Other formats

    TEI XML file   ePub eBook file  

Connect

    mail icontwitter iconBlogspot iconrss icon

Tuatara: Volume 13, Issue 1, April 1965

Some Discontinuous Distributions of Plants Within New Zealand and their Ecological Significance — Part II: Disjunctions between otago-southland and nelson-marlborough and related distribution patterns

page 9

Some Discontinuous Distributions of Plants Within New Zealand and their Ecological Significance
Part II: Disjunctions between otago-southland and nelson-marlborough and related distribution patterns

An Earlier Account by Wardle (1963) presented an hypothesis of Pleistocene extinction of plants in the central South Island, and the likelihood of the presence of Pleistocene refugia in the extremities of the Island, among other places. The present writer, independently, had come to similar conclusions from study of distributions of species in the South Island. This paper was written in the belief that it could add considerable detail to the chain of evidence presented by Wardle and elucidate some of the problems of distribution.

Two groups of plants are disjunct** over wide geographical distances in New Zealand. The fact that numbers of species are disjunct together makes it most probable that the causes of separation are common for all of these species. The first disjunction is between Auckland and Nelson and will be considered later.

The second striking series of disjunct species consists of plants with a conspicuous gap in their known distribution in the central South Island. Wardle listed a few of these and drew on the presence of a relatively large number of endemic plants in each of Otago-Southland and Nelson-Marlborough, with correspondingly few endemics in the central South Island, as support for his hypothesis of Pleistocene extinction in the latter area. An implication of the hypothesis is that Nelson, Marlborough, Banks Peninsula, coastal Southland. Stewart Island and offshore islands were refugia where large numbers of species including some forest plants survived the glacial maxima of the Pleistocene.

* E. L. Hellaby Research Fellow.

** For plant geographical terminology see Cain (1944).

Readers are referred to A. H. McLintock 1960: A Descriptive Atlas of New Zealand for place names.

page 10

In examining, the present distribution patterns of species disjunct between the northern and southern parts of the South Island, and of endemics in both these areas, the present writer came to the conclusion that a simple statement of Pleistocene extinction in the central South Island did not explain all the facts of distribution. The detailed knowledge of distribution of plants and vegetation, together with information from non-botanical fields, allows elaboration of the hypothesis as formulated by Wardle. In this article evidence is brought forward in support of a modified hypothesis.

It is proposed here to deal with species fully disjunct between Nelson-Marlborough (and/or North Island) and Otago-Southland (and/or Stewart Island); with closely related (vicarious) species, whose distributions do not meet in the central South Island; with species which are abundant to north and south of the South Island and less common between; with the patterns of distribution of species endemic to the north and south of the South Island; with some distributions which indicate that the populations of these plants have undergone recent contraction: and with some disjunct animal distributions.

Full Disjuncts

The species fully disjunct between the northern South Island (and/or North Island) and southern South Island (and/or Stewart Island) in a few cases extend to the Auckland or Campbell Islands. The gap in distribution varies in width and is obvious from distribution maps (Figs. 1, 2, 3 and 4). The extremes in separation are those for populations of Juncus procerus (Edgar, 1964), 1650 miles, from Auckland to Southland) and for Drosera pygmaea, from the Volcanic Plateau to Bluff Hill (more than 600 miles). In the South Island proper there are patterns held in common by numbers of species. Many of the species are confined to the Gouland Downs-Tasman Mountains-Mt Arthur-Paparoa Range areas in the north and to Stewart Island-Fiordland-Longwood Range-Blue Mountains-Maungatua in the south. The ranges of other species are further to the south in Nelson-Marlborough and to the north in Otago-Southland with limits in various places, a few extending as far as Canterbury-Westland. One species (Pimelea aridula)‡ is centred on Central Otago and central Marlborough, and lives on dry, rocky hillsides. It extends also to Hawkes Bay. The other full disjuncts include Mitrasacme novae-zelandiae*, Petriella thomsoni*, Hemiphues suffocata*, Pseudalepyrum pallidum*, Tetrachondra hamiltonii*, Gentiana saxosa†. Drosera pygmaea*, Drapetes laxus†, Senecio reinoldii‡, Celmisia petriei‡, C. traversii†, Ourisia macrocarpa var. macrocarpa† (found only once in northwest Nelson). O. modesta* (a rare plant). Parahebe catarractae‡, Myosotis tenericaulis‡, page 11
Fig. 1

Fig. 1

page 12 Pimelea gnidea‡, Elytranthe colensoi**, Melicytus macrophyllus‡ (found in one locality near Dunedin), Juncus procerus‡, Nothofagus fusca‡, N. solandri var. cliffortioides‡ (hereafter called N. s. cliffortioides), Lyperanthus antarcticus*, Rubus australis‡, Hebe salicornioides‡, Atriplex novae-zelandiae‡, Lepidium tenuicaule‡. Those species marked * are confined to boggy habitats. The implication that bogs are also limited to the same areas is generally true for the cushion bog type which is the habitat for most of these plants. Extensive cushion bogs in the South Island are found only north of about the Taramakau-Waimakariri Rivers and south of about the Haast Pass. Extensive raised sphagnum bogs are also limited in much the same way although a few occur in central Westland where conditions for their formation are very favourable. An hypothesis to account for the bog disjunction will also account for disjunction of the individual plants. The other strictly habitat limited plant (marked **) is an Elytranthe, usually parasitic on Nothofagus menziesii which itself has wide gaps in distribution but is found in a few places in the central South Island. The other species (marked ‡ and †) are not apparently habitat limited at present and there seems to be no disability which prevents their potential migration into the central South Island. Species marked ° are normally alpine and those
FIG 2

FIG 2

page 13
FIG 3.

FIG 3.

page 14 marked ‡ are lowland to subalpine plants of various habitats. It is possible that some of the species listed may be discovered in the central South Island in future but this would not affect the general pattern of disjunction. The mapped distributions (as for all other figures) are from herbarium specimens and the writer's own observations. The general impression when considering these distributions is that there have been centres of dispersal in Nelson-Marlborough and Stewart Island-Fiordland-Coastal Southland Within the limitations of the tolerance ranges of the species and limitations of space and pathways for migration, there tend to be some eccentric dispersal patterns radiating from north west Nelson and from the far south. In some cases minor evolution has brought about slight morphologic differences in populations of some of these species but separation has probably not been for long.

Vicarious Species

There is a group of several pairs of morphologically similar species in which one of the pair is distributed in Otago-Southland and the other in Nelson-Marlborough or North Island. There is, thus, a gap in the central South Island in which the pairs of species do not meet. These pairs seem to be closely related or vicarious species. It may be speculated that they have experienced evolutionary divergence comparatively recently (probably following separation earlier than the disjunct species were separated). They are enumerated below.
Nelson-Marlborough and/or north islandOtago-Southland and/or stewart island
Celmisia dalliiC. holosericea
C. rupestrisC. hectori
Dracophyllum densumD. politum
Euphrasia cheesemaniiE. australis
Gentiana sp. aff. lineataG. lineata
G. sp. aff. montanaG. montana
Pimelea sericeo-villosaP. pulvinaris
Raoulia rubraR. buchananii
Senecio adamsiiS. revolutus
Chionochloa junceaCh. acicularis
The distribution patterns of these vicarious species are similar to those of the fully disjunct species (Fig. 5).

‘Partially’ Disjunct Species

The distribution patterns of a further group of species show that they are generally abundant in the north or south of the page 15
FIG.4

FIG.4

page 16 South Island, but that they are rare, or less abundant in the central South Island (Fig. 6). Within this gap there may be varying degrees of disjunction. Nothofagus menziesii occurs in the north west and south west and south of the South Island but is found in at least three scattered localities — near Burkes Pass, at Lake Heron and in the Karangarua Gorge, between the main areas of occurrence. The Armstrong herbarium of the Canterbury Museum contains two specimens of each of Nothofagus menziesii and N. fusca labelled as having been collected in the Upper Rangitata and Upper Ashburton Rivers in 1869 but these distributions require verification. Adenochilus gracilis nearly always accompanies Nothofagus fusca. There is one known occurrence of this orchid in the Whitcombe River, a tributary of the Hokitika River. Festuca matthewsii is an important component of alpine and subalpine grasslands in Marlborough (Wraight, 1963) and about and south of the Waitaki River (Connor, 1961), but although occasionally present is very inconspicuous in grasslands between. Geum leiospermum, Liparophyllum gunnii, Schizaea fistulosa var. australis and Oreostylidium subulatum have similar patterns of distribution, the
FIG. 5

FIG. 5

page 17
FIG.6

FIG.6

page 18 three latter being found occasionally in pakihi bogs in central and south Westland or occasionally in boggy patches east of the main divide and much more commonly in boggy ground or on wet, acid soils in the north and south. This kind of pattern may possibly be ascribed to the same causes as for full disjunction but recent contractions of area might also be involved in some cases.

Endemics of Nelson-Marlborough and Otago-Southland

There are sufficient endemic vascular plant species in the South Island confined within the areas north of a line from about Greymouth to Motunau and the south of a line from about Bruce Bay to the Waitaki River mouth to enable the northern and southern floras to be regarded as distinct from the rest of the South Island flora (see Wardle, 1963a; approximately 110 endemic species in Nelson-Marlborough and 90 in Otago-Southland excluding the Foveaux Strait species). The central South Island proper, however, is characterised by relative poverty in endemic species, with only about 14 such species being confined approximately within the boundaries described above, four of them on Banks Peninsula. Some of these central South Island endemics are plants of rupestral habitats and others favour periglacial habitats up to about 9000 feet in the high Central Alps (Myosotis explanata, Ranunculus godleyanus, R. grahamii). Rupestral, scree and high altitude species of somewhat wider distribution will be discussed later. Individual distribution patterns of the endemics which have been grouped here into Nelson-Marlborough and Otago-Southland floras resolve the endemics into minor groups centred on focal areas within both northern and southern South Island. Thus there is a group centred on northwest Nelson and a smaller group in eastern Nelson, and Marlborough including the Kaikoura ranges. Similarly in the south a group centres on Fiordland and another on Central Otago. These may be designated ‘western’ and ‘eastern’ facies respectively, but individual species transgress any boundaries which may be drawn. The fully disjunct species and vicarious pairs of species have similar distributions. It is not intended to enumerate the distributions of these endemic species in detail. The generalised patterns of distribution will be described with the aid of a map (Fig. 7). From Fig. 7 it is seen that northern patterns may be grouped into at least nine ‘types’. These are designated only for convenience in description. Individual distribution patterns show great complexity, with each species tending to behave independently. Each ‘type’ is named here according to the absolute southern limit of the species within it and is characterised by the presence of at least three species.

page 19
1.Tasman Mountains: Poranthera alpina, Hebe ochracea, Pseudowintera traversii, Astelia spp. (several undescribed), Pittosporum dallii.
2.Buller Valley-Clarence River headwaters: Hebe coarctata, Anisotome deltoidea, Nototlaspi australe.
3.Paparoa Range: Celmisia dallii, C. lateralis, C. parva, Euphrasia townsonii, Dracophyllum townsonii, Senecio hectori.
4.Taramakau-Waimakariri Valleys: Chionochloa australis, Parahebe cheesemanii, Hebe decumbens.
5.South Westland: Rubus parvus. Euphrasia disperma, E. cockayniana, Hebe gracillima.
6.Wairau Valley-Eastern Nelson: Celmisia rutlandii, C. cordatifolia, Hebe rigidula, Scutellaria novae-zelandiae, Pimelea suteri.
7.Kaikoura Ranges: Ewartia sinclairii, Wahlenbergia matthewsii, Celmisia cockayniana.
8.Mount Terako-Lower Waiau Valley: Olearia coriacea, Haastia pulvinaris, Helichrysum coralloides. Pachystegia insignis, Hebe hulkeana. Senecio monroi.
FIG. 7 Limits of distribution of endemic species from northern and southern South Island, (see text)

FIG. 7 Limits of distribution of endemic species from northern and southern South Island, (see text)

page 20
9.Middle Waiau Valley: Helichrysum microphyllum, Pimelea concinna, Schizeilema roughii. The southern patterns fall into at least nine ‘types’,* each named here according to its northern limit.
10.Fiordland: Olearia oporina, O. crosby-smithiana, Euphrasia integrifolia. Chionochloa teretifolia, Dracophyllum fiordense, Senecio bifistulosus.
11.Mt. Earnslaw-Mt. Aspiring: Aciphylla congesta, Celmisia argentea, C. bonplandii, Anisotome capillifolia.
12.South Otago: Celmisia lindsayi, Pimelea lyallii, Centrolepis strigosa (not endemic but restricted to this area). Euphrasia repens.
13.Central Otago: Mysotis pulvinaris, M. albosericea, Celmisia prorepens, Aciphylla simplex, Pygmea myosotoides. Lepidium kirkii.
14.Waitaki Valley: Pachycladon novae-zelandiae, Anisotome brevistylis, Dracophyllum muscoides.
15.Kirkliston Range-Southern Hunters Hills: Parahebe tetragona, Ourisia glandulosa, Celmisia ramulosa.
16.Mt. Cook-Northern Mackenzie Country: Celmisia densiflora, C. brevifolia, Carmichaelia petriei, Pimelea pulvinaris.
17.Rangitata Valley: Aciphylla dobsoni, Dracophyllum prostratum, Hebe hectori.
18.Hokitika River-Arthurs Pass-Rakaia River: Lobelia linnaeoides, Raoulia hectori, Bulbinella gibbsii var. balanifera.

In all the patterns so far described the species involved are relatively abundant and continuously distributed within their area. Later in this account there are also recorded a few cases of species from northern and southern New Zealand which show considerable disjunction between continuous populations and other known localised occurrences. This is probably to be related to contraction of areas.

The general situation as expressed in Fig. 7 is that from several centres (which are probably centres of dispersal), species distributions radiate outwards. Common limits are reached by groups of species, but patterns tend to be different for each individual species. This is to be expected in view of the likely differences in tolerance ranges between different species. The patterns tend to be eccentric about the centres of dispersal. There are tendencies for species to be restricted towards west or east and this is reflected in the presence of ‘western’ and ‘eastern’ facies. but some of the species extend into both western and eastern areas. It seems clear that species have migrated outward

* Since this article was written further field work reveals that another group of southern species has its northern limit in the Haast Pass-Lake Ohau region. The species include Dracophyllum menziesii, Coprosma astonii, Ranunculus buchananii.

page 21 from (and in some cases contracted towards) the centres of dispersal. Some limitations have been placed by the narrowness of the island, by the narrow extent of particular environments in west or east, by the presence of suitable routes for migration and probably by the rate of development of mature soil profiles in some cases. In spite of these, the eccentric distribution patterns are fairly clear. Although there is overlap in the central South Island of distributions of some endemics from north and south, there is no meeting of the total floras and the floristic gap traceable by examining distributions of these endemics is real enough. Agreement between generalised limits of endemics, disjuncts and vicarious species is good.

Disjunctions Caused by Recent Contractions

A further situation complicates the position with respect to the hypothesis of Pleistocene extinction in the central South Island. Outside the main areas of distribution of numbers of species there are found distant, isolated occurrences. The ‘partial’ disjuncts described previously are of this type, but many others also exist which are not dicentric. Again, in these monocentric patterns the tendency is for a distribution to tail-off into the central South Island. The gaps between populations seem too great to have been bridged by long distance dispersal. In most cases, too, the apparent lines of dispersal lie athwart prevailing wind gradients or likely pathways such as river valleys and mountain ranges. Contractions of area are believed to be the explanation for these distributions. Species which appear to have contracted northward are: Celmisia allanii, continuous to the Hurunui River, one occurrence in the Godley River: Hebe cheesemanii, continuous to Hurunui River, occurrences at Mt Peel, South Canterbury, Kirkliston Range and Mt Alta, Otago; Ranunuculus insignis (‘monroi’ type), continuous to Waimakariri River; occurrences at Mt Hutt, Mt Peel and Four Peaks; Leucogenes leontopodium, local but not uncommon on North Island and Wairau Mountains, isolated populations at Hurunui River and Mt Peel; Hoheria sexstylosa, common to about 40 deg. 30 min. south latitude, occurrences on Banks Peninsula (possibly now extinct) and near Gore; Pittosporum patulum, local but not uncommon north-west Nelson to middle Clarence Valley, occurrences near Lake Ohau. Species which appear to have contracted toward the south are Euphrasia dyeri, not uncommon in bogs in Stewart Island, Fiordland, Southland and South Otago, occurrences at Mt Kyeburn and Mt Somers; Coprosma intertexta, common in Southland-Otago, occurrences at Cass, Waiau River, Rahu Saddle, north-west Nelson; Olearia moschata, common along the main divide to about Mt Cook, less common in Rangitata and Rakaia Rivers, rare at Arthurs Pass, represented by hybrids with O. page 22 avicenniaefolia (= O. haastii) in the Waiau River and probably by hybrids with O. nummularifolia at Boulder Lake in north-west Nelson. It may be noted here that the dicentric Nothofagus fusca which also has outlying populations in the central South Island, is sometimes represented there by hybrids with N. s. cliffortioides where no pure individuals of the former species are now found. Wind-blown pollen could account for this in some cases.

It seems probable that these monocentric and some or all of the ‘partially disjunct’ dicentrically distributed species have undergone contraction later than the separation of the species regarded as full disjuncts. Some of the latter may also have had their areas contracted.

Animal Distributions

Some animal distribution patterns fit into the general patterns of plant distributions outlined above. Powell (1957) described the following marine animals with disjunctions in their distributions between North Island and South Island: the mollusc Amphidesma ventricosum (toheroa) and at least one other mollusc and a species of brachiopod. Fleming (1950) described a molluscan fauna from the coast of Fiordland containing three genera and 16 species otherwise known only from North Auckland or North Island. Mr P. Johns (pers. comm.) states that although there are no species known by him to be disjunct between Nelson-Marlborough and Otago-Southland, certain groups of species of the terrestrial animals, ground beetles, millipedes and cockroaches are closely related in the two areas, separated by a gap in the central South Island. They are probably equivalent to the pairs of vicarious plant species. Similarly, the distributions of individual species extending into Canterbury from north and south are parallel in many cases to plant distributions. Lee (1956, 1959) recorded that one genus of earthworms in New Zealand was represented only in Fiordland and Auckland Province, and three species of Taranaki-Wanganui are most closely related to species of the south-western South Island and subantarctic islands. Comparatively unspecialised faunas consisting of widespread species are present in the eastern South Island. Westland is likewise unspecialised with respect to earthworms but there are some endemic species and the area had been colonised from both north and south. Banks Peninsula is also an area with a small number of endemic earthworms. The Wellington-Nelson area is regarded as a centre of dispersal for these animals.

Conclusions

Distributions of many plants and some animal groups suggest that there is a real floristic-faunistic gap in the central South page 23 Island. The explanations for the interrupted patterns of distribution as described above cannot be simple, however, and various factors seem to have been involved. Glaciation during the Pleistocene seems to have been a primary cause, but glaciation was multiple and there were long interglacial periods with mild climate so that the effects of ice maxima and glacial recessions on plant distributions must have been repeated again and again.

A brief consideration will now be made of the role of glaciation and post-glacial climatic changes in bringing about discontinuous distributions. A more detailed account of these changes, together with interpretations of the available pollen analyses will be published elsewhere. References for the statements made below are included in Burrows (1964 M.S.).

The last major glaciation, known as the Otira Glaciation (Gage 1961), was preceded by an interglacial in which there is evidence for mild climates and vegetation similar to the present in the South Island. This probably is the time when most of the plant species now disjunct were distributed throughout the Island. The onset of the first stadial of the Otira Glaciation brought about their disjunction. The vicarious pairs of species may have evolved after disjunction brought about by one of the earlier Pleistocene glaciations. Even at the height of each stadial some vegetation would have survived throughout the South Island. This would have included scrub and tall and short tussock grassland composed of unspecialised species now widely distributed. Specialised species of scree, rock and ‘periglacial’ sites would also have been present in the many suitable habitats. Many other species including forest trees would have been limited by cold to refugia in extremities of the island. During the interstadials before the last (Blackwater 2 and Poulter) ice advances of the Otira Glaciation (Gage and Suggate 1958) forest species including Nothofagus s. cliffortioides and N. menziesii expanded into the central South Island (as they probably did during each interstadial, to be pushed back as the next ice advance began). This vegetation was probably not completely dislodged from the centre of the Island by the subsequent ice advances. It was, however, likely to have been severely limited. Soon after the retreat of Poulter ice about 15,000 years B.P.* (McKellar 1960) the central South Island again was fully vegetated. Various assemblages of forest trees including beeches and podocarps were present in the areas north of about latitude 42 deg. 30 min. and south of about 44 deg. probably forming discontinuous forest, but mosaics of the smaller podocarps (Phyllocladus alpinus, Dacrydium biforme) and the larger

* B.P — before 1950 A.D

page 24 Podocarpus hallii, with scrub and grassland existed east of the main divide in the centre of the island proper. In central Westland the vegetation was much as it is at present, although podocarp stands on the young surfaces were probably denser. It is likely that the alpine species with present restricted distribution began to recolonise the central South Island from the northern and southern refugia contemporaneously with the lowland species. Many other more widespread alpine species would have been drawn from the flora of grassland vegetation which occupied the lowlands during glacial maxima. They would have occupied various open habitats during the reforestation of deglaciated areas. In the light of this early reforestation, the failure by some species to complete the colonisation of the central South Island is puzzling. Nothofagus species for example have had ample time to enter this part of the island, since some of them had reached Lake Hawea from the southern refugium somewhat later than 15,000 years B.P. (McKellar 1960) but earlier than about 8000 B.P. Beech species had also reached the upper Waiau-uha River from the north by about 14,100 B.P. (Dr R. P. Suggate pers. comm.). They probably had also colonised the easternmost ranges of Canterbury as far south as the Waimakariri River early in the post-Poulter period.

The explanation of this apparent anomaly is thought to be two subsequent events. The first of these was a renewed ice advance, less extreme than the Poulter advance. By analogy with events in the northern hemisphere and in South America (Flint 1961, Auer 1956, 1958) this short-period stadial, known in New Zealand as the Birch Hill advance, occurred about 10,000-11,000 B.P. The moraines left by it are most evident in the central South Island, and there were probably only cirque glaciers in Nelson and Fiordland. Soon after the Birch Hill ice advance there began a period of cool (upland) to warm (lowland) moist climate known as the ‘climatic optimum’. This effectively prevented further beech advance at the time. Cranwell and von Post (1936) and others have demonstrated how podocarp-broadleaved forest during this time largely superseded beech forest. Beech was, however, present at all times and must have been restricted to higher altitudes and various somewhat extreme habitats. At the fronts with mixed forest in central Westland, the Taramakau and Paringa Rivers, beech species remain limited to the present day by efficient competition chiefly from broadleaved trees such as Quintinia acutifolia and Weinmannia racemosa. The high rainfall occurring in central Westland is an important controlling factor. In headwaters of Canterbury rivers from Mt Cook to the Wilberforce tributary of the Rakaia the upland podocarp forest was maintained in the same way and beech immigration is probably of recent origin. The beech forest front with mixed forest in these valleys is approximately at the position of moraines of presumed Birch Hill page 25 age. Other forest types appear, thus, to have been maintained in the central South Island, at the expense of beech, by high rainfall in the period from about 9,000 B.P. to about 5,000 B.P. (and in some cases to the present). N. menziesii forest, which dominates in Fiordland, supplanted mixed forest (Harris, 1963), probably because of its presence in many places during the ‘climatic optimum’ and because of the local features of cool climate and extremely wet soils to which it is well suited. A renewed glacial advance, probably about 6,500 B.P., may have had some bearing on the subsequent vegetational changes. By analogy with these known situations, other species including some alpine plants could have been restricted by climatic factors.

Some of the discontinuities in populations of ‘partial disjuncts’ and monocentric species which have undergone contraction may have been brought about by the same causes. It is difficult to conceive of any phenomenon so profound that it could cause extinction in the central South Island on the scale described in the opening part of this article later than the last main series of glacial advances of the Pleistocene. The Birch Hill and later advances do not seem to have been intense enough to cause primary disjunction. It seems probable however that post-Poulter climatic changes have inhibited plant migrations from refugia and have also in more recent times brought about contractions in areas. It is probable that subsequent to the period of moist climate with mild temperatures between about 9,000 B.P. and 5,000 B.P. (for dating see Deevy and Flint, 1957) climates began to become at first warmer and then increasingly cooler and drier. Some specific data on distributions of certain plant species support this theory of climatic deterioration probably since about 2,500 B.P. A peat sample from the Grey River valley dated at about 8,300 B.P. (Bowen, ex Grant Taylor and Rafter, 1962) contained Metrosideros robusta pollen (a little south of its present range) and a small amount of Agathis australis (Dr W. F. Harris pers. comm.). Cranwell and von Post (1936) mention the occurrence of Alseuosmia pollen in a sample from Southland. This sample is presumably from a time equivalent to the above date — during the ‘climatic optimum’. Agathis is not now found south of the Waikato, and Alseuosmia not south of Nelson so that major contractions of area are evident. Some time after about 5,000 B.P. beech pollens begin to supplant podocarp pollens in the peat columns, apparently in a warm, dry period (Cranwell and von Post 1936, Harris 1963). This process was accelerated after about 2,500 B.P. (Deevey and Flint 1957, Harris 1963). Since this time there must have been a decline in temperature and also in rainfall. Changing climate may well be the most important cause of the contractions of many species. Dry, even arid climates seem to be involved in the limitation of species such as Nothofagus menziesii. page 26 Biologists and others should look for evidences of a period of aridity in post-glacial times. The arid period may have occurred within the last few thousand years.

Scattered discontinuous distributions of various other widespread species throughout the South Island may have resulted from the same causes of contraction as the examples described above. One cause of many such discontinuities in the eastern South Island, however, undoubtedly is fire during the Maori era (see e.g. Molloy et al. 1963). Subfossil remains demonstrate that forest has been extensively fragmented by fire within the last thousand years. Species dependent on a forest environment are limited by this (see Burrows 1961). Fire is not the only likely cause of discontinuity, nor even the primary one since the evidence for climatic deterioration in the last few thousand years is quite clear. Climatic fluctuations seem to have continued up to the present (Holloway 1954, Wardle 1963b).

Brief consideration of the Auckland-Nelson disjunction (Wardle 1963a) may now be made. During the early Otira Glaciation the southern and central North Island apparently were subjected to severer climatic conditions than was north-west Nelson. The latter area was at the western side of an extensive plain exposed by low sea levels (Fleming 1962). General extinction of forest occurred in the southern half of the North Island although scrub and grassland were present at lower altitudes. During the Poulter stadial beech forest was probably extensive. The question as to whether the tender kauri associate species survived the whole Otira Glaciation in north-west Nelson is still an open one. The influences of Birch Hill cooling have not yet been recognised in this area nor in the North Island. The ‘climatic optimum’ climates would have enabled expansion southward of the kauri associates (and probably kauri itself) so that disjunction of these species may be subsequent to it. Vulcanism in the central North Island (see e.g. Taylor 1953) is almost certainly an important contributor to limitation of many species north of the 38th parallel. Widespread ash showers have probably been as efficient in causing extinction in the centre of the North Island as glaciation in the centre of the South Island, but their invocation as a cause of complete extinction in the whole southern North Island poses various problems. Why did some of the species concerned not survive in the western North Island or near Wellington?

The detailed information presented above supports the general hypothesis of Wardle (1963a) that the Pleistocene glaciation caused extinctions of plants (and animals) in the central South Island and contractions into refugia. In the north, the refugia are likely to have been the coastline and hills of north-west Nelson to the Paparoa range, the coastline of Marlborough and, by virtue of a few plants endemic there, Banks Peninsula. In the south, refugia page 27 are likely to have been Southern Fiordland, the coast of Southland to South Otago, and Stewart Island. The latter area and its extended coastline is likely to have been of primary importance as a refuge. A group of specialised endemic species in Central Otago, another in northeast Nelson and eastern Marlborough and another found throughout the South Island mountains suggests that these areas also carried some vegetation at least throughout the latter part of the Otiran Glaciation. Some distribution patterns suggest comparatively recent contractions of the area of a considerable number of plant species.

Floristic Areas

As a result of this detailed knowledge of distribution some doubt is cast on the validity of the botanical district concept as framed by Cockayne (1928). On the basis of numbers of endemic species three main floristic areas may be discerned in the South Island (Fig. 8). Each of these contains at least 80 endemic angiosperms. They are: 1. Nelson-Marlborough. 2. Otago-Southland-Stewart Island, 3. An area which overlaps with both of these and runs the length of the island along the mountain chains from Nelson to Southland. There is a concentration of endemics in this latter area mainly east of the main divide and many of
FIG 8 FLORISTIC AREAS IN THE SOUTH ISLAND (see text)

FIG 8 FLORISTIC AREAS IN THE SOUTH ISLAND (see text)

page 28 the species are rupestral, scree plants, or plants of other open habitats. In the light of the habitat preferences of the many endemic species found throughout the South Island mountains and of the unspecialised habitat requirements of many other widespread plants not confined to the South Island it seems probable that plants survived the height of the last (Otira) Glaciation in many places. It is likely, even, that many species were favoured by the large number of open habitats available. The refuge concept is applicable, however, to the more restricted plants which have not yet adjusted to post-glacial conditions for various reasons. It also applies, in the main, to tree species. Detailed analysis of distribution patterns of subspecific variation in widespread species awaits study but there are indications from field observations and from Allan (1961) that some such variation is linked to the same areas discussed above as refugia. Minor evolution would be enhanced by isolation during the glacial maxima and the subsequent re-meeting of populations of species. These phenomena, it is apparent, have occurred many times during the Pleistocene. Within the three floristic areas described above (which are defined solely by the presence of a large number of endemic species) there may further be delineated floristic sub-areas, defined arbitrarily by the presence within each of at least ten endemic species. Areas with smaller numbers of endemics could be demarcated but it is not considered worthwhile to take this subdivision further here. The sub-areas in Nelson-Marlborough include 1(a) North-west Nelson to the Paparoa Range; 1(b) North-east Nelson and eastern Marlborough; 2(a) Fiordland; 2(b) Central Otago; 2(c) Coastal Southland-South Otago-Stewart Island. No sub-areas may clearly be discerned upon this basis within area 3. In this, the main South Island floristic area, there is, however, some segregation of species into western and eastern facies, in part controlled by rainfall differences.

Acknowedgments

Grateful acknowledgment is made for permission to consult specimens from those responsible for the herbaria of the Canterbury Museum and the Botany Division, D.S.I.R. I have also to thank Mr A. P. Druce and Mr N. Elder for information about existing North Island distributions of Pimelea aridula. Several of my colleagues have offered useful criticisms of the script and I express my thanks to them. I am especially grateful to Dr W. F. Harris for permission to use some pollen analytic information, to Dr R. P. Suggate for permission to use carbon dates and, with Dr M. Gage, for their willingness and generosity in discussing problems of glacial geology.

page 29

References

Allan, H. H., 1961. Flora of New Zealand, Vol. I, Wellington.

Auer, V., 1956. The Pleistocene of Fuego-Patagonia. I. The Ice and Interglacial Ages. Ann. Acad. Sci. Fennicae 45, 226 pp.

——, 1958. The Pleistocene of Fuego-Patagonia. II. The History of the Flora and Vegetation. Ann. Acad. Sci. Fennicae 50, 239 pp.

Burrows, C. J., 1961. The Forest Flora of Canterbury: Ecological Inferences. Proc. N.Z. Ecol. Soc. 8, 23.

Cain, S., 1944. Foundations of Plant Geography. London.

Cockayne, L., 1928. The Vegetation of New Zealand. 2nd edition, Leipzig.

Cranwell, L. M., and von Post, L., 1936. Post-Pleistocene Pollen Diagrams from the Southern Hemisphere. Geog. Annaler 3-4, 308.

Deevey, E. S., and Flint, R. F., 1957. The Post Glacial Hypsithermal Interval. Science, 125, 182.

Edgar, E., 1964. The Leafless Species of Juncus in New Zealand. N.Z.J. Bot. 2, 177.

Fleming, C. A., 1950. The Mollusca of the Fiords of Western Southland. N.Z.J. Sci. and Tech. B. 31, 20.

——, 1962, New Zealand Biogeography: A Palaeontologist's Approach. Tuatara 10, 53.

Flint, R. F., 1957. Glacial and Pleistocene Geology. New York. 553 pp.

Gage, M., 1961. New Zealand Glaciations and the Duration of the Pleistocene. J. Glaciol. 3, 940.

Gage, M., and Suggate, R. P., 1958. Glacial Chronology of the New Zealand Pleistocene. Bull. Geol. Soc. America 69, 589.

Grant-Taylor, T. L., and Rafter, T. A., 1962. New Zealand Radiocarbon Age Measurements 5. N.Z.J. Geol. Geophys. 5, 331.

Harris, W. F., 1963. Palaeo-ecological Evidence from Pollen and Spores. Proc. N.Z. Ecol. Soc. 10, 38.

Holloway, J. T., 1954. Forests and Climate in the South Island of New Zealand. Trans. Roy. Soc. N.Z. 82, 329.

Lee, K. E., 1956. Natural Areas in New Zealand — Earthworms. Proc. N.Z. Ecol. Soc. 4, 12.

——, 1959. The Earthworm Fauna of New Zealand. N.Z. D.S.I.R. Bull. No. 130, 486 pp.

McKellar, I. C., 1960. Pleistocene Deposits of the Upper Clutha Valley, Otago, N.Z. N.Z.J. Geol. Geophys. 3, 443.

Molloy, B. P., Burrows, C. J., Cox, E. J., Johnston, A., Wardle, P., 1963. Distribution of Subfossil Forest Remains, Eastern South Island, N.Z. N.Z.J. Bot. 1, 68.

Powell, A. W. B., 1957. Shells of New Zealand. 3rd edition, Auckland.

Taylor, N., 1953. The Ecological Significance of the Central North Island Ash Showers. The Soil Pattern. Proc. N.Z. Ecol. Soc. 2, 11.

Wardle, P., 1963a. Evolution and Distribution of the New Zealand Flora, as Affected by Quaternary Climates. N.Z.J. Bot. 1, 3.

——, 1963b. The Regeneration Gap of New Zealand Gymnosperms. N.Z.J. Bot. 1, 301.