phyla—Mollusca, Brachiopoda, and Foraminifera. Certain conconclusions have inevitably been drawn as to their respective usefulness in correlation, ease of discrimination of the genera and species, reliable identification of these in our Tertiary deposits from both perfect specimens and fragments, and availability in the maximum number of localitics and lithologies. This is not the place to discuss these conclusions, but it has also been observed that there is no discrepancy in their evidence, if due regard is paid to biological principles and imperfections in the fossil record—the former affecting them equally, the latter unequally. In every case where discord seemed present at first, consistency resulted from fuller investigation.

Recurrence of Faunas.

Another opinion that has gained some mention in literature is that identical faunas might appear in different horizons if lithologies were similar, and thus upset correlations. This would leave the way open for much fanciful speculation about surface geology, but statements such as “determinations of micro-faunas in Arkansas afford other examples of recurrence of faunas” (Stephenson, Bull. Am. Assoc. Pet. Geol., vol. 11, p. 12, 1927) are obsolete in micropalaeontology. A much more modern conception and treatment of this problem of unvarying sediments has been finely presented by Hedberg (1937) in dealing with the massive Carapita formation of Venezuela. There is no example in New Zealand remotely suggestive of recurrence of identical faunas, though similar backgrounds of hardy long-ranging species are of course common. There is always a difference if the age is different; if there is no difference in faunas of reasonable size their age is the same.

Post-Jurassic Complexity.

The problem of dividing up our Cretaceous and Tertiary beds into well defined stages is much more difficult than the subdividing of the Triassic and Jurassic. In these the faunules are small, uniform throughout the country, and they occur in easily defined, strictly limited succession. For instance, when one finds the widely spread Carnic Maoria problematica (Zitt.), one can be reasonably certain of finding the Noric Monotis richmondiana (Zitt.) a few hundred feet higher in the section, each species practically comprising the whole faunule. Similarly, no trouble has been found in accepting the range of Pseudaucella marshalli (Trech.) as about the middle of the Liassic; it has never been found in any embarrassing company, and Inoceramus haasti Hoch. is not encountered until several hundred feet higher in the beds. Obviously we are dealing with entirely different conditions from those of the Tertiary, where the trouble inclines to be a plethora of riches. The time and labour involved in winnowing the grain has naturally led some workers to seek a simpler way out, and the suggestion has been made, apparently by analogy with the old faunas, that only the common species need to be considered. Substitute “characteristic” for “common” and we agree; synonymise the two and we disagree.

Probably owing to unfavourable stations, our Cretaceous molluscan faunas are few in number; but, as far as can be seen, they tend to show the diversity characteristic of the Tertiary faunas.

Whether these observed differences in the straightforwardness of the fossil record are due to general climatic change is too large a question for discussion here, but it seems significant that they apparently synchronise with the world revolution in the vegetation so markedly initiated between Lower and Upper Cretaceous. Whatever the cause, it soon becomes evident to the practical palaeontologist that stimulating and successful methods of dealing with ancient faunas and strata have to be modified or even discarded in a study of the Tertiary.

Percentage of Recent Species.

The writers have not given the percentage of Recent species associated with each stage. It would involve, for present or future correlation, the naming of total faunas, of which a large proportion will consist of non-significant species. The percentage method of estimating the age of Tertiary faunas is of little or no value in defining or correlating stage divisions (see, for example, Laws's difficulties with regard to Kaawa Creek; 1936, pp. 39–42). Of course, the nearer a given fauna is in age to the Recent, the more it is likely to have in common, so that a general indication as to early, middle, or late Tertiary age can be obtained. It is significant, considering the very extensive use now made of micro-faunas in correlation by that severest taskmaster of stratigraphy—the Oil Company—that the writers can recall only a single mention of “percentage of Recent species” in connection with Foraminifera—that of Vaughan as far back as 1923.

Specific divisions are so unequal in value that mathematical statements based on them assume an analogy with unit numbers that does not really exist. The very basis, therefore, of the method is unsound. If the results agree with the other evidence, they are generally regarded as confirming it, but if they disagree, they are always readily explained away as due to differences in rate of evolution, contact with or isolation from other faunas, and so on. In such a realm as palaeontology, the writers are inclined to look askance at too much application of statistical methods; the “unknowns” and “controls” seem to them too debatable and too likely to vitiate the outcome. In practice the principle of key species and delimitation of ranges appears to give superior results, and has been used throughout this paper. The method has for some time been used in guiding the palaeontological work of two major oil companies, and the results obtained over wide areas have been entirely satisfactory.

Type and Standard Sections.

Allan (1933) has drawn attention to the need of a better stratigraphical basis for the New Zealand stages, and has more accurately defined the limits of several of them. There still remains, however, much to be done, especially more detailed field work at the type localities.

Palaeontologists, schooled as they are in the value of type selection in zoological procedure, may be biassed in its favour, but is seems likely that, in the end, uniformity and precision of usage can be achieved only by this means.

Considerable divergence of opinion prevails among stratigraphers, not only as to the need of type localities for stage (and series) names, but also as to the manner in which these types should be designated. The terms type locality, type section, unit type, standard locality, standard section, control section, and reference section are all in use, but uniformity in their application is lacking. The general practice, no doubt, is to regard the “type” as something fixed and not subject to change, but that this is not always so is shown by A. J. Butler's “type section” of the Wenlock Limestone (1939, Q.J.G.S., vol. 95, p. 41), which “is used for convenience to denote the complete reference-section …. and, of course, bears no significance of priority over sections described by earlier authors.”

There has also been much criticism as to the poorness of some type sections, and refusal to recognise any merit in their use. Objection to the use of types probably has arisen through confusion of two separate processes—(1) stabilising the stage or series name by fixing it to a physical entity; (2) selection of the best all round section (at least at the time of selection) as a basis of comparison with other localities in order to determine their age. The writers propose to keep the two underlying ideas quite distinct by applying the terms Type Locality and Type Section to the former, and the terms Standard Locality and Standard Section to the latter. The term reference section can be regarded as practically synonymous with the Standard, but J. E. Eaton's “control section” (1931, Bull. Amer. Assoc. Pet. Geol., vol. 15, No. 4, p. 372) represents rather a number of type sections or of standard sections combined into one standard column.

The terms type section and type locality, then, should be rigidly applied, permanent, and reasonably subject to the law of priority.

The terms standard section and standard locality should be applied to the best section that, with the available data, can be confidently correlated with the type, where the two are necessary. If additional evidence upsets this correlation, then a new stage name or a new standard section must be selected.

This use of the standard, as distinct from the type, section will secure the results aimed at by Allan regarding the Bortonian and the Awamoan, while avoiding complications with regard to the validity of type designations and the accuracy of stage definition.

The writers have found some difficulty in evaluating the Bortonian and Hutchinsonian, and in fitting the faunal facts into the conventional usage, so that their treatment of these stages may seem unduly involved. The foraminiferal evidence suggests that the Bortonian, as represented at Hampden and elsewhere can be subdivided, while both molluscs and forams indicate that a new stage or stages may be necessary between the Bortonian and Wangaloan. This should await the elaboration of field evidence brought to light during the

recent survey of the Hampden district by D. A. Brown. Owing largely to facies differences, the exact relationship of the Waitakian to what is here called Lower Hutchinsonian is not yet clearly understood, nor is the best section of the Lower Hutchinsonian yet determined to serve as a standard. The position has been purposely left vague until further faunules have been collected or new stratigraphic evidence brought to light.

Scope of Lists.

The following lists are concerned mostly with restricted species and first and last appearances of important fossils. In the case of the molluscs, records of shallow-water and rock forms and the very small shells, such as Rissoids and Liotids, have mostly been omitted as too sporadic and uncertain. Except in the case of some outstanding genera, only described species are dealt with, but many index new species are known.

The Foraminifera so far validly known from the Dominion have been fairly fully dealt with here, and the ranges include the latest information on all the species discussed so far by Finlay in the “Key Species in Stratigraphy” papers, most of the species described by Stache and Karrer in the “Novara” volume, all Kreuzberg's new species from the Ihungia of Whakatu Stream, most of Parr's new forms in various recent papers and of Chapman's in Pal. Bull. 11, and a number described by Cushman and collaborators. Records of mollusca are not nearly so complete as regards the very many fossil species now described by Zittel, Hutton, Suter, Wilckens, Trechman, Woods, Marshall, Murdoch, Marwick, Finlay, Powell, Allan, Laws, King, and others; but are reasonably complete in the case of genera of any significance. Records of the brachiopods have been taken from the published papers of Thomson and Allan and from personal knowledge.

It would not, in any case, have been practicable to present as full a specific account of the Mollusca and Brachiopoda as has been done with the Foraminifera. There are so very many cases of “species groups”—closely allied forms which have been named as distinct in the literature, coming from different localities of about the same age and certainly in the same Stage. Even within the same district, the diversity of species of such genera as Turritella, Pelicaria, Verconella, Parvimitra, Volutid genera, Pachymagas, Rhizothyris, Liothyrella, etc., sufficiently indicate the multiplicity of local faunules, which have not much real bearing on Stage correlation. Marwick and Thomson have frequently referred to this. It is then the group of species rather than the single name which becomes important, and it is not possible to delimit all these except in a check-list. Subdivision of the foram species may not have been so minutely attempted (though the personal equation should be absent here), but it is likely that their abundance has allowed of more accurate estimation of variability, with the consequent suppression of over splitting. Perhaps because of this, perhaps because their record is much fuller, perhaps