Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 59, 1928
This text is also available in PDF
(6 MB) Opens in new window
– 571 –

The Volcanic Deposits of Scinde Island. With Special Reference to the Pumice Bodies called Chalazoidites.

[Read before the Hawke's Bay Philosophical Society, 20th August, 1926; received by Editor, 16th April, 1928; issued separately, 23rd November, 1928.]

Plates 68, 69, 70.

Scinde Island is situated on the East Coast of the North Island of New Zealand, and has a large part of the residential section of the town of Napier built upon its hills. The island is an irregular ellipse in shape, its longest axis measuring about 1.85 miles and running almost in a north-eastern and south-western direction. Its widest part, situated close to the north-eastern extremity, measures about three-quarters of a mile. At this extremity it reaches its greatest height of 330 ft. above sea level. Running through the length of the island is a central watershed from which valleys descend on either side. In many places it is bordered by cliffs, and there are numerous quarries and cuttings which facilitate a study of its complicated geology. Although still called an island it is now in reality a peninsula, since within recent times plains have formed between it and the mainland. These plains are largely due to the deposits of silt and shingle brought down by the Tukituki, Ngaruroro, and Tutaekuri rivers. The island mainly consists of highly-fossiliferous limestones which may provisionally be regarded as of Mid-Pliocene age. At some time during this period or later it was raised above sea-level, and during late Pliocene or early Pleistocene times was subjected to extensive denudation, when it obtained very much its present contour. Deposited unconformably on these limestones are pumiceous clays and beds of pure volcanic ash.

The volcanic deposits in Scinde Island may be classified as follows:—






Pumiceous clays and beds of volcanic ash of subaqueous deposition


Pumiceous clays and beds of volcanic ash of subaerial deposition




Layers of pumice were stated by McKay (1) to be interbedded with the limestones in the cliff known as the Bluff. This was denied by Hill (2). There is no doubt, however, that McKay is correct. About 30 ft. above the level of the road close to the breakwater is a stratum extending for some distance of pure pumiceous material.

– 572 –

It is about a foot in thickness; in its lower part it is of coarse texture, and in its upper part it is extremely fine. Microscopical examination demonstrates that the great bulk consists of rounded quartz grains—showing that it had been well washed about by the sea—and that volcanic glass and possibly magnetite are present. Other and more extensive but inaccessible beds are almost certainly of the same nature.

Yet other beds of pure pumiceous material, according to Mr. Hill, were encountered in the “blue clay” which underlies Scinde Island during the course of an experimental boring for oil put down close to the breakwater, and this is in agreement with the fact that beds of pumice are found in corresponding deposits in the country adjoining Napier (3).


Resting unconformably on the Napier limestones are pumiceous clays interbedded with layers of pure volcanic ash.

(A) Beds of Sub-Aqueous Origin. Par the formation of these beds it is necessary to assume that after the island had been elevated above sea-level and the limestones denuded it was once more sub-

Picture icon

Fig. 1
a = Lower Napier limestones.
b = Pumice deposit of sub-aqueous deposition.
c = Pumice deposit of sub-aerial deposition.
d = Pumiceous clay containing greywacke pebbles.
e = Pumiceous clays.

merged and formed the floor of an almost currentless freshwater lake. Another but less likely possibility is that these beds represent an estuarine deposit. Fig. 1 is a diagramatic section of the cliffs on the south-eastern aspect of the island facing Hyderabad Road, including the brickyards of Morse & Robertson. Here there is a pocket of pumiceous material between 70 and 80 ft. in thickness. This is the deepest section that has yet been exposed, and probably the deepest that exists.

The lowest layer of clay extends about twelve feet below the level of the present floor of the quarry, and contains a stratum of waterworn greywacke pebbles. Similar pebbles are found in a few other localities in Scinde Island, as in Northe's quarry at the end of Faraday Street, where they occur in an unusual whitish clay; in Battery Point, and in Burns Road quarries. At Park Island, the

– 573 –

nearest point on the old mainland to Seinde Island, the same stratum is present. To the left of the roadway leading to Quarantine Island near Petane there is a hillock with a considerable stratum of pebbles interbedded with the pumiceous clays. This particular deposit appears to be of great importance since it probably represents part of the vast deposits of pumice which alternate with the shingles, clays, and lignite-beds, the typical section of which appears in the cliffs forming the southern shore of Hawke Bay, and which Mr. Hill has called the Kidnappers section. This remarkable deposit is found, over a great extent of country, and was brought about mainly by the action of fresh water. At Redcliffe, near Taradale and 6 miles from Napier, this deposit has pumiceous clays and volcanic ash, both of sub-aerial origin, resting unconformably upon it. Mr. Hill regards the Kidnapper section as of Pliocene age. It is certainly younger than, the Napier limestones and may be classified as belonging most likely to the later Pliocene or to early Pleistocene times.

Above this layer of pumiceous clay with its greywacke pebbles is another layer of a chocolate-brown colour containing much organic matter—a point of some commercial importance since it is found that in the manufacture of bricks much less fuel is needed for material obtained from this layer than for that from other layers. On extracting with ether an oil was obtained, but its properties have not been determined. It seems most likely to be of vegetable origin.

The next layer of interest above this is about 4 or 5 ft. in thickness. In its upper part there are a large number of dark globular bodies, which in the lower part of the layer are larger and more irregular, and are known to the quarrymen as “ironstone.” The surface of the small ones is often mammilated, and there is evidence of concentric lamination. These bodies also occur in the cliff above Sea View Terrace, in Havelock Road, and also at Park Island.

The next succeeding layer to be considered consists almost entirely of pure volcanic ash. It is deeper in the middle of the “pocket” than at the sides, and averages about 10 feet in thickness. The lowest part consists of coarse rounded quartz grains, but the upper part is very much finer in texture, the appearance suggesting a sedimentary origin. Microscopical examination shows it to consist almost entirely of angular fragments of quartz and felspar and glass laths. The rounded quartz grains show that some of the material was sorted under water, but the presence of long thin laths of glass show that a certain amount of the material was not transported any distance. The deposit may have been due to pumice falling into fairly calm water, such as a lake, into which some sand was being brought; but the material is not what one would expect from a purely river-borne deposit. This same deposit occurs on the western side of Griffin's quarry, and in the valley attains a maximum thickness of between 20 and 30 ft.

Various other beds of clay having different colours, textures, and properties for brickmaking occur before reaching the beds that are definitely of sub-aerial formation. It seems certain that the beds already described at one time formed a continuous sheet over the whole district, but that as a result of denudation only certain portions of them remain in Scinde Island.

– 574 –

(B) Beds of Sub-Aerial Origin. These beds are present all over the island and accurately follow its contour. They are deeper in the valleys than on the hilltops, and although this would be expected from their method of formation it could not be definitely established that they were thicker on certain aspects than on others. This is perhaps explained by the fact that denudation was in progress during the relatively slow formation. They are stratified, and towards their upper-part are the layers of volcanic ash which it is proposed to describe later in some detail. These volcanic ash-layers, averaging about 3 ft. 6 in. in depth, are sharply demarcated from the underlying clays, but above merge into the two or three feet of fine light-coloured clay below the surface-soil layer. There can be no doubt that all these clays were originally pumice which has undergone decomposition. Why certain interbedded layers present volcanic material relatively fresh as compared with that of other layers is probably largely explained by a varying rate of deposition. Assuming that the materials forming the pumiceous clays were deposited at a relatively slow rate they would be subjected to weathering conditions which would result in their mechanical disintegration and decomposition, while if the layers of pure volcanic ash were deposited much more rapidly they would not have been subjected to the same extent to such influences. It is also, possible that the clays themselves have been derived from pumice having a slightly different chemical nature which would render them-more liable to decomposition, or perhaps they are wind-borne from a much older deposit of volcanic ejectamenta in the surrounding country. The fact that the upper layers of pure volcanic ash merge gradually into a clay somewhat similar to some of the subjacent clays, and that interbedded with the volcanic layers themselves are pumiceous clays, is perhaps evidence that the change has taken place in situ, although, of course, these interbedded clays may also be derived from older deposits during a period of volcanic intermission. The appearance of the formation is, however, rather against this. Moa-bones have been found in these clays, so that it is reasonably certain that their formation was a somewhat slow process. Probably a great deal of the clays in the North Island have their origin in decomposed pumice. Jameson (4), in dealing with the Auckland clays, regards them as being formed from rhyolitic pumice brought down by the Waikato river from its upper reaches. Hill suggests that the blue clays,” or more correctly blue argillaceous sandstones, are themselves-simply volcanic ejectamenta that have been altered by aqueous influences.


This is represented by the surface soil which, in an article on volcanic dust-showers in Napier (5), Mr. Henry Hill showed to consist largely of material of volcanic origin. At a later date (6) he recorded a volcanic dust-shower in Napier on the night of 14th December and early in the morning of 15th December, 1896. The wind had been blowing from the north-west while an eruption of Mount Tongariro was in progress. The dust that fell was of a light-grey colour, and as Mr. Hill pointed out the shower bore full testimony to the truth of his statement that “in and around Napier a large percentage,

– 575 –

in fact the larger portion of the soil is of volcanic origin.” Subsequent analyses have amply confirmed this.

Picture icon

Fig. 2
Volcanic Beds Of Sub-Aerial Origin.

The thickness of these beds varies with the amount of denudation that their upper layers have undergone prior to the deposition of the overlying stratum of pumiceous clay. Measurements from the base of

– 576 –

the volcanic deposit to the lower border of layer 8, the chalazoiditic layer, both clearly defined points and representing a thickness from the uniformity and constancy of the intervening layers that has probably not been appreciably altered by denudation, were made in a large number of places on the island. From these it was clear that the deposits are certainly thicker in the valleys than on the hilltops, and probably thicker on the northern and western than on the southern and eastern aspects. The following is the most constant order in which the layers are deposited, and is applicable also to those areas that have been examined on the mainland within a short distance of Scinde Island. (See Fig. 2).


The lowest, a most constant characteristic layer, is one to which Mr. Hill was undoubtedly referring when he described the pumice as being “finer than the finest flour and as white as snow.” This layer is about 2 ins. in thickness. It rests on an uneven surface of brown pumiceous clay, and in consequence often has a sinuous outline. This layer contains little if any magnetite.


A layer of about equal thickness, greyish in colour, and composed of particles which are rather coarser than in the preceding layer.


A layer stained red or brownish by iron oxides, about ½ in. in thickness, is very constant. In some places where the deposits are thick this layer may be double and separated by the greyish material of layer 2.


A layer of about 3 in. in thickness, coarser in texture than the preceding layers, and becoming coarser in its upper part, often apparently forming two different strata. This layer contains fragments of pumice sometimes about a quarter of an inch in thickness.


A well-defined brown layer of about 1½ ins. in thickness which is relatively resistant to denudation and is the lowest layer in which sparse small chalazoidites are present.


A layer about 10 in. in thickness containing larger pieces of pumice than any of the others. The largest are in the upper part of the layer, one fragment in the dry state being over 900 mg. in weight. These pumice fragments are irregular in shape and their angles are smooth and rounded. The upper part of this layer contains chalazoidites intermingled with the pumice fragments.


A layer stained a reddish colour due probably to limonite, which is not very constant, and when present averages about a quarter of an inch in thickness. Sometimes it is very well marked, as in the cutting in Napier Terrace, and contains small rather sparsely-scattered chalazoidites.


A layer which, on an average is 12 in. in thickness, in which the chalazoidites are most abundant. The matrix appears the same throughout. It is divisible into two fairly well-defined layers, the lower part con-

– 577 –

taining the largest chalazoidites and the upper part containing smaller ones of a remarkably constant size.


A very tough layer resisting denudation more than the preceding layers, and measuring on an average about 2 in. in thickness. This layer, often divided into two, is also packed in its lower part with small chalazoidites, apparently continuous with the subjacent stratum.


A layer of fine pumice dust averaging about 3 in. in thickness.


A tough red-coloured layer about 2 in. in thickness, similar to layer 9.


A layer of fine pumice dust about 4½ in. in thickness.


A clay layer of about 1 in. thick.


Following on this is a deposit of fine ash which is sometimes traversed by another layer or two of pumiceous clay and merges gradually into a layer of pumiceous clay between 2ft. and 3 ft. in thickness. Resting on this is the soil-layer.

This order is applicable to all the deposits found in various parts of the island, although from layer 10 onwards they may be absent wholly or in part. On examination of a section exposing these layers, more or less vertical cracks are seen running sometimes from the clay above to the clay below, more often extending through only a few layers. They are probably due to tension during drying and consolidation.


One of the most interesting features of these deposits is the presence of ellipsoidal-shaped bodies composed of pumice varying in size from about a pin's head up to about half an inch across. They have been formed, according to the theory to be stated later, in very much the same manner as a hailstone, and have a similar laminated structure. For this reason I have called the theory of their formation the “hailstone theory,” and the bodies “chalazoidites” from the Greek + resembling hail, and + a stone. The term “pisolite” has been used, but is unsatisfactory as it refers to a different structure having a different origin. They have also been called “lapilli” but this, too, is a misnomer. There is a distinct need for a new term and “chalazoidite” seems to have the advantage in accurately describing their suggested method of formation.

Chalazoidites occur in layers 5 to 9, but are present in greatest abundance in layer 8, which is therefore called the chalazoiditic layer. There seems no reason why they should not occur in the other layers, and it is possible they may have been present originally but have suffered disintegration. This appears to have definitely occurred in the upper part of layer 9, which, as will be mentioned later, probably represents an interval of intermission of volcanic activity, when weathering influences would result in their destruction. In some sections, as in Amner's quarry in Milton Road, chalazoidites are found in other layers besides those mentioned; the extraordinary density which characterizes the whole section in this locality may explain their preservation in these layers.

– 578 –

In the chalazoiditic layer itself as a general rule the chalazoidites appear to be diffusely scattered through the dust-matrix; the largest ones are found towards the base, and the smaller ones towards the top, the transition from the large to the small being rather sharply defined. There is evidence that the chalazoidites were not always diffusely scattered through the matrix, but were themselves arranged in layers. This is particularly well seen in sections in the valley through which Burns Road runs where the chalazoiditic layer is well developed, measuring 17 or 18 in. in thickness. Here in the lowest part the chalazoidites are quite small and arranged in several horizontal layers; above this is a stratum where the largest and also the smallest chalazoidites occur, and this in turn is succeeded by the layer of smaller ones of uniform size. In a few other places on the island the same arrangement can be observed, and is probably explained by the fact that this condition is usually found in sheltered valleys where wind action would not affect stratification.

It is probably impossible to determine what was the original relationship of chalazoidites to matrix, as the latter consists to an unknown extent of disintegrated chalazoidites. On passing a large quantity of the upper part of the layer and a similar quantity of the lower of the layer, through a “30 mesh” sieve, i.e., one with 30 holes to the linear or 900 to the square inch (a size which is very effective in the separation), it was found that in the case of the lower sample a little less than half by weight was retained in the sieve, and in the case of the upper sample a little more than half. The residue left in the sieve consisted entirely of chalazoidites, fragmented chalazoidites, “tubules,” and small fragments of pumice. It is a very extraordinary fact that the proportion of chalazoidites should be roughly the same whether they were removed from the lowest part of the layer where the largest occur or from the upper part of the layer, where they are almost uniformly of small size. Still further, it appeared in both cases that about one half of what was retained in the sieves consisted of fragmented chalazoidites. It might be expected that the proportion of chalazoidites would differ in different localities for the same particular stratum, such as the chalazoiditic layer. For at least several parts of Scinde Island the proportion appears to be constant. If one or more of the conditions necessary for chalazoiditic formation were absent, then they would be absent, as appears to be the case in places west of Napier.

Appearance: The larger chalazoidites are of a light-grey colour with small dark rectangular patches very similar to those on a bird's egg, and on sectioning, similar black markings are sometimes seen running in the line of the laminae. The smaller chalazoidites are darker in appearance, and the dark areas are not noticeable.

Weight: The heaviest chalazoidite, a very exceptional specimen, found in the western portion of the island, weighed when dry 1,063 mg., and the smallest that appeared definitely to be a chalazoidite weighed 1 mg. If 10 grins, of the lower part of the chalazoiditic layer (the specimens in this case were taken from near Jellicoe Ward at the Napier Hospital) is examined, there will be found on the average one chalazoidite weighing 300 mg., two weighing 200 mg., six weighing 100 mg., and about 40 weighing 35 mg., the remainder,

– 579 –

including a large number of fragmented specimens, would average 15 mg. If 10 grammes of the upper part of the chalazoiditic layer, is examined it will be found that the chalazoidites are much more uniform in size, the average weight being about 15 mg. Fig. 3 illustrates graphs showing the numbers of intact chalazoidites of certain weights differing successively by 5 mgs., present in an equal mass first from the lower (broken graph-line), and then from the upper part of the layer (solid line). The solid graph-line shows that the largest number of chalazoidites in the upper part of the layer weigh 15 mg., and none in this particular sample was found weighing more than 35 mg. The broken graph-line shows that here, too, the greatest number of chalazoidites weight 15 mgs. There was no evidence that the smaller chalazoidites were derived to any appreei-

Picture icon

Fig. 3

able extent from the larger chalazoidites owing to exfoliation of the capsular layers, nor was there any evidence that the larger chalazoidites were built up by accretion from the smaller ones.

Shape: A large number of chalazoidites of varying sizes were measured, and it was found that they are all ellipsoids having varying eccentricities with no constant relationship between them. Some are almost perfect spheres, while others are shaped very much like a grain of wheat. During descent they probably tended to be spherical and were saturated in varying degrees with water, which would render them plastic. The impact on striking the ground, if it were an even surface, would tend to flatten them out, and it is noticeable when some of the larger flattened specimens are examined in situ, that the short axes are perpendicular to the plane of the stratum, although this is by no means a constant feature. In some cases small dents or depressions can be made out, almost certainly due to the fact that in their fall, instead of striking the soft pumice

– 580 –

matrix, the chalazoidites came into contact with others that had already fallen, or in their turn were struck by others falling later. The weight of the materials deposited above the chalazoidites might also be a slight factor in causing the flattening, but it seems more probable that in order to remain unbroken they must have been at least only moderately saturated, and therefore only moderately plastic. If they had been, saturated in excess they would have undergone destruction immediately on striking the ground. There is no evidence that this was not the fate of some of them, nor is there any evidence as to the length of time involved in the formation of the layers themselves, although in all probability it was extremely rapid, and weight might therefore cause slight flattening of the chalazoidites after deposition, and before drying out or setting took place. Fig. 6 shows chalazoidites which have been flattened on one aspect, doe apparently to the impact on striking the ground. This shows definitely that they must have been in a plastic state. In these cases of a definite flattening on one aspect, it is noticeable that this is at one pole of the longest axis, showing that they fell with the long axis vertical, and probably had a spinning motion on this axis. The appearances suggest that there was no violent wind at the time of their deposition, and that they fell vertically. If, in the process of setting, there was contraction, then the tendency of a spherical body to assume a tetrahedral shape might also be invoked as an explanation for some of the irregularities. The fact that chalazoidites show considerable variations in the ratio of the short axes to the long axes, and the presence or absence of dents, is largely explained by the different degrees of saturation resulting in different degrees of plasticity, whilst in many instances the irregularities are due to exfoliation of the outer capsular layers. Still one other irregularity particularly well marked with the larger chalazoidites is the presence of a small “tail” of what appears to be adherent matrix, which can in most cases easily be removed from the capsule by a little pressure. It is, in all cases quite distinct from the outer layer. This may be analogous to the “tail” of a falling drop of water, but the comparatively loose attachment is very suggestive that it was formed secondarily. The more probable explanation is that the water in the plastic chalazoidite on reaching the ground would, by capillarity, collect at the lower pole, and owing to the chemicals in solution would cause to adhere a portion of the adjoining matrix, thus forming a “tail.” A stained section through a chalazoidite (see Fig. 7) sometimes shows a meniscus where the stain is more intense, suggesting that there has been water which has caused slight chemical alterations. If a chalazoidite is cut across and part of the cut surface is placed in contact with a dye the fluid is absorbed and forms a meniscus with the convexity upwards, which is explainable by surface tension and the greater porosity of the central nucleus. The appearance is very similar to that seen in the stained specimens. Why it is not seen more often probably depends on the manner of the section, and it is also possible that it may be due to water that has entered the chalazoidite long after its formation. An examination in situ to determine on which aspect the “tail” was situated was, as might have been expected, unsatisfactory.

– 581 –

Density: No exact measurements were made to determine the hardness of specimens of different sizes in the dry and in the wet state. Most of them are considerably denser than those mentioned by Pratt (7) as occurring at Mount Maquiling in the Philippines, which could be “broken only with difficulty between the fingers.” Some of the larger ones were able to withstand a pressure of well over 100 lb. without fracturing. In the saturated condition they will not stand so much pressure, but show no evidence of deformation before fracturing occurs.

Effects of Heat: On moderate heating the chalazoidites become somewhat dark in colour. This appearance is due to organic material that has entered into their composition secondarily. If the eruption were intermittent it is likely enough that a slight vegetation of lichens and mosses sprang up similar to that which occasionally grows on exposed sections at present. The decomposition of this vegetable matter would account for the organic material, although it seems more likely that it has entered at a later date from the period of intermission represented by the superjacent clay stratum. At a temperature of about 800° C. the outer layers become partially vitrified, but a temperature of above 1,000° C. was needed to ensure complete vitrification.

Resistance to Disintegration: One of the most extraordinary features of the chalazoidites is their resistance to disintegration. For their preservation it seems necessary that they should fall on soft ground, otherwise they would be broken in pieces or deformed out of all recognition. Before and after drying, certain chemical changes must have occurred cementing them into a solid mass. What exactly these changes are it is difficult to state. It has been pointed out (8) that 0.150% Si O2 is dissolved from fine tuff after standing in a vessel for six weeks. From this it might be assumed that the silica was reprecipitated and cemented the chalazoidite together.

Something similar is described by Hewitt (9) in the cliffs northwest of Arapuni along the northern side of the Waitete Valley. These cliffs are composed of pumiceous breccia, and there has developed a hard outer skin up to 2 in. in thickness due to the deposition of silica. The same explanation may be used for the density of the layers which represent periods of intermission. The more or less vertical cracks in these volcanic layers are often bounded on each side by similar dense layers.

That there have been chemical changes since the formation of the chalazoidite admits of no doubt. At one time it must have been plastic, but no amount of saturation at present will reproduce that condition. The larger and medium-sized chalazoidites are particularly resistant, but the smaller ones are not so resistant. This is probably due to the fact that in the smaller chalazoidites their surface area in relation to their mass is relatively greater than in the larger chalazoidites, and also, as will be mentioned later, their porosity is greater. Their ellipsoidal shape would also help them to withstand pressure. The nature of the matrix in which the chalazoidites are imbedded is also, of importance. When the matrix is soft, as it usually is, evidence of disintegration is shown, by the large number of fragmented specimens present on sieving. When it is of the

– 582 –

peculiar density seen in Amner's quarry disintegration appears to have been absent or almost negligible. There would appear to be two methods of disintegration. In the case of the larger chalazoidites exfoliation of the capsule seems to be the commoner method; in the case of the smaller chalazoidites, complete fracturing. Pratt (7), in discussing this question in relation to the chalazoidites that occur in the Philippine Islands, states, “It would appear equally remarkable that they should retain their form upon falling into water. Yet it is beyond question that the tuff series into which the wells at Banan and Taal penetrated is in great part water laid, and it is to be presumed that the mud balls encountered in the wells at these towns fell into the sea originally.” It would be expected that deep borings on the plains round about Napier would also show the presence of chalazoidites, but up to the present they have not been noticed, probably because they have not been looked for. In a section of these volcanic layers a whitish efflorescence appears which is crystalline and soluble in water. Its exact chemical nature has not been determined, nor what part it plays, if any, in the “rotting” of pumice.

No experiments have been carried out to ascertain the elasticity, the electrical properties, etc., of chalazoidites, but no doubt if these were determined for varying sizes very interesting graphs would be obtained.

Structure: On naked-eye examination of sections it was found that the larger chalazoidites showed, as the result of weathering, slight evidence of a nucleus, which in some cases had apparently fallen out, leaving a cavity; and more rarely that there was a distinct suggestion of concentric lamination. Usually a cut section of a chalazoidite appears perfectly homogeneous, except perhaps for a slightly less dense and more brownish structure in the middle as compared with the outer parts.

In order to determine whether all chalazoidites showed a definitely laminated structure, sections were treated with carbol fuchsin and then thin scrapings removed from the surface. By this method it was found that the dye had penetrated more deeply into the central area or nucleus in rather a punctate manner, and that there were concentric stained laminae separated by other non-stained layers.

On treating intact chalazoidites with Canada balsam and grinding down a thin section, it was found on microscopical examination that the deeply staining nucleus and the absorbent concentric laminae were due to their texture being less dense than the non-staining laminae. The dye was able to penetrate more deeply into the nucleus and stained rings than in the non-staining rings. Other methods of staining, such as treating a section with silver nitrate and then applying formalin, shows the absorbent concentric laminae as very distinct black rings. No satisfactory methods were devised for showing up the laminae by staining-methods in the smaller chalazoidites, but they are undoubtedly present as is shown by those specimens that have been exposed to weathering conditions, and also by the occasional presence of a darkly-stained laminae due to iron oxides. From these various methods of examination it was found

– 583 –

that all chalazoidites have the characteristic structure of a nucleus surrounded by concentric laminae, but there were differences in the size and shape and position of the nucleus, and differences in the number and thickness of the laminae. It was quite clear that the laminae themselves differed somewhat in thickness in some places a given layer may be absent and its place taken by outer succeeding layers (see Fig. 8).

Radiological examinations were made to ascertain whether these would throw any light on the structure of the chalazoidite. As has already been mentioned, incomplete fractures were occasionally revealed in the apparently intact chalazoidite. Thin sections of from ⅛ in. to l/16th in. were taken from the thickest part of the largest specimens, and plates taken at varying distances and varying-exposures. From these it was found that the nucleus was always less radio-opaque than the outer portions, and there were faint but-distinct evidences of laminae in the capsular portion. Very often a very much less radio-opaque nuclear point or nucleolus could be-demonstrated. It was small and usually rounded, but sometimes oval and elongated. In Fig. 8 B there are two densely radio-opaque-particles, probably quartz, in the nucleolus. Examples of matrix and crushed chalazoidites were X-rayed, and it was found that the proportion of radio-opaque material was greater in the matrix than in the chalazoidite.

Nucleolus or Nuclear Point: This represents the first stage in the formation of a chalazoidite. That it represents an area of much looser texture than the surrounding nucleus is undoubted, both from the microscopical examinations and from the staining reactions, which sometimes show it up as a small intensely-stained area. Whether it has any peculiar chemical composition was not determined.

Nucleus: The nucleus is always eccentric in position, and varies in size with the size of the chalazoidite. No evidence of its ever being double was found, although one might expect this to occur occasionally. Its relation in mass to the outer capsular portions will be discussed later. With the larger specimens the nucleus showed no evidence of lamination.

Capsular Layers: The number and thickness of these layers-varies with the size of the chalazoidite. The laminated appearance depends on the varying density of the different layers. A layer of less density absorbs the stains more deeply than the layers which have a greater density. These layers appear to alternate regularly unless one portion of a layer is absent. In sections of the chalazoidites it is rarely found that there are oval air-spaces—lacunae—which run in the line of the laminae. These are more frequent in the outermost layers. These may possibly be due to certain soluble constituents having been leached out, or even due to contained air-bubbles during the formation of the chalazoidite.

Outermost Layer: It is clear that in many cases this layer is of much greater density than any of the other layers; in some specimens the surface can be scratched only with a knife, and this extreme density is suggestive of vitrification.

– 584 –

Relationship Of Nucleus To Capsular Layers.

From the evidence of structure shown by staining, microscopic enlargement, and radiological treatment, it is clear that the nucleus is of looser texture and has a higher porosity than the outer or capsular portions. In view of this fact an attempt to determine the relationship in mass of the relatively porous portion was made by estimating the total porosity of chalazoidites of varying sizes. Since chalazoidites of the same approximate weights show differences in structure, large numbers of a given weight were taken for each estimation. The method adopted was inexact, but the results shown in Fig. 4 seem to be fairly uniform. By calculation it is clear that the size of the nucleus or more porous part varies but little, whilst that of the capsular layers or less porous part increases rapidly with the increasing size of the chalazoidites. From 5 mg., the lowest limit of weight of chalazoidite used, to about 70 mg., the porosity falls

Picture icon

Fig. 4.

from 25.3% to 23.25%, which seems to imply that the capsule is increasing very rapidly as compared with the nucleus, and that the masses are comparable. From about 70 mg. onwards the curve becomes more flattened, implying that the mass of the capsular layers now considerably overshadows that of the nuclear portion.

Microscopical Examination.

Mr. L. I. Grange reports as follows: “The constituents are typical of rhyolitic pumice. They consist largely of colourless particles of glass with felspar, quartz, and finer material that was not resolvable.” Other minerals such as rutile and limonite are also present. An examination of a large chalazoidite from the banks of the Waikato near Taupo to determine whether there was any difference between the nuclear and capsular portions showed nothing very marked. “Both consist mainly of glass, comminuted pumice, with a few minute grains of felspar and more rarely tiny prisms of a green mineral and minute black specks. The green mineral seems to be either hornblende or a pyroxene, but the crystals are too small to be identified with certainty. I suppose the minute specks of black to be magnetite, as you have demonstrated that mineral. In the outer shell the glass fragments are perhaps ever so slightly finer in grain than in the nucleus, and the shell shows perhaps more of the flocculent almost clay-like patches, which I suppose represent aggregates of the very finest dust particles.” (Campbell Smith).

– 585 –

Matrix Of Chalazoiditic Layer.

On the hailstone theory it was thought that probably all chalazoidites had their origin at a low and fairly constant level in the pumice-cloud, and that the formation of the larger ones was due to their being carried by vertical air currents into higher levels of the atmosphere where finer volcanic dust predominated, from which they received many of their additional laminae. As seems proved by their porosity and by a study of stained specimens, growth in size depends largely on the increase in the number and thickness of these capsular layers, and not so much on the size of the nucleus. It was reasoned that it was likely that the size of the particles, and therefore possibly the chemical composition of the particles, would vary in the lower part of the matrix as compared with those in the upper part, and still further that the chalazoidites of varying sizes might also show chemical differences. Mr. F. M. Saxton very kindly graded samples from various layers. His report is as follows: “I have numbered the samples submitted as follows:—

  • No. 1.—No. 1 layer.

  • No. 2.—No. 2. layer.

  • No. 3.—No. 6. layer.

  • No. 4.—Matrix containing small chalazoidites.

  • No. 5.—Matrix containing large chalazoidites.

Each sample was first dried and then gently powdered between spatula and paper in order to free concreted fine particles without actually grinding them to a finer state than already obtained. No. 2 was in the form of a single large lump; this was easily friable, and after gently breaking was treated as the other samples.

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

No. 1 No. 2 No. 3 No. 4 No. 5
% % % % %
Retained ¼-in. mesh sieve 2.7
" ⅛" " 1.0 11.2 26.5
" 1/10" " 1.0 12.5 7.2
" 1/40" " 3.0 20.5 41.0 22.4 20.0
" 1/80" " 13.2 35.0 34.6 10.6 10.0
" 1/200" " 15.3 14.6 14.5 8.3 5.1
Passing 1/200 " " 68 5 29.2 8.0 34.9 28.5
100.0 99.3 100.0 99.9 100 0

The total of 99.3% in the case of No. 2 is on account of the very small quantity of material that was available.

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

No. 4: The 11.2% and 12.5% retained up to 1/10 in. mesh consisted almost entirely of chalazoidites, and about half of the 22.4% retained by the 1/40 in. mesh was mostly of intact chalazoidites, so that the total percentage was 34.9%.

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

No. 5: The 2.7%, 26.5%, and 7.2% up to 1/10 in. mesh were almost entirely composed of chalazoidites, 25% of the 20% retained in 1/40 in. mesh consisted mostly of intact chalazoidites, so that the total percentage was 41.4%.”

– 586 –

It will be noted that the proportions of chalazoidites given in these figures differ from the proportions given previously, where it was found to be roughly one half in each case. The difference is explained by the method of crushing between a spatula and paper. This would destroy a large proportion of the smaller fragmented specimens which are easily friable. From these figures it is clear that the greatest proportion of the finer particles is in the upper layer.

No method was found of determining the sizes of the particles entering into the formation of the chalazoidites. As would be expected with any coloured homogenous substance, crushing would render it lighter in colour; or if fractions were separated by sifting, the finer portions would be lighter in colour than the coarser, merely by reason of their fineness. It was very noticeable that the matrix, on being passed through the sieves of varying sizes, became progressively lighter in colour as it became finer. This might in part be explained by the differences in chemical composition. On thorough crushing of a number of small chalazoidites and a number of large ones, the mass from the large chalazoidites was much lighter in colour than that from the small. The differences in colouration appear to be largely due to the quantities of magnetite and ferro-magnesium minerals present.

Estimations of the amount of magnetite in the matrix and in the chalazoidites appeared to be of some importance, and attempts to ascertain these were made by means of the electro-magnet. It was found that no accurate estimations were possible, owing to the fact that the magnetite grains were often imbedded in, or adherent to, other substances. In addition, limonite was freely present and this, too, is feebly magnetic. It is stated that as a rule coarse grains of magnetic substances may be more easily removed than fine grains of this material, since fine grains usually have a larger percentage of non-magnetic material adhering to them, and hence a greater magnetic force is necessary to remove them. The result of these examinations showed that much more material was extracted by the electro-magnet from the matrix than from the crushed chalazoidites, and more from the small than the large chalazoidites.

An explanation of these results is difficult. Pumice dust consists of several minerals of specific gravities varying from magnetite 5.18 and rutile 4.2, to quartz 2.65 and felspar 2.58–2.76. It might appear that there had been some gravitative selection whereby the heavier minerals fell more rapidly than the lighter ones, but bodies of equal mass and equal air resistance will fall with the same velocity from the same height. This is true up to a certain stage of subdivision, but even beyond this the differences in the rates of fall would be very slight, possibly not more than 1/1,000th of the distance travelled. Other factors come into the question, such as the height to which the various constituents would be ejected from the volcano, and, most important of all, the effect of wind, which would have a great influence in causing a gravitative selection. It would be expected that a given layer of volcanic dust traced outwards from the center of eruption could not only show differences in texture, but, from this “gravitative selection” also differences in chemical composition,

– 587 –

the minerals of lighter specific gravity predominating. Whether this has been proved I am unable to learn, but the question could easily be settled by tracing the chalazoiditic layer, or any other well-defined layer of these beds inland and having analyses made of the material at different points. The material ejected from the Tarawera eruption of 1886 should be very suitable for examinations of this sort. It would be expected that the material which floated in the air for such long periods of time following the Krakatoa eruption was very different in some respects from that which fell near the volcano.

In dealing with the magnetite present in the chalazoidites it must be kept in mind that if the amounts could be accurately estimated they would not necessarily bear any relation to the total iron shown on chemical analyses. Other compounds of iron are relatively easily susceptible to solution processes, while magnetite is extremely resistant. It might happen that in a chalazoidite of a given size 50% of the iron might be in the form of magnetite, and the other 50% in other compounds of iron; in a chalazoidite of another size 80% might be in the form of magnetite and 20% in the other minerals.


Another structure met with is a tubule of volcanic ash. This is found in nearly all layers. The fact that in some instances these tubules have intimately adherent to them imperfectly formed chalazoidites is very suggestive that they have an aerial origin. Another possibility is that they may have formed round roots of plants that have grown during intervals between the depositions of these layers, but this does not seem likely. It is noticeable, however, that the roots of recent plants that have died in these beds become encased in a rather dense pumiceous envelope. On looking through the literature, nothing very similar to it appears to have been described. The tubules are of varying lengths, shapes, and sizes. A large majority are branched, and many are adherent to imperfectly-formed chalazoidites. They range in length from 1 mm. up to 8 or 9 mm., and in thickness from 1 mm. to 2.5 mm. Occasionally they are much larger. They show no evidence of vitrification. Tubules of ice have been described as falling during hailstorms, but no very satisfactory explanation has been given to account for them. It seems possible that these tubules are analogous in their origin to the ice-tubules, whatever that origin is. In Fig. 6 B 1 and 2 are tubules. On X-ray examination of 1 the central tube was seen to bifurcate towards the right extremity, and at various points along its length were small openings through the wall, indications of which can be detected in the figure.

Aerial Fulgurites.

Yet other structures that are fairly common in certain localities, as in Havelock Road, are fragments of what at first sight look like silicified wood. These occur most frequently in the chalazoiditic layer, but are found in almost any of the layers. They are from about an inch to a foot or more in length, and from about a quarter

– 588 –

of an inch to an inch or more in thickness. On cross section their central portions show distinct evidence of vitrification, and also a longitudinal tubular structure apparently due to the formation of gases. Very often they are in the form of single large tubes. Imperfectly-formed chalazoidites are often densely adherent to their exterior surface. It seems highly probable that these structures were formed by the electrical discharges that must have accompanied the volcanic eruption and the formation of the chalazoidites. The usual formation of fulgurites is by the electrical discharge entering the ground, so that they are usually more or less vertical in position. It is suggested that these fulgurites may have been formed from fusion of the materials suspended in the air, although it is certainly difficult to visualize such a density of particles as would make a welding of this nature possible. The attitude of these fulgurites is always horizontal, which may be expected from their suggested manner of formation, although it is conceivable that in the loose matrix they would assume this position after formation by the usual method.

Microscopical examination of these fulgurites shows them to have a homogenous appearance with no evidence of any cellular structure.


Mr. Morgan, late Director of the Geological Survey, kindly had the following analyses made by Mr. F. T. Seelye of the Dominion Laboratory.

No. 1 is an analysis of No. 1 layer.

No. 2 is an analysis of the larger chalazoidites from the lower part of the chalazoiditic layer.

No. 3 is an analysis of the matrix in which they were embedded.

No. 4 is an analysis of the smaller chalazoidites from the upper part of the chalazoiditic layer.

No. 5 is an analysis of the matrix in which they were embedded.

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

No. 1 No. 2 No. 3 No. 4 No. 5
% % % % %
Silica SiO2 72.46 70.73 71.01 70.27 69.72
Alumina Al2O3 12.23 12.14 12.18 12.42 12.47
Ferric Oxide Fe2O3 0.95 1.38 1.38 1.43 1.50
Ferrous Oxide FeO 0.94 0.85 0.94 0.83 0.94
Magnesia MgO 0.34 0.56 0.61 0.54 0.63
Lime CaO 1.54 1.59 1.76 1.51 1.74
Soda Na2O 3.65 2.83 2.75 2.87 2.97
Potash K2O 3.01 2.74 2.74 2.63 2.54
Loss on ignition 3.93 4.69 4.43 4 65 4.63
Water lost below 105°C 0.70 2.08 1.94 2.41 2.42
Carbon Dioxide CO2 none none none none none
Titanium Dioxide TiO2 0.20 0.23 0.24 0.23 0.24
Manganous Oxide MnO 0.0 0.04 0.05 0.04 0.05
Sulphur S. 0.01 0.025 0.04 0.025 0.04
Baryta BaO 0.08 0.08 0 08 0.08 0.08
100.10 99.97 100.15 99.94 99.97
– 589 –

In some analyses carried out by Mr. A. E. Aldridge to determine the amounts of silica in chalazoidites of different sizes, he obtained the following interesting figures:—

1. Chalazoidite weighting 840 mg. 69.76%
1. Chalazoidites of large size 68.03%
3. Chalazoidites of small size 67.90%
4. Chalazoidites of differing weights, 5 mg. 64.68%

These differences in the silica-content may be partly explained by the decomposition that has gone on in the outer layer of the chalazoidites. Assuming that the outer capsular layer were equally thick for the small as for the larger chalazodites, there would be a greater proportionate bulk in the case of the former. In acidic rocks the result would be reduction in the content of silica and slight increase of Al2 O3, with increase, of course, of water.

Mr. F. T. Seelye, to whom my thanks are due, has very kindly read through this paper, and in a private communication has discussed the problem of the nature of the cementing-substance. “The analyses show that it cannot be a carbonate nor calcium sulphate. It may, of course, be silica or even clay, but whatever it is now the change would seem to have started at the same time as the pumice was formed, or very shortly afterwards, for it seems almost impossible for water alone to render ordinary pumice dust cohesive enough to resist fracture, even if the mass dropped into a bed of pumice dust from a height. As you mention, superheated steam would indeed have a marked effect on the chemical constitution of pumice, and if gelatinous silica could be liberated by such action on the surface of the particles, this might prove a very effective binding material. I have also wondered if the acids and other gases so often evolved at the time of a volcanic eruption should not also be considered as playing a prominent part in modifying the properties of the finest pumice dust. Jaggar possibly has this in mind, as your reference to his paper shows. That large amounts of acid and other active gases are evolved during many eruptions, and especially at the beginning, seems to be generally accepted. Sulphurous gases seem commonly to accompany explosions, whilst hydrochloric acid and chlorine vapours are common, and are especially associated with high temperatures and energetic action. Even dilute hydrochloric acid solution at 100° C. quite appreciably attacks pumice, yielding chlorides of iron, aluminium, etc., and the attack would probably be much more intense at the very high temperatures prevailing during an explosion. In a recent paper E. G. Zies, calling attention to the vapour-phase activity of an igneous body, states that in the ‘Valley of Ten Thousand Smokes’ in Alaska, approximately a cubic mile of rhyolitic pumice was deposited after having been blown through the old floor of the volley. Many fumaroles are located in the pumiceous area. The persistence of relatively high temperatures (97° C. — 650° C.), and great volumes of steam since the eruption of 1912, indicate that the emanations derive their heat from a deep-seated igneous body. The emanations (now) contain over 99% of steam together with approximately 0.12% of hydrochloric acid, 0.03% of hydrofluoric acid, and 0.03% of H2S. These acid gases

– 590 –

have greatly altered the pumice in the vicinity of the vents of the fumaroles.

“If we suppose the surface of the particles altered by these vapours and soluble salts of the bases formed, together with gelatinous silica, there is a possiblity that such substances might in some way or other, perhaps at a certain degree of hydration, form a cementing medium; if too much water did not condense on the surfaces of the particles.

“If now the particles coalesce and the mass increases in size in the way you have described, then after the deposit has been formed on the earth we would very likely have conditions necessary for the quick ‘rotting’ or decomposition of some of the pumice beds, the salts of iron and aluminium decomposing to form hydrates, and the aluminium ultimately giving rise to claylike minerals such as kaolinite.

“It seemed to me that several facts mentioned in the paper might just possibly be explained by the presence in the pumice of soluble salts formed by the agency of acid gases, these salts still being present on the surfaces of the particles at the time of their deposition; for example, from Fig. 1 the presence of interbedded layers of relatively fresh volcanic material might be explained by the absence of such salts at the time of the ejection of the material. Again, the curious alternation of pumiceous layers with (usually iron-stained) tough highly-resistant layers, the latter seemingly forming a not-very-sharply-separated ‘cap’ to the pumice layer, suggests the idea that each resistant layer might have been formed largely from material derived from the lower layer; for example, by the salt solutions working upwards and depositing their salts at the surface (just as this occurs in dry irrigated areas, e.g., Central Otago). Certain salts of iron and aluminium are easily decomposed and would give rise to hydroxides, and even to clay-like material possibly. (In the neighbourhood of geysers the aluminium is ultimately deposited as alumite or as kaolin.)

“It is significant, too, that the chalazoidites apparently have the toughest and most resistant layer at their surface. Facts in favour of this are:—


The cementing of the ‘tail’ has already been attributed to such a cause.


Iron oxides (secondary and hydrated?) are at least occasionally present in the laminations of the chalazoidites.


The nucleus (and nucleolus) and the absorbent parts of the laminations might just possibly consist of secondary material produced by the decomposition of the salts, or of gelatinous silica liberated during the formation of the salts—it is indeed just the gelatinous silica, of course, that takes the stain in staining an ordinary rock slice.


The very hard and resistant outermost layer of many chalazoidites suggests, as mentioned before, the possibility of this having been formed in a similar way to the ‘cap’ of the pumice layers, and in Campbell-Smith's descrip-

– 591 –
  • tion of the microscopic properties of a chalazoidite he refers to the patches of clay-like material in the outer layers.

“All this is purely speculative on my part. However, I thought it might be worth while further examining some of the samples you sent in for traces of chlorides—sulphur not being strongly in evidence as shown by the analyses. I tested samples 1, 2, and 3 only. Sample 1 had very little chloride soluble in water, but samples 2 and 3 gave very strong tests indeed. However, such chloride would most likely have been derived from the salt spray carried from the sea by the wind over Scinde Island. After removing this chloride, practically no more could be extracted by hot dilute nitric acid, but I was rather surprised to get relatively large amounts of chloride again from all the residues, after decomposing the latter with hydrofluoric and nitric acids, showing that some difficultly-decomposed chloride mineral is present in all. The tests were only qualitative on 1 gm. of sample, but I should think that the amounts of this latter chloride were of the order of 0.1% or somewhat more. Of course, chlorides are not unknown in acidic lavas such as rhyolites and obsidians, but are usually present in traces. But an obsidian from Iceland, for example, is recorded which contains 0.1%, and one tested recently in the Dominion laboratory had somewhat more than this. Further, it is possible, I suppose, for sea-spray salt to have formed a complex mineral with the pumice. So the evidence is quite inconclusive; nevertheless, it would be interesting to know if chalazoidites from other districts, and particularly from inland districts, contained relatively large amounts of chloride or of sulphur.”

In order to settle this point with regard to the large amount of chlorine I forwarded to Mr. Seelye samples from a recently-exposed section close to the entrance of the Napier Hospital, and also some chalazoidites from Taupo. The results of the analyses which he kindly undertook are as follows:—


Unbroken chalazoidites from upper part of chalazoiditic layer.


Matrix of same.


Unbroken chalazoidites from lower part of layer.


Matrix of latter.


Chalazoidites collected in Taupo.

Chlorine in Water—Soluble Chlorides.
1. 2. 3. 4. 5.
CL. 0.04 0.03 0.03 0.03 0.02 per cent.

Chlorine in chlorides decomposed by dilute nitric acid:—none.

Chlorine in chlorides decomposed only by hydrofluoric acid:—

CL. 0.11 0.11 0.12 0.12 0.11 per cent.

Mr. F. T. Seelye in commenting on these results states that in the published analyses of pumice the presence of chlorine is scarcely mentioned, and probably in most cases has not been looked for.

Washington in his tables of all the published analyses of igneous rocks has only about twenty-six analyses of pumices of various types.

– 592 –

In one vulsinite pumice (silica 5.8%, alkalies 11.8%) a trace of chlorine is present, whilst two analyses of lapilli ejected during the Tarawera eruption show 0.05% and 0.4% of chlorine respectively (S. P. Smith, “Eruption of Tarawera,” 1887, p. 76).

On re-examination of a sample of pumice recently sent in by the Geological Survey, Mr. Seelye found only a trace of water soluble chloride and 0.08% in form only decomposed by hydrofluoric acid.

The presence of chlorine in the Tarawera lapilli, in the obsidianite mentioned above, in the Rotorua pumice, and in the present specimens makes it seem quite possible that this element is of widespread occurrence at least in the North Island volcanic district.

The larger chalazoidites from the lower layer were tested for the presence of fluorine. “It is certainly not present in more than very small amounts, and as the method of estimations and even the detection of very small amounts is not very satisfactory, one would hesitate to say that it was present at all.”

Dealing with the complete analyses Mr. Seelye states:—

“I have calculated the results of the analyses made here on the five samples into the amounts of those Standard (or Normative) minerals used by many petrographers in calculating the ‘norm’ of a rock from the analyses. Somewhat similar calculations are used in calculating the theoretical or standard mineral composition of a clay from the analyses. Further information about a rock or clay can sometimes be gained in this way, but the results are to be used with caution. The standard minerals are not necessarily those actually present. I am enclosing the results on a separate sheet. The amounts of silica and of the various felspars may probably be taken as fairly representative of the amounts of these minerals actually present, or potentially so in the pumice glass. The amounts of corundum (Al2O3), may possibly here be taken as indicating the relative amounts of clay produced. Beneath the table are given the amounts of the clay molecule kaolinite corresponding to these amounts of Al2O3. On this assumption, decomposition has proceeded to a far greater extent in the chalazoidites themselves and in their matrices than in layer 1. The differences in the amounts of (Na2O+K2O) between sample 1 and the other samples might then also be accounted for in this way, as also the less amount of H2O in sample 1. The amounts of magnetite (FeO, Fe2O3) are not here truly representative—no doubt, much of the Fe2O3 of this ‘normative’ mineral is really in the form of hydrated oxide of iron, while some of its FeO, and also some of the ilmenite (FeO, TiO2) is much more likely to be present together with the normative hypersthene in the green and greenish-brown ferro-magnesian minerals that are actually present in all the samples.

“Both the analyses and the ‘norms’ show that the tendency is for the chalazoidites to have rather less of the bases, especially FeO, CaO, MgO, and MnO, than the corresponding matrices, but the differences are very small. It would seem to suggest the greater attack and removal of ferromagnesian minerals in the chalazoidites.”

– 593 –

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

Results of Analyses of Nos. 1–5 Calculated to the Usual Standard or “Normative” Minerals.
1 2 3 4 5
Quartz SiO2 35.58 39.78 39,72 39.18 37.98
Orthoclase K2O, Al2O3, 6SiO2 17.791 16.12 16.12 15.57 15.01
Albite Na2O, Al2O3, 6SiO2 30.92 Felspar 56.49 23.58 Felspar 47.76 23.58 Felspar 48.32 24.63 Felspar 47.71 25.15 Felspar 48.78
Anorthite CaO, Al2O3, 2SiO2 7.78 8.06 8.62 7.51 8.62
Corundum Al2O3 0.10 1.63 1.53 2.04 1.73
Hypersthene × FeO, y MgO, (x+y) SiO2 1.43 1.40 1.63 1.40 1.73
Magnetite FeO, Fe2O3 1.39 2.09 2.09 1.86 2.09
Ilmenite TiO2, FeO 0.46 0.46 0.46 0.46 0.46
Haematite Fe2O3 0.16
Water (total) 4.63 6.77 6.37 7.06 7.05
Minor constituents 0.15 0.15 0.17 0 15 0.17
100 23 100.04 100.29 100.02 99.99

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

The amounts of corundum, Al2O3, would correspond to the following amounts of the clay mineral:
Kaolinite Al2O3, 2SiO2, 2H2O (The SiO2 for this mineral would have to be deducted from the Quartz of the norm, and the necessary water from the total water.) 0.26 4.13 3.87 5.16 4.39

Site Of The Volcano Or Volcanoes.

The question as to which volcano or volcanoes were concerned in the formation of these pumice-beds is one of very great interest, and there can be little doubt that it will be answered and throw much interesting light on the past volcanic history of the North Island. The most important method is to determine exactly the distribution of the chalazoidites. Although it is extremely probable that other periods of volcanic activity have been associated with their formation, if corresponding beds (but thicker than those in Scinde Island) could be traced amongst the vast pumice deposits of the centre of the island, and those along the East Coast, the evidence would be exceedingly valuable.

Chalazoidites occur at various points along the Napier-Wairoa road, at Tangoio the pumice fragments are larger than in Scinde Island, along the Napier-Taupo road, along the Napier-Taihape road, on the Napier-Puketitiri road near Patoka, and on the foothills of the Ruahines. Mr. Hill; since the reading of this paper, has drawn

– 594 –

my attention to the fact that large specimens up to 2.5 cm. in diameter are met with on the banks of the Waikato. An examination of the cutting on the road leading from Taupo, just before the bridge over the Waikato is reached, shows them to be sparsely scattered through a matrix of coarse pumice fragments. Fig. 6 shows some specimens from this locality. They are identical in all respects with specimens from Napier, except that they are larger. Nothing similar to a chalazoiditic layer such as occurs in Scinde Island was found, either here or on a cursory examination in the surrounding district. There were conglomerations of masses of pumice-dust

Picture icon

Fig. 5.

several inches across in the midst of the layers of coarse pumice-particles. Of more than passing interest is the presence of water-worn greywacke pebbles in these beds. Mr. L. I. Grange has found chalazoidites at several points in the Rotorua district. Professor Bartram points out that Hewitt (9) in his account of the geological features of the Arapuni district, states that

– 595 –

near the diversion tunnel some small concretions have been noted in the soft pumiceous silt. They are described as being about the size of marbles, and form in little clusters which occur in definite planes of the silt. It is suggested that they were probably formed during floods in sluggish waters by the rolling along of initial sand-grains with accretion of smaller particles. I have not had the oportunity of examining these so-called concretions, but it seems more than likely that they will be found to be chalazoidites. Further examinations will undoubtedly show them to have a widespread distribution. When this distribution has been worked out in detail it will be possible to locate the volcano responsible. In at least one place on the Napier-Rissington road the various layers shown in Fig. 2 can be identified, including the chalazoiditic layer, but it contains no chalazoidites. This is certainly not due to any process resulting in their disintegration, but is explained by the fact that this part of the volcanic cloud contained insufficient water to form chalazoidites, or that some other condition necessary for their formation was absent.

It seems that a study of the prevailing winds might help in locating in which direction the volcano was situated. Fig. 5 is a diagram showing the winds in Napier and their relative frequency of duration. It will be seen that the westerly winds predominate, and if the winds in later Pleistocene times were the same as at present, and there is no reason to suppose any great differences, it would be reasonable to assume that the volcano was situated to the west or north-west of Napier. This would point to the present centres of volcanic activity. The farthest south I have as yet encountered the chalazoidites is towards Moteo, but they extend much farther in a northerly direction. It is clear that even with a constant west wind the pumice deposit in which the chalazoidites occur would spread fanwise from the erupting volcano, so that taking a point due west of Napier would not necessarily locate the situation of the volcano; it could equally be west of points many miles farther north, and several miles farther south. When the depth of these deposits at various points, together with the distribution of the chalazoidites of this particular era of volcanic activity, are approximately known, it will then be possible to locate the volcano with a fair degree of accuracy. The identification of deposits elsewhere must be based on continuity with beds on the mainland having a similar order of deposition to those in Scinde Island, and to a considerable extent on the presence of chalazoidites. If in addition a diagram showing the winds at the suspect volcano could be applied to a map of the North Island, the distribution and thickness of these pumice beds would be found in agreement.

It might appear that each different layer in Fig. 2 owed its origin to a different volcanic outburst, or even that several volcanoes were concerned. It seems unnecessary to assume more than one volcano, or perhaps a group of two or more close together, discharging very similar material. The difference in texture of the layers, and even chemical composition, could easily be explained on the theory of only one volcano. If a volcano discharged perpendicularly a vast quantity of pumice-dust high up into a relatively still atmo-

– 596 –

sphere, forming the well known “mushroom” volcanic cloud, it is clear that, with the crater as a centre, radii could be drawn in all directions, and that at any given point along any radius the coarseness of the pumice-dust would be the same as at any other similar given point along another radius. Let us assume that layer 1, Fig. 2, represents the deposit that would be formed on the 80 mile circle— a distance certainly too great for the deposit of material from the “mushroom” eruption suggested, but since this is roughly the distance of Scinde Island from the present centres of volcanic activity it will do for the purpose of analogy. If now the volcanic eruption increased in violence the lighter particles would be carried farther, and at the 80 mile circle layer 2, a coarser, and for an equal duration of time, a stratum of greater thickness, would be deposited. Similarly, a diminution in intensity of the eruption would result in the formation of a thinner and finer stratum of material. A still more important factor would be the variations in force and changes in direction of the prevailing winds. In addition, changes in the chemical nature and in the sizes of the particles of the material ejected might occur at various phases in the volcanic activity, although this need not be invoked since the selective action of the wind on the various constituents of pumice-dust would be sufficient. The presence of varying quantities of super-heated steam would, however, have a very marked effect on the chemical constitution of the pumice, and this is probably the reason for the differences in the relationship of the amount of magnetite to hydrated oxides of iron, such as limonite, in the various layers. It would also undoubtedly accelerate the change from pumice into the dense layers of pumiceous clays. The effect of varying quantities of water adherent to the particles of pumice on their rate of descent would also be important. Still further, it is likely that periods of intermission of activity would occur, and during these intermissions alterations due to weathering conditions would take place more rapidly in the most superficial layers deposited. For these reasons there seems no need to postulate two or more volcanoes at widely separated points, as with the various factors mentioned all the differences in texture and in chemical composition could be amply accounted for. At present it cannot be definitely stated that more than one volcano was not engaged in the process of formation of these layers, although it seems unlikely and unnecessary.

Duration Of Volcanic Disturbance.

In studying these volcanic deposits it is interesting to speculate as to what length of time was taken in their formation. Was it a period of time covering centuries, or were they deposited comparatively rapidly in the course of a few months? It seems that the latter estimate is the truth. Certain of the layers, such as the chalazoiditic layer, were probably formed in the course of a day or two—perhaps even in a few hours. That there were periods of intermission seems beyond doubt. Apart from the fact that intermissions would be expected during the course of the deposition of beds comparatively so vast, probably 80 miles from the centre of activity, there is evidence from an examination of the beds them-

– 597 –

selves that intermissions actually occurred. The unusual constancy of the layers throughout the whole area of the island seems to be strong evidence of a comparatively rapid deposition. If each different layer were due to a distinct volcanic outburst separated by long periods of time it would be expected that weathering influences would have altered their nature and their stratification to a considerable extent. Certain layers, under the influence of weathering conditions, have become converted into extraordinarily tough reddish clay, which in exposed sections at the present time are particularly resistant to denudation. These, it is suggested, correspond to intervals of intermissions. On this view the first interval of intermission is represented by layer 5. The next would correspond to layer 9. This particular layer merges gradually into the underlying chalazoiditic layer, but it is sharply demarcated above, and its surface presents markings which seem to indicate the action of running water. In the lower part of this layer chalazoidites are firmly imbedded, and since these are of aerial formation there can be no doubt that there were chemical changes from pumice to clay subsequent to their deposition. These chemical changes, it is suggested, would not have taken place except for the influences of atmospheric and aqueous agencies, continued over a period of some time. Layer 13 also corresponds to an interval of intermission, and sometimes one or more beds of clay are found in the pumiceous ash above. These beds, corresponding to periods of intermission, are red or brownish in colour owing to some hydrous oxide of iron (probably limonite), and magnetite does not seem to be present to any extent. Since magnetite is extremely resistant to ordinary weathering conditions it seems probable that the change actually took place in the volcanic crater or in the pumice cloud, and not to any extent in situ. There are other iron-stained layers which do not have the nature of pumiceous clay. As has been pointed out, the upper layers of these beds are inconstant, so that probably intervals of intermission were longer, or the nature of the beds rendered them less liable to be converted into pumiceous clays, so that they were more subject to denudation.

Moa-Footprints And Moa-Remains.

During the course of excavations for the high reservoir on the Bluff Hill, Mr. Hill found footprints of the Moa in these beds, possibly in one of the layers which represents an intermission. These footprints may be used as evidence that intermissions did occur, and that some type of vegetation must have developed upon which the moa subsisted.

The evidence of Mr. R. H. Finch in dealing with the chalazoidites of Maunaiki, as will be mentioned later, shows that human beings would not necessarily be overwhelmed by the falling volcanic ash, since the impressions of bare feet have been preserved in this layer. The moa would therefore not necessarily be destroyed at once, but would die from starvation.

– 598 –

Method Of Formation Of Chalazoidites.

In considering the method of formation of chalazoidites there appear to be three possibilities:


They may have been formed in the crater of the volcano.


They may have been formed in situ.


They may have been formed in the air, either above the volcanic crater or some distance from it.

The first possibility may be excluded at once. They are not “lapilli” and show no definite evidence of vitrification, except possibly in the outermost layer. In dealing with the second possibility Pratt states, “That these aggregates have not resulted from solution processes is evidenced by the facts that they contain no calcium carbonate nor any other extraneous cementing agent, and that these beds in which they occur have certainly not experienced metamorphism.” The question as to whether they were spherulites which had undergone changes coincident with or following on the disintegration of the parent rock was considered. The absence of the radial structure typical of spherulites, and the nature of the beds themselves, overruled this possibility.

Dr. Ohashi, of the Akita Mining College, Akita, Japan, states in a letter that he has studied the method of formation by the following simple experiment. Fine, dry volcanic ash is put on to a dish and then a drop of water is allowed to fall on its surface. By rolling the drop of water about a chalazoidite is formed. I had tried this experiment previously with fine, dry volcanic ash from Scinde Island, and although it is true that a small globular body can be made, it lacks the cohesion and the structure of the true chalazoidite. In the upper part of layer 6 chalazoidites occur intermingled with coarse particles of pumice. It is clear that on the explanation suggested their formation is impossible unless they had been transported by water or wind action from elsewhere, which, in view of the constancy of the layers and their contour following that of Scinde Island, is impossible. There are also the differences in chemical composition of the matrix and the chalazoidites, part of them probably being primary and part due to weathering influences, which would be difficult to account for on this theory. There is also the final and conclusive evidence of Dr. Hovey (9) who found them plastered on the walls of the houses in Precheur following the Martinique eruption, and also the statement of Pratt, already quoted, who has proved that they keep their form on falling into water. It may be taken that the usual method of formation of chalazoidites is not that described by Dr. Ohashi, although it seems a possible explanation under certain conditions. A chalazoidite when once formed, might conceivably shortly after its formation collect further layers in the manner described by Dr. R. Ohashi, and in fact the “tail” that is often present might be evidence that such accretions could occur, although this is not intimately adherent.

Mr. Guthrie-Smith (10) has an interesting paragraph in his book, Tutira, in describing the formation of what he called pilules due to raindrops. “About the bases of bare scarps—the unnealed sears of hillside slips—quantities of the finest dust accumulate in dry seasons. On these miniature screes of powdered soil fall the

– 599 –

first great drops of a western shower. The dust slope can neither retain the drops nor instantaneously absorb them. Striking the slope they gather earth particles on their downward course. While thus in motion, as if by miracle they change from liquid to solid. Metamorphosed first into ashen grey, and then into brown balls, these earthen pilules, preserving their shape but changing their substance, race madly downhill, bound downhill no longer clear drops from heaven, but minute circular solid globes of soil. With a faster fall of raindrops the process ends perforce; the dust heap becomes a mud torrent.”

Hailstone Theory.

(a) Theory of Ascents and Descents: We are now left with the last and only possibility—viz., that they are of aerial formation. One theory I wish to suggest is that the formation of a chalazoidite is analogous to the formation of a hailstone. Quoting from the Encyclopaedia Brittanica, 1st edition, “All hail is probably connected immediately with whirlwinds more or less developed, and it is when the hailstorm is one of the phenomena attendant on the tornado or on a great thunderstorm that it assumes its most destructive form. The theory of the formation of hail has been stated by Ferrel in his Meteorological Researches for the Use of the Coast Pilot. The vapour carried aloft by the gyrations of the tornado is below a certain height condensed into cloud and rain, but above that height into snow. Let the raindrops formed below be carried up into the snow region by the powerful ascending currents of the tornado and be kept suspended there a little while and they become frozen into hail. If now these be thrown quite outside the gyrations of the tornado they fall to the earth as a shower of compact homogeneous hailstones of clear ice of ordinary size. If, however, they are caught in the descent and carried towards the vortex by the inflowing current on all sides they are again rapidly carried aloft into the freezing region. A number of such revolutions of ascent and descent may be made before they fall to the earth. While high up in the snow regions the hailstones receive a coating of snow, but while traversing the region lower down where rain yet unfrozen is carried up they receive a coating of solid ice. Thus, alternate coatings of snow and ice are received, and the number of each sort indicates the number of revolutions described before the hailstones fell to the ground.”

In the various descriptions of hailstones their size and structure vary in the same way as the chalazoidites.

At the time of the deposition of the chalazoiditic layer there was a vast cloud of pumice over the whole district, including Scinde Island. In this drifting cloud the coarser and heavier particles would be in process of descent at a more rapid rate than the finer and lighter particles at a higher level. The temperature would be relatively high, as pumice is a poor conductor of heat. Each particle of pumice at what might be called the “water-level” in the cloud would be surrounded by a thin film of water. In addition the capacity of air for holding vapour would increase with the temperature. At 30° F. the maximum weight of water absorbed by air in

– 600 –

grams per cubic foot would be 1.94, at 70° F. 7.98, while at 100° F. it would be 19.7. If as a result of a decrease in temperature of the cloud due to dissipation of its heat during its journey towards Scinde Island, or to the meeting of colder currents of air coming from the sea—the land breeze—then condensation of its moisture would occur. These particles of pumice with their moisture under the influence of surface-tension would tend to coalesce to form a nucleolus, and this in turn would aggregate to itself other particles to form a nucleus. This would account for the relatively loose structure of the nucleus. If the graph in Fig. 4 is accurate it is possible to infer that chalazoidites up to about 70 mg. in weight may have had their origin at roughly the same level in the cloud of water-charged pumice-dust, and that the variations up to this size were dependent on the amount of water-vapour immediately available. Chalazoidites more than about 70 mg. in weight owe their size to additional laminae the dense ones possibly composed of finer constituents possessing a lesser specific gravity, and probably with lesser quantities of water, whilst the less dense ones had coarser dust and a greater proportion of water. It is even possible that the particles of pumice-dust forming the nucleus and capsular layers were originally of about the same size, but those that were in the lower part of the cloud were there, partly owing to an increased amount of water, accelerating their descent. This idea alone would explain satisfactorily the differences in density of the laminae, but not the differences in chemical composition, assuming that these differences were present at the time of formation of the chalazoidites. It would not explain the fact that chalazoidites below 70 mg. in weight are all definitely laminated in structure. Such factors as the varying specific gravity of the minerals present, the size of the constituent particles, and the amount of water-vapour at various levels, would be operative and need consideration. If the small chalazoidites were carried by an ascending current of air into a layer of more finely-comminuted dust, or dust with a lesser amount of water-vapour, then the dense lamina would be accounted for. If they then fell back through the cloud into the “water-level” a coating of dust, each particle surrounded by water, would result in a lamina that was not so dense. If this process were repeated several times the complete chalazoidites would be accounted for. The essential difficulty of this theory is that if the ascending current were strong enough to raise the partly formed chalazoidite to a higher level it would also carry the coarser particles that composed it to the same level.

For the purpose of testing this theory it was thought that estimations of the amounts of magnetite would be valuable. It would be expected from its more rapid descent, due to its greater specific gravity, that it would be in greater abundance in the matrix than in the chalazoidites, and greater in amount in the small ones than in the large ones. On testing, as explained previously, this actually appeared to be the case. If the possibility of the leaching out of certain constituents, or the possibility of the addition of others were excluded, the differences would be accounted for on the theory that has been outlined. Examination of a large chalazoidite from Taupo

– 601 –

showed the glass-laths to be finer, and the flocculent clay-like particles to be more numerous in the capsular layers than in the nucleus. This again would be valuable evidence in favour of the theory, but these capsular layers are most liable to the decomposition that would produce a similar appearance.

(b) Theory of Direct Descent: Ordinary summer hail is merely frozen raindrops, so that this method of formation for the chalazoidites may be invoked. It could be assumed that the chalazoidite commenced as a central nucleus at a level in the pumice-cloud where water was abundant, and that in its descent it passed through various clouds of pumice-dust having varying humidities, and that these each in turn gave alternately a dense or less dense lamina. Another possibility is that a “drop” of water formed around a nucleus of pumice particles and that in its descent it collected films of pumice dust which would result in laminae. On this theory the sizes of the chalazoidite would be dependent on the following factors:—


The height of the water-charged pumice-cloud in which it originated, which would be dependent on the temperature.


The quantity of water and the amount of dust immediately available in this cloud to form a larger or smaller nucleus.


The number and thickness of pumice-clouds through which it passed in its descent.


The relationship of the amount of pumice-dust to the amount of water necessary to form a dense or less dense lamina, or no lamina at all, in passing through these clouds.

Rather against this theory of direct descent are the more or less regular alternate laminae of dense and less dense rings. This would seem to presuppose that clouds containing varying quantities of water occurred at alternate and fairly regular intervals, and that the difference in density of the laminae is dependent on the varying amounts of water surrounding the particles of pumice. A still further curious characteristic is that the laminae appear to be of rather constant thickness, whether dense or less dense. It seems very unlikely that alternate clouds of pumice would occur, as required in this theory. In dealing with this question it seems clear that the smaller chalazoidites do not represent the nucleus of the larger chalazoidites, since they present evidence of lamination, indeed the nucleus of the larger chalazoidites appears to be homogeneous except for the nucleolus.

It seems even possible that both theories of formation may have been operative. The upper part of the chalazoiditic layer, where the chalazoidites are of such constant size, shape, and structure, suggests that their manner of formation was identical, and that the theory of direct descent might be the correct one. In the lower part of the layer, where there is such great divergence in size, the conditions of formation were possibly not identical, and the theory of ascents and descents may be correct. Even then the differences in the amount of water available and the size of the initial mud ball

– 602 –

may be explainable on the theory of direct descent. The drops in a sudden tropical downpour of rain are often first of large size, but later become of uniform size. It is, however, clear in Scinde Island that the first drops were small, and were succeeded later by a mixture of large and small ones, and again later by small drops. Then, of course, the sizes of the drops may have varied in different localities even a comparatively short distance apart. In favour of this theory is the fact that the particles entering into the formation of the nucleus of both large and small chalazoidites appear to be about the same size, and appear to have the same porosity. The argument of the fineness of the constituents of the capsular layer, and the differences in the amount of magnetite cannot be used, as they may equally be due to decomposition influences. The only method of obtaining any satisfactory evidence on this point would be the examination of recently fallen chalazoidites if they were available.

The question now arises as to whether these chalazoidites were formed in the air above Scinde Island or actually over the mouth of the crater. In either case one or both methods described may have been involved in their formation. If they were formed above the mouth of the crater there would be no difficulty in their transportation, as pieces of pumice (even in the dry state) as heavy as the large chalazoidites are abundant in other parts of these layers. There are, however, no coarse fragments of pumice in the chalazoiditic layer, and it is clear from the table showing the gradings that all the material reaching Scinde Island at that time was of fine texture. When the distribution of the chalazoiditic layer is known farther inland it will almost certainly be proved that the pieces of pumice will become progressively larger as the site of the volcano is approached. Were this evidence available it would be proved from their weight that the chalazoidites were formed over Scinde Island. As the position stands at present there is very little doubt that this is what actually occurred. The possibility that the chalazoidites were formed above the volcanic vent, and that the extraordinary density of the capsule is due to the effects of heat and not due to solution processes would supply a simple explanation for their cohesion, but for the reasons already given this seems unlikely for Scinde Island specimens. There is no reason, however, why this explanation should not apply in some cases, particularly to those chalazoidites formed above and falling in the vicinity of the erupting volcano. One other possibility as an explanation for the semivitrification of the outermost layer of some of the chalazoidites has to be considered. The pumice-cloud, as already pointed out, would be at a relatively high temperature. It seems probable, with the high-tension electric conditions present, that more heat might be generated from electric discharges round about some of the chalazoidites after their formation, and might result in their outer coats being semi-vitrified. This would explain why some at least retain their cohesion even on falling into water. The evidence against this is, however, conclusive. Any explanation for the cohesion must be universally applicable—the above is not, since not all chalazoidites show evidence of semi-vitrification. In the Martinique eruption the chalazoidites were found plastered against the walls of the houses,

Picture icon

Fig. 6.
A. Chalazoidites of varying sizes from Scinde Island.
B. On the left are Chalazoidites from Scinde Island. The one on the extreme left shows a flattened area, the next shows the “tail” of matrix, the next is a Chalazoidite showing a flattened area due apparently to the impact on striking the ground while on the other aspect is a concave area possibly due to the impact of another falling Chalazoidite, 1 and 2 are tubules; 1 is of unusual size; 3 shows a concave area in the matrix uniting two Chalazoidites; 4 shows two partially disintegrated Chalazoidites with central holes representing the nucleus in each case, while the other is a specimen of the peculiar oat-shaped forms.
C. 1 is a fulgurite with Chalazoidites intimately adherent. The fractured ends show vitrification. The remaining Chalazoidites are from Taupo, the two on the left and the one on the extreme right have been stained with eosin.

Picture icon

Fig. 7.
Chalazoidites from Scinde Island somewhat enlarged which have been sectioned and stained. The one in the centre shows the meniscus to which attention has been drawn in the text.

Picture icon

Fig. 8.
(A) Enlargement of chalazoidite at right end of row C, Fig. 6. (B) An X-ray photograph of a section of the same chalazoidite. A great deal of the detail shown in the X-ray film has been lost in printing.
1. The nucleolus in both photographs. In the X-ray two densely radio-opaque shadows are present in the translucent nucleolus. In other X-rays the nucleolus is always translucent, and densely radio-opaque material has not been noted before.
2. A densely-staining lamina which is indicated in the X-ray as a less radio-opaque lamina. The line in A has passed too far towards the centre.
3. An incomplete fracture formed either at the line of impact or more probably during the process of drying.

– 603 –

and in the island of Maunaiki the chalazoidites have been found squashed out by human footprints. A study of their shapes in Scinde Island and elsewhere is all in favour of most, if not all, being in a plastic condition on striking the earth. The semi-vitrification, whether a primary condition or a secondary condition due to solution processes, is not a sufficient explanation of their cohesion.

There are at least three essential conditions necessary for the formation of chalazoidites:


The presence in the atmosphere of pumice-dust in a fine state of subdivision, and in a certain degree of concentration.


The presence of a considerable proportion of water.


Certain chemical conditions must be present to allow of the chalazoidites becoming coherent on drying.

Exactly what is the relationship between these factors offers a very interesting field for further research. There is no evidence to show that all the water in the pumice-cloud was used up in the formation of chalazoidites; if so, a considerable amount may have descended with the pumice-particles that constitute the matrix. There is possibly another factor involved in the formation of the chalazoidite—viz., the presence of magnetite. In chalazoidites from Vesuvius, as would be expected from the basic nature of the material ejected from that volcano, magnetite is present in larger quantities than in Scinde Island specimens.

After the formation of the chalazoidite one other factor has to be considered—viz., the surface on which they fell. “It is probable that only when they fall on soft unconsolidated beds of recently-fallen tuff is their form preserved under sub-aerial conditions.” (Pratt).

Literature Dealing With Chalazoidites.

It is an extraordinary fact that in spite of the large amount of literature that has appeared in the past concerning Scinde Island geology, no mention has, up to the present, been made of chalazoidites or of their occurrence in other parts of New Zealand. Mr. W. Kerr drew attention to them in an address to the H.B. Philosophical Society about 20 years ago, but regarded them as having been formed by the action of rain drops following on the pumice dust. There is no doubt that they are of widespread distribution, not only in New Zealand but in other parts of the world, but references in the literature are very few. At the time of reading this paper I was unaware of any previous work having been done on this subject, and it was not until I communicated with Dr. J. Marwick, Government Palaentologist, to whom I am deeply indebted for many useful suggestions, and to Mr. L. I. Grange, Government vulcanologist, that I learned of the existence of the literature, the following references to which they supplied me. As the subject is new to New Zealand and seems worth much further study, I propose to give rather full extracts from the papers that are accessible. The references are:— Hovey in American Journal of Science, 14 (1902); Lacroix, La Montagne Pelee, Paris (1904); Pratt, Journal of Geology, vol. 24 (1916); Jaggar, Bulletin Hawaiian Volcano Observatory, 9 (1921);

– 604 –

and Perret, “The Vesuvius Eruption of 1906,” Carnegie Institution Publication No. 339 (1924).

The account of Dr. Edward Otis Hovey (11), who observed these “drops of mud” after the eruptions on Martinique in 1902, is as follows:—

“In addition to the showers of dry dust and ashes, there fell during the eruption an immense amount of liquid mud which had been formed within the eruption cloud through the condensation of its moisture. This mud formed a tenacious coating over everything with which it came in contact. The drops of mud, too, formed in the air and fell as a feature of the eruption, is proved by the conditions of the walls of the houses in Precheur, on which I found flattened spheroids of dried mud which could have formed only in the manner indicated. These flecks of mud were two, four, and even six inches across, where two or more had coalesced. They occurred mostly on the northern and eastern walls of the houses. The testimony of the people as to the occurrence of rain during the great eruption is conflicting, but the evidence of the coating and these drops of mud proves that much aerial condensation of steam accompanied these outbursts.”

As previously pointed out, this account is the most conclusive as to the actual method of formation of chalazoidites.

Wallace E. Pratt, in an article entitled, “An Unusual Form of Volcanic Ejecta” (7), has given the most complete account hitherto of the chalazoidites. In the course of a study of the eruption of Taal Volcano in south-western Luzon, Philippine Islands, during the month of February, 1911, when 1,335 people were killed, he commented on the presence of spherical bodies in the ash fall at the time as follows:—

“An interesting feature of the fall of the ejecta is the formation of drops or balls of mud. These were observed most abundantly on the island itself, but were seen at Talisay and Banadero also. They range in size from large shot to hazelnuts, and when broken sometimes show concentric markings. Apparently they fell late during the activity, being found just below the surface of the deposit. These mud balls cannot be classed as lapilli in the strict sense of that term, since they are built up, probably through the condensation of steam into drops of water.”

It will be noted that the chalazoidites are larger than any found in Scinde Island, and that some presented concentric markings. It is certain that staining methods would show this structure to be universal, a fact which Pratt later regarded as characteristic. In a figure of a vertical section taken on the south-west slope of the volcano his chalazoiditic layer is about five inches in thickness; in the lower part the larger chalazoidites occur, while in the upper part are smaller chalazoidites of more nearly uniform size, an arrangement similar to that found in Scinde Island. The two layers are separated by fine volcanic ash.

Later chalazoidites were found on the slopes of Mount Maquiling, an extinct volcano 20 kilometres north-west of Taal. These, in rare specimens, attained a diameter of 4 centimetres, thus being comparable in size with the chalazoidites observed by Dr. Hovey.

– 605 –

On Bondoc Peninsula, Tayabas, and near the Santa Lutgarda iron-mine at Angat Bulacan Province, widely separated parts of Luzon, he found chalazoidites in slightly-indurated tuff dating back probably to the late Miocene. In his description he describes the chalazoidites as volcanic hailstones, and mentions that they are dependent on volcanic disturbances in which large quantities of water-vapour with fine pumice-dust are ejected.

The reference by Jaggar is entitled “Fossil Human Footprints in Kau Desert.”

In the spring of 1920 during visits of Maunaiki, Mr. R. H. Finch discovered the prints of naked human feet in old beds of volcanic ash about 6 miles from Kilauea.

“In current exploration of the desert these ancient trails have been photographed and knowledge of them in increasing. The prints are preserved by solidification of the ash mud through the agency of a carbonate or sulphate crust. This is the pisolitic ash which increases in thickness nearer to Kilauea. The pisolitic spheres give evidences of mud rains, and the footprints have invariably been found in the layers showing the solidified raindrops. The squashing out of the mud from under the bare feet is shown in the hardened impressions (Figure 22)…. There is no possibility of these footprints being modern, for the natives of this high country rarely go barefoot, this ash does not make mud patches to-day, and the solidified shell in which the footprints are lithified was a product of the acid volcanic gases, mingled with mud rains, which were distinctive in the 1790 eruption.”*

This is an extremely interesting account, as the flattening out of the chalazoidites under the bare foot shows that they must have been very soft when they struck the earth. It also demonstrates that the concentration of pumice-dust in the air was not so great as to overwhelm a human being. It is interesting also that fossil moa footprints should be found in the volcanic layers of Scinde Island.

In his account of “The Vesuvius Eruption of 1906” (13), Frank A. Perret states:—

“Still farther on, and more especially upon the mountain itself, the down-sweeping ash became conglomerated through condensation of the water-vapour in the crater-cloud, forming balls of soft mud, some as large as an egg. During the two following days the production by this means of ‘pisolites’ was on a gigantic scale, although they were of small size compared with others found by the writer at Kilauea, and among the ancient ejected material of Vesuvius (Fig. 33).”

[Footnote] *Similar pisolites are abundant in the ash scattered over the Kau desert by the great explosive eruption of Kilauea in 1789. See T. A. Jaggar, Bull. Hawaiian Volc. Observ., 9, p. 114, 1921. Specimens of these collected by H.S.W. in 1920 vary from 2 to 5 mm. in diameter. Pisolites were found also by Lacroix in some ash-beds of Mont Pelée (La Montagne Pelée, Paris, 1904, p. 419), and by Pratt at Taal, Luzon (Jour. Geol., vol. 24, p. 450, 1916). (H.S.W.).

[Footnote] †Note by J. Marwick: The reference given by Perret in the original is Bull. Haw. Vol. Observ., 8, p. 114, 1290. This is an error. I have given the correct reference above. The H.S.W. stands for H. S. Washington, an American geologist.

– 606 –

In the South Kensington Museum in London I recently had the opportunity of seeing some specimens of chalazoidites which fell at Bellavista, near Portici, during the Vesuvius eruption of 1906. These are about the same size as those occurring in the upper part of the Chalazoiditic Layer in Scinde Island, and present about the same proportion to the matrix. They are, however, very much darker in colour, in response to the more basic material of which they are composed, and there is an increased amount of magnetite. Examination under the microscope shows them to be formed of a “crystal tuff,” in contradistinction to Scinde Island specimens, which are formed of a “vitric tuff.” In the Monticelli collection from Vesuvius there were specimens which fell during the eruption of A.D. 79, when Herculaneum and Pompeii were overwhelmed. These are remarkably similar to those found by Mr. Hill at Taupo, in appearance, size, and structure. The “tail” to which attention has been drawn was often present. I am informed by Dr. R. Ohashi that chalazoidites are well known in the volcanoes of Japan, particularly Oshima in the Pacific, near Tokyo, although apparently nothing has been written concerning them. From the Volcano Letter (14) it appears that chalazoidites have been described in western Germany.

While this paper was awaiting publication, Dr. J. Marwick kindly forwarded me an account by Drs. A. R. Richards and W. H. Bryan of volcanic mud-balls from the Brisbane Tuff (15). The authors have found chalazoidites, or spheroids as they call them, at Castra on the right bank of the Tingalpa Creek, twelve miles east-south-east of Brisbane. The beds in which they occur are very similar to the “Brisbane Tuff” of Upper Triassic age which is typically developed in and about the city of Brisbane. The account is rather brief, but from the details that are given some very interesting comparisons can be made. The larger chalazoidites were found as seems usually to be the case, in the lower part of the tuff, and the smaller ones towards the top. The largest specimens measured 25–30 mm., and are thus comparable to the Taupo specimens. In separating the chalazoidites from tuff in the middle of the section, by means of sieves with meshes of 11, 9, 6, and 4 mm. respectively, the largest number were in Group C—6 + 4 mm. Practically all the specimens collected from the higher part of the section belonged to Group D—6 + 4 mm. With regard to Scinde Island specimens, I have pointed out the same remarkable uniformity in size and weight of the chalazoidites in the upper part of the Chalazoiditic layer. The explanation of this fact must be left to the meteorologist, and opens up questions of the very greatest interest. The concentric structure was noted, although some specimens were thought to be homogeneous. Staining methods were not employed. I am inclined to believe that concentric lamination is a constant feature of all chalazoidites of whatever size. The density of the outermost layer is mentioned as “a skin of material with a glazed surface” and its probable effect in accentuating differences in chemical composition between the chalazoidite and the tuff in which they are imbedded is suggested. The authors suggest the “hailstone theory” as the probable explanation of the formation of the chalazoidite. They point out that in the volcanic outbursts of Castra and those in

– 607 –

Luzon, where the Taal volcano is actually within a lake, close proxmity to large bodies of water seems to be one of the factors in chalazoiditic formation. This seems likely enough, and should the volcano responsible for the formation of chalazoidites in Scinde Island be found to be to the south of Lake Taupo it would be another case in point.

It was suggested to the authors that these mud-balls from Brisbane, after being formed high above the volcano, became dried and hardened in their final descent through the hotter air nearer the volcanic vent, and the outermost parts in particular became glazed and indurated, and thus gave the mud-balls sufficient strength to withstand the shock of impact, whether they fell into the water of a Triassic lake, or into loose finely-comminuted volcanic ash. They have noted that the short axes are perpendicular to the bedding-planes of the tuff, which they suggest is the result of the pressure of the sediments added to the series. They consider the constant flattening of their chalazoidites to be due to pressure, which seems unlikely to have had much effect, since deformation of a body with a hard glazed outer skin imbedded in volcanic tuff would be difficult. The probability seems that they were in a plastic state on striking the earth.

Pratt has suggested that conditions peculiar to the tropics, such as high temperature and perhaps excessive humidity, are essential in the formation of chalazoidites. The authors are rather in favour of this view. They point out that the cosmopolitan nature of the floras of Rhaetic and Jurassic times, when little difference existed between the plants of such widely separated places as Greenland, Ceylon, and Antarctica, suggests that there was a uniformly moist and hot climate spreading from pole to pole. It was in these tropical conditions that the chalazoidites of Castra were formed.

The occurrence of chalazoidites as far south of the equator as Napier, and as far north as Naples in Pleistocene and in Recent times respectively, disproves the view that tropical conditions are necessary, though a high temperature of the pumice-cloud itself may be important. On the conditions that I have postulated for their formation they should be found in the ejecta of the volcanoes of Antarctica and Alaska as frequently as in the volcanoes of tropical countries. The statement that chalazoidites are a rare form of volcanic ejecta is probably incorrect. It is likely that now that these bodies are begininng to be recognized, they will be found in every volcanic country, and that knowledge of their distribution, both in recent and in geological times, will increase rapidly.

Richards and Bryan loc. cit. have considered the question of the centre of the eruption which resulted in the formation of their chalazoidites, and locate the source of activity to the east and south-east of Brisbane (i.e., towards Castra) in a region now foundered below sea-level. It is very interesting that the chalazoidites in Scinde Island are about 80 miles from the nearest present centre of volcanic activity—a greater distance than any hitherto recorded. If the conditions necessary for their formation that I have outlined are correct, there is no reason why they should not be found at even greater distances.

– 608 –

Illustrations show that the Brisbane chalazoidites are very similar to Scinde Island specimens in size and shape, although some are much larger. The series showing decortication resulting from weathering is interesting since fractures are also present. The polished sections show a well defined capsular layer, and also in some specimens, an eccentric nucleolus or nuclear point.


It seems that a study of volcanic layers will acquire more importance as knowledge of them increases. In an eruption, for example, in Miocene times, where volcanic material had covered a widespread area of country, it seems extremely probable that much valuable information would be obtained as to the contemporaniety of various, deposits, and what effect influences such as climate, depth of water, etc., have had in altering the fauna and flora, if this particular volcanic deposit could be identified by its continuity and its physical and chemical peculiarities. The account that has been given of the chalazoidites is incomplete and so far unsatisfactory, but further research will undoubtedly solve many of the intricate problems they present. There is no doubt that chalazoidites show differences in chemical composition, and in physical properties, depending on their sizes, and very interesting graphs could be prepared showing these differences. The explanation of these graphs, and the solution of the problem as to the cause of the cohesion of the chalazoidites, might not only be of value commercially in employing pumice for constructional works, but also give help in explaining their method of formation. If these researches are undertaken the object in writing this paper will have been attained.

In conclusion I wish to offer my thanks to the various people mentioned in this paper, and also to Prof. John A. Bartrum, Mr. and Mrs. Harwood, Mr. P. L. Hickes, Mr. E. A. Aldridge, and to Drs. R. S. R. Francis and A. R. Ford, for very valuable help.


1.—McKay, A., Geological Reports, 1879.

2.—Hill, H., Trans. N.Z. Inst., vol. 20, p. 301, 1888.

3.—Hill, H., loc. cit., p. 304.

4.—Jameson, N.Z. Journ. of Sci. and Tech., vol. 2, p. 209, 1917.

5.—Hill, H., Trans. N.Z. Inst., vol. 19, p. 385, 1887.

6.—Hill, H., Trans. N.Z. Inst., vol. 29, p. 571, 1897.

7.—Pratt, W. E., Journ. of Geol., vol. 24, 1916.

8.—Henderson, I., N.Z. Journ. of Sci. and Tech., pp. 225–226, November 1920.

9.—Hewitt, T. C., N.Z. Journ. of Sci. and Tech., vol. 9, p. 277, 1928.

10.—Guthrie-Smith, H., Tutira, p. 44.

11.—Hovey, E. O., Amer. Journ. of Sci., vol. 14, 1902.

12.—Jaggar, T. A., Bulletin Hawaiian Volcano Observatory, vol. 9, 1921.

13.—Ferret, F. A., The Vesuvius Eruption of 1906, Carnegie Inst. Publication No. 339, 1924.

14.—The Volcano Letter Kilauea Report, No. 856, June 21st, 1928.

15.—Richards, A. B., and Bryan, W. H., Pro. Roy. Soc. Queens., vol. 39, No. 5, 1927.