tion with Fox River; horizontal sandy mudstone unconformably overlies tilted coal measures which dip north-west at 20°. The other outcrop is on the opposite side of the stream 15 chains farther upstream, where severely crushed coal measures are involved in a major fault. No well-defined seam was observed, but fragments of fault-involved coal are not uncommon (A5 in table of analyses). The significance of these two outcrops in the interpretation of the structure is discussed later and shown by the cross-section.

North of Henniker Creek the belt of coal measures widens, and detrital coal becomes common and more widespread. Bacon Creek is the name given to a small stream that enters Henniker Creek 40 chains above the Henniker Creek-Fox River junction. In this creek fragments of coal are common and are derived from two seams exposed on the eastern side of the major fault mentioned above and half a mile from the mouth of the creek. Downstream from the fault Bacon Creek flows through a narrow gorge formed by 200ft. cliffs of sandy mudstone, but upstream from the fault the mudstone is replaced by crushed coal measures and the cliffs are abruptly replaced by gentle slopes. A coal seam five chains from the fault is crushed and not well exposed, but may be as much as 10 ft. thick. A less crushed seam five chains farther upstream was sampled (A4 in table of analyses) and traced for two chains across a low spur near the head of the creek.

At Fox River, 30 chains north, the two rock types are still clearly distinguishable by their characteristic topography, the position of the fault-contact between the coal measures and the sandy mudstone being marked by the termination of the steep cliffs and the beginning of the even slopes. Detrital coal is again common east of the fault, but only two thick seams were observed, both on the south side of the river. The more westerly, 25 chains east of the fault, dips east at 20° and forms part of the following section exposed in a bluff at near river level five chain from the present river bank.

The other seam, 15 chains upstream from the one just described, has recently been exposed by a slip. The seam is vertical and is probably the 26ft. seam mentioned by Henderson (1918, p. 204). The slip shows:—

Breccia is exposed at a few places between the 22 ft. seam and the first granite outcrop 10 chains farther upstream. The breccia contains angular and sub-angular fragments of granite with a large range in size, fragments of pegmatite with inch flakes of mica being

conspicuous. It is more indurated than the breccia at the mouth of the Fox River, but shows no other significant difference.

The area between Fox River and Four Mile Stream, two miles north, was not examined, but the presence of much detrital coal in a small stream that drains the area, and enters Fox River 25 chains below the junction of Henniker Creek, makes it probable that further coal outcrops will be found in this direction.

Content and Relation to Underlying Reds.—The coal measures contain no quartzose beds and consist entirely of arkositic sediments which range in grade from shale to coarse sandstone and conglomerate. Breccia grades up into rounded granite conglomerate, and the conglomerate into coal-measure sandstone, both at Fox River and six miles south, outside the mapped area, at Bovis Creek. Both breccia and coal measures are composed of granite-derived material with an admixture of hornfelsie greywacke, and both probably were derived from the same source. Although the contact was not seen, it is almost certain that the breccia rests directly and unconformably on the underrmass.

Age and Correlation of the Coal Measures.—The Paparoa coal measures of the Greymouth district (30 miles south) are stated by Morgan (1911, p. 56) to grade up from a basal conglomerate known to be similar to that of the Fox River district. The Paparoa beds contain many seams of coal; all have a sulphur content of less than one per cent. Above the Paparoa beds, and perhaps resting on them unconformably, are the quartzose Brunner beds with medium-to-high-sulphur coal. The Brunner beds grade up into the marine Island Sandstone (Bortonian) and the Kaiata Mudstone (Kaiatan).

There is no evidence against correlating with the Paparoa beds the low-sulphur arkositic coal measures, and breccia of the Upper Fox River; and with Brunner beds the high-sulphur quartzose coal measures near the coast, which unconformably overly the basal breccia and grade up into marine beds of Kaiatan age. The above correlation is indirect, but no better attempt can be made until the coal-measure leaf impressions have been reliably described.

Appearance of Coal.—The thick high-rank 22 ft. seam has attracted the greatest attention. The coal, only slightly dirty to the touch, is bright and has a sub-metallic lustre. According to the American Society for Testing Materials proposed classification of coal by rank (1938, Supplement to book of A.S.T.M. Standards) this coal, having more than 86 and less than 92 per cent, fixed carbon on a dry ash-free basis, is placed in the semi-anthracite group of the anthracite class of coal. The high rank of this coal is of some interest, for with the exception of the Canterbury anthracites, which are lignites locally changed by igneous intrusions (Evans, 1899), no other thick seams of anthracite are known in New Zealand.

Although the coal is classified as a semi-anthracite, it shows in hand specimens a greater resemblance to the low volatile bituminous coals produced from the Paparoa Mine than it does to true anthracites, being soft and friable and intersected by numerous shear planes, and not hard like a full anthracite.

It might be thought that these numerous shear planes are solely a result of the intensity of deformation suffered by the coal measures, but this is probably not the case, for Campbell (1930, p. 694) when discussing the stages in the progressive metamorphism of coal has shown that in the United States bituminous coal becomes progressively more cleaved and friable as the volatile content is reduced and reaches maximum friability or “softness” at low-volatile bituminous or semi-anthracite rank; with increasing metamorphism the individual fragments become re-cemented to form typical hard anthracite. It would appear likely that the friability of the Fox River coal is as much due to its rank as to its extreme deformation; for it is probable that the generalizations made by Campbell—which are irrespective of special deformation—will apply also to New Zealand. Consequently any other high-rank seams found, even if less deformed, will also be friable.

Analyses and Rank of Coal.—In spite of the soft nature of the coal and its inaccessability its obvious high rank has attracted several prospectors and many ill located samples have been analysed. Several of these analyses have been assembled by Henderson (1918, p. 204). It will be shown later, however, that the rank of the coal changes so rapidly in a direction normal to the Lower Buller Fault Zone that individual analyses are of little value unless the precise locality from which the samples were obtained is known, consequently the only anaysis reproduced in the table below is the one which was collected from a known locality by Henderson himself. This analysis is of some importance, for it shows (in agreement with the other analyses quoted by Henderson) that the sulphur content is low and considerably less than one per cent. The sulphur content of coals from different New Zealand formations show large and consistent differences and the low-sulphur content of this coal has already been referred to when discussing the correlation of the coal measures.

With the exception of A2 quoted from Henderson (1918, p. 204) to give the sulphur content, all of the following analyses have been made by Mr. Doherty of the State Mines Department, Rewanui, and are published here by the courtesy of that department. The origin, of the samples is marked on the plan with the analysis numbers given in the table below.

• A1. Picked lump of hard coal from 22 ft. seam, upper Fox River.

• A2. Sample near large seam (Henderson, p. 204, No. 7).

• A3. From 6ft. seam, 25 chains east of fault (picked sample).

• A4. Seam at Bacon Creek.

• A5. Fault-involved coal, Henniker Creek.

• A6. Loose coal, Bovis Creek.

The wide variation in the value of the fuel ratios (ratio of fixed carbon to volatile matter) indicates that coals of very different ranks occur within the limited area of the upper Fox River Basin. The analyses when plotted on a map show a rapid but regular decrease

in rank from east to west, the most easterly samples from the Fox River being a semi-anthracite, and the most westerly sample from Bovis Creek being a non-coking sub-bituminous coal, the intermediate samples being coking bituminous coals. No direct reason for this unusually rapid change in rank is apparent, but many of the explanations invoked elsewhere (for example, age, permeability of enclosing rock, contact metamorphism, or original composition) can be shown to be non-effective here. Difference of age is often given as a cause, but it is highly improbable that the Bovis Creek coal measures can differ appreciably in age from those at Fox River for both grade up from the basal breccia and neither can be as young as the quartzose coal measures at the coast (Fx 4).

The well defined differences in the degree of induration of the coal measures no doubt cause corresponding differences in permeability. But these differences are the opposite of those needed to explain the differences in rank, for the high-rank low-volatile anthracite is associated with the more indurated coal measures, and the low-rank high-volatile coal with the less indurated. An early view of Campbell's (1905) was that joints significantly increase the permeability, permit the escape of volatile matter, and increase the rank of coal; but no direct relation between coal rank and jointing can be shown at Fox River. In any case, a general and uniform increase in the rank of coal with depth is usually admitted (Hilt's Law) and, in general, increase in rank means decrease in volatile, matter, but this is directly opposed to the supposed effect of permeability, for the deeper the coal the more difficult will it be for the volatile products to escape. If permeability be important, we should expect the volatile matter to increase with depth, and yet no such regular increase of volatile matter with depth has been reported from any bituminous coalfield, and neither permeability nor jointing can be accepted as a cause of increase in coal rank. The most probable explanation of the increase of induration and jointing with rise of rank is that all three variables—induration, jointing, and rank—are the result of increasing metamorphism; jointing thus having indirect but no direct connection with the amount of volatile matter in the coal.

Anthracite in Canterbury has been convincingly explained as lignite altered by intrusion of igneous rock. (Evans, 1899.) The thermal effect of these intrusions is always local and often irregular and some part of the intrusive rock is usually found without difficulty. But deep erosion has exposed no such intrusive rock in the Fox River area, and, indeed, the regular way in which rank changes show that we have to deal with regional rather than contact metamorphism.

Differences in the coal-forming vegetation have also been suggested as the cause of variations in rank, but it is unlikely that anthracite was deposited as such. Also, if the original material is important, why should variation be regional and irrespective of seam?

Although most of the coal measures in the upper Fox River Valley have been extremely crushed and deformed, a few small blocks have escaped, but the distribution of these uncrushed blocks is irregular and certainly shows no obvious relation to the change in rank. The table of analyses makes this clear, for pulverized coal from the