Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 54, 1923
This text is also available in PDF
(953 KB) Opens in new window
– 106 –

Art. 8.— Fossiliferous Limestone at Dowling Bay.

[Read before the Otago Institute, 8th November, 1921; received by Editor, 5th December, 1921; issued separately, 8th February, 1923.]

Plate 11.

The occurrence of a limestone at Dowling Bay has long been known, but no fossils have yet been recorded from it. A few species were found on the occasion of a visit of the Otago University geology class, and, at the suggestion of Dr. Benson, we undertook to make as complete a collection, of fossils as possible, and to describe the locality.

The place was visited by Blair (1879), who found five seams of limestone which he stated might be of use for building purposes. Analysis, however, showed that varying amounts of sand (from 3 to 28 per cent.) were present, and the rock has been rejected as a source of lime, though it has been used for building-stone in Port Chalmers.

There is no record of any other examination of the locality till it was inspected by Morgan (1919). He found the strikes ranging from S. 30° W. to W. 10° N., with westerly and northerly dips of from 25° to 60°. He also noticed the extreme variation in the composition of the rock.

As may be seen from the map, the outcrops of sedimentary deposits are limited on one side by the coast-line, and on the other by an overlying series of basalt-flows. The marine beds have been exposed by the removal of the basalt by erosion, and except along the sea-border are almost completely covered with soil. It is thus' difficult to determine the line of separation between the basalt and the calcareous sandstone, but a series of springs was taken as giving evidence of the dividing-line.

Fossils were obtainable only from a road-cutting at the western end of Dowling Bay, the remainder of the fossiliferous strata being obscured by two masses of landslip rubble and soil. The Limopsis-bearing limestones, which are the most prominent parts of the outcrop, were found to dip about “W. 10° N. at angles varying from 36 to 45. The fossiliferous layers are scattered through a thickness of about 160 ft., with no well-defined order of occurrence. The limestone varies largely in appearance and character from a hard grey to a softer black rock, and in small laminae is even of a “shaly character. The Limopsis bands are up to 1 ft. in thickness, while those in which Flabellum and Turritella are abundant have a thickness in some cases of several feet.

In a cliff on the road between the two heaps of rubble, blue-grey calcareous sandstones are seen to rest on the fossiliferous beds, and these in turn. on yellow sandstone like that at Caversham. Here the dip of the grey limestone is 35°. From the second rubble-heap to the mass of phonolite at Otafelo Point the sandstone is yellow, unfossiliferous, and stands in high cliffs, though without much sign of stratification. The beds here appear to dip vertically, and it was noticed that this displaced stone is in line with the fault that cuts off the phonolite. The phonolite adjacent to the faults is fractured and broken, especially the northern mass at Otafelo Point.

– 107 –

The section along AB indicates the relative position of the beds. It can be seen that the fossiliferous limestone overlies the Caversham sandstone without unconformity, and therefore, if the age of these beds can be determined from palaeontological evidence, a limit is placed on the age of the Caversham sandstone. It may be noted also that somewhat similar fossiliferous limestones appear among the volcanic rocks of Sandymount, on the southern side of the harbour, and these also appear to be underlain by the Caversham sandstone.

The trachytoid phonolite, mapped as such by Dr. Marshall in 1906, has been described for us by Dr. Benson in the following terms: “ Microscopically this rock resembles the phonolite of Logan's Point, near Dunedin, but has a rather greater grain-size. The dominant mineral is a soda sanidine, which occurs in small phenocrysts up to a length of 0.8 mm., and much more abundantly in narrow laths in general fluidal arrangement, but with sheaf-like radiating groups. No trace could be observed of anortho-clase, or, indeed, of any polysynthetically twinned feldspar. Interstitially there is a small amount of zeolite material with bright polarization colours, probably replacing nepheline,’ which latter could not be distinguished in

– 108 –

the unstained section. The pyroxene is a pale weakly pleochroic aegyrine augite, occurring rarely in small idiomorphic phenocrysts, but generally in minute prisms without terminal faces, or in less regular wisps. Magnetite occurs rather more abundantly than is the rule in Dunedin phonolites, ranging from very small grains to those about 0–1 mm. in diameter. Possibly this mass of phonolite is the continuation of the much thinner flow of phonolite described by Dr. Marshall (1914) near the base of the great series of flows at North Otago Heads—i.e., about a mile farther to the north-east. Several intermediate outcrops of phonolite have been found suggesting this.”

The Dowling Bay limestone is rather rich in shells, but, unfortunately, most of them are either rolled or fractured. The fossils have a characteristic black or rusty crystalline appearance. In many of them the ornamentation is obscured, and frequently much excavation of the matrix is necessary before any identification can be made. The limestone is generally compact, the shells mostly occur in thin carbonaceous beds which do not split out easily, and the exposure of complete specimens is of rare occurrence.

It has been possible in almost every case to compare the species identified from Dowling Bay with actual specimens from Oamaru, especially Target Gully, in the author's cabinet, and so it is hoped that as few errors as possible have been incurred.

A definite idea of the horizon of this bed may be obtained most readily by tabulating records and estimating percentages after the methods of Marshall (1919, p. 247) and Thomson (1920, pp. 389 et seq.).

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

Dowling Bay Fossils. 1. 2. 3. 4. 5. 6. 7. 8. 9. Remarks.
a b c d
Calliostoma suteri n. sp.* N N N
Turritdla of. patagonica Sow. X X
Turritdla roaea(?) Q.&G. X X X X X X Fragment.
Calliostoma semiconcava Sut. X X X X X X
Sigapatella novaezdandiae (Less.) X X X X X X X X X Several.
Crepidula densistria (?) X X X
Sigapatella cosata (?) (Sow.) X X X X X
Sigapatella gregaria Sow. X X X X X
Sigapatella monoxyla (Less.) X X X X X X X X X
Sigapatella striata (Hutt.) X X X X X
Ampullina suturalis (Hutt.) X X X X X X X
Cymatium sp. Fragment, but not any N.Z. species so far described.
Epitonium zdebori (Dkr.) X X X X X X X X X
Fusinus macrotegens n. sp.
Aethocola spiniferan. sp. N N N Spire-fragment.
Dentalium solidum Hutt. X X X X X X X X X X Many.
Nucula sagittata Sut. X X X X X X Many.
Nuculana semiteres (Hutt.) X X X X X X
Malletia australis (Q. & G.) X X X X X X

[Footnote] * Described at p. 101 in this volume— “Some Remarks on New Zealand Callio-stomidae, with Descriptions of New Species.

[Footnote] † See description later.

– 109 –

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

Anomia huttoni (?) Sut. X X X X Much worn specimens. Radiate ribs cannot be seen, but may have been obliterated.
Anomia huttoni undata Hutt. N N X X
Anomia huttoni trigonopsis Hutt. X X X X N X X X
Anomia huttoni Placunanomia incisura (?) Hutt. X X X X X Fragment.
Limopsis producta n. sp.
Limopsis producta zealandica Hutt.(= L. aurita of previous lists) X X X X X N X X X Very plentiful.
Limopsis producta zitteli Iher. X X X X X X
Limopsis producta catenate, Sut. X X X X X X
Mytilus huttoni Cossm. X X X X X
Pecten polymorpkoides Zitt. X X
Pecten polymorpkoides hutchinsoni Hutt. X X X X X N
Pecten polymorpkoides cf. hilli Hutt. Fragment.
Crassatellites sp. X X X X Hinge-Fragment Either amplus (Zitt.) or attenuatus (Hutt.), not obesus (A.Ad.)
Venericardia ponderosa Sut.
Venericardia ponderosa subintermedia Sut. n. var. X X Same as th Target Gully variety.
Lucvnida laminata Hutt. X X X X
Divaricella cumingi (Ad., & Ang.) X X N X X X X Several.
Macoma edgari (?) Iredale X X X X X X Fragment.
Zenatia acinaces (Q. & G.) X X X X X X X
Chama huttoni (?) Hect. X X X
Ventricoloidea vellicata Hutt. X X X X
Chione Mesodesma (Q. & G.) X X X X X X X Hinge-Fragment.
  • Column 1.—Wangaloa and Hampden beds. Marshall (1917, p. 451; 1919, p. 235).

  • 2.—Waiarekan stage of the Oamaruian.*

  • 3.—Ototaran stage of the Oamaruian.

  • 4.—Hutchinsonian stage of the Oamaruian (Lower Hutchinsonian of Park).

  • 5a—Awamoan stage of the Oamaruian.

  • 5b—Basal stage of the Awamoan (Upper Hutchinsonian of Park).

  • 5c—Ardgowan “ shell-bed.

  • 5d.—Target Gully “ shell-bed.”

  • 6—Kawa Creek beds. Bartrum (1919, p. 104).

  • 7.—Waitotaran stage.

  • 8.—Castlecliffian stage. Marshall and Murdoch (1920, pp. 120 et seq.).

  • 9.—Recent.

[Footnote] * The records in columns 2 to 5d are from Park's lists in N.Z. Geol. Surv. Bull. No. 20 (n.s.), pp. 90, 94–96, 97–104. Uttley (Trans. N.Z. Inst., vol. 52, p. 176, 1920) has shown that Park's Upper Hutchinsonian cannot be separated from the Awamoan, and it is here so considered. The Hutchinsonian records in column 4 are taken only from Park's Lower Hutchinsonian lists, but the fauna of his upper beds is listed for comparison in column 5b. The letter N in any column denotes a new record made by the author and so far unpublished.

[Footnote] † Taken from various Waitotaran lists in N.Z. Geol. Surv. Pol. Bull. No. 8, 1921.

– 110 –

Besides these were several unidentifiable fragments, a few Balanus plates and echinoderm spines, very numerous specimens of Flabellum laticostatum T.-Woods, and a single Hemithyris species, too much worn to be specifically identified, but very similar to H. nigricans (Sow.). No traces were found of other brachiopods, though careful search was made.

Over 140 specimens were examined, yielding forty-one species. Twenty-five of these are lamellibranchs, a percentage of 61. This predominance of lamellibranchs shows that the conditions of deposition were somewhat akin to those of some of the Trelissick beds, which contain a very similar fauna (Speight, 1917, pp. 352–55). In view of the correlations made later in-this paper, it may be useful to quote Marshall's words, used in connection with the relative poverty in gasteropods often noticed in the New Zealand Tertiary beds: “ So far as observations have gone up to the present time, this striking difference does not appear to be due to the depth of the water or any other of the ordinary conditions that control the deposition of sediment” (Marshall and Murdoch, 1920, p. 128). The absence of small gasteropods may be due to bad preservation, or, as shown by the limestone, the prevailing muddiness of the sea—bottom.

As regards relative abundance of species, most of them are sporadic in occurrence, but Limopsis zealandica Hutt. and Flabellum laticostatum T.-Woods are extremely plentiful. This last point is interesting in view of the fact that the Dowling Bay limestone is the probable equivalent of the Sandymount limestone on the other side of the harbour, which was included by Marshall (1916, pp. 93, 96) in his “ younger limestones of New Zealand.” The non-occurrence of Flabellum, &c, in these limestones is drawn attention to, so it appears that in this respect the Dowling Bay beds are exceptional. A curious point is the complete absence so far of the Toxo-glossa-Turris. Surcula, Drillia, Terebra, &c, being generally plentiful in the Oamaruian. Here again the fauna agrees closely with that of the Trelissick beds, this absence being commented on by Speight (1917, p. 356). A consideration of the habitats of the Mollusca enumerated leads to the conclusion that the beds were deposited probably at a depth of about 50 fathoms, and this is in accordance with the impure sandy character of the limestone.

There are only three species in the above list peculiar to the locality.* Omitting forms doubtfully identified, also Pecten cf. hilli Hutt. and Veneri-cardia ponderosa Sut., the horizons of which are not definitely known, it can be shown, by using Thomson's method of implication, that the percentage of records in the various stages is as follows:—

Recorded from Column. Percentage.
Wangaloa and Hampden fauna 1 22
Waiarekan stage 2 41
Ototaran stage 3 53
Hutchinsonian stage 4 53
Awamoan stage 5a 88
Upper Hutchinsonian of Park 5b 72
Ardgowan “ shell-bed 5c 59
Target Gully “ shell-bed 5d 88
Kawa Creek beds 6 50
Waitotaran stage 7 50
Castlecliffian stage 8 31
Recent species 9 30

[Footnote] * Since the above was written the author has found Limopsis producta n. sp. also at Target Gully, and fragments of the same Cymatium sp. at Target Gully and Pukeuri, so that only one species, Fusinus macrotegens n. sp., now remains peculiar to the locality. This strongly supports the correlations here made.

– 111 –

Marshall (1919, pp. 237–47) has discussed reasons in favour of using this Lyellian percentage method of age-determination, also the minimum size of a collection that is of value on palaeontological grounds, and he considers that about forty species give a fair indication of the age of a bed, if one uses the percentage of Recent; species. When the collections are small, Thomson (1920, p. 394) considers the relation of the fauna to typical fauna of other Cainozoic stages, rather than the percentage of Recent species alone, but shows that correlation is unduly favoured with a stage, such as the Awamoan, in which a great number of species occur. Now, although the percentage of Recent species at Dowling Bay (30) is below that usual for Awamoan horizons, it is considerably above that for typical lower Hutchinsonian localities, (Park 1918, pp. 73–85,. 94, &c), and probably further collecting would raise the percentage of Recent species. If the other percentages are compared it is seen, that they point strongly to correlation with either the Upper Hutchinsonian (basal Awamoan of Uttley) or even more with the Awamoan, but not with the Lower Hutchinsonian. That the higher percentage of Awamoan forms is not due simply to the greater number of Awamoan species can be seen by studying the actual species themselves and comparing the-percentages of columns 5c and 5d. It is then at once apparent that the Dowling Bay fauna is essentially the same as that at Target Gully. Only six of the species in the former” do not- occur in the latter; and three of these are new. Even the percentage of records from Ardgowan, another Awamoan locality, falls' far below that from Target Gully. Calliostoma suteri, Aethocola spinifera, Venericardia subintermedia var., Lucinida laminata, &c, show a distinctly Awamoan facies, and the whole collection reminds one forcibly of Target Gully. The abundance of the coral Flabellum and the presence of echinoderm spines is at first suggestive of Ototaran or Hutchinsonian horizons, but the writer has collected the same species of Flabellum rather commonly at both Argowan and Target Gully, also spines and Balanus plates indistinguishable from Dowling Bay specimens. Also, Pecten huttoni Park, fish-teeth, and other corals have not so far been found at Dowling Bay, and these are. generally abundant in horizons just below the Awamoan. The brachiopodal evidence, though scanty, is also worth considering. Uttley (1920, p. 176) remarks that “ the [Pachymagas] ‘parki ’ band marks a definite horizon, the close.of the Hutchinsonian.” Now, as noted before, a single species of Hemithyris is-the only brachiopod yet found at Dowling Bay: there is no sign of the Terebra-tellidae so abundant in the Ototaran and Hutchinsonian. Considering these facts, we may conclude that the Dowling Bay limestone is on the horizon of either the Target Gully “ shell-bed ” (Awamoan), or less probably of the” glauconitic sandstone that lies conformably beneath it (Upper Hutchinsonian), and suggest that it correlates better with the former of these.

The chief importance of this conclusion lies in the inferences that may be drawn from it regarding the age of well-known formations near Dunedin. An extensive sandstone horizon (Caversham and Waikouaiti sandstones) and numerous volcanic outflows are a conspicuous feature of the geology of the district, and until the last few years the ages of these horizons were very uncertain. Park (1904) referred the Waikouaiti sandstone and the underlying foraminiferal clays simply to the Oamaruian, and placed “ the first outbursts in the later period of the Miocene, or in the beginning of the Pliocene.” Grange (1921, p. 161) places the lavas as Post-Awamoan, since they rest unconformably on a denuded surface of Caversham sandstone,

– 112 –

which, he believes, extends through the Awamoan. The fact that basalt rests also unconformably on the Dowling Bay limestone shows that volcanic activity did. not commence till a considerable time after the Awamoan, for in most localities sufficient time must have elapsed for this limestone (which is the highest marine bed in the district) to have been denuded away altogether, and part of the Caversham sandstone besides.

As regards the Caversham. sandstone, the Dowling Bay section presents important fresh evidence, and corroborates views held by some previous observers. The Waikouaiti sandstone, which is believed to be the equivalent of the Caversham sandstone, has not yet been thoroughly investigated, though there are indications that several zones with determinable fossils are present (Park, 1904, p. 422). Thomson (1918), basing his views mainly on the early Tertiary age of Waiarekan horizons and the finding at Waikouaiti of Waiparia intermedia Thomson, concludes that “ the Waikouaiti sandstone is not Ototaran, but Upper Oamaruian, and may well be Awamoan, and the underlying clays Hutchinsonian.” The palaeontology of the Caversham sandstone itself is also very meagre and inconclusive. Grange (1921, pp. 166–67) places it, together with the underlying small greensand bed, as representing Awamoan, Hutchinsonian, and Ototaran stages. It may be noted that the few molluscs found in it agree quite as well, if not. better, with the fauna of the Upper Hutchinsonian (basal Awamoan) than with that of the true Hutchinsonian. Grange records Pachymagas parki (Hutt.) and Thomson (1920, pp. 375–6) mentions P. cottoni Thomson as occurring in the Caversham sandstone, and both these range into the Awamoan. These facts, together with its general poverty in brachiopods, suggest it is of a higher age than Hutehinsonian. On the other hand, it is unlikely that the Caversham sandstone extended beyond the so-called Upper Hutehinsonian, since it lies conformably beneath the limestone at Dowling Bay, for which a low Awamoan age has been suggested. It is not yet possible to make a certain correlation, but the points discussed lead us to suggest that the Caversham sandstone is a basal Awamoan horizon, and that formerly limestones now for the most part eroded off followed conformably and were of the same age as the Target Gully beds. The highest members of the Awamoan series are without representatives near Dunedin, unless they should occur beneath the basalt;north of our line of section.

In conclusion, we wish to thank Dr. Benson for drawing the map and describing the rock-section, and also for his kindly assistance and advice in the preparation of this paper.

Limopsis producta n. sp. (Fig. 1.)

Shell moderately large, elongated, and produced posteriorly, thick, convex, with fine radial striae and strong concentric growth-lines. Beak small, pointed, incurved. Anterior end much shorter, with practically vertical slope, dorsal margin straight and slowly descending. Posterior end much produced, and subtruncated, forming a blunt angle. Basal margin broadly curving and horizontal only near posterior end. Sculpture of numerous concentric growth-lines and few radial striae, similar to the

Picture icon

Fig. 1. — Limposis producta n. sp. Natural size.

Picture icon

Fig. 1.—Fusinus macrotegens n. sp Holotype. Natural size.
Fig. 2.—Aethocola spinifera n. sp. Holotype below; two paratypes above. Natural size.

– 113 –

catenate ornamentation of L. zealandica Hutt. Margins smooth and bordered by a wide bevel edge as in L. zealandica. Hinge-plate strong, with about six teeth on each side, but much obscured by matrix. Ligamental area obscured, but produced on each side into rather prominent ears, the anterior being especially elevated. Interior filled with matrix.

Height, 22.3 mm.; length, 20 mm.; diameter of single valve, 7.5 mm.

Holotype in the author's collection.

Locality.—Dowling Bay limestone (Awamoan).

Remarks.—Closely allied to L. zealandica Hutt., seeming at first sight to be only a crushed specimen of that species; but it is difficult to see how the shape could have been so materially altered and the anterior ear so produced without breaking the shell. The specimen is certainly a little distorted, but, even allowing for this, it does not seem referable to any New Zealand or Australian species.

Fusinus macrotegens n. sp. (Plate 11, fig. 1, and text-fig. 2.)

Somewhat like F. tegens (Hutt.), but much larger and the sculpture not so bold. Keel tubercles not spinose and serving only to mark posterior ends of rounded axial costae, which are twelve per whorl; interstices a little wider. Body-whorl with about twenty prominent spiral ribs extending

Picture icon

Fig. 2.—Fusinus macrotegens n. sp. Natural size.

from suture to neck of canal, anterior ribs getting much wider apart, interstices two to three times width of spirals. Carina on spire-whorls about half-way up the whorl. Outer lip thin, regularly convex, apparently not denticulate inside; a groove separates inner lip from spiral chords. Other details as in F. tegens (Hutt.).

Height, 32 mm.; width, 15 mm. (holotype).

The unique holotype is rather crushed, protoconch and small portion of canal broken off. It is in the author's collection.

Locality.—Dowling Bay (Awamoan).

Aethocola spinifera n. sp. (Plate 11, fig. 2.)

Shell moderately large, fusiform, ventricose, strongly shouldered, thin, with short, open, notched canal, turned to the left and slightly backwards. Protoconch minute and smooth, of little more than two whorls, nucleus flattish, depressed. Post-embryonic whorls have strong axial riblets, average about twelve per whorl; the holotype (43 mm. X 23 mm.) has eight whorls, and fourteen axials on the body-whorl. Axials almost vertical, much stronger than in A. nodosa (Martyn), vanishing only near canal, interstices a little wider; they arise between keel and suture and continue down to anterior suture; on keel itself they carry strong vertically compressed

– 114 –

spines, continuing on upper keel of biangulate body-whorl, lower keel of which bears a row of similar but smaller nodules. There is a tendency for one or two lirae below this second keel to be slightly spinose. Spiral sculpture formed of fine, unequal, inequidistant threads, stronger and weaker alternating, several more prominent lirae on base; the whole reticulated by fine growth-lines. Spire lower than aperture with canal, outlines straight. Spire-whorls carinated at middle, straight below keel, very slightly concave above it; body-whorl concave between the two keels, rapidly contracted and slightly convex below them. Suture wavy, indistinct. Aperture subvertical, triangulate, fasciole prominent. Outer lip thin, not callous, with only short, linear, and distant lirae. Inner lip rather strongly callous, spreading a little beyond columella and over parietal wall, limited by groove which also truncates basal spirals.

Height (estimated), 43 mm.; width, 23 mm. (holotype).

Holotype (almost perfect) and two paratypes (imperfect) from Ardgowan in author's collection.

Remarks.—Closely allied to Siphonalia nodosa acuticostata Sut., for which it was at first taken. Suter's description and figure of the latter do not enable it to be separated from this variety at all easily, but Mr. J. Marwick assures the writer that specific differences may be noted in actual specimens. S. nodosa acuticostata Sut. has stronger lirae inside the aperture and a more concave base; its axials carry simple angles at the periphery. Mr. Marwick finds this species also at White Rock River, Pareora; while the author has recorded it from Dowling Bay in this paper, and has specimens from Awamoa and Target Gully.

List of Papers Cited.

Bartrum, J. A., 1919. A Fossiliferous Bed at Kawa Creek, Trans. N.Z. Inst., vol. 51. pp. 101–6.

Blair, W. N., 1879. Building Materials of Otago, pp. 103, 110, 112, 115.

Grange, L. I., 1921. An Account of the Geology of the Green Island Coalfield, Trans. N.Z. Inst., vol. 53, pp. 156–74.

Marshall, P., 1914. Quart. Jour. Geol. Soc., Dec., 1914.

Marshall, P., 1916. The Younger Limestones of New Zealand, Trans. N.Z. Inst., vol. 48, pp. 87–99.

Marshall, P., 1917. The Wangaloa Beds, Trans. N.Z. Inst., vol. 49, pp. 450–60.

Marshall, P., 1919. Fauna of the Hampden Beds and Classification of the Oamaru System, Trans. N.Z. Inst., vol. 51, pp. 226–50.

Marshall, P., and Murdoch, R., 1920. Tertiary Rocks near Wanganui, Trans. N.Z. Inst., vol. 52, pp. 115–28.

Morgan, P. G., 1919. The Limestone and Phosphate Deposits of New Zealand, pt. 1, p. 272.

Park, J., 1904. On the Geology of the North Head, Waikouaiti, and its Relation to the Geological History of Dunedin, Trans. N.Z. Inst., vol. 36, pp. 418–30.

Park, J., 1918. The Geology of the Oamaru District, N.Z. Geol. Surv. Bull. No. 20 (n.s.).

Speight, R., 1917. The Stratigraphy of the Tertiary Beds of the Trehssick or Castle Hill Basin, Trans. N.Z. Inst., vol. 49, pp. 321–56.

Thomson, J. A., 1918. On the Age of the Waikouaiti Sandstone, Trans. N.Z. Inst., vol. 50, pp. 196–97.

Thomson, J. A., 1920. The Notocene Geology of the Middle Waipara and Weka Pass District, Trans. N.Z. Inst., vol. 52, pp. 322–415.

Uttley, G. H., 1920. Remarks on Bulletin No. 20 (New Series) of the New Zealand Geological Survey, Trans. N.Z. Inst., vol. 52, pp. 169–82.