Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 11, 1878
This text is also available in PDF
(2 MB) Opens in new window
– 493 –

Art. LXXXIV.—On the Geological Structure of Banks Peninsula,

[Read before the Philosophical Institute of Canterbury, 7th March, 1878.]

Gentlemen,—Being called again by your vote to the honourable position of presiding at your meetings, the agreeable duty devolves upon me to address you to-night at the opening of the session 1878. It has been the custom of your newly elected President either to offer you a review of the progress of science in New Zealand, to treat of some special branch of scientific research, or to lay before you the results of his own investigations into the zoology, geology, or ethnology of these interesting islands.

With your permission, I shall follow the latter course, and venture to offer you some remarks upon the geological features disclosed to us by the piercing of the Christchurch and Lyttelton Railway Tunnel, a gigantic work, ever creditable to the energy and forethought of the Provincial Government of Canterbury in those days when only a small population had settled here, and the work to be undertaken was looked upon by many as far beyond our means. I shall preface the description of the tunnel, of which a section on a scale of one inch to twenty feet hangs at the wall, by some observations on the genetic history of Banks Peninsula, and upon the remarkable system of dykes, by which the older caldera walls have been intersected.

When standing on the Canterbury plains the most striking feature in the landscape is Banks Peninsula, rising so remarkably above the sea horizon, that its regular form at once attracts our attention. First we observe a series of mountains, of which the summits are all nearly of the same altitude, which, as it appears to us, as far as our eye can follow their outlines, form nearly a circle, from which a great number of ridges slope with a nearly uniform gradient towards south, west, and north. Above them, in the centre, stands conspicuously a higher truncated mountain with precipitous escarpments, assuming, according to the position of the traveller, a different aspect. The rim of the lower mountains in front rises to an average height of 1,600 feet, whilst the central system attains an altitude of 3,050 feet. On reaching Banks Peninsula from the sea, we find that several

– 496 –

deep indentations, forming splendid harbours, enter far into the outer rim of the mountains, passing for a considerable distance along the higher central range. Similar indentations are also found to exist towards the Canterbury plains, but they have either been already filled by alluvial deposits forming fertile valleys, such as the Kaituna valley, or they appear in the form of a lake (Lake Forsyth). In examining the nature of the rocks of which the system under consideration is composed, we find that, with the exception of a small zone at the head of Lyttelton Harbour, the whole is composed of volcanic rocks; that the deep indentations are ancient crater walls, so-called calderas, into which a channel with precipitous walls, the barranco, leads; and that they consist of a series of lava streams, with agglomerates consisting of scoriæ, lapilli, ashes, and tufas interstratified with them. These beds have all a qua-qua versal dip, that is to say, they all incline outwards from the centre of the cavity. The higher mountains in the centre consist also of volcanic rocks of a similar composition, which appear either horizontal or, when the direction of the lava-streams composing them can be ascertained, are found to flow into the calderas previously formed, from which we can at once conclude that they are of younger origin. Finally, we find mostly in or near the centre of these deep cavities, or calderas, either a small island or a peninsula stretching so far into these harbours. They consist also of volcanic rocks, having been preserved above the last centre of eruption. This last sign of vulcanicity is on a smaller scale than the previous ones. The whole of Banks Peninsula, measuring along its longest axis from north-west to south-east, has a length of 31 miles, with a greatest breadth of 20 miles, and if we do not take the numerous indentations into account, it has a circumference of 88 miles, which corresponds closely with that of the base of Mount Etna.

Having thus given an outline of the general features of the volcanic system under consideration, I shall now proceed to offer a short history of its origin.

The oldest rocks in Banks Peninsula form a small zone of palæozoic sedimentary strata, possessing a slightly altered structure, many of them forming beds of chert, others, peculiar light-coloured brecciated schists; however, sandstones and dark clay-slates are also represented. This zone has a north and south direction, and reaches to the southern watershed of McQueen's Pass, which leads from the head of Lyttelton Harbour to Lake Ellesmere. Near this pass, slates appear as high as 600 feet above the sea-level. On the western slopes of Castle Hill, the south-western cotinuation of Mount Herbert, 2,900 feet high, which rises so conspicuously above Lyttelton Harbour, they reach an altitude of nearly 1,000 feet, where they are overlaid by the older lavas, forming the Lyttelton Harbour caldera.

– 497 –

Thus a sub-marine hill stood here in the young mesozoic sea of which portions of the summits and the slopes were gradually covered by agglomerates and brecciated beds. These beds were formed during and after the eruption of quartziferous porphyries, of which here and there portions of the coulées have been preserved. Some of these quartziferous porphyries resemble in every respect those from the Malvern Hills and Mount Somers. They are also accompanied by pitchstones, porphyritic from the presence of numerous well-formed crystals of sanidine or glassy felspar, and occasionally of garnets. Other portions of the quartziferous porphyries, as for instance, the whole coulée of which Manson's Peninsula is formed, have a rougher, more trachytic matrix. They are full of grains and small crystals of white greyish or smoky quartz. The brecciated beds have a hard felsitic matrix, and the angular fragments of rock enclosed in them belong to a variety of eruptive rocks of many colours, and of different texture, often forming a rock of striking character. They appear conspicuously on the summit of Gebbie's Pass, having been washed into cliffs of picturesque forms, and covering the palæozoic sedimentary beds from one side, of the pass to the other. After the formation of the brecciated agglomerates, new eruptions of acidic rocks took place, now in the form of rhyolites, the highly liquid matter reaching the surface through broad channels, of Which one has been preserved as a large dyke, forming a beautiful section on the northern side of Gebbie's Pass, not far from the summit. The dyke is here about 100 feet thick, half of which is formed by the central portion, consisting of a whitish rhyolite with a fine laminated structure, breaking in prismatic blocks; the rest on both sides, where in contact with the agglomerates, has cooled more rapidly, and has assumed the character of an obsidian. This obsidian is greenish or brownish-black, very brittle, and imperfect crystals of sanidine are enclosed in it. This dyke can be traced for a considerable distance upwards. Where overflowing and covering the agglomerates it forms the highest peak on the western side of Gebbie's Pass, well visible from Lyttelton Harbour. The rock here is divided into small pentagonal columns, with a vertical arrangement; lower down the pass, the same coulée has a tabular structure.

The oldest crater, of which the principal boundaries can be traced a the present time, is the Lyttelton Harbour caldera, having a general diameter of about two miles, the centre of which is situated a little to the south of Quail Island. The general structure of this crater, even before the Christ-church and Lyttelton Railway tunnel was entirely pierced through, could easily be made out by studying the numerous sections exposed in many directions, and by ascending the steep escarpments of the caldera wall, where a succession of streams of stony or scoriaceous lava, interstratified

– 498 –

with beds of agglomerates, ashes, tufas, and laterites can be traced to the very summit. Still clearer sections are open to our inspection if we follow the barranco or entrance into the harbour, forming sometimes vertical cliffs of considerable altitude, and where the whole series of beds can easily be followed. However, the most interesting and complete insight was obtained in the railway tunnel passing through the caldera wall, and of which, as the work gradually advanced, I prepared a careful section. The succession and dip of the lava streams and the intervening beds can also be made out by following the slopes of the ridges between the deep valleys washed out on the outer side of the crater wall, where it will be found that the lava streams forming the lip of the crater have generally a slighter inclination than those lower down, the dip of the upper ones being only nine degrees on the average. In the tunnel the dip is greater, an inclination of twenty degrees not being uncommon. It is evident that the building up of such a huge system during numerous eruptions, often of great magnitude, could not be accomplished without a great destruction of portions of the beds previously formed taking place, the point of eruption in the crater shifting continuously about the centre. If, at the same time, we examine the lava streams and the interstratified agglomerate and ash beds along the water's edge, we have to come to the conclusion that all the eruptions by which the caldera wall was formed from summit to bottom, occurred under the same physical conditions.

Examining into the formation of the Lyttelton caldera, and beginning our observations in the harbour, we find that many lava streams have been preserved which have cooled in their ascent; others lie horizontal for a short distance, and are then seen to descend, conforming to the gradient of the underlying lava streams or agglomerate beds. In many instances we have also clear evidence that considerable destruction of the beds previously formed had taken place before new streams flowed over the lip of the crater, or before beds of ashes, scoriæ and lapilli, were deposited anew. The tunnel section in this respect is also very instructive. Thus, in course of time, the great crater wall was formed, rising to an altitude of nearly 2,000 feet, and having a diameter of more than five miles at its crest. It is clear that close to the vent, from which scoriæ and ashes were thrown out in large quantities, the greatest thickness of the agglomerate beds ought to be formed, and this, in fact, is the case, as the largest beds, having sometimes a thickness of several hundred feet, are situated within the inner side of the caldera wall. The lava-streams here between these agglomerates are irregular in their direction, and mostly of small dimensions. The more we advance towards the outer slopes of the caldera wall, the less frequent become these agglomeratic or tufaceous layers, whilst the lava-streams,

– 499 –

which towards the centre have the greatest bulk, and are very stony and compact, become now gradually more and more numerous, but of smaller size and more porphyritic or scoriaceous, according to the laws by which the flow, dimensions and cooling of the lava-streams are regulated. It is moreover, evident that many of them, owing to want of material, scarcely reach half way down the slopes of the caldera wall, that others rapidly thin out, and that many which, for some distance after flowing over the lip of the crater, had been of large dimensions and stony, become, long before its outer edge is reached, thin and scoriaceous, so that here streams of five feet in thickness are not uncommon. Although the tunnel does not offer us the necessary data to judge of the breadth of the lava-streams, we have for that purpose ample evidence in many other localities. There are streams which are 500 feet broad, others only 30 to 40, but all without exception are somewhat scoriaceous on the bottom, where the lava flowing over cold ground cooled more rapidly. In many instances this is well exhibited by the existence of a small bed of laterite, a brick-red coloured rock, sometimes only a few inches thick, which doubtless was a layer of soil on the decomposed upper portion of the lava-stream or agglomerate bed exposed for a considerable time to atmospheric action before the new eruption took place. The lava in the larger streams, and in its central portion principally, very stony and of a blackish colour, gradually becomes, as we approach the surface, more porphyritic, with a more open texture, and assumes pinkish or lilac tints, till it changes into scoriæ. The decomposition or alteration is here often so great that it is impossible to trace the top of the line of contact between the surface of the stream and the bottom of the overlying bed, both forming a layer of coarse agglomerate. In other instances the rough, uneven scoriaceous surface of the lava-streams has been well preserved, the hollow spaces being filled up by ashes and ejecta, in which case they resemble many of the recent lavastreams which I examined in Mount Vesuvius and Mount Etna shortly after they had issued from the crater.

The lava of which the caldera wall under consideration has been, built up, consists of basic rocks, changing from a dolerite to a fine-grained basalt. Some of the lava-streams, however, as previously pointed out, show also a remarkable difference in the structure of the rock of which they are composed, the central portion being a compact basalt with a few crystals of augite, basaltic hornblende, labradorite, whilst the upper portion consists of a lighter coloured porphyritic dolerite, sometimes so replete with good sized crystals of labradorite that the greater portion of the rock is formed of that mineral.

– 500 –

Returning to the orifice or orifices from which the material for the formation of the caldera wall was ejected, and to which also the numerous dykes, mostly having a vertical position, intersecting it, can be traced, it appears that the principal focus of eruption was situated a little to the south-west of Quail Island, as the greatest portion of the dykes radiate from here, and the eastern and southern sides of Quail Island, and the shores near Charteris Bay, are formed of tufaceous agglomeratic and brecciated beds, in which a number of angular blocks of rock are enclosed, having all a very bleached appearance.

It would go beyond the limits of this address were I to follow the further genetic history of Banks Peninsula in all its details. I can, therefore, only indicate here in a few words how the whole, in course of time, has been built up. Simultaneously with, or shortly after, the Lyttelton caldera, the Little River caldera, of which only a small portion remains, was formed in the same manner. The formation of the largest of the whole series, the Akaroa caldera, is next in age, which, with the exception of a small portion of its northern rim, is perfectly well preserved. After their formation, new eruptions, and of a different form, took place south of the Lyttelton caldera and north of the Little River and Akaroa calderas, during which the highest portion of Banks Peninsula was built up—Mount Herbert, 3,050; Castle-hill, 2,900; and Mount Sinclair, 2,800 feet; only portions of the craters of these younger systems are still preserved, but easily recognized when standing on the summits of these mountains. The southernmost portion of the Lyttelton caldera was partly destroyed or covered by lava-streams belonging to the Mount Herbert system, also of a basic (basaltic) nature, of which a whole series flowed into it, now forming the huge spurs descending from the summit of Mount Herbert into the harbour between Charteris and Rhodes Bays. The last eruption, of a submarine character, took place in the centre of the Lyttelton caldera, by which Quail Island was formed.

I shall now proceed to offer you some observations on the system of dykes, which are so well developed in the Lyttelton caldera. The most striking facts in connection with the system of dykes of the caldera, and to which I have devoted considerable attention, are their size, longitudinal extent and constancy in direction. From the researches of numerous observers, it has been proved that all the dykes of Mount Vesuvius and Mount Etna do not extend much beyond the centres of eruption, so that they advance only a short distance, and, rapidly thinning out, soon disappear, a fact which my own observations along the crater walls of both mountains have amply confirmed. However, I have no doubt that other volcanoes similar in construction to Banks Peninsula, and differing as considerably from

– 501 –

these two European volcanic mountains, will be found to posses their systems of dykes developed in the same manner. During a number of years, it has been well ascertained by me that the dykes radiating from the several centres of eruption situated not far from each other, continue in many instances without notable interruption from the former mouth of the crater to the outer slopes of the caldera, where they disappear below the sea, or under the deposits now forming the Canterbury plains. Very often the principal dykes rise nearly 2,000 feet above the sea level. They are well visible from the harbour to the summit of the rim of the caldera wall, above which, in some instances, they stand prominently as a wall, often six or eight feet high. Where proper measurements of the same dyke can be obtained for a long distance, it has been found that generally, as it advances towards the outer circle, it diminishes in breadth; however, in other instances this is not the case, as repeatedly I have found some which, after narrowing on their outward course, considerably enlarge again before reaching the foot of the caldera. Thus to give a few examples, the large dyke of trachyte, which is crossed in the railway tunnels, about 29 chains from the Heathcote end, is first seen west of the town of Lyttelton, near Naval Point, where it is nearly 40 feet thick. On the summit of the caldera wall, not far from the top of the Bridlepath, it has narrowed to 23 feet 9 inches, after which it gradually gains in proportion, so that in Thompson's quarry it has enlarged to 26 feet, a breadth which it still has in the tunnel. A mile beyond the quarry the spur along which its course can be followed runs out in the Heathcote valley, where it disappears below the Loess.

Two remarkable dykes, reaching the summit of Dyke Hill, about 2,000 feet high, west-south-west of Castle-hill, are very conspicuous. They both project boldly from the mountain, with a space of 35 feet between them. The eastern one is 18 feet, and the western 12 feet broad. Two similar dykes exist on the opposite side, and run up the caldera wall behind Raupaki. To mention a few others, there are some important dykes south of Dyer's Pass, which, after crossing Manson's Peninsula, are again met with at Ohinitahi (Governor's Bay), and of which several, after ascending to the very summit of the caldera, reach to the foot of the peninsula near Cashmere, being extensively quarried in different localities along their course. These dykes, like many others which cross the caldera wall towards the Canterbury plains, mostly all radiate from a point lying in the centre of the bay, formed by Manson's Peninsula on the one side, and Potts' Peninsula on the other, both of which consist of quartziferous porphyries, and between which this newer focus has been formed after the greatest portion of the caldera wall had already been built up. There is also the large dyke which crosses

– 502 –

the Lyttelton-Summer road at right-angles, on the very summit of Evans Pass, and which is repeatedly passed by the road winding in and out of the different bays before reaching that pass. It can be followed to Taylor's Mistake. Everywhere along the sea cliffs at and near the entrance of Lyttelton Harbour, numerous dykes, mostly all in a vertical position, can be seen pointing towards the centre of that harbour. A few, however, stand in a slanting position, and others have a tortuous course. As one of the remarkable changes which some of the dykes have undergone since their formation, I may also mention one which is well exposed in the sea cliffs at Ohinitahi, Governor's Bay; here a dyke of domite, about nine feet broad, crosses in a nearly vertical position the so-called trachyte sandstone de-posited on the slopes of the quartziferous porphyry. After its solidification, a new fissure, about three feet broad, has been formed parallel to its direction, and running along its centre, which has been injected from below by domitic matter, but slightly different from the former; however, instead of continuing to the top of the cliff, about twelve feet above the sea level, the dyke is seen to turn from its vertical to a nearly horizontal position, and to thin out considerably at the same time, disappearing altogether when it touches the side wall of the bed-rock. The older dyke, above this change of direction, is considerably shattered and broken.

Before proceeding, it will perhaps be useful if I offer a few remarks on the causes which led to the formation of these remarkable dykes. I consider this the more important, as nowhere, as far as I am aware, do they exist in such great numbers, nor do they possess such a large longitudinal extent, as in the volcanic system under consideration. It appears to me that the immediate cause of the formation of a radiating system of dykes may be traced to the choked-up vent or chimney of a volcano, the mouth of which, after an eruption of considerable dimensions, is thoroughly filled up, either by its sides falling in, by the cooling of ascending lava-streams, or by both causes combined. When, from abyssological origination, masses of steam and gases have collected below this vent, and new matter is ready to be erupted, an enormous effort of nature will be necessary to clear out the old, or form a new chimney, which cannot be accomplished without a series of violent earthquakes, succeeded by an enormous explosion, by which the mouth of the volcano is cleared out or newly formed, and of the magnitude of which we can scarcely form a conception. A similar effect, on a gigantic scale, must have been produced repeatedly by the compressed masses of gases and steam during the formation of the Lyttelton caldera wall, when the upper portion of the closed-up volcano was not only removed, but vast quantities of ashes, scoriæ, and lapilli were thrown out, together with lavastreams which flowed in various directions. Before, or during these

– 503 –

eruptions, molten matter in a high state of fusion generally rushed up in the fissures which had been formed at the time, radiating from the focus like the spokes of a wheel. An examination of these dyke rocks will show at a glance that most of them are quite different in composition and character from those of which the lava-streams have been formed. The latter, as already explained, with one notable exception,* all consist of true basic rocks—basalts often assuming a dolertic texture, the dyke rocks being generally acidic, having either the composition of a trachyte or domite. We are able to judge of the more or less high state of fusion in which the molten matter ascended the open fissures from the effect produced on the walls on both sides. The trachytic matter forming the dykes, which are principally developed on the eastern side of the caldera wall, has evidently been in such a condition that it could exercise a most powerful effect on both walls of the fissure, the rocks often, for several inches, being changed to tachylite, a peculiar basic volcanic glass, quite distinct from obsidian. This change in the character of the rock is most observable when the dykes pass along tufaceous or agglomeratic beds. Here the reddish or light purple rocks have been altered to a black vitreous mass, containing small crystals of felspar. The domitic dykes, mostly confined to the western half of the caldera wall, seem not to have exercised such a great influence as the former, as in most instances the walls on both sides of the dykes are only slightly hardened. However, there is no constant rule; large dykes, as for instance the huge domitic dyke at Governor's Bay, running for a considerable distance parallel to the coast, and forming such a conspicuous object along the picturesque beach road lately constructed, have scarcely made any alteration on either side, whilst smaller dykes of the same rock, only a few feet in thickness, are sometimes accompanied by a well-defined selvage of tachylite. The same may be said of the basaltic dykes, of which, however, by far the greatest part has caused no visible alteration along the walls on either side. The trachytic varieties,

[Footnote] * This exception consists of a trachytic lava-stream of considerable size, and having an average thickness of eighty feet, which is interstratified between two others of a basic character. This peculiar stream occurs between Lyttelton and the pass to Summer. It is the only trachyte lava known to me as having flowed from any of the different centres of eruption of Banks Peninsula, all the other acidic rocks, as I shall show in the sequel, having been ejected into fissures of more recent date. This lava-stream consists of a white vesicular trachyte rich in quartz, resembling closely some of the domites of the Auvergne, from which, however, it is distinguished by its larger amount of silica, although it approaches it again in its considerable percentage of potash. A vertical dyke, about eight feet thiik, of a peculiar flaky silky trachyte, passes through this lava-stream, narrowing, however, in its upper portion. Although this acidic lava is rather soft and friable in small pieces, it has nevertheless resisted the disintegrating agencies at work far better than the hard basaltic lavas and agglomerates in its neighbourhood.

– 504 –

of which most of the dykes on the eastern side of the Lyttelton Harbour consist, are formed generally of a peculiarly lustrous and flaky rock, some-times vesicular, with small crystals of sanidine. This rock has a light greyish colour, and its small cavities are lined by sphærosiderite. On both sides of the dyke the rock is generally tabular—parallel to the direction of the flow, and is massive in the centre with polyhedric joints, of which the principal ones appear at right angles to the flow. There are also a few trachytic dykes, principally small ones, where the sides, for half an inch to one inch, consist of a rather brittle obsidian, doubtless the effect of rapid cooling. Some very thin thread like dykes, about one to two inches thick, consist entirely of that peculiar form of acidic volcanic rock.

In studying the position of the dykes it becomes manifest that they have been formed at different times; however, the altitude of their uppermost portion does not indicate their age. I have no doubt that many of them, which scarcely reach above high water-mark, are not older than others of the same petrological nature, which reach to the very summit of the caldera wall. In the present state of our knowledge it is impossible to solve this interesting question in all its bearings, and I can therefore only suggest that dykes containing rocks of exactly he same lithological character have most probably been formed during the same eruption. It is also evident that a number of dykes were formed long before the whole of the caldera wall was built up, and that they were partly destroyed during one of the next eruptions. One clear instance of the occurrence of such older dykes is to be found near Cliff's Cove in Lyttelton Harbour, where several trachy doleritic dykes were injected when the rest of the caldera wall was at least 1,00 feet lower than at present. They pass through a basaltic lava-stream, which latter was afterwards partly destroyed along with them, the whole possessing now nearly a straight surface, upon which a large, bed of agglomerate has been deposited. However, what is of the greatest interest in the history of the volcanic systems under consideration is the predominating acidic character of the dykes when compared with the basic lava streams. In Vesuvius and Etna all the dykes are formed by the same kind of rock as the lava-streams are composed of, but they are generally more of rock compact, having, as Lyell suggests, cooled and consolidated under greater pressure. It is evident that they owe their existence to the same subterranean efforts by which the lava-streams were ejected from the mouth of the crater, the fissures in which they were formed being evidently filled up from the same focus, and about the same time as the eruption of the lava-streams took place. But such a simple process cannot be admitted for the greater portion of the dykes of Banks Peninsula, which must owe their existence to paroxysmal perturbations in the earth's crust distinct from

– 505 –

those during which the caldera walls were built up. It is evident that a great portion of the lava-streams and agglomeratic beds which once formed the crater of the volcanic system of Lyttelton Harbour, must have been blown away, or at least removed during one of those violent outbursts of subterranean forces necessary to clear the choked vent of the volcano, similar to those by which in recent times the upper portions of active volcanoes have repeatedly been destroyed under the eyes of the trembling population in the neighbourhood.

For an explonation we might go back to Durocher's views, that all igneous rocks, even the most modern lavas, are derived from two distinct magmas which co-exist below the solid crust of the globe, each of them occupying a well-defined position. According to this distinguished French chemist, the uppermost portion is occupied by the acidic magma, which, be-sides being of lighter specific gravity, possesses a larger amount of silica and less iron oxyde than the other or basic magma. From the upper layer the granites, porphyries, and trachytes, according to his views, are-derived the zone of contact producing rocks of an intermediate character, such as trachydolerites. If this theory is correct, we have to admit that not only the dyke rocks were injected in rents formed during earthquakes, or immediately before volcanic eruptions had taken place from the opened chimney of the volcano, but that in each case the molten matter was furnished both from the upper and lower stratum of incandescent matter below the hard crust of the globe. There is, however, one great difficulty which crops up here, and which I wish to point out, and that is the presence of dykes of basic rocks and of others of an intermediate character. If all the radiating fissures without exception had been filled up by acidic rocks, this would go far to prove the existence of such an upper acidic incandescent magma; in which case we should be forced to the conclusion that the chimney of the volcano reached lower down to the lower or basic layer. But it is difficult to under-stand how all the radiating fissures over an area of 12 miles in diameter could pass through the solid crust of the earth and through the fluid acidic magma, and how the lower basic rocks could be injected into them from below without disturbing the acidic magma, which certainly should have been forced up before. This difficulty might, however, be met by the suggestion that the radiating fissures in this instance did not reach so far down as the fluid acidic magma, and that the material for the formation of the dykes had been furnished from the crater itself, but it is scarcely conceivable that for a distance of six miles and for an altitude of several thousand feet the molten matter would have been forced in all directions from the central axis of eruption along these fissures, often only a few feet wide. Mr. R. Mallet,*

[Footnote] * Transactions of the Royal Society. Phil. Trans. 1873.

– 506 –

has proposed another theory, namely, that the principal cause of vulcanicity is to be sought in the compressing and crushing action taking place beneath the crust of the earth, and by which such a great amount of heat is generated that a fusion of rocks, often on a large scale, is easily produced. This theory would so far explain very well the difference in the composition of the rocks varying according to the depth where the crushing action was actually taking place; thus, if the same action were to act upon granites, trachytes, and other acidic rocks, the result would be the production of trachytes, whilst if basic rocks were fused, basalts would ascend towards or to the surface. Here, however, another great difficulty presents itself in the fact that, although the number of volcanic eruptions during which the caldera walls were built up must have been very great, no trachytic lava streams, with one single exception, have made their appearance, the whole series being of a basic, whilst most of the principal dykes are of an acidic nature. In such a case, the crushing of acidic rocks would have exclusively taken place when the dykes were being formed, and never when lava-streams issued from the crater's mouth, which is altogether improbable.

Although I have carefully read every work accessible to me in English, German, French, and Italian, treating on vulcanicity, I have not been able to find either any account of the existence of dykes in other volcanic regions converging so regularly to a few centres close to each other, or continuing over such a large area (always keeping the general direction with which they set out), as do those of the Lyttelton caldera; or again, offering an explanation for the difference in the composition of the dyke rocks when compared with the lava-streams or agglomeratic beds through which they pass. Mr. R. Mallet's excellent paper on the “Mechanism of Production of Volcanic Dykes,”* and of those of Mount Somma, in which an exhaustive account of the physical features of the dykes in the old caldera wall of Mount Vesuvius is given, unfortunately does not contain any physical theory to account for the mode by which fissures are produced, forming, when filled, volcanic dykes. If we take the heterogeneous nature of the material of which the caldera wall has been built up into account, it is astonishing that the dykes show such a remarkable regularity, always starting from a few points not far from each other, from which they radiate in all directions. It is still more remarkable to observe that all dykes which are cut by the Christchurch and Lyttelton railway tunnel have such a constant direction that they all, with one or two exceptions, appear to converge to one single axis behind Quail Island, a fact worthy of note if we consider the distance, which is more than four miles, measured to the most distant dyke in that tunnel. The only dyke with which I am acquainted,

[Footnote] * Quarterly Journal of the Geological Society of London, No. 128, Nov., 1876.

– 507 –

showing some remarkable irregularity, is the one in which the so-called Ellis Quarry is situated. This dyke, which strikes nearly east and west, goes out about 400 feet below the summit, where a saddle intersects the spur. Shortly above its lower termination it sends off a smaller branch in a south-west direction, also ceasing after a short course. Whilst the main dyke does not appear any more above the surface, the smaller south-western branch crops up again on the other side of the depression, now gradually changing its direction, so that, in its lower course, about 300 feet above the plains, it crosses the spur in a south-east and north-west direction. The whole system of dykes in the Lyttelton caldera wall is thus very different from the dykes of Mount Somma, of which, in his paper, Mr. R. Mallet gives us such a lucid and suggestive account, and of which many are fractured, displaced, and crushed, and have at the same time a wedge-shaped form. We can, therefore, assume that the fissures and dykes in the Lyttelton caldera were only formed after the latter had been so thoroughly consolidated that, after the formation of the fissures and their filling up by the principal dykes, no more changes of any importance took place in them; and that, moreover, the forces by which the walls of the volcano were starred from top to bottom, must have been far deeper-seated and more effective than the agencies by which Mount Somma was rent.

In conclusion, I wish to lay before you a few notes on the geological features of the Lyttelton and Christchurch railway tunnel, of which I made a careful survey during a number of years, as the work of the miner advanced. I watched this interesting and instructive work with great attention, this being the first time that a caldera wall of a large extinct volcano was to be pierced through. I prepared at the time a section on a scale of lin. to 20ft., which I have great pleasure in laying before you.

The direction of the tunnel is N. 14° W. The first trial shaft was commenced in January, 1860, and the permanent works under contract with Messrs. Holmes and Co., began in July, 1861. The tunnel was laid out, and its execution solely superintended by Mr. Edward Dobson, C.E., Provincial Engineer. It was brought to a successful termination on May 25, 1866, when both adits met near the centre. The opening for railway traffic took place on December 9, 1867. The total length of the tunnel is 8,598 feet, and if we deduct from this 365 feet on the northern or outer side, and 105 feet on the southern or inner side, formed by slope deposits and loess, there remains 8,128 feet of rock of volcanic origin, of which the caldera wall has been built up. Classifying the rocks according to their lithological character, we find that the crater above the present waterline consists of—

61 lava-streams, having the character of a stony compact or porphyritic basalt.

– 508 –

54 lava-streams formed of a scoriacepus basaltic and doleritic lava, some of them changing so gradually into agglomeratic beds that the line between them cannot be clearly defined.

39 beds of agglomerates, a few of them changing into scoriaceous lava, but most of them consisting of scoriæ, lapilli, and other ejecta, imbedded in ashes. A few of them have a brecciated appearance.

19 beds of laterite, clays, and slope deposits, partly or wholly burnt by overlying lava-streams, and

1 small layer of bolus—together 174.

These beds are intersected by 32 dykes, 18 consisting of trachyte lava (of which five do not reach to the roof of the tunnel), and 14 of a basic nature (five of them being intermediate in character, trachy-dolerites). One of them comes from the top of the tunnel.

Beginning at the southern or Lyttelton side of the tunnel, we observe that a large bed of loam has been deposited upon the volcanic rocks, being thickest on the lowest third of the caldera wall. This peculiar rock, which, when in small pieces, is easily pulverized between the fingers, has a remarkable consistency and solidity when in large masses, and is of subaerial origin. It may be designated as loess, an expression now extensively used in Europe for similar deposits. It owes its origin to various processes, of which rain, wind, and vegetation are the principal factors. This bed of loess, which in some localities is more than 100 feet thick, changes gradually before we reach the volcanic rock to a true slope deposit, consisting of fragments of rock more or less rounded, the lines of junction being often impossible to trace, owing to the decomposition of the volcanic rocks immediately below the slope deposits. The greatest amount of agglomerate, consisting of scoriæ, lapilli, and ashes is, as might be expected, congregated on the inner side of the caldera wall, not far from the focus of eruption. These more or less incoherent beds, of which each was probably formed during one eruption, have generally an inward as well as an outward dip, of which the beds 232 to 241 close to the entrance of the tunnel at Lyttelton form a notable instance. They were without doubt deposited on the lip of the crater. Near the Lyttelton end they are much disturbed. Two stony lava-streams cross these agglomerate beds, and we have to assume that after No. 231 was formed, the lava-stream 233, ascending from the mouth of the crater, had consolidated over it, being in its turn covered by a new talus of ejecta sloping inwards to the crater's mouth. After these latter beds 234 and 234A were formed, a new stony lava-stream, No. 237, ascended, in which case Nos. 231, 234, and 238 to 241 were three distinct agglomerate beds, covered and preserved on their

– 509 –

inner slope by stony lava-streams, consolidated during their ascent. Or, to offer another explanation, we might regard these two stony lava-streams, 233 and 237, as having broken through the huge accumulations of ejecta which were heaped up all round the crater's mouth—a phenomenon frequently observed during violent volcanic eruptions, when a huge cinder cone is formed in a short time. A similar occurrence seems to have taken place more towards the centre of the tunnel, about 60 chains from the Lyttelton side, where a large stony lava-stream, No. 167, is seen to ascend though the agglomerate bed or beds, Nos. 166 to 168. The lava-stream, 163, in close proximity, might be considered to be the continuation of the former, which here flows down the steep side of the cinder cone. Gradually, as we retreat from the focus of eruption, the agglomerate beds decrease in number and size, but they still are occasionally present even close to the mouth of the tunnel near to the Heathcote entrance. Some of them consist in their lower portion of fine ashes, or lava d'aqua, and above of scoriæ and lapilli, so as to suggest that first fine ashes had been thrown out or had been brought down the side in the form of a mud stream, on the top of which large ejecta were afterwards deposited. Another agglomerate bed having an anticlinal or saddle arrangement is 22A, 17 to 20 chains from the Lyttelton end; it was evidently deposited on the rim of the crater, of which the uneven surface is well visible in its lower portion. After its formation, two more agglomerate beds were deposited over it, 216 and 227, and 211 and 228 in the section, each being separated from the other by a bed of laterite. Moreover, it is clear that, whatever may have been its origin, the lowest portion of this and several other agglomerate beds must have been deposited when in a state of high temperature, as the argillaceous bed below it has been burnt red, so as to take all the characteristics of a laterite. All round Banks Peninsula agglomerate and ash beds are visible in the cliffs, but they are like the lava-streams of small vertical extent only, and we have to approach more towards the centre of eruption when we wish to see them in their greatest dimensions.

The largest and most numerous stony lava-streams are met with towards the centre of the tunnel, where the basalt of which they are composed possesses the greatest hardness and crystalline texture. More towards the boundaries of the volcanic system, the lava-streams are much thinner and at the same time more porphyritic, amygdaloidal or scoriaceous, and it is very instructive to follow some of the lava-streams which form clear sections in the deep valleys radiating round the peninsula, from the summit of the caldera wall to their termination at its foot, and to note the gradual change in their size, and in the texture of the rocks of which they are composed. I have already alluded to the lava-stream 237, nine chains from the Lyttelton

– 510 –

end, but in connection with it I may here mention that the first shaft sunk by Messrs. Smith and Knight, the English contractors, unfortunately reached it soon below the surface of the ground, and continued all the way through it to the roof of the tunnel. This was one of the principal causes that the firm, being unacquainted with the formation of the crater wall, abandoned the contract so soon. The first stony lava in the tunnel, flowing down the slopes of the crater wall, is a small stream, No. 214, about 22 chains from the Lyttelton end. Several others of similar dimensions follow, till we reach stream 206, which might be the continuation of No. 237, 11 ½ chains from the Lyttelton end. This stream throws a great deal of light by its configuration on the manner of the flow of liquid lava. After flowing down the slopes, we see it shortly afterwards ascend again (No. 202) over a bed of agglomerate, and, after having reached the apex of the latter, descend again (No. 200), diminishing rapidly in size, the rock now becoming highly porphyritic and lighter in colour. The largest stony lava-stream of the whole series begins about 41 chains from the Lyttelton end, and continues without interruption to 52 ½ chains. Consequently, taking its angle of dip into account, it is more than 500 feet thick. More or less porphyritic on both sides, the whole central portion consists of a very hard basaltic rock, ringing to the hammer, irregularly jointed, with here and there a tendency towards spheroidal structure. This huge stream gains an additional interest from the existence of three caves in its centre, which, however, have partly been filled up with thin plates of basalt of the same texture as the lava-stream, and which lie more or less horizontal. They are coated over and often cemented together by sphærosiderite. Sometimes they lie in such regular order, and so loosely upon each other, as if they had been artificially placed in that position. The open space, or cave proper, is always on the southern side of each cavity. The only explanation I can offer as to their formation is that gases have been enclosed in this portion of the lava stream, which in course of time were absorbed, and that liquid matter from the upper portion of the stream found access to the cavities, gradually filling them up, but that the channels of communication were stopped before the whole of the gases still remaining in the southern parts of each had been absorbed.

Another stream of large dimensions is No. 14, beginning 20 chains from the Heathcote end. It is over 100 feet thick, has a jointed structure, the central portion being spheroidal, with concentric layers. All the stony streams in the tunnel above the latter are very thin, but it is possible that the scoriaceous basaltic lava (the violet beds of the section) which overlie them, are only their upper portion, the bottom of the streams, owing to their thinness and to the distance from the centre of eruption, not having

– 511 –

been able to cool to the stony compact form. I may, however, observe that the boundary line between both kinds of rock is, in many instances, very distinct and clearly defined. These scoriaceous beds occur throughout the tunnel; they are sometimes of considerable dimensions, some of them being over 100 feet thick. In speaking of the formation of the Lyttelton caldera, I have already pointed out that it has been built up by volcanic rocks belonging to two distinct divisions, of which the basic rocks have furnished all the material for the lava-streams, agglomeratic and tufaceos beds, whilst the principal portion of the dykes owe their origin to the acidic division. As might be anticipated, the dykes are most numerous near the focus of eruption; thus we find the greater portion of them near the Lyttelton side, several of them not reaching to the roof of the tunnel. Of these dykes, No. 29 is the most important. It consists of a soft flaky and lustrous trachyte, and possesses, like most of the other acidic dykes, the characteristic feature that it is acoompanied on both sides by a selvage of tachylite, sometimes two or three inches thick. This change in the character of the bed rock is especially visible when the dykes pass through agglomeratic or tufaceous beds. It shows clearly that the volcanic matter ascending by these fissures was in such an intense state of fusion that it was able to alter the rocks on both sides so thoroughly for such a distance. In some instances the dyke rocks themselves have a selvage of tachylite, the bed rock being unaltered. It is worthy of notice that the basaltic dykes have not produced the same effect, the rocks on both sides being generally unaltered. Large beds of loess, similar to those deposited on the inner side of the caldera wall, have also been passed through on the Heathcote side. Of minerals of secondary origin found in the tunnel, the most diffuse is sphærosiderite, which usually coats the pores and cavities of scoriaceous lavas. Of others, calcareous spar and aragonite are the most conspicuous. The latter is younger than the former, having often been deposited on the surface of the calcareous spar coating the small geodes. In a few localities, hyalite fills small clefts, or is found in a stalactitic form.

I shall close this address by offering a few observations on some other physical features of the beds through which the tunnel has been excavated, and as I noted them on the large section during the survey. Forty chains from the Heathcote end, a scoriaceous lava-stream, fifteen feet thick, and accompanied on both sides by beds of laterite and agglomerate, was passed, which was so loose and full of water that the ground had at once to be heavily timbered. All the cavities in the lava are lined with sphærosiderite, on which crystals of calcareous spar have been deposited. At 40 ¾ chains on the same side, in a bed of laterite, four feet above the floor of the tunnel, a small spring was struck, drying up a few months after; 35 ½ chains from

– 512 –

the Lyttelton end, the lava-streams, when first passed through, were so wet that the workmen could scarcely continue the work. In these streams all the cellular cavities were either lined with sphærosiderite or filled with calcite. Sixteen chains from the Lyttelton entrance, in the agglomerate bed No. 228, and from a fissure reaching from the roof of the tunnel, a copious, spring flows which has a constant temperature of 65.20 degrees Fahrenheit, consequently 12.20 degrees above the mean temperature of Christchurch—about 53 degrees. Several eels have been caught near this spring in the drain which runs from here to the mouth of the tunnel. There being no connection with any other watercourse, these eels must have ascended by the spring; they belong to the species Anguilla aucklandii, Rich., and have properly developed eyes. During the construction of the tunnel it was frequently observed in the north, or Heathcote end, that the water rose in the floor before a south-west gale, and subsided before the gale lulled; no observations could be made to ascertain whether the state of the tide had anything to do with this. The height to which the water rose was some-what under half an inch. After the earthquake of August 17, 1868, this spring in the tunnel increased to such an extent that it laid the rails slightly under water; after a few days it decreased again to its former volume.