The Flora of the Waipaoa Series (Later Pliocene) of New Zealand.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 59, 1928, Page 287

The Flora of the Waipaoa Series (Later Pliocene) of New Zealand. By W. R. B. Oliver, M.Sc., F.L.S., F.N.Z., Inst. [Read before the Wellington Philosophical Society, 24th October, 1926; received by Editor, 9th March, 1928; 10th August, 1928.] Plates 44-45. Material: The plants described in the following account consist of two small collections made in the same locality, namely, near Ormond, in the Poverty Bay District, New Zealand. The first was made by Mr. H. Hill, about 1910, the second by Mr. M. Ongley, of the New Zealand Geological Survey, during 1914-16. Both collections are the property of the Geological Survey, and were placed in the hands of the author for identification by the late Director, Mr. P. G. Morgan. The specimens are at present deposited in the Dominion Museum, Wellington. The material consists of carbonized remains and impressions of leaves and in some cases fruits on a fine-grained volcanic silt. Many of the specimens were too fragmentary for identification, but from those sufficiently well preserved, thirty-one species have been described. Associated with the leaves were several impressions of fish skeletons and fins. In size they agree with the Recent fresh-water fish Gobiomorphus. Geological Horizon: The fossils were found in a finely-laminated pumiceous silt which outcrops along an escarpment for about half a mile. Impressions of leaves occurred here and there along the western end, but only in places were the leaves found in any quantity. The author visited the locality in 1926, but found only a few leaf remains. The beds lie in a Tertiary valley, and with others of similar lithological character and stratigraphical position are classed as the Waipaoa Series. Age. (1) Palaeozoological. The information under this head has been kindly supplied me by Dr. J. Marwick, Palaeontologist to the New Zealand Geological Survey Department. In their report on the Geology of the Gisborne and Whatatutu Subdivision, Henderson and Ongley give a list of mollusca supposed to have been found in beds of the Waipaoa Series in the Kaiti Hills and said to be of older Pleistocene age. Dr. Marwick informs me that the authenticity of the collection is challenged, that it is very like Castlecliff material, and that recent search has failed to find a similar fauna in the locality. Further, it is of later Pliocene age. In the Kaiti Hills Dr. Marwick could find only impressions of Chione stutchburyi (Gray). The only other locality in which beds of the Waipaoa Series contain marine fossils is a hill east of the junction of the Oweka and Maddox streams. Here the following mollusca were collected by the officers of the Geological Survey:—Nucula hartvigiana Pfr.,

Ostrea angasi Sow., Dosinia grayi Zitt., Chione stutchburyi (Gray), Bassina Yatei (Gray), Cominella aff. adspersa Brug., Acteon sulcatus (Hutt.). The presence of Acteon sulcatus gives evidence of the beds being of Castlecliffian age (Later Pliocene). (2) Palaeobotanical. The evidence of the plant-fossils as to the age of the beds supports that of the mollusca, for although seventeen of the thirty-one species have been described as not now existing, thirteen of them are referred to existing New Zealand genera, and in most cases there is a close relationship to New Zealand species. However, the fact that a considerable proportion of the species including some of extra New Zealand affinities are extinct, would indicate an age not younger than later Pliocene. Relationships with the early Tertiary are shown by the following lists in which certain plants of the Waipaoa Series are compared with species described by Ettingshausen as early Tertiary:— Waipaoa Series. Early Tertiary. Lomaria proceroides. Lomariopsis dunstanensis. Goniopteris pennigera. Aspidium tertiaro-zelandicum. Typha angustifolia. Zamites sp. Rhopalostylis sapida. Seaforthia zelandica. Beilschmiedia ovata. Daphnophyllum australe. Apocynophyllum novae zelandiae. A. Mackinlayi. Ceratopetalum pacificum. C. Gilesii. Comparisons may be made with the Cromer Forest Bed, England, usually considered as of late Pliocene age, and in which none of the 150 species are extinct, and with the Potosi flora of Bolivia described by Berry (Proc. U.S. Nat. Mus., vol. 54, p. 113, 1919) in which 54 of the 66 definitely-determined species are closely related to existing species. Berry determined this flora as of later Pliocene age. The proportion of extinct forms in the Waipaoa flora is similar to that of the Potosi flora and greater than that of the Cromer Forest flora. Climate: The climate indicated by the flora of the Waipaoa Series is warm temperate, possibly slightly warmer than that at present found in north New Zealand. The following genera show a distinctly warm facies:—Platycerium. Rhopolostylis, Litsaea, Beilschmiedia, Knightia, and Carmichaelia. The presence of Nothofagus does not necessarily indicate a cold climate as N. fusca occurs living in the north of New Zealand. Plant Formations: With one or two exceptions the flora of the Waipaoa Series represents a rain forest. In fact, with the exception of three species, all are such as would be expected in a forest. Three species, namely, Typha angustifolia, Pteridium esculentum, and Leptospermum pliocenicum, indicate a more open type of vegetation, such as a lake-shore which would provide the requisite conditions for the preservation of the fossil flora, and perhaps some scrub which might grow in its vicinity. Relationships: Of the thirty-one species herein described fourteen have been referred to Recent New Zealand species, thirteen

have been attributed to Recent New Zealand genera, two to extra-New Zealand genera (Platycerium, Ceratopetalum), and two to a genus (Apocynophyllum) of which the New Zealand species are all extinct. The flora is thus in the main related to the Recent flora of New Zealand, the only extraneous element being of Australian or Malayan affinity, for Ceratopetalum is an Australian genus, and Platycerium a common tropical genus occurring widely in Australia. The genera include such distinctive New Zealand plants as Rhopalostylis, Knightia, Carmichaelia, Melicytus, Nothopanax, Pennantia, Hebe and Coprosma. Relationships are shown with South America (Coriaria), Australia, Malaya and the Pacific Islands, but not to any extent different to those of the Recent flora. This late Pliocene flora is thus to be considered as of the same origin as the Recent New Zealand flora. Literature: The fossil flora of the Waipaoa Series is referred to by Hill (Trans. N.Z. Inst., vol. 20, p. 300, 1888) who states that he found forty-six species of plants in the Poverty Bay pumice-beds, and again by the same author (Trans. N.Z. Inst., vol. 21, p. 320, 1889) when describing the occurrence of moa feathers in the same beds as the leaves. Morgan, in Henderson and Ongley's report (N.Z. Geol. Surv. Bull., No. 21, p. 50, 1920) gives a brief account of the flora and concludes that it is of Pliocene age. Dryopteris novae zelandiae Oliver n. sp. (Fig. 1). Impression of portion of pinna. Allowing for the tip the pinna is narrow-deltoid, the upper segments being narrower and shorter than the lower. The lower segments are separated almost to the rhachis, the upper, about three-fourths the way. Segments rhomboid, blunt pointed, both margins crenulate. The largest segment is 10 mm. long by 4 mm. wide. Venation: The mid-rib of each segment enters near the lower edge and crosses the segment diagonally as an undulating line to the apex. From it are given off on either side 3 or 4 secondary nerves which fork once or twice, the veinlets running free to the margin, each marginal tooth always having a vein terminating in it. Sori: The position of the sori, which apparently were small, is on the lateral veins which may be produced beyond them or not. Each lateral vein has a sorus on its inner branch. There are 4 or 5 sori on each side of the mid-rib and about midway between it and the margin. This fern belongs to the typical section of Dryopteris. The Recent New Zealand species are not closely related as all have much narrower and distinctly serrated segments. Lomaria proceroides Oliver n. sp. (Fig. 24). Several incomplete pinnae. 20-23 mm. wide. Nerves free, a few fork at or near their origin; terminating in the ends of the serrations at the margin of the pinna; 1 to 1.2 mm. apart. Pinna lanceolate, narrowing gradually towards the base, then turning in more abruptly. The serrations are not so marked at the basal end.

Fig. 1.—Dryopteris novae zelandiae Oliver n. sp. Fig. 2.—Goniopteris pennigera (Forst.) Spreng. Fig. 3.—Platycerium morgani Oliver n. sp. Fig. 4.—Nothofagus fusca (Hook. f.) Oerst. Fig. 5.—Litsaea calicaris (A. Cunn.) Benth & Hook. f. Fig. 6.—Beilschmiedia ovata Oliver n. sp. Fig. 7.—Clematis obovata Oliver n. sp. a, upper surface; b, lower surface. Fig. 8.—Apocynophyllum novae zelandiae Oliver n. sp. Fig. 9.—Plagianthus antiquus Oliver n. sp. Fig. 10.—Rubus australis Forst.

This species agrees with Lomaria procera (Forst.) Spreng. in everything save the shape of the base of the pinnae. In L. procera the base is truncated or even dilated, although in very young pinnae it is somewhat narrowed. Pinnae of the same size as the fossil specimen are, however, never narrowed towards the base. Pteridium esculentum (Forst.) Ckn. (Fig. 25). Carbonized impressions of both sides of a portion of a frond. Pinnule 17 mm. long, lowest segment 4 mm., breadth of segment 1 mm. The segments are linear, straight or slightly curved, obtuse. Pinnule with about 8 segments on either side. The impression of the under surface shows a central rib and two lateral lines apparently corresponding to the indusia. The veins are well seen in several places. They are free, and once forked near the base. Pinnules sessile, alternate, lanceolate, broadest at the base, the lowest segments longest and springing from near the rhachis. There does not appear to be a long terminal segment. Upper basal segment shorter than the lower. The specimens agree with Pteridium esculentum in most of the characters. It differs in the rather narrower segments though P. esculentum when dried sometimes has segments 1 mm. wide. No long terminal segment is shown in the fossil, which, however, may be imperfect. Goniopteris pennigera (Forst.) J. Smith (Fig. 2). Fragments of pinnae consisting of several more or less imperfect pinnules and a portion of the rhachis. Segments obtuse or subacute slightly curved towards tip of leaf, margin entire. Rhachis as a double line, indicating the upper surface of the leaf. Segment 12 mm. long. Nerves free, parallel, 8-12 on each side of midrib, sub-opposite to alternate. The lowest pair united with those on adjacent segments. The veins are close being slightly less than 1 mm. apart. In another specimen the segments measured 14 mm. long. Another specimen shows sori on the veins, mostly midway between midrib and margin, though some are nearer the midrib, in one instance there is a sorus on the lowest vein. This specimen matches almost exactly Recent specimens except in the nerves being closer together. Platycerium morgani Oliver n. sp. (Fig. 3). The specimen consists of the impression of a fragment of a frond showing distinctly the major and minor veins. The sides diverge at an angle of 25°. Along a length of 3 cm. the leaf increases from 4 cm. to 5 cm. in width. The venation consists of anastomosing veins of two orders. The deeper impressions make lacunae 5 to 10 mm. in width and none are complete in the length of frond (5 cm.) represented. Over the whole of the interspaces is a finer network giving elongate more or less hexagonal meshes 1 to 2 mm. wide and 4-6 mm. long.

Another and larger impression shows a precisely similar type of venation. The veins are very clear, and in several a short vein extends and ends freely in a slight dilatation. I place this species with little doubt in the genus Platycerium as the type of venation agrees very closely with P. alcicorne. It is dedicated to the late Mr. P. G. Morgan, Director of the New Zealand Geological Survey, who kindly placed the present collection in my hands for description. Typha angustifolia Linné. Impression of leaf, both surfaces. The impression is of a linear leaf 12 mm. wide with longitudinal striation. The fragment is 7½ cm. long. There is no indication of ribs, and it is therefore referred to Typha. The striations are fine and of a similar type to those of T. angustifolia. Dried leaves of Typha angustifolia show about 12 ribs but if the impression was of an unshrunken leaf the ribs would not show Rhopalostylis sapida (Forst.) Wendl. & Drude (Fig. 26). Impressions of leaves of this palm are quite common in the pumice silts. The best specimen consists of the midrib of a leaf, 25 cm. in length, showing 4 pinnae arising on one side. A leaflet 10 cm. from its origin is 3 cm. in diameter. The stem is about 1 cm. in width. It shows no definite markings. The pinna converges towards the base where, opposite its inner junction with the midrib, it is 2 cm. in width. The pinnae show close parallel veins, with a prominent midrib, or on the upper surface, depression, and usually one prominent rib on each side of the midrib near the margin. The pinnae are placed at the same distance apart as in the Recent species, but they appear to be joined by a wider base. Nothofagus fusca (Hook.) Oerst. (Fig. 4). Leaf impression, upper surface, basal portion incomplete. Length along midrib 29 mm. Distance midrib to right side 12 mm, to left side of 9 mm. Leaf ovate narrowing in the distal half to a 3-dentate lobe, coarsely and irregularly serrate. The secondary nerves are fairly regularly arranged, diverging at an even angle on either side of the midrib, running approximately parallel, and ending in the apexes of the teeth. There are 4 on the right side and 5 on the left. The lower one on the right side has two strong branches on its under side, the upper one is forked near the base, the middle one on the left side is forked about half-way to the margin. Under the microscope the whole lamina appears finely reticulated, the larger nervules crossing between the secondary nerves at right angles to them. The leaf corresponds in all its characters with that of Nothofagus fusca, though the angle at which the secondaries spring is not quite so acute.

Paratrophis cuneata Oliver n. sp. (Fig. 27). A hardened piece of silt contains portion of a leaf impression. The complete leaf would be elliptical, tapering gradually from the centre towards base and apex, but evidently broadest nearer the apex. The margin is coarsely sinuous, not bounded by a conspicuous nerve. Length of lower portion 85 mm., breadth 40 mm. Estimàted length of entire leaf 14 cm. The midrib is strong and of considerable width, being 2 mm. wide at the base. The secondary nerves are weak and irregularly spaced, and branch frequently, the main portion arching forward to join the nerve above. The whole surface of the leaf is finely and distinctly reticulated into 4 to 6-sided polygons, in many of which the nerve endings appear to terminate freely. The secondaries spring at an angle of about 70° from the midrib. In its sinuated margin and type of venation the leaf agrees with Paratrophis. It is much larger than P. opaca, and a point of difference is the method in which the secondaries arise from the midrib. In the fossil species the base of each curves away from the midrib, in P. opaca it meets it abruptly. The fossil also narrows more decidedly from the broader distal end towards the base. Its closest ally, however, seems undoubtedly to be P. opaca. Litsaea calicaris (A. Cunn.) Benth. & Hook. f. (Fig. 5). Impression of both sides of leaf. Leaf regularly elliptical, tapering rather abruptly at both ends, the sides unequal and the base also slightly unequal, quite entire. Apex blunt. The leaf was apparently smooth and thin. The midrib and secondary nerves are well marked, all the others are easily seen with a hand glass. Of the secondaries about 6 are well marked on either side diverging at an angle of about 60°, nearly opposite. Tertiary nerves with coarse reticulation, the lacunae entirely filled in with minute 3-5 sided cells. The secondary nerves arch round and join the ones in advance well in from the margin. Length of petrole 9 mm., lamina 60, breadth left of midrib 17, right of midrib 15 mm. The leaf appears to agree in every particular with that of Litsaea calicaris. Beilschmiedia ovata Oliver n. sp. (Fig. 6). Several impressions of leaves giving details down to the minute surface reticulation. Lamina, regularly elliptical, obtusely pointed at apex and base. Some specimens show the apex truncated and split so as to appear retuse. Margin entire. Lamina 74 × 37 mm., 61 × 37 (truncated apex). Midrib wide and well impressed, striated. Secondaries: 4 or 5 prominent ones on either side, arising at an angle of about 40° from the midrib and curving slightly towards the apex. They approach but do not quite reach the margin, curving round almost parallel to it. A few weak tertiaries are present springing from the

Fig. 11.—Knightia fossilis Oliver n. sp. Fig. 12.—Coriaria arborea Lindsay. Fig. 13.—Ceratopetalum pacificum Oliver n. sp. Fig. 14.—Pittosporum oblongum Oliver n. sp. Fig. 15.—Apocynophyllum inaequilaterale Oliver n. sp. Fig. 16.—Melihcope elliptica Oliver n. sp. Fig. 17.—Quintinia waipaoaensis Oliver n. sp. Fig. 18.—Nothopanax reticulatum Oliver n. sp.

distal outer sides of the secondaries. Weaker veins cross fairly regularly between the larger veins, and there is a fine reticulation over the whole surface. In their shape and type of venation these leaves correspond very closely with the Recent Beilschmiedia tarairi. The chief differences are: the lower secondaries springing at an acute angle instead of nearly a right angle as in B, tarairi; obtuse apex not showing the mucro of the young leaf of B. tarairi; lees prominent and fewer principal nerves. Knightia fossilis Oliver n. sp. (Fig. 11). (a) Impression of a leaf, base imperfect. Leaf oblong-elliptical, slightly curved, distally tapering rather suddenly to the apex which is not quite complete; margin wavy and irregularly serrate. Length 115 mm., breadth 30 mm. (b) Portion of leaf, both base and apex missing, margin wavy and irregularly serrate, the serrations much more pronounced than in specimen (a). Breadth 30 mm. These two specimens are associated together on account of the type of surface and venation which correspond exactly. They only differ in the margin. The midrib is well marked; all the other veins are weak. The secondaries spring at an angle of 45-50° from the midrib and at irregular distances. They fork into two main branches rather nearer the margin than the midrib. The whole surface is finely and clearly reticulated, the nerves terminating freely in the meshes. The type of venation corresponds precisely with that of Knightia excelsa R. Br., and hence the specimens are referred to Knightia. The difference from K. excelsa lies in the character of the margin which is irregular in the fossils, but fairly regularly dentate in the Recent species. Coriaria arborea Lindsay (Fig. 12). Impression of large leaf. Leaf obovoid, broadest about ¼ from base, sides beyond the broadest part rather straight, gradually tapering and then turning in towards apex which is produced into a blunt point. Margin entire. Rounded at the base. Lamina 105 × 50 mm. Seven longitudinal ribs present. Midrib strong. On either side is a strong rib, confluent with the midrib at the base for about 1 cm., then arching out and again converging at the apex. On either side of these are two longitudinal ribs, considerably weaker. The inner ones can be traced nearly to the apex, but the marginal ones can only be followed 3 or 4 cm. The secondaries arise from the longitudinal nerves, and join up to the longitudinal nerve or margin outside them. The leaf is faintly reticulated with rather large meshes. The leaf agrees entirely with the Recent C. arborea though it is very large for that species. There are also present in the collection two imperfect impressions of smaller leaves.

Clematis obovata Oliver n. sp. (Fig. 7). Leaf impression of both surfaces, complete except for portion of one side and the petiole. Lamina—length 50 mm., breadth 38 mm. The midrib not in the centre, the distances to the margins being 20 and 18 mm. The shape is broadly ovate, truncate at the base, broad for more than half the length, then tapering to the apex, which is not quite complete, but appears to have been produced into a short point. The nerves are few, prominent, and wide. Two strong secondaries arise from near the base of the midrib, diverge at an angle of about 40°, and extend to nearly three-fourths the distance to the apex. At more than half-way up two other strong ribs, nearly opposite, diverge at a wider angle and arch round towards the margin about three-fourths the way to the apex. Near the apex is a third pair of strong nerves nearly opposite. The secondaries branch with the strongest branches on the outside, bnt branches also cross from secondary to secondary, and from secondary to midrib at an angle of 90 to 100° from the midrib. There are indications of a fine reticulation over the whole surface. The outer branches of the secondaries arch round towards the apex. This leaf does not resemble closely any plant at present living in New Zealand. In its shape it recalls Clematis indivisa Willd., but the venation is different. The venation is, however, not unlike that of C. hexapetala (Forst.). Hence rather than refer it to any genus not found in New Zealand, I am describing it as an extinct species of Clematis. Two impressions of achenes of Clematis are in the collection and may be described here, although, of course, they are not attached to any leaves, and therefore the proof that they belong to the same species as the leaf above described is lacking. In one the achene is 6 mm. in length, with a style 40 mm. long. The achene is elliptical, more convex on one side than the other, and is shortly stalked or at least produced at the base. It differs from the achene of C. indivisa in being slightly larger and in possessing a short stalk. Apocynophyllum novae zelandiae Oliver n. sp. (Fig. 8). Two well preserved impressions of leaves, in both cases with the upper portion of the leaf wanting. Lamina lanceolate, narrowing gradually to the base, unequal sided, in the middle portion the sides nearly parallel. Margin entire. Midrib broad and shallow at the base narrowing above. Secondaries all fine, numerous, and with many short fine nerves arising from the midrib between them. The principal secondaries take a slightly sinuous course to near the margin, then turn parallel to it to join up with the next distal secondary. Ultimate reticulation of small meshes faintly marked. Width of lamina 23 mm.; estimated length about 90 mm. I am unable to refer this species to any Recent genus of New Zealand plants. In the type of venation, namely, the numerous secondaries, connecting near the margin, it seems to approach the Tertiary Apocynophyllum, and hence is provisionally referred to that genus. Of the published figures to which I have had access, it resembles most A. Mackinlayi of Vegetable Creek, New South Wales.

Fig. 19.—Pseudopanax crassifolium (A. Cunn) C. Koch. Fig. 20.—Parsonsia obtusa Oliver n. sp. Fig. 21.—Pennantia corymbosa Forst. Fig. 22.—Hebe salicifolia (Forst.) Pennell. Fig. 23.—Coprosma vulcanica Oliver n. sp.

Apocynophyllum inaequilaterale Oliver n. sp. (Fig. 15). Distal portion of impression of a leaf. Lamina broad, narrowing gradually to a broad blunt apex. Apex conspicuously unequal sided, the angle of midrib and margin being acute on one side, but about a right angle on the other. Measurements of the fragment 50 mm. long, 42 mm. broad. The entire leaf is estimated to be 100 mm. long, and 50 mm. broad. All the veins are faintly marked, indicating that they were not prominent in the living leaf. The midrib is broad. The secondaries are all narrow, and many fine ones are interspersed among the more prominent ones. These latter arch round on nearing the margin to join up with those in advance. The ultimate reticulation consists of small meshes faintly marked. The present leaf impression resembles Griselinia in its general outline and unequal-sidedness. The type of venation is, however, altogether different and more nearly approaches that of Apocynophyllum. Plagianthus antiquus Oliver n. sp. (Fig. 9). Imperfect impression of a leaf. The leaf appears to have been broken, for the midrib is bent at about 100° near the base. Lamina roughly oblong, coarsely and irregularly serrate, the serrations at the base becoming progressively smaller towards the petiole. The teeth are rather bluntly pointed. Length 60 mm., but apex missing. The leaf is estimated to have been about 80 mm. in length. Midrib broad and shallowly impressed. Secondary veins weak, arising from the stem at an angle of about 60°, irregularly spaced, and quite irregular in their branching. Some branch near the midrib, others distant from it, while they take irregular courses. Each terminates in a tooth at the margin, and the principal branches may also end in teeth. A fine reticulation covers the whole surface of the leaf. In the general shape of the leaf, and the type of venation, the fossil agrees with Plagiamthus reqius (Poit.), but it differs in being more oblong, and in the secondary veins being less regular in their courses and in arising from the midrib at a much wider angle. Leptospermum pliocenicum Oliver n. sp. Branching specimen containing several capsules. The branches are slender, with a few longitudinal ribs. No leaves are preserved. The capsules are small, a little broader than high with convex sides and convex crown. At the outer angle are slightly projecting points evidently representing the calyx segments. There is, in most cases, a short persistent style. The pedicel is slightly longer than the height of the capsule. The capsules arise singly at short intervals on the twigs. Length of capsule 2 mm., of pedicel 2.5 mm. The capsules agree in size, shape and length of pedicels with young fruit of the Recent L. ericoides A. Rich. The ridging of the young stems is also similar in the two forms. The arrangement of

the capsules in the stem is, however, different. In L. ericoides they occur in groups either on the main branchlets or on short lateral branchlets; in the fossil they occur singly. Carmichaelia australis R. Br. (Fig. 28). Two large impressions of branches. The largest specimen is 25 cm. long, and consists of a branch bearing several short branchlets, each ramifying in the way characteristic of the genus. The impressions are bad, but here and there the striations of the branches are evident. The branches are 3-5 mm. in width. So far as the details are preserved, that is, in habit, size and striation, the specimens resemble the Recent C. australis, and so may provisionally be referred to that species. Rubus australis Forst. (Fig. 10). Impression of leaf, complete. Lamina lanceolate, broadest near the base, thence gradually tapering to the apex, inequilateral, and unequal sided at the base, the wide side diverging at a narrower angle and higher up than the narrow side. Coarsely and fairly regularly serrate. Length of lamina 65 mm., width 44 mm., 5 mm. from midrib to one margin, and 9 mm. to the other. Midrib well defined. Secondaries wide apart, branch off at an angle of about 40° and terminate in the teeth. The whole surface finely reticulated. Two other smaller specimens agree essentially with the above described leaf. In its general characters these leaves agree with the Recent Rubus australis. In the secondary nerves leading direct to the serrations, however, it approaches R. schmidelioides A. Cunn., so that perhaps in the Pliocene as now hybridization was going on in the New Zealand species of Rubus. Pittosporum oblongum Oliver n. sp. (Fig. 14). Impression of both surfaces of leaf, that of upper not good, that of under surface one side nearly perfect, the other (left) partly missing. Apex missing. Leaf broadly elliptical rounded off rather abruptly at apex and base. Margin entire. Length of lamina 57, width from midrib to right side 16 mm. Midrib and secondaries prominent, remainder obscure. The secondaries arise at an angle of 70° to 75°, arch little or not at all for most of their course, and turn rather abruptly near the ends towards the apex of the leaf. There are 11 principal secondaries on the right side, and between these here and there are shorter ones. Secondaries alternate or opposite. Tertiaries weak, but show fairly large reticulations. The leaf has the characters of a Pittosporum and agrees with P. colensoi Hook. f., except that the basal angle is broader and the secondary veins spring from the midrib at a wider angle.

Ceratopetalum pacificum Oliver n. sp. (Fig. 13). Impression of the greater portion of one side of a leaf and the base. Lamina narrowly obovate, base rounded. Margin irregularly serrate, generally a larger tooth alternating with a smaller one. Length estimated at 4.5 cm., breadth 1.8 cm. Midrib broad and well marked. Secondaries numerous but weak, breaking up before reaching the margin, and their branches terminating in the serrations. The ultimate reticulation shows large meshes. I cannot refer this species to any New Zealand genus. It is, however, not unlike the Recent and Tertiary Ceratopetalum of Australia, and may be provisionally placed in that genus. Melicope elliptica Oliver n. sp. (Fig. 16). A well-preserved impression of a leaf with the base imperfect. Lamina broadly elliptic, unequal-sided, the broadest part of one side being just in advance of the centre of the leaf, while that on the other side is behind the centre. Apex rather acutely pointed. Margin slightly irregular but entire. Width of leaf 23 mm.; estimated length 50 mm. Midrib well marked by a dark shallow depression. Margin marked by a continuous dark line. Eight to ten secondaries on each side arising from the midrib at an angle of about 55°. They usually break up by dividing into two a little nearer the margin than the midrib. The ultimate reticulation is faintly marked with small meshes. Within each mesh there is seen under a strong lens, a dark spot. These spots do not seem to be connected with a nerve ending, and hence probably represent oil glands. The leaf agrees with the Recent Melicope ternata Forst. in the type of venation, marginal line and presence of oil glands. It only differs in the shape of the lamina, being more acutely pointed than in the Recent species which is broadest near the distal end giving an abruptly pointed apex, while it narrows gradually to the base. Quintinia waipaoaensis Oliver n. sp. (Fig. 17). Leaf impression imperfect at the tip. Lamina oblong, obtuse at apex, tapering rather abruptly at base. Margin with large regularly spaced serrations. Petiole short. Length of lamina 57 mm., breadth 20 mm., petiole 5 mm. Midrib well marked as also are 10-12 secondary nerves on either side. These diverge at an angle of about 65°, are straight for threefourths of their way to the margin, when each breaks up sending a branch above and below to serrations. The basal portion of the secondary nerve usually points to a position between two of the marginal teeth. The leaf surface is faintly reticulated with small meshes. The leaf is referred to Quintinia because its type of venation agrees with the members of that genus. It differs from the Recent species in the shape and outline of the lamina.

Fig. 24.—Lomaria proceroides Oliver n. sp. Fig. 25.—Pteridium esculentum (Forst.) Ckn. Fig. 26.—Rhopalostylus sapida (Forst.) Wendl and Drude

Fig. 27.—Paratrophis cuneata Oliver n. sp. Fig. 28.—Carmichaelia australis R. Br.

Melicytus ramiflorus Forst. Impression of portion of leaf. Lamina ellipitical, tapering in the distal half, regularly dentate, the serrations having a longer curved lower side and a shorter upper side. Width from midrib to margin 2 cm. The estimated size of the lamina is 10 × 4 cm. Midrib strong. Secondaries strong, arising from the midrib at various angles 50° to 65° each curves distally as a prominent vein and, running nearly parallel to the margin, joins up with the vein in front. From this marginal festoon, nerves run towards the margin curving forwards to the teeth. A fine meshed reticulation is evident in places. In shape, margin and venation the impression agrees with the Recent Melicytus ramiftorus. Nothopanax reticulatum Oliver n. sp. (Fig. 18). Impressions of two leaves both imperfect as regards apex and base. Lamina broadly lanceolate, gradually narrowing towards the tip, but more abruptly towards the base; margin entire. Width 40 mm., length estimated to be 12-14 cm. The midrib is broad and well impressed. The secondaries are numerous and arise at fairly regular intervals and at an angle averaging 50° from the midrib. Near the base the angle is wider, except in the case of the lowest two or three veins on either side. On nearing the margin the secondaries arch round and join those in front. The ultimate reticulation is well marked with fairly large polygonal meshes. The present leaves resemble those of the Recent Nothopanax Edgerleyi, but differ in the following points: the base is more abruptly narrowed, the reticulation is finer and does not show the regular loops of the Recent species. Pseudopanax crassifolium (A. Cunn.) C. Koch. (Fig. 19). Impression of upper surface of leaf, imperfect at each end. Leaf lanceolate, gradually tapering towards the base. The margin coarsely and sharply serrated. The lower teeth 20-22 mm. distant, the upper 8-12 mm. Between the larger teeth are, in most cases, one to three much smaller ones; absent in the lowest portion, 3 in the longer intervals and 1 in the upper portion of the leaf. The teeth are sharp pointed. Total length of preserved portion of leaf 120 mm., width 23 mm. The complete leaf would be not less than 150 mm. in length. The midrib is well marked, just over 1 mm. in diameter at the base and tapering towards the tip. Secondary nerves very faintly marked. They spring from the midribs at an angle of 25° to 30°. The ends are directed towards the notches but just short of them they fork, the lower branch going towards the apex of the tooth, the upper continuing parallel with and near the leaf margin. The leaf agrees in all its characters with Pseudopanax crassifolium.

Parsonsia obtusa Oliver n. sp. (Fig. 20). Several leaf impresions, some with the carbonized remains of the leaf still attached. Lamina broadly or narrowly eliptic, blunt pointed, base broadly cuneate. Margin entire. Lamina 37 × 21 mm.; 30 × 14 mm. The midrib and secondary veins are deeply impressed, and therefore must have been prominent in the leaf. There are a few prominent branches, otherwise the veinlets are weak. There are five secondaries on either side of the midrib, branching at an angle of 40° 45′ and curving forward near the margin, but not quite reaching it. Tertiaries given off mainly from the outer side of the secondaries similarly approach the margin. A fine reticulation covers the entire space between the prominent veins. The position of these leaves is doubtful. They are referred to Parsonsia because of the characters of the secondary and tertiary veins, but they differ in shape, non-mucronate apex, and in the presence of a fine surface reticulation. Pennantia corymbosa Forst. (Fig. 21). Impression of both surfaces of a leaf. Leaf oblong, base rounded, apex truncate, sides nearly parallel, with six rounded blunt teeth on the distal half, of which one appears to be mucronate. Length 26 mm., breadth 18 mm. The midrib and secondary ribs are well marked but not deep. There are four principal secondary ribs on either side of the midrib arising at an angle of 50°, nearly opposite, the upper two on each side terminating in the apexes of the teeth. The remaining nerves are obscure, but there are indications of a large meshed reticulation. The leaf agrees very well with intermediate leaves of Pennantia corymbosa, except that the base is more rounded than is usual in the Recent form. Hebe salicifolia (Forst.) Pennell (Fig. 22). Impression of both surfaces of a portion of a leaf. Leaf broadly lanceolate, entire. The length of the impression is 7 cm. It represents about half a leaf. Distance from midrib to margin 13 mm. The midrib is well marked, fairly broad and not high. The secondary ribs are very faintly marked. One near the base branches at a small angle and runs parallel to the margin. Others further up branch at wider angles, arching round until they are nearly parallel to the margin. The leaf impression agrees in all points with the Recent species, Hebe salicifolia, and may accordingly be entered as such.

Coprosma vulcanica Oliver n. sp. (Fig. 23). Portion of a leaf, about ¾, with impression of both upper and lower surfaces. The part preserved measures 93 × 43 mm. In the broadest portion the distance from margin to midrib is 26 mm. The complete lamina is estimated to be 130 × 52 mm. The margin is entire. Only one side of the apex is preserved. It is obtuse and bends in towards the termination of the midrib. If this is not due to an injury or loss in preservation it indicates an emarginate apex. The general form of the leaf is elliptic, the broadest portion being distal to the centre. Midrib strong, prominent. Secondaries: 4 or 5 strong ones on each side, with smaller ones between. They arise from the midrib at a wide angle, curve regularly, though somewhat fastigiately, and are united near the margin by a branch of the secondary distal to each. The tertiaries are fine and indicate a fairly large meshed reticulation. The discoloured margin indicates a marginal vein. Domatia are indicated at the bases of most of the secondary veins. The leaf agrees essentially with the large leaved Recent species of Coprosma. It perhaps comes closest to C. australis (A. Rich.) (= C. grandifolia Hook. f.) especially in the venation. It differs in the following points: the angle of the secondaries and midrib is wider, the shape is different being less cuneate at base and apex, the apex is apparently emarginate.