518 W. D. MATTHEW CORRELATION BY FOSSIL VERTEBRATES 



than external form, for it is more indicative of affinity and less of adapta- 

 tion. Complex structure affords a better guide than simple, for it is 

 more subject to change and the alteration more readily recognized. The 

 more complete, the more complex, the more the internal structure is rep- 

 resented, the safer guide is the fossil as to the affinities and position of 

 the animal whose existence it records. 



Marine invertebrate faunas have generally been regarded as the stand- 

 ard basis of geologic correlation. For this there are excellent reasons. 

 The marine succession is far more widespread and uniform than most 

 terrestrial formations. Its invertebrate fossils are abundant, character- 

 istic, wide-ranging ; the evolution, migration, and succession of the faunas 

 have been thoroughly studied, rest on a vast mass of evidence, and in most 

 cases must be regarded as finally settled, at least in their broader outlines. 



Fossil vertebrates, on the other hand, are comparatively rare, usually 

 fragmentary, and owing to the scarcity of the evidence and the complex 

 problems presented by their evolution and migration they may frequently 

 be misinterpreted in correlation work. 



There are nevertheless certain advantages in correlation by means of 

 fossil vertebrates to offset these disadvantages. 



In the first place, the skeleton of a vertebrate is an internal structure, 

 highly complex in its construction as compared with the hard parts of 

 most invertebrates. For this reason the affinities of a fossil vertebrate 

 can be more exactly and more securely determined than most fossil inver- 

 tebrates, provided that our specimens are sufficiently complete. 



In the second place, the geologic range of species and genera of verte- 

 brates is more restricted than that of most invertebrates and plants. 

 They enable us, therefore, to make more precise correlations. This is 

 especially true of Tertiary mammals, and in particular of certain groups 

 of mammals of progressively specialized adaptation and wide geographic 

 range. 



It is in the correlation of land and fresh-water formations that verte- 

 brates are chiefly useful. Fresh-water invertebrate faunas are compara- 

 tively limited in variety, and nearly all of them relatively simple in the 

 structure of their hard parts and apparently very slow to change. Plant 

 remains are similarly limited in their usefulness for correlation purposes, 

 so that while invertebrates and plants are in general far more abundant 

 and easy to find, the vertebrate faunas, if sufficiently complete, Avill afford 

 more exact conclusions. 



The great difficulty in using fossil vertebrates for correlation work is 

 in the identification of fragmentary material. Most of our material is 

 fragmentary, and its value varies widely according as it is more or less 



