The economic geologist and stratigraphical palaeontologist both demand that their units shall be accurately and clearly defined. Recent work in the British Isles has demonstrated very clearly that stratigraphy can be a very exact science.^*

When the palaeontologist is able to study fossils which have been collected from well-defined units he is able, more often than not, to provide the stratigrapher with very accurate information as to age or correlation of the strata in question. The stratigrapher gains from his exact work. This is nowhere more apparent than in recent developments in oil-geology, and has a great economic significance.

It seems to the writer that the stage-names in New Zealand are imperfectly defined, with the result that, in spite of the palaeontological data now available, exact correlation is practically impossible in many cases.

Too much attention has been paid to individual fossils, their description, and the determination of their range in time. The tendency of modern palaeontology is towards (a) the study of restricted groups of organisms as they develop in time; and (b) the study of faunal communities or assemblages rather than of individuals.

Neither of these methods has yet been applied in the Dominion. The first means more than merely recording the succession of different forms in time. The object is to find the stratigraphical value of stages in development in particular lineages. The guiding principle is that stages of development in any one lineage are successive in time. This science of stratigraphical palaeontology demands of its devotees a greater knowledge of fossils in the field than do the older methods. The focus of modern work is in the field rather than in the museum. The significant changes in many lineages are so intimately related to changes in lithology or facies that they can only be appreciated fully in the field.

In a report on the Lower Devonian fossils of Reefton, soon to be published, the writer has attempted to show the value for purposes of correlation of faunal communities based upon characteristic fossils. It is not necessary, therefore, to provide full details of this type of study here. It will be sufficient to state that a given horizon of uniform facies can be treated in terms of its characteristic fossils alone. For practical purposes a characteristic fossil of a given horizon and facies is one which is there abundant. It is not necessary to know the complete fauna, nor is it vital to determine the exact range of any individual species.

Fossil communities, in the same way as recent animal and plant communities, are bounded by certain physical conditions, and are constant only where the physical conditions are constant. A change in environment results in a change in the community as far as the characteristic members of it are concerned.

It is thus of great importance to realise that a community characteristic of a certain stage in the type locality will be characteristic of that stage only in areas where it is represented by strata of the same facies.

I would submit, therefore, that before correlation of faunas of distinct facies is possible one must understand clearly the characteristic fauna in its relation to the facies of the type area.

In very few cases is it possible in the Dominion to enumerate standard characteristic communities. This is because in most cases the type localities for our stages are not fixed, or because the fossils from type localities are imperfectly known. Until these deficiencies are made good it is impossible, in the writer's opinion, to expect to get any accurate data as a basis for sound correlation.

Before either of these methods, viz., the study of restricted groups of organisms as they develop in time, and the study of faunal communities based upon characteristic fossils, can be applied in New Zealand there must be a more rigorous definition of stratigraphical units. This is attempted in the sequel.

For each stage I propose to select a definite type locality which will form a standard for that stage. Each stage will be based upon definite lithological units which will be defined as accurately as possible. When they are known the most characteristic fossils will be listed. Stress will be laid upon the community rather than upon the individual.

In selecting type or standard localities the clearness of the section and the state of preservation of the fossils will receive consideration, but underlying the selection the law of the superposition of strata will be the guiding factor as far as it is applicable in any one area.

I should like to suggest that if the stages herein employed prove to contain more than one community in sequence in the standard area, then each might fittingly be spoken of as a zone. Each zone would be indicated by selecting an index-fossil, thus the “Zone of Stethothyris sufflata”; or, more simply, the sufflata-zone.

It is therefore possible, in some cases certain, that the stages will be subdivided into zones. This method is more convenient than the excessive introduction of stage or substage names. For gaps between the stages as herein defined, however, new stages will be both necessary and desirable. Such a scheme should make for permanence of nomenclature—a most excellent aim.

It appears to the writer that the first task of Dominion students is not to attempt correlation problems, but to study in detail the faunas of the standard areas. Only after these are thoroughly known in relation to their particular facies can correlation be attempted. Areas other than standards could well await study until knowledge of the standards is reasonably complete.

A second task concerns the study of changes in faunal assemblages in relation to changes in facies. This can only be attempted by palaeontological studies in the field after the standard faunas and facies are determined.