are apparent, but there is even greater difficulty in applying New Zealand Tertiary stage names, regional formation names, or local formation names to the rock groups. Tertiary stage names can only be applied to the marine part of the section, and any future changes in these stages would make the correlation ambiguous and cause confusion. An adequate number of suitable unique names for local formations are difficult to find, and unless applied to the most complete section of the region, they encumber the literature without being of much permanent value. Probably the most complete section in Westland is in the Greymouth district, but it has not been sufficiently closely divided or clearly enough defined for regional application. The numerical system employed herein for the Koiterangi area is to be regarded only as provisional, pending the establishment of a standard Westland section.

In Bell's grouping, the conglomerates (1) correspond with our K.2. and K.3.; the coal measures (2) with K.5., K.6. and K.7.; the limestone (3) with K.8. and K.9.; and the basalt with K.4. Neither Bell nor Morgan described the mudstone and overlying conglomerate K.10. and K.11.

K.1. Greywacke and Argillite.

These have been fully described by Morgan (1908, p. 96) as the greywacke and argillite with quartz veins composing the Greenland Series. They are much older than the overlying beds, from which they are separated by a major unconformity.

K.2. Basal Breccia.

This, the most extensive formation, is exposed continuously along the north, west, and south flanks of the hill, but does not show on the east side, where younger beds dip below surface gravels. Along the north side, its lower contact is generally obscured by recent gravel, but near the western end of the hill a contact or near contact was mapped by Bell but not described. On the south side Morgan mapped a fault contact between these beds and the greywacke undermass (p. 193). The authors examined a section through the formation in a stream flowing north from the saddle which exposes bare rock for a large part of its length. At the head of the stream, a slip exposes over 100 ft. of the upper part of K.2., all of K.3., and the base of the overlying basalt K.4.

The breccia impresses by its lithologic uniformity, which is such that a description of one part applies equally to the entire mass. It consists of well-indurated red breccia or conglomerate almost entirely composed of sub-angular to sub-rounded fragments of greywacke, argillite and quartz, indistinguishable from the corresponding rocks of the undermass K.1. The only other variety seen was a single 9-inch piece of spotted schist. The bulk of the material is composed of fragments ranging in size from 3 in. to 1 ft., with a small proportion of boulders as much as 10 ft. in diameter, all embedded in a matrix of finer fragments and sand. No bands of silt or mudstone were seen. Bedding is generally obscure, but rare bands of gritty sandstone show the general attitude. Some of the breccia differs from the main mass only in having a grey or blue colour.

Our correlations, some of which have been proposed by other authors, are given below. The references are to the most complete description of each correlated formation.

Lower part of Pakawau Series, Collingwood. (Ongley and Macpherson, 1923.)

Hawkes Crag Breccia, Buller, Inangahua and Murchison districts. (Morgan and Bartrum, 1915; Henderson, 1917; Fyfe, 1928.)

Basal part of Paparoa beds, Greymouth. (Morgan. 1911.)

Basal conglomerate, South Westland.* ^*

Horse Range conglomerate, North Otago.

Basal part of Kaitangata Series. Otago. (Ongley. 1939.)

Basal part of Kyeburn Series, Central Otago. (Williamson, 1939.)

The basal bed at Koiterangi K.2. and the above correlatives have most of the following features in common:—

• (a) The mass of the material is unsorted and unbedded.

• (b) Angular blocks over 6 ft. in diameter are not uncommon, and are not confined to the base of the formation.

• (c) The average degree of rounding is such as could be produced by present-day river transport of less than one mile.

• (d) The fragments were Little weathered before or during deposition, and are now often coloured red by ferric iron.

• (e) The bulk of the material is composed of 3 inch to 12 inch fragments, and fine sand and clay form a very small proportion of the whole.

• (f) Rocks known to be older, and now outcropping only a few miles away may be either poorly represented, or absent, and the beds are almost entirely composed of material from the closely adjoining undermass.

• (g) The beds are assumed to rest directly upon the undermass, but contacts have rarely been observed.

Concerning the origin of K.2. and the similar formations elsewhere in New Zealand, geologists have previously advanced theories involving glacial action, for example, Park (1909, p. 43) and several earlier writers for the Henley beds, which are now grouped with the Kaitangata Series. Although Morgan (1908, p. 104) advocated a glacial origin for the Koiterangi beds, he later advanced strong reasons for abandoning the glacial theory in the case of the similar Hawke's Crag Breccia (Morgan and Bartrum, 1915, p. 77), which he subsequently correlated tentatively with the Koiterangi breccia, prefering to regard them as pluvial deposits, while Henderson (1917, p. 83) adopted Morgan's views of pluvial origin when dealing with the Hawke's Crag Breccia as developed in the Inangahua district. It is the writers' view that all the arguments for pluvial origin advanced by Morgan and Bartrum are equally applicable to the K.2. beds. These arguments can be summarized:—

• (1) Relative scarcity of material of clay and fine sand grade.

• (2) Uniformity of rock types, representing very locally derived material.

• (3) Absence of sorted and rounded detritus, such as is always associated with piedmont glacier deposits.

The deposit is a fanglomerate, laid down close to its place of origin by heavily overloaded streams. The conditions most nearly equivalent in New Zealand to-day are those governing the accumulation of steep coalescing fans of angular unsorted rock waste in the Canterbury Alps. A necessary requisite would appear to be little or no vegetation to retard denundation, and this in turn would probably require low or moderate rainfall. The preservation

of a thick deposit of this type may perhaps be ensured by deposition of sediments derived from rising fault-blocks in adjacent sinking fault angle depressions. The fanglomerates would fail to be preserved on the positive blocks which supplied the waste, so that continuous sheets of breccia can not now occur. Supply of coarse detritus must have been uninterrupted and the land always above water level, for there are no bands of fine sand or mud and no evidence of water sorting. Neither the K.2. breccia at Koiterangi nor the other breccias with which it is correlated contain internal evidence of age, and although the coal measures which separate the breccias from the overlying marine beds contain leaf impressions, these have not been recently described and are of no assistance. The fossils from the overlying marine beds are of different ages, and if our correlation of the breccias is correct, all the breccias must be older than the oldest marine beds. The value of the correlations depends upon the improbability that the extreme relief needed to produce the breccias would be either repeatedly produced or would continue much later in some parts of the South Island than in other parts.

The earliest marine bed above the basal conglomerates of the West Coast is the Bortonian (Mid-Eocene) Island Sandstone (Finlay and Marwick, 1940, p. 106) which at Greymouth is separated from the top of the basal breccia by about two thousand feet of coal measures. But in Otago, the Horse Range and Kaitangata breccias are both overlain by beds of Cretaceous age. The Brighton limestone, considered to lie above the Kaitangata Series and the Herbert beds above the Horse Range breccia, both contain belemnites of Piripauan age (upper Cretaceous) (Finlay and Marwick, 1940, p. 103). In both cases there are thick intervening coal measures, and the basal conglomerate could not have been deposited immediately prior to the marine beds. The maximum age is less clearly defined, for the undermass upon which the basal breccia rests is granite, greywacke or schist of uncertain but very probably pre-Triassic age.

K.3. Conglomerate.

The basal breccia is overlain both on the north and south sides of the hill by a band of conglomerate composed mostly of 3 in. to 5 in., rounded to sub-rounded, greywacke fragments and a few, scattered, red-stained, rounded, quartz boulders. The contact between K.2. and K.3. is not conspicuous and was not examined in detail. There may be a transition, as has been described for beds in the same relative stratigraphic position in other areas, but a comparison of the thickness of the Koiterangi section with that in the Paparoa Range shows that either deposition was slower at Koiterangi, or beds which were deposited at the Paparoa Range are not represented at Koiterangi, and that an unconformity may be represented at this contact.

K.3. conglomerate is more rounded, and more closely resembles present day river gravel than does K.2., suggesting that the extreme topographic irregularity of basal breccia times must have been lessened during or before the deposition of K.3.

K.4. Basalt.

The contact between the basalt and the upper part of the K.3. conglomerate is fairly well shown in the top of the slip below the saddle, but both the basalt and the conglomerate are weathered, and much of the basalt has disintegrated into blocks and may have slumped a few feet. The lower contact of the basalt is in place, and several square feet have been exposed. The contact is not a smooth surface, for the basalt has irregularly penetrated a few inches into the conglomerate. A thin, soft band of red clay, already described by Bell (1906, p. 84), marks the actual contact. The conglomerate did not appear to be thermally altered and the basalt is not vesicular. About 10 ft. of basalt is exposed above this contact. The lowest observed part of the overlying conglomerate K.5. is about 20 ft. higher, and no more than 30 ft. of basalt can be present at this place. From the saddle the basalt extends continuously to the basalt-capped plateau ten chains west, where it is over 100 ft. thick. On the south side of the hill, south-east of the saddle, a sharp contact between K.3. conglomerate and overlying dark quartz grits K.6. is clearly exposed in the east bank of a small stream which flows south from the trigonometrical station to the Hokitika River. Both the basalt and the overlying K.5. conglomerate are absent.

The contact at the upper surface of the basalt was not seen, and it is uncertain from the field relations whether the basalt is in the form of a sill or a flow. If it is a sill it will be younger than the overlying K.5. conglomerate and perhaps much younger, but if it is a flow, as was Bell's opinion, it will be younger than the K.3. conglomerate and older than the K.5. conglomerate. It could then be contemporaneous with basalt of the Paparoa Range (Morgan. 1909, p. 81; Hutton, 1943, p. 64).

Recent field work at Greymouth, as yet unpublished, shows the basalt to be a flow and to be younger than the basal Paparoa breccia, but older than the bulk of the coal measures. A petrological description of the Koiterangi basalt was kindly supplied by Dr. C. O. Hutton and is given in Appendix I.

K.5. Conglomerate.

A few chains west of the previously mentioned slip in which the basalt is exposed, K.5. conglomerate overlying the basalt forms a low cliff which extends for a few chains just below the top of the ridge. It is brown-grey in colour, and composed of rounded fragments of greywacke and quartz with an average diameter of 2 in., dispersed through a sandy matrix. The upper surface of this conglomerate was not seen, but quartz sand and well-rounded pebbles of quartz scattered along the top of the ridge suggest that the next overlying bed is chiefly composed of quartz. The conglomerate is about 6 ft. thick.

K.7. Coarse Sandstone and Coal Measures.

A hard band of coarse sandstone, apparently overlying K.5. conglomerate, forms a steep 50 ft. cliff about half way along the ridge from the saddle to the trigonometrical station, extending with increasing height around the north side of the hill for several chains. It is the basal member of the coal measures as exposed

in the headwaters of a small stream which rises at the foot of a limestone cliff, two chains north of the trigonometrical station. This stream flows north over the limestone cliff to join the stream, in which occurs the basal breccia section already described. The following coal measures are exposed in the stream bed:—

K.7. Conglomerate.

2ft. Moderately haid, well bedded quartz sandstone, sharply defined upper limit.

1ft. Coal.

30ft. Moderately soft, dark-blue, micaceous quartz sandstone, a few vegetable impressions.

10ft. Soft, white, quartz sandstone.

40ft. Micaceous, carbonaceous sandstone and shale, 4 thin coal seams(1ft., 6in., 9in., and 1ft.).

200ft. Massive, hard, white or grey, coarse, quartz sandstone.

The coal is mostly rather dull, with poorly developed conchoidal fracture, it resembles that of the James seam of the Greymouth Coalfield, although one of the thin seams was more finely cleated. The uppermost seam appeared soft and dirty, but this may be merely an outcrop condition. Both Bell and Morgan describe 6-8 ft. outcrops of impure coal, but these were not seen. Analyses of the coal are given by Bell (1906, p. 79).

K.7. Conglomerate.

The stratigraphic position of the K.7. conglomerate is given in the above section. It is composed of well-compacted sub-angular to sub-rounded quartz and greywacke up to 3 in. in diameter alternating with bands of quartz sandstone as much as 1 ft. thick. Sixty feet of this conglomerate is exposed.

No more than 20 ft. of beds can separate the highest part of the conglomerate (K.7.) from the base of the limestone at the foot of the cliff just north of the trigonometrical station. Probably the same conglomerate was seen in the headwaters of several small streams on the south-west slopes, at one place overlain by about 10 ft. of soft fine quartz sandstone. If this conglomerate is correctly correlated with K.7., then the top of the overlying sand at this place cannot be separated from the base of the limestone by more than 10 ft.

K.8. Marine Carbonaceous Sandstone.

On the eastern slope approximately 40 chains east of the trigonometrical station and not far below the eastern limit of limestone, an outcrop of carbonaceous-looking sandy mudstone and sandstone yielded a poor microfauna suggested by Dr. H. J. Finlay to be of Kaiatan or Whaingaroan age (F. 6275). It marks the onset of marine conditions. The occurrence is isolated; and the relationships obscure. The material was not observed on the south side, where quartz sandstone appears to separate K.7. conglomerate from limestone. Although not observed, it could have been present in the 20-ft. gap in the section at the foot of the limestone cliff below the trigonometrical station.

The age limits given, although vague, indicate a possible Brunner age for the underlying coal measures.

K.9. Limestone.

The whole thickness of the limestone was not seen at any one place, and its thickness is not accurately known. About 50 ft. is exposed to the north of the trigonometrical station and the total may well exceed 100 ft. The quality is variable, ranging from fairly pure to sandy near the base. At one locality it was seen to contain a moderate amount of fine glauconite. Except within a few feet of the upper contact, it is uniformly hard, contains no soft bands, and only rare, poorly preserved macro-fossils.

The upper part of K.9., together with K.10. and K.11., is well exposed in the headwaters of a small stream about a quarter of a mile south of the trigonometrical station, the beds being little disturbed and forming an almost continuous outcrop. They are separated by a fault from low-dipping sands and conglomerates which are several hundred feet lower in the sequence, but which outcrop at about the same level in the headwaters of two small streams five and ten chains to the west.

At this place only the upper 20 ft. of the limestone is exposed, the underlying beds being obscured by a litter of limestone blocks. The lower part of the exposed limestone is sandy, hard, and almost devoid of fossils, and is thus similar to the bulk of the Koiterangi limestone. The upper part, however, contains a 6-ft. band with numerous brachiopods together with rare, worn, phosphatised, internal casts of gasteropods. These fossils have been examined by Dr. Marwick. who writes as follows:—

“No interior features of the brachiopods could be determined, but the exteriors show them to be probably:—

Stethothyris aff. tapirina (Hutt.)
Liothyrella aff. neglecta (Hutt.)

“The original S. tapirina was described from ‘Cobden,’ presumably from the Cobden limestone, but from what horizon in it is unknown. The two Koiterangi specimens (Geol. Surv. loc. 3153) have more prominent beaks than the Cobden ones, so that identification as tapirina is not justified. However, they are near enough to tapirina to suggest that this upper part of the Koiterangi limestone is equivalent to some part of the Cobden limestone, probably also the upper part. According to J. A. Thomson (1926, N.Z.J.Sc.Tech., vol. 8, p. 153) S. tapirina is a fairly consistent member of the ‘landonensis fauna’ (Duntroonian), ranging throughout Ototataran and down into the Waiarekan (Whaingaroan and Kaiatan), but not extending up into the Hutchinsonian. The association of the two forms, S. tapirina and a small Liothyrella, suggests that we are dealing with representatives of the ‘landonensis fauna’ and that the bed is Duntroonian.”

K.10. Mudstone.

Immediately above the fossil band, the limestone is softer, and 3 ft. higher there is an abrupt change to glauconitic phosphatic mudstone, which at the base contains a few scattered quartz granules. The limestone, for 1 ft. below the upper contact is also phosphatised and penetrated by glauconite-filled borings.