The Tertiary Flora of the Kaikorai Valley, Otago, New Zealand

Transactions and Proceedings of the Royal Society of New Zealand, Volume 66, 1937, Page 284

The Tertiary Flora of the Kaikorai Valley, Otago, New Zealand By W. R. B. Oliver, D.Sc., F.R.S.N.Z. Director, Dominion Museum, Wellington. (With Notes on the Diatoms by N. Ingram Hendey, F.L.S., and on the Sponge Spicules by Maurice Burton.) [Read before the Wellington Philosophical Society, April 22, 1936; received at the Editor, May 20, 1936; issued separately, December, 1936.] Material. The plant impressions described in the present paper consist of a collection made by myself in April, 1929. All were taken by digging into a bank to a depth of five or six feet on the north side of Fraser'a Creek, which flows into the Kaikorai stream, near Dunedin. In 25 cases the impressions were sufficiently representative for description. In addition there are few fragments that could not profitably be taken into account. The impressions in many cases show every detail of the nervation, though seldom is the entire outline of the leaf preserved. Each is a thin, yellowish to dark-brown layer of carbonaceous material preserved between the layers of a fine more or less redistributed volcanic tuff of a creamy-white colour. The leaves are not distributed evenly throughout the tuff, but are found thickly at certain levels, and then may follow lamina totalling 1–2 cm. in thickness without fossils at all. Some skeletons of fresh-water fish, apparently Galaxias, were included; but no insects were found. The material is preserved in the Dominion Museum, Wellington. A collection of fossil leaves made in the same locality about 1901 by Dr P. Marshall, and referred to by him in Trans. N.Z. Inst., vol. 34, p. 585, 1902, was, so Dr Marshall informs me, sent to Australia for determination. The description of these was almost completed by Mr Henry Deane, but unfortunately both the collection and manuscript were lost after his death. Professor W. N. Benson has submitted for examination some specimens of oil shale containing leaf impressions, collected at the water supply intake on Morrison's Creek in the Upper Leith Valley, where they occur in beds which are the continuation of those containing the leaves at Fraser's Gully. With the exception of a fragment which is quite similar to some specimens of Fagus australis the leaf impressions are indeterminable. Some specimens of conglomerate from Pine Hill, overlooking Woodhaugh Gardens, also were received from Professor Benson. The age of these is between the median and newest stages of the post-trachytic volcanic rocks mentioned below. A large-leaved Coprosma, close to C. vulcanica Oliver, of the Upper Pliocene, is recognisable. Professor Benson has kindly read the introductory remarks to this paper and has made some valuable suggestions which have been adopted. He has also supplied copies of the notes on diatoms and sponge spicules appended to this report.

Geological Position. The position of the leaf-bearing beds in the volcanic series in the vicinity of Dunedin has been determined by Professor Benson as between the oldest and median stages of his post-trachytic rocks. The following extracts from an abstract of his paper on the Dunedin Districts are from the Quarterly Journal of the Geological Society:—“On the eroded surface of the basement schists rest Upper Cretaceous Coal Measures, followed by a series of Upper Cretaceous to approximately Late Oligocene or Early Miocene marine sediments containing seven faunal divisions. Slight flexure and faulting with a north-east and south-west strike and general south-eastward (oceanward) inclination was followed by regional peneplanation, truncating these formations obliquely. “Volcanic eruption commenced presumably in Pliocene times by developments of anorthoclase trachyte, breccias and lavas, chiefly about foci in the eastern centre of the region mapped. These were invaded by trachyte-porphyry dykes. But before trachytic activity had ceased, basaltic activity had commenced. The innumerable post-trachytic volcanic rocks are essentially relatively thin and extensive plateau-flows, of a total thickness approximating to 2,000 feet. They are dominantly either basaltic or phonolitic and their eruption may conveniently be divided into three stages. “Oldest Stage. Basaltic agglomerates and flows … also subordinate trachy-basalts and an extensive flow of kaiwekite … also flows of trachytoid phonolite…. Anticlinal folding took place along a north-north-west and south-south-east axis through the centre of the area. Its eastern limb is the steeper. A series of great explosive vents opened along the axis. The fragmental material was widely distributed by mud-flows, producing a ‘phonolitic conglomerate’ with felspathic tuff containing leaves, fishes, and diatomaceous earth.*The material herein described from Fraser's Gully, Morrison's Creek and Millar's Creek came from these beds. “Median Stage. Trachytoid phonolites more alkaline than before. Basalts were especially well-developed…. After the eruption of a widespread olivine dolerite, there were poured out in the later part of the period trachybasalt (widespread) and trachyandesite intercalated with phonolite. Warping on north-east and south-west axes recommenced, with development of a second widespread ‘phonolitic’ conglomerate in a synclical depression….† The leaves from Pine Hill occurred in sediments of this series. “Newest Stage. Basalts, chiefly in the eastern area, with agglomerates and vitric tuffs, were followed in the south-west by ‘trachy-dolerites’ and phonolitic rocks…. Folding … became strongly accentuated … and the modern topography was determined … by this post-volcanic dislocation … modified by the erosive action of the streams on the dislocated surface.” Age. There being no remains of marine animals in any strata coming into relation with the plant-bearing beds, the age can be determined only by considering the relations of the plants themselves. Fifteen

of the species are referred to 13 recent genera. One of these is Fagus australis, which is compared with F. Moorei of Australia; and another is Ceratopetalum kaikoraiense, referred to an Australian genus. The others belong to genera found in New Zealand. All the Kaikorai species are, however, described as differing from recent species; and this in itself would indicate a place in the Pliocene. Lomaria proceroides is identical with a later Pliocene species, and Ceratopetalum kaikoraiense is related to C. pacificum Oliver of the late Pliocene of New Zealand. The remaining ten species of the Kaikorai Valley beds are related to species described from strata referred to the early and mid-Tertiary. Six belong to the family Fagaceae (Fagus, Parafagus). The others are Ulmophylon, Knightia, Calpidia, and Kaikoraia. The Kaikorai Valley species are compared with species referred to the Eocene of New Zealand, Oligocene of Seymour Island, and Miocene of Australia. Hence, an age as early as the Miocene is indicated; but, taking into consideration the affinity of most of the species of the Kaikorai Valley beds with recent species, the whole flora might indicate an age at least as old as Lower Pliocene. This age would be consistent with the history of the beds as quoted above from Benson's paper, where the topmost marine strata on which the volcanic rocks rest are stated to be Late Oligocene or Early Miocene, and the volcanic eruptions are presumed to commence in Pliocene time. Professor Benson in a letter to me stressed the extensive dislocation and dissection which has affected the volcanic series since the close of the newest era of activity. This suggests “that the series must be older than Pleistocene, while the peneplation of the Oligocene rocks which preceded the volcanic activity similarly separates it for a long period from the Oligocene. These two physiographic arguments support the relegation of the fossiliferous beds to the Pliocene.” Climate. The large proportion of species of the family Fagaceae indicates a climate which might be as cold as the present climate of Dunedin, but yet could be as warm as the climate of the Wellington Province at sea-level. The presence of the following genera, however, stamps the climate as warm temperate: Knightia, Calpidia, Ixerba, and Metrosideros (especially M. laeta). It is safe to say, therefore, that a climate warmer than that at present prevailing in Dunedin obtained in that locality when the Kaikorai Valley plant remains were embedded. Plant Formations. The leaves most frequently encountered in the Kaikorai Valley beds are those of the various species of Fagus and Parafagus. Thus, a beech-forest with a considerable admixture of other broad-leaved trees is indicated. Calpidia and Senecio would indicate proximity to the sea. Laurelia shows that a stream was nearby. The skeletons of Galaxias confirm this, while the leaves of Typha show that there was a swamp or pool in the valley.

Relationships. The two outstanding features of the Tertiary flora of Kaikorai Valley are (1) that more than half of the species are closely related to recent New Zealand species; (2) that the Fagus type includes a much larger proportion of the species than it does in the recent flora. Taking first the element related to the recent New Zealand flora, eleven species (44 per cent.) are described in the following account as related to New Zealand endemic species. This is a sufficient proportion on which to base the statement that the recent flora had developed its main characteristics in New Zealand by mid-Tertiary time. The following genera, of importance in the recent flora, were represented in the Kaikorai Valley flora :—Senecio, Coprosma, Genio-stoma, Nothopanax, Metrosideros, Coriaria, Ixerba, Laurelia, and Podocarpus. In the same New Zealand element must also be included Lomaria proceroides and Typha robusta. There is also an endemic New Zealand element in the Fagus group. Thus Fagus novaezelandiae, Parafagus, three species, Ulmophylon pliocenica, and, probably, Kaikoraia gracilia, form a New Zealand assemblage not to be closely related to Australian species. The remaining species show a definite relationship with Australia. Thus Fagus australis resembles the recent F. Moorei, F. kaikoraiensis is related to the Tertiary F. Benthami, Calpidia zelandica is an ally of “Artocarpidium” Stuarti, and Ceratopetalum kaikoraiense belongs to a genus now found living only in Australia. There are two species, Fagus maorica and Knightia oblonga, which can be compared with Seymour Island Oligocene species; but too much weight should not be attached to these. With regard to the variety of species of the Fagus type present in the Kaikorai Valley flora (Fagus, four species; Parafagus, three species) it is to be noted that this feature also characterizes the Australian Tertiary flora. Ettingshausen stressed this point and stated that the New Zealand Tertiary flora is part of that universal original flora from which all living floras of the globe descend. The implication here seems to be that the same flora once occupied both Australia and New Zealand, and, of course, other countries as well. While this may have been true in mid-Mesozoic times, the evidence of the Tertiary floras does not support this theory. The Kaikorai Valley flora when analysed in the light of the recent New Zealand and Australian floras shows not only that in mid-Tertiary time the floras of Australia and New Zealand were almost as distinct as they are at the present time, but also that each had developed the main characteristics which now distinguish them. Each had, it is true, a common element in the Fagus group, but the main flora of New Zealand was just as much Malayan in its base at it is at present. Even if there was a Fagus type of flora universal in Cretaceous times, it was not the origin of the present New Zealand flora, which in the main is related to a non-Fagus flora which developed in the Malayan region. The conclusion to be drawn from a study of the Kaikorai Valley flora, therefore, is that in mid-Tertiary times there had developed in New Zealand a flora similar to the present flora and,

therefore, presumably, with Malayan relations. Mixed with it was a flora with Fagus types predominating, and this flora had affinities with a similar element in the Australian flora. This element is also present in the Tertiary and in the recent flora of temperate South America. Possibly, therefore, it is the remains of a widespread Antarctic or perhaps universal flora which developed in Cretaceous time, but was mostly replaced by a flora composed of higher dicotyledonous types which specialised in various parts of the world into groups which formed the foundations of the recent floras. The New Zealand flora at present dominant thus did not arise from the supposed widespread Cretaceous flora of Fagus type, but developed from a higher type of flora which gradually replaced the Fagus flora. The Oligocene flora of Seymour Island has its main elements corresponding with those of the Tertiary floras of Australia and New Zealand, namely, an element related to the recent South American flora and a Fagus group. Fig. 1.—Lomaria proceroides Oliver. Nat. size. Fig. 2.—Podocarpus obtusifolius Oliver. × 3. Fig. 3.—Typha robusta Oliver. Nat. size. Descriptions of the Species. Family Polypodiaceae. Lomaria proceroides W. R. B. Oliver. (Fig. 1.) Trans N.Z. Inst., vol. 59, p. 289, 1928. Portion of leaf-impression; length 40 mm., width 22 mm. Nerves 11 to the space of 10 mm. Mostly separate, but a few forking at or near the base. Leaf-margin serrated, the anterior margin of each tooth short, steep, opposite the nerve, the posterior margin nearly as long as the distance between the nerves. This leaf does not show the base, but the blade corresponds with that of the Later Pliocene L. proceroides from the Waipaoa series at Ormond, Poverty Bay. The Early Tertiary Lomariopsis dunstanensis Ett. is a narrower-leaved species with similar nervation and margin. On account of the spots on the upper surface Ettingshausen thought it to have been covered with sori.

Family Podocarpaceae. Podocarpus obtusifolius W. R. B. Oliver, n.sp. (Fig. 2.) Impressions of a twig 25 mm. long with more than 10 leaves attached. Leaves 10–12 mm. long; 2.5 mm. wide. Leaves elongate-oval, base unequal-sided, the lower margin nearly straight, the upper margin expanding abruptly into the full width of the leaf; sides parallel, tip rounded. In its oblong form and blunt, rounded tip this species differs from all the New Zealand Recent and Tertiary species of Podocarpus. Family Typhaceae. Typha robusta W. R. B. Oliver, n.sp. (Fig. 3.) Two impressions of portions of leaves; width 21 mm. About 14 ribs, between which are faint parallel striations. The ribs are distinct in the central portion of the leaf; near the margin they are not traceable continuously. The characters of these specimens are precisely those of the leaves of Typha. The leaves are wider than in the Recent T. angustifolia, and have a few more ribs. In order to distinguish these differences the name robusta is proposed. Family Fagaceae. On leaf characters the seven species from the Kaikorai Valley belonging to this family are divided into the following three genera :— Fagus—secondary nerves straight, teeth one or two to each nerve. Nothofagus—secondary nerves straight, teeth several to each nerve. Parafagus—secondary nerves curved. These characters are perhaps artificial, but they serve to distinguish groups which are convenient in the present state of knowledge of Tertiary floras. Furthermore, those referred to Fagus and Nothofagus can be compared with Recent species classed under these genera. When describing the Seymour Island Oligocene flora Dusen used both Fagus and Nothofagus, apparently referring leaves of leathery texture to Nothofagus. This also is an artificial distinction. Fagus novae-zealandiae W. R. B. Oliver, n.sp. (Fig. 4.) Several impressions of leaves, some showing the apex, others the base. The largest leaf-blade, almost complete, measures 70 by 35 mm. Leaf ovate, the base rounded. Apex gradually tapering to an acute produced tip. A single small acute tooth projects at each secondary nerve termination; between the teeth the margin is straight or very slightly incurved. The teeth point towards the apex of the leaf. Midrib straight or very slightly flexuous in the middle third. Secondaries 8–10 on either side of the midrib, arising at regular intervals at an angle of 40°–45°, straight, meeting the margin at the teeth and there turning slightly forward; not strong and not closely

placed near the tip of the leaf. In places the tertiary reticulation is indicated, the principal nerves crossing at right angles between the secondaries. This species resembles in venation and margin the northern Fagus sylvatica and accordingly appears to be a true Fagus. A Miocene species from Idaho, F. pacifica, is very like it in shape, venation, and margin. F. novae-zealandiae bears a general resemblance to F. risdoniana, of the Upper Tertiary of Risdon, Tasmania. The venation is similar, also the produced apex; but the base is a little wider. F. risdoniana is described as being sometimes doubly dentate near the apex. Fig. 4.—Fagus novae-zelandiae Oliver. Nat. size. Fig. 5a, 5b.—Fagus marorica Oliver. Nat. size. F. novae-zealandiae differs from F. maorica in its thinner texture, weaker secondaries and produced apex. Fagus maorica W. R. B. Oliver, n.sp., (Fig. 5.) A number of leaf-impressions all showing similar characters and easily recognised. Average size 45 by 25 mm. Leaf ovate with acute apex and rounded or emarginate base. A single prominent acute or acuminate tooth terminating each secondary nerve. Midrib strong. Secondaries strongly impressed. parallel, about 12 on either side of the midrib, closely placed. Tertiary nerves much weaker than the secondaries, finely reticulated, with the principal nerves at right angles to the secondaries. The dominant features of this species are the strong, close, parallel secondaries, and the conspicuous teeth. It appears to be quite similar to F. obscura Dusen from the Oligocene of Seymour

Island. Dusen's figure shows the same close parallel secondaries, but the margin between the teeth is more deeply incised than in maorica. F. maorica approaches F. novae-zealandiae, but differs in the less produced apex and much stronger ribs, which become closer and not less prominent towards the tip. It apparently belongs with F. novae-zelandiae and F. pacifica to the F. sylvatica group. Fagus australis W. R. B. Oliver, n.sp. (Fig. 6.) Several leaf-impressions varying a good deal in general shape. Two leaves, almost complete, measure 80 by 40 mm. and 55 by 26 mm. Leaf ovate, or broadly ovate, apex gradually tapering to an acute tip, base rounded or widely cuneate; margin sharply serrate with a mucronate point at the termination of each secondary, and a less prominent point between them but nearer the apically-situated secondary. Midrib strong, very slightly flexuose about the centre. Secondaries arising at an angle of 45°, straight, 10 to 12 on either side of the midrib. Tertiary reticulation fine with the larger nerves at right angles to the secondaries. Figs. 6a, 6b.—Fagus australis Oliver. Nat. size. Fig. 7—Nothofagus kaikoraiensis Oliver. Nat. size. Differs from F. novae-zealandiae in the less regular shape, the midrib being generally curved to one side at base and apex; the more tapering apex, stronger, more pointed marginal teeth, and extra tooth between the secondaries. A near ally appears to be F. Benthami of Vegetable Creek, New South Wales, in ironstone shale, which has a doubly dentate leaf, broad at base, but the petiole is longer and thinner than in F. australis. Of Recent species with two teeth to each secondary, F. Moorei F. v. M. resembles F. australis in its general shape. Nothofagus kaikoraiensis W. R. B. Oliver, n.sp. (Fig. 7.) Several leaf-impressions, all with the margins imperfect. Size 55 by 30 mm., 42 by 26 mm.

Leaf ovate, apex gradually narrowed and with the tip produced, base broadly cuneate Margin finely and irregularly serrate, 3 to 4 teeth to each secondary. Midrib slightly flexuose. Secondaries 8–9 on either side of the midrib, strong, nearly straight, ending in a tooth on the margin, lower ones arising from the midrib at 50°, the angle decreasing towards the apex, where it is 35°. In general shape and serration this species resembles the Australian Tertiary Fagus Benthami Ett., F. Muelleri Ett. and Betula derwentensis Ett., all of which probably would be better classed under Nothofagus. It does not resemble closely any of Ettingshausen's species of Fagus with serrated leaves from the New Zealand Tertiary, nor any Recent species of Nothofagus. Parafagus W. R. B. Oliver, n.gen. Leaves with dentate margin, one tooth terminating each secondary nerve and one served by a branch from its lower side. Nervation craspedodromus, secondaries arched forward, tertiary venation well marked. Type: P otakouia W. R. B. Oliver. Kaikorai Valley, Otago, Pliocene. The leaves of this genus resemble those of Fagus in their craspedodromus venation and dentate margins; but the secondaries are curved and the tertiary reticulation conspicuous. Ettingshausen referred this type of leaf, for instance, Quercus Blamingi, to Quercus; but that genus differs in the regular, almost parallel, tertiary nerves stretching at right angles from one secondary to the next. It is impossible, of course, in the absence of flowers or fruit, to say to which family Parafagus belongs; but, on account of its craspedodromous venation and general resemblance to the Fagus type of leaf, I place it in the family Fagaceae. It appears to take in the Tertiary of New Zealand and Australia the place of Quercus in the northern hemisphere. Parafagus otakouia W. R. B. Oliver, n.sp. (Fig. 8.) Well-preserved impressions of both sides of a single leaf. Length (estimated) 60, breadth 27 mm. Leaf broadly oblong, apex and, apparently, base, narrowing rather abruptly. The margin with prominent teeth pointing towards the apex, outer side of teeth regularly sinuous. One tooth at the termination of each secondary nerve, and another, generally smaller, between the secondaries but near the apical end. Midrib and secondaries strong; remaining nerves much weaker. Secondaries, about 7 on either side of the midrib, alternate, slightly flexuous, curved towards the apex and terminating in the tips of the teeth. A fairly strong nerve branches off near the end of each secondary and passes to the intermediate tooth. Tertiary reticulation irregular. The distinctive features of this leaf are the broad short form, craspedodromous venation and two large teeth to each secondary nerve. It differs from P. pinnata and P. angustifolia in the broader form, fewer secondaries, and relatively larger teeth.

Parafagus pinnata W. R. B. Oliver, n.sp. (Fig. 9.) Several leaf-impressions. Length (estimated) 65 mm., breadth 30 mm. Leaf oblong, apex acute, base rounded. Margin regularly serrate, there being two, rarely three, prominent teeth for each secondary; teeth pointing towards apex of leaf, outer margin sinuous. Midrib and secondaries distinct; tertiary reticulation much less so. Secondaries, about 12 on each side of the midrib, alternate and opposite, regularly arched towards the apex, and ending in the sinus near the tip of each tooth. A branch from each secondary leads to the intermediate tooth. Tertiary venation irregular. Differs from P. otakouia in the narrower form of the leaf, more numerous secondaries and less prominent serration. Fig. 8.—Parafagus otakouia Oliver. Nat. size. Fig. 9.—Parafagus pinnata Oliver. Nat. size. Fig. 10.—Parafagus angustifolia Oliver. Nat. size. Parafagus angustifolia W. R. B. Oliver, n.sp. (Fig. 10.) A few leaf-impressions. Length (estimated) 45 mm., breadth 17 mm. Leaf narrow oblong, base cuneate (apex missing). Regularly dentate, there being two teeth to each secondary nerve; the tips of the teeth directed apically; margin between teeth regularly sinuate. Midrib and secondary nerves distinct; tertiary reticulation indistinct. Secondaries about 9 on each side of the midrib, alternate and opposite, regularly arching forward, and ending in the sinus next the tip of each alternate tooth, a branch from the basal side of the secondary serving the intermediate tooth. Tertiary reticulation irregular. Differs from P. pinnata in being proportionately narrower, and in the fewer secondaries and cuneate base.

Family Ulmaceae. Ulmophylon pliocenicum W. R. B. Oliver, n.sp. (Fig. 11.) Impressions of two leaves, both imperfect. Estimated size 60 by 40 mm. Leaf ovate, apex acute, base broadly cuneate. Margin irregularly and closely serrate, the teeth at the termination of the secondaries and their main branches being larger than the others. Midrib strong, flexuous. Secondary nerves 5–6 on either side of the midrib, widely spaced, arising from the midrib at an angle of about 40°. The lowest secondaries bear on their basal sides three branches, the remainder one or two branches, near the tip; these all reach the margin at prominent teeth. The tertiary venation consists of a fine network strengthened by nerves crossing nearly at right angles to the secondaries. Ulmophylon was founded by Ettingshausen*Denksch. Akad. Wiss. Math. Nat., 53, bd. 1, p. 44, 1887. for two species from the early Tertiary of New Zealand. The first species U. latifolium Ett., which may be selected as type of the genus, seems to be congeneric with the present species. It has, however, broader leaves with the secondaries arriving from the midrib at a much wider angle. Family Proteaceae. Knightia oblonga W. R. B. Oliver, n.sp. (Fig. 12.) Impressions of a leaf with base and margin of one side complete. Estimated size of leaf 90 by 38 mm. Leaf oblong, the base widely diverging. Margin wavy, with regular blunt teeth directed outwards. Midrib rather strong, showing longitudinal surface wrinklings. Secondaries arising at regular intervals of 6–7 mm., at angles of 65°–70°, nearly straight, generally forking in the marginal fourth, with indications of the branches of adjoining nerves uniting. Tertiary reticulation not preserved. K. oblonga is perhaps more like K. Andreae Dusen, of the Oligocene of Seymour Island, than any New Zealand species; it resembles it in the marginal dentations and the angle at which the secondaries arise. The looping of the nerves is, however, much more clear in K. Andreae. The broad oblong form of the leaf of K. oblonga separates it from both the recent K. excelsa B. Br. and the Later Pliocene K. fossilis W. R. B. Oliver. Family Mirabilidaceae† As Nyctago Juss. is a synonym of Mirabilis L., I suggest this name for the family instead of Nyctaginaceae. Calpidia zealandica W. R. B. Oliver, n.sp. (Fig. 13.) Several leaf-impressions, all imperfect. Estimated size of largest leaf 16 × 7 cm. Leaf apparently thin, oblong, the apex obtuse and rounded (specimens imperfect), base abruptly widening from the petiole to the nearly parallel sides; margin entire. Midrib broad, 4 mm. wide at base, flat, slightly expanded beneath the origin of the secondaries.

Fig. 11.—Ulmophylon pliocenicum Oliver. Nat. size. Fig. 12.—Knightia oblonga Oliver. Nat. size. Fig. 13.—Calpidia zealandica Oliver. Nat. size. Secondaries arising at various angles, 40° to 50°, from the midrib; some are regularly arched forwards; others are straight; they may fork in the distal half; they do not reach the margin, but curve round and unite with the vein nearer the apex. There is a fine tertiary reticulation, the principal nerves being irregular and mostly at right angles to the secondaries.

This species resembles in its main features the leaf described by Ettingshausen under the name of Artocarpidium Stuarti, which has a produced apex, but the midrib and venation are extremely like those of C. zealandica. A Stuarti is classed as Miocene by Winkinson and as Eocene by Ettingshausen. It is an Australian species and should be included in Calpidia. The leaves here described are quite distinct from those properly included under Artocarpidium. They agree essentially with those of the Recent genus Calpidia, a widely spread tropical genus. Until quite recently the species were included in the genus Pisonia. Family Monimiaceae. Laurelia cuneata W. R. B. Oliver, n.sp. (Fig. 14.) Impression of two imperfect leaves. Estimated size 70 by 35 mm. Base of leaf cuneate, the sides diverging at an angle of 35° from the midrib. The basal margin is entire, but the sides of the leaf bear regular teeth with blunt, rounded tips directed towards the apex. Midrib wide; petiole long and stout. The secondary nerves are numerous and weak; they arise from the midrib at an angle of about 50° and extend in an irregular course to the margin, which they meet on the inner side of the teeth. The leaf is not, however, strongly craspedodromous, as there are -prominent cross nerves, and the secondaries sometimes branch towards the margin; but there is not a prominent series of loops as in a camptodromous leaf. The tertiaries give a fine irregular reticulation of quadrangular or polygonal lacunae. This leaf can be compared with that of the Recent L. novae zealandiae A. Cunn., in the cuneate base, serrated margin and general type of venation. It differs in the stout petiole and midrib, narrower base with longer entire portion, shallower indentations between the teeth, and less regular secondaries. In L. novae zealandiae the secondaries arise at regular intervals and, though mostly traceable to the margin as in L. cuneata, they form fairly regular arches as in the typical camptodromous leaf. L. cuneata comes nearer to L. novae zealandiae than to the Oligocene Seymour Island L. insularis Dusen, which, according to its author, is related to the South American L. sempervirens. Family Saxifragaceae. Ixerba semidentata W. R. B. Oliver, n.sp. (Fig. 15.) Impression (45 × 20 mm.) of basal portion of leaf. One side is crumpled, with the veins near the margin obscured; the base is twisted towards the same side. Leaf narrow-oblong, base cuneate. Base entire, but sides bear distant low teeth directed apically and with a shallow sinus within the blunt points. Midrib strong, narrow. Secondaries weak, regular, arching round in camptodromous fashion and meeting the veins in front. Tertiary reticulation consisting of a rather open network.

This-leaf resembles in essential characters the Recent I. brexioides A. Cunn. It differs in the more abruptly narrowing base, more distant teeth with shallower identations and more widely spaced secondaries. Ceratopetalum kaikoraiense W. R. B. Oliver, n.sp. (Fig. 16.) Two leaf impressions, both imperfect. Estimated size of leaf 60 by 18 mm. Leaf narrow-oblong; base cuneate; its margin forming an angle of about 45° with the midrib; margin regularly serrate, about two teeth to each secondary nerve, the tips of the teeth produced and apically directed. Midrib narrow; all other nerves weaker. Secondaries, about 8 on either side of the midrib, arising at an angle of about 45°, flexuous, traceable to the marginal teeth, connected by cross nerves, but not definitely looped, forking about two-thirds the distance to the margin. Tertiary reticulation evident in places. Fig. 14.—Laurelia cuneata Oliver. Nat. size. Fig. 15.—Ixerba sedimentata Oliver. Nat. size. Fig. 16.—Ceratopetalum kaikoraiense Oliver. × 2. The oblong shape, serrated margins and weak secondaries without pronounced looping place this species near other Tertiary species from Australia and New Zealand referred to the genus Ceratopetalum. Possibly they are related; but this cannot be known with certainty in the absence of flowers. C. kaikoraiense has much the appearance of the Later Pliocene C. pacificum W. R. B. Oliver, but the sides are more nearly parallel and the serration comes closer to the petiole. C. praearbuttoides Ett., of the Tasmanian Tertiary, approaches C. kaikoraiense very closely in shape and serration. Ettingshausen's Myrica proxima also resembles C. kaikoraiense.

Family Coriariaceae. Coriaria latepetiolata W. R. B. Oliver, n.sp. (Fig. 17.) Impression of portion of large leaf and of a smaller one. Estimated size of large leaf 100 × 40 mm. petiole 10 × 4 mm. Leaf ovate, base cuneate and forming a wide angle with the sides, which, at the middle of the leaf, are nearly parallel to the midrib; margin entire. Petiole long and broad, formed by the combined midrib and leading lateral nerves. Midrib wide and strong. From the top of the petiole on either side a strong rib runs about halfway between midrib and margin; it is about half the width of the midrib and slightly flexuous, being bent where cross nerves meet it. The reticulation consists of (1) between the midrib and leading lateral, secondaries arising from the midrib at an angle of 50° and extending to the laterals; they are weak and irregular and connected by weaker nerves forming a coarse network: (2) between the laterals and the margin arise nerves similar to the secondaries and about the same distance apart; generally they arise near the termination of the secondaries, but are not direct extensions of them; these nerves loop forward, joining those in front and thus give rise to an irregular intra-marginal nerve well away from the margin; the reticulation is like that between the midrib and laterals. Fig. 17.—Coriaria latepetiolata Oliver. Nat. size. Fig. 18.—Metrosideros laeta Oliver. Nat. size. Fig. 19.—Metrosideros pliocenica Oliver. × 3. C. latepetiolata agrees with the Recent C. arborea in the large size of the leaf, the disposition of the leading lateral nerves, the presence of an intra-marginal looped nerve, and in the fact that the

petiole is broad and formed of the combined midrib and laterals. It differs in the cuneate base and longer petiole, in the wider spaced secondaries, and in the absence of a second pair of laterals arising from the base. Family Myrtaceae. Metrosideros laeta W. R. B. Oliver, n.sp. (Fig. 18.) Impression of leaf imperfect at apex. Assuming the apex to be acute the complete leaf is estimated to measure 50 × 24 mm. Leaf ovate, rounded at the base, the sides regularly bowed with the widest part basal to the middle. Superficially only the midrib is evident and thus in life the leaf would have been smooth. In reflected light other nerves are indicated. The secondaries are about the same width as the midrib; they are weak, more or less parallel and extend to a faint marginal nerve which arises just above the insertion of the petiole. The general shape of the leaf is that of the Recent M. robusta A. Cunn., except that the widest portion is below the middle. The arrangement of the nerves is precisely that of M. robusta, but this species dries with the nerves rather prominent, whereas in the fossil they are only faintly indicated in places. Metrosideros pliocenica W. R. B. Oliver, n.sp. (Fig. 19.) Impressions of incomplete leaf. Estimated size 22 × 12 mm. Leaf ovate, base narrowed rather abruptly to a short petiole. The apex is missing, but the direction of the marginal nerves indicates that it was acute. Midrib strong. On either side near the base arises at an angle of between 25° and 30° a leading lateral nerve which curves slightly and leads towards the margin approaching it about halfway to the apex and then continues as a marginal nerve, The lower part of the leaf is served by marginal nerves arising adjacent to the leading lateral nerves. Tertiary reticulation consisting of a fine network supported by stronger nerves arising from the midrib at an angle of 50°–60° and others connecting the leading lateral nerves with the lower marginal nerves. This leaf belongs to a group of species of Metrosideros having small leaves. In the disposition of the lateral and marginal nerves it approaches M. diffusa (Forst.) Oliver (= M. hypericifolia), but differs in the lateral nerves, being straighter and reaching the margin further from the base. The base of the leaf in M. pliocenica is acute, whereas in M. diffusa it is often expanded and cordate. Family Araliaceae. Nothopanax distans W. R. B. Oliver, n.sp. (Fig. 20.) Impression of a single leaf or leaflet showing the venation and the margin of one side. Estimated size of complete leaf 80 × 28 mm. Leaf ovate, base cuneate, gradually narrowed to the petiole. On the margin are preserved three distant teeth, the first being less than halfway to the apex. Teeth mucronate, directed apically; margin between the teeth straight. Midrib strong. Secondaries prominent,

about 8 on either side of the midrib, subopposite, arising at an angle of 35°, flexuous and curving forward in their distal portions, branching well within the margin, terminating in the marginal teeth. In its general shape, venation and marginal toothing this leaf resembles the Recent N. Sinclairii (Hook. f.) Seem., from which it differs in its larger size, more oblong shape and more distantly placed teeth. Fig. 20.—Nothopanax distans Oliver. Nat. size. Figs. 21a, 21b.—Koikoraia gracilis Oliver. Nat. size.

Family Sapotaceae. Kaikoraia W. R. B. Oliver, n.gen. Leaves lanceolate, entire, margin slightly sinuous; secondaries numerous, reaching the margin and with conspicuous connecting nerves forming a wide-meshed reticulation. Type: K gracilis W. R. B. Oliver. Kaikorai Valley, Otago, Pliocene. The leaves of K. gracilis bear a resemblance to those of Sideroxylon in their entire, slightly sinuous margins and the characters of the secondary nerves, except that in Sideroxylon, the secondaries for the most part loop round and only the branches from the loops reach the margin (camptodromous venation). It is because of the definitely craspedodromous type of venation that I provide a new genus for the Kaikorai Valley fossil. To Kaikoraia may be referred Apocynophyllum elegans Ett. (Denkschr. Math. Nat. Kaiserlich Acad., Wiss. 53, p. 28, 1887), in which the secondaries are figured as reaching the margin. A third species to be referred to this genus may be A. berwickense Deane (Rec. Geol. Surv. Vic., 1, p. 25, 1902), of which the figure shows a shape and venation essentially like K. gracilis. Kaikoraia gracilis W. R. B. Oliver, n.sp. (Fig. 21.) Several impressions of leaves. One almost complete measures 140 × 28 mm. Leaf lanceolate, apex acute, base gradually narrowed to a wide petiole; margin entire, very slightly sinuous. Midrib broad throughout. Secondaries numerous, rather prominent, roughly parallel but irregular, arising at an angle of 60° from the midrib and reaching the margin of the leaf; everywhere connected by tertiaries forming a large-meshed network. Differs from Apocynophyllum elegans Ett., from Landslip Hill, Otago, in the narrower shape, stronger secondaries and wider-meshed reticulation. Family Loganiaceae. Geniostoma oblongum W. R. B. Oliver, n.sp. (Fig. 22.) Impression of a leaf 83 × 33 mm., petiole 8 mm. Leaf oblong, tapering equally at either end, apex produced into a blunt lobe. Margin entire. Petiole slender. Midrib faintly impressed, narrow. Secondaries very faintly impressed, arising from the midrib at an angle of 50° and slightly curved forwards. This leaf is placed in the genus Geniostoma because it resembles the leaf of the Recent G. ligustrifolium A. Cunn. in general shape, thin texture and lightly impressed nerves. It differs from G. ligustrifolium in being more regularly oblong (that is, narrower base and wider apex), in the tip being less produced, and in the secondaries arising at a narrower angle.

Family Rubiaceae. Coprosma praerepens W. R. B. Oliver, n.sp. (Fig. 23.) Impression of portion of a leaf. Estimated size of complete leaf 50 × 36 mm. Leaf broadly ovate, widest about the middle, apical end expanded and truncate, with a shallow emargination, base making a distinct angle with the side and narrowing abruptly to the petiole. Margin entire, with a well-marked thickened edge. Midrib broad, slightly flexuous. Secondaries, about 6 on either side of the midrib, weak, arising at an angle of 50°-55° and looping round near the margin to meet the one in front. Tertiary venation weak, with large lacunae, indicating a semi-succulent texture. This leaf agrees in all essential points with Coprosma of the group of C. Baueri. It resembles C. repens Rich. in shape, but has more secondaries, which are weaker and arch to meet those in front nearer the margin. Fig. 22.—Geniostoma oblongum Oliver. × ½. Fig. 23.—Coprosma praerepens Oliver. Nat. size. Fig 24.—Coprosma pliocenica Oliver. Nat. size. Coprosma pliocenica W. R. B. Oliver, n.sp. (Fig. 24.) Four leaf-impressions. Estimated sizes of complete leaves 75 × 28 mm., 85 × 35 mm. Leaves obvate, apex rounded, obtuse or slightly emarginate, base cuneate, gradually narrowed into a stout petiole; margin entire. Midrib strong, the tip curving to one side in two of the specimens. Secondaries, about 6 on either side of the midrib, arising at an angle of about 50°, arching regularly forwards and consisting of rather strong cross nerves enclosing a finer network. The closest ally of this species appears to be the Recent C. robusta Raoul, with which it agrees in the shape of the basal portion and in the type of venation. The apex is more rounded than in C. robusta and approaches that of C. repens Rich. C. pliocenica resembles the fossil C. pseudoretusa Penseler, but has more curved and more distantly placed secondaries, a more obtuse apex and a narrower base.

Family Compositae Senecio pliocenicus W. R. B. Oliver, n.sp. (Fig 25.) Single imperfect impression of a leaf. Estimated size of complete leaf 60 × 43 mm. Leaf broadly oblong, the sides rounded, abruptly contracting to base and apex which were probably rounded; margin regularly sinuous. Midrib broad. Secondaries weak, arising at an angle of about 55°, 5 or 6 main ones and several smaller ones on either side of the midrib; these fork about two-thirds the way to the margin, and are connected submarginally, but there are no prominent loops. The whole leaf surface is covered with a regular reticulation of polygonal cells, in many of which free nerve-endings may be seen; near the margin is in places a nerve plexus, while the margin itself is broad and well-marked. Fig. 25.—Senecio pliocenicus Oliver. Nat. size. This leaf resembles quite closely those of the Recent S. rotundifolius Hook f. In both the lamina is suborbicular, regularly reticulate with free nerve-endings in the lacunae, and the loops between the secondaries not pronounced. S. pliocenicus differs from S. rotundifolius in the regularly sinuate margin, better defined intermediate secondaries, wider angle at which the secondaries arise from the midrib, and the more pronounced margin. Diatoms. Notes by N. Ingram Hendey, Esq., F.L.S. Specimens of the diatomaceous earth associated with the leaf-bearing beds at Fraser's Gully and Morrison's Creek, and also from the exposure of the same series of beds in Millar's Creek south-west of Waitati, were submitted by Professor Benson to Mr N. Ingram Hendey, of the British Museum of Natural History, who reported concerning them as follows:— “Considerably less than 50% of the silica is diatomaceous, and the most important species I have interpreted widely as Melosira granulata. This species is very common in similar deposits throughout the southern part of the Middle East, e.g., Arabia, India, and the

East Indies.” (It also occurs in the Tertiary diatomaceous earths of New South Wales.) “The other species are typical European fresh-water diatoms. The deposits are undoubtedly of fresh-water origin and appear to have undergone denudation and secondary deposition, or considerable pressure at some time or other. The extreme fragmentation of all the forms present suggests that the material has undergone considerable rough treatment. Millar's Creek, Lower Waitati Valley. Melosira granulata (Ehrenberg) Ralfs. Morrison's Creek. Melosira granulata. Fraser's Gully. Melosira granulata. Navicula placentula forma latiuscula (Grun.) Meister. Pinnularia viridis (Nitsch) Ehrenberg.” Navicula cryptocephala var. intermedia Grun. Sponge Spicules. Notes by Maurice Burton. Associated with the diatoms were traces of fresh-water sponge spicules which were submitted by Mr Hendey to Mr Maurice Burton, of the same museum, and his remarks thereon may here be recorded. “I find in the three samples from New Zealand one predominant spicule form, the spined oxeote. There is only one thing that can be said with certainty, and that is that these spicules belonged to fresh-water sponges. It is impossible to name a species or even a genus with any degree of certainty, because the all-important spicules from the taxonomic point of view, namely, the gemmule spicules, appear to be entirely absent. The absence of these spicules even in Recent fresh-water sponges makes it very difficult indeed to identify a specimen, and one has to fall back on such features as external form, the construction of the skeleton and the texture of the sponge, all of which are obviously missing in the scattered remains of a fossil sponge. So far as it is possible to hazard a guess the spicules belong to a sponge representing perhaps the genus Heteromeyenia, and closely resemble in size, shape, the variations they present, and the monstrosities they occasionally give rise to, those spicules figured by Jane Stephens in the Freshwater Sponges of Ireland, Proc. Roy. Irish Acad., Dublin, vol. 35, section B, no. 11, 1920, plate 28, fig. 2a. These spicules figured by Stephens belong to Heteromeyenia ryderi, a species commonly found in North America, occasionally in Ireland, and rarely in some parts of Scotland. It is doubtful, therefore, whether the New Zealand specimens belong to this species, but they do belong to a closely related species. “There are in addition two more categories of siliceous bodies which have puzzled me somewhat. To one belong a number of spheraster-like bodies about .008 mm. in diameter, and to the other belong a number of irregular siliceous bodies very much resembling the fragments of Lithistid sponges. As both of these suggest the spicules of marine sponges, and especially as it is not possible to get a good view of the first or a complete example of the second, I have assumed that they are derived from some other source than the skeletons of the sponges, and have accordingly ignored them.”