Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 64, 1935
This text is also available in PDF
(8 MB) Opens in new window

Acid Rocks of the Taupo-Rotorua Volcanic District.

[Read at meeting of Wellington Philosophical Society, October 25, 1933; received by the Editor, November 3, 1933; issued separately March, 1935.]


  • Foreword.

  • Previous descriptions.

  • Occurrence.

  • Physiographic features of the field occurrence of ignimbrites.

  • Petrography of the ignimbrites.

  • Hinuera type.

  • Paeroa type.

  • Arapuni type.

  • Putaruru type.

  • Origin of ignimbrites.

  • Rudimentary crystallisation in ignimbrites.

  • Temperature of the material when deposited.

  • Chemical composition.

  • Changes in the glass of ignimbrites after their deposition.

  • Pectinate structure.

  • Radial or spherulitic structure.

  • Development of pectinate and radial structures in pink wilsonite.

  • Development of structures at Maraetai 5.

  • Development of structures at Te Toki Point.

  • Compaction of ignimbrite at Te Toki Point.

  • Experimental evidence in regard to crystallisation in acid materials.

  • Goranson's experiments.

  • Possible relation between spherulitic rhyolites and ignimbrites.

  • Other areas of rhyolitic rock. Mount Somers, N.Z., Brisbane, Yellow-stone Park, Lake District of England.

  • Use of ignimbrites as building stones.

  • Suggested classification of ignimbrites.

  • Summary.

  • Explanation of illustrations.

  • Table of references.


In order to avoid difficulty and confusion in nomenclature in speaking of the rocks that are dealt with in this paper, this preliminary statement is made. These rocks are regarded as formed by an eruptive process similar in its nature to that which was described by Fenner as acting at Katmai in Alaska. In other words, they are thought to have been deposited from immense clouds or showers of intensely heated but generally minute fragments of volcanic magma. The temperature of these fragments is thought to have been so high that they were viscous and adhered together after they reached the ground. Since this type of rock mass has a wide occurrence in New Zealand, but has not been recognised before except by Fenner, who calls it “indurated sand flow rock,” it appears that a new term is required for its designation. Rocks which are thought to have been formed in this manner are here called Ignimbrites.

The following quotations will show that these ignimbrites are different from the vitric tuffs described by Pirsson (Pirsson, 1916, p. 894): “The expansion of gas and rupturing began in a liquid medium and falls as a rigid glass. We do not know, of course, the

– 324 –

exact march of events between these points; but it is clear that the magma ruptures into separated masses of different sizes as the volume of mingled gases and molten glass rush out of the conduit. The separated masses are themselves swelling and flying apart into smaller ones as they ascend, and this continues until the stiffening of the glass and the lessening of the expansion of the gas through cooling bring the process to an end.”

It is clear from this extract that Pirsson, who is the originator of the term “vitric tuff,” did not in this vivid account of the eruption of tufaceous matter visualise the idea of the particles of melted glass, suspended in the gases that had emanated from the volcanic matter, and retaining a temperature so high that they remained viscous even after they fell. In other words, his term “vitric tuff” does not comprise ignimbrites. It would seem that the possibility of particles of volcanic matter retaining their viscosity in minute sizes and in immense volume until they had actually fallen was not imagined until after Fenner's account of the Katmai eruption.

I wish to express my deep thanks to Mr E. T. Seelye, of the Dominion Laboratory, who with the permission of the Dominion Analyst, Mr W. Donovan, made a complete analysis of ignimbrite (tridymite rhyolite) from Waihi. He also was good enough to make many experiments on the effect of submitting ignimbrite material to temperatures as high as 1050 degrees Centigrade.

Previous Descriptions.

The earliest explorers of the North Island of New Zealand at once recognised the wide occurrence of volcanic rocks. Dieffenbach said: “We passed cliffs of pumice stone… In some places there was a greyish lava of a striped and variegated appearance resembling jasper.” (Dieffenbach, 1843, vol. 1, p. 371.)

Taylor remarked: “There is much obsidian and pumice near Taupo.” (Taylor, 1855, p. 244.)

Hochstetter made many observations and collected a number of specimens in this region. In particular, he published a sketch of Horohoro mountain near Rotorua (von Hochstetter, 1864, p. 106). Horohoro has a special interest as this hill illustrates better than any other in the district the flat-surfaced, steep-sided forms which in greater or less degree are characteristic of the elevations composed of rocks that are the main subject of this paper. In the map on page 20 the volcanic rocks are classified by Hochstetter as rhyolite lava. This classification is based on the work of Zirkel, who was entrusted with the examination of the rock specimens that were collected by the expedition. He distinguished between glassy, spherulitic, and banded types of rhyolite as well as tuffs.

Hutton, in 1889, in a paper on the eruptive rocks of New Zealand (Hutton, 1889, p. 115 et seq.) described a number of the rocks as enstatite rhyolite and enstatite trachyte.

– 325 –

Rutley, in the Quarterly Journal of the Geological Society of London (Rutley, pp. 449–469, and 1901, pp. 493–510) gave petrological descriptions of a number of these rocks, mainly from the Waihi district. He describes at some length the various types of spherulitic structures that are found in them. In the latter paper (Rutley, 1901, pp. 496–7) he classifies two of the rocks as tufaceous rhyolite lava. He remarks on the probable similiarity in origin to those of the Dufton Pike (Rutley, p. 509).

Sollas has described a number of these rocks in detail. It is clear, however, from p. 23 that he has difficulty in deciding whether some of the rocks are flow lavas or tuffs. (Sollas, 1905, p. 23.)

More recently, Morgan, and later Henderson and Bartrum, and afterwards Morgan again, have published descriptions in the Bulletins of the New Zealand Geological Survey (Morgan, 1912; Henderson and Bartrum, 1913; Morgan, 1926). A good list of references is given by Henderson and Bartrum in their publication, p. 70. They give a useful summary of the statements that have been made by various authors in regard to certain of the rocks that are found near Waihi.

Within recent years these acid rocks have been mentioned in several publications (Henderson and Ongley, 1923, p. 56): “Rhyolitic tuffs in places hundreds of feet in thickness cover large areas.” The source “undoubtedly lay to the east.” The rocks consist of “very fine particles of glass containing glass flakes in abundance as well as crystals and fragments of sanidine, albite, and oligoclase. Quartz crystals are rare.” (L. I. Grange, 1927, p. 39): “Well cemented rhyolitic tuff occurs in the Ongaruhe Valley”; and in the Annual Report of the Geological Survey (Grange, 1927–28, pp. 8, 9): “The rhyolitic tuff which is the predominant rock … since the surface of this deposit except where disturbed by later movements is almost flat, it is thought that the fragmental material was laid down in water.” J. H. Williamson speaks of cemented and well-jointed rhyolite tuff. (Williamson, 1930, p. 9.)

The author, in 1929 (Marshall, 1929, p. 31) wrote: “A vitric tuff that occurs widely in the Waikato district might be used with advantage for building purposes; it is relatively fine grained, and is mainly formed of minute particles of glassy rhyolite which were ejected by the expansion of steam in melted volcanic rock and fell back to the ground while still hot, and united together sufficiently to form a moderately compact rock. With the small fragments of glass there are large numbers of quartz and felspar crystals more or less rounded in form.”

In 1931 the author read a paper before the Wellington Philosophical Society giving descriptions of these rocks and of the supposed eruptive actions to which they were due. These were regarded as probably similar to those of the Valley of Ten Thousand Smokes at Katmai in Alaska. This eruption was described by C. N. Fenner, who mentions an “indurated sand flow rock” which seems to be of the same nature as some of the rocks that are described here. The most characteristic feature of this eruption was afterwards called the “nuée ardente Katmaienne” by Lacroix.

– 326 –

This address, being then incomplete, was not published; but a short abstract of part of it was printed. (Marshall, 1932, p. 198.) The present paper is in effect the address of 1931, but many additions and some alterations have been embodied in it.

In the references that have been cited above no mention is made of flow structure. So far as these rocks are concerned, the flow structure is here interpreted as due to the pressure of the overlying tuff directly after its deposition. The structures called in this paper radial and pectinate are not mentioned, though the latter is particularly prominent at Ongaruhe and elsewhere in that neighbourhood. The rock is described by Grange as cemented, and the reference to deposition beneath water seems to imply that cementation by material dissolved in water is regarded as operative. There is no reference to welding.

H. T. Ferrar says: “The rhyolitic material like that ejected at Katmai in Alaska shows no distinct craters and was probably erupted through a number of vents, not one of which has at present been located… The tuff usually contains angular fragments of obsidian and pumice, which being hot and sticky when erupted, adhered to form a massive rock.” (Ferrar, 1930, p. 3.)


When the size of the North Island of New Zealand is considered, it can be said that these acid rocks or ignimbrites are found over a wide area. On the Wanganui River near Taumarunui and on the eastern side of Lake Taupo they reach their southern limit. The northern extremes are at Mercury Bay in the Coromandel Peninsula and at Mercury Islands, and in the Hinuera Valley in the Middle Waikato region. In the west they reach to Te Awamutu, and even to Puketiti, 15 miles west of Te Kuiti (specimen kindly given to me by Mr Ferrar). In the east they are found as far as Tauranga, the lower Rangitaiki, and almost to Tarawera in the Mohaka Valley. Within this wide area, however, they do not constitute the striking volcanic cones which are the most conspicuous feature of this region. The several large mountains Ruapehu, Ngauruhoe, Tonga-riro, Pihanga, Maungatautari, Tauhara, and Edgecumbe are none of them composed of acid volcanic rocks, but of andesites of various types. The area over which scattered outcrops of acid rocks occur amounts roughly to 10,000 square miles. This is the “rhyolite plateau” of New Zealand geology.

Physiographically it may be said that within this district the outcrops of the ignimbrite rocks can generally be recognised in the field, even before they are closely approached. Usually they have a surface that is approximately level; while the margin of the areas are markedly precipitous. Here massive outcrops of rock are frequent. These ignimbrites have in nearly every case a coarse prismatic jointing, and, as their nature, though not extremely hard, is distinctly tough and they are resistant to atmospheric action, the huge columns stand out in the most striking manner. Little broken rock is shed from them, and the outcrop thus acquires a noticeable and striking form. Plate 69, figs. 1, 2, 3.

– 327 –
Picture icon

Marshall.—Acid Rocks Taupo-Rotorua Volcanc District.

– 328 –

This condition is most arresting on the margins of the old Waikato Valley at Hinuera, and is well seen in the present gorge of this river at Arapuni. An illustration of this most characteristic type of outcrop is given by. Morgan. (Morgan, 1913, Pl. X.) A striking illustration in Reise der Novara, p. 80, of the columns at Motuara Island, Mercury Bay, and entitled “Saulenformige Trachyt” is, however, misleading. The actual locality is at the north end of Great Mercury Island. The rock is hypersthene basalt. Everywhere within the area of these volcanic acid rocks there has been some activity since the ignimbrites themselves were deposited. One phase of this activity has been the covering of the ignimbrites with a superficial coating of pumice of varying but of ten of considerable thickness. This material, too, covers all the country in the lower land, between the hills that are formed of ignimbrite rocks. Another phase of the later volcanic action has been the formation of the great cones which attain their culminating point in Mount Ruapehu, which has an altitude of almost 10,000 feet. These cones are with few exceptions formed of hypersthene andesite.

In contrast with these andesitic cones it is a fact that there are no cones of acid rock in spite of the wide occurrence of this type of material. The highest point at which this material is found is perhaps Mount Tarawera at 3600 feet, but no full petrographic description has yet been given of its occurrence in that locality. It is, however, certain that acid rocks were not concerned in the great eruption of that mountain in 1886, except as ejected blocks of previously consolidated material. Tarawera has not a true conical shape, and there are no lava flows on its flanks. The rock emitted during the eruption of 1883 was basic andesite. Wherever these ignimbrites have been intersected by stream courses steep narrow gorges have been developed. This is clearly a result of the pronounced prismatic jointing in the absence of horizontal division planes; but it is partly due also to the remarkable chemical resistance of the rock to the action of the atmosphere. Where a larger river intersects an outcrop of these rocks massive blocks are from time to time carried away and are transported far down the river valley. Large blocks of this nature are conspicuous objects in the Wanganui River as much as forty miles below the outcrop of any acid material. This again emphasises the unusual toughness of the rock, in spite of its softness, and its comparative freedom from separation planes and irregular joints.

The thickness of the ignimbrites varies considerably. It is unusual, however, to find any single outcrop with a greater thickness than 200 feet; though this is certainly exceeded at Ongaruhe, and at Motutere on the east side of Lake Taupo, and on the coast south of Mercury Bay. In the Waikato Valley 30 miles south of Arapuni it is at the least 500 feet thick. Plate 70, fig. 1. On the other hand, in the Hinuera Valley the thickness is 60 feet, and in general it appears that formations of the ignimbrites are not often thicker than 100 feet.

– 329 –

The eruption of ignimbrites probably extended over a considerable period of later Tertiary and Post Tertiary time at irregular intervals.

Physiographic Features of the Field Occurrence of the Ignimbrites in the North Island Area.

Before considering the detailed mineralogical and microscopic characters of the ignimbrites, it is as well to mention some physiographic and field facts with regard to the disposition and occurrence of these rocks in general. Remarks that may be made under this head must not be considered as applicable in all respects to every outcrop of these rocks.

(1) It has already been stated that the upper surface in many of the localities where the rocks occur approaches the horizontal plane. The lower surface, too, is generally regular and shows little or no roughness such as that which is generally seen on the lower surface of an ordinary lava flow.

(2) These ignimbrites appear to be never definitely related to any volcanic cone. Although they occur over such a wide area, there is no culminating point which can be regarded as the point from which they were ejected. Actually, as previously stated, some of the thickest of the deposits that the author has seen were at the south end of the area at Ongaruhe, at the north end at Mercury Bay, and on the east at Motutere on the east side of Lake Taupo. If there is any distinction to be made, it would appear that the greatest thickness of deposits is to be found in the marginal portions of the area, though at present no importance can be assigned to this distribution, even if it is rightly stated.

(3) No scoria was found anywhere on the upper surface of these rocks. Molten material of this highly siliceous nature is notably viscous; and in all cases it is probable that it contains large quantities of water vapour. These are well known to be the conditions that particularly promote the development of scoriaceous matter. The very general occurrence of pumiceous matter of a highly vesicular nature emphasises this in New Zealand. The entire absence of scoria on these rocks is therefore a notable fact. The covering of pumice that is generally found on them has certainly not been developed from the ignimbrites; but it has probably been projected over large distances by violent explosive eruptions of the Krakatoan type. The fragments of pumice never show any attachment to the surface of the ignimbrite on which they rest. They are more or less rounded as though from attrition, and have no resemblance to the scoriaceous development of a lava surface. It is abundantly clear that the pumice has fallen in showers onto the surface of the ignimbrites after their formation and by independent eruptive action.

(4) The lower surface of these acid rocks wherever it has been observed, notably at Arapuni, has been found to differ from the upper parts but little, in any respect except coherence; for the very base is formed of incoherent matter which actually consists of minute

– 330 –

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

particles of volcanic glass among which there are some crystals of felspar and quartz of relatively large size. In this part of the rock, too, there is no development of scoriaceous matter. The grading of this fine matter which forms the base of the ignimbrite at Arapuni is as follows:—1/30″–1/40″ 6.60, 1/40″–1/50″ 3.90, 1/50″–1/60″ 2.33, 1/60″–1/80″ 3.77, 1/80″–1/100″ 4.73, 1/100″–1/200″ 18.40, <1/200″ 59.60%. Examination of microscopic preparations shows that this is practically the grade of the particles throughout the rock also, even where it is solid and compact.

(5) The regular jointing and consequent display of columnar structure demands exceptional conditions for its development in flow rocks. It is necessary that there should be a condition of complete rest and stillness whilst cooling and solidification are in progress. The attainment of such conditions in a lava flow implies great fluidity and relatively rapid flow before a necessary containing basin is reached. It has not yet been shown that any of the occurrences of these ignimbrites are situated in such pre-existing basins. Their actual positions perhaps point to the reverse. At any rate, great fluidity for some time after the emission of the molten material would be required in order to account for the present position of the rock masses and for their habitual wide extension. Yet in the case of these rocks chemical composition is such as to preclude the possibility of such great fluidity as is obviously required: for the percentage of silica exceeds 70.

When all of these facts of field occurrence and of detailed structure are considered, it appears certain that these extensive formations of acid rocks were never real lava flows. It is therefore misleading to speak of any of them as rhyolites. Before expressing any other opinion as to their origin, however, it is necessary to describe in some detail the main facts of their lithological and petrological structure.

Lithological Nature of the Ignimberites.

A very great variety of lithological features is to be seen even on the most cursory examination of a series of hand specimens of these rocks. The type that is generally called wilsonite is perhaps the most arresting. In this rock there are lenticules or ovoid patches of dark coloured material, usually glass, often three or four inches long, which are embedded in a matrix of stony appearance. This structure appears to be typically eutaxitic, as usually defined, though this term was first applied by von Fritsch and Reiss to nephelinitoid phonolites in 1868. Yet the whole rock sometimes has a vitreous lustre on a fractured surface. The transition from this extreme form to the completely fine and even grained lithoidal or stony rock is probably a gradual one; though it is true that no complete transition has yet been observed in any one locality. This lithoid or more finely textured type of rock has generally been called rhyolite, though Henderson, Grange, and others have recently written of occurrences at Ongaruhe and elsewhere as tuffs. This type of rock has usually a white to grey colour and has always rather a harsh feel. It has a very general occurrence throughout this area.

– 331 –

Even in hand specimens it can always be seen that all these rocks contain a large number of glassy crystals of felspar and some of quartz. Usually the rock breaks round the crystals, not across them. The rock is, however, relatively soft, though it does not fracture with ease, as it is quite tough. Even with a lens the stony base cannot be resolved into any particular components. At or near the under surface of this stony lithoid rock the material may have a darker colour; and as at Arapuni it may have a vitreous lustre; but for the most part it is notably granular, and may in its upper portion become almost earthy. A sample of the surface material at Arapuni was broken down with a rubber pestle under water and was found to have the following grading:—

> 1/40″ 16.34
1/40″ 1/50″ 4.17
1/50″ 1/60″ 3.00
1/60″ 1/80″ 2.22
1/80″ 1/100″ 3.78
1/100″ 1/200″ 12 65
< 1/200 54.37

Although thus very generally similar over a wide area, a difference in colour, or compactness, or texture will often distinguish the samples from different localities. Thus specimens from Hinuera will often contain small fragments of real rhyolite rather than glass, somewhat vesicular indeed, but far more compact than pumice. Specimens of a fine-textured ignimbrite from Ongaruhe have a yellow or even an orange tint. Those from Arapuni are almost white. The wilsonite from Putaruru is pink; while that from Owharoa is pale grey, and samples from Waikino are light brown in colour. All of the great outcrop at Motutere on the eastern shore of Lake Taupo is a brownish grey in tint.

The actual extent of any one of these outcrops is not easy to discover. The heavy superficial covering of pumice obscures them for wide distances over the surface of the country, and it is only where streams have cut valleys through the pumice beds above them that the underlying ignimbrite can be seen. It can hardly be asserted that the stream valleys really delineate the boundaries of a single area of these acid rocks, especially in the absence of detailed petrographic work, which is at present wanting. The isolated and steepsided hill Horohoro, the form of which has generally been ascribed to faulting, is a striking feature of the country on the south-west of Lake Rotorua. Hochstetter originally took this view of its origin, (von Hochstetter, 1864, p. 196.) It is, however, possible that Horohoro is the scene of a separate ejection of ignimbrite; though such a suggestion is at this time a pure hazard, for we have no detailed description of the hill, nor of the rock of which it is formed, at present.

The general impression gained in the area over which these acid rocks occur is that any one lithological type has not a wide lateral extension. The margin of an ignimbrite wherever it is

– 332 –

exposed has, with a few exceptions, the form of steep cliffs; but it never has the heavy scoriaceous slopes that mark the termination of a true lava flow. While it is true that these exposurels arte always on the margin of areas of erosion, the impression is nevertheless conveyed that these rock masses terminate rather abruptly; though originally they may probably have passed laterally into softer and more incoherent matter which was an easy prey to the agents of erosion and was soon removed.

Petrography of the Ignimbrites.

Wilsonite type. (Fig. 2, plate 66; figs. 1 and 2, plate 67.)

The nature and origin of the wilsonite type of ignimbrite have been much discussed. The specimens that were first described by Rutley were called by him pumice tuffs. (Rutley, 1899, p. 457, also 1900, p. 494.) Sollas, in the volumes on the Hauraki Goldfields (Sollas, 1905, vol. 1, plate facing p. 66, pp. 123, 124) remarks: “I am quite prepared to accept it as having once been in a state of flow.” But in a footnote: “Some of the specimens are formed of material that has fallen through the air.” Further (Sollas, 1905, vol. 2, p. 46): “If, as the field evidence seems to suggest … this is a flow rock, then the form of the shreds of glass can no longer be regarded as characteristic of a tuff.” Again (Sollas, 1905, vol. 2, p. 46): “The microscope offers no disproof of the theory that this rock was once flowing.” He also states that the rock had undergone an extreme amount of internal brecciation. Bell and Fraser (1912, p. 48): “The rock is evidently a flow with a considerable amount of internal brecciation.” Henderson and Bartrum (Henderson and Bartrum, 1913, p. 72, analysis p. 73) review previous opinions and state that they incline to the view that the rock is a tuff and was deposited very close to the point of ejection; but that its components were in the viscous state. Arguments for and against this idea are given in full. This opinion seems to have been accepted by Morgan (Morgan, 1924, p. 65), though he had previously stated that the rock was a lava, not a tuff (Morgan, 1911, p. 273). Morgan sometimes refers to the rock as displaying brecciated flow structure. It seems that the rock was first mentioned by Cox (Cox, 1882, p. 20): “A rhyolitic rock containing a mixture of pumice and obsidian.” A similar description was given by Hutton (Hutton, 1889, p. 116).

All of those that have studied typical wilsonite from Waikino or Owharoa are agreed that it is composed of relatively large fragments of glass embedded in a matrix of stony appearance in which many clear and colourless crystals of felspar and quartz can be distinguished. Actually the stony matter is found to consist mainly of small shreds of glass which may show a parallel arrangement—the so-called flow structure. Most of these small fragments of glass are colourless, though some of them have a colourless border around a brown core in which the colour seems to be due to numerous minute globules or granules too small to identify, though they may be gaseous. The felspar crystals are generally fractured or merely broken pieces, and are sometimes rounded. Oligoclase is the most usual species.

– 333 –

though there is a little sanidine, and some andesine. The grains of quartz are rounded and show no crystal faces. Small crystals of hypersthene are common, and they show crystal faces. Sometimes there is a little biotite, or greenish brown hornblende. The large fragments of glass in wilsonite are found when studied in microscopic preparations to be traversed by many capillary air spaces.

Hinuera type. (Plate 69, fig. 3; plate 62, fig. 2.)

In this type of the rock the field occurrence has massive columns which attain as much as eight feet in diameter and sometimes they are forty feet in height. The rock in the typical locality at Hinuera has a buff colour, but fragments of conspicuous white rhyolite that is too compact to call pumice are noticeable and frequent. These may be three or four inches in diameter, but they are not lenticular like the glass fragments in the wilsonite, and they are stony rather than glassy. In microscopic preparations the fragments of rhyolite are seen to consist of colourless glass traversed by numerous thin capillary pores. They are embedded in material of a fine texture in which numerous broken crystals of felspar and a few of quartz can be seen. As in many of the other rocks there are occasional narrow crystals of hypersthene, and rarely of hornblende. The matrix in which they are embedded is of a pale yellow tint. When studied in microscopic preparations the matrix is found to consist of irregular but largely linear glass shreds lightly welded together. Here again they are often brown in the centre, though this feature is far more noticeable in the larger fragments than in the smaller.

It is suspected that the colour is due to densely crowded minute gas globules, though the small size prevented them from being clearly resolved under the microscope except in a few instances. (Plate 62, fig. 2.) In this rock there is no flattening of the glass fragments, and no flow structure has been observed, but only a small portion of the outcrop has been examined at present. There appears to be a strong tendency for the brown fragments of glass to lose their colour. In the upper part of this formation in the Hinuera valley, that is at any point more than ten feet above the base, the particles have all lost their vitreous nature and show a rudimentary axiolitic structure. Even the extent of the Hinuera deposit is at present unknown. Certainly it occurs over a distance of at least three miles on both sides of the Hinuera valley; but how far it extends backward from the side of the valley has not yet been determined. The universal covering of pumice, here as elsewhere, offers very considerable difficulty in the details of field work.

This Hinuera type of ignimbrite occurs at Mercury Bay and in the valley of the Mangakino River, a tributary of the Waikato, as well as in the Hinuera valley and many other localities.

Paeroa type. (Plate 62, fig. 1.)

This rock is of a brownish colour and is less coherent than the other types—in hand specimens it seems almost earthy. It contains fragments of a great variety of volcanic rocks, sometimes as much as four inches in diameter. However, the earthy appearance is

– 334 –

actually misleading; for even when boiled in sulphuric and hydrochloric acids there is no disintegration, and the rock actually gains in hardness and solidity from the treatment. The lapilli embedded in the rock consist of different varieties of rhyolite, and andesite, which are not glassy, but are already well crystallised. Sometimes there is a little alteration, and some of the ferro-magnesian minerals may be partly chloritised; in this respect resembling the rock from the Okaro crater as described by Hutton. (Hutton, 1889, p. 137.) Despite this structure, the rock in the field has a well-marked coarse columnar development.

When the rock is studied under the microscope, the usual crystals of felspar and quartz are found, with their usual colourless transparency. As in other localities, there is some hypersthene in small columnar crystals and a little hornblende, and occasionally pale green augite; but in the fragments these may be converted into chloritic and serpentinous substances. In the andesite there are in addition a few grains of pale green augite such as Hutton recorded from Atiamuri. The fine portion again consists of shreds of glass, which actually in this case have undergone some amount of crystallisation since they were deposited. On their inner margin there is a fine felt of minute felspar needles which at times extend a third of the distance across the shred. The glass shreds do not bend round the crystals, as can occasionally be seen at Hinuera; but the solidity of the rock shows that they are welded together to some extent, and the partial crystallisation shows that their temperature was high when they were deposited.

This type is typically shown in the Paeroa Range on the west side of the Rotorua–Taupo road, between Waiotapu and the Waikato River.

Arapuni type. (Plate 63, fig. 1.)

This is a typical fine textured rock of the ignimbrite group. It is fine and even grained; but has a harsh surface. This type is always light coloured, grey to creamy, or even pale yellow. Conspicuous clear glassy crystals of felspar and quartz can always be distinguished even in hand specimens.

Samples of this type of rock collected in the underground workings of the Waihi mine were called by Morgan “tridymite rhyolite” because of the wide occurrence of tridymite in that locality, though it is revealed by microscopic examination only. (Morgan, 1911, p. 273.) Such study also shows that the felspar is generally triclinic, oligoclase being the most usual species. Sanidine occurs rather rarely, and all of the felspar crystals are usually incomplete with jagged margins. Often they are mere fragments. Quartz is far less frequent than the felspars. Hypersthene, hornblende, and biotite are sometimes found, and occasionally small crystals of magnetite. The tridymite occurs very generally, sometimes as small rosettes in cavities, but more frequently along the centre line of axiolites, but also sporadically among the finer elements of the rock.

– 335 –

The mineral tridymite appears to have been first recognised by Rutley in these fine grained rocks. He remarks of them in general: “Reheating of already solidified lavas has been, as we have frequently had occasion to remark, a by no means uncommon feature in the history of these rocks.” He also speaks of “a globulitic devitrification” in connection with this. (Rutley, 1899, p. 465.)

Sollas disposed of this idea of refusion so far as certain features of the more coarsely spherulitic rocks are concerned; but substituted for it the explanation that a certain peculiar decomposition had taken place. (Sollas, 1905, p. 121, vol. 1.) While Rutley seemed to regard the fine grained rocks as lava flows, with two exceptions which he called “tufaceous rhyolites.” Sollas treated them all as lavas; and this opinion was apparently held by Morgan, Henderson, and Bartrum also. (Morgan, 1912, 1913, 1924.) In the last of these bulletins Morgan gave a peculiar explanation of an unusual feature of these “tridymite rhyolites.” All observers had noticed the occurrence of triclinic felspars almost to the exclusion of monoclinic forms, and Morgan had previously recorded the presence of carbon in some of these types. In the publication quoted he refers to the crystals as “clearly showered on the lava, or surface derived, like the fragments of carbonised wood previously mentioned.” (Morgan, 1911, p. 67.) It is merely remarked at the moment that the common occurrence of plagioclase, and the far less common occurrence of quartz, suggests that the magma from which the ignimbrites were derived was of the nature of a highly acid dacite.

This type has its most typical development at Arapuni. It is also found at Ngutuwera and widely near Waihi.

In all of the samples of this fine-grained type or arapunite that were examined by Rutley, Sollas, Morgan, Henderson, and Bartrum, it seems that fragments of glass were practically absent. In other parts of the district, however, it is found that glass shreds and curved glass fragments of minute size compose the whole of the fine grained matrix of the rock; though in hand specimens the rocks can hardly be distinguished from the typical “tridymite rhyolites” of Waihi. Though actual shreds and fragments of glass are not seen in these “tridymite rhyolites,” there can be no question that the ill-defined axiolites, which constitute their main material, were originally glass shreds; though in them incipient crystallisation was developed at the time of their deposition.

These minute particles of glass at Arapuni always have their greatest dimensions in the horizontal plane. They are pale brown in colour, and in section are seen to bend round the corners of the felspar crystals, thus giving the typical appearance of the so-called flow structure. (Plate 63, fig. 1.)

Fayalite occurs in types of these acid rocks at Ohena Island, Mercury Bay, at Pohaturoa, Atiamuri, and is perhaps altered to magnetite in many other localities.

– 336 –

The Origin of Ignimbrites.

Before entering upon any further discussion of the composition and structure of these rocks in different outcrops, their origin may be considered. It is as well to take the simplest case first, such as the rock which occurs at Arapuni. Here the lowest part of the rock, above the basal layer of fine sand, consists of the usual crystals of felspar and quartz with a little hypersthene which are embedded in a matrix of fine shreds of glass, the particles of which have that form and arrangement which are commonly considered as characteristic of flow structure, and have been so interpreted in a large number of instances. In many cases the phrase “corrugated flow structure” has been applied to them. The very nature of the rock, however, precludes the possibility of mass flow, for it consists of minute but quite distinct individual fragments of glass which are lightly fused together. It comes, however, within the range of similar rocks that have been described by Iddings. (Iddings, 1899, part 2, p. 404, and pl. 50, A, B, and C.) This kind of rock is stated by him to be formed of collapsed pumice. It is here considered, however, that the uniform nature of the rock over wide distances, the absence of all remnant of pumiceous condition, the absence of a sufficient grade for flow before the assumption of the pumiceous state, the occasional presence of carbon, the obvious occurrence of minute separate fragments of glass, and the absence of any other evidence of collapse of pumice in rock masses throughout the district, individually and collectively, preclude the application of such an explanation in this case.

It is thought that the observations which have recently been made by Fenner in regard to the burning sand flow in the Valley of the Ten Thousand Smokes afford a satisfactory explanation of the deposit of these ignimbrites. A similar explanation has already been made by the author in a pamphlet on the building stones of New Zealand. (Marshall, 1929, p. 31.) This was before Fenner's statements in explanation of the “indurated sand flow rock” had been noticed. (Fenner, 1925, p. 198, and footnote.) At the Katmai eruption, as described by Fenner, the material of the sand flow was of such a high temperature when ejected that the fragments of glass of which it was composed, in places, welded together and thus formed the “indurated sand flow rock.” The actual mechanical state of this sand flow is compared by Fenner to a cloud of heated basic carbonate of magnesia in a finely powdered state. Lacroix, however, suggests that the condition of the erupted material was more properly to be compared with that of milk boiling over. (Lacroix, 1930, p. 460.)

That a mixture of incandescent powder suspended in a heated gas behaves as a fluid is now generally admitted. Lacroix states that the nuée ardente from Mount Pelée flowed towards the sea with the velocity of 159 metres per second, equivalent to almost 360 miles per hour.

It is clear that a velocity much less than this would enable a “nuée ardente” to rise far on the flanks of an opposing acclivity and to cover an area of deposition different in extent and configuration from any that could be associated with an ordinary flow

– 337 –

of a liquid. Though it would seem that none of the ignimbrites in New Zealand flowed down such a declivity as partly caused the high velocity of the nuée ardente at Mount Pelée (the nuée ardente Peléenne of Lacroix) it yet appears to be a fact that the material of the “nuées ardentes” from which the New Zealand ignimbrite rocks were formed was of far finer grain than that at Katmai. This in itself may be supposed to imply a more rapid evolution of gas and consequently a higher initial velocity than that of the material of the Valley of Ten Thousand Smokes. I am indebted to Dr Fenner for a sample of his “indurated sand flow rock” which has enabled me to make an actual comparison which reveals a close resemblance to the New Zealand ignimbrites. At Arapuni there is a layer of fine sand three inches thick beneath the ignimbrite. This is regarded as the portion of the fiery cloud that was cooled rapidly by contact with the atmosphere and the ground. Immediately above this sandy layer the rock has a glassy lustre, and microscopical examination shows that the glassy particles are lying with their longer axes horizontal and were compressed by the weight of the overlying material and welded together into a moderately compact glassy rock. The particles of glass were bent round the angles of the solid crystals in such a manner as to give the effect of actual flow movement. It is thought that this may be the origin of the flow structure that has been described in many rhyolite areas. It is also suggested that this explains the features that were ascribed by Iddings to the collapse of pumiceous structure in the Yellowstone region. When the glassy fragments are of conspicuously unequal size as in the wilsonite the resultant rock would give the impression of “brecciated flow structure,” a phrase which with “corrugated flow structure” is so widely employed in descriptions of rhyolitic rock. It has always seemed to the writer a matter of extreme difficulty to comprehend the physical conditions that are described as the cause of the development of this structure. A brief description of this structure and its origin is given by Holmes in these words: “A term describing lavas in which fragments of partly solidified magma produced by explosion or flowage have become welded together or cemented by the still fluid parts of the same magma.” (Holmes, 1920, p. 100.)

Even if such a structure were developed in the manner described, it appears to the writer that a continuation of the flow would quickly change it into a highly streaky form of rock or would entirely destroy the structure described. It is thought that the relatively large particles of glass, incandescent at the time of deposition, which occur in large numbers in the wilsonite, imply rather less gaseous emanation and expansive force than is evidenced by the structure of the finergrained ignimbrites. However, the super-hurricane velocity of nuées ardentes as described by Lacroix would allow of even these large particles being carried and distributed over wide distances. Whether a wilsonite in its lateral portions gradates into an arapuni type or other finer material on the margins of its area has not yet been determined. Unfortunately, the difficulties of field work have

– 338 –

up to the present prevented the necessary investigations for solving this question from being made. It is, however, the author's opinion that such variation does not occur.

Lacroix has defined thenature of the material of a nuée ardente and its physical condition in the following words (Lacroix, 1904, p. 203):—

Les nuées ardentes sont constituées par un mélange intime, une sorte d'émulsion des matériaux solides en suspension dans de la vapeur d'eau et dans la gaz, portés les uns at les autres á haute temperature.

P. 350: Chacune des parties ou des particules solides qui les constituent rayonne de la chalure et doit ětre entourée par uno atmosphère de gaz et de vapeurs extrèmement comprimée au début mais se dilatant rapidement; c'est cette atmosphère qui, empěchant les particules solides de se toucher, maintient l'ensemble dans un état de mobilité permettant de couler sur les pentes presque a la facon d'un liquide.

P. 358: II est donc probable que l'explosion est due a la brusque détente de la vapeur d'eau contenue dans le magma … on doit admettre que la tension des gaz inclus dans celui-ci augmente a mesure que sa consolidation s'effectue sans nfluence du refroidissement.

Rudimentary Crystallisation in Some Ignimbrites.

It has already been stated that the lowest part of the ignimbrite at Arapuni is a fine sand consisting of minute glass shreds, and that overlying this there are eight or ten feet of a compact rock that has a vitreous lustre. The upper portion of the fifty feet of thickness of this Arapuni type changes gradually into a rock that in hand specimens has a more stony appearance, but also a more pulverulent structure. With microscopic examination it is found that in this portion the small glass shreds are less transparent because of a rudimentary crystallisation which has been set up. While it has not been found possible to identify the actual minerals that have been formed by this action, comparison with samples from other localities in this district in which the structure is more fully developed indicates that felspar and tridymite are present in an association such as that which will be subsequently described as occurring at Ongaruhe and which is typical of a number of the rocks of this district. The gradual change from the more compact and distinctly glassy condition of the shreds of the lower part of the outcrop at Arapuni to the less compact and slightly crystallised condition of those of the upper part is probably due to difference of temperature originally, rate of cooling, evolution of gaseous matter, and decrease in the amount of pressure. At the base, cooling would have been relatively rapid because of the proximity of the cold substratum on which the erupted material rested; the dissolved gases in the material would have become disengaged at once; and the pressure of the overlying matter would have been considerable. These conditions would tend to promote the formation of a relatively rigid rock. The

– 339 –

material ten or twenty feet higher up may have had a slightly lower temperature. This, however, is too uncertain to provide any basis for definite conclusions. Even if the temperature of the material when it was emitted had been the same, the conditions of cooling before the point of deposit was reached may have been very different. After deposition the rate of cooling was probably not the same as that of the lower material. The fall in temperature was probably less rapid because of the greater distance from a cooling surface. Heated gases would probably be emitted from the lower part of the ignimbrite for a considerable period and would maintain the upper part at a higher temperature than the lower. At the same time, the uprising of the gaseous matter might tend to prevent the small paricles of glass from welding together as readily as they would if they were undisturbed, an effect that would be augmented by the smaller pressure in this higher portion. It is worthy of note that Lacroix quotes experiments of Brun which show that deformation of particles of volcanic glass of an andesitic nature takes place at a temperature of 938 degrees centigrade and that the glass can be drawn into threads at 1050 degrees. (Lacroix, 1908, p. 54.) He further states that the formation of spherulites will take place in such glass until the temperature has fallen to 460 degrees. It seems reasonable to think that in these highly siliceous and viscous materials development of rudimentary crystalline structure would take place very slowly. Consideration of these results leads to the opinion that the temperature of the glass fragments when they fell exceeded 1000 degrees centigrade and that the rudimentary crystalline development in the upper part of the ignimbrite at Arapuni and at other localities where these ignimbrites have been examined resulted from the material maintaining a high temperature for a considerable time. The comparative incoherence of the upper part of the deposit would be due to the smaller pressure of the overlying matter, and to the disturbing effect of gases rising from the material below.

Specific Gravity and Texture of Ignimbrites.

It is clear that, if the pressure of the accumulating material is the cause of the so-called flow structure of the lower part of the deposit, one would expect that the specific gravity of the material that constitutes the lower part of one of these formations would be greater than the specific gravity of the rock in the higher portion. A test of this was made in connection with the formation at Arapuni itself, where the specific gravity was determined at different depths in the ignimbrite with the following results:—(1) Two feet from the base, 2.09. (2) Six feet from the base, 2.20. At eight feet, 2.16. At 14 feet, 2.00. At 20 feet, 1.95. At 30 feet, 1.81. At 45 feet, 1.57. This progressive decrease from the base upwards well illustrates the extent to which compaction took place as the result of the pressure of the overlying material. The lowest specimen is slightly vesicular, presumably because the rapid cooling did not permit of the escape of the liberated gases completely. It is clear that such a gradation of specific gravity could not develop in a lava flow.

– 340 –

It was found that the material of the upper surface of the ignimbrite at Arapuni was so incoherent that it could be broken with a rubber pestle in water, though it was in no way decomposed. The grading of the material treated in this way is nearly identical with that of the fine sand at the base of the rock. These have been previously stated, and are now tabulated. Lacroix's grading of the Katmai sand flow is added for comparison.

> 1/40 6.60 16.34
1/40–1/50 3.90 4.17
1/50–1/50 2.33 3.00 > 1.65 mm. = 1/10′ 14.57
1/60–1/80 3.77 2.22 0.42 mm.–1.65 mm. 17.70
1/80–1/100 4 73 3.78 < 0.42 mm. = 1/40′ 67.00
1/100–1/200 18 40 12.65
< 1/200 59 60 54.37

Grading of material at the base of the arapunite (Arapuni).


Grading of material near the surface of arapunite (Arapuni).


Grading of Katmai sand flow (Lacroix, 1930, p. 465).

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

The similarity of A and B is close enough to indicate that the deposition of the material took place so rapidly that no sorting took place. The high proportion of B coarser than 1/40′ is probably due to incomplete disintegration.

Differences Between Ignimbrites and Lavas.

Previously it was stated that the very general classification of these ignimbrites as volcanic lavas did not satisfy the field conditions of their occurrence in these five respects at least:—


Their upper surface is approximately horizontal.


There are no volcanic cones which can be regarded as the source of the supposed lava.


There is no scoria on the upper surface of these rocks.


The base is formed of an entirely incoherent but narrow band of fine sand, the individual grains of which are of the same nature as that of the rock itself.


There is a general and pronounced vertical jointing. (Plate 70.)

It is well now to consider these field characters in relation to the “fiery shower” origin that has been suggested above for these rocks.

1. It is obvious that the upper surface of a widely distributed dust shower of considerable thickness would be of an approximately level nature if the ground previously was not particularly irregular. In general, the ignimbrites themselves lie on widely distributed pumice tuffs which are often nearly horizontal. It may here be noticed that the surface of the sand flow rock at Katmai as described by Fenner was nearly level. He publishes photographs which show a gently sloping surface, and specially mentions its level nature in his description. (Fenner, 1925, p. 196.)

– 341 –

2. The absence of volcanic cones formed of rhyolite rocks in this plateau area in New Zealand is a striking fact. Here, again, comparison must be made with the Katmai region as described by Fenner. He particularly mentions that the sand flow material did not issue from any of the cones, but from fissures that are now filled up by some of the material that issued from them; and that their orifices nowhere form any surface mounds or hills. In New Zealand, though the ignimbrite area is often much intersected by stream valleys and gorges, no structure that suggests a fissure through which extrusion might have taken place has yet been described. A locality where it is now thought that such a structure may be seen is eight miles north of Taumarunui. It must, however, be mentioned again that the superficial covering of pumice greatly hampers field observations. The following is Fenner's statement in regard to the sand flow at Katmai: “During the period of eruption a broad Y-shaped valley bounded by abrupt mountains opened in numerous fissures. These are believed to have been the source from which a great volume of hot pumice and fragmental glass (‘the sand flow’) was poured out and forming a phase of eruption similar in some respects to the ‘nuées ardentes' or glowing clouds of the West Indian eruption of 1902.” The area that was covered by this sand flow rock was approximately twenty-two miles by eight. (Fenner, 1925, p. 194.) Fig. 5 shows the flat surface of the sand flow, and Fig. 6 indurated sand flow rock.

3. Absence of scoria on the upper surface of ignimbrites. It is clear that on the surface of a sand flow rock no scoria would be formed, for the gases originally present would generally have escaped in large proportion during the very formation and eruption of the glassy shower. In addition, the glassy particles seldom if ever adhered together with sufficient completeness to prevent the free escape of gases through the spaces between them. Even if complete welding of the particles took place the process would be slow enough to allow much of the gas to escape. The absence of a scoriaceous surface is therefore a normal and necessary condition of the formation of an ignimbrite. Its surface is thus wholly different from the scoriaceous one that would certainly develop on a lava flow of rock with the chemical characters of these ignimbrites. Within New Zealand, comparison with the Auckland basaltic lava flows and with the andesitic lavas of Tongariro, Egmont, and still more with that of the obsidian of Mayor Island, shows a complete contrast. The rough, irregular, crevassed, hummocky surface of these lavas is thus completely different from the smooth surface of the ignimbrites.

4. A rock that was formed from a fiery shower would obviously not disturb or deform the surface on which it was deposited. At Arapuni, where alone the base of an ignimbrite could be studied, it was found, as stated before, that the lowest layer consisted of a fine sand composed of a few crystals of felspar and quartz mixed with a multitude of minute but free particles or shreds of volcanic glass. This would certainly be the actual condition of the material that formed the base of a deposit from a fiery cloud. The particles of glowing glass that formed the bottom of such a cloud would be

– 342 –

cooled to such an extent during transit that they would lose their viscosity and would not adhere together after they fell. The base of the ignimbrite has therefore the precise character that would be expected of material deposited from the base of a fiery cloud of this nature.

5. The peculiarly general and sometimes regular vertical jointing (Plate 70, fig. 1) is clearly due to the contraction during cooling of a heated rock that was completely at rest and effectively solid. The extreme viscosity of such an acid glass, and the amount of cooling that it had already undergone, would practically prevent it from developing any mass flow after it was deposited. It would, however, after its deposition cool through an interval of temperature amounting to approximately 1000 degrees centigrade. This implies a large amount of contraction which would best be satisfied by the development of a series of vertical joints producing a roughly hexagonal prismatic character. A noticeable point is the difference in the diameter of the vertical rock columns in various deposits of the ignimbrites. At Arapuni they are about one foot in diameter; at Ngutuwera their diameter is about three or four feet; at Hinuera they are as much as eight feet in diameter.

This difference in the diameter of the columns may well be due to the initial temperature of the ignimbrite as it was deposited in different localities—the higher the temperature, the more closely the vertical joints would be spaced. There is internal evidence in favour of this. At Arapuni, where the columns are of narrow diameter, the glassy particles of the ignimbrites were of such a high temperature that they fused together almost rigidly and show well-developed parallel arrangement or so-called “flow structure.” At Ngutuwera the columns are far larger, and here no parallel arrangement can be seen. At Hinuera, where the dimensions of the columns are very great, the glass particles have much of the characteristic form of the particles in ordinary tuffs—in other words, they were nearly solid when they reached the ground.

The question obviously arises here as to the extent to which welding of the particles of glass might take place, under the most favourable conditions, after they had fallen, and to what extent reconstituted glassy lavas might be formed. The possibility of the crystallisation of various minerals of igneous rocks in such material also arises. Further reference to both of these questions will be made subsequently.

Sollas appears to have had a suspicion that the material of some of the Hauraki tuffs reached the ground in a viscous state, as the following extracts show (Sollas, l.c., vol. 2, p. 66): “The slice is seen to be composed of contorted flow shreds of the form supposed to be characteristic of tuffs, the interstices filled with darker, more granular glass. The tuff-like bodies contain globulites and spirally curved thread-like crystallites which would be called trichytes were they not transparent.” “There can be no manner of doubt about the original glassy character of the shreds. Their crystalline character at present is obvious, and thus we appear to have definite

– 343 –

evidence of secondary devitrification. The only escape from this inference lies in the possibility that the contorted threads were not wholly solid, but simply viscous, after they acquired their form.” “Rhyolite with plagioclase showing contorted flow lines.”

The Temperature of Ignimbrite Material.

The following considerations will give some indication of the temperature of the material of the ignimbrites when deposited:—

  • (1) Quartz crystals are found in practically all of these rocks. In all ordinary instances these crystals are phenocrysts that have been formed before eruption.

  • (2) Tridymite has often been formed in the later stages of cooling after eruption.

  • (3) The glass particles were generally in a viscous state.

  • (4) In some cases the glass was drawn into threads.

  • (5) The glass sometimes contains small crystallites and globulites that were formed after eruption.

  • (6) Spherulitic structures of several types have often been developed in the glass after it fell.

  • (7) Samples of these rocks raised to a white heat in a blacksmith's furnace fused to a highly viscous material. However, samples heated to a temperature of 1000 degrees centigrade by Mr E. T. Seelye were unaltered.

The temperatures indicated by these observations are as follows:

(1) Not more under surface conditions than 870° C.
(2) Not less than 870° C.
(3) In andesitic material 938° C., in this material probably 1100° C.
(4) In this rock probably not less than 1200° C.
(5) No definite indication as minerals could not be identified.
(6) This could take place at any temperature above 460° C.
(7) The rock fuses to a highly viscous glass at 1200° C.

On the other hand, no crystobalite was seen, which implies that the temperature was less than 1470° C.

The low temperature that is suggested by (1) may possibly be due to different molecular conditions which permitted the crystallisation of silica into quartz at high pressures in spite of high temperatures; or, conversely, that union of gaseous constituents developed high temperatures when pressure fell. At any rate, the abundance

– 344 –

of quartz crystals in association with this acid glass cannot be denied; the grains are large and rounded, and were certainly formed before eruption.

Lacroix maintains that at Mount Pelée the quartz crystals were formed after the consolidation of the rock. The corroded and cracked nature of the quartz precludes this explanation in regard to the ignimbrites that are here described. (Lacroix, 1908, p. 52.)

Chemical Composition.

Few analyses have been made of these ignimbrites. Morgan, however, quotes five of them. (Morgan, 1924, p. 70.) Two of them are given here. Henderson gives some others (Henderson, 1923, p. 57); Grange also gives one (Grange, 1927, p. 40). Some analyses of the Mayor Island comendite are given for comparison. Mr F. T. Seelye has kindly given me analysis K.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

SiO2 73.08 72.89 72.40 75.46 70.10 67.83 72.30 69.61 72.82
Al2O3 13.50 12.83 10.00 11.27 13.76 14.68 12.50 15.53 13.53
Fe2O3 2.60 1.04 6.17 1.17 2.64 4.79 2.12 1.49 2.06
FeO 0.13 0.38 0.93 2.05 0.79 4.79 0.47 0.83
MgO 0.15 0.05 none 0.27 0.17 0.69 0.10 0.32 0.06
CaO 1.07 1.25 0.22 0.53 1.33 2.81 1.35 2.27 1.66
K2O 3.19 3.92 4.54 4.88 3.08 2.90 3.58 2.76 3.86
Na2O 3.95 2.81 5.43 3.45 3.42 3.46 3.25 3.86 2.93
H2O—105° 1.33 0.96 0.29 0.28 0.88 0.66 0.46 1.13 0.71
H2O+105° 3.46 0.29 0.07 3.74 1.87 3.54 1.86 1.45
TiO2 0.62 0.12 0.05 0.26 0.43 0.12 0.37 0.23
P2O5 tr 0.03 0.02 0.31 0.06 0.04
BaO 0.06 0.09 0.08
MnO 0.02
99.68 99.83 100.00 99.48 100.17 100.12 100.20 100.17 99.37

Rhyolite (lithoidal) 100 feet down No. 1 shaft, Grand Junction mine, Waihi. Lower down on p. 70 this rock is called tridymite rhyolite.


Rhyolitic tuff (wilsonite) from quarry, Waitekauri stream crossing the old Waihi-Paeroa road.


Obsidian, Mayor Island. Washington, N.S.G.S. Prof. Paper, No. 99.


Pantelleritic rhyolite, Mayor Island. von Wolff, 1904.


Rhyolite breccia, quarry N.W. Otorohanga Railway Station. (Henderson and Ongley, 1923, p. 57.)


Rhyolite breccia, Arapuni Gorge. (Henderson, 1923, p. 57.)


Rhyolite breccia, near Waikino. (Henderson, 1923, p. 57.)


Rhyolite tuff, Lower Pleistocene. (Grange, 1927, p. 40.)


Tridymite rhyolite, opposite gasworks, Waihi. Mr F. T. Seelye.

Sample A from Waihi is substantially the same in composition as the other samples which have been obtained from localities far apart. All the analyses show that the rock is a rhyolite in composition, though the lime content is rather higher than is usually the case in such rocks. The alkalies, too, have rather different amounts, but in each case the sum has nearly the same value in the other rocks, with the exception of the alkaline types (comendites) from Mayor Island.

– 345 –

Sample G is the typical wilsonite. There is nothing in its composition that distinguishes it from the other types of these tufaceous rocks.

Analyses C and D of the pantelleritic or comendite rocks of Mayor Island are quoted for comparison only. Lava flows of this type are well developed, and have a thick surface of obsidian, while the interior parts of the lavas are of a stony nature. This feature as well as the roughness of surface and inclination of the formation at once distinguishes these lavas of Mayor Island from the ignimbrites which have been described above.

Changes in the Glass of Ignimbrites after their Deposition.

Hitherto the ignimbrites have been described as composed of relatively large crystals of felspar and quartz embedded in a maze of minute glass shreds. It is, however, in a few places only that the glass shreds still retain their vitreous characteristics. Actually, so far as the very numerous specimens that have been examined are concerned, it is only in the wilsonite from Owharoa, and in that from Waikino, and in the lower layers at Putaruru, and in the lower ten feet on the fine grained rock at Arapuni, as well as that from the lower part of the deposit at Hinuera that this vitreous material is still dominant. In practically all of the other samples the glassy nature has been lost; though the form of the originally glass particles has been retained, or is still visible in a maze of other structures. The variety of these structures is very great, and their development is often so pronounced as sometimes, at first sight, to conceal the structural features of glass shreds which reveal the ignimbrite origin.

It is thought that the temperature of the ignimbrite material was so high, even after deposition, that some crystallisation took place within the glass fragments; sometimes, indeed, to such an extent as to dominate the rock structures.

Pectinate Structure.

Such changes are seen in their most rudimentary development in the rock of the Paeroa ranges, close to the crossing of the Waikato River, on the road from Rotorua to Taupo; and to some extent also in the lower part of the rock at Hinuera. This most minute change is the development of very slender colourless needles at right angles to the surface of the glass shreds and extending inwards from it. A resemblance between these needles and the teeth of a comb suggests the use of the term pectinate which is here applied to them. This pectinate structure in the instances that have been mentioned has been developed without any simultaneous alteration of the rest of the material of the glass shreds. The alteration usually proceeds further than this, and in the majority of instances it has been complete, and no glassy residue remains. The sharpness of the teeth in the pectinate structure usually becomes less pronounced as the structure develops, and whilst this change takes place in the margin

– 346 –

of a glass shred, the central portion also changes, and soon becomes an indefinite mass of feebly birefringent particles which are thought to be tablets of tridymite. This development has proceeded to such an extent in many of the rocks as to effect the entire obliteration of the glass. This is the case in the upper part of the Arapuni occurrence, as well as that of Maraetai 5, and along the course of the Mangakino River, and at Motutere on Lake Taupo. In other localities, notably at Puketiti west of Te Kuiti, at Tarawera on the Napier-Taupo road, and near Maungatautari, the whole local development of the rock appears to be changed in this way and without any further advance.

It is to be noted, however, that this structural alteration has been effected without any change in the form of the shreds originally glass, which in the instances mentioned still retain their irregularly disposed arrangement. At times it almost seems as though this alteration had taken place before the glass particles had reached the ground. The pectinate development often continues far, even in those parts of the rock that were subject to such pressure from overlying material that the glass shreds have a parallel arrangement and “flow structure” becomes pronounced.

In other localities the felspar fibres become far stouter, the core of tridymite becomes distinct, and a so-called axiolitic structure is developed. Within this region such a structure is dominant in many places. In the large occurrence near Ongaruhe, notably at Waimiha, it is found consistently throughout the whole mass of the ignimbrite, which here is from 60 to 100 feet in thickness. (Plate 64, fig. 1.)

Radial or Spherulitic Structure.

There appears to be a tendency for some of the pectinate structures to grow at the expense of others, and the felspar needles or fibres may extend from their own glass shred across the finer dusty substance that intervenes between the glass shreds and join up with those formed in another. There seems to be a general tendency to round the structures off, and negative radial structures or spherulites result. A few of these spherulitie or radial areas are to be found in most of the rocks in which this general pectinate character is dominant. This structure is quite distinct in patches of the “indurated sand flow rock” from Katmai given to me by Dr Fenner. The radial structures, however, are at first markedly irregular, though a rock of this kind would normally be called a spherulitic rhyolite. Even when the radial structure is well developed the boundaries and many of the features of the original glass particles can still be distinctly seen.

In some rocks that at first seem to be typical spherulitic rhyolites, as for instance in the quarry on Mount Ngongotaha on the road to the summit, distinct dusty lines cross the fibres of the well-formed spherulites, and comparison of these with the structures in the lower rocks at Motutere suggests that even this spherulitic rock at Ngongotaha may have an ignimbrite origin. (Plate 68, fig. 1.)

– 347 –

Development of Pectinate and Radial Structures in Pink Wilsonite.

The pink wilsonite from Putaruru shows in the most striking manner the development of radial structures. As this change is at first associated with the large fragments of glass in the rock, a separate description is necessary. The large glass fragments, which are sometimes as much as 20 centimetres in length, are narrowly lenticular, a form due to their viscosity when deposited, and to the vertical pressure of overlying matter. These glass fragments contain numbers of capillary gas pores which lie in the direction of the greatest diameter of the glass lenticle. (Plate 66, fig. 2.) In specimens of rock from the base of the deposit these lenticles have no crystalline structure, but are pure glass, except for a few larger crystals of quartz and felspar which belong to an earlier period, and occur generally in all of these rocks. The rest of this wilsonite is composed of irregularly arranged shreds of brown glass, with others of smaller size almost colourless, and a good deal of dusty matter between them. A few feet above the base of this wilsonite there is a complete change in the detailed structure. The dark lenticles are still quite distinct in hand specimens and in section; but in all of them innumerable felspar fibres have been formed with their axes at right angles to the capillary pores and directed into the glass like the teeth of a comb. A number of axiolites have thus arisen within the glass lenticles each with a central line of tridymite more or less distinct. (Plate 67, fig 1.) Sometimes the felspar fibres extend across two or three of the capillary gas pores, and sometimes rounded radial groups of spherulites have been formed. (Plate 67, fig 2.) In some examples the lenticular fragments have formed spongy masses of spherulites. (Plate 68, fig 2.)

The rock has thus become a distinct pectinate type, with here and there some radial development. In hand specimens occasionally spherulites large enough to be distinctly seen with the naked eye are clearly visible when one of the lenticles has been broken across. Occasionally some of these lenticles become vesicular, and even in hand specimens projecting spherulites can be easily seen. This structure is maintained to the top of the pink wilsonite, though the texture becomes rather less compact and the rock softer. It is at once evident that crystallisation begins sooner and is more active in the larger lenticles of glass than elsewhere in the rock.

Development of Structures at Maraetai 5. (Plate 70, Figs. 1, 2.)

At Maraetai 5 of the Perpetual Forests' estate the Waikato River flows in a gorge which here is 350 feet deep. This can be well seen from the trig station 1003, which is approximately 1258 feet above the sea level. From the level of the trig to the bottom of the gorge the rock is ignimbrite with distinct columnar structure, and no discontinuity could be seen in it of such a nature as to suggest that there was more than one period of deposition. The rock at 350

– 348 –

feet above the water level is of a pink colour, and is not very compact, with a specific gravity of 1.91 when saturated with water. At the 200 feet level the rock is less pink, but more compact, and its specific gravity has risen to 2.03. At the water level the pink colour has been lost, and the rock is now compact, with specific gravity of 2.26. These three rocks have porosity of 28.4, 25.9, and 13.4 respectively.

The structure of the sample from the 350 feet level is poorly pectinate, but the component particles have all of that fluffy, irregular arrangement that is found in characteristic tuffs. (Plate 63, fig 2.) This material, when it was deposited, contained relatively large rock or glass fragments. These have undergone considerable alteration. In hand specimens they are soft and pulverulent. In section they consist of distinct and separated spherulites with much tridymite. In a few instances the pectinate structure has developed distinctly parallel to the capillary pores. It is interesting and important to note that the aspect of this rock in hand specimens is closely similar to that of the “indurated sand flow rock” of Katmai; for a sample of which I am deeply indebted to Dr Fenner. The resemblance between the two rocks in hand specimens is maintained in micro preparations, even to the peculiar elementary pectinate structure and the radial or spherulitic structure of some of the patches, which are thought to have been glass originally. This seems to support the probability of a similarity of origin in the most definite manner. In itself it demonstrates that this rock of Maraetai 5, at least, has been formed from the material of a fiery shower.

At a level 150 feet lower, that is, about 200 feet above the river, the tuffaceous character is still dominant in the rock; but the pectinate character is more developed, and the original larger glass fragments are less spongy and show traces of capillary pores more clearly. Parallel arrangement of the shreds or “flow structure” is shown but slightly.

At the water level the pectinate structure is very distinct, with a radial tendency, and the rock would normally be called an axiolitic or even a spherulitic rhyolite. The bending of the pectinate structures round the crystal edges and angles is characteristic, giving a vivid impression of “flow structure.” The original large glass lenticles are now quite dense and are formed of spherulites in close contact, while the original capillary pore effect is not visible. Throughout the rock a radial development is now frequent; but usually the form and arrangement of the original glass shreds can be seen in the spherulites even when they are well formed and circular in section. (Plate 64, fig 2.)

Development of Structures at Te Toki Point, Lake Taupo.

Near Motutere Point, on the east shore of Lake Taupo, at the spot where the road from Tokaanu to Tapuaeharuru passes round the rocky bluff, the development of pectinate, and radial, as well as vitreous structure is well seen. The rocky bluff is approximately

– 349 –

250 feet high. Fourteen specimens were taken from this cliff for examination. As far as possible, the spots from which these were taken were evenly spaced from the top to the bottom of the cliff.

The top sample, which is a soft rock, has a typical tufaceous nature. The glass shreds of which it is composed have a feathery arrangement and a rudimentary pectinate structure. There are a few patches of small size, some of which have a spherulitic nature. Each sample from a spot successively lower in the outcrop shows a higher development of pectinate structure, while the shreds are more compactly arranged. At No. 7, which is approximately 120 feet from the surface, the pectinate structure has become dominant and the feathery character has become almost lost, for the shreds have taken up a roughly parallel arrangement and already can be seen to bend round the angles of the included crystals, and thus give the appearance of flow structure.

In the samples from still lower levels the pectinate structure becomes stronger, and more fully developed. The parallel fibres of felspar are longer and more distinct, and occasionally the fibres in two adjacent glass shreds unite, while the terminations of the shreds have a tendency to round off, and thus develop an irregular spherulite which is still of the transparent or negative type. At the same time, especially from specimens near the base, the streaming appearance of the pectinate representatives of the original glass shreds now presents an impression of typical and pronounced flow structure. Examination of this series of preparations gives no room for doubt that the whole outcrop is a rock of ignimbrite origin, in which the pressure of overlying material has developed flow structure in the lower portions, while the temperature of the material, after it fell, was sufficiently high to allow of development of axiolitic and spherulitic structures. (Plate 65, fig 1.) The lowest rock of all, which is now unfortunately hidden by detritus deposited from roadmaking operations, illustrates still other changes that the ignimbrite material underwent after its fall.

One often finds in this lowest rock at Te Toki that there are irregular patches of clear, colourless glass, sometimes of considerable size. One of these measures 1.2 cm. by 0.4 cm. (Plate 65, fig 2; plate 66, fig. 1.) These patches often have a ramified shape, and the branches penetrate far into and among the original glass shreds. These patches of colourless glass have irregular and ragged borders. Dusty matter contained in them sometimes follows the lines of those brown glass shreds across which the patch extends. Near its border a few small microlites may occur similar to those in the fine dusty matter between the glass shreds. It is impossible to resist the conclusion that this colourless glass has resulted from the fusion or welding in place of original glass shreds of the ignimbrite. In this colourless glass bubbles of gas are not infrequent. In almost all cases they have a perfect rounded form, though occasionally they are slightly oval in shape. This shows clearly that no movement took place in the colourless glass after it had been formed. The dusty

– 350 –

lines are not flow lines in the glass, but are merely inherited structures from the ignimbrite. It is interesting to note that, in the colourless secondary glass, strings of margarites are sometimes found, and occasionally they develop into the curved radiating groups which form such a picturesque micro-feature of the Aratiatia rocks. Spherulitic structure with the dark brown positive spherulites has developed more strongly and regularly in this colourless secondary glass than in the glass shredded ignimbrite material; but lithophyses are not so common. Some, however, were observed with crystals of tridymite projecting from their sides. The cause of the fusion of patches of the rock cannot be certainly stated. It is, of course, probable that the temperature was somewhat uneven at the moment of deposition. Gaseous reactions too probably took place to a considerable extent, and may have caused local increase of temperature.

These observations and descriptions are, of course, opposed to the statements of Weinschenk and Clark (1912, p. 333, fig. 235): “Rock glasses often present the appearance of decided flow structure. In these glasses various coloured bands, mixed with each other in multifarious ways, form the principal constituent. In other cases they flow round the large crystals that have separated out, so that it appears that the different parts were not miscible with one another even in the liquid condition.”

Seeing, however, that clear glass has been formed in this Motutere rock from ignimbrite material after it fell, the question arises as to whether the glassy rocks that contain spherulites have also been formed in this manner from incandescent showers. It is, of course, probable that a far more glassy structure than that observed at Te Toki might be developed.

The field geology of these acid volcanic rocks in this district is of such a uniform nature that it suggests a common method of eruption. Each of the separate areas, however, requires detailed examination before definite statement can be made. At present it can only be said that the structure of the rock of the isolated hill Kaimanawa, which is a granular—not a solid obsidian rather suggests such an origin. The trichites, too, in this rock are of precisely the same nature as those formed in the clear welded glass of Motutere. Again, the well-known rock of the Hemo Gorge near Rotorua with its numerous spherulites enclosed in a somewhat granular glass has a structure which in some respects closely resembles that of the colourless portion of the rock at Motutere.

Compaction of Ignimbrite at Te Toki.

In order that this apparently distinct evidence of compaction in the structure of the rock at different levels should be tested, the specific gravity of the different samples from top to bottom of the deposit at Motutere was determined with the following results. The samples were soaked with water before estimation of the S.G.

– 351 –
Specific Gravity.
Porosity. When soaked. Calculated dry.
1 28.6 1.96 1.67
2 25.2 2.09 1.83
3 23.4 2.11 1.87
4 22.0 2.15 1.93
5 20.0 2.17 1.97
6 19.8 2.16 1.96
7 23.1 2.13 1.90
8 18.4 2.19 2.01
9 9.2 2.38 2.28
10 10.7 2.24 2.13
11 9.9 2.36 2.26
12 8.4 2.39 2.31
13. 8.0 2.40 2.32
14 3.4 2.41 2.37

This table gives remarkable evidence of the compaction of the rock as the distance from the surface increases. The two apparent exceptions in no way detract from this conclusion. The specimens had been collected, without any intention to use them for this purpose, and both No. 5 and No. 9 had a considerable portion of the weathered surface adhering to them. This external surface always has a considerable amount of secondary silica deposited in the pores. This, of course, decreases the porosity and increases the specific gravity. The actual amount of this effect was estimated in a sample from the locality Maraetai 5. In this instance the specific gravity of the exterior surface was 2.04 with porosity 7.96, while the inner portion had porosity 12.01 and specific gravity 1.83.

It may perhaps be said that consideration of the above observations shows that the temperature of the ignimbrite when it first reaches the ground may be as high as the fusion point of this acid glass and may remain at this temperature long enough to allow of the formation of positive spherulites. In other portions under less pressure negative spherulites may be formed in such numbers as to develop a structure that cannot be distinguished from that of a typical spherulitic rhyolite. Mr E. T. Seelye, of the Dominion Laboratory, was good enough to submit some samples of several of these rocks to a temperature of 1200 deg. C., but microscopic examination of these specimens failed to discover any difference in structure, even when the treatment was continued for six hours. Many spherulitic rhyolites in the Rotorua district have definite lines extending through the spherulites. Examination of some of those observed suggests that the lines mark the direction of original glass shreds.

That spherulites may form in the glass of rhyolites as distinct from material in the state of fusion was recognised by Harker. “When they” (spherulites) “are developed in a glassy or devitrified matrix the flow lines are seen to pass uninterruptedly through

– 352 –

the spherulites, and, indeed, the latter may sometimes be seen to have formed subsequently to brecciation of the rock. In such a case the matrix was a glass rather than a liquid when the spherulites crystallised.” (Harker, 1909, p. 275.)

Experimental Evidence in Regard to Crystallisation in Acid Materials.

Until quite recently there has been little or no information in regard to the conditions under which crystallisation might take place in an acid magma, and of the actual temperature required for fusion at the probable depth from which such a magma might be derived. Recently experiments which are of great importance in this connection have been made by Roy W. Goranson. (Goranson, 1932, p. 227.) These experiments have involved pressures extending to 1500 bars, and in some cases the temperature reached 1000 degrees centigrade. These conditions were in some cases maintained for 160 hours. Finely ground granite powder was used in most of the experiments, but in some instances finely ground granite glass was used, which had been previously prepared by fusion of samples of the same granite. Various quantities of water were used with the granite in the different experiments.

Goranson found that at 1000 deg. C. and 960 bars no glass was formed, though with 15 per cent. of water 80 per cent. of the rock became glass. Even at 704 deg. and 960 bars 70 per cent. of the material was changed into glass when 4.4 per cent. of water was present. In this case the grains that remained unfused were all quartz. In other experiments the quartz grains that remained were highly corroded. Even at 600 degrees in 460 hours a small amount of glass was formed. On the other hand, granite glass at 600 degrees and 385 bars with 2.6 per cent. of water became a mass of birefringent grains. It is particularly interesting to note that silica glass treated for 3 hours at 900 degrees and 1000 bars crystallised directly to quartz.

As a result of his experiments, Goranson concluded that a granite magma with 1 per cent. of water at a depth of 10 kil. will begin to crystallise at 1025 C. When the temperature reaches 700 C. 85 per cent. of the original matter will have crystallised. If such a magma were at a depth of 4 kil. crystallisation would take place while the temperature fell from 1025 to 950 C. About 65 per cent. will then have crystallised, and any further crystallisation will be accompanied by an ebullition of water. A similar result would be effected if the magma at 10 kil. contained 3 per cent. of water, and was cooled to 700 C. Fifty per cent. would then have crystallised and ebullition would take place. These results are particularly illuminating in connection with the acid rocks that are now being considered. Not only would crystals be formed including those of quartz; but at the lower temperatures the quartz crystals, that might have formed, could be much corroded. In addition to

– 353 –

this, the residual glass would be highly charged with water throughout its mass. If pressure were reduced, as might well be the case, it would seem that the natural result would be the shattering of the highly viscous magma. This condition is of the precise nature required for the production of an ignimbrite as here defined.

Goranson's experiments with granite glass show that crystallisation may take place in this acid material even at a low temperature, though he certainly employed a pressure far greater than that which would operate in connection with an ignimbrite, the materials of which after eruption would have been subject to pressures of a few atmospheres only.

Here, however, the facts that have been recorded by Penrose in regard to the crystallisation that takes place on the internal surface of adjacent glass sheets in an annealing oven are of special interest. He describes axiolitic and spherulitic crystallisation in artificial glass. (Penrose, p. 112, pl. 1.) A further note on p. 425 is based on a letter from R. L. Frink, a technical engineer who has made a special study of the glass industry. He points out that the axiolitic structure can be produced at will in the flattening ovens by laying sheets of glass one upon the other and submitting them to the action of heat for varying intervals of time. The longer the time, the more opaque the glass will become. This observation of physical processes at glass works is clearly of a similar character as that supposed in this paper as the cause of the pectinate structure, which is so often observed in the earlier stages of the crystallisation process, in the series of glass shreds of which ignimbrites are mainly made up. This important statement of Penrose was overlooked until long after similar conclusions had been reached from a study of these ignimbrites. Penrose suggests that the layer of air between the glass plates aids in the process of crystallisation. It is, however, probable that in an ignimbrite derived from a nuée ardente all the gases present would be of magmatic derivation and more active than air in virtue of their nature and their origin.

Possible Relations Between Spherulitic Rhyolites and Ignimbrites.

The descriptions and statements that have been given above render it clear that rocks which ordinarily would be classed as axiolitic, spherulitic, and flow rhyolites, as well as pitchstones, may in some cases at least have an ignimbrite origin. The area over which rocks of this nature occur in the North Island of New Zealand is large, and their thickness in portions of it at least is surprising. Obviously the question arises as to whether the rocks of some of the areas of spherulitic rhyolite, of relatively coarse texture, in which no remnant of ignimbrite origin is evident, have actually been formed in this way. The sample of spherulitic rhyolite from the quarry on the south-west side of Ngongotaha shows distinct remnants of original structure which suggest an ignimbrite history. (Plate 68, fig 1.)

– 354 –

The frequent occurrence of spherulitic rhyolite rocks in beds which are almost horizontal, without a scoriaceous surface, without obsidian selvages, with a columnar structure, and with a relatively crumbly nature show at least that such huge masses as that at Horohoro, eight miles in length and from two to six hundred feet in thickness, if of lava origin, must have had an extreme fluidity, and must have come to complete rest before they solidified. It is extremely difficult to conceive that such conditions could have actually occurred.

Mr E. T. Seelye has kindly submitted some of the New Zealand volcanic glasses to various temperatures with the following results.

Samples of glass or obsidian from the upper part of comendite lava flows at Mayor Island, where there is a selvage of obsidian sometimes four feet thick on the upper surface and about the same thickness on the lower surface. Other samples were from Rotorua and from Waihi. In both of these cases, however, the samples were taken from distributed boulders.

At a temperature of 600–660 C. after four days the samples from Mayor Island became finely vesicular and puffed out. The Waihi and Rotorua samples, however, were not affected. At a temperature of 900 degrees for three days the bubbles had escaped from the Mayor Island samples, but a strong vesicular structure had now developed in the samples from Rotorua and Waihi. The surface tension and viscosity apparently even at this temperature had been so great as to prevent the gases from escaping. At a temperature of 1000 to 1020 C. a very vesicular pumice was developed from the Waihi specimen, while the Mayor Island samples simply fused and the Rotorua samples were not tested.

The Mayor Island obsidian thus fuses completely with elimination of bubbles at 900 degrees, though it retains bubbles and becomes pumiceous at 600–660 degrees. True lavas of this rock have definitely flowed, and they have a rough, irregular surface with an obsidian coating above and below with a body of regularly crystalline material.

The obsidians from Waihi and Rotorua retain abundant bubbles at 900, and in the former case at 1000 degrees also, in such amount as to be regularly pumiceous, and they show no indication at this higher temperature of the pronounced flow of the Mayor Island obsidian at 900 degrees.

It would seem, then, that if Horohoro was formed of lava rock this huge mass of material must have flowed through its present extent, though certainly extremely viscous and very slow moving, without its temperature in its upper or lower portions falling below 1200 degrees at the very lowest.

Such conditions seem to be beyond the imagination and to transcend the possibilities. On the other hand, there are masses of rock of ignimbrite origin with dimensions of a similar order. At Maraetai 5, for example, as a centre, there are immense masses distinctly of ignimbrite origin: though these have a less complete spherulitic structure than the Horohoro mass, but at present the

– 355 –

latter is not completely known in detail. Similar remarks apply to such large but isolated masses of spherulitic rhyolite as that at Tumunui, on the Taupo-Rotorua road, and perhaps even at Arotiatia.

It is clear, however, that much more field work and very much more petrographic work is required before any assertion can be ventured in regard to the origin of these rocks; though the facts that have been brought forward in this paper appear to the author to establish a strong possibility that they may have an ignimbrite origin.

Other Areas of Rhyolitic Rock.

One is tempted here to make comparisons with the rocks and conditions in other areas in which rhyolitic rocks occur. The author, however, has little personal knowledge of these.

Within New Zealand rhyolitic rocks occur over a large area in the South Island, with the well-known Mount Somers as a centre. Here, however, the rhyolites are older than the Upper Cretaceous. Professor Speight, who has devoted particular attention to this region, has remarked on the wide slightly inclined beds of rhyolites found there and of the occurrence of pitchstones at the base, of a nature that is closely similar to the glassy facies at the base of the Motutere outcrop.

At Brisbane the tuffs have presented difficulties because on the one hand of the evidence of high temperature; and on the other the columnar form with a certain flow structure have been thought by some to indicate that the formation was an actual rhyolite lava. Here, again, the rocks have a wide extension and are nearly horizontal.

Rhyolites have a wide occurrence in the Yellowstone Park of the United States of America. The rocks of this area have been described at some length by Professor Iddings. (Iddings, 1889.) It is evident from a mere inspection of Plates 50, 51, that Iddings encountered many rocks that are almost identical in their features with those called ignimbrites in this paper. He seems to have experienced great difficulty in accounting for their various structures.

On page 403 he states: “The colourless glasses free from microlites are in almost every instance highly pumiceous, so that the glass solidified in thin rods or films.… In some instances it is evident from the confusedly twisted and curved arrangement of the glass fibres and films that the inflated glass mass settled back upon itself or collapsed after the escape of much of the gas.… Hence, in a moving stream of rhyolitic lava, portions which have been inflated to pumice may be forced whilst yet plastic into more compact masses by movement of the lava, and they may be expected to exhibit some indication of their former pumiceous condition. When we remember the former enormous extent of many of the streams of rhyolite in this region we may easily imagine the formation of pumice over the surface of an intensely heated area of lava thus permitting of its subsequent welding.”

– 356 –

P. 465: “In numerous cases a pumiceous character is entirely wanting. The mass is a compact glass, but it consists of irregularly shaped shreds and patches of different colour. These twist and curve round one another and appear like a perfectly welded mass of strips and ribbons and irregular fragments of variously coloured glass. In some cases their shape closely resembles that of fragments of pumice pressed closely together and welded.”

Iddings again deals with this subject in his volume on Igneous Rocks. (Iddings, 1909, p. 331.) “In case the collapse has not been complete there may be porous spaces, and if the lava has flowed enough to draw out the welded glass the structure appears as in fig. 21. In a completely welded pumice that has been long drawn out the structure appears as in fig. 22. In this rock spherulitic crystallisation has arranged itself along the axes of the welded threads—a structure known as axiolitic, already described in connection with spherulitic crystallisation.” The rocks seem very similar to those in New Zealand, and the explanation as collapsed pumice is quite unsatisfactory.

The rhyolites of the Lake District of England appear to present some problems similar to those discussed here. J. F. N. Green, for instance, says: “The frequent exposure of a thin bed of bedded ash below this rock points to virtual horizontality.… These examples merely illustrate the general impression of horizontality and wide extension formed when traversing the lava outcrops.… Nothing in the Lake District is more noticeable, or at first sight more surprising, than the great proportion which weather with a brecciated structure sometimes invisible on a freshly broken surface, but detectable under the microscope by variations in size of the little felspar laths and in the colour of the glass.

“Further for some reason vesicular structure is very rare in Lake Country rhyolites so that the increase of density with solidification has full effect. This may account for the fact that included fragments usually rounded or lenticular are found throughout nearly all the rhyolites. They are often drawn out and bent, following flow lines, and there is a complete gradation from rhyolites with rounded or subangular enclosures, having a disquieting resemblance to a tuff when weathered, to a well-banded rock.” (Green, 1919, p. 163.)

These quotations indicate that very general but perplexing problems similar to those met with in the Rotorua-Taupo region are encountered in other districts where rhyolite rocks are found. It would seem that some, if not all, of these difficulties would be resolved if an ignimbrite origin could be assigned to them.

Ignimbrites as Building Stones.

It has been stated that ignimbrites are almost entirely composed of crystals of felspar and quartz embedded in a multitude of shreds of glass. Each of these substances is, of course, immune from the destructive effect of weathering within the lapse of a historical

– 357 –

period. There can be no question as to the lasting qualities of a rock formed of these substances when exposed to atmospheric conditions. In ordinary tuffs the particles are merely pressed together or are cemented together by such a substance as carbonate of lime. Such a cementing substance is subject to gradual solution, and the rock which depends on it for solidity will crumble.

The fact that the glass particles of an ignimbrite are welded together confers on the aggregate a resisting power immensely greater than that of a cemented tuff or sandstone. In the field evidence of this resistance is abundantly shown on the exposed out crops; for there is no disintegration and no flaking on them. Actually, the exposed surface has a hard crust due to the deposition of silica which is presumed to have been derived with extreme slowness from the finest elements of the ignimbrite.

The resistance to distintegration is well shown by the behaviour of the rocks when treated with acid. When boiled in hydrochloric and in sulphuric acids each of the five types that have been tested has been found to harden to a decided extent. This is the case after repeated boiling in acids and drying of certain samples over a period of four years.

Ignimbrites from Hinuera and Ngutuwera have been used in small quantities in buildings at Auckland and Litchfield as much as forty years ago, and to-day the stone shows no sign of deterioration. This is the case where the stone is in contact with wet ground, as well as fresh water and salt water. Ignimbrites from various outcrops have been largely used in railway works for fifty years, especially for platform curbings, lining of culverts, and bridge approaches, and in no case has any deterioration been observed. The stone has considerable porosity and absorption in common with all relatively soft rocks. In New Zealand climates this cannot be regarded as an important matter. It is only in regions where extremely sharp frosts immediately succeed heavy rainfall of some days' duration, which completely saturates the stone, that any serious effect would result. There are no inhabited localities in New Zealand where such an association of conditions is experienced. The columnar jointing causes the stone to break out easily in a quarry, and it is so soft that it is readily squared without undue expenditure. It is practically a free stone, and can be worked with equal ease in all planes. The great variety of texture and tone will give opportunity to develop almost any architectural effect that is desired.

Classification of Ignimbrites.

The following classification of ignimbrite rocks is suggested:—

Ignimbrites.—Igneous rocks of acid or perhaps intermediate composition which have been formed from material that has been ejected from orifices in the form of a multitude of highly incandescent particles which were mainly of a minute size.

– 358 –

It has, however, to be borne in mind that in any single deposit formed from such material many different types of structure or phases may be developed. These differences in structure are thought to be due to the pressure of overlying material and to the distance from a cooling surface; in other words, to the relative time during which a high temperature was maintained. It is not possible, therefore, to name any one structure as characteristic of any one deposit of ignimbrite.

Since structure fails, it has been found advisable to use texture as the basis of the main divisions of ignimbrites. Subdivisions or phases, however, are based upon structure. Ignimbrites are usually thick rock masses with well-developed vertical prismatic jointing.

A. Pulverulites.—Essentially fine grained rocks. Composed of fine dust like shreds of glass surrounding crystal grains of quartz, felspar, and some hypersthene, hornblende, or biotite.

B. Lenticulites.—These rocks contain conspicuous lenses usually of dark material often drawn out. These are embedded in a matrix of fine glassy shreds amongst which are some crystals more or less rounded in outline.

C. Lapidites.—Pieces of rock not drawn out or lenticular are embedded in fine material which is composed mainly of shreds of glass.

Each of these classes of ignimbrites may have the following phases:—

1. Vitreous.—Apart from crystals all the fine matter and lenses if present are formed of glass.

2. Radial.—Crystallisation in the form of radial groups has developed quite independently of the boundaries of the original particles of glass.

3. Pectinate.—A comb-like development of extremely fine felspar needles with their axes at right angles to the margins of glass shreds. There is often fine granular matter along the middle line. In the more highly developed examples this fine granular matter is found to be tridymite.

4. Plumose phase.—The minute particles are lightly arranged and are in no way deformed. Their arrangement suggests downy matter.

A. Pulverulites.—Thirty to as much as four hundred feet thick. Examples at Arapuni, Ngutuwera, Toki Point (Lake Taupo), Waihi.

1. Vitreous phase.—This is usually restricted to a thickness of a few feet near the base of the deposit. The incandescent particles in most cases were of such a high temperature when deposited that they were viscous and the pressure of overlying matter bent them round the angles of crystals and thus gave rise to the appearance of flow structure. The glass particles weld together, and sometimes so completely that the glassy matter becomes continuous (Plate 63, fig. 1.)

– 359 –

2. Radial phase.—Felspar fibres develop in radial groups giving rise to a spherulitic structure. When this radial structure is developed it is restricted in most cases to a relatively small thickness of rock above that which has the vitreous phase. At Toki Point, Lake Taupo, the radial and vitreous phases occur in association in the same specimen. This association is found at low levels only, and is soon succeeded by rocks in which the radial phase only is found. (Plate 65, fig 1; part of plate 64, fig. 2.)

3. Pectinate phase.—This phase occurs in the greater thickness of the deposit at Toki Point, Lake Taupo; and nearly all of that at Waimiha. It is poorly developed at Arapuni. When well developed, this has been called axiolitic structure, but its development has formerly been ascribed to wholly different causes. (Plate 64, fig 1; part of plate 64, fig. 2.)

4. Plumose phase.—This is commonly the nature of the upper part of pulverulite rocks wherever they have been examined. This phase is usually combined with a rudimentary pectinate development. (Plate 63, fig 2.)

B. Lenticulites.—The lenses of dark rock vary greatly in number and size. Good examples are found at Waikino and Putaruru.

1. Vitreous phase or wilsonite.—All particles large and small consist of glass; but the larger are penetrated throughout by subparallel gas pores of capillary dimensions. The smaller particles often bend round crystals. It apears that this phase constitutes the whole deposit of ignimbrite at Waikino. At Putaruru only the bottom two feet in the total thickness of forty feet are true wilsonite in the strict sense defined above. (Plate 66, fig 2.)

2. Radial phase.—This is moderately developed at Putaruru. The large lensoid fragments of glass develop in this way far more readily, than the small ones and than the fine grained matter. (Plate 67, fig 2.)

3. Pectinate phase.—This is the main phase at Putaruru. The large lensoid fragments of glass develop the structure first; especially on the margin of the capillary pores by which they are traversed. (Plate 67, fig 1.)

4. Plumose phase.—This has not been seen in this class of ignimbrite.

C. Lapidites.—The structure is dominated by the presence of angular fragments of rock of an acid nature. Examples are found at Hinuera, Paeroa Range, Mangakino Valley, and Mercury Bay.

1. Vitreous phase.—Restricted to the base of these deposits. The small particles of glass are not so strongly welded together in this type of rock as in the others. (Plate 62, fig 1.)

2. Radial phase.—This has not been seen in any of the lapidites. Apparently the material was at too low a temperature for such crystallisation to take place.

– 360 –

3. Pectinate phase.—A rudimentary development of this phase is found directly above the vitreous phase at Hinuera and in the Paeroa Range. It is the dominant phase throughout the valley of the Mangakino.

4. Plumose phase.—Combined with a poorly pectinate phase in many localities. (Plate 68, fig 2.)


Ignimbrite is used as a name for a tufaceous rock of acid composition that has been formed from a “nuee ardente Katmaienne” in the nomenclature suggested by A. Lacroix. This type of rock has a wide occurrence over 10,000 square miles in the North Island of New Zealand.

Ignimbrites are distinguished from ordinary tuffs by the following characteristics:—

1. They have a uniform and normally fine texture.

2. There is an absence of bedding.

3. They show a pronounced prismatic jointing.

4. The rocks are coherent and have an effective solidity.

5. In micropreparations the rocks typically show “flow structure.” This “flow structure” is explained as due to the bending of viscous glass shreds round previously existing crystals.

Ignimbrites are distinguished from lavas by the following field characters:—

1. The deposits have a disposition that is in ordinary instances approximately horizontal.

2. There is an absence of glassy selvages.

3. A scoriaceous surface is wanting.

4. Though there is no scoriaceous structure, the specific gravity is low.

5. A thin bed of extremely fine glass dust occurs below the formation in typical localities.

6. There is an increase in the specific gravity from the top to the bottom of each formation of the rock.

7. There is no indication of mass flow.

– 361 –

The following are the evidences of the high temperature of the ignimbrite material when it was deposited:—

1. Fragments of included charcoal are sometimes found.

2. The minute particles of glass of which the ignimbrites consist are welded together more or less completely.

4. Occasionally blebs of glass as much as 2 mm. in diameter have been formed from the welded shreds.

5. There is often a development of crystalline structure after the ignimbrite material reached the ground.

The Development of Axiolitic and Spherulitic Structures in Ignimbrite Deposits.

1. The upper portion of an ignimbrite deposit shows a feathery arrangement of glass shreds with incipient “devitrification.”

2. There is a gradual increase in the distinctness and dimensions of the marginal felspar fibres and the central line of granular tridymite in the original glass shreds.

3. The feathery structure becomes less pronounced and gives place to a linear arrangement of the glass particles.

4. Distinct axiolitic structures become apparent.

5. A development of spherulites as well as axiolites becomes apparent.

6. In places at the base of a deposit the glassy shreds of various colours become fused together. In these blebs of glass there are spherical bubbles of gas as well as globulites and beaded trichites. In such material positive brown spherulites may develop in either the original glass shreds or in the newly formed colourless glass.

This complete transition has been found in all of its details at Motutere only, but in every place where the structure of an ignimbrite has been observed in microscopic preparations it has shown a gradual development of this type throughout its thickness.

In some spherulitic rhyolites the dusty lines which traverse the spherulites and other structures are thought to be inherited structures from the glass shreds of original ignimbrites.

It is maintained that deposition from a nuée ardente provides a satisfactory explanation for all of these features of occurrence and structure.

– 362 –


Microscopic preparations of these ignimbrite rocks are extremely transparent and show very little contrast. It is therefore a matter of considerable difficulty to obtain negatives which illustrate the described features in a distinct manner. This difficulty is greatly increased by the extreme fineness of the details of the structure, while some of the main features are on a relatively coarse scale, but have an outline which is wanting in sharpness. In attempting to overcome the difficulties it has been found necessary to use rather larger plates than is usual.

Explanation of Illustrations.

Plate 62. Fig. 1.—Paeroa type of ignimbrite from the Paeroa Hills close to the crossing of the Rotorua-Taupo road. Lapidite. X 120. Broken crystals of felspar (oligoclase) and quartz and rarely hypersthene embedded in irregular shreds of glass of most unequal size and distributed without any regular arrangement. Some of the shreds are vesicular. Particles of glass are very lightly welded together. In some of the glass shreds minute needles of a mineral, presumably felspar, can be distinguished with their longer axes at right angles to the margin of the shred.

Plate 62, Fig. 2.—Ignimbrite from two feet above the base of the Hinuera occurrence. Lapidite. X 120. A few corroded crystals of quartz and more commonly crystals of felspar. The largest of these has some small inclusions of hypersthene. One large crystal of hypersthene with small crystals of magnetite included in it. Shreds of glass of varying sizes generally irregularly arranged, though there is slight bending round the crystal of hypersthene. Portions of the larger glass shreds are brown in colour, prohably because of minute gas bubbles which are too small to be seen separately.

Plate 63, Fig. 1.—Pulverulite. X 120. From near the base of the outcrop at Arapuni. A large crystal of oligoclase with corroded margin. The greater part of the photograph consists of shreds of colourless glass with parallel arrangement; c.f. N.Z. Journal of Science and Technology, vol. 13. p. 199, 1932. In this case the glass shreds after falling were sufficiently viscous for pressure to give them a parallel arrangement and to make them bend round the angles of the crystal, thus presenting an appearance of flow structure. The temperature, however, was not maintained for a sufficient time or other conditions were not favourable for the development of incipient crystallisation or devitrification in the glass shreds, and they are therefore moderately distinct.

Plate 63, Fig. 2.—Pulverulite, Plumose phase. X 120. From ten feet below the upper surface of the ignimbrite at Te Toki, near Motutere, on the east side of Lake Taupo. Then usual crystals of felspar are embedded in extremely

– 363 –

irregular shreds of glass in which a good deal of crystallisation or devitrification has taken place. This has obscured and largely obliterated their outlines. It can, however, be seen that they have no parallel arrangement. The pressure at this point so close to the surface of the deposit was not able to affect their form.

Plate 64, Fig. 1.—Pulverulite, Pectinate phase. X 120. From Waimiha, Auckland-Wellington railway line, twenty miles north of Taumarunui, ten feet from the base of the deposit. The usual crystals of quartz and felspar, the former of which are much corroded. The shreds originally of glassy material have a marked parallel arrangement showing that they were viscous when deposited and yielded to pressure. The material of the shreds has crystallised, though the crystalline particles are so fine that they are seen indistinctly, even with high magnification. It can, however, be discovered that they have a margin of felspar fibres with their axes arranged at right angles to the surface of the shred. There is a central portion of minute tridymite tablets. The irregular vein-like structure is composed of small transparent spherulites. It is clear that the devitrification of the shreds took place after they fell and that the vein with its spherulites was also formed subsequently to deposition.

Plate 64, Fig. 2.—Pulverulite. Radial phase. X 120. From 300 feet from the surface of the formation close to trig 1003 near Maraetai station, Perpetual Forests' estate. At one side of the photograph the parallel arrangement of the original glass shreds can be distinctly seen. The greater portion of the material has been crystallised into indistinct transparent spherulites. At one place a small deposit of iron oxide has darkened the rock. This rock, if examined separately, might well be called a spherulitic and axiolitic rhyolite with flow structure. Its field occurrence and upward variation to a typical glass shredded rock show clearly that it is an ignimbrite deposited at such a high temperature that the parallel arrangement of the shreds, formation of axiolitic structure, and of spherulites could take place after it fell.

Plate 65,Fig. I.—Pulverulite, Radial phase. X 120. From near the base of the formation at Te Toki, near Motutere, east side of Lake Taupo. In this rock the outline of the original glass shreds can be seen with difficulty. Their material has been completely changed into axiolitic and spherulitic structures. In the two hundred feet of outcrop between this rock and that shown in Plate 63, Fig. 2, microscopic preparations show a complete and gradual transition from the one type to the other.

Plate 65, Fig. 2.—Ignimbrite from the base of the outcrop at Te Toki Point near Motutere, east side Lake Taupo. X 30. The greater part of the dark portion of this photograph consists of glass shreds which are too small and too indefnite in their outlines to be distinctly seen with such a low magnification. A. Lithophyse bordered with a narrow maigin of dark brown spherulitic material. There are small crystals of tridymite T. projecting inwards from the margin of the lithophyse. Brown opaque spherulites are seen at S and S. Q. is quartz which is much corroded and in places is invaded with “gulfs” of colourless glass. Oligoclase felspar at F. At B.B. there

– 364 –

is a large area of colourless glass in which felspar and quartz are embedded. This has a distinctly frayed margin, and it is traversed with very indistinct dusty lines which in place can be distinguished as continuations of brown glass shreds which border the colourless glass. Numerous gas bubbles are embedded in the colourless glass.

Plate 66, Fig. 1.—A portion of Plate 65 enlarged 120 diameters. A portion of the side of the lithophyse shows the dark spherulitic matter with the projecting crystals of tridymite. The shreds of glass can now be distinctly seen between the lithophyse and the colourless glass bleb. The irregular border of the colourless glass has become distinct. It is seen to be traversed with perlitic cracks and it contains a large number of bubbles of gas.

Plate66, Fig. 2.—Wilsonite from the base of the formation at Putaruru. X 120. The greater part of the photograph shows shreds of glass of an indefinite shape and with an indefinite arrangement, and generally of a pale brown tint. One of the larger fragments of glass extends diagonally across the photograph and ends irregularly among the smaller glassy fragments. It is traversed by a large number of capillary pores. In hand specimens such glass fragments are black and have a vitreous lustre. The remainder of the rock is of a pale pink colour because of the presence of dispersed iron one in it. Crystals of felspar and quartz are relatively rare in this rock. It is thought to be evident that this large fragment was in a viscous state when the material was deposited and that it owes its linear form in part at least to pressure of the material that was deposited on the top of it.

Plate 67, Fig. 1.—Wilsonite, Putaruru, from the same locality as the above, but from a point 25 feet above the base. X 120. In this specimen the disseminated glass shreds are devitrified. They have become changed into the usual felspar needles forming a comb-like structure at right angles to the margin of each shred, while the median line consists of a fine granular substance which is probably tridymite. The devitrification here as elsewhere is thought to be primary and to have occurred at the time of the deposit of the ignimbrite. It destroys the outline of the glass shreds, and the structure becomes so highly indefinite that it has become almost impossible to give an adequate idea of its nature in a photograph, especially as there is no parallel arrangement of glass fibres in this instance. A large inclusion of glass extends across the centre of the photograph comparable with that in the previous photograph, and capillary pores can still be distinguished in it, but its structure is now wholly axiolitic with felspar fibres everywhere, their axes at right angles to the line of the pores. At times, or even generally, the felspar fibres extend over considerable distances and cross the capillary pores, which are so fine that they do not attain the thickness of the section. The rock still has a pink tint due to finely disseminated iron oxide, while the inclusion is black, but is now largely wanting in the vitreous lustre.

Plate 67, Fig. 2.—Wilsonite from Putaruru ten feet from the base of the deposit. X 120. Here the original glass shreds have a parallel arrangement, but are quite devitrified—to such an extent, indeed, as to give an axiolitic appearance to the whole rock. The greater part of the photograph shows a

– 365 –

section through a black inclusion now mainly changed into a spherulitic structure. The margin of a large spherulite can be seen extending across a considerable width of the field. This structure somewhat obscures the original capillary development of the inclusion, though this can still be seen in the rather indefinite lines which extend across the imperfectly formed spherulite. Other parts of the inclusion have their structure to a great extent obscured by spherulitic and axiolitic development. The spherulites can actually be seen more distinctly in hand specimens than in microscopic preparations.

Plate 68, Fig. 1.—Spherulitic rhyolite from road cutting on Mount Ngongotaha, Rotorua, on the road to the summit at an elevation of 700 fect. X 120. The radial structure in the large spherulties in this section is crossed by numerous narrow and discontinuous bands. These are of an indefinite nature, and it is thought that they are the remnants of an original ignimbrite structure. If this is the case, the ignimbrite material was clearly of such a high temperature when deposited that a complete spherulitic structure developed subsequently, and now these indefinite lines alone reveal the original ignimbrite structure.

Plate 68, Fig. 2.—Lapidite from the junction of the Waikato and Mangakino Rivers. X 120. The ignimbrite material has a plumose structure, though but little of it is visible in this photograph. The greater part of the photograph is a portion of a typical soft inclusion which is characteristic of lapidites over a wide area. Apparently this inclusion had cooled to such an extent before it reached the ground that a crust had formed round it. It was therefore not compressed, and gaseous matter was unable to escape from it. Its internal structure is therefore soft and loose and is altogether wanting in that compact structure found in the black inclusions that have been illustrated in previous plates. In this case crystallisation took place in radial aggregates arranged on the inside of the hard marginal crust of the inclusion. The interior of the inclusion consists of loosely arranged spherulites which are composed of felspar fibres with which many tablets of tridymite are associated.

Plate 69, Fif. 1.—Outcrop of pink wilsonite near Putaruru. The cliff is thirty-five feet high. Fig. 2.—Quarry face of pulverulite at Ngutuwera. The face of the rock is twenty feet high. Fig. 3.—Outcrop of lapidite at Hinuera. The cliff of lapidite is fifty feet high. The scale is given by a figure standing near the centre of the photograph.

Plate 70, Fig. 1.—Gorge of the Waikato River at Maraetai 5. The cliff face on the opposite side of the river is four hundred feet high, and is formed of a single deposit of ignimbrite. Fig. 2.—Part of the lower portion of the cliff in Fig. 1.

Table of References.

Bell, J. M., and Fraser, D., 1912. N.Z. Geol. Surv. Bull, No. 15.

Cox, S. H., 1882. Reports N.Z. Geol. Surv.

Dieffenbach, E., 1843. New Zealand. John Murray, London.

Fenneb, C. N., 1925. Journ. Geol., vol. 39.

Ferrar, H. T., 1931. Annual Rep. N.Z. Geol. Surv.

– 366 –

Goranson, R. W., 1932. American Journ. of Sci., vol. 31.

Grange, L. I., 1927. N.Z. Geol. Surv. Bull., No. 31.

— 1928. Annual Rep. N.Z. Geol. Surv.

Green, J. F. N., 1919. Presidential Address, Proc. Geol. Assoc., vol. 30.

Harker, A., 1909. Natural History of Igneous Rocks. Methuen, London.

Henderson, J., and Bartrum, J. A., 1913. N.Z. Geol. Surv. Bull., No. 16.

Henderson, J., and Ongley, M., 1923. N.Z. Geol. Surv. Bull., No. 24.

Holmes, A., 1920. Nomenclature of Petrology. Murby, London.

Hutton, F. W., 1889. The Eruptive Rocks of New Zealand. Trans. Roy. Soc. N.S.W., vol. 23.

Iddings, J. P., 1909. Igneous Rocks. Wiley, New York.

— 1899. U.S. Geol. Surv. Monograph, 33. Washington.

Lacrolx, A., 1930. Liv. Jub. Cen. Geol. Soc. de France. Paris.

— 1908. La Montagne Pelée apres ses éruplrons. Paris.

— 1904. La Montagne Pelée et ses eruptions. Paris.

Marshall, P., 1929. Bull. Dept. of Scientific and Industrial Research, No. 11. Wellington, New Zealand.

— 1932. New Zealand Journ. of Sci. and Tech., vol. 13.

Morgan, P. G., 1911. Trans. N.Z. Inst., vol. 43.

— 1912. N.Z. Geol. Surv. Bull., No. 15.

— 1926. N.Z. Geol. Surv. Bull., No. 26.

Penrose, L. V., 1910. Amer. Journ. Sci., 30.

Pirsson, L. V., 1916. Amer. Journ. Sci., 40.

Rutley, F., 1900. Q.J.G.S., vol. 55.

— 1901. Q.J.G.S., vol. 56.

Sollas, W. J., 1905. Rocks of the Hauraki Goldfields. Wellington: N.Z. Gov.

Taylor, Rev. R., 1855. New Zealand and Its Inhabitants. London: Wertheim and Macintosh.

von Hochstetter, F., 1864. Reiss der Novara, Geol. Teil. Bd. and Abt. 1, Wien.

von Wolff, von F., 1909. Centralblatt fur Min. Geol. Palae. Stuttgart.

Washington, H. S., 1917. U.S. Geol. Surv. Prof. Paper, No. 99. Washington.

Weinschenk, E., and Clark, R. W., 1912. Petrographic Methods. McGraw Hill, New York.

Williamson. J. H. 1930. Annual Rep. N.Z. Geol. Surv.

Picture icon

Fig. 1.
Fig. 2.

Picture icon

Fig. 1.
Fig. 2.

Picture icon

Fig. 1.
Fig. 2.

Picture icon

Fig. 1.
Fig. 2.

Picture icon

Fig. 1.
Fig. 2.

Picture icon

Fig. 1.
Fig. 2.

Picture icon

Fig. 1.
Fig. 2.

Picture icon

Fig. 1.
Fig. 2.

Picture icon

Fig. 1.
Fig. 2.