Discontinuous plant distributions have been one of the main preoccupations of plant geographers (see e.g. Cain, 1944). They are also of interest to plant ecologists. The implications of separation of a species into populations with gaps between them are manifold. Grouping of kinds of discontinuity enables one to distinguish some of these implications more clearly. A quotation from Cain will exemplify this. ‘(1) Within their areas, species populations do not have absolutely continuous distributions. The so-called continuous areas are only relatively so, the disjunctions being less than the normal dispersal capacities of the diaspores of the species. (2) Environmental discontinuities — consisting of regions of particular topographic, climatic, edaphic or biotic characteristics separated from each other by regions of different character — constitute a sufficient basis for floristic discontinuities, since organisms are limited by their tolerances to the particular kinds of habitats to which they are adapted; but these facts provide no explanation of how … members of the same species got into separated regions. (3) Minor discontinuities of areas probably frequently result from recent migrations, but major disjunctions seem almost exclusively to have resulted from historical causes which have produced the migrations, in a once more nearly continuous area, through destruction or divergent migrations caused by climatic or some other changes.’

It is probably best to retain the term disjunction for the widely separated populations of a species and to refer to locally separated populations with an ecologic basis and potential free seed and/or pollen exchange as local discontinuities. One implication not stressed by Cain in the above statements is that there are possibilities, in cases of wide disjunction, of long distance dispersal of diaspores. The close similarity of New Zealand moss, fern and orchid floras (all with very small disseminules), for example, to those of Australia, and the presence of favourable westerly wind streams is highly suggestive. In other cases this explanation seems unlikely and discontinuities must be accounted for by a theory of removal of part of the area and possible shrinkage of the remaining disjunct groups. Another factor which may complicate the issue is the difference in physiology between discontinuous populations. In trying to account for discontinuous distribution, plant geographers may need to invoke ecological, palynological, or evolutionary evidences, as well as information from outside the botanical disciplines. The interest of New Zealand plant geographers has often been applied to problems of distributions of taxa of various ranks between New Zealand and some other land mass but within the country there are equally interesting discontinuities some of which have had attention from botanists in the past. It is intended to reopen discussion on some of these, pointing out their extent in more detail.

Some ecologic and other information will be adduced in an attempt to explain the circumstances of the discontinuities.

(1) Local Discontinuities
Distributions of Species on Mountain Tops and Lowlands

Local discontinuities often occur in Canterbury between stony mountain tops with fellfield (wind desert) vegetation and riverbed or well drained river flats. The mountain top at 5000-6000ft. of altitude may overlie the lowland at 1000-2000ft. but often there is dense forest between the two sites. Species discontinuous in these circumstances are Poa lindsayi, Pimelea sericeo-villosa, Raoulia australis, Cyathodes fraseri, Muehlenbeckia axillaris. The simplest explanations for such distributions are similarities in habitat in the different sites. In each case there is a well drained gravelly substratum which may be relatively stable or liable to be disturbed and both sites are cold in winter. The mountain tops are blown free of snow and experience very low temperatures while the valley floors are areas of inversion where frost is also heavy. Intermediate sites are warmer and have deep, mature soil profiles. Many other mountain species, if the opportunity arises, descend to low levels. Species such as Podocarpus nivalis, Celmisia spectabilis, page 127 C. angustifolia, Chionochloa australis, Notodanthonia setifolia, are often found on lowlands below high peaks in Canterbury. As was noted by Burrows (1960) this phenomenon is associated with destruction of forest which formerly occupied both lowlands and mountain slopes up to about 4500ft. In some cases there is discontinuity over several thousands of feet but in others distributions are more or less continuous. There is a natural tendency, even where closed forest separates the mountain tops and lowlands, for many mountain species to migrate down stream beds, avalanche chutes and other open areas. Van Steenis (1962) has discussed this phenomenon in tropical mountains. One clue to the nature of the relationship of alpine vegetation with forest in New Zealand is evident from the way in which many of these mountain species are restricted to open habitats if they descend to the lowlands. They potentially are able to inhabit lowland sites but normally are confined above the treeline because they cannot compete with the trees. There are, possibly, few true microtherms, restricted rigidly to alpine habitats. In Fiordland on the only open ground available, along stream sides and on slips, or even on rocks at sea level, separated by 2000ft. of forest from other populations on the mountain peaks, are species such as Celmisia coriacea, C. holosericea, Phormium colensoi, Forstera tenella and Ourisia macrocarpa. Leptospermum scoparium, better known as a lowland species, is found at treeline and at shoreline in parts of Western Fiordland. In Westland some inverted vegetation sequences are found. As one descends a hillside one passes from alpine grassland through subalpine scrub to treeline, then subalpine forest, then midslope forest, once again to meet subalpine forest and scrub bordering the open valley flats. Among these discontinuously distributed subalpine species are Olearia ilicifolia, O. lacunosa, Libocedrus bidwillii, Dacrydium biforme, Hoheria glabrata, Phyllocladus alpinus and Dracophyllum longifolium. If an avalanche path or stream bed traverses forest in this area some of these species may be found on their margins.

These patterns, where similar vegetation of subalpine zone and valley floor is separated through 2000ft. or more of altitude by forest with completely different composition, presumably are primarily due to climatic similarities in the two sites. The valley floors are subject to strong night-time temperature inversions and heavy frost in winter. This parallels the climate of the subalpine zone. The midslopes, however, are warmer sites, especially at night (see Geiger, 1958). Another controlling factor of such patterns is the release from competition which is found in habitats more open than the forest proper.

(2) Disjunct Distributions

Distributions of Species in Coastal and Inland Sites

Cockayne (1906, 1928) discussed some examples of species disjunct between mountains and coasts. He concluded that there is, in some cases, close ecological resemblance between habitats inland and on the coast. He also suggested that the plants may be confined to these habitats by competition from other communities and tentatively invoked conditions in former ages related to shrinkage of land and extension of coastline inland to account for some of the distributions. The following discussion elaborates on some of these suggestions in more detail. Earlier writers (Kirk, 1870; Petrie, 1895) had also remarked on such distributions and still others are recorded by Allan (1961) as a result of recent collections. Correlations with a range of different factors of the environment are evident when these distributions are examined in detail. One such group of species is found in wet habitats. Coastally they grow in tidal estuaries, brackish lagoons or salt marshes and most are commoner here than inland. Inland they are found submerged or around margins of lakes or tarns or in wet patches in grassland. The physiognomy of the inland vegetation around tarns is very similar to the short turf of salt marsh meadows on the coast: Selliera radicans, Leptocarpus simplex, Scirpus americana, Cotula dioica, Ruppia maritima, Zannichellia palustris, Triglochin striata var. filifolia. Another page 129

Fig. 1.
Inland stations of some coastal species

page 130 group of species inhabits sites rich in mineral salts. The coastal habitats are salt marshes or areas swept by salt spray. Inland, some of these species live in solonetz soils with sodium chloride accumulation in Central Otago: Apium filiforme, Lilaeopsis novaezelandiae, Glossostigma elatinoides, Chenopodium ambiguum, Samolus repens, Ranunculus acaulis. Other species of this group live on limestone, other young sedimentary rocks, or on volcanic rocks in their inland sites: Apium australe, Asplenium obtusatum, Myoporum laetum, Rhagodia triandra, Tetragonia trigyna, Gingidium (Angelica) geniculatum. A third group of species inhabits sandy or gravelly sites. On the coast these live on sand dunes, stony or gravelly beaches. Inland they occupy riverbeds and gravel flats or, in some cases, screes: Raoulia hookeri, R. australis, Zoysia pungens, Eryngium vesiculosum, Muehlenbeckia ephedroides, Claytonia australasica (in wet gravel). A fourth group of species found on the coast and inland inhabits cliff faces or steep hillsides in both sites: Metrosideros excelsa, Phormium colensoi, Linum monogynum, Helichrysum selago, Aciphylla colensoi. Fig. 1 shows inland distributions of some of the species mentioned above which are more generally distributed on the New Zealand coastline.

From the observed correlations of habitats it is clear that, if the plants in coastal and inland sites are physiologically equivalent, a range of factors must control their existence in different sites. Not all of them are dependent on saline habitats. In fact, some of the inland sites are low in available bases. It is fairly certain that physiologic races will be found in some of the species. The morphology of Selliera, for example, is different in its different sites, and several forms of Craspedia uniflora, Triglochin striata var. filifolia and Claytonia australasica exist, so their physiologic requirements may also be different. On the other hand, Leptocarpus, Muehlenbeckia ephedroides, Apium australe and others from coastal and inland areas appear to be morphologically identical. The species in wet habitats presumably do not require saline conditions. Some of them are equipped with aerenchyma tissue and they are suited to waterlogged soils with low oxygen and high carbon dioxide tension. Those species found inland in saline soils must exhibit a high tolerance for sodium and satisfy a high demand for other cations. They are true halophytes but, as Cockayne (1906) suggests, the ability of some of them to grow in cultivation in non saline soils shows that even these may be less dependent on saline conditions than is often supposed. Plants found on coasts and inland on young sedimentary or volcanic soils may reflect, in their inland habitats, high requirements for calcium, magnesium and potassium ions. In some cases the sedimentary rocks are saline. In others (estuarine mud deposits) there may be other similarities of habitat. The species of gravel, sand and cliff habitats seem to be controlled by good drainage conditions. page 131 In a few cases there is potential connection of inland and coastal sites by way of riverbeds.

Some of the species noted above may have broad ecological amplitudes which enable them to inhabit a wide range of habitats. An hypothesis to account for disjunct distributions of some of the species is that they are confined to their open and more or less extreme habitats through competition from species better adapted to the conditions of modal terrestrial habitats. The variations in habitats and their open nature in many of these cases suggest that this is so. Pigott and Walter (1954) and Pearsall (1964) have described discontinuous distributions of plants of open habitats in the British Isles. Such distributions are interpreted as being due to the successive increase in closed vegetation following deglaciation at the end of the Pleistocene period and, in particular, in the effects of the post glacial ‘climatic optimum’ period. Species adapted to the open habitats which prevailed after the retreat of ice have progressively been confined, by competition from other plant communities, to cliffs, sand dunes, mountain tops and other such habitats. The post glacial history of New Zealand is not substantially different from that of Britain and similar discontinuities of plant distribution here suggest that there are common causes for limitation of plants to a variety of open habitats.

Species Confined to Limestone or Other Mineral Rich Rocks

There is, in New Zealand, no well developed calcicolous vegetation. Only a few species are obligate or near-obligate calcicoles. One reason for this may be the comparatively recent emergence of extensive limestone habitats (Pliocene-Pleistocene). Another reason is that, in western areas heavy rainfall favours the formation of acid soil, even directly above limestone. The small group of ‘calcicoles’ displays some interesting disjunctions of the plant populations. The species include Poa acicularifolia, Gentiana astonii, Gingidium enysii, Asplenium lyallii. (See Fig. 2 for distributions of some of these species). As well as being found on limestone or marble, Poa acicularifolia is present on the magnesium rich rocks of Dun Mountain and Gingidium enysii is found on the volcanic rocks of Banks Peninsula, and occasionally on non-calcareous screes.

The wide separation of these species epitomizes one of the puzzles of plant geography. It seems possible that their seed may be disseminated over long distances. The same is likely to be true of the coastal-inland disjuncts. The hypothesis of restriction to open habitats following the post glacial increase in closed page 132

Fig. 2.
Distribution of some species largely confined to limestone

page 133 vegetation is likely to hold true, especially since some species inhabit a variety of open sites. Nevertheless, supply of cations is likely to be high in the case of the habitats of all these ‘calcicoles’.

Acknowledgments

Some of the detailed distribution data for this article was gathered from specimens in the herbaria of the Canterbury Museum and the Botany Division, D.S.I.R. Thanks are due to those in charge of the herbaria. Mr. G. Brownlie, Mr. W. Silvester and Dr. J. Soons kindly read the script and made useful suggestions. Miss P. Carpenter helped to prepare the figures.