Art. XXX. — On the Subdivision of the Lower Mesozoic Rocks of New Zealand.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 36, 1903, Unnumbered Page

Art. XXX. — On the Subdivision of the Lower Mesozoic Rocks of New Zealand. By Professor James Park, F.G.S., Director Otago University School of Mines. [Read before the Otago Institute, 11th August, 1903.] Plates XXVIII. to XXXI. Introduction. An examination of the Trias in Nelson, Southland, and alpine Canterbury has led me to the conclusion that the subdivision and correlation of the horizons of this formation present a much more difficult problem than hitherto supposed. Not only do the rocks occur in many disconnected basins, but even in what appears to have been a continuous basin the strata have been so wasted by denudation or disturbed that it is often impossible upon stratigraphical grounds to correlate group with group except within a very limited area. To these difficulties there is added a wonderful uniformity of life throughout the system; and, although the assemblages of life in claystone and sandstone often possess distinctive forms, there is evidence in some places that repetitions of similar conditions of deposition had been attended with a reappearance of the life characteristic of those conditions. The minor subdivisions of a formation are only of local value when the limits of characteristic life and stratigraphical connection can be seen to coincide. It is evident that a classification is of little value that follows too closely the great subdivisions as they occur in Europe, subdivisions, perchance, grouped on the mere presence of a few genera whose association in a particular horizon may have been the result of peculiar local conditions prevailing at the time of deposition. It has been shown, both by geologists and palæontologists alike, that there is a singular lack of homotaxial parallelism between the marine and terrestrial organisms of Australasia and Europe, which, moreover, seems to become more marked and significant as we advance through the Mesozoic and Tertiary periods up to the present time. For example, there is a greater characteristic relationship between the Secondary flora of Australia and the Carboniferous of Europe than there is

between the known Carboniferous floras of these respective-regions; or, to quote a more striking illustration, the Miocene floras of New Zealand and Tasmania, consisting principally of oaks, elms, beeches, alders, laurels, &c., show a closer relationship to the existing vegetation of Europe than to the present floras of these islands. As would naturally be expected, the marine faunas show a closer parallelism than the floras, a fact probably due to the greater facilities for rapid diffusion possessed by marine inhabitants in a continuous sea compared with the slower diffusion of terrestrial organisms checked by physical obstructions such as wide stretches of sea and mountain-chains. The true explanation of this seeming lack of parallelism of life in geologic time is at present a subject for much conjecture and speculation, and doubtless possesses a significance-which is not now very obvious to us. The theory which is thought to satisfy our present knowledge of the facts is that which supposes that the Southern Hemisphere was the cradle of organic life, from which life gradually diffused itself northward by a process of slow migration, differentiated in its progress by accidents of climate and changes in the distribution of land and water. Attractive as such a theory must be to southern geologists, I am inclined for my own part, after a perusal of recent discoveries in arctic Europe and America, to favour the existence of two independent and widely separated centres of distribution, from which migration slowly radiated into the far-distant regions where types of two distinct ages appear to commingle. The notorious likeness existing between the marine faunas of the different geological periods in the Southern Hemisphere with the faunas of the corresponding periods in the Northern Hemisphere, from the Cambrian, when the first assemblage of life suddenly burst upon the geological horizon—life that in many respects was complex and comparatively highly organized, whose appearance was unexpected and perplexing because there were no known probable ancestors—from then upward to the close of the Tertiary, this likeness may be held to support the bipolar theory of distribution. The rate of dispersion, as pointed out by Huxley, Wallace, and other writers, is governed by two dominant factors, of which the position of the centre of origin of each organic unit is one, and the distribution of land and water at the time of dispersion the other. Moreover, in a mixed fauna the powers of migration would not be equal, nor the ability to withstand varying climatic and other adverse conditions. Hence the rate of advance would be unequal; but of this we have no certain evidence. On the contrary, it is probable that the faunas of the different geological periods were similar and

universal throughout the globe in the Palæozoic and Secondary epochs, and that differentiation began with the dawn of present life in the early Eocene. But this contention brings us no nearer the solution of the problem of dispersion, and throws no light on the perplexing difficulties which surround the processes of migration. Furthermore, it is well known to palæontologists that at certain geological epochs, apparently throughout the world, there was a sudden appearance of animals and plants belonging to orders, and even sub-orders, that were not represented by any probable ancestors in older strata. For example, there was the sudden and almost, if not quite, simultaneous appearance in the Upper Cretaceous of acan-thopterygian fishes and dicotyledonous angiospermous plants, now the common existing types throughout the globe. Dispersion to the remoter parts from the centre of origin, or even to the meeting-points from two or more centres of migration, would occupy countless ages measured by our telluric standard of time. A thousand years is but a fraction of a geologic day; and a migration starting at the dawn of a new day would probably reach the remotest regions before the advent of the coming day. Considering the question of distribution over wide areas from a biological standpoint, it seems unlikely that all species originated in one particular hemisphere. On the contrary, it is probable that some species originated in the Northern Hemisphere and other species in the Southern. If we entertain the unicentral theory of origin we must suppose that migration progressed in a succession of great waves; that each wave encountered no insuperable barrier to its forward movement; that each wave maintained its distinctive characteristics notwithstanding the diverse life, latitude, and climate encountered in travelling from one hemisphere to the other; and, further, that each wave was endowed with a migratory instinct that impelled it forward through uncongenial latitudes until it finally arrived at its natural habitat. The problem of the distribution of life is surrounded with many difficulties. Existing conditions are doubtless a reflection of the past, and until we are able to explain the present there seems little hope of unravelling the more tangled problems of geologic time. There is considerable evidence in favour of the belief that, there existed a synchronous parallelism of life in both hemispheres—that is, in corresponding biological provinces and bathymetrical horizons—up till later geological times, when the extreme differentiations of modern types became the dominant inhabitants of the globe. It is, at any rate, certain that the continuous succession of

fossiliferous marine formations in both hemispheres throughout the geological record does not leave much room for chronological divergence; and in applying European time-names against the place-names of our New Zealand formations we are pursuing a course that seems to be in no way inconsistent with the facts that are known relating to the distribution of life in the different geological formations. Present Classifications. At the present time we have two classifications of the Lower Mesozoic and Upper Palæozoic rocks of New Zealand—namely, that formulated in 1878 by Mr. S. Herbert Cox, F. 'G. S., late Government Assistant Geologist,* Reps. Geol. Expl, 1877–78, p. 28. and that by Captain Hutton in 1899.†Hutton, Trans. N. Z. Inst., 1899, vol xxxii., p. 183. In his classification of 1873 Hutton placed the Maitai series in the Lower Jurassic period; but in his classification of 1899, in deference to the opinion of the Director of the Geological Survey, he transferred it to the Permo-carboni-ferous.‡Hutton, i. c., p. 183. I have shown in another paper that the Maitai series must be referred back to the Jurassic system, where it was originally placed by Captain Hutton in 1875. Mr. Cox, in his excellent geological survey of the Hokonui Mountains, in Southland, in 1877 (in which he was assisted by Mr. McKay), subdivided the Lower Mesozoic rocks into seventy-three horizons, which he subsequently gathered into seven groups or series of beds, as follows: Upper Oolite Mataura series. Middle Oolite Putataka series. Lower Oolite Flag Hill series. Lias Bastion series. Upper Trias (Rhætic) Otapiri series. Middle Trias Wairoa series. Permian Kaihiku series. In Captain Hutton's classification of 1899 the Lower Mesozoic rocks are grouped as follows: Hokonui System— Mataura series Lower Jurassic. Wairoa series Triassic. Taking the Jurassic rocks first, it is apparent that the “Mataura series” of Cox and the “Mataura series” of Hutton differ very widely in their meaning. In Captain Hutton's table the Mataura series includes the whole of the Jurassic rocks, and in this respect is the same as the Mataura

series of Dr. Lindsay, and of Sir James Hector up till 1873. On the other hand, the Mataura series of Mr. Cox includes only the higher beds of the Jurassic system. The same divergence is also seen in respect to the Wairoa series of Hutton and Cox. With Captain Hutton the Wairoa series includes all the Triassic and Permian of Cox and Hector, while the Wairoa series of Cox embraces only the beds of the Middle Trias. In 1886, acting under the instructions of the Director of the Geological Survey, I revised the work of Cox in the Hokonui Hills, and discovered that his Mataura and Flag Hill series referred to the same group of beds.* Park, Reps. Geol. Expl., 1886–87, p. 146. His Putataka series I was unable to locate as a separate group. At the time of my examination of the Hokonuis in 1886 I was convinced that Mr. Cox had attempted a subdivision that could not be maintained, not only because of the overlapping of his own groups, but also for the reason that his groups were not sufficiently defined either on stratigraphical or palæontological grounds to be of service for correlative purposes. An examination of the Nelson Trias having shown the subsequent date of the Maitai formation, the way seemed clear for a rearrangement of the Lower Mesozoic and Upper Palæozoic formations on a broader and safer basis than that employed by Cox in 1878. With that object in view I have made a special examination of the typical sections of the Trias at Nugget Point, North Otago, Upper Rangitata, and Nelson, so as to gather sufficient data to enable me to arrange the different beds in distinctive groups of correlative value. After a study of the characteristic features of each locality, I have finally selected the Shaw Bay section at Nugget Point, on account of its stratigraphical clearness and numerous marine zones, as the most typical for the subdivision of the Trias and associated beds. The classifications of Captain Hutton, Mr. Cox, and of the author are shown below in tabulated form:— Hutton, 1899. Cox, 1878 Park, 1903. Mataura series (Jurassic) Mataura series Putataka series Flag Hill series Bastion series 1. Mataura series Wairoa series (Triassic) Otapiri series Wairoa series Oreti series Kaihiku series 2. Shaw Bay series— (a.) Athyris or Clavigera beds. (b.) Mytilus and Monotis beds. (c.) Trigonia beds. (d.) Halobia beds. (e.) Spiriferina beds. (f.) Nugget Point plant beds.

Nugget Point District, Otago. From the promontory on which Nugget Point Lighthouse stands southward to False Island, at the mouth of Catlin's River, in a distance of nearly four miles, there is exposed along the coast-line a thickness of over 18,000 ft. of strata, forming a continuous succession of beds dipping to the southwest at angles that are always very high. South of Catlin's River the higher beds occur in gently undulating folds, and in this manner pass southward toward Waikawa. This great assemblage of beds consists of claystones, sandstones, greywacke, and conglomerates, occurring sometimes in thin alternating layers, sometimes in thick beds. Organic remains are not wanting, but are seldom abundant-Plant-remains, generally broken and fragmentary, are scattered from the base to the summit, and in the higher beds at Catlin's large masses of silicified driftwood are not uncommon. Marine shells occur at wide intervals, but are never plentiful except in a few of the lower zones in Shaw Bay, which is now locally known as Roaring Bay. The character of the sediments and contained fossils clearly indicate that fluvio - marine or estuarine conditions prevailed during the deposition of these beds; and the abundance of granite and related rocks in the lower conglomerates tends to show that the area of erosion whose dissection provided the materials lay to the south-west, probably in the direction of Stewart Island. The fossils and continuous succession of stratified sediments prove that deposition extended uninterruptedly from Upper Palæozoic to Jurassic times. The lower beds at the north end of Shaw Bay are intruded by a narrow dyke of porphyrite, and excepting this the whole system is singularly free from evidences of contemporary volcanic activity. The rocks exposed on the coast-line between Nugget Point and the south head of Shaw Bay, embracing a thickness of nearly 5,000 ft. of strata, belong to the Trias and Upper Palæozoic periods, and the remaining 13,000 ft., lying between the south head of Shaw Bay and Catlin's River, to the Jurassic. In 1873 Mr. McKay reported the existence in this district of a “Productus formation,”* Reps. Geol. Expl., 1872–73, pp. 63–65. and my first object was to locate this formation, with the view of ascertaining its relation to the Trias. I made a careful examination of the so-called Productus beds, and was unable to discover any fossils that

could be referred to the genus Productus. I was therefore compelled to conclude that an error had been made in the determination of the genus. Before discussing the classification of the Lower Mesozoic rocks in its wider aspect, I shall first detail the sequence and fossil contents of the different beds as we find them exposed at Shaw Bay and Catlin's River and in the typical Trias sections at Nelson and Upper Rangitata, in order that the facts upon which my subdivisions are based may be placed on record. In my examination of the Nugget Point and Catlin's River sections I was accompanied by Mr. A. Hamilton, Director of the Colonial Museum, to whom I was indebted for much valuable assistance in the collection and identification of the fossils. Previous Geological Examinations.—The Nugget Point district, so far as I can gather, was first geologically examined in 1869 by Dr. W. Lauder Lindsay, who classified the rocks in Shaw—that is, Roaring—Bay as Triassic. Mr. McKay, in 1873, made extensive collections of fossils here for the New Zealand Geological Department, but in his report he does not refer the rocks to any particular age. However, the Director of the Department, in referring to Mr. McKay's collections in the Sixth Annual Report of the New Zealand Institute, considered that the fossils indicated a range from Jurassic to Upper Carboniferous. * Trans. and Proc. N Z. Inst., 1875, p. 564, The next examination of this district was made in the summer of 1874 by Captain Huttou, F.R.S., who grouped the Shaw Bay beds with his Putataka formation of Lower Jurassic age.†“Geology of Otago,” 1875, p. 43. Since that date no further geological examination has been made of this district till the present year, covering an interval of twenty-nine years. Kaka Point to Hay's Gap. The Trias rocks are well exposed at Kaka Point and many places on the beach before Hay's Gap is reached, but the outcrops are small and so isolated by stretches of sand that no continuous section can be obtained. For that reason no time was spent in making observations where so much obscurity existed, and where the results could not be other than doubtful and possibly conflicting. From the beds of Wiltshire Beach McKay collected what he calls a “spirifer,” but from his description it was probably a species of Spiriferina which I found there associated with Halobia, Rhynchonella, &c. ‡L.c., p. 69.

In the claystones forming the first line of low reefs on the beach south of Campbell's Greek I identified Halobia and Spiriferina, which would indicate a repetition of the Wiltshire beds in that direction. Hay's Gap to Nugget Point. The distance between these points is about two miles, the direction from Hay's Gap for the first mile being almost due south (true meridian), and thence, for the second mile, southeast. At Hay's Gap the Trias rocks consist of sandstones and claystones, in places alternating in thin laminæ, often highly indurated, jointed and seamed with veins of iron-peroxide Their strike is E.—W., and the dip south at an angle of 58°. Some 5 chains south of the Gap the strike changes to a direction between E. and E.N.E., with a nearly S S.E. dip at an angle of 49°; at 6 chains south the strike is about E.S.E.-W.N.W., and the dip nearly S.S.W. at an angle of 78°; while at a distance of 7 ½ chains the strike is exactly between E. and E.S.E., with the dip still southerly at an angle of 57°. Thus in a distance of less than 170 yards the strike has changed through an arc of about 30°. The nipping of the beds caused by this sudden change of strike has resulted in a steepening of the dip towards the centre of the nip. Proceeding southward, the strike bends still more to the southward, and at a distance of half a mile from the Gap is N.W.-S.E., and the dip S.W. at an angle of 42°. Towards the Nuggets the sandstones become coarse in texture, and in places assume the character of greywacke. They form high rocky points and numerous isolated flat reefs on the beach below high-water mark, separated by stretches of sand. Immediately north of “Boat-landing” the strike varies between N.W. and W.N.W., while the dip is still southerly at angles generally about 65°. At the point where the road leading to the lighthouse leaves the beach—that is, about 300 yards south-east of Boat-landing—the strata are intruded by a dyke of augite porphyrite which is well exposed in the south side of the roadcutting. This dyke is of especial interest, as it is the only igneous rock found associated with the Trias formation in this district. It runs almost parallel with the bedding-planes of the enclosing rocks, and can be traced across the saddle into Shaw Bay, a distance of 50 chains. Near Boat-landing it is about 9 ft. thick, but in two places on the ridge dividing Boat-

landing from Shaw Bay it bulges out into great lentil-shaped masses resembling the “laccolites” of Gilbert. The actual line of contact between the porphyrite and sandstone is clearly exposed in the road-cuttings, and, so far as can be determined, the latter seems to have suffered very little alteration, the only noticeable difference being a somewhat greater degree of induration than elsewhere. At the north head of Roaring Bay the dyke throws out ramifying branches into the enclosing sandstone. The porphyrite is granitoid in texture, and weathers to a soft crumbling yellowish earthy brown rock of a dark greyish-green or blue colour, in which the large well-developed phenocrysts of plagioclase become conspicuous when partially kaolmised. It is tough, but not hard, and ought to form a building-stone of superior quality. In polarised light, the rock is seen to be composed principally of plagioclase, almost to the exclusion of the augite. The latter surrounds the feldspars like a cementing medium, filling the interspaces and consequently occurring mostly as narrow irregular aggregates. The plagioclase is generally kaolinised, opaque, and turbid. The augite is mostly altered to greenish serpentinous matter and secondary magnetite. At the point where the road leading to the lighthouse leaves the beach, near Boat-landing, there is a bed of clay-stone a few feet below the dyke which contains a considerable number of marine forms in a fair state of preservation, occurring principally as casts. Among the genera identified here were Spiriferina (Psioidea of Hector), Spiriferina (two sp.), Epithyris, Rhynchonella, and Pleurotomaria. Besides these, fragments of a bivalve shell resembling Modiolopsis were not uncommon. This fossiliferous horizon is of especial value, as its contents establish the Triassic or Permo-triassic age of the formation, and enable the position of the associated beds to be correlated with the Nelson and Southland Lower Trias or Permian beds. In the sandstones and underlying the Spiriferina beds there is a thin stratum of indurated sandstone crowded in a few places with casts of single valves of a thin bivalve shell. On the surface of rounded boulders, when worn at right, angles to the plane of deposition, the casts are seen in section. The profile of the single valves gives a concave appearance somewhat resembling the ventral valve of a Productus, and the resemblance to that genus becomes more deceptive when two valves happen to fit into each other. The casts are those of a very thin gibbose conchiferous shell from 0.5 in. to 0.75 in. high and about the same width.

The remains are too fragmentary even for generic identification. The Spiriferina beds are underlain in descending order by sandstones, often coarse and gritty, and sandstones and clay-stones alternating in thin laminæ. The latter form the “Nuggets” and the promontory on which the lighthouse stands. Here the beds strike W.N.W. - E.S.E., and dip S.S.W. at very high angles, at places becoming nearly vertical. On the islet, in the centre of the Nugget group of rocks, the strata are sharply folded and pinched. In the road-cutting leading to the lighthouse, at a point about 15 chains beyond the keeper's house, the strike of the beds is between E. and E.S E., and the dip about S S.W. at an angle of 70°. Here the sandstones are interbedded with beds of soft shaly claystone containing numerous plant-remains, but too broken and fragmentary to be identified. “The plant beds are about 850 ft. below Spiriferina beds. Shaw Bay Section. From the head of Shaw Bay to the headland separating that bay from Sandy Bay, a distance of half a mile, the strike of the beds is W.N.W. - E.S.E almost without variation, and the dip continuously S.S.W. at very high angles, generally between 75° and 85°. The coast-line trends almost at right angles to the strike of the beds, and, the angle of dip being high, bed succeeds bed with wonderful regularity, appearing in the sea-cliffs which enclose the bay as clearly defined as books standing on a shelf. There is exposed along this line of section a thickness of fully 2,500 ft. of strata in one continuous sequence, forming the most complete assemblage of Lower Mesozoic rocks in New Zealand, singularly free from faults or complicated foldings that might cause obscurity or lead to doubtful interpretation. Furthermore, the fossiliferous zones are numerous, and so distinctive as to be easily recognised. Beginning at the north end of Shaw Bay, the beds exposed along this line of section and their approximate thickness are as follows in ascending order:— 1. Thin laminated claystones and sandstones with plantremains, probably 200 ft. thick. 2. Indurated sandstones and greywacke, 150 ft. 3. Spiriferina beds—dark-blue claystones, 100 ft. 4. Porphyrite dyke. 5. Greywacke. 6. Halobia beds about 210 ft. thick—crumbling claystones containing Halobia lommeli, Wissm., Retzia, Epithyris. Seen in road-cutting descending to the bay.

7. Sandstones and claystones, 250 ft., not exposed on beach. 8. Breccia-conglomerate, angle of dip 77°, 20 ft. 9. Highly indurated claystones, 55 ft. 10. Hard greywacke, 15 ft. 11. Thinly laminated claystones and sandstones, 24 ft. 12. Breccia-conglomerate, slaty and granitic, 27 ft. 13. Gritty sandstone, 15 ft. 14. Indurated sandstones with occasional beds of claystone, 200 ft., nearly vertical. 15. Granitic conglomerate, 10 in. thick; angle of dip 86°. 16. Indurated claystone, 50 ft. 17. Trigonia bed—claystones, 10 ft. Contains an ammonoid shell, Trigonia, Pleurotomaria, Terebratula, Pentacrinus, &c. 18. Sandstones and claystones with subordinate beds of gritty sandstone, 450 ft. 19. Oyster and Mytilus bed, 29 ft.; angle of dip 78°. 20. Coarse sandstones, 600 ft. 21. Athyris beds, 10 ft.—sandstones containing. Athyris (Clavigera of Hector), Spiriferina (Rastelligera of Hector). 22. Plant beds—sandstones and claystones. The Athyris beds are standing nearly vertical or incline slightly to the N.N.E. for a few yards. Around the rocky headland the dip resumes its normal S.S.W. dip at very high angles, and from this point to the south side of Sandy Bay the dip is continuously in that direction, exposing a thickness of probably not less than 8,000 ft. of strata. There are three zones of Athyris separated by beds of barren sandstone, the sequence in ascending order being as follows:— 1. Dark-blue indurated claystones. 2. Bed of granitic conglomerate, 4 in. thick. 3. Athyris bed, 40in., richly fossiliferous. 4. Sandstone, 4 ft. 5. Athyris bed, 2 ft. 6. Sandstone, 4 ft 7. Athyris bed, 18 in. 8. Indurated sandstones. The upper Athyris bed (No. 7) is a coarse gritty, sometimes pebbly, sandstone containing shells of Athyris of at least two different species in such abundance that the rock is in places moderately calcareous. At this rocky headland the strata are nearly vertical, and, the overlaying thin layer of conglomerate and claystones having been worn down by

denudation, the lower fossiliferous Athyris bed (No. 3) stands up like the wall of a house, presenting an even surface 70 ft. high and 100 ft. long thickly incrusted with shells of Athyris which have weathered out of the matrix. By far the greater proportion of the forms are impunctate Athyris, but the large Spiriferna (Rastelligera), possessing fine comb-like dentition along the hinge-line, is not uncommon. In the middle and upper fossiliferous zones Athyris are sparingly scattered in a sandstone matrix. The Athyris beds form the natural close of the Trias formation. On stratigraphical and palæontological grounds they may be safely correlated with the Athyris beds forming the close of the Trias at the Wairoa Gorge and Eighty-eight Valley in Nelson, where the sequence of fossiliferous horizons shows a remarkable correspondence with that of Shaw Bay. It is impossible in such widely separated districts to correlate bed with bed, but a comparison of the sequence clearly demonstrates a general harmony in the order of succession of the different zones. In Eighty-eight Valley the sequence of the Trias, as indicated by the fossiliferous zones, is as follows:— 1. Athyris beds. 2. Mytilus beds. 3. Trigonia and Spirigera beds, with Mytilus. 4. Halobia and Spiriferina beds. 5. Plant beds. The Athyris is a very unusual form, and is of great zonal value. It has been identified at the close of the Trias throughout New Zealand. Its distinctive characters are referred to in my paper on the Jurassic age of the Maitai formation, which appears in another part of this volume. Monotis salinaria var. richmondiana, Zitt., which is so abundant in some parts of the Trias of Nelson and Southland, generally occurs in a horizon above the Trigonia beds. In his report of 1873 Mr. McKay mentions that the upper conglomerate divides an Ammonite bed into two parts. The Ammonite bed is described as being “very rich in chambered shells of various genera”; but, unfortunately, we were unable to find any of the genera reported by McKay, and succeeded in discovering only one fragment of an ammonoid shell among the beach debris. Dr. Marshall, who visited this place with a party of his students in July, was more successful. He discovered a number of Ammonites, but they were all broken and fragmentary. The great assemblage of sandstones, claystones, and slaty

shales which conformably follow the Athyris beds in the direction of Sandy Bay have been referred by Captain Hutton to the Mataura formation, and I have elsewhere shown that this formation is the equivalent of the Maitai rocks which overlie the Trias of Nelson conformably. On the occasion of my visit to Nugget Point the coast-line was followed as far south as the south headland of Sandy Bay, and in that distance the dip was observed to be southwards at angles seldom under 75°. Catlin's River District. From the south head of Sandy Bay to the centre of False Island the dip of the beds is to the S.S.W. at varying high angles, which are not often under 70° or 75°. In this distance the thickness of strata exposed along the coast-line is probably not less than 13,000 ft. At False Island the higher beds form a syncline, and proceeding southward from Catlin's Heads the same beds form a series of gently undulating anticlines and synclines, thereby preventing the exposure or repetition of the lower Triassic beds. The beds consist of alternations of claystone, often shaly, sandstone, greywacke, and conglomerates. Three distinct fossiliferous horizons were found between Cannibal Bay and Sandy Bay, all previously described by McKay. In the first bay south of Sandy Bay, marked “Boat-landing Bay” on the survey map, McKay collected a number of fossils, but he did not mention any of the genera represented in his collection. In the same beds at Glenomaru quarry reserve, which is two miles inland, we collected, in indurated sandy claystones, a small species of Ammonites, Pinna, Arca, Panopæa, and Pholodomya. This is apparently the lowest fossiliferous horizon of the Jurassic system, and it is of interest to note that it lies about 7,500 ft. above the Athyris bed at the top of the Trias in Shaw Bay. The intervening beds are principally claystones, grey sandstones, or greywacke devoid of organic remains, except a few scattered plant-remains. In Tuck's Bay—that is the bay between Boat-landing Bay and Cannibal Bay—the rocks are principally sandstones, often coarse and pebbly. At the quarry reserve near the saddle overlooking Tuck's Bay we collected a number of Inoceramus haasti, Inoceramus labiatus, Schloth., and a small Pecten, which were found in the hard greenish-blue sandstone in which the road is excavated at that place. This horizon is about 2,000 ft. above the lower Ammonite horizon.

From the quarry reserve southward the rocks are principally coarse sandstones, which extend almost to Cannibal Bay, when they are overlain by dark-blue indurated claystones, which strike between E. and E.N.E., and dip about S.S.W. at an angle of 82°. These claystones form the north side of Cannibal Bay. In them we found Trigonia, Ostrea, Anomia, Inoceramus, Pholadomya, and a Venus. Besides these, Mr. McKay reports the occurrence of Ammonites and Belemnites at this place, but they must be scarce, as we were unable to find any trace of them. Between the north end of Cannibal Bay and False Island the rocks are obscured by sand. The lowest rocks on north end of False Island are conglomerates, followed by coarse pebbly sandstones, and these in turn by sandstones and clay-stones. About the centre of the island the dip changes to the northward, whereby the sandstones and conglomerates are repeated on the south side of the island. The southern end of the island is traversed by a fault which repeats the conglomerate and brings into view some underlying sandstones. The thickness of the conglomerate is about 145 ft. From the mouth of Catlin's River southward to Jack's Island the Cannibal Bay beds form the sea-cliffs, lying horizontal or occurring in gently undulating folds. In the indurated dark-blue claystones at the north of Makura Bay were found Ammonites novo-zealandicus, Hauer, Inoceramus haasti, Pholadomya, a Venus, and a number of imperfect casts of conchiferous bivalves. A number of indistinct bivalves occur in the sandstone cliffs opposite the south end of Jack's Island; and in the flat shelving rocks in Catlin's River, immediately opposite Pouna-wea Township, were found a considerable number of indistinct casts imbedded in a soft yellowish-brown sandstone. The Trias of Nelson. The Trias formation in Nelson is clearly exposed in three typical sections— namely, the section from Waimea Plains across Mount Heslington to the Wairoa Gorge; that from Wakefield to Sellen's run; and that along the course of Well's Creek, at Eighty-eight Valley, a few miles west of Wakefield. These sections are separated by intervals of a few miles, and as each crosses the same basin the sequence of rocks exposed in each is practically the same. The clearest and most complete section is exposed in Well's Creek, where the sequence is as follows in descending order:—

O. Clavigera (Athyris) bed, calcareous pebbly sandstone. 1. Claystones. 2. Upper granite conglomerate. 3. Mytilus sandstone and claystone. 4. Spirigera sandstone. 5. Mytiius sandstone. 6. Spirigera and Trigonia sandstones. 7. Lower granite conglomerate. 8. Halobia claystones. 9. Spiriferina claystones and sandstones. 10. Plant beds, with Tæniopteris, Asplenium, &c. The parallelism of the facies existing between the Nelson and Nugget Point sections is remarkable when we remember the distance which separates these places. The details of the Nelson sections are given in my paper “On the Jurassic Age of the Maitai Formation,” and need not be repeated here. Mount Potts District, Upper Rangitata. The Permo-trias and succeeding Jurassic rocks, extending from Mount Somers across the Harper Range westward to the Southern Alps, occur in a series of sharp folds. The core of the western anticlinal fold is well exposed in Tank Gully, a deep chasm excavated on the western flanks of Mount Potts, situated in the upper Rangitata. Here are seen the lowest beds of the Permo-trias, which consist of black carbonaceous slaty shales and sandstones containing a number of well-preserved plant-remains. A collection of this fossil flora made by Sir Julius von Haast was submitted to the European palæobotanist Baron von Ettingshausen, who distinguished six different genera, of which the species were, he stated, mostly analogous to Tri-assic ones.* Trans. and Proc. N.Z. Inst., 1890, vol. xxiii., p. 241. The outcrop of the plant beds is very small, not much ex ceeding an acre in extent. It is situated in the higher part of Tank Gully, along the bottom of which it stretches for a distance of about 150 yards. The lower end of the outcrop is about 3,150 ft. above sea-level, and the upper part perhaps 200 ft. higher. The thickness of beds exposed is about 110 ft. The plant beds are deeply involved in the core of a sharp anticline, and from their yielding character are much crushed and slickensided. They are surmounted by thousands of feet of sandstone and claystone, which along the axis of the anticline are also greatly bent and crushed. Where the strata did not bend they have been subjected to lateral shearing stress, the effects of which can still be traced

in some of the high precipitous rock-faces. A notable example is seen on the south-west side of Tank Gully, where the strata forming the summit of the ridge have been sheared along a nearly horizontal thrust-plane for some 17 ft., thus displacing them relatively to the strata below to that amount. The strike of the beds on the north side of Tank Gully is N.E.-S.W., and the dip N.W. at very high angles, which are seldom under 80°. On the south side the strike is the same, and the dip S.E. at angles varying from 60° to 75°. Mr. McKay, in 1877, examined this locality and noted the N.W. and N.E. dips, but he apparently failed to recognise the anticlinal arrangement of the beds.* Reps. Geol. Expl., 1877–78, p. 94. In a section across Tank Gully he shows the dips correctly; but he marks the beds on the N.W. side of the anticline as Trias and those on the N.E. side as Carboniferous (Rimutaka series), and extends the latter in his sketch so as to pass unconformably below the former. It is noteworthy, however, that his description of the strata in Tank Gully does not correspond with his section. The plant beds are succeeded by a series of greyish and greenish sandstones, slaty shales, and claystones. The shales and sandstones alternate both in thin laminæ and in thick beds. The greenish sandstones are generally highly indurated, and weather to a greyish colour on exposed surfaces. They are of all degrees of texture from fine-grained, which predominate, to coarse and gritty. They occasionally become pebbly, and pass into thin beds of sandstone conglomerate. In a few places the finer-grained sandstones are brecciated with small angular fragments of black claystone scattered sparingly through a thickness of a few feet. The sandstones and claystones between Tank Gully and Rocky Gully are interbedded with a few beds of red slaty shale and a bed of pale-green aphanitic sandstone, which is often streaked with thin interlacing veins of white quartz. Sir Julius von Haast, in 1861, examined the upper Rangitata,†Von Haast, “Geology of Canterbury and Westland,” 1878, p. 268. and on that occasion discovered the fossiliferous beds in Rocky Gully. He subsequently revisited the same locality, in 1872, ‡ Reps. Geol. Expl., 1872–73, p. 1. and made a collection of marine shells and saurian bones, the latter including a number of vertebræ, one of which was referred by Sir James Hector to the genus Ichthyosaurus. § Trans. N.Z. Inst, vol. vi., p. 334. At the same time a collection of plantremains was made at Tank Gully. In his collection from Rocky Gully Sir Julius von Haast reported the occurrence of Orthis, Productus, Atrypa, and

Euomphalus, all Palæozoic genera, which led him to place his Mount Torlesse formation, of which these beds formed a part, in the Carboniferous period. * Reps. Geol. Expl., 1872–73, p. 6. These genera have never been identified in other collections made at Rocky Gully, and we can only conclude that the forms to which these names were applied belonged to other genera. The Rocky Gully marine beds are about 2,000 ft. higher in the series than the plant beds, being separated from the latter by the sandstones, shales, and claystones just described. The conformable relations of the Rocky Gully Spiniferina beds to Tank Gully plant beds seems to admit of little doubt, and is a point upon which Sir Julius von Haast and Mr. McKay were both agreed. The fossils occur in indurated claystones and sandstones, which are interbedded with a pebbly bone bed varying from 2ft. to 4ft. thick. The bone bed contains a large number of broken shells and rolled fragments of bone. The sandstones, on each side of the bone bed, contain a few fossils, mostly Brachiopoda, scattered through a thickness of several feet. A more varied assemblage of marine shells occurs in a bed of black claystone underlying the sandstone zone. This claystone has been subjected to great lateral thrust, with the result that the contained fossils are often flattened and distorted; while the stress has induced in the rock itself a tendency to break on exposed surfaces into long prismoidal pencils, an effect which adds considerably to the difficulty of securing perfect unbroken shells. In places the fossiliferous sandstones and claystones are slightly calcareous. I collected from the lower horizon Halobia(?), Megalodon(?), Nuculana, Athyris, Rhynchonella, and Spiriferina, three sp.; and from the upper or bone-bed horizon Orthoceras, Pleurotomaria, Turbo, Pentacrinus, several bivalves not determined, and some fragments of saurian bones. Two small vertebræ were noted, but not collected, in a cavity in a large boulder. They were exposed in elevation so as to show the full size of the disc, which in the case of one was ¾in. in diameter, and of the other 1/16in. less. They were flat or nearly so. From Rocky Gully the section was followed up the Clyde Valley to the foot of Mount Goethe, on the main alpine divide, a distance of nine miles as the crow flies. The rocks consist of a great succession of sandstones, greywacke, shales, and claystones. The sandstones are palegrey, greenish-grey, and yellowish-green in colour, and vary, in texture from fine-grained to coarse, gritty, and pebbly, in places passing into conglomerate. They are siliceous, and

argillaceous, and frequently feldspathic, the latter passing into greywacke. They are mostly hard, being seldom soft, except where crushed, much jointed and shattered. Veins of quartz are common in the sandstones, especially in the yellowish-green variety. The shales and claystones are present in endless variety, ranging in colour from pale-grey to yellowish-grey and green, from dark-green to black; and from chocolate-colour to red and purple. They are argillaceous and occasionally siliceous, but, like the sandstones, rarely calcareous. They are generally much jointed, and, when interlaminated with bands of the harder and more unyielding sandstone, are often crushed into a crumbling and drossy material presenting shining and slickensided surfaces, which led Sir Julius von Haast to describe them as “serpentinous.” Some thin bands of slaty shale are silky in lustre and quite fissile. These rocks dip at high angles in one continuous succession of beds to the north-west, thereby exposing between Mount Potts and Mount Goethe a thickness of not less than 40,000 ft. of strata. The strike of the strata, from Rocky Gully to the junction of the Clyde and Lawrence Rivers, and thence up the Clyde for a mile and a half, is uniformly N.E.-S.W., and the dip N.W. The angle of dip varies from 65° to 70° at Rocky Gully to 83° at the river-junction, and in the Clyde for the first two miles from 85° to nearly vertical. The strike, about a mile and a half above the junction, trends more towards the north, gradually bending from N E. to N.N.E. until it is about N.-S. two miles above the Lawrence. From Maiden Flat Creek to Mount Goethe the strike is N.N E.-S.S.W., and the dip W.N.W. at angles varying from 55° to 85°, the flatter angle being observed near Maiden Flat Creek. The most careful search failed to discover organic remains in this magnificent pile of clastic rocks, excepting some obscure markings in the bands of black silky slate, which might be referred to annelid trails. No less remarkable than the absence of organic remains was the absence of limestones or calcareous beds, a circumstance commented upon by Sir Julius von Haast, who examined the same rocks in the Rakaia Valley and neighbourhood of Mount Cook. The Jurassic and Triassic formations of New Zealand are everywhere connected in one continuous stratigraphical succession, and where the Upper Trias are fossiliferous, as at Nugget Point and Nelson, it is not difficult to establish a limit to the one and a beginning to the other. But in the Mount Potts section the Permian alone are fossiliferous, and it becomes a matter of difficulty to determine where the Trias

shall end and the Jurassic begin. And, after all, it is not a matter of very great importance whether we classify this great pile of strata as wholly Triassic or partly Triassic and partly Jurassic. The stratigraphical relations of the different beds must always remain the same wherever they are placed in the geological scale. On the other hand, the well-defined limitations of the Trias in Otago and Nelson leave little doubt that the Clyde River beds should be correlated with the Mataura and Maitai formations of Jurassic age. Triassic Marine Rocks. The position of the marine zones associated with the Triassic beds at Nelson, Hokonui Hills, and Nugget Point have now been accurately determined. The character and thickness of the sediments separating the zones were necessarily governed by the local conditions prevailing at the time of deposition, and therefore are found to vary greatly in the different basins; but the succession and relative position of the zones are always the same, clearly indicating that the different basins were connected by a continuous sea. The zones contain a number of characteristic forms of fossils, and many of the species have been determined. Where the species have not been determined the genera of the Brachiopoda are often sufficiently distinctive for correlative purposes. The prevalence of Brachiopoda in the lower and higher marine zones is conspicuous in all the Triassic basins throughout New Zealand. Jurassic Marine Rocks. The marine zones associated with the Jurassic beds of New Zealand are few in number and seldom rich in fossil forms. Characteristic fossils are rare, and the majority of the species have not yet been determined. In his survey of the Hokonui Hills, in 1878, Mr. Cox subdivided the Jurassic rocks into four different groups or series of beds, but a subsequent examination which I made of his sections showed that his groups contained no characteristic marine zones, while in some cases they were found to overlap, each other. Groups that cannot be identified in their typical localities are necessarily valueless for purposes of correlation. The brachiopods which are so abundant in the Upper-Trias in Nelson and Nugget Point basins are sparingly represented in the overlying marine zones of the Jurassic; and gasteropods are equally rare. As late as 1899, Captain Hutton attempted no subdivision of the Jurassic rocks, and in the absence of sufficient data I

do not now propose to do so, but shall refer these rocks to my Mataura series, which is the same as Hutton's Mataura series of 1899. * Hutton, Trans. N.Z. Inst., 1899, vol. xxxii., p. 183. Hokonui System—Hutton† When I placed this paper before the Institute I included the Mataura, Shaw's Bay, and Kaihiku series in my Rangitata system. The name Rangitata seemed to me at that time to be more suitable than either the name Mount Torlesse or Hokonui; but on further consideration I have decided to adopt Captain Hutton's name Hokonui, using it in a somewhat more extended sense than originally intended by its author. The Hokonui system will now include all the rocks that are stratigraphically connected from the Permian to the Jurassic.(Syn., Rangitata System— Park). The vast assemblage of sedimentary strata just described comprises wholly, or in part, the Mount Torlesse formation of Von Haast. The rocks, unfortunately, in Mount Torlesse contain none of the fossiliferous horizons present in Nelson, upper Rangitata, or Nuggeet Point, with the exception of some indistinct plant-remains and a calcareous flattened tubular body which has been doubtfully referred to an annelid. In the absence of shell beds it is impossible to fix the position of the plant and annelid beds in relation to known horizons elsewhere. The Mount Torlesse plant beds may be contemporary with the Mount Potts plant beds, but in the absence of satisfactory evidence this correlation may be called in question at any time. The Mount Torlesse formation, according to Von Haast, has a thickness of at least 25,000 ft. ‡ Haast, “Geology of Canterbury and Westland,” 1879, p. 279. It was placed by him up till 1865 in the Silurian period, but in 1878 he referred it to a time ranging from Palæozoic to Lower Mesozoic. § L.c., p. 279. The Carboniferous age of the lower beds was apparently based on the supposed identification of Orthis, Productus, and other Palæozoic genera in the Mount Potts Spinferina beds at Rocky Gully. ∥L.c., p. 272. The assemblage of strata exposed in the upper Rangitata attains a thickness not less than 40,000 ft. It possesses, too, well-marked fossiliferous groups of beds which are absent in the Mount Torlesse district. I am therefore reluctantly compelled to set aside Von Haast's formation-name, chiefly for the reason that the strata at Mount Torlesse do not afford the data necessary for their subdivision into groups and series of beds. The main object of my investigations in the past three years has been to determine the succession and stratigraphical

relations of the Trias and associated rocks, and to establish a classification that shall be a safe standard for the correlation and grouping of strata throughout the colony belonging to the Lower Mesozoic and Upper Palæozoic periods. The nomenclature I have adopted is as follows:— Jurassic … … 1. Mataura series. Triassic … 2. Shaw's Bay series. (a.) Clavigera beds. Hokonui system. (b.) Mytilus problematicus and Monotis salinaria beds. (c.) Trigonia beds. (d.) Halobia lommeli beds. (e.) Spiriferina beds. (f.) Nugget Point plant beds. This classification is an extension of that formulated by Captain Hutton in 1899, and a modification of that adopted by the Geological Survey in 1878. The boundary between the Oreti and Kaihiku series of the Geological Survey is not very well defined, and in some places these series seem to cover the same set of beds. In other places, as in Nelson, the Wairoa series includes beds that are elsewhere included in the Oreti and Kaihiku series; and the Clavigera beds that in the Hokonui district constitute the Otapiri series have in Nelson been referred by the Geological Survey to the Kaihiku series.* Reps. Geol. Expl., 1878–79, p. 117. The Wairoa series, as defined by Sir James Hector in his “Outline of New Zealand Geology,” 1886, pp. 72–74, apparently includes all the beds that are elsewhere considered to constitute the Oreti and Kaihiku series of the Geological Survey. (f.) Nugget Point Plant Beds.—These are the lowest beds exposed at Nugget Point and Eighty-eight Valley, and they are correlated with the Mount Potts plant beds, which are also the lowest rocks exposed in the upper Rangitata. At the two former places they underline the Spiriferina (Psioidea) beds; and at Mount Potts the Tank Gully plant beds appear to underlie the shell beds at Rocky Gully. At Nugget Point the plant-remains are broken and indistinct; at Hunt's farm, in Eighty-eight Valley, they are broken, but distinguishable; while at Tank Gully the beds contain a varied and well-preserved flora. As I have just mentioned, the Tank Gully beds seem clearly enough to underlie the shell beds at Rocky Gully. They are readily accessible, easily found, and occur in such a position that their outcrop cannot be obscured or obliterated by cultivation or the con-

structions of man. It was, moreover, from this place that Sir Julius von Haast obtained the collection of plant-remains-sent to Europe for identification. The place has therefore claims to be considered a typical locality, and, but for the doubt that still remains as to the age of the Mount Potts shell beds at Rocky Gully, I should have preferred to designate this group of beds the “Mount Potts plant beds,” instead of “Nugget Point plant beds.” The Nugget Point beds form the long narrow promontory on which the lighthouse stands, and the outlying rocky islets. They are followed conformably by the Spiriferina beds. The dip of the beds is to seaward, and consequently the lowest beds of the series are not seen. The thickness exposed to view at low water is about 350 ft. Sir James Hector identified among the plant-remains collected by Mr. McKay at Hunt's farm, in Eighty-eight. Valley, Glossopteris, Zamites, and Rhacophyllum. * Hector, “Outline of New Zealand Geology,” 1886, p. 74. Among the plant-remains from Tank Gully, Mount Potts, Baron Von Ettingshausen determined Tæniopteris pseudovittata, Asplenium hochstetteri, Palissya podocarpiodes, Baiera australis, and others, all more or less related to European Triassic forms. † Trans N.Z. Inst., vol. xxiii., p. 237. The form from Mount Potts identified by Sir James Hector as Glossopteris was determined by Ettingshausen to be Tæniopteris, and probably the Nelson Glossopteris also belongs to that genus. (e.) Spiriferina Beds.—These embrace the lowest marine beds of the Hokonui system. Among the fossil shells found in them at Nelson, Nugget Point, and Hokonui Hills are Spriferina, three sp., including the new sub-genus Psioidea, ‡ Trans. N.Z. Inst., vol. xi., p. 537. created by Sir James Hector to embrace a small distinctive form resembling Spiriferina cristata, Schloth.; and Epithyris, Rhynchonella, and Pleurotomaria. The long-winged species of Spiriferina found in these beds at Nugget Point, Hokonui, and Mount Potts have been referred by the Geological Survey to the genus Trigonotreta of Konig. § Hector, “Outline of New Zealand Geology,” 1866, p. 75. (d.) Halobia lommeli Beds.—The principal forms in this horizon are Halobia lommeli, Wissm., Nuculana, Retzia, Epithyris, and Psioidea. The evidence as to the age of the Mount Potts beds is, unfortunately, not very satisfactory; but the presence of Halobia lommeli, of bones supposed to belong to Ichthyosaurus, and the apparent absence of true Spirifera seem to correlate

these beds with the Halobia and Spiriferina beds of Nelson, Nugget Point; and Southland. Sir Julius von Haast collected from the Mount Potts beds in 1861 and in 1872, and in both his report of 1872 * Reps. Geol. Expl., 1872–73, p. 6. and “Geology of Canterbury and Westland,” 1879, p. 272, enumerates the “following genera and probably species” which he says were identified by the late Professor MeCoy: Orthis spinigeia, Spirifera lineata, S. lata, S. oviformis, S. duodecimocostata, Producta, Atrypa, Euomphala, Murchisoma, Orthoceras, and Encrinites, forms which characterize the Carboniferous coalfields of New South. Wales. In 1877 Mr. McKay made a large collection of fossils at Rocky Gully, † Reps. Geol. Expl, 1877–78, p. 94. but, curiously enough, his collection contained, he said, no Orthis, Productus, Atrypa, or Euomphalus; and Sir James Hector, in commenting on the age of the Mount Potts beds, says that Mr. McKay's collection contained no true Spirifera, the spirifer-like forms being all Spiriferina.‡Reps Geol. Expl, 1885, p. xxvi. The fossils collected by myself at Rocky Gully in the early part of this year seemed to agree with the identifications of the Geological Survey rather than with those of Professor McCoy. The saurian remains seem to make the matter no clearer. From the bone bed with the shell beds Von Haast collected a number of saurian bones, some of which he thought resembled “the vertebræ of ganocephalous reptiles of the Carboniferous period; § Haast, “Geology of Canterbury and Westland,” 1879, p. 272. as, for instance, those of Denroerpeton, Hylonomus, &” Among these vertebræ Sir James Hector found one which he thought belonged to an Ichthyosaurus, which he named I. australis.∥ Trans. N.Z. Inst, vol. vi., p. 334. Sir Julius von Haast, however, subsequently challenged this identification, saying, “I fail to see any resemblance between the vertebræ and those of Ichthyosaurus.” ¶ Haast, “Geology of Canterbury and Westland,” 1879, p. 272. Mr. McKay, in his “Report relative to the Collection of Fossils from the Mount Potts Spirifer Beds,” in 1878, says, “Among the additions contained in this new collection were portions of ribs and a vertebral centrum, evidently reptilian in character, which, closely resembling the vertebræ of Ichthyosaurus in general outline, was described as I. australis by Dr. Hector, who at the same time referred the beds to the Triassic period.” ** Reps. Geol. Expl., 1877–78, p. 105. Again in 1890, when discussing the relative age of the

Mount Potts and Clent Hills beds, the Director of the Geological Survey says, “The Mount Potts beds, besides an abundance of Brachiopoda, abound in reptilian remains, the true character of which has not yet been exactly determined, but from the nature of the vertebræ it would appear that these remains have some affinity to the genus Ichthyosaurus.” * Reps. Geol. Expl., 1890–91, p. 147. Then follow some details of the saurian remains found by Mr. McKay in 1877. The Director continues, “The vertebræ are much restricted in length, and the ribs flattened and hatchet-shaped, but not divided at the articulation. Some of the vertebræ are of enormous size, as much as 15 in. in diameter. One fine specimen not yet collected shows a convected series of sixteen vertebræ and thirty-two ribs of the character described. The centra of the vertebræ are 4½ in. in diameter, and the ribs are 3 ft. 6 in. in length.” The above conveys the impression that Sir James Hector entertained some doubts of his previous identification of Ichthyosaurus. In 1879 he so far modifies his opinion as to refer the remains to the genus Eosaurus of Marsh. † Reps. Geol. Expl., 1878–79, p. 11. So far as I could judge from the fragments of bone I saw and collected, I was of the opinion that at least two different genera of saurians were represented in the Rocky Gully bone bed. The Mount Potts plant beds, which are acknowledged by Sir Julius von Haast and Mr. McKay to underlie the shell and bone beds at Rocky Gully, contain a flora which Ettings hausen described as being Triassic and Professor McCoy as Jurassic.‡ Trans. N.Z. Inst., vol. xvii., p. 333. On the other hand, Professor McCoy is credited with having referred the exuvia of the shell beds, which appear to overlie the plant beds, to the Upper Devonian or Lower Carboniferous period. § Loc. cit., p. 333. It should be mentioned, however, that Professor McCoy, in a memorandum to Sir James Hector dated the 12th August, 1885, distinctly denies that he had “heard of saurian bones mingled with Carboniferous shells in Mount Potts beds in New Zealand, and had not determined them.” He admits, however, that he did make determinations of Carboniferous limestone Mollusca and Echinodermata submitted to him by Dr. Von Haast from Fossil Creek, Clyde, Rangitata, Southern Alps, and says that he has “no reason to doubt the age suggested for those beds as above.” ∥ Reps. Geol. Expl., 1885, p. xxi. That a saurian-bone bed does occur at the close of the

Rocky Gully fossiliferous shell beds is beyond all doubt, and, with respect to the diverse opinions and evidence quoted above, we can only conclude that a misunderstanding arose between Sir Julius von Haast and Professor McCoy in the transmission and description of the fossils. Briefly, we find that the plant beds are held to be Triassic or Jurassic; while the overlying marine beds, on the shell determinations of Professor McCoy, are Carboniferous, or, according to the saurian determinations of Sir James Hector, Triassic. (c.) Trigonia Beds.—In most places these are characterized by the presence of Trigonia, Estheria minuta, Edmondia, Ammonites (sp. not determined), a nautiloid shell, Murchisonia, Conularia, and Spirigera (Athyris) wreyi, Zittel. These beds are well developed in Well's Creek, Eighty-Valley and Mount Heslington sections. In the Shaw's Bay eight Valley, and on southern slopes of Mount Heslington, Nelson, in Shaw Bay, and many parts of Hokonui Hills. (d.) Mytilus problematicus and Monotis salinaria Beds.—These beds generally consist of a compact mass of Mytilus or Monotis, to the exclusion of all other forms. The Mytilus beds attain a great thickness in Eighty-eight section the Mytilus bed is composed principally of a confused mass of oyster-shells. At Nelson the cast of a large Gryphaea is occasionally found in the Mytilus beds. Monotis salinaria, Zittel, is present at Mount Heslington, French Pass, Okuku (in Canterbury), and many parts of Hokonuis. It is absent in the Shaw's Bay section. In the Hokonui sections Monotis seems to take the place of the Mytilus, so abundant in the Nelson Trias; and throughout that region the Monotis beds follow the Trigonia beds, as do the Mytilus beds in Nelson and southern Otago. (a.) Clavigera Beds.— These beds everywhere form the close of the Trias. At Shaw's Bay they contain only two distinct shell-forms—namely, an Athyris-like form and Spiriferina. The Athyris is rather an unusual form. It is smooth, bisulcate, has a long slightly curved hinge-line, and possesses the habit of a spirifer rather than that of an Athyris. It corresponds closely with the Clavigera of Hector, which is a subgenus of Athyris created to include this unusual form. * Trans N.Z. Inst., vol. xi., 1878, p. 537. The Spiriferina possesses a straight hinge-line with comblike dentition, and is the sub-genus Rastelligera of Hector. † Loc. cit., p. 537. In Well's Creek, Nelson, the Athyris horizon contains several species of Spiriferina (including Rastelligera, which is

always a characteristic form), Athyris (Clavigera of Hector), Patella, Pentacrinus, fragments of saurian bones, and corals. The Ciavigera beds correspond approximately with the Otapiri series of the Geological Survey. Contemporary Eruptive Rocks. At Nugget Point a dyke of augite porphyrite is intruded between the Spiriferina and Halobia beds. It has already been referred to in that portion of my paper dealing with Nugget Point district. Elsewhere the Shaw's Bay series seems to be free from igneous intrusions, although it should be noted that in north Otago and upper Rangitata they contain massive beds of greywacke. Nomenclature. The name “Shaw's Bay series” was first applied to the Trias of Nelson and Southland by Dr. Lauder Lindsay as far back as 1857. Afterwards, in 1873, Sir James Hector set Lindsay's name aside, substituting the name “Wairoa series,” which at that time embraced the whole of the Trias. * Reps. Geol. Expl., 1876–77, p. v. Subsequently, in 1877, the name “Wairoa series” was restricted to the Upper Trias, as follows:—† Loc. cit., p. v. Trias (a.) Wairoa series. (b.) Te Anau series. In this sense the name was used up till 1878, when Mr. Cox subdivided the Trias into “Upper Wairoa series” and “Lower Wairoa series,” which together now embraced the whole of the Trias. In this new classification, however, the Clavigera (Otapiri) beds were referred to the Lias and the Kaihiku beds to the Permo-carboniferous. ‡ Reps. Geol. Expl., 1877–78, p. 37. In the Geological Survey classification of 1879 the Wairoa series is placed in the Middle Trias, with the Otapiri series above and the Oreti series below, as follows:— § Reps. Geol. Expl., 1878–79, p. 10. “X.—Trials. “(a.) Otapiri series. “(b.) Wairoa series. “(a.) Oreti series.” In the classification of 1880 the Wairoa series appears in the Lower Trias, ∥ Reps. Geol. Expl., 1879–80, p. iv. and in that of 1884 again in the Middle Trias, between the Otapiri and Oreti series. ¶ Reps. Geol. Expl., 1883–84, p. xiv. In the last

classification of the Geological Survey Sir James Hector once again referred it to the Lower Trias. * Reps. Geol. Expl., 1890–91, p. 121. Captain Hutton, in his latest classification of New Zealand formations, includes the whole of the Trias and Permian rocks of the Geological Survey in his Wairoa series, which is thus the equivalent of the Shaw's Bay series of Lindsay. The Wairoa series of Captain Hutton has a very different meaning to the present Wairoa series of the Geological Survey, and therefore to avoid the confusion that might arise from the use of that name I have thought it advisable to discard it for Lindsay's original name, “Shaw's Bay series,” which not only possesses the right of priority, but can claim to take its name from a place where the development of the Trias is as perfect as anywhere in New Zealand, and free from faulting, folding, and tectonic disturbance that might lead to doubtful interpretations of the succession and relations of the different marine horizons. Age The marine beds at the base of the Nugget Point section, embracing the Spiriferina beds (the Productus formation of McKay), have since 1878 been referred by the Geological Survey to the Permian Kaihiku series, but no stratigraphical or palæontological evidence has been adduced to support this view. On the contrary, it has been shown that the Kaihiku series always forms a part of the Triassic succession of strata; and Sir James Hector, when discussing the age of the Kaihiku series, says, “The occurrence of these saurian remains ” [Eosaurus of March], “together with the survival of many Permian forms into the Wairoa, and even the Otapiri, series, and the absence of true spirifers, Productus, and other usual Palæozoic elements of a Permian fauna, would seem to connect the Kaihiku series rather with the Mesozoic than the Palæozoic formations of New Zealand.” † Reps. Geol. Expl., 1878–79, pp. 11 and 12. In Nelson, Otago, and Southland the Spiriferina beds are followed by the Halobia (Daonella) lommeli beds, and in the European alpine Trias this is one of the distinctive forms of the lowest division of the Upper Trias. In 1869 Mojsisovics arranged the Trias of North and South Tyrol Alps so as to place the nodular limestone with Halobla lommeli at the base of his Noric division, but we have it on the authority of Zittel that the stratigraphical researches of Hauer, Richthofen, and Gumbel had shown that this arrangement arose from a series of stratigraphical blunders on the part of Mojsisovics.

So far as I can discover, there is no authentic record of Halobia lommeli from the Lower Trias of Europe, and, if we assume that a homotaxial parallelism existed between the Trias of the Pacific and Europe, it is evident that the Halobia beds of New Zealand must be referred to the lower part of the Middle Trias. I have shown in a paper “On the Jurassic Age of the Maitai Series” that the Clavigera (Athyris) beds in Well's Creek, referred in 1879 by Mr. McKay to the Permian, were the closing beds of the Trias, and occupied the same position as the Clavigera beds at Shaw's Bay and Otapiri. The theory of inversion of the whole Triassic system in Nelson, by means of which Mr. McKay sought to establish the inferior position of the Maitai series, brought many difficulties with it, and rendered it impossible to correlate the Nelson marine horizons with the succession so clearly established in Southland and southern Otago The many attempts to reconcile the inverted succession in Nelson with the normal succession elsewhere were unsatisfactory even to the Geological Survey, judging from the frequent rearrangement of the groups subsequent to 1878; and, naturally, the blending of the two resulted only in confusion. The persistence of. Clavigera, Rasteltigera, and other brachiopods possessing an Upper Palæozoic facies right up to the base of the Jurassic must always impress the New Zealand Trias with Palæozoic affinities, and tend to correlate it with the Alpine Trias of Europe rather than the beds of the great German Triassic basin. Palæontology will doubtless provide general systematic types, which will assist arrangement into smaller divisions, and afford a safe basis for correlations with distant lands. Mataura Series. (Syn., Mataura Series, Lindsay, Hector, and Hutton; Putataka Formation, Hutton; Mataura, Pulataka, Flag Hill, and Bastion Series, Cox; and Catlin's River, Hector.) The Putataka formation of Hutton included this formation and a portion of the Shaw's Bay formation, and also the original Mataura series of Lindsay and Hector. The Mataura series of the latter was, in 1878, subdivided by the Geological Survey into five series, as follows:—* Reps. Geol. Surv., 1877–78, p. vii. (a.) Mataura series. Jurassic (b.) Putataka series. (c.) Flag Hill series. Liassic (a.) Catlin's River series (Hector). (b.)Bastion series.

The Catlin's River series of Hector was interpolated in Cox's classification, being placed sometimes above and sometimes below the Bastion series. In my revision of the classification adopted by Cox in 1878 for the formations in Southland, I showed that the Mataura series and Flag Hill series referred to the same group of beds, * Reps. Geol. Expl., 1886–87, p. 146. and, after an examination of Catlin's district, I am of the opinion that the beds included in the Catlin's River series of Hector will be found to belong to the Mataura series of Cox's classification. The beds included in the Mataura series from the south headland of Shaw Bay to False Island exhibit a thickness of not less than 13,000 ft. In the upper Rangitata their thickness probably exceeds 30,000 ft. In these localities they are devoid of all trace of organic remains, excepting doubtful worm-trails in the Clyde Valley beds, and indistinct plant-remains and three thin marine zones between Sandy Bay and Cannibal Bay. The lithological character of the beds tends to show that they were of estuarine origin. The most characteristic genus in the Cannibal Bay beds is Inoceramus, which occurs in the two highest zones. The beds grouped in this series in Nelson contain an Inoceramus zone about 100 ft. above the Triassic limestone, also a band of fissile slaty shale crowded with supposed wormtrails. The total thickness of the strata exposed on the flanks of Dun Mountain is about 25,000 ft. In the Geological Survey classification the strata included in the Mataura series are subdivided, as we have seen, into five different groups or series. Three, and probably more, of these series overlap each other, and until more definite information is available I think it would be unwise to perpetuate a multiplication of place-names that at present have no meaning, and are useless for purposes of correlation. Relationship to Underlying Strata. In 1886 I showed that the Mataura series and Flag Hill series of Cox referred to the same group of beds. † Loc. cit., p. 146. At the same time I showed that the Mataura plant beds at Mataura Falls followed the Upper Trias granite-conglomerate (2 ft.) and sandstones of Ship Cove conformably, and were underlain by a marine horizon containing Belemnites, Ostrea, Pecten, Pholadomya, &c. In dark-blue slaty shales overlying the shell bed I found what I then identified as Pecopteris grandis, probably the Pecopteris proxima of Ettinghausen; Asplenites, probably Asplenium hochstetteri, Ettings.; two species of Tæniopteris,

probably T. pseudo-simplex and T. lomariopsis, Ettings.; and a species of Taxites. * Reps. Geol. Expl., 1886–87, p. 146. The conformable relation of the Mataura Falls beds to the Trias is clearly seen in the section from Mataura to Gore. The conformable relations of the Mataura series (syn., Matai series) of Sandy Bay and Catlin's beds to the Trias of Nugget Point have already been described. Age of Mataura Series. Baron Von Ettingshausen, who examined a collection of plants from the Mataura Falls beds, stated that “the species bear more or less the facies of those of the Triassic flora,” † Trans. N Z. Inst., vol. xxiii., p. 42. a conclusion quite in harmony with the stratigraphical evidence, which shows that the plant beds follow the Upper Trias. The boundary between the two formations is quite arbitrary, and might easily be fixed so as to place the plant beds in the Trias. On the other hand, the presence of Inoceramus, Beleminites, and Ammonites in the Mataura beds, and Inoceramus in the Maitai and Dun Mountain beds, clearly establishes a relationship to the Jurassic rather than to the Triassic period. No attempt has been made to subdivide the Jurassic and Trias rocks into minute groups of beds. This is a work that can only be successfully accomplished when the fossils have been determined by a palæontologist. Minute subdivision which is not supported by accurate detail and clear definition is certain to be abandoned, and this has been the history of many attempts at subdivision in New Zealand where the facts were not sufficiently ascertained or supported by accurate data. To avoid confusion in the literature of the subject I have discarded many of the old names, some of which were used by different authors with different meanings. The local names adopted by me refer to places where the rocks thus designated are typically developed, and where their relationship to the associated formations is clearly defined. Distribution of Mataura Series. In my paper on the Jurassic age of the Maitai series I showed that the rocks of that series, as typically developed in Nelson, overlaid the Trias conformably, and contained the Secondary genus Inoceramus in abundance. On these grounds I contended that the Maitai rocka could not be Carboniferous, but must be subsequent to the Trias; and I therefore referred them back to the Jurassic period, where they were originally placed by Captain Hutton in 1875. The evidence seems conclusive that the Maitai and Mataura rocks are contem-

porary formations, and must therefore be grouped together under one name, which for reasons already specified must preferably be the place-name Mataura. It is probable that all the rocks hitherto referred to the Carboniferous by the Geological Survey are Jurassic or Juror triassic, with perhaps the exception of the gold-bearing rocks at Reefton, concerning the age of which no definite information is available, and certain aphanitic sandstones, breccias, and slates in the mountains west and north of Lake Wakatipu. The Jurassic Mataura series, now comprising the Maitai series, has a wide distribution in both Islands, and is in most places so intimately associated with the Trias in the formation of the great tectonic features of our mountain-chains that the limits of the Trias may broadly be said to define the distribution of the Jurassic. In Southland the Mataura series occupies a wide tract of country, which extends south of a line drawn from the Mataura Falls to the south head of Roaring Bay, near Nugget Point. Again, passing northward into Canterbury, the enormous thickness and disposition of the Jurassic rocks in the upper Rangitata seems to justify the conclusion, of Von Haast, expressed in 1873, that the great mass of the Canterbury Alps, extending from Mount Cook to the sources of the Rakaia, was composed of rocks of Jurassic age. In some parts of Nelson, and more particularly in the vicinity of Dun Mountain range, the limits of the Jurassic and Triassic rocks are fairly well defined; but elsewhere in Nelson, and in the North Island generally, it is impossible with our present information to define the respective limits of the two systems, excepting in a few isolated patches, as at Kawhia, Raglan, and Waikato Heads. The Mataura series is the important member of the great Hokonui system of Juro-permian age. General Distribution of Hokonui System. This vast assebblage of sedimentary strata comprises a stratigraphical system ranging in time from Permian to Jurassic in one unbroken succession. It is the principal mountain-builder in New Zealand, and in its vastness, in the character of its sediments, age, and tectonic importance, bears a marked resemblance to the celebrated Gondwána system in India. Its nearest equivalent is the great Juro-permian system of New Guinea. In the North Island of New Zealand it forms the Tararua, Rimutaka, Ruahine, and Kaimanawa Mountains, extending in one almost unbroken chain from East Cape to Cook Strait. Outlying patches occur at Coromandel, Tuhua, Kawhia, Waikato Heads, and north of Auckland.

From the French Pass it stretches to Nelson, where it forms all the massive mountains lying between Golden Bay and Marlborough. It occupies the Kaikoura, St. Arnaud, and Spencer Mountains; and thence, spreading southward, forms the greater part of the Southern Alps, as well as the great subsidiary ranges that descend eastward toward the Canterbury Plains. From South Canterbury rocks belonging to-this system cross into North Otago, whence they stretch westward to the main divide. In Southland they occupy the greater part of the land-surface, and sweep round the south end of Otago to Nugget Point and lower Clutha. [Note.—Since the above was written collections of fossils from the Trias in Nelson and Permian at Mount Potts, made by me early in the year, have been sent to Professor G. Boehm of Freiburg University for determination, to whom also were sent at the same time collections of shells from the Jurassic and Triassic rocks at Nugget Point and Catlin's River made by Mr. A. Hamilton and myself—The Author.] Reference To Plates XXVIII.-XXXI. Plate XXVIII.—Geological Sketch-Plan And Section of Coast-Line between Nugget Point And Catlin's River. Plate XXIX.—Section from South Head, Shaw's Bay, toNugget Point. 1. Claystones and sandstones. 2. Athyris sandstones. 3. Coarse sandstones. 4. Oyster and Mytilus bed. 5. Banded claystones and sandstones. 6. Trigonia beds. 7. Granite conglomerate. 8. Sandstones with claystone bands. 9. Breccia conglomerate. 10. Claystones and sandstones. 11. Greywacke. 12. Indurated claystones. 13. Breccia conglomerate. 14. Claystones and sandstones. 15. Halobia claystone. 16. Dyke of porphyrite. 17. Greywacke. 18. Black Spiriferina claystones. 19. Claystone plant beds. A. Track to Roaring Bay. B. Nugget Point Lighthouse. Plate XXX.—Section from Mount Cook Range to Potts River. 1. Sandstones with beds of claystone. Jurassic 2. Sandstones, greywacke, and slaty claystone. 3. Sandstones and slaty shales. Triassic 4. Fossiliferous claystones and conglomerate. 5. Tank Gully plant beds. 2a. Inoceramus, Pentacrinus. Plate XXXI. Upper Figure.— Section on South Side of Tank Gully. (a.) Thin-bedded shales and sandstones. (b.) Grey sandstones. Lower Figure.—Section across Tank Gully. (a.) Sandstones. (b.) Black Carbonaceous shales with fossil fern-leaves.