleton have taken place in a certain definite and irreversible 

 order ; and the relative age of two skeletons of the same type 

 of animal of widely different periods can readily be determined 

 at a glance by an expert. For example, among the early 

 Palaeozoic fishes there are many heavily armoured forms with 

 very little ossification in the endoskeleton and only incipient 

 vertebral centra. The endoskeleton does not begin to ossify to 

 any noteworthy extent until the exoskeleton atrophies ; event- 

 ually at a later period the bony endoskeleton is the all-im- 

 portant part of the framework. Now, it so happens that in 

 certain Mesozoic Pycnodont fishes (Mesturns) and the Tertiary 

 coffer-fishes (Ostraciontidse), the rigid dermal armour is again 

 acquired; but it is apparently contrary to law for the endo- 

 skeleton to revert to its primitive state as observed in the 

 Palaeozoic fishes just mentioned the parts of the axial skeleton 

 simply become rigidly fixed and are nearly as well ossified as in 

 the unarmoured contemporaries and relatives. When the 

 vertebral centra have once become fully formed, indeed, they 

 never degenerate to allow the unconstricted notochord to 

 persist again. To take another example, the lobate fins with 

 endoskeletal supports in the earliest fishes always appear to 

 tend towards atrophy, while the dermal rays surrounding them 



