The Geology of the Mokohinau Islands, North Auckland

Transactions and Proceedings of the Royal Society of New Zealand, Volume 78, 1950, Page 255

The Geology of the Mokohinau Islands, North Auckland By C. A. Fleming, N.Z. Geological Survey [Read before Wellington Branch, June 12, 1948; received by the Editor, June 28, 1948.] Part 1—General Geology* Part II, “The Petrology of the Igneous Rocks,” by Dr J. W. Harris, is in preparation; preliminary petrological notes have been supplied by J. J. Reed. Contents Abstract. Introduction. Acknowledgments. General Description and Physiography. Geology: Sequence of Rock Formations. (1) Mokohinau Formation (extruded banded rhyolites). (2) Burgess Formation (pyroclastic rocks). (3) Lighthouse Formation (intrusive andesite). (4) Later acid intrusives. (5) Fanal Formation (banded flow rhyolites) (6) Recent and Pleistocene Deposits. Correlation and Age of Formations at Mokohinau. Abstract The Mokohinau Islands, east of North Auckland Peninsula, consist of an extrusive cumulo-dome of glassy rhyolite at Burgess Island, surmounted by pyroclastic rocks and intruded by a plug of andesite and by rhyolite dykes. Fanal Island is formed of banded flow rhyolites and obsidian. The acid rocks are believed to be late Pliocene in age and the later intrusives Pleistocene. Introduction During the war a Royal New Zealand Air Force Unit was stationed at Mokohinau. In October, 1944, shortly before the station was abandoned, arrangements were made by the Secretary, Department of Scientific and Industrial Research through the courtesy of the Commanding Officer, 60 Squadron, R.N.Z.A.F., for me to visit the islands for a week to investigate the geology. I landed on Burgess Island on October 10th, 1944, and spent the next six days at the group. Little information has previously been available on the geology of Mokohinau, which is, of all the island groups of the outer Hauraki Gulf, the least accessible. F. Sandager, one of the early keepers, wrote (1889) that the Mokohinau Islands were “all of volcanic formation, more or less precipitous, and divided from each other by narrow, rarely deep channels.” Acknowledgments Acknowledgment is made to Dr E. Marsden, Secretary, Department of Scientific and Industrial Research, Dr J. Henderson, then Director, Geological Survey, and to the Royal New Zealand Air Force authorities, who arranged my visit to Mokohinau. I am grateful to

Pilot Officer T. Marsden, O.C. Unit I, 60 Squadron, and to the men of Unit I, whose co-operation and assistance to me made my stay at the Unit both enjoyable and profitable, and to the Surveyor-General for permission to publish the accompanying aerial photographs. General Description and Physiography. The Mokohinau Islands lie off the east coast of the North Island of New Zealand some sixty-three miles north of the city of Auckland and thirty-two miles east of Whangarei Heads. (Fig. 1a). The nearest point on the mainland is Bream Head, 27 miles west-north-west of Burgess Island; the nearest islands are all about 15 or 16 miles away, the Chickens, to the west, Little Barrier, to the south, and Great Barrier, to the south-east. (Plate 26.) Text-fig. 1—(A) Map of part of the North Island to show the position of the Mokohinau Islands. (B) Map of Mokohinau Islands. The Mokohinau Islands are the most distant of the coastal islands at the entrance to the Hauraki Gulf. The continental shelf of the east coast of North Auckland is broad and gently sloping, and the Mokohinau Islands rise steeply from a sea bottom some 60 to 70 fathoms deep. (Fig. 2a). The 100-fathom isobath lies about 12 miles further to the north-east. The Mokohinau group (Fig. 1b) embraces about a dozen small islets and many rocks and reefs of negligible area; they extend from Groper Rock* Several of the place names used do not appear on any published maps, but have currency among yachtsmen and servicemen at Mokohinau. to the north-west, to Fanal Island, in the south-east, a distance of about 6 ½ miles, as an irregular belt of islets and rocks about 2 ½ miles wide from Cathedral Rocks* to Navire Rock. About three miles south of Navire Rock is the isolated low projecting reef of the Simpson Rock. Burgess Island, the site of the Mokohinau lighthouse, has a covering of rough pasture and is occupied in normal times by lighthouse keepers of the Marine Department. The other islands are uninhabited and support coastal scrub. Groper Rock, the north-eastern member of the archipelago, is a low dome-shaped islet

less than 20 chains in diameter, with some vegetation. It was not visited. Burgess Island, some 50 chains long and 20 chains in width, is dominated by the lofty plug of Lighthouse Hill (366ft.); elsewhere the island does not greatly exceed 250 ft. in height. The coastline is deeply indented by bays and narrow, steep-walled gulches, and there are numerous outlying reefs and stacks. (Plate 27.) Text-fig. 2—(A) Profile of continental shelf from Bream Tail to the Mokohinau Islands. Vertical scale greatly exaggerated. (B) Mokohinau Islands from the sea, view looking N.E. from a field sketch. The irregular outline of Burgess Island is due to marine erosion of what must have been at one time a much larger island. The rocks forming the island are not homogeneous: resistant intrusives penetrate weaker altered igneous rock intersected by joints of somewhat irregular pattern. Wave-attack has been selective and has emphasized the structural features by eroding narrow gulches along joints and dyke contacts. Thus a gulch twelve chains long, a chain wide, with vertical walls over 150ft. high extends eastwards across the north end of the island and almost meets a roughly rectangular gulch-system on the east coast. (Plate 28.) The amphitheatre-like embayment on the west coast is bounded by straight lines, possibly controlled by joints, and north-west-striking vertical joints determine the line of skerries which include Elephant Rock: parallel joints further south-west determine a deep gulch penetrating the island for at least five chains beyond the coast as a sea-filled cavern. There are similar deep indentations in the south-west, and the bay at the landing jetty seems also to be structurally controlled. The andesite plug of Lighthouse Hill stands high above the less resistant rhyolitic and fragmental rocks of the rest of the island, where the surface is, in general, undulating and smooth, at heights of from 100–250 ft. above sea level. Vertical structure in the igneous rocks of the island has favoured the development of sheer cliffs, locally extending down into deep water: only at the bay near the landing is there a limited area of boulder beach, backed by storm beach and raised beach material filling an originally deeper embayment.

South-west of Burgess Island, and separated by a few chains of strait, is a group of two smaller unnamed islands and some dozen islets and stacks. Like Burgess Island, these are cliff bound, irregular in outline, and deeply penetrated by caverns and embayments; they reach a height of about 250ft. (Plates 26, 27.) Fanal Island, 2 ½ miles south-east of Burgess Island, is the largest of the Mokohinau group and appears from the sea as a long, low, flat-topped mass, bounded by steep cliffs (Fig. 2b). No detailed map is available, but on the Admiralty Chart, Fanal is shown as over half a mile in diameter, with an irregular outline. The flat-topped appearance is deceptive, for there is considerable relief on the bush-covered summit. “Cathedral Rocks,” a reef a mile north of Fanal Island, comprise two low tabular rocks, the larger supporting a little vegetation; Navire Rock, the same distance to the south, is an isolated cone. Simpson Rock is a double-peaked reef, dark above storm-wave level and capped with a veneer of white guano. None of these rocks exceeds a few square chains in area, nor thirty feet in height. Geology Sequence of Rock Formations. Except for a small area of raised beach conglomerate, all the rocks so far recorded at Mokohinau are igneous. The oldest rocks (Mokohinau Formation)* Capitalisation of the initial of the word “formation,” when used in a formal sense, is “in harmony with prevailing practice in British countries,” and follows a recommendation of the Australian Code of Stratigraphic Nomenclature (Aust. Journ. Sci., 11: 7–9, 1948). are rhyolites of varied facies which appear to have been extruded from a more or less linear north-east striking vent at Burgess Island as an obsidian-capped cumulo-dome. Unconformably overlying the rhyolites are subaerial tuffs and agglomerates of the Burgess Formation, products of an explosive phase of activity, and including what may be flow rocks of more basic composition. Intrusive into the Mokohinau Formation, and possibly in part contemporaneous with the Burgess Formation, are the dark glassy andesites of the Lighthouse Formation, including the large plug of Lighthouse Hill a sill (?) penetrating the Burgess pyroclastics, and a further intrusive in the islands to the south. Dykes of grey glassy rhyolite penetrate the Mokohinau rhyolites, but are possibly truncated by some of the Burgess pyroclastics; later leucocratic rhyolitic dykes penetrate all formations. (1.) Mokohinau Formation. The name Mokohinau Formation is given to the rhyolites which are the oldest and most widespread rocks of the archipelago. Since future work may necessitate subdivision, the strongly banded almost vertical glassy lavas, striking about 35°E. at sea level on the north coast of Burgess Island are nominated the type section of the formation (P. 10247–10250). They include perlitic glasses with scattered phenocrysts. The rocks included in the Mokohinau Formation are vitreous acid extrusives, usually with well-developed flow banding. Typically, on

the north coast of Burgess Island, vertical bands, a few inches to several feet wide, of black glass with light amygdales, alternate with lighter, more weathered rusty orange-brown, grey, and whitish banded rhyolites in which the flow banding is parallel to the contacts with adjacent bands. The rocks strike roughly from 20°W. to 35°E., and are practically vertical. Similar almost vertical banded amygdaloidal glassy lavas from sea level to varying heights up the cliffs form the coasts of Burgess Island except below Lighthouse Hill, and, with some variation to be noted, form most of the islets and stacks to the south. (Plate 28, Plate 30, fig. 2.) Text-fig. 3—Burgess Island, Mokohinau. View looking south towards Fanal Island across deep gulch eroded in vertical rhyolites, spines of which (a, a) penetrate into a carapace of fragmental rocks (foreground and on face opposite). Drawn from a photograph. In the deep gulch penetrating the northern part of Burgess Island from the west, and on the steep cliff to the north, the relations of the vertical north-striking glassy perlitic lavas (P. 10231, 10251–3) to overlying rocks are clearly seen. (Plate 30, fig. 2.) The bands rise from sea-level, varying in width, locally contorted, weathered on exposed faces into ropey and twisted forms, to heights of about 150 ft. (at the gulch), where irregular vertically-elongated pockets of coarse agglomerate are in places included between adjacent bands. Higher up the cliff, coarse angular agglomerate is penetrated by tongues of

ropey contorted lava from below (Text-fig. 3). The agglomerate carapace is about 50 ft. thick and the degree of penetration by lava-tongues from below varies. In places bending and fragmentation of the tips of such tongues, as if by movement of the agglomerate, can ben clearly seen. The carapace is formed of angular fragments of the same glassy lavas as the underlying flow rocks, some exceeding 3 ft. in diameter, packed in a coarse, light matrix of finer material. Some of the boulders in the carapace are themselves composed of agglomerates, and others of glassy banded flow rocks enclosing xenoliths of similar material. Finer, well-bedded, gently-dipping tuff and agglomerates of the Burgess Formation overlie the agglomerate carapace with apparent unconformity. (Plate 30, figs. 1, 2.) Text-fig. 4—Burgess Island, Mokohinau. Coast of embayment in north-east coast looking south to the lighthouse. The lower part of the section is composed of vertical banded glassy rhyolites visible in foreground. Above (a, a), bedded pyroclastic rooks penetrated by an andesite sill (b) dip to the south-west, where they are cut off by a fault occupied by a rhyolite dyke (c). Drawn from a photograph. The fragmental carapace of the Mokohinan rhyolites can be recognized all round the north coasts of Burgess Island. The carapace is highest to the north-west, and falls, with the overlying pyroclastics, to the south and east. On the west coast it falls almost to sea level and is cut off by a later dyke occupying a fault with downthrow to the north-east (Text-fig. 4). South and west of the line of this fault,

which strikes south-east across the narrow waist of Burgess Island, the Mokohinau Formation rhyolites differ in several ways from those to the north-east. They lack darker fresh obsidian-like phases, have less well-defined flow-banding, and tend to be uniformly pale rocks with coarse spherulitic structure, but they have the same generally vertical or steeply dipping attitude on a strike ranging from north to north-east (P. 10240, 10255–6, 10260–1). Some phases are massive and microcrystalline, others coarsely spherulitic, coarsely amygdaloidal, or vesicular, the vesicles filled with secondary crystals. Locally, in the north-west bay, gently dipping joints or discontinuities, emphasized by differential weathering, give an appearance of more horizontal sheet-like structure, suggesting normal flows, but no inter-flow fragmentals nor clear contacts could be found, and closer investigation showed the persistence in such apparent sheets of steep flow-structure on a consistent strike, with some well-defined bands extending uninterrupted, across the horizontal pseudo-stratification, from top to bottom of the cliff. Such horizontal layers as are not gently dipping joints within the steep dipping lavas, are probably zones of alteration related to past or present water-table levels and to differential crosion at the level of maximum wetting and drying near sea level. The southern peninsula of Burgess Island, south-west of a line from the jetty to Elephant Rock, consists chiefly of vesicular, locally contorted, banded, pale rhyolites and variable but chiefly massive weathered rhyolite (P. 10261) dipping steeply or vertically on a north-easterly strike. North-east of the intrusive plug of Lighthouse Hill, massive white igneous rock, vertical and striking about 35°E., is considered mineralized Mokohinan Formation rhyolite, but may be a dyke (P. 10255). The reef east of the Lighthouse, at the foot of a 300 ft. cliff of intrusive andesite (Plate 29, fig. 1) is composed of vertical north-east striking rhyolite (P. 10256) separated from the andesites by a few yards of water. South of Lighthouse Hill, Lizard Island is composed of similar vertical rhyolites, and so also, judging by their appearance, are the Sphinx Rocks still further south. The group of stacks and islands west-south-west of Burgess Island expose in their coastal cliffs pale, steeply-dipping, poorly or moderately banded lavas, similar to those of the southern portion of Burgess Island (Plate 27). At three points closer inspection showed that the lavas included massive white types (P. 10260) with twisted flow structure and coarsely amygdaloidal phases with idiomorphic crystals of quartz filling spherical cavities up to 5 cm. in diameter, apparently once occupied by spherulitic material (P. 10277). Measured strikes on sub-vertical flow-banding ranged from 4° to 45°E. Examination of these outliers was limited to a circumnavigation by dinghy and landings at a few points, but the clear coastal sections gave little suggestion of much variation in structure. Thirty-eight measured strikes in the rhyolites of Mokohinau Formation range from 30°W. to 75°E., but 75 per cent. of these fall between 5° and 40°E., the mean direction of strike being about 20°E. The dip is in few places less than 70°, but dips as low as 45° to the

Text-fig. 5

east, were measured in the south-west portions of the island. There seemed to be no simple pattern in the direction or amount of dip. The rhyolitic rocks of the Mokohinau Formation are interpreted as remnants of a complex cumulo-dome, exuded from a linear vent striking roughly 20° east of north. The consistently steep dips and the presence, locally, of a carapace of fragmental rocks penetrated by tongues of glassy lava from below, can hardly be otherwise interpreted. Possibly “coulées” flowed out from the cumulo-dome, and may be preserved somewhere in the archipelago, but it seems unlikely that any appreciable amount of lateral flow could occur without producing horizontal structure in the viscid lavas. Of such there is little evidence at Burgess Island, where excellent coastal sections amply demonstrate that movement of successive tongues and spines of stiff glassy lava was chiefly upwards. In the north of the island successive injections of lava into the growing dome have differed in character, and, in a limited area, three or more generations of lava can be recognized, the younger usually penetrating planes of weakness which are parallel to planes of flow-banding in the older, but, more rarely, transgressing such planes. It is not suggested that a great time elapsed between successive injections. On the contrary, contortion of banded rocks of an older generation near the contact with a later tongue suggests that the earlier rocks may have still been viscous when the succeeding ones were intruded into them. No lava completely penetrating the fragmental carapace has been observed, so that a cumulo-dome “chiefly of endogenous growth” (Cotton, 1944, p. 163) is represented. Theoretically, the flow or displacement surfaces within such a structure should be arranged in an upwardly fanning pattern (Cotton, p. 170, 176): differences in the inclination of the Mokohinau rhyolites could be thus interpreted, but it did not seem, in the field. that there was any simple pattern. Only in the north-east of Burgess Island has a fragmental carapace been recognized. Elsewhere, apparently, it has been up-faulted to an unknown extent and eroded; but the general continuity of strike suggests that one eumulo-dome extended over all the group of islets southwest from Burgess Island. This archipelago is elongated in the same direction as the strike of the rhyolites within it, and the inference drawn is that the cumulo-dome originally extended over them all, and thus had a length of at least 2 ½ miles: probably the volcanic field extended further to embrace Groper Rock (which photographs show to consist of contorted steeply dipping pale igneous rock) to the west and Fanal Island to the east. The rocks of Fanal Island (see below) are flow rhyolites differing in habit from the Burgess Island rocks, but quite possibly coeval. (2.) Burgess Formation: Pyroclastics. Overlying the fragmental carapace of the Mokohinau Formation in the northern peninsula of Burgess Island are well-bedded pyroclastic rocks, dipping southeastward from the high north-west cliff. There, the unsorted carapace agglomerate is overlain by four feet of fine sandy tuffs and then by a further four feet of coarser breccia with angular ejectamenta of volcanic rocks up to 3 or 4 inches in diameter. The layers are separated by a strong iron pan. The bedded tuffs north of the barracks (Plate

30, fig. 2) are named as type locality of the Burgess Formation. In the “waist” of Burgess Island similar tuffs and agglomerates up to 30ft. thick, dipping south-east, are grouped as the same formation. Dykes and sills penetrate the tuffs, which are indurated at the contacts. On the south-west peninsula of Burgess Island bedded fragmentals, dipping gently eastward, in the upper parts of the coastal cliffs, are conspicuously unconformable on the lower Mokohinau rhyolites, which locally dip steeply westward. Some of the fine tuffs and weathered agglomerate near the surface in the south have locally broken down to form bluish clay-like masses. In the field, water-worn pebbles in the soil overlying tuffs were at first taken to indicate interbedded conglomerate laid down in water, but they most probably represent Maori oven-stones, and no other sign of subaqueous beds was seen. On the nearest of the larger islands south-west of Burgess Island observed fragmental beds appeared to be merely talus accumulations, and no definite bedded tuffs were seen in other islets of the group. The tuffs mapped as Burgess Formation include almost white beds of sandy texture and purple-grey and brown ash (P. 10245). Fragments in the coarser pyroclastics are chiefly of glassy banded rhyolites of Mokohinau Formation, but there are, in addition, obsidians not matched in outcropping rocks, and dark fine-grained igneous rocks comparable with those of the Lighthouse Formation. Some tuff beds appear to transgress unconformably over platy grey rhyolitic dykes; other tuffs are certainly intruded by dykes of similar lithology, and peripheral sills of the Lighthouse plug appear to penetrate the tuffs. For these reasons it seems possible that more than one group of pyroclastic rocks has been confused. At present it is certain that the rocks grouped as Burgess Formation are younger than the rhyolites of the Mokohinau Formation, which they overlie unconformably, and older than the andesite of Lighthouse Hill and than most of the dykes. (3.) Lighthouse Formation. The name Lighthouse Formation is here given to the dark glassy andesites outcropping on the vertical sea cliffs east and south of the Mokohinau Lighthouse (Plate 29, fig. 1), and is also applied to rocks of similar lithology elsewhere in the group. The contrast between the high steep-sided mass of Lighthouse Hill (366 ft.) formed of coarse columns of blackish igneous rock exposed in the seaward faces, and the lower gently sloping terrain developed in pale-weathering rhyolites, is the most striking topographic characteristic of Burgess Island. The rocks of Lighthouse Hill are andesites, disposed in rude columns inclined steeply to the north. Lighthouse Hill appears to be a plug more or less circular in plan, penetrating the rhyolites of Mokohinau Formation which almost competely surround it. Nowhere was the actual contact between the darker intrusives and the pale rhyolitic country rock seen. To the north-east massive greyish-white rhyolite (P. 10255) is separated from the andesite by several yards of talus slope. On the east, the present cliff (Plate 29, fig. 1) is developed very near the contact, for off-lying reefs near sea-level, composed of altered grey and white perlitic glass (P. 10256) are

separated by only a few yards of water from the foot of the andesite cliffs. In the south the plug has been reduced by marine erosion, and the contact lies between Lizard Island (rhyolite) and the stack to the north (andesite). On the west the boulder-filled bay at the landing has apparently been eroded to the contact, and no rhyolites in situ were found at the east end of the beach, but, some 8 chains to the north-east, leucocratic lavas near the keepers' houses wrap around the plug. North of the hill what appear to be sills arising from the boss penetrate agglomerate at the knife-edge ridge followed by the road to the barracks. Near the inferred contact east of the jetty the andesites are dark glassy rocks (P. 10262–3) locally scoriaceous (P. 10264) with scattered phenocrysts (but none of recognizable ferro-magnesian minerals) and a great many included xenoliths of paler rhyolitic lava. Up the western slopes, some dark glassy phases, apparently representing the selvedge of the boss (P. 10265–6) are with difficulty distinguished from the darker rhyolites, which, however, do not outcrop nearby. On the summit of the hill, above the eastern cliff, halfway between the Lighthouse and the barracks road, the rocks are glassy and purplish-brown with scattered visible phenocrysts and locally abundant xenoliths of country rock (P. 10267–8). At sea-level below the eastern cliffs the andesites (P. 10257–9) are dark purplish-brown, locally limonitic, with patches of irridescent pyrite. Seaming the vertical cliff face are bands of clastic material: friction breccias developed between adjacent rock columns or perhaps agglomerate-filled vents. The conspicuous later acid dyke intruding the plug has followed one such clastic layer in the andesites. North of Lighthouse Hill, thick bands of glassy, igneous rock interbedded in the fragmentals of the Burgess formation and dipping south-east with them at a moderate angle (15°), are exposed on both sides of the narrow “waist” of Burgess Island. The rocks (P. 10273–4) bear a close resemblance, in hand specimen, to the more glassy phases of the Lighthouse andesites (P. 10266–8) and, further, seem to be in continuity with the main plug high in the eastern cliff due north of the Lighthouse. The impossibility of reaching the relevant part of the cliff face prevented adequate study of these relations, but from a distance, and from the sea, it appeared that a peripheral sill from the boss penetrated the fragmental rocks to the north and rose within them for some distance. The alternative is to regard such gently-dipping bands as flows from a source north-west of Burgess Island (Text-fig. 4). A further intrusive body of more basic nature than the surrounding rhyolites was located in an arcuate reef south-west of the southern islands of the Burgess group. The rock (P. 10278) is a purplish-black, glassy andesite with scattered visible phenocrysts: it is tentatively correlated with the Lighthouse Formation (Plate 26, fig. 3). In this photograph the larger stacks in the foreground and on the extreme right are, however, pale rhyolites of the Mokohinau Formation. (4.) Later Acid Intrusives. Steeply-dipping dykes up to 50ft. thick outcrop freely in the coastal cliffs and on the surface of Burgess

Island. Some clearly penetrate the rhyolites of Mokohinau Formation, but are transgressed above by tuffs of the Burgess Formation; others have, intruded and baked tuffs and agglomerate (Plate 30, fig. 1); others intrude the Lighthouse neck (Plate 29, fig. 1). Most of the dykes, as can be seen from the map (Fig. 5) strike north-eastward and dip either east or west at high angles; fewer dykes with a west or north-west strike were mapped. Lithologically, the dyke rocks are of two types, a more leucocratie type in which flow structure is poorly developed, and a darker more glassy type with flow structure imparting a rough parallel fracture. In the first type falls a wide dyke trending south-west along the eastern side of Burgess Island for some 20 chains, forming conspicuous white surface outerops (Plate 29, fig. 2). Similar in lithology are the dykes intruding Lighthouse Hill, which have the same strike and steep north-westerly dip. The dyke occupying the fault crossing the middle of Burgess Island may also be classed here (Text-fig. 4). Petrological examination of specimens of the above dykes (P. 10231, 10232, 10254, 10275) shows them to be partly-devitrified colourless glassy rhyolite and glassy cryptocrystalline rhyolite (J. J. R.) The remaining dyke rocks of Burgess Island are of more platy aspect. The most conspicuous of such dykes trends from the barracks south-westward, parallel to the more leucocratic dyke noted above. Near the barracks several separated but apparently related intrusions are mapped, and their presence seems to have at least partly determined the erosion of the steep gulches of the north-east part of the island. Similar but smaller dykes have been mapped in the northern peninsula of Burgess Island and another in the west of the island may have controlled the erosion of a deep straight-sided west-striking bay. Of the specimens collected from these dykes (P. 10230, 10237–8, 10239, 10246) some show prominent trachytic structure, others are perhaps more basic, with rhyolitic inclusions in a felted ground mass (J. J. R.). No rhyolite dykes were recognized away from Burgess Island. (5.) Fanal Formation. The west coast of Fanal Island was examined from the boat, and a landing made in the south. Seen from the sea, the rocks exposed in the western cliffs are pale or brownish, and show well-defined stratification. The banded rocks dip irregularly, but not steeply, are locally contorted, and are intersected by closely-spaced joints striking east-west. No vertical flow-structure comparable with that described at Burgess Island was seen. At the south-west, point the rocks were fine-grained, well-banded, flow rhyolites, light chocolate and fawn in colour, the alternating laminae up to several millimeters in width (P. 10280–10282). Apart from minor contortions, the flow-banding is flattish or gently dipping. Similar banded rhyolites were encountered all the way up the ridge from sea-level to a height of about 200 ft. (P. 10283) and no inter-flow fragmentals were recognized either on the outcrops or from the boat. In the gully draining the summit plateau of Fanal Island two rock-types were recognizable among the angular boulders of the dry creek bed. Finely-banded rhyolites like those described dominated,

Fig. 1—Looking W.S.W. over Burgess Island (foreground) and Groper Rock (a) to the mainland of North Auckland, showing Hen Island (b), the Chickens (c), and Bream Head (d). Fig. 2—Looking south over Burgess Island, showing Fanal Island (e) and Little Barrier Island (f). Fig. 3—Looking E.N.E. towards Burgess Island (with lighthouse), showing Cathedral Rocks, north of Fanal Island (g). The islands are composed of vertical pale glassy rhyolites penetrated by more basic intrusions which form Lighthouse Hill and the festoon of rocks in the south-west (h). R.N.Z.A.F. Official Photographs—Crown copyright reserved. Aerial photographs of Mokohinau

Mokohinau Islands View from an unnamed island of the group, looking E.N.E. to Burgess Island (with lighthouse) and its outliers, Elephant Rock (left) and Lizard Island (right of centre). The islands in the foreground are composed of light-coloured perlitic rhyolite with vertical flow structure on a north-easterly strike, away from the camera. Most of Burgess Island similary constituted, but an intrusive boss of resistant andesite forms Lighthouse Hill, 366 ft. (centre middle distance) and the conical stack immediately to the south; bedded pyroclastic rocks cap parts of the ridges visible on the left of the lighthouse. Marine erosion has sculptured gulches, skerries and stacks from the vartical rhyolite columns, and the low islets and western promontory on the left of Burgess Island may be part of a Pleistocene wave-eut bench 100–150 ft. above sea-level. Photo.: G. A. Buddle, Auckland.

Fig. 1—Looking W.S.W. to show the steep cliffs and deep gulches of the north-east coast. The changes of slope above the cliffs (left centre) occurs at the passage from vertical rhyolite to a carapace of fragmental rocks. Fig. 2—Looking N.W. over the Lighthouse Hill andesite plug (left centre), which is partly surrounded by vertical glassy rhyolites forming the reef below the lighthouse, and the peninsulas beyond and to the right of the central isthmus. R.N.Z.A.F. Official Photographs—Crown Copyright reserved. Aerial photographs of Burgess Island.

Fig. 1—Intrusive neck of andesite forming Lighthouse Hill, Burgess Island (366 ft.), viewed from the north-east. Vertical rhyolite, part of the cumulo dome through which the andesite was intruded, outcrops at Lizard Island (a) and at b, b. A peripheral sill of andesite extends outward and upward (c) over the underlying rhyolites. Later rhyolite dyke (d) penetrates the andesite. Photo.: C. A. Fleming. Fig. 2—Surface of Burgess Island, looking south from the north-eastern end across the joint-controlled gulch almost bisecting the island. Bedded tuffs in foreground; late dyke on left of buildings intruding rhyolite and tuff; light colored rocks outcropping beyond are another rhyolite dyke. Photo.: C. A. Fleming.

Fig. 1—View looking west along vertical-walled gulch developed on joints and dykes penetrating the Burgess Island cumulo-dome, the fragmental carapace of which is visible in the left foreground and at the break of slope on sky-line. Bedded tuffs outcrop on rounded hill right. Photo.: C. A. Fleming. Fig. 2—Steep northern coast of Burgess Island, showing vertical rhyolites of the cumulo-dome (a) with spines penetrating its fragmental carapace (b), overlain by bedded pyroclastic rocks (c). Relief about 250 ft. Photo.: C. A. Fleming.

but there were considerable numbers of large pieces of obsidian (P. 10284–5). Obsidian was not seen in situ, but must outcrop somewhere on the island judging by the number of pieces seen. The Fanal Island obsidian is greenish black, slightly amygdaloidal, more opaque and less homogeneous than Mayor Island obsidian, lacking perfect conchoidal fracture. The black amygdalar glassy phases of the rhyolites at Burgess Island are softer and more brittle, and could not be flaked as knives by the Maori: the Fanal Island obsidian, on the other hand, though inferior to Mayor Island material, was used by the Maori, for flakes and pieces of it were seen in middens in the surface soil. No dykes or other intrusives were noted at Fanal Island, and the relation of the banded flow rhyolites to the Burgess Island rocks is not known; they may be flow equivalents of the Mokohinau extrusives, but could also be flow equivalents of the later intrusive rhyolites of Burgess Island. (6.) Pleistocene and Recent Deposits. Behind the bay at the anchorage at Burgess Island, a tract of flat is formed by accumulations of beach boulders at no great height above sea-level, and probably within the range of storm-wave transport. Behind the bay gentle slopes lead up to a low saddle in the ridge forming the divide above the north-western coast. Here, at a height of about 50 ft. above sealevel a small outcrop of somewhat cemented pebbly conglomerate, apparently veneering the slope, is considered a Pleistocene raised beach deposit. Correlation and Age of Formations at Mokohinau. There is no evidence for the age of the Mokohinau igneous formations other than what may be deduced from the indirect correlation of the rocks and their sequence with similar rocks and sequences elsewhere. In North Auckland, dacites and rhyolites were extruded at two or more periods in the Tertiary (Bartrum, 1936. p. 527); the rhyolitic phases are of limited extent, and not directly comparable with Mokohinau rocks. From the Poor Knights Islands, Bartrum (loc. cit.) has described highly potassic rhyolites with spherulitic structure which may be related to the Mokohinau occurrences, and aerial photographs of the Sugar Loaf, south of the Poor Knights, show vertical structure similar to that Mokohinau. At Great Barrier Island, Bartrum (1921) has mapped and described finely-banded pinkish-grey rhyolites and what he considered was “probably a weathered coarsely perlitic obsidianitic flow.” Obsidian is known to occur on Arid Island (Hutton and Kirk, 1868) and coarsely spherulitic rhyolites occur freely as boulders in stream beds between Wangaparapara and Fort Fitzroy (C. A. F.). At Great Barrier, the rhyolitic rocks overlie an eroded surface of andesitic volcanics (correlated with the “second period” or “Beeson's Island” andesites of Coromandel Peninsula, probably of Miocene age), and were, therefore, correlated by Bartrum with the “Third Period” acidic volcanic rocks of Coromandel. The descriptions of the Great Barrier rhyolites and perlitic obsidianite are sufficiently like the same types of rock at Burgess and Fanal Islands to make correlation a reasonable inference. No later intrusive andesites have been recorded at Great Barrier.

“Third Period” acidic volcanic rocks are widespread on Coromandel Peninsula, and locally (Table Mountain. Thames Subdivision, Aroha Subdivision, Waihi Subdivision) more basic intrusives (hypersthene-andesite) have been described penetrating them. Some of the Mercury Islands, which have not so far been geologically examined in detail (but see Marshall, 1935, pp. 326 and 328) are largely composed of pale igneous rocks similar to those of the “Third Period” at Whitianga (classified as ignimbrite by Marshall), but superficial examination of Ohena and other islets to the north showed the presence of local dark lavas intruding the acid rocks (C. A. F., September, 1946). Marshall (loc. cit., p. 328) has identified hypersthene basalts of columnar habit at the north end of Great Mercury Island. In default of more precise data, therefore, the Mokohinau and Fanal Island rhyolites may be correlated with the “Third Period” rocks of Coromandel Peninsula, and thus, in a general way, with the ignimbrites and rhyolites of the central volcanic plateau of the North Island. A late Pliocene age for such rocks is generally accepted. The intrusive andesites of the Lighthouse Formation are correlated with the still later hypersthene andesite intrusives of Coromandel Peninsula, and they, with the still later rhyolite dykes at Mokohinau, may be Pleistocene in age. Macpherson (1947. p. 9) has grouped the Poor Knights. Mokohinau, Fanal, Great Barrier, Arid, Cuvier, Aldermen and Mayor Islands as an are of volcanic islands with linear distribution indicating a north-west trending fold axis along which vents have been active in Recent time. The north-easterly strike of the Mokohinau rhyolites, and the implied north-easterly trend of the linear vent from which they were erupted, is transverse to the main axis of this marginal fold of North Auckland, a feature in accord with Macpherson's conclusion that the recurved geanticline of New Zealand was fragmented by transverse faults in late Pliocene times. Literature Cited Bartrum, J. A., 1921. Notes on the Geology of Great Barrier Island. Trans. N.Z. Inst., vol. 53, pp. 115–127. — 1936. Notes on the Geology of Three Kings and Other Outlying Islands of Northern New Zealand. N.Z. Journ. Sci. Tech., vol. 18, no. 6, pp. 520–530. Cotton, C. A., 1944. Volcanoes as Landscape Forms. Whitcombe and Tombs. Hutton, F. W., and Kirk, T., 1868. Description of Arid Islands, Hauraki Gulf. Trans. N.Z. Inst., vol. 1, p. 163. Macpherson, E. O., 1946. An Outline of Late Cretaceous and Tertiary Diastrophism in New Zealand. D.S.I.R. Geological Memoir, no. 6. Marshall, P., 1935. Acid Rocks of the Taupo-Rotorua Volcanic District. Trans. Roy. Soc., vol. 64, pp. 1–44. Sandager, F. S., 1889. Mokohinau Islands and the Birds which Visit Them. Trans. N.Z. Inst., vol. 22, p. 286. Sollas, W. J., and McKay, A., 1905. The Rocks of Cape Colville Peninsula. Government Printer, Wellington.