Some Problems of Distribution of Indigenous Plants in New Zealand. By A. WallM.A., Professor of English, Canterbury College. Read before the Philosophical Institute of Canterbury, 1st April, 1925; received by Editor, 15th June, 1925; issued separately, 17th November, 1926.] The student of the flora of New Zealand in its wild state is confronted with very many problems arising out of the distribution of particular species. A certain plant, for example, will be found only in a very restricted area with no nearly related forms in its vicinity; or another will be found most nearly related only to one which grows many hundreds of miles away, both being quite rare; or yet another will be found commonly growing in two or more different localities at a great distance from one another, yet entirely absent in the intervening spaces. An attempt is here made to offer provisional explanations of some of these apparent anomalies. Geological Background. In order to consider problems of present-day distribution it is useless to go further back than the great post-Tertiary orogeny to which “and the consequent denudation is due almost all of the present relief of New Zealand, excepting that of the volcanic mountains” (Benson, 1921). At this time the component elements of our flora, the Paleo-zelandic, Antarctic, Malayan and Australian were all present; the surface of the country resultant from the block-faulting must be thought of in early Pleistocene times as “a group of differentially elevated earth-blocks” with high plateaux, “intermontane basins,” and narrow rectilinear rift-valleys, subsequently brought to its present state by a long process of denudation and sedimentation. We must think of the early covering-strata of marine origin, once a continuous sheet, as removed by erosion and remaining only here and there in corners and pockets. We must think of the climate attendant upon the Pleistocene elevation, and its extension into Pleistocene time, as “glacial” on the high plateaux and “steppe” or “semi-desert” on the Eastern side of the Southern Alps. We must think of New Zealand as extending, for the last time, far beyond its present limits and including indeed, the Chathams, Auckland, and other outlying Islands. Then we must think of the subsequent depression of the land, the wearing down of the high Alpine plateau to its present form in the Southern Alps, the retreat of the glaciers, and the return of a milder and more pluvial climate. Following Cockayne (1921) we note during this epoch the rise of the intense xerophytes “by epharmonic change” during the Pleistocene and again at the end of the Pleistocene, as a result of the depression of the land and the great climatic changes, “new forms of plants, descendants of the ancient species acted on by the novel and diverse environments.” At this time, thinks this authority, did Celmisia, Veronica, Epilobium etc. “burst forth into multiplicity of forms” and this process is not even yet complete.

From this account it is clear that we must regard our present species as resultant partly from mutations as demanded by De Vries (whether such-mutations be due to any determinable external cause or not) and partly from “epharmonic variation,” the forms responding to the stress of new conditions and becoming modified or adapted into true new species; partly to “reversions”; and partly, according to Dr. Cockayne's latest pronouncements (or even entirely) from hybridisation. This, in bare outline, is the past history to which we must look for the explanation and elucidation of our modern problems of distribution. It may be necessary, however, in certain cases to appeal to a still more remote past, and here may be mentioned, as offering a solution when all else fails, the theory that in “Oligocene-Miocene” time New Zealand was reduced to a group of small islands, when necessarily many species, especially alpines, must have become extinct “and plastic species, perhaps the sole survivors of large genera, would find a haven of refuge on rock-faces etc.” (Cockayne, 1921). This phrase “plastic species,” however, implying as it does inheritable variability, would now, I suppose, be withdrawn by the writer of this passage and replaced by some such words as “aggregates including many microspecies capable of intercrossing.” The natural difficulty of these problems is much increased by the too frequent absence of sufficient data. In the case of lowland species the activities of man have destroyed so much that we can never feel ourselves on firm ground in making inferences about the original distribution of particular species, while on the other hand in the case of alpine species, especially very small forms like the dwarf grasses and Carices, an immense amount of hard observational work must yet be done before we can feel at all certain that we have all the facts. In the writer's own experience it has happened more than once that the chance of his taking a particular line in descending through a high shingle basin has resulted in an extension of the range of a particular species by over a hundred miles. The same is true in a lesser degree of all small plants and all that grow in out-of-the-way or inaccessible situations. A further cause of complication is the destructibility (or the reverse) of the habitat. A rupestral, high alpine, maritime, or bog plant has natural bulwarks and will generally survive (e.g.) grazing or burning, when a species of the grassland or shrubland is in danger of extinction. So great a part of the lowland and hill districts of New Zealand has been subjected to these risks that we are often quite unable to say what was the original distribution of a given species. Attention has been directed in this paper chiefly to species in whose case these disabilities are absent, or as nearly so as possible. We may, for instance, be quite sure of our data in the case of such a conspicuous and famous plant as Olearia insignis which is purely rupestral and of very conspicuous and striking appearance; not so sure, but fairly so, in the case of Celmisia Traversii, which is alpine and subalpine. About fifty of the typical cases of difficulty are here tabulated:

A Species B Locality or Localities. C Nearest Affinity or Affinities. D Locality of Affinity. E Geographical Relation of A. to C. Species Locality or Localities. Nearest Affinity or Affinities. Locality of Affinity. Geographical Relation of A. to C. Ranunculus paucifolius Castle Hill, Waimakariri R. chordorhizos Macaulay R. to Mt. Hutt or Mt. Torless Both occur or occurred in Waimakariri District. Ligusticum flabellatum Stewart Id. L. Enysii Canterbury and Otago (rare) Never occur together. Angelica trifoliolata Torelsse and L. Tennyson A. decipiens and A. gingidium (?) S. Alps A. gingidium not far away, A. decipiens at Mt. Torlesse. Olearia insignis Parts of Marlborough No near affinity O. angustifolia (?) S.S. Island and Stewart Island Canterbury and Otago Alps Never together, nor within 200 miles. Celmisia rupestris Mt. Peel, Nelson C. Walkeri Canterbury and Otago Alps Never together, about 150 miles apart. C. Gibbsii Mt. Cobb, Nelson C. Walkeri and C. rupestris As above C. rupestris and C. Gibbsi neighbours. C. Walkeri not nearer than 150 miles. C. Traversii (1) Nelson and Hanmer Alps (2) S.W. Otago C. spectobilis and group Passim. C. Travesii and C. spectabilis neighbours. C. cordatifolia Nelson Mts. C spectabilis and group Passim. C. cordatifolia and C. spectabilis neighbours. C. Rutlandii Mt. Stokes C. petiolata “allied to C. petiolata” T.F.C. Central and Western Alps Never together, nor nearer that about 100 miles. C. Mackoui Banks Peninsula C. coriacea S. Alps and passim. Never together now, Mt. Grey nearest station to C. Mackaui about 50 miles.

A Species B Locality or Localities. C Nearest Affinity or Affinities. D Locality of Affinity. E Geographical Relation of A. to C. C. Petriei Otago and Southland; Colling wood? C. Lyallii S. Alps and passim. Together in Otago Alps. C. McMahoni Mt. Stokes, Fishtail, and Richamond. C. Hectori S. Alps from Mt. Cook southwards. Never together nor nearer than about 200 miles. C. parva Heaphy R. extreme N.W. C. Hectori S. Alps from Mt. Cook southwards. Never nearer than 200 miles. C. vernicosa C. Campbellensis Auckland and Campbell Island C. spectabilis group (?) S. Alps, passim. Affinity only on mainland R. only on outlying Islands. Roulia Goyeni Stewart Island R. bryoides S. Alps, passim. Not together on Stewart Id. Helichrysum coralloides N.E. District of S. Alps H. selago S. Alps, passim. Grow together. Cotula linearifolia Wairan C. pyrethrifolia S. Alps, passim. Grow as neighbours. Senecio saxifragoides Port Hills S. lagopus S. Alps etc., passim. Grow separately, but as near neighbours. S. glaucophyllus Mt. Aruthur S. latifolius Western S. Alps Not near neighbours probably. S. Greyii S.E. of North Id. S. latifolius S. Nelson Alps Never together or near. Helichrysum dimorphum Waimakariri Dist. H. depressum S. Alps, passim. Grow together. Wahlenbergia cartilaginea N.E. District of South Island W. saxicola Passim. Grow near together, but at different altitudes. Logania depressa Onetapu Geniostoma N. Island Sole representative of its genus in New Zealand. Myosotis decora Castle Hill, Mt. Torlesse M. antarctica S. Alps etc., passim. Not very far apart. Myosotis explanata W. Canterbury Alps M. capitata S. Island, cliffs on coast Never within 200 miles of each other.

A Species B Locality or Localities. C Nearest Affinity or Affinities. D Locality of Affinity. E Geographical Relation of A. to C. Myosotidium nobile Chatham Id. Myosotis (?) Passim. Stands alone. Veronica Dieffenbachii Chathams V. speciosa N. or N. Island Nowhere near each other. V. Barkeri Chathams V. speciosa N. or N. Island Nowhere near each other. V. anomala Upper Rangitata V. buxifolia Passim. Grow probably near each other. V. Gibbsii Mt. Rintoul, etc. V. Pinguifolia Nelson to Otago Grow together on Ben Nevis, etc. V. tumida Nelson Mts. V. tetrasticha and V. quadrifaria Passim, S. Alps Grows alone. V. Lavaudiana Banks Peninsula V. Roouli Nelson, Canterbury Grows alone, but V. Raoulii grew formerly with it. V. Hookeriana Volcanoes V. Lyalli Passim. Not together, but not far apart, and at different altitudes. V. spathulata Volcanoes Scutellaria novae-zelandiae Nelson No near affinity Muehlenbeckia Astoni Wellington and S. Id. Coast M. australis Passim Grow together. Poranthera microphylla P. alpina Nelson Province Euphorbia (?) Coastal, passim. Never together. Uncinia nervosa Mt. Arthur, Nelson U. compacta Passim in S. Id. Probably together or nearly. Carex pterocarpa Otago Alps and Canterbury, 4,500-6,500 ft. C. Kirkii General and Alps to 4,000 ft. Not far apart but at different altitudes. C. Hectori Old Man, Otago C decurtata C. uncifolia L. Tekapo, S. Alps, passim. Not together, nor at all near. Not together.

A Species B Locality or Localities. C Nearest Affinity or Affinities. D Locality of Affinity. E Geographical Relation of A. to C. C. Dallii Nelson Mts. C. Petriei S. Alps, passim. Not together. Hirochloe Brunonis Auckland and Campbell Islands H. redolens Passim Not together in Aucklands, but together in Campbell Id. Agrostis Petriei Otago A. Dyeri S. Alps, passim. Not together but not far apart. A. tenella Otago and Canterbury (?) Deschampsia novae-zelandiae D. Chapmoni D. Tenella D. pusilla Very scattered localities in S. Island and D. caespitosa General Not together. Danthonia pungens Stewart Island D. australis General Not together. Poa acicularifolia Limestone in Nelson and Canterbury P. Colensoi N. and S. Alps. passim. Not together, but near neighbours. P. pygmaea Otago Alps P. Colensoi (?) Geneal Agropyrum Enysii Canterbury Alps A. scabrum General Close together.

In order of difficulty the possible cases may be thus arranged: (1) The R (Restricted Distributant) and the A (Affinity) occur together or in fairly close proximity. Thus Helichrysum coralloides(R.D.) and H. selago; Senecio saxifragoides (R.D.) and S. lagopus; Celmisia Mackaui and C. coriacea may be placed here (?) (2) The R and A belong to the same district while generally growing in such different habitats that they rarely or never actually occur together. Thus Carex pterocarpa (R.D.) 4,500-6,500 ft.; and C. Kirkii to about 4,000 ft.; Wahlenbergia cartilaginea and W. saxicola. (3) The R and the A both grow in New Zealand proper but are separated from one another by great distances. Thus Celmisia McMahoni (R.D.) and C. Hectori; Veronica, rigidula and V. Colensoi (of the Manual). (4) The R occurs in New Zealand only with no near affinities, or in Stewart Id. or the outlying Islands with affinities only on the mainland of New Zealand. Thus Poranthera microphylla and P. alpina (R.D.) with no near affinities; Danthonia pungens (R.D.) on Stewart Id. and D. australis on the mainland; Celmisia vernicosa and C. Campbellensis (R.D.). This is still more striking in the case of Pleurophyllum for no representative of the genus occurs on the mainland. (5) The genus which includes one species only with no close affinities in New Zealand. Thus Logania, Xeronema, Tetrachondra, Myosotidium, and probably Olearia insignis. Of these there are sub-classes or special cases of which a few may be mentioned. (6) Celmisia Traversii generally restricted to parts of the Northeast Botanical District, is most nearly related, no doubt, to C. spectabilis and its group, as in (1) above, but occurs also in the far Southwest. Such apparent anomalies as this may be due to imperfect observation or record, or both, but hardly in the case of so striking a plant as this. This is also the case of Ranunculus crithmifolius: affinity, R. Haastii, widely distributed. (7) Celmisia rupestris and C. Gibbsii, both most closely related to C. Walkeri, grow in different parts of the same restricted area (in Nelson Province), from which C. Walkeri itself is absent. The case of C. McMahoni resembles this, but that species grows on different isolated areas within a single district. Considering these cases in their order:— (1) The first offers little or no difficulty. Helichrysum coralloides may be considered to have originated, whether by mutation or crossing or by “evolution,” from the commoner plant, and its restricted distribution may be set down to its comparatively recent origin. (So far as has been observed it shows no tendency to spread any further, however). This would appear to be the natural or normal procedure of nature; it is to be expected that the older or original species should exist with or not far from the younger; our difficulties begin when this does not occur. The Senecios, S. saxifragoides and S. lagopus, offer a harder problem, but they are very slightly differentiated from one another. They occupy, each on its own district, conterminous areas, without ever commingling. (Wall, 1918). The case of Cel-

misia McKaui confined to Banks Peninsula, while C. coriacea, its affinity, is generally distributed, but not found on the peninsula, may be one of the most difficult cases to explain, but we cannot certainly say that C. coriacea was never on the Peninsula, so that our data are possibly insufficient. C. coriacea is on Mt. Grey about 40 or 50 miles distant. As it tends to be a “fireweed” it seems improbable that it has existed on Banks Peninsula in recent times. Fire has been the great destroyer here, and it would rather encourage than repress it. (2) The case of the Carices (C. pterocarpa and C. Kirkii) is not very difficult. We may infer that C. pterocarpa originates from C. Kirkii (or of course both from one ancestral form) and that the differences between them are due (in part) to the difference of habitat, one being adapted for life at very high, the other at comparatively low altitudes. This does not preclude the possibility of an origin by crossing of microspecies. C. pterocarpa has only very recently been observed beyond Otago and it is so small that it may well have been overlooked and have a far more general distribution than is at present recorded. There must be many similar cases to this, but the reason why a certain species should in a certain district or districts give rise to a new form while refraining from doing so in others will probably always remain unexplained, though the “new stimulus” of changed conditions (as postulated by Cockayne 1921 in re return of Alpines to mountains in a period of elevation) may be called upon. (3) The case of C. McMahoni, on Mt. Stokes only, and C. parva, also in the Northern mountains but at a very considerable distance, or C. Hectori which seems not to be recorded in the Northern mountains at all, is very difficult. Here we may assume that the forms descend from a common ancestor and that although the great species-forming activity in Celmisia is comparatively recent (Cockayne, 1921) it must go back to a period when geographical conditions were entirely different from the present. The summits of the Northern mountains like Mt. Stokes, Richmond, Rintoul etc. are all in the nature of islets, each supporting an alpine flora isolated in vast areas of forest in which no such flora can exist. The Pleistocene period at least would seem to be demanded for this, though the forest may, of course, be comparatively recent. There are many such cases, where the closest affinity of a rare plant is separated from it by very great distance, especially in Veronica and Celmisia, and in all of these a truly recent origination seems to be barred. (4) The Celmisias of the outlying Islands may be taken as typical of this class. It is impossible to say exactly which species of the mainland Celmisias ought to be considered as their affinities. They belong to the herbaceous group; and possibly C. spectabilis may be taken as akin to some ancestral form of both C. Campbellensis and C. vernicosa. Judging by leaf-characters the two are not very closely akin; by flower-characters they are very near relatives. No other Celmisia is recorded from either of these groups of islands. Here we must push back still further the period at which these species can conceivably have originated, even as far, probably, as the time when these Islands and New Zealand formed one continuous land surface which cannot have been later than the Pleistocene (?).

An alternative is that the seeds of mainland Colmisias have been transported by wind or birds to these Islands and that on their establishment there the plants have developed their own individuality under the new conditions. But these are, however, not “new lands” but remnants of old which renders this last hypothesis improbable. Either alternative seems to demand the (Neo-Lamarckian?) development of distinctive characters by the fixing of “variations” however caused, rather than by hybridisation. The case of Pleurophyllum is not very dissimilar to this. The genus is represented by three species all of which are confined to the outlying islands of New Zealand; none occurs on the mainland or on Stewart Island. It seems to demand a still greater backward extension of time, and we are irresistiby drawn to take into consideration the old hypothetic Antarctica and regard Pleurophyllum as one of the most distinctive features of its flora, preserved to us by the continued existence of certain small portions of its area. (5) The case of a genus representing a single species also demands a further backward extension in time. It is useless to speculate upon a period so remote that a plant like Logania, or Tetrachondra could originate as a member of a group whose other members have now entirely disappeared, leaving the species almost perfectly isolated. Euphorbia can hardly be conceived as giving birth to Poranthera, and we can only conjecture what may have been the generic and specific links which will unite these now far-separated forms. The contemplation of these takes us back to the very early history of the Angiosperms when the very existence of land in these regions is problematical. This is the case of “relict endemism”; such species cannot possibly be conceived as “new.” (6) It is difficult to see how the distribution of Celmisia Traversii can be explained otherwise than by the hypothesis of polygenesis. The same parent species has apparently become mutable, or varied by crossing of microspecies in two widely separated places, and has given rise in these places and in no others to the same new form. The conveyance of a chance seed from one locality to the other over intervening great mountain chains and a distance of over 200 miles would seem to be incredible. Another alternative would be the hypothesis that the species has once been generally distributed and destroyed in every other locality except these two, but against this commonsense rebels. Yet another hypothesis is suggested by the similar case of Ranunculus crithmifolius, where the wide separation of the stations also appears, (Mt. Arrowsmith and the Wairau River, about 200 miles apart) but allowance must be made for imperfect observation. In this and similar cases we might look to the conception of a high plateau (up to 10,000 or 12,000 feet) with tectonic hollows capable of supporting an Alpine flora, such a plateau as is posited by Cotton and Benson, 1921, in mid-tertiary time. During the process of denudation of such a plateau and its dissection into the ranges as we know them to-day, with long narrow river-valleys separated by high ridges, isolations of particular species may well have been brought about. A plant like Ranunculus crithmifolius, originating in such hollows at a considerable altitude might move farther and farther up the neighbouring slopes as the land subsided and the plateau

was dissected out. Its place of origin may have been entirely destroyed, as practically the whole of the covering limestone cap was destroyed, and the plant itself may then survive only in a few widely separated stations where it appeared to have established itself in places which were not doomed. (7) The Celmisias, rupeslris and Gibbsii, might be supposed to be descended either from C. Walkeri or from an ancestral form common to all three. The absence of C. Walkeri in their neighbourhood is hard to explain. If all descend from a common ancestor, however, the two rafe plants may be supposed to have developed on different lines each in its own restricted sphere in comparatively recent times, whether by mutation or epharmonic variation or crossing with other species we cannot tell. One very clear and certain result of this study is that Willis's rule of “age and area” will by no means hold good for particular species in many cases. All depends upon affinities. Olearia insignis is seen to have a very restricted distribution and according to the “law” it should therefore be of recent birth; but that is impossible unless it arose from some one of the other Olearias by a very rapid series of gigantic mutations, or from some other species or groups which have all quite recently disappeared without cause shown. This reasoning will also apply to Logania, Mitrasacme, Poranthera, Tetrachondra and many others, including practically all monotypic genera. The theory of mutation as set forth by De Vries or of hybridisation as set forth by Lotsy can help to explain some of the difficult cases very fairly. A further consideration is here advanced with great diffidence as affording an explanation of some of the more puzzling distributions. According to De Vries a species or group of species in a state of mutability may throw off swarms of new microspecies in a particular district, but the species itself (or group) does not cease to exist. The mutations are confined to certain individuals and to a certain percentage only of the seeds of a given individual, so that the mutant remains (in nature) while its descendants are to be found in its vicinity. Further, he makes it clear that a particular species may be in a mutable state in a certain district while it is perfectly stable in others. If, however, it could be found that a particular species had become so mutable within a certain area that all the seeds of every individual bore mutated forms, then the species itself would perish in that area and be represented only by the new form. For example we may conceive an ancestral form antecedent to Ranunculus Haastii, R. paucifolius, R. chordorhizos, and R. crithmifolius reaching a period of very active mutability and throwing off, in one district by means of every seed, R. paucifolius, and itself perishing completely; in another R. crithmifolius, and so on. The new species would then exist each in its own area isolated, the “parent” species being absent. The case of Oenothera as studied by De Vries is not so very different from this. The single species, in his example, throws off a variety of microspecies at the same time and within the same area, we might ask whether it is not at least possible that it should throw off only one successful microspecies by means of every seed and then itself disappear. Such a theory presupposes the possibility of a certain unanimity on the part of the species itself or of what we vaguely call “Nature”

which is attractive, no doubt, to the poetic or imaginative mind, but seems hardly to fall within our conception of scientific or proven truth. “Nature,” says Cockayne, “recognises only the individual.” May it not be true, however, that “she” only recognises the species? Also there are not wanting some indications in “nature” of a certain invisible bond uniting all the individuals of a given species wherever they may happen to be scattered over the surface of the world. The case of a species of Bambusa may be called in evidence. At a certain time in a certain year all sufficiently mature individuals flowered wherever they were, while for many years all had been inactive. This is explained, however, by the fact that all were grown from cuttings and were really the same plant. The case of the common English scented musk may be also cited; all plants in Europe apparently have lost their scent, however, wherever cultivated, and all with comparative suddenness. And the New Zealand plants have nearly lost it also. In the animal world there are, of course, many well-known examples of this race-solidarity; such as the instincts governing the migration of insects, birds and small mammals, and changes in the habits of animals and birds; for instance, the case of the starling which became a fruit-eater almost simultaneously in England (Europe) and in New Zealand and Australia where it had been introduced. Conclusions. In most cases of restricted or anomalous distribution an explanation is to be sought in one or other of the following:— (1) Comparatively recent mutations, or crosses, or adaptative modifications producing new forms which are found only within a certain area: e.g. Helichrysum coralloides and H. selago; Senecio saxifragoides and S. lagopus; Celmisia McKaui and C. coriacea; Carex pterocarpa and C. Kirkii. (2) Changes in land-surface isolating certain forms, species or genera, these changes being of widely varying age, some comparatively recent (e.g. Celmisia McMahoni and C. Hectori), others more ancient (e.g. Poranthera and Euphorbia), others still older (e.g. Pleurophyllum). (3) Carriage of seeds to great distances by wind or birds or other agencies, followed by adaptive modifications or crosses (e.g. Celmisia vernicosa). This explanation is not favoured but mentioned only as a remote possibility. (4) Polygenesis seems inevitably suggested by certain cases, the same form having independently originated in widely separated localities whether by mutation or crossing, or epharmonic variation, e.g. Celmisia Traversii, and possibly C. Petriei. (5) The hypothesis of a universal or unanimous change of one species (all individuals) into another, by mutation (as of De Vries) or otherwise, is put forward tentatively as a suggestion in certain very difficult cases, e.g. Celmisia Traversii. The subjoined diagrams are intended to illustrate the particular problems here investigated. R stands for the plant of restricted distribution and A for its nearest affinity. Thus in (1) R stands for Celmisia McKaui, restricted to Banks Peninsula, from which C. coriacea, the affinity, is absent, while C. coriacea is present over the very wide area of the Alps.

Diagrams. (1) C. McKaui & C. coriacea. (2) H. Coralloides. H. Selago. (3) Carex pterocarpa C. Kirkii or W. Cartilaginea. W. saxicola. (4) V. rigidula. V. Colensoi (5) C. McMahoni C. Hectori. (6) D. pungens. D. australis. Myosotidium? (7) Pleurophyllum. (8) Poranthera. (9) C. Traversii and C. spectabilis. (10) C. rupestris. C. Gibbsii. C. Walkeri.