the dyke where it cuts the road to the north-east. A similar feature is referred to later in the case of another multiple dyke, where the order of intrusion is definite.

A little further on is a 12 ft. dyke of basalt, and then a complex occurrence which presents some difficulty. On the east side there is first of all 7 ft. of andesite, reddish in colour, and jointed parallel to the walls of the dyke, but faulted and crushed, with inclusions of crush breccia; then 6 ft. of andesite showing little disturbance and with cross-jointing fairly well developed; this is succeeded by about 3 ft. 6 in. of country rock of wedge shape; and finally there is 6 ft. 6 in. of dark andesite with an interior zone showing prophyritic crystals of feldspar. This last also contains slicken-sided surfaces and gives evidence of movement. I think that this is one large dyke with an inclusion of country rock, but the suggestion is advanced with all diffidence.

There is other evidence of rock-movement in the locality and some of the less competent fragmentary beds show slickensided surfaces. These I consider to be due to movement within the beds themselves and not to change affecting the hillside as a whole, the phenomenon being analogous to the failure of unfaulted coal and the production of soft coal.

There is a large trachyte dyke (18 ft.) near the saddle where Bridle Track crosses (fig. 2). This appears to be thicker than 18 ft., but the discrepancy is due to a flange-like expansion of the dyke on the east side. Further on is the great dyke known as Castle Rock, which forms the backbone of the ridge bounding Heathcote Valley on the west. It does not appear on the road in proper alignment, for it takes a pronounced turn when approaching it and cuts the road as two dykes to the east of the ridge, with a possible splinter on the west side.

Section 3 extends from the great dyke at the head of Raupaki Valley known as the Giant's Causeway to Dyer's Pass. It contains several large dykes including the Causeway itself (35 ft.). This is a trachyte and stands up as a great wall, and so has little resemblance to the classic occurrence in the north of Ireland. Considerable lengths of this part of the road cut in solid rock show few or no dykes, and this is specially the case where massive flows of andesite have to be broken through, such as those which occur north of Trig. AA, and on the northern slope of the Sugarloaf. In connection with this it is to be generally noted that, with some striking exceptions, dykes are more common on the saddles and less common on the ridges which mark the edge of the crater-ring. It may be that the harder rock has determined the position of the radial ridges and intrusions found this difficult of penetration, or perhaps that the saddles with their numerous dykes were more susceptible to weathering and were therefore more rapidly reduced.

A large hornblende trachy-andesite dyke greenish grey in colour on the margins and brownish grey in the interior occurs west of Trig. AA, and a large dyke (25 ft.), a little further on along the road, has margins 10 ft. and 8 ft. thick with an internal core of

basalt 7 ft. thick, the basalt being apparently the later intrusion. Another trachy-andesite dyke (15 ft.), just east of the Sugarloaf ridge, is interesting for containing inclusions of hornblende diorite up to 12 in. in length as well as of hornblende as aggregations and in individual crystals, a striking instance of the association of large hornblendes with rocks of this class. Some of the dykes have an internal zone, brown in colour, and speckly with feldspars, with a much finer grained dark margin.

Thirty-six dykes were measured in this section, having an aggregate thickness of 220 ft., which gives an expansion of the cone resulting from their intrusion of 2.1 per cent.

Section 4 from Dyer's Pass to Cass Peak has 76 dykes with a total thickness of 398 ft., indicating an expansion of 3.3 per cent. This is lower than the real value since the road is cut in clay or loess for considerable distances and intrusions, if existent, do not show. The visible ones range up to 39 ft. in thickness. This section is remarkable for the occurrence of large dykes of trachytoid character. The first to be encountered is about half a mile from Dyer's Pass. It is an excellent example of a multiple dyke. Before the road was widened the following succession was to be seen at this point:—

In this section the repetition of vi and viii is clearly due to faulting, as the slickensided surfaces show. The strike of this fault is almost parallel to that of the dyke, but in this case the faulting is post-intrusion, so that it can hardly have been responsible for the intrusion unless the present fault marks the reopening of an old fissure. As a result of the widening of the road the section has become different. No. ii has thinned out till its maximum thicknes is only 6 in.; iv shows as before; v has thickened to 9ft; and there are no wedge-shaped masses of andesite (vi and viii) in this part but one intrusion 2 ft. 9 in. thick. The zone of platy jointing in the trachyte on the east side is still 8 ft. thick, but on the west side owing to repetition due to faulting it now shows a total of 12 ft., with the andesite 2 ft. 9in. intruded parallel to the jointing (fig. 3). In the quarry a chain or so south-east, and off the road, this andesite does not show.

The order of succession in this dyke is that the andesite was intruded after the trachyte had been consolidated. On both sides the trachyte exhibits a marginal facies with platy jointing and an interior zone with columnar cross jointing, and the intrusion has followed up the plane of weakness at the contact of the two types. The fact that the junction of the two types is a plane of weakness is well exemplified by the basaltic dyke which lies about four yards to the east of this dyke. It is of variable thickness, but averages about 4 ft., with an internal zone of cross jointing, and the contact between this and the outside platy jointing definitely fluted on both sides, and also the fluting is repeated at varying distances from the margin of the dyke within the zone of the platy facies. This indicates some movement of the middle zone while the exterior was relatively stable and the rock was sufficiently hard to retain markings impressed on it. The surfaces are not polished in any way, so the rock was not quite solid, but probably in a viscous condition. The fact that successive plates of the external zone are also fluted upon themselves suggests that platy jointing may be partly due to flow and not entirely attributable to chilling and consequent contraction of successive layers of cooling rock.

Similar features were noted in basaltic dykes elsewhere round the road, and they are apparently analogous to the occurrence near Auckland noted by Bartrum (1928, pp. 23–5), and to the horizontal flutings in dykes in the Hopi country, Arizona, as recorded by Howel Williams (1936, pp. 121–2).

The large Hoonhay trachy-andesite dyke is cut by the road some 20 chains further on. Where quarried above the road it is 19 ft. thick, but on the road this is reduced to 14 ft. This dyke is one of the most persistent to be noted in the district, as it can be traced for three miles down the outer slope of the cone till its possible further extension is masked by the gravels of the Canterbury Plains.

Further on there are several multiple dykes, a notable one showing on the road as it passes round the spur leading up to Mt. Ada. This shows the following sequence in order from north to south:—(i) trachyte, 11 ft.; (ii) country rock, 6 in.; (iii) trachyte, 7 ft., with inclusions of country rock; (iv) country rock, 18 in.; (v) trachyte, 7 ft. It is not clear whether the inclusions of country rock, which are wedge-shaped, are due to faulting or whether they are masses caught up during intrusion, or yet again whether they have been isolated from their parent rock by the splitting of the intrusion.

A little further on down the cutting is a basaltic dyke (5 ft.), then after a short interval comes a very large greenish trachyte (39 ft.), which is similar in mineral composition to the large dyke quarried many years ago for building stone at various points below the road to the west. Although this has a thickness in places of 60 ft. lower down the hill it does not appear on the road where it should do so. They are both trachytes with soda-amphibole in addition to the soda-pyroxene, and they both contain inclusions of the andesite country rock. It is clear that the fissure up which the dyke

material came was not continuous, but sent off a splinter from the main direction; a clear case of side-stepping. Reaching down the ridge from Cass Peak is a fairly large dyke (10 ft.) of apparently related character, but it is much more basic in composition, although it may have been injected at the same time.

The dominant material of the dykes of this section is trachytoid; basaltic dykes are usually small, though they are perhaps more numerous than the more acid variety.

Section 5, stretching from Cass Peak to Coopers Knobs, includes about one-third of cutting in loess or clay with consequent obscurity. However, 26 dykes were measured, having a total thickness of 137 ft., which gives an expansion of 1.4 per cent. If all of the dykes could be seen, it would be much higher, probably over 2 per cent. The material is mostly trachytic, and none of the dykes are large, the largest being only 12 ft. in thickness; at least one of them is multiple.

In section 6 the road follows the inside of the crater-ring and gradually falls from 1500 ft. to about 800 ft., where it meets the rhyolite and greywacke of the Gebbies Pass ridge. It has been cut in solid rock almost continuously for approximately three-quarters of a mile, so that the sequence of dykes is reasonably complete. Altogether 41 were counted, 30 of these being in the upper part of the road, with a definite falling off in number as the road becomes lower in level, a point to be noticed seeing that this part of the road is nearer the internal base of the cone where more intrusions are to be expected.

The largest dyke measured in this section was 16ft. in thickness; one or two multiple dykes were seen; they are mostly basaltic. The total thickness is 177 ft., giving an expansion of 3.2 per cent. This high value can be plausibly explained by the fact that the road is here at a lower level than in other parts of the crater-ring and dykes which could not reach the upper levels have been cut by the road, but this cannot be maintained for certainty since they are relatively infrequent in the lower parts of the road, and, further, there is a tendency all round the crater-ring for the dykes to be bunched in certain sectors. Although dykes do occur in the greywacke and rhyolite overlying it further along the road towards the summit of Gebbies Pass, they were not considered in this examination since their circumstances are somewhat different in that they do not belong to the cone itself as they show at present.

To summarise, the number of dykes cut by the road and measured totals 220, with a combined thickness of 1233 ft., giving an expansion of the cone corresponding to road level of 2.13 per cent., an amount which is of the same order as that given by each section indicated previously. Allowing for parts of the road which are obscure, the total expansion of this height of the cone is not likely to exceed 2.5 per cent., but it will be in the vicinity of that amount. There is not definite evidence available for the expansion of the cone at lower levels, but it will certainly be greater nearer to the centre of eruption and considerably less for those portions of the periphery at lower levels.

The only records with which I can compare this amount of expansion are those given in the case of the dykes of the islands of Arran and Mull on the west coast of Scotland, and these are not analogous in that they deal with parallel dykes and not those radial to a cone. The value in the case of the former is 7.1 per cent. over a breadth of 14.8 miles, and in the case of the latter 3.8 per cent. over a breadth of 12.5 miles. Of course, it is possible in both cases that the crustal adjustments necessary for these intrusions have taken effect over a greater breadth than that where the dykes are actually located, and for comparative purposes the percentages are perhaps to be slightly reduced, whereas in the Lyttelton case the circuit is a closed one; and, further, owing to the presence of obscurities the value given is a minimum one. It should be mentioned, too, that the number of dykes on the south side of Quail Island necessitates a much higher expansion, estimated from 5 to 10 per cent., perhaps more, but this affects only a limited area of the floor of the caldera.

I have not been able to compare the conditions of Lyttelton with those of any other volcanic cone with the exception of Vesuvius and Etna. In the case of the former the dykes are irregular in orientation, whereas in the case of Lyttelton, with the exception of the dykes near the centre of intrusion, they are quite radial. In the case of Etna, Lyell mentions radial dykes, and the statement is amplified by Geikie, but sets of parallel dykes are mentioned as well, so that the case is not quite analogous. The only occurrence in Great Britain which might be cited is the island of Rum with its beautiful arrangement of radial dykes, but the sequence includes plutonic rocks of various types and the island is not a simple volcanic cone. In the case of the island of Mull the structure is complex, with cone-sheets, ring-dykes, and a parallel arrangement of ordinary dykes, although in some cases there may be an approach to radial orientation.

In America, too, the only cases recorded to date that I am aware of which might be considered in this connection are those of the Spanish Peaks, the Shiprock in the Navajo volcanic field, and the Highwood Mountains. The first of these is a deeply eroded double stock with plutonic rocks (Knopf, 1936), the second a volcanic neck with a small number of radiating dykes (Howel Williams, 1936), and the last, according to a private communication from Professor Daly of Harvard University, a complex assemblage of dykes round a number of centres of eruption including at least one laccolite. They are thus not analogous to the case under consideration.

The general explanation of the mode of formation of volcanic dykes as given by Barrell, Iddings, Daly, and others is that they result from fissure formation following on stresses in the cone, and that these fissures have been filled with liquid material injected from below, or laterally by hydrostatic or other pressure from the central crateral cavity or from some reservoir beneath the mountain. No completely demonstrable explanation has been advanced to account for the regular orientation, though suggestions have been made of its association with (1) faulting on radial lines, (2) pre-existing

major fault fissures, (3) fissuring on radial lines, (4) that the magma has just made its own way. A corollary to any of these modes of formation is the suggestion made by Barrell, Daly, and others that the injection took place rapidly, for if it had been slow the magma would have cooled and so lost mobility.

There is no evidence available from Lyttelton which indicates the presence of major faults or faulting on radial lines prior to intrusion. Although faulting is associated with a few of the dykes the movements have taken place subsequent to intrusion. The dykes occasionally show local irregularities in direction while preserving their general orientation, and the inclusion not only of small pieces, but of large masses, of country rock are both features which support the contention that they are associated with fissures and not with faults, and the latter observation suggests that the formation of the fissure and the intrusion of the dyke material were synchronous. The side-stepping of dykes and local irregularities in direction seem to indicate that the dykes themselves largely determine the formation of fissures through which the liquid material has been forced.

An expansion of the crust amounting to nearly 2.5 per cent. suggests that dykes may be located on tension ruptures, but the intrusion of the material from some crateral or sub-crateral cavity under internal strains may be responsible for the fissures as well as for their subsequent or simultaneous injection. It is well known that doming up of a volcanic cone is a common precedent to eruption, and this must cause tension in the mountain. From this arises, therefore, the force responsible for the fissuring and probable synchronous injection. It may be considered as certain that an expansion of the cone took place on many occasions during the period of active vulcanicity, but it is also likely that bunches of dykes with similar or related composition were formed synchronously in a limited sector of the cone, so that it would follow that the number of injection periods might be considerably less than the number of dykes, and may indeed have been only a small fraction of the number.

One point should be noted, viz., that the orientation of the dominant joints in the lava flows is the same as the strike of the dykes. These joints are no doubt due to the formation of tension ruptures accompanying shrinkage on cooling, and this will take place chiefly from the sides of the flows, and as they are generally radial so will the dominant joints be radial too; it is of course recognised that there will be joints with other orientation. But the dominant radial joints provide planes of weakness which may have some effect in directing the paths of intrusions. It is clear that joints do so operate in the case of multiple dykes as noted earlier, and instances were seen where the direction of single dykes coincided with that of dominant radial jointing in closely adjacent rocks, especially when the flows are massive and naturally somewhat resistent to intrusions. These cases may after all be chance coincidences.

The position of the locus of the origin of injection must have been at some depth, and not a mere surface crateral manifestation, for the material of the dykes is not directly related to that of the

• 3. More basic trachytes, with light green-grey or deep green colour containing few phenocrysts.

• 4. Trachy-andesites, greenish or greyish in colour, frequently with brown hornblende and olivine.

• 5. Andesites, basic in character, and grading into 6.

• 6. Basalts, some acid and others very basic; in some of the basic forms hornblende is a notable constituent, and they appear to have some relation to the trachytes.

The hornblende is perhaps the most striking mineral in these rocks. It first of all occurs as large brownish or brownish green crystals and aggregations, the crystals ranging up to 10 cm. in length, but usually less than 1 cm.; they are sometimes unaltered, but occasionally are resorbed partially or completely. Another type of amphibole is a soda variety resembling arfvedsonite, which occurs sparingly as micro-phenocrysts in some of the more alkaline types. The augite which is most common is a greenish, slightly pleochroic aegirineaugite which occurs as phenocrysts, microphenocrysts, or disseminated through the base as grains and in stumpy crystal forms. A greyish, slightly purple, not appreciably pleochroic augite also occurs in subordinate amount. The olivine is sometimes unaltered but usually stained with iron oxides along cracks and marginally, and occasionally it is entirely replaced by iddingsite.

The feldspars are usually of the acid type, basic forms occurring only in the basic andesites and basalts. They frequently show zonal structure and denticulate margins, the last feature being specially characteristic of the base of the more acid rocks. Some of these denticulate feldspar laths have a lower index of refraction than balsam and may therefore be anorthoclase especially in the sodic rocks, but in other cases the index is higher than that of balsam and the minerals must be oligoclase or andesine, most likely the former.

It has not been found possible to section all the dykes, but fifty have been examined microscopically, and to the best of my knowledge no important type has been omitted. I quote the analyses kindly supplied by the Dominion Laboratory, and none previously published are listed here, but all that have appeared to date are quoted by the present author (loc. cit., pp. 130–1) in his paper on the Intrusive Rocks of Banks Peninsula. These recent analyses furnish a useful supplement to the list there given.

A perusal of the list given herewith indicates the presence of trachytoid rocks with a moderately high percentage of soda, which is accounted for by the presence of oligoclase and andesine and by aegirineaugite. The basic rocks do not show any special abnormality in composition, except a fairly high percentage of titanium. The water content of the specimens submitted is very high considering that all were obtained from rock cuttings and every care was taken to get as fresh a sample as possible; in some cases some amount of decomposition appeared to be inherent in the rocks as if it contained some percentages of unstable mineral.

[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. Trachyte, Dyer's Pass ridge, I″. 4. ″2. 3. (4), Toscanose.

No. 2. Trachyte, Cass Peak, (I). II. 5. 2. ″4., Akerose.

No. 3. Soda Trachy-Andesite, Dyer's Pass ridge, II. 5. 2. 4., Akerose.

No. 4. Basic Andesite, occurring as an intrusion in No. 1., Dyer's Pass ridge, II. 4″. 3. 4., Tonalose.

No. 5. Basalt, Mt. Ada, II (III). 5. 3. 4., Andose.

No. 6. Feldspar Basalt, near Hoonhay Dyke, (II) III. 5. 3°. 4°., Camptonose.

No. 7. Dolerite. 4th dyke down cutting from Coopers Knobs, ″III. 5. 3″. 4., Camptonose.

The most acid trachyte encountered is the soda-amphibole trachyte of Kennedy's Bush (Speight, 1908, p. 176), and it should be noted that this rock contains a considerable amount of soda-augite in addition to the micro-phenocrysts of the amphibole.

The most common type of trachyte is brown in colour, and there is also a frequent occurrence of a stain of oxide of iron which occasionally appears to be developed along spheroidal cracks and ordinary joints. A typical example is the large dyke (fig. 3) on the summit of the Dyer's Pass ridge. Crystals of feldspar show plainly in the hand specimen, and under the microscope they prove to be andesine with very occasional sanidine. The base is composed of short stumpy laths of sanidine, and there are as well a fair number of forms with denticulate margin and nearly straight extinction, both forms having a lower index of refraction than balsam, so anorthoclase is probably present in addition to sanidine. In between the laths are numerous grains of magnetite, some of which are no doubt derived from the augite. Towards the margin of the dyke the texture of the base becomes progressively finer, and this applies to the part with platy jointing, while at the margin of the dyke there is a selvage which is definitely glassy. In this facies the feldspar phenocrysts become more numerous and are apparently entirely of andesine; greenish crystals of aegirine-augite also occur, and the glass has developed well-defined perlitic cracks. This facies is the chilled margin of the mass and the increased proportion of phenocrysts is noteworthy. Analysis No. 1 is from this dyke. This is similar in all respects with the analysis of a dyke from Heathcote also cutting the Summit Road (Speight, 1923, pp. 130–1). I have little further comment to make on this rock except that the slides of specimens taken from the summit of the ridge show very little hornblende, no mica, and much more aegirine-augite as phenocrysts and microphenocrysts, and also there is an approach to a bostonitic habit in the base (fig. 5).

The intersection of Castle Rock, the large dyke that forms the crest of the ridge on the western side of Heathcote Valley, shows variations which are characteristic of the main dyke. First of all the interior is definitely trachytic, the feldspars being chiefly sanidine and occasional andesine; the microphenocrysts are of aegirine-augite; and the base of long laths of feldspar showing flow structure and also radiating sheaf-like forms (fig. 6), among which are small grains of greenish grey augite and magnetite. The feldspars have lower index of refraction than balsam and most of them show straight extinction, but there are a number which have an angle of extinction for albite, so in all probability both are present.

In a 6 ft. splinter of this dyke the features are similar except that the augite phenocrysts are definitely purplish, and olivine occurs unaltered and again replaced at times by iddingsite. The base is composed of microlites, some with straight edges and others with denticulate margins both with index less than balsam and with extinction angles of sanidine and albite. The texture is definitely bostonitic and it is probable that anorthoclase is present as well;

there are also frequent patches of tridymite. As noted previously (loc. cit., pp. 136–8) this dyke is very variable in character both along the strike and across it.

The type of trachyte mentioned first occurs as dykes at Giants Causeway, further west near the 4 mile post, on the eastern side of the Sugarloaf, on the ridge leading up to Mt. Ada, and elsewhere.

Basic Trachytes.

In addition to these normal trachytes there are others morebasic in character, and it is possible that they might be assigned to the trachy-andesites, andesine occurring as phenocrysts more commonly than sanidine. These rocks are usually grey or greenish grey in colour, an excellent example being that quarried on the north side of the road passing over the shoulder of the spur leading up to Cass Peak. This is very vesicular, sometimes showing in the hand specimen white crystals of feldspar and very occasionally small crystals of hornblende and also olivine. Under the microscope it shows phenocrysts of andesine and microphenocrysts of sanidine. The base is holocrystalline and consists chiefly of feldspar laths from .3 to .4 mm. in length, some with straight edges and others markedly denticulate; the index of refraction of some of the laths in about that of balsam, some a little lower, and the angle of extinction measured with the length is generally straight or nearly so, but some of the laths are twinned and show an extinction angle which with the index of refraction just below balsam shows them to be albite-oligoclase. The untwinned laths show straight extinction and are no doubt sanidine, but the denticulated margins suggest the presence of anorthoclase, though it was not determined for certain (fig. 7). In between the laths are numerous short crystals and laths of aegirine-augite, green-grey in colour and frequently stained with decomposition products of iron-oxide. Neither hornblende nor olivine appeared in any of the slides made. The chemical composition of this rock is given in analysis No. 2, which shows that it has a fairly high percentage of alkalies with dominant soda and a moderate amount of lime, which is in keeping with the absence of a basic feldspar. Composition, mineral content, and texture therefore indicate that the rock should be called a trachyte. It should be noted that the high percentage of water shown in these analyses reduces considerably the importance of other constituents in the percentage composition. The dyke which occurs just east of the main trachyte dyke on Dyer's Pass ridge (fig. 3b) might possibly be placed in this category, but the composition is definitely more basic, and it will be best to assign it to next section.

Trachy-andesites.

The rock just referred to belongs here. It occurs just east of the large trachyte dyke on Dyer's Pass ridge referred to earlier, marked b in fig. 3. In the hand specimen it is dark grey in colour but with a suggestion of very dark green as it occurs in the road cutting. On the joint surfaces it shows brown owing to a stain of iron oxide. It is generally slightly vesicular and in parts markedly so. Under the microscope it appears to be composed chiefly of

prismoids of feldspar with denticulate margins and of lengths from .3 to .4 mm., and with index of refraction slightly higher than that of balsam. The extinction angle is very low, so they must be oligoclase, with probably more acid outgrowths, the whole texture being strongly reminiscent of bostonite (fig. 8). Some few of the prisms of the base have an index of refraction lower than that of balsam and with approximately straight extinction, so some sanidine or anorthoclase is probably present. In between the feldspar laths are small grains and stumpy crystals of aegirine-augite, sometimes stained brown with oxide of iron, and there are grains of magnetite as well as some brownish alteration product. In this base are very occasional microphenocrysts of feldspar with refractive index higher than balsam, an extinction angle, as observed, of andesine, but the feldspar may really be more basic. The chemical composition of the rock is given in analysis No. 3. The importance of soda as compared with potash and the high percentage of water are the two most striking features. The former is reflected by the amount of albite in the norm, and the lime is reflected in the amount of anorthite. The rock has certain relations to andesite on account of the oligoclase in the base, and to trachyte by the presence of the more acid feldspar, even if in small amount. There is a close resemblance in composition to other undoubted trachy-andesites of the area, such as those at Evans Pass, Castle Rock, and Hoonhay (see Speight, loc. cit. pp. 130–1) and the group is extremely well represented. They are usually rocks of light colour, grey or greenish grey, as a rule, and show freely crystals of hornblende which reach 1 cm. in length, though they are usually less. Under the microscope the phenocrysts are hornblende, augite, plagioclase, olivine in that order of importance, the plagioclase being either andesine or labradorite. The hornblende is brown with a faint tinge of green, at times resorbed on the margins; the augite is grey and then again faint purplish in colour, the olivine fresh or altered to iddingsite, either as crystals up to 1 mm. in diameter or as smaller grains between the feldspar of the base; microphenocrysts of sanidine also occur. The groundmass is composed of microlites of sanidine either quadratic in section or in lath-shaped forms, and there is an admixture of forms with higher index of refraction than that of balsam and with the extinction of oligoclase-andesine, and frequently with denticulate margins. In among them lie stumpy forms of greenish augite and grains of magnetite.

The rocks are related to the ciminite of Washington, but they have too much soda in relation to the potash, too much olivine, and the feldspar is somewhat too acid as well; it will therefore be best to call them merely trachy-andesites. They furnish another instance of the association of hornblende with rocks of this class.

A good example (13 ft.) can be seen on the rock cutting east of Heathcote Valley, another (10 ft.) on the saddle east of the Sugarloaf, the latter showing numerous inclusions of large hornblende with and without matrix attached, as well as a foreign block of granitoid rock 12 inches in diameter. This inclusion is formed chiefly of brown hornblende but contains some diallage as well, a little

biotite, and frequent crystals of apatite and ilmenite, and a subordinate amount of granular feldspar, the granular form being apparently the result of pressure; some of this is andesine, and none was observed with a lower index than that of balsam. This rock is in all probability a diorite or diorite gneiss and may be the deep-seated equivalent of the trachy-andesitic material of the dykes just referred to.

Basic Andesites and Basalts.

Typical andesites are apparently not represented but basic andesites verging on basalts occur freely. They are found first of all as narrow dykes, dark in colour, and frequently as one of the members of a multiple dyke. A case illustrating this is furnished by the assemblage associated with the large trachyte dyke on the Dyer's Pass ridge about half a mile west of the pass. In the hand-specimen this is dark coloured with a fine-grained base in which crystals over 1 cm. in length show up occasionally; the rock is deeply weathered, a feature which applies in other cases. Under the microscope the phenocrysts prove to be labradorite with very occasional augite; the groundmass is composed of feldspar laths the great majority of which have the low extinction angle corresponding with oligoclase and often have denticulate margins, but there are microlites as well without such borders and with the extinction-angle of andesine. If only one plagioclase can be present in such a groundmass they must all be andesine. There are as well grains of magnetite and patches of material with no effect on polarised light, in all probability a glass. Numerous cavities also occur lined with fibrous chalcedony marginal to a brownish aggregate occupying the interior with an index lower than that of balsam, and having some effect on polarised light at minute points, which are occasionally arranged in definite parallel lines in one section of the amygdule and with a similar arrangement but oriented differently in another section. Analysis No. 4 is of this rock. It shows a high percentage of water, and when this is allowed for the composition becomes that of a basic andesite; the norm, too, shows a very high percentage of quartz, no doubt accounted for by the presence of chalcedony.

Similar rocks marginal to trachytes occur on the crest of the ridge at the head of Heathcote Valley, and with some glass in the base as well as numerous vesicles full of yellowish green alteration products. The microlites of the base appear in two forms, one lathshaped and the other more quadrate, both with index higher than that of balsam. These occurrences appear to be restricted to only short lengths of trachyte dykes, and they appear to have no close relations in composition to them, that is, they are in no sense composite dykes arising from differentiation of a common magma.

Andesite forms the complex dyke east of the Heathcote Valley saddle and occurs much decomposed on the side of the road crossing the ridge leading up to Mt. Ada and near Rhodes's Bush on the south side of Cass Peak. An undecomposed rock also occurs on the southern side of the Sugarloaf as a moderate-sized dyke (6 ft.) cutting the road twice. This contains a few phenocrysts of plagioclase

In a hyalopilitic base with lath-shaped microlites of andesine and grains of augite; and andesites of this grade occur on the cutting leading down from Coopers Knobs to Gebbies Pass. One of these is composed of phenocrysts of labradorite in a base of oligoclase microlites, augite grains and broken-comb forms of ilmenite and patches of brownish glass as well as amygdules of fibrous chalcedony. Other cases of this type show oligoclase microlites with denticulate margin in the base, and in one case the phenocrysts are of labradorite and augite in a base of andesine microlites often arranged with radial grouping and at times showing flow structure, and in between the laths are grains of augite and patches of alteration products green and brown in colour, the amount of F.M. mineral present being unimportant.

A special case of basic andesite occurs as a dyke on the north-east face of Witch Hill just east of the Giants Causeway. The rock is green-grey in colour and under the microscope shows phenocrysts of feldspar up to 1 cm. in diameter, slightly purplish augite, occasional olivine in grains up to 0.1 cm. in diameter, and grains of brownish-black magnetite. The base consists of laths of oligoclaseandesine usually showing flow arrangement, grains and short crystals of greenish to purplish augite, and grains of magnetite. There appears to be a very occasional feldspar which has an index of refraction slightly lower than that of balsam, but by far the great majority of the microlites have a higher index. Amygdules of chalcedony and of tridymite also occur as well as small patches of a greenish anisotropic mineral.

These rocks grade into undoubted basalts, and probably some of them might be classed as such, but it is difficult without analyses to say definitely at times what group they should be assigned to. One of these border-line cases is the basalt with groovings referred to elsewhere in this paper. This is an even-grained rock, without phenocrysts, and composed of laths of oligoclase, grains and short crystals of augite, and grains of magnetite. Sections taken with orientation parallel to the walls of the dyke show no difference in composition or texture from sections taken at right angles. Segregations of coarser texture, most composed of feldspar, also occur.

Undoubted basalts form the 12 ft. dyke on the east side of Heathcote Valley saddle, the core of the multiple dyke near mile post 4, and a dyke cutting the road on the south side of Mt. Ada. The last of these is dark in colour with feldspar and olivine showing occasionally in the hand-specimen. The feldspar proves to be labradorite, but it often has a more acid border. The F.M. minerals are faint violet augite and olivine in grains and aggregations up to 2 mm. in diameter, the latter sometimes fresh and again partly replaced by iron oxides. The holocrystalline base is composed of feldspar laths, augite grains, small olivines and grains of magnetite. The feldspar which makes up the bulk of this rock is of two types, the first, and more common, has approximately straight extinction and has denticulate border, while the other has an extinction angle corresponding to basic andesine, both with index of refraction higher

than that of balsam. If only one feldspar can be present it must be andesine, but if two can co-exist, then there appears to be oligoclase in addition, the extremely large proportion which shows the extinction angle of oligoclase being in support of the contention that both are present. No appreciable difference in texture or composition of the inside from the outside of the dyke was disclosed in slides selected from various parts, although the hand-specimens did suggest a difference.

Analysis No. 4 is of this rock. The most striking feature is the amount of normative albite, which seems to bear out the hypothesis that oligoclase is present in the base.

Some of the rocks which look like andesites in the hand-specimen prove to be undoubted basalts when examined microscopically. A good example of this is the dyke (6 ft.) which crosses the road a few yards south of the Hoonhay dyke and which presents a facies strongly reminiscent of the olivine andesites or feldspar basalts so characteristic of the effusive rocks of the area. In this there are numerous phenocrysts of labradorite 1 cm. in length; olivines up to 2 mm. in diameter with the usual alteration products; somewhat occasional augite; in a base composed of microlites of andesinelabradorite frequently with denticulate margin, grains of augite, small brown olivines, and grains of magnetite. Its composition is given in analysis No. 5, which shows that it is definitely a basalt.

The most basic dyke judged microscopically is the fourth down the cutting from Coopers Knobs and 12 ft. in thickness. The phenocrysts take up fully 50 per cent. of the slide and consist of labradorite, angite, and olivine in sub-equal proportions. The augite is light brown in colour, the olivine sometimes fresh and sometimes seamed and bordered with iron oxide. The base is holocrystalline and composed of short laths of labradorite, grains of augite, olivine, and magnetite; needles of apatite and alteration products of iron-oxide occur as well. Analysis No. 7 gives the composition of this rock.

I have left to the last all reference to the hornblende basalts which present certain interesting features. This type has been referred to by Bartrum (1917, p. 416) and by the present author (1923, p. 144), the rock referred to being a dyke in Sumner Valley which does not appear to cut the Summit Road. An excellent example can be seen in a dyke (6 ft. 6 in.) on the rock cutting east of Heathcote Valley (fig. 4). In the hand-specimen this rock is grey with a slight tint of green; it is very vesicular and shows crystals of hornblende up to 5 mm. in length, also very occasional feldspars in a base which shows well-marked flow structure. Under the microscope the phenocrysts appear as brown hornblende, sometimes resorbed; light purplish augite in sporadic crystals and in concentrations; olivine, fresh and stained with iron-oxide; very occasional microphenocrysts of plagioclase with higher index of refraction than balsam; grains of faintly brownish magnetite (?titaniferous); all in a base of lath-shaped microlites and more quadratic forms of andesine, twinned and untwinned, with definite flow arrangement, forms of greenish grey augite and grains of magnetite. This basalt is entirely different in facies from the other basalts cut by the road.