phyllites. This part of the area is almost surrounded by a more or less continuous ring of rhyolite, which forms an important part of the crest of the ridge to the west, and after a break reaches as a broad band across the lower parts of the streams which rise in the divide lying in the direction of Lyttelton Harbour, and which converge to form the main stream running through the valleys.

There is no visible occurrence of andesite between the ridge east of McQueens Pass and the proximal end of Gold Spur, although it may exist under the veneer of rhyolite. The thickness of the beds as exposed on the spur east of the pass is comparatively small, so that it is quite possible that the andesite thins out under the rhyolite, and may not exist at all at the present time, although it may have occurred there originally. Andesite forms the main portion of Gold Spur, but is divided from the rhyolite at the proximal end by a low saddle. Likewise a saddle marks the junction of the mass of rhyolite ending the spur, and the andesite to the north of it (Plate B). This series forms the end of the spur beneath the rhyolite, the occurrence of greywacke noted by Speight (1917, p. 377, and map) as occurring there being a mistake.

The series is developed on both sides of Gold Valley downstream from the saddle at the proximal end of the spur. On the east it forms a definite shelf, practically accordant in height with the summit of Gold Spur and that of the series as developed west of McQueens Valley.

Outside the McQueens-Gold Valley watershed the series is exposed at two places on the road from the pass eastward to the new wireless station, and again to the west of the pass on the side of the road 2–3 chains north of the summit. These exposures are of no great area, and if other masses occur concealed for some distance west of the pass they must be thin, since the greywacke and overlying rhyolite are in close proximity and do not allow space for any development. This also applies to the proximal end of Potts Point peninsula, where the Charteris Bay sandstone rests on greywacke.

Another occurrence no doubt belongs here, viz., one about halfway up the ridge between the valleys leading down to Governors Bay on the west and to the Head of the Bay on the east. This consists of a thick mass of basic andesite, but with quartz amygdal-oids, and of different facies from the Lyttelton andesites. It lies right down on greywacke, and is covered at the S.W. end of the exposure by Lyttelton basic rocks, while in the middle and towards the N.E. end it is definitely overlain by rhyolite. It is thus between the rhyolite and the greywacke, and it is not intrusive.

On the whole, the McQueens Valley series is somewhat thin. It attains its maximum thickness, some 500 feet, on the flanks of McQueens and Gold Valleys. No doubt it was once continuous over

a wider area, and the isolation of the various masses is to be attributed to the erosion to which the area was subjected at various stages in its subsequent history.

The exposed beds belonging to this series consist chiefly of flows, ash-beds being relatively unimportant; but this may be due to the fact that the latter weather more easily and their surface is more completely covered with soil and slip-material.

The general trend of the bedding appears to be E.-W. with dips in varying directions and of varying amount. Occasionally they lie flat. Some of the folding was certainly pre-rhyolite, since the latter lies in places across the denuded edges of the andesite, but as will be noted later the rhyolites have been folded on apparently the same lines, and therefore there are two periods at which distortion has taken place. There is a marked syncline in ash-beds on the east side of Gold Valley.

A special characteristic of this series is the presence of masses of quartz, not only as amygdaloids, but also like reefs. In the ‘60’s and ‘70’s of last century the latter were believed to be gold-bearing and numerous adits were driven in the hope that payable gold would be obtained. This hope proved fallacious, as can be clearly seen from the results of the analyses of stone from this district published in the fifth Annual Report of the Colonial Laboratory, 1870, pp. 21–3. All the same, the most eastern valley in the district was named Gold Valley. Traces of the drives occur in it and on both sides of McQueens Valley in this volcanic series as well as in the older Trias-Jura sedimentaries, usually following along the reef-like masses of quartz. In addition to this form of quartz there is the somewhat frequent occurrence of chalcedonic silica as amygdaloids closely resembling those occurring in the andesites of the Malvern and Clent Hills.

The age of the beds has to be very carefully considered. First of all, the maximum age is determined by the fact that they rest unconformably on the sedimentaries of presumably Trias-Jura age. Their minimum age is determined by their relationship to the Charteris Bay sandstones, which are Awamoan, that is, Miocene. These latter certainly overlie the McQueens Valley Series on the spur south of the wireless station, just as they themselves underlie the rhyolite.

The relation of these beds to the rhyolites is a matter of considerable importance. There seems to be no doubt that rhyolite rests on the andesite and basalt of the McQueens Valley Series. This is clearly seen on both ends of Gold Spur and on McQueens Spur, where a rhyolite outlier is completely surrounded by andesite, the former resting definitely on the latter. Other cases occur on the same ridge, in which rhyolite occurs at a higher elevation and faulting seems impossible to account for any apparent reversal of age. There is besides the presence of andesite fragments in breccias

which belong to the rhyolite series. In no single case of all the outcrops does andesite or basalt of the McQueens Valley Series overlie rhyolite.

Therefore the conclusion is arrived at that this series is prerhyolite and also anterior to the Charteris Bay sandstone—to be referred to more particularly later. This last has been determined as Miocene (Awamoan) in age; there is thus a wide range in time to which they can be assigned. For more precise designation we are driven to the uncertain test of lithological resemblance to other occurrences whose age has been more definitely determined, viz. to those of the Malvern and Clent Hills. The former are certainly pre-Senonian, since they underlie Senonian coal-measures. Therefore relying on this somewhat uncertain evidence we may tentatively assign these beds to the Upper Jurassic or Lower Cretaceous period. The Clent Hills volcanics do not afford any certain evidence for an assignment to narrower limits.

Although the great majority of the beds belonging to this series as exposed in the typical locality are andesitic, some of the more basic should certainly be called basalts. This applies to those beds resting close down on the older sedimentaries, e.g., those occurrences just east of the road over Gebbies Pass, that on the summit of McQueens Pass and the exposure near the lagoon south of the wireless station. They are all doleritic in character, and this suggests an intrusive origin. Their field-occurrence is somewhat against this hypothesis, for there is no exposure of similar beds on the same line or in close proximity in the underlying sedimentaries, as there should be if the beds were intrusive, unless, indeed, they have forced their way along the junction of the McQueens Series and the Trias-Jura beds.

The greater proportion of the andesites belonging to the McQueens Valley Series are dark-grey to blackish in colour, but there is one variety of limited occurrence which is much lighter in tint and might be taken in the hand-specimen for a rhyolite; it shows departures in character from the general run. They are usually stony in appearance, but at times they are vitreous in texture; the former show phenocrysts freely, but the latter are compact and phenocrysts are very rare.

Under the microscope they prove to be almost invariably pyroxene-andesites, ranging in texture from those which are definitely glassy to those in which the base is holocrystalline. The feldspar is a medium labradorite; the pyroxene is usually augite, but there is an occasional hypersthene, with its margin outlined at times with augite. Olivine does not occur in the true andesites as far as can be seen, though it occurs in the basic earlier occurrences.

The following are analyses showing the composition of representative rocks of the area:—

Analyses by F. T. Seelye.

The first four of these analyses give the general composition of the andesites of the McQueens Valley Series. They show a marked difference from those of the Lyttelton Series and a close approximation to those of the Mount Somers and Malvern areas, thus emphasizing the fact that they must be considered apart from the Lyttelton rocks. The high content of silica may be partly due to small amygdaloids of quartz, but every effort was made to secure samples for analysis free from them. No. 4 is of a very fine-grained type, and the analyst notes: “Some qualitative tests made on the soft yellowish-brown material in the vesicles of this rock showed the presence of fairly large amounts of iron oxide, alumina, and magnesia, together with a small amount of lime.”

Analysis No. 5 is closely comparable with those of the somewhat alkaline basalts which are typical of the later phases of the Lyttelton and Akaroa eruptions, and supports the hypothesis that the McQueens Pass outcrop and other similar ones are really intrusions.

F. The Rhyolite Series. (See Map and Plates 37 and 38.)

This is best developed outside the area under special consideration on the peninsulas stretching down into the upper part of Lyttelton Harbour. Almost the whole of Mansons Peninsula and a considerable portion of the ridge it terminates, as well as almost the entire peninsula of Potts Point apart from the sandstones which form its basement are composed of rhyolite. Rhyolite extends from Gebbies Pass to the west and to the east. In the former direction it forms a well-defined band along the road following the ridge and north of the greywacke basement. To the east of the pass it forms the greater part of Gebbies Spur and portions of the crest of Thompsons Spur, and extends therefrom after a slight break as a broad band across the middle portion of McQueens Valley into the head of Gold Valley. An isolated mass lies across the end of Gold Valley, forming the bluff at the end of the spur and the slopes opposite it (Pl. B), an occurrence which probably links up under the floor of McQueens Valley with the rhyolite which forms the distal end of Gebbies Spur—except the very end—and links up again with the occurrences to the west of Gebbies Valley. These extend up to the wall of Lyttelton volcanics to the west of the valley for a considerable distance, the contacts being obscured so that the precise boundary cannot be determined. A small outlier lies at the extreme end of the greywacke ridge through the pass, but further south there is no sign of its continuity in that direction, since the Lyttelton volcanics rest directly on greywacke. No doubt all these separate occurrences were once continuous, but the connection has been broken by erosion preceding and following on the folding.

In the neighbourhood of Gebbies Pass the rhyolites have suffered a folding on approximately E.-W. lines. One anticlinal axis passes through the summit of the pass. Another, parallel to this, lies about half a mile to the south of the wireless station, cutting Gebbies Spur in the vicinity of the sheep-yards which lie just north of Trig. R.; another lies across the middle portion of Thompsons Spur, connects up with the one just mentioned, and continues across the middle portion of McQueens Valley to the east, while yet another passes through the middle portion of Gold Spur, so that the rhyolite which terminates it lies on the southern wing of an anticline, and probably links up with that forming the portion of Gebbies Spur south of the homestead. As a result of the folding the beds exhibit a somewhat steeply dipping stratification, so that they occasionally resemble dykes in appearance and disposition, a fact noted by Haast (1879, p. 327), who was impressed by their wall-like character and columnar structure. This conclusion was formerly endorsed by the present author (1917, p. 374), but more mature opinion based on detailed observation is that these hogsbacks are merely tilted flows or ash-beds. The hogsback form of the mass crossing Gebbies Spur through Trig. R., and that of the knobs north-west of the pass are most striking.

The beds of this rhyolite as exposed near the pass consist of solid material, with interstratified fragmentary layers, and pitchstone which is presumably intrusive. The rhyolite is usually white in colour, sometimes pinkish, and very occasionally greenish. It is

stony in texture, with flow-lines clearly showing; these are very frequently puckered. It is at times beautifully spherulitic, as for example the outcrop west of the road on the south side of the pass. Lithophysae occur occasionally, a notable instance being the knob 25 chains north-west of the pass; and the rock is in places micro-spherulitic as at the quarry just inside the gate on the summit of the pass. The rock exhibits at times a marked silicification, which shows as irregular layers of quartz mosaic, and again in larger masses of chalcedonic silica.

One facies of the rhyolite is that showing phenocrysts of smoky quartz and sanidine, typically developed on the road over Mansons Peninsula (Speight, 1922, p. 80). This also occurs on the road north of the pass.

The rock is frequently jointed and brecciated as the result of earth movements subsequent to consolidation. The occurrence inside the gate at the pass illustrates this. The autoclastic brecciated rhyolite overlies ordinary breccia, the contact being at a steep angle and dipping south-west. This is so well defined that an intrusive origin is indicated, but there is no selvage of pitchstone, only clear-cut smooth contacts and slickensided surfaces, which suggest faulting or other earth movement.

The basal bed of rhyolite varies in different places. In some cases it is apparently solid rhyolite, but in the greater number of cases it consists of fragmentary material. This is of two types—(1) a purely rhyolite-breccia, and (2) a tuffaceous deposit, usually brownish in colour with a considerable percentage of andesitic and greywacke fragments. The former of these is most widely distributed especially to the east of the pass; in the vicinity of the wireless station it forms very thick beds. It also occurs on Potts Point and Mansons Peninsula. In the Gebbies Pass area it is whitish or pinkish in colour, and composed of angular pieces of rhyolite and pitchstone of all sizes up to 3 feet in diameter. As a rule there is a fairly large percentage of fine material.

There is a frequent occurrence near the base of the deposit of fragments of a special type of rhyolite. This contains reddish crystals which in many cases are mistaken in the hand-specimen for garnet; but when examined in thin section, they prove to be either grains of quartz stained with iron oxide or alteration-products after some easily decomposable mineral (? fayalite). In all the sections made I have never found a true garnet; of course, this does not mean that they do not occur, but if so they are rare.

The second facies of the fragmentary material is best developed west of the pass, although it does appear in places south of the wireless station. It is best exposed in the road-cutting about 25 chains west of the pass, and it appears in the upper part of the gully just north of this; round the northern side of the knob west of this it is exposed in places on the roadside, and again beyond the greywacke ridge at the foot of the steep slope of Lyttelton andesites and basalts which bounds the area in that direction. In the road-cutting in the first-named locality it is composed chiefly of brownish material, fine-grained, with fragments of greywacke, chert, and

andesite usually not more than 3 inches in diameter, and pieces of rhyolite with reddish minerals and pitchstone with well-developed flow-structure, these pieces ranging up to 8 inches in diameter. Were it not that this bed contains fragments of rhyolite, it might be supposed to belong to the upper portion of the McQueens Valley Series, and this is still possible if the existence of a pre-Tertiary or even an Early Tertiary rhyolite in the neighbourhood can be demonstrated. In the gully just north of the cutting in addition to the rhyolite in the breccia there are large fragments of pitchstone, in some places forming a notable percentage of the bed.

The exposures on the roadside round the knob to the west of the cutting are obscure, but the bed outcrops again near the western end of the greywacke ridge. In places rhyolite-breccia rests on grey-wacke and again on an erosion surface of the brown tuff, which dips 30° W. Both are affected by small faults in which the downthrow is to the north-west, and there is one reversed fault with an upthrow of about 25 feet in the same direction. Actual movement in these cases is indicated by a layer of pug on the contacts. The rhyolite-breccia is penetrated by trachyte dykes, and the brown facies by a sill, 18 inches thick, jointed across, very basic, and with xenocrysts of red garnet and hornblende.

One or two occurrences, formerly classed as intrusives, merit further reference. The first of these is the bed forming the northern slope of the knob about half a mile north-west of the pass. This dips N.N.E. at moderately high angles. Although well exposed on the hillside and in the road-cutting, its upper and lower surfaces cannot be seen anywhere. One or two features are however quite clear. First of all it is usually well laminated, the laminations being extensive and flat in places, and again frequently with contortions and puckering on a small scale. The dividing planes are close together—frequently less than an inch—and the rock splits freely along them so that it is almost as fissile as a slate, certainly as fissile as some shales (Pl. 38). The lamination is in places persistent for several yards, and bent up into anticlines and synclines. Although it is inclined usually at moderate angles with the horizontal, in places it lies almost, if not absolutely, vertical. This structure must be due either (1) to flow or (2) to difference in deposit in a mass formed of a shower of fragments, or (3) to pressure—direct or shear; but it is clear that whatever the condition of deposit the material was laid down in a viscous condition and experienced some movement while in that condition.

Secondly, the rock is now jointed (Pl. 38). The jointing is of two classes: (1) irregular jointing, such as might be produced in a rock as the result of pressure, that is, ordinary autoclastic brecciation; and (2) polygonal jointing. The first produces irregular angular fragments in a jumbled mass. This must have been impressed on the rock after it had consolidated, but it may have been done immediately after deposition, while the rock was still hot. The polygonal jointing is at first sight analogous to columnar jointing in basalt, and this led former observers to consider the mass to be intrusive. The columns are occasionally from 10 to 12 feet in length, and are usually from 4 to 6 inches in diameter. They are

generally five-sided, though four-sided and six-sided prisms also occur. They are irregular in direction as regards the planes of lamination of the rock. These planes are generally in excess of 60° with the axes of the prisms, and sometimes they are at right angles. The faces of the prisms are sometimes continuous with well-developed and somewhat regularly oriented cracks in the rock, 3 to 4 feet apart, while in the space between these cracks the faces of the prisms are not so oriented, but disposed as in the case of ordinary columnar jointing. The cracks gradually fade out, and then the columnar structure disappears also. The prisms are cut across by more or less continuous joints at right angles to or at high angles with their length, which continue as clear-cut planes for several yards and then disappear (Pl. 38).

It appears that both shrinkage on cooling and subsequent earth movements are responsible for the columnar jointing. Dr. Marshall has suggested to me that this might arise when a bed was formed by aerial deposition in a more or less plastic condition on a somewhat steep slope so that slumping took place differentially within the mass, thus causing puckering, and that the prisms were formed by the usual process by shrinkage proceeding at right angles to the cooling surface, this being variable in direction within the mass of the rock consequent on differential movement down a slope. This will certainly account for the general features of the jointing, but in addition there has been definite brecciation after consolidation. This can be attributed to earth movements which were responsible for the general deformation of the beds throughout the area. The mass has also been affected by small faults.

Under the microscope the rock appears to consist almost entirely of an even-grained matrix of feldspar crystals without definite form and indefinite granophyric masses all with lower index of refraction than balsam. There are also occasional small granules of quartz and numerous small shreds of mica. In this base are included larger grains of quartz and very occasional feldspar usually without definite outlines and appearing as if they were fragments broken from larger crystals. The quartz sometimes shows a corroded margin. The groundmass contains spherulites, some of which are hollow and others without any cavity; there are also numerous microspherulites and granophyric patches. The spherulites are frequently axiolitic in form, with the crystalline fibres at right angles to the elongated vesicles; the length of the fibres varies. The parting planes of the rock appear to run along these axiolites in many cases as well as along elongated cavities with no spherultic rays at right angles to them. At times too the rock appears to part along lines of crystalline fragments which are finer than the usual run and arranged in parallel layers. This finer mosaic suggests some acceleration in the normal rate of crystallisation of the rock mass.

The chemical composition of this rock is given in analysis 6 on page 319. This shows it to be a very acid rhyolite, and confirms the observations on thin sections.

There is the further case to be considered as to whether this bed is a flow or a tuff. In order to obtain an authoritative statement

I submitted slides to Dr. P. Marshall, of Wellington, and to Drs. H. T. Stearns and Powers, of Hawaii, and all concurred that the evidence from the slides was inconclusive. Also the field evidence is not definite, since one cannot see either the upper or the lower surface of the bed. However, on the knob about 20 chains S.E. of this occurrence there is a similar bed with the same orientation, indicating that the two parts now separated were probably continuous, but that the connection had been destroyed by erosion. This last occurrence shows the same puckering and lamination, also the polygonal and irregular jointing, and its lower surface passing down into an undoubted rhyolite explosion-breccia without any break whatsoever so that both seem to be part of the same deposit; the joints of the upper finer facies pass without any break into the lower. This lower bed is from 30 to 35 feet thick, and is not a flow-breccia or one formed as the result of earth movement, but undoubtedly one formed as the result of a rain of angular fragments, and the conditions support the contention that the finer beds developed here, and by inference that on the knob to the north-west, were formed as the result of aerial showers.

Underneath the breccia just referred to is one of finer grain about 6 feet thick, well stratified, striking N.W. and dipping N.E., that is, parallel to the overlying bed, but containing spherulites and lithophysae. Over it is a thin layer of pitchstone, and under it is a solid bed with flow-lines and puckered lamination. After an obscurity of some yards, breccia appears again also with thin layers of pitchstone. Owing to the irregularity of the beds and the covering of soil their relations are obscure, but it is very likely that the occurrence just referred to is the same as that mentioned earlier in this paragraph. These layers of pitchstone were at first considered as tongues proceeding from a closely adjacent mass of pitchstone, no doubt intrusive, which is to be referred to later, but it is also reasonable to think that they are only a facies of the fragmentary deposit laid down at a high temperature.

This is supported by evidence from an exposure on the road just below the easterly knob. The greywacke occurs a short distance away, and then after a few feet of obscurity there is a facies of rhyolite such as is typically developed on the road over Mansons Peninsula (Speight, 1922, p. 80). This passes up without a break into a stratified tuff suggesting a common origin for both, as is the case just referred to from near the summit of the knob. This tuff is well stratified with outstanding crystal-fragments of quartz and feldspar. Streaks and thin layers of pitchstone ¼ inch thick and extending laterally for 18 inches to 3 feet, also take part in the stratification and in such positions that they cannot have been formed in any other way than by having been deposited in a liquid condition or at such a high temperature that they consolidated into a glassy form immediately on deposition. They are not intrusive, and isolated fragments of pitchstone also occur, these in all probability of different origin. The first facies of pitchstone confirms the hypothesis that the tuff had been laid down while the materials were at a high temperature.

Rhyolite with corrugated and plane lamination occurs on the south side of the summit road 20 chains from the summit of the pass, and also in the quarry just inside the gate at the top of the pass. It is not maintained that these are all part of the same deposit, though this may be true, or that the time of deposition was synchronous, but the conditions of deposition are substantially the same.

In this connection an examination was made of the rhyolite exposed on Potts Point. Along the western shore-line of this peninsula the exposures are excellent, and the rock is everywhere of fragmentary nature. At the terminal portion, for a distance of 60 chains, the beds are entirely stratified tuffs striking approximately E.-W. and folded so that the dip is alternately N. and S. at moderate angles. They show crystals of quartz and feldspar in a finer matrix. Fragments of chert, greywacke and the mudstone facies of the Charteris Bay Series occur all through and not merely at the base of the deposit. This emphasizes its tuffaceous origin. In one place streaks of pitchstone like those on the roadside near Gebbies Pass were observed, and occasionally spherulitic structure. In some places the stratification is distinct and can be seen clearly on the hillside, but in most cases this is only visible where the rock is exposed at the shore-line and the fragments of quartz and feldspar stand out on the weathered surface. It should be mentioned that some difficulty in interpretation arises from the network of trachyte dykes which penetrate the tuff, but it is clearly evident that the deposit is of great thickness—certainly some hundreds of feet—uniform in character, and with no indication of the presence of flow rocks. It does not show any columnar structure, but is well jointed at right angles to the stratification.

On the proximal part of the peninsula the shore-line shows a development of coarser breccia, but with large masses of finer tuff included irregularly in it. These masses are so large that they must have been formed in position, and the explosive origin of both facies is clear. The breccia is not a flow-breccia formed under a lava stream. It rests on sandstone directly.

On the eastern side of the peninsula the exposures are not so clear owing to the labyrinth of trachyte dykes which pentrate the rhyolite, but facies occur similar to those described, with the coarse one resting on sandstone. There is the same inclusion of underlying sediments, and there is no occurrence which can definitely be asserted to be a flow. It might be mentioned that the exposures of the rhyolite on Quail Island, which is in close proximity to Potts Point, are similar in character.

An examination of the shore-line of Mansons Peninsula was also made. The proximal portion gives evidence similar to that from Potts Point. There is the same stratified tuff with fragments of feldspar and quartz crystals freely showing, the same rapid alternation of the direction of dip, the presence of undoubted foreign fragments, and the passage of breccia into the finer material without a definite break. On the distal end of this peninsula the occurrences become complicated, and while the fragmental origin of certain beds is clear, that of others is not so.

However, the general evidence from these two peninsulas, taken in conjunction with that from the roadside on the pass, is strongly in favour of the facies of the rhyolite as exposed on the roadside over Mansons Peninsula being also tuffaceous in origin. The smoky quartz and feldspar fragments are the same in all cases. There is little or no polygonal jointing in the Potts Point and Mansons Peninsula shore-line tuffaceous rhyolite, while that on the road over the latter peninsula does show it, the prisms being approximately a foot in diameter. It may be noted here that the Stratified Tuff of the Brisbane area, as described by Richards and Bryan (1934), shows no polygonal jointing, and in that respect resembles the tuff of Potts Point, whereas their massive tuff does show this structure. The bed on the summit of the Gebbies Pass ridge does show lamination, which may be looked on as a form of stratification and also shows polygonal jointing. In both this case and that of the stratified facies of the Brisbane Tuff there is no definite evidence of high temperature at the time of deposition (loc. cit. p. 55) although there are features which are suggestive. I have seen no carbonised fragments such as characterise the latter, but the spherulitic structure and the occurrence of interstratified streaks of pitchstone are suggestive of a high temperature.

These appear to me to be the only lines of evidence bearing on the question of the temperature at which these tuffs were deposited. There is none of the vitric material such as is described by Marshall in connection with the tuffs of the North Island and by Richards and Bryan in connection with the Brisbane Tuff, for the groundmass of all rocks microscopically examined is definitely crystalline. Dr. Marshall kindly lent me an extensive suite of his slides of North Island tuffs, and the difference is most marked, the only one of the series at all resembling the Gebbies Pass rock from the western knob being a rhyolite from Atiamuri. But the frequent occurrence of spherulitic structure and lithophysae in doubtful rocks as well as in those which are definitely fragmentary seems to point to their not having been laid down in the cold, but to indicate that they have had sufficient heat to allow the re-arrangement of molecules after the rocks had fallen.

In making this statement concerning the tuffaceous origin of these rocks, it is not maintained that all the rhyolites have this origin, but that certain rocks formerly considered flows are definitely tuffs, while others are probably tuffs, and this materially reduces the volume of material in this area which must be credited to a flow origin. It may ultimately be decided that all the rhyolitic material is tuffaceous.

Then there is the pitchstone. This occurs freely as a constituent of breccia near the base of the series, but appears absent from those higher up. The most notable occurrence in position is on the south side of the knob N.W. of the pass, where there is an oval exposure nearly 100 feet in width and some 4 chains in length in an E.-W. direction. This is flanked on both sides by pure rhyolite-breccia and on the northern side this is interstratified at two levels, perhaps at more, with layers of pitchstone up to a foot in thickness. These have been referred to earlier.

syncline, with axis running E.–W., and the andesite has also experienced the same movement. But nowhere are the typical Lyttelton rocks involved in folding. So there is a prima facie case for the hypothesis that there lies between the rhyolite on the one hand and the andesites and basalts of the Lyttelton series on the other, another andesitic series, not the mere basal facies of the latter. Further, on the western shore of the peninsula there is a purplish fragmentary rock associated in places with a thin sheet of olivine-augite-andesite interstratified with white rhyolite and exposed discontinuously for half a mile near sea-level, generally near the crests of the anticlines into which the rhyolite has been folded. Andesite shows clearly at the end of the spur leading up to Trig. KK, and in the little island off the shore. These rocks are not intrusive, but are definitely inter-stratified, probably at various levels, the lowest being low down near the base of the rhyolite series. The occurrences suggest either a change in the composition of the material thrown out from the volcano, or more likely the overlapping of the material from adjacent vents discharging material of different composition. The beds are very thin, and apparently thin out still more when traced into the ridge. There are no similar exposures on the eastern shore, the only occurrence being that near the point which is definitely synclinal, and at a higher level.

The feldspar phenocrysts of the andesite are a medium labradorite, the augite is pale brown in colour, and the olivine sometimes fresh, and again serpentinized. Grains of magnetite occur, and the pilotaxitic groundmass consists of small, thin laths of andesine. In some of the purplish fragmentaries sanidine crystals show as well as the plagioclase, and the rock seems to be intermediate between an andesite and a rhyolite.

Another occurrence of a more basic facies than the ordinary rhyolite lies on the shore at the southern end of the peninsula, where a peculiar mass of fragmentary material appears on a small projection of the shore-line near the end of the fence running out into the flat. Under the microscope this shows broken crystals of plagioclase and fairly numerous fresh and corroded crystals of biotite, but no sanidine or quartz. The rock has weathered brown, but passes into white rhyolite without an apparent break, but its relations are somewhat obscure since the exposure is only a few square yards in area and is bounded by faults, one with a well-marked fault-breccia. It seems, however, reasonable to place it with the andesite occurring interstratified with the rhyolite near the end of the peninsula. There is no doubt that it is an ash-bed.

This occurrence of andesite suggested that the rocks on the foreshore of Governors Bay, usually supposed to belong to the basal portion of the Lyttelton series, might have an analogous position. They are exactly similar petrologically to the rock on the point of Mansons Peninsula, and they do not show the outward dip from the centre of the volcano characteristic of Lyttelton rocks. Along the southern stretch of the road along the foreshore they appear to dip towards the harbour at low angles, but round the point immediately south of the wharf they are involved in definite folds, to be seen clearly in the fragmentary bed which underlies the lava-flow quarried

for road-metal. These folds cannot be credited to local slumping of the fragmentary beds. Though irregular locally, they appear to have a general N.N.E.-W.S.W. or E.-W. trend, and their axes would pass in the vicinity of the end of Mansons Peninsula if they continued to the east. Some of the folding is acute and attended with faulting. A definite fault occurs on the south side of the point at the wharf, and the lava-flow quarried for road metal has been thrown down till it butts against the underlying fragmentary. When considering the question of faulting, due regard was paid to the necessity of demonstrating actual movement at these discordant junctions, and in the case just mentioned, if evidence of movement did not exist, it might quite well be thought that the lava-flow had filled a gully eroded in the underlying fragmentary. Evidence of folding appears on the foreshore north of the point where the large trachyte dyke cuts the road; it is thus not a mere local phenomenon. If therefore we assign these rocks to the Lyttelton Series we have to postulate a folding and a faulting movement contemporaneous with the initial stages of Lyttelton vulcanicity, a movement which apparently ceased and did not affect the overlying rocks belonging to that series. It seems more satisfactory, however, to assign these rocks to an intermediate episode.

It is unfortunate that owing to the covering of soil the contact of these rocks with the overlying Lyttelton Series can never be seen, but they appear to form the well-marked shelf which extends along the western side of Governors Bay, and which has been dissected by the streams running into the harbour from that direction.

One point should be mentioned with reference to the fragmentary beds near the wharf, viz., they contain numerous amygdaloids of chalcedony and occasional fragments of jasper. In this they resemble the occurrence on the end of Mansons Peninsula, and differ from Lyttelton rocks, which never show such amygdaloids; if they do occur in the latter, they are always of calcite, arragonite, or chabazite. These siliceous amygdaloids would connect the occurrence with the McQueens Valley Series, but this definitely underlies the rhyolite, whereas the occurrence at Mansons overlies it or is interstratified with it, and the Governors Bay set of beds is more definitely connected with the latter than with the former.

The foreshore was examined in the direction of Lyttelton to see if these beds continued in that direction, but in every case the dip was radially outwards from the centre, and the facies of the rocks definitely that of the Lyttelton Series. This is very clearly seen at the Harbour Board quarry near the dock. However, in the section disclosed by the Lyttelton Tunnel, as recorded by Haast, there is an appearance of beds dipping definitely towards the harbour as far as about 10 chains from the Lyttelton end. Haast notes this, but accounts for it by supposing the beds to have been plastered on the inside of the crater. This explanation of course supposes that the point where they occur was in the vicinity of the wall of the crater when the volcano was in eruption, a supposition which is extremely improbable when the enlargement of the crater by explosion and by subsequent erosion is allowed for. It seems to me possible that these rocks may be, at least tentatively, considered