Art. XXI.—On the Distribution of Senecio saxifragoides Hook. f.′ and its Relation to Senecio lagopus Raoul.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 50, 1918, Page 198

Art. XXI.—On the Distribution of Senecio saxifragoides Hook. f.′ and its Relation to Senecio lagopus Raoul. By Professor A. Wall, M.A. Communicated by R. Speight, M.Sc. [Read before the Philosophical Institute of Canterbury, 5th December, 1917; received by Editors, 31st December, 1917; issued separately, 10th June, 1918. Plates XI–XIII. 1. Introduction. (a.) General. The problem to be attacked in this paper is suggested in the following passage from L. Cockayne (“Notes on the Plant Covering of Kennedy's Bush, and other Scenic Reserves of the Port Hills,” Report on Scenery Preservation, Parliamentary Paper C.-6, 1915) concerning S. saxifragoides: “It also is a most striking plant. Now, an almost identical species, named Senecio lagopus, also occurs on the main mass of Banks Peninsula, which differs from S. saxifragoides merely in the possession of numerous bristles on the leaf, whereas in the latter such are absent. Yet, so far as is known, S, lagopus does not occur on the Port Hills, nor S. saxifragoides on Banks Peninsula proper. If this is truly a fact, the distribution of these two species, each equally well suited to the rock-conditions of the area, is one of the most remarkable cases of plant-distribution in the world.” The same authority, in his description of his new species, Senecio southlandicus (Trans. N.Z. Inst., vol. 47, p. 118, 1915), further says, “The species is, indeed, far more distinct from S. bellidioides and S. lagopus than are these from one another. The classification of the whole series, including those already mentioned, together with S. saxifragoides Hook. f. and S. Haastii Hook. f., is in a most unsatisfactory position. Specimens are constantly coming to me from various correspondents which it is impossible to place with any degree of satisfaction. There are undoubtedly a number of well-marked forms, which demand, at the least, varietal names. Even one fixed character may serve quite well as a specific mark. This is illustrated in the case of S. saxifragoides and S. lagopus (the type from Akaroa), where the presence of numerous bristles, or their absence, on the upper surface of the leaf is the sole distinguishing character, so that, so far' as large plants of the two are concerned, if this character were not present no one could consider them in any degree different.” In this paper attention has been directed to these two species solely as they occur on Banks Peninsula. Banks Peninsula is situated in lat. 43° 32′ S. and long. 175° 30′ E., and forms a rough elliptical salient on the east coast of the South Island of New Zealand. Its diameter in a N.W.—S.E. direction is about twenty-five miles, and its breadth at right angles thereto about eighteen miles. Some forty miles to the westward stretches the main chain of the Southern Alps, from which the peninsula is separated by the gently inclined expanse of the Canterbury Plains, so that it is almost as completely isolated as regards the distribution of subalpine vegetation as if it had been separated from the mountain region by the sea. The oldest rocks within its limits consist of Trias-Jura sedimentaries overlain in places by a thin veneer of Cretaceous rhyolites, but the main

Plate XI Photographic reproduction of relief map of Banks Peninsula, from original in Canterbury Museum by S. C. Farr.

mass of the peninsula was built up in mid-Tertiary times by flows of basalt and fragmentary material of similar lithological character, poured out from two vents situated somewhere near the centres of Akaroa and Lyttelton Harbours. A third focus of activity lay near Mount Herbert (3,012 ft.), but it was of less importance, although it was responsible for the formation of the highest peak in the area. The high cones thus formed were subject to paroxysmal explosions of moderate intensity, and their surface was modified by the establishment on their outer slopes of a well-developed system of radiating valleys. Volcanic action ceased in all probability long before the end of the Tertiary era. After the stream-system had reached a mature stage the land sank, and the sea entered the floors of the enlarged craters and extended a considerable distance up the lower reaches of the valleys, and these now form marked indentations of the coast-line. Owing to the prolonged weathering the land is now covered with a rich and fertile soil, and steep rock-faces occur only on the coast and at higher levels, where the more resistant basalts form at times precipitous cliffs—the' characteristic habitat of the senecios under consideration. Map of Banks Peninsula and Port Hills, showing distribution of the two species of Senecio. L, Senecio lagopus; S, Senecio saxifragoides A, rhyolite escarpment where S. lagopus occurs; B, rhyolite escarpment where neither species occurs. The following are the most important geological considerations affecting the distribution and ecological conditions of plants established in the locality:— (1.) The isolation of the region from neighbouring mountain areas since it was first formed.

(2.) The uniformity of the lavas which form the majority of rocks in the area. No anomalies of distribution can be interpreted in the light of lithological differences in these rocks. (3.) As a result of prolonged denudation the crater-ring of the Lyttelton volcano has been broken down at its south-western side and a sector completely removed, so that a stretch of comparatively low country, nowhere over 875 ft. in height, and consisting of exposed rhyolites and sedimentaries, separates the northern part of the crater-ring from the other part of the peninsula. This northern part forms the low range usually called the “Port Hills,” and is referred to throughout this paper, as the habitat of Senecio saxifragoides, by this name. For a fuller account of the geological features of this area see J. von Haast, Geology of Canterbury and Westland, 1879, and R. Speight, “The Geology of Banks Peninsula” (Trans. N.Z. Inst., vol. 49, pp. 365–92, 1917). (b.) Historical. Senecio saxifragoides was first described by Hooker in 1853 (Flora Novae-Zelandiae, vol. 1, p. 144), and in the Handbook its discovery is accredited to Lyall (Handbook, p. 159). Hooker, Kirk (Students' Flora, p. 339), and Cheeseman (Manual, p. 372) agree in describing S. saxifragoides as distinguished from S. lagopus only in respect of the leaf, which is described as “clothed with shining silky and woolly hair “(Hooker), “silky or villous” (Kirk), “silky or villous” (Cheeseman), upon the upper surface only, and wanting the stout bristle which characterizes S. lagopus and S. bellidioides. Kirk says, “The leaves are often glabrous or glabrate on the upper surface, but never bristly as in S. lagopus.” Cheeseman says, “A handsome species, separated from large states of S. lagopus, some of which approach it very closely, by the much stouter habit, more copious villous hairs, and larger thicker leaves, which are silky above and never show the stout bristly hairs so characteristic of S. lagopus and S. bellidioides.” All agree that these three species, S. lagopus, S. bellidioides, and S. saxifragoides, are very closely allied. Hooker (Handbook) says, “This [S. lagopus] and the two following [i.e., S. bellidioides and S. saxifragoides], though most dissimilar in their usual states, appear to me to be united by intermediate forms,” and (Flora Novae-Zelandiae), “This and the two following are closely allied and very singular species.” The distribution of S. saxifragoides is given by Hooker as “Port Cooper”; by Kirk as “Port Lyttelton, Banks Peninsula”; and by Cheeseman as “Port Lyttelton and other localities on Banks Peninsula.” All agree in describing the leaf of S. saxifragoides as broader or more nearly orbicular than that of S. lagopus; but they do not quite agree as to the relative size. Hooker makes the leaf of S. lagopus 2 in. to 4 in. long (!); that of S. saxifragoides 3 in. to 5 in. long. Kirk makes the leaf of S. lagopus 1 in. to 8 in. long (excluding the petiole), and that of S. saxifragoides 3 in. to 6 in. long. Cheeseman makes the blade of S. lagopus 1 in. to 5 in. long, and that of S. saxifragoides 3 in. to 6 in. long. Raouls description of S. lagopus (Choix, p. 21) gives the leaf about 1 decimetre (4 in.) long and from 7 to 9 centimetres (3–3 ½in.) broad. Hooker's and Raoul's descriptions would seem to have been based upon comparatively small specimens of both species.

The distribution of S. lagopus is given by all authorities as from the Ruahine Mountains to South Canterbury. The original description of S. lagopus by Raoul, and his plate (Choix, pl. 17), must here be referred to, as of the greatest importance in the study of the two species. In describing the petiole of S. lagopus, Raoul says, “Petioli … canaliculati in vaginam semiamplexicaulem dense lanatam dilatati “; and in describing the leaf he says, “Folia … pilis rigidis grossis, spinescentibus praesertim ad. margines inspersa.” His plate shows a plant with four large and several small leaves. Of the four large leaves three are glabrate (as the old leaves of both S. lagopus and S. saxifragoides always are): the fourth bears the characteristic “bristles” very thickly close to the margin all round the leaf, or nearly so; and near the apex; less thickly upon the upper third of the leaf or thereabout; the lower part of the leaf bears the hairs only, very thickly distributed. The hairs and bristles occur together over some portions of the leaf, about the middle and towards the apex, but at the apex itself and in its immediate neighbourhood the bristles alone occur. The dual occurrence of hair and bristle* The term “bristle” is kept throughout, as that employed by previous authorities though the organ is really a glandular hair. on the same leaf, which no subsequent authority describes at all, will appear to be of great importance to this inquiry; and it may be added that my descriptions of the variant forms of S. saxifragoides given below were fully made before I had seen Raoul's plate. The species are further thus referred to by-Laing and Blackwell (Plants of New Zealand, pp. 437–38, 1906): “The handsome S. saxifragoides, sup posed by Kirk to be confined to Banks Peninsula, is undoubtedly the typical S. lagopus of Raoul. It still produces its large-leaved rosettes on the southern faces of cliffs, where Raoul found it, near Akaroa. It is also plentiful behind Lyttelton, often growing in altogether inaccessible localities, and it is the only Senecio which haunts these situations on the Peninsula.” 2. Special Characters of the Two Species Under Consideration. 1. All round the margin of the leaf of S. lagopus, S. saxifragoides, and S. bellidioides occur at very regular intervals—i.e., at the ends of the veins—rounded glandular protuberances of a very dark red or purple colour. Microscopical examination shows these to be typical hydathodes. 2. The petiole of the leaf of S. lagopus bears quantities of dark-red or purple bristles, generally spotted or pied with white; these are continued up the back of the midrib nearly to the apex of the leaf, and are also present along the margin all round the leaf. 3. S. saxifragoides also shows this purple or pied bristle upon the petiole and the back of the midrib exactly as in S. lagopus, and also bears this bristle all round the margin of the leaf, making a continuous fringe. The young leaves of S. saxifragoides have yellow marginal bristles, which change gradually into purple and continue to deepen in colour up to maturity. The yellow bristle, however, is usually present together with the purple, the yellow being upon the upper surface of the leaf just within the margin, the purple being upon the” margin itself; but they are sometimes more or less mixed together. 4. Forms of Senecio lagopus and Senecio saxifragoides which grow in situations shaded by other vegetation, as among long tussock-grass, or on

the edge of forest, or beneath large plants of Linum monogynum, show a complete or almost complete absence of the purple colouring-matter in the glandular hairs, and also tend to be larger, to have a much longer petiole than usual, and to be less thickly covered with silky hairs or bristles (as the case may be) than the usual form. 5. Though the plant of S. saxifragoides is probably on the average a little larger than that of S. lagopus, the difference is not great. Kirk's measurements do not agree with the others. A leaf of S. lagopus 8 in. long without the petiole would be most exceptional, but leaves 6 in. long with a petiole of from 3in. to 4in. are common—e.g., on Mount Herbert—and S. saxifragoides can hardly ever be much larger than this, though its leaves are generally broader and more substantial. I have a very strong impression, which I hope to verify by future observation, that the plants of S. lagopus in the neighbourhood of Akaroa Harbour are in general distinctly smaller than those of the Mount Sinclair to Mount Herbert area. This would explain the small measurements of Raoul's type. Some of the large individuals of the Mount Herbert area, indeed, almost seem to be intermediate states such as Hooker speaks of. The largest leaf of S. lagopus measured by me shows the following dimensions: Length of blade, 6 in.; length of petiole, 3 in.; breadth of blade, 4 in. This plant grew on the south-western peak of Mount Herbert, and was exceptionally large. Its measurements equal those of S. saxifragoides in any authority and exceed most of them. A most exceptionally large plant of S. saxifragoides, however, gave the following measurements: Length of blade, 7 ½ in.; length of petiole, 4 ⅛ in.; breadth of blade, 6 ¾in. This is much above the average of the species, the plant being shaded by plants of Linum monogynum and tussock-grass. 6. Thus the leaf of typical S. lagopus is found to bear six different types of structure—(a) The thick brownish “wool” of the rootstock, which covers the base of the petiole and seems to pass gradually into (b) long white silky hairs, which clothe the petiole and are continued up into the sinus and on to the lower portion of the leaf; (c) the characteristic stout bristle which occurs, as described below, on the margin and upon the upper part of the blade especially; (d) the dark-red or purple bristle which is thickly intermixed with the white hairs upon the petiole, from the point where the brownish “wool” passes into white hairs up to the apex, or nearly, on the back of the midrib; (e) the glandular marginal purple protuberances; (f) the white tomentum upon the back of the leaf. 7. Many plants of S. lagopus bear the silky hairs as well as the characteristic stout bristles. The silky hairs usually occur very thickly on the petiole and at the base of the leaf and in the immediate vicinity of the midrib; less thickly, if at all, on the rest of the leaf, as depicted in Raoul's plate. 8. Many plants of S. saxifragoides bear the stout bristles which have been hitherto considered to be characteristic of S. lagopus and S. bellidioides. The bristles occur in S. saxifragoides under these conditions:— (a.) They occur near the apex of the leaf upon the upper surface of about one-fourth or one-third of the whole—not near the base, and but rarely on the lower half of the leaf at all (as in Raoul's plate of S. lagopus), though specimens have been observed with the bristles fairly evenly distributed over the whole surface. (See Plate XII.) (b) They occur regularly and as a permanent character all round the margin exactly as in S. lagopus and S. bellidioides.

Plate XII Young plant of Senecto saxifraffaides, showing bristles, from Mount Pleasant. Port Hills

Plate XIII Plants of senecio lagopus photographed in ate on Mount Sinclair, Banks Peninsula,

(c.) They occur sometimes, not infrequently, near the margin, upon the surface of the leaf, to a distance of about ¼ in. or 1/5 in. from the margin all round the leaf, as in Raoul's plate of S. lagopus. (d.) They occur most frequently upon small, ill-nourished, or depauperated individuals. They occur more frequently upon the lower and outer leaves than upon the upper and inner leaves. Leaves. bearing many bristles have been often found upon plants which are in every other respect typical specimens of S. saxifragoides some whole plants bearing such leaves have been preserved in the collection mentioned below. The stout bristles may be almost certainly observed upon all individual plants which are found growing alone, apart from the large masses in which they generally cluster, such plants being generally small and unfavourably situated as regards aspect of soil. Many leaves of such plants have been preserved in the collection, and in many cases the bristles are to be seen as thickly congregated, as evenly distributed, and as stout as in typical examples of S. lagopus and S. bellidioides. Such leaves are, however, all small, much below the average size of the species, but they are from undoubted examples of S. saxifragoides which are seedlings from neighbouring masses of quite normal specimens. A small plant of S. saxifragoides from Mount Pleasant, Port Hills, which is now under cultivation at Canterbury College, has, its leaves thickly and evenly covered with the stout glandular hairs. here mentioned, and is in no respect to be distinguished from specimens of S. lagopus of the same age. (See Plate XIII.) 9. Microscopical examination of the so-called bristle yielded the following results:— (a.) The bristles are simply typical glandular hairs. (b.) The bristles have exactly the same structure in both species. (c.) The bristles upon the margin of each species differ from those of the blade only in the length of the stalk, except that in the case of S. saxifragoides there is a slight difference in colour. (d.) The “silky hairs” of both species have exactly the same structure. The “hair” arises from a single cell, the “bristle” from several. (e.) The glandular organ on the margins of the leaves is a typical hydathode. (f.) The variegated appearance of the bristle is due to the arrangement of the colouring-matter, which is present in some cells, absent in others, with no definite arrangement. 3. Distribution on Banks Peninsula. For the purpose of this study the forms of Senecio lagopus were observed on all the chief peaks on both sides of Akaroa Harbour, such as Brasenose, Stony Bay Peak, Mount Bossu, and Carew Peak; then on the principal high points on the ridge connecting the Akaroa Harbour heights with the main mass of Mount Herbert, such as Mount Sinclair and Mount Fitzgerald. These points are all between about 2,500 ft. and 2,700 ft. high. Both peaks of Mount Herbert, 3,000ft. and 2,800ft. respectively, were visited on several occasions. Specimen leaves of S. lagopus were collected from various points, especially from Mount Herbeit, and preserved. The forms of Senecio saxifragoides were next observed, and found to occur on all the peaks of the Port Hills from that south-west of Cooper's

Knob to Mount Pleasant, from all of which specimens were collected. These points are mostly from 1,600 ft. to 1,800 ft. high. Senecio lagopus was found to occur plentifully in all favourable localities —that is, especially on and about all steep rocks which face south and south-west—practically continuously from the Akaroa Heads, on both sides of Akaroa Harbour, to within a mile of the Port Hills. There is hardly a space of even two miles in extent anywhere along this main line in which S. lagopus does not occur; it occupies an almost continuous line between the points mentioned, and does not here differ generally from the form of S. lagopus as known elsewhere in New Zealand. The most rema kable point about this distribution concerns the gap of about four miles, a low undulating neck, which connects the south-west peak of Mount Herbert with the Port Hills. This area is mostly under cultivation, and offers but few escarpments. There are, however; three or four possible localities, rhyolite escarpments or peaks, from 600 ft. to 875 ft. high, chiefly on the northern side of the gap. Upon the highest and most likely of these stations, rocks between Gebbie's and McQueen's Valleys, about 875ft. high, neither has been found; but upon two of them (about 600 ft. to 700 ft. high), one just north of the road from Teddington to Gebbie's Valley and the other just north of that, Senecio lagopus occurs —not plentifully, but undoubted typical S. lagopus. Between these two points and Mount Herbert, a distance of about three miles, neither plant is found, the ground being nearly all under cultivation. The northernmost of these points is barely one mile from the nearest peak of the Port Hills, that south of Cooper's Knob (about 1,600 ft. high), and here S. saxifragoides begins to appear. This seems to me a most striking and puzzling fact. Upon the peak next again to the south (on the Port Hills), though it offers an ideal locality, neither plant occurs. We may conjecture that under the original conditions some part of this gap at least was occupied by both species, but at present S. lagopus only is found there, and that only upon the northern side, and upon what is virtually a spur of the Port Hills. Mr. R. M. Laing, however, informs me that he' has collected plants of S. saxifragoides at various points on Banks Peninsula proper, in the neighbourhood of Akaroa, and it has been seen above that both Kirk and Cheeseman give Banks Peninsula as a locality for the species. It may be suggested that states of S. lagopus, “some of which,” as Cheeseman says, “approach it very closely,” have been mistaken for S. saxifragoides. In any case, I cannot say that I have found any examples of undoubted S. saxifragoides, having no bristles at all, on Banks Peninsula proper. On the other hand, no plants of typical S. lagopus were found by me anywhere along the Port Hills, though, as shown above, plants were frequently found in that locality which showed more of the special characters of S. lagopus than have hitherto, apparently, been observed or recorded. Except for a doubtful record in the Kaikoura neighbourhood, Seneciosaxifragoides seems to be confined to this locality. Forms observed by me in April, 1917, on Mount Fyffe and some other points on the Seaward Kaikouras were all typical S. lagopus. The specimens collected from Banks Peninsula and the Port Hills were preserved and fixed in such a manner as to make confusion impossible, and were then submitted while still fresh to Dr. Cockayne, together with certain inferences and conclusions to be drawn from them. While both plants show a preference for high rocky situations and dark, cold faces, neither is by any means restricted to such localities. Both are

found growing among tussock and occasionally on northerly faces, and both descend nearly to sea-level in situations backed by high hills, but apparently not otherwise. It may be conjectured that both were far more widely distributed (within the limits here described) before the advent of white men, for where the ground has been permanently fenced and protected from stock S. lagopus in particular, grows freely at a distance from rocks, especially on steep slopes and those facing south and south-west (as on the old Summit Track, where it rises to the ascent of Mount Sinclair on the south side, and where it runs along the southerly flank of Mount Herbert). Both also are found to grow among grass, & c., at the foot of cliffs and crags; but here, being accessible to stock, they live but precariously, and cannot form the large masses in which they cluster upon the rocks themselves. Thus both species, while now almost purely rupestral in their habit, were probably present in quantities over a very large area where they can now obtain no foothold. 4. Conclusions.* Regarding these conclusions I have consulted Dr. Cockayne, and they owe their present for to his suggestions. 1. Cockayne's surmise in regard to the restriction of the habitat of S. saxifragoides is proved to be absolutely correct, though it is possible that the plants of the Mount Herbert district are intermediate or even hybrid forms. 2. The “bristles” of Hooker's, Raoul's, and subsequent descriptions are simply glandular hairs, and both species bear them, though in varying quantity and differently distributed upon the leaf in the two cases. 3. Both species, in common with S. bellidioides and S. Haastii, have typical hydathodes at the ends of the veins, and both bear purple glandular hairs which are not mentioned in previous descriptions. 4. Neither species has the glandular hair or “bristle” as a distinctive character. The two species differ from one another only in respect of the frequency and locality of occurrence of both glandular hairs (“bristles”) and silky hairs. Those differences in degree, being certainly hereditary, constitute true unit characters. The two kinds of hairs are thus unit characters common to the two species, but the abundance or sparseness of such hairs is a unit character peculiar to either species, as the case may be, and their sole distinction, 5. As the two groups of individuals keep their individuality, each in its isolated, fairly wide area, they are almost certainly microspecies, and they should be grouped together as an aggregate, with saxifragoides as the name of one, group and some other name for the lagopus group. It would seem advisable that S. bellidioides, S. Haastii, and S. southandicus should also be brought into this aggregate, since they have in common with them the woolly rootstock, marginal hydathodes, and glandular hairs upon the surface and the margin of the leaf. 6. Such varieties as the two under consideration, which have every distinguishing character in common, and which differ only in the hereditary degree of intensity, or the distribution of such characters, form a class of varieties (microspecies) different from those which are usually considered such through their possession of one or more quite distinct characters. 7. The question whether the remarkable variation of S. saxifragoides can be explained at all is only approached here with extreme caution and

diffidence. As it is well established that the occurrence of silky hairs in great quantity is a xerophytic phenomenon,* E. Warming, Oecology of Plants, pp. 114, 193, 1909. it might be suggested that this character in S. saxifragoides is of climatic origin. The Port Hills, upon which S. saxifragoides flourishes, are nowhere higher than 1,800 ft., and most of the seven or eight high points upon them are between 1,500 ft. and 1,800 ft. in height; while the Akaroa peaks are on the average about 800 ft. higher than this, and Mount Herbert just exceeds 3,000 ft. As a consequence, the rainfall on Banks Peninsula proper is, and presumably has been for ages, considerably greater than on the Port Hills, the annual rainfall at the Convalescent Home station on the Port Hills being 25–52 in., while that of Akaroa is 44–72 in. There is no geological evidence to show that, since the formation of Lyttelton and Akaroa volcanic areas, Banks Peninsula proper and the Port Hills have not always stood in the same relation to one another as at present in respect of altitude, rainfall, and climate generally, though when the level of the whole was higher than at present, as it once undoubtedly was, the rainfall upon the Port Hills might have been more greatly reduced, relatively, than that upon Banks Peninsula proper. It might thus be argued that the drier climate of the Port Hills has directly determined the development of S. saxifragoides as above outlined. If this were the case we should expect to find similar forms developed in other dry localities, but it is doubtful whether any equally suitable situation exists within the limits of distribution of S. lagopus. If the Port Hills form a unique locality in this respect one could understand how S. saxifragoides has such a narrowly restricted range. Upon this theory S. saxifragoides and S. lagopus would be classed as two varieties of the same plant, differing only in the degree of efficiency reached, under stress, in the development of their xerophytic apparatus. Presumably, also, the other peculiarities of these two plants, such as the woolliness of the rootstock and petiole, might be assigned to the same general cause. Presumably all the six structures above described upon the leaf of S. lagopus, except, perhaps, the marginal glandular structures, would perform a similar function, though attention has here been confined to those two which characterize the upper surface of the blade of the leaf. I desire to express my acknowledgements, first, to Dr. L. Cockayne, F.R.S., to whom I owe the original suggestion of this paper, and without whose kindly encouragement and invaluable aid the work could not possibly have been carried out by me; also to Miss E. M. Herriott, M.A., assistant in the Biological Laboratory, Canterbury University College, who made the microscopical examinations of the various structures and described them (as above) in the most able manner; and, finally, to Mr. R. Speight, M.Sc., F.G.S., Curator of the Canterbury Museum and Lecturer in Geology at Canterbury University College, who supplied the geological history of Banks Peninsula here given, and also very kindly photographed for me the plants of Senecio lagopus on Mount Sinclair (Plate XIII).