The Bryophytes of Stewart Island. Part II By William Martin [Read before the Otayo Branch, September 13, 1949; received by the Editor. October 18, 1949.] Introduction Part I of this paper (1949) dealt exclusively with the northern half of Stewart Island. This was due to inability to gain access to the country further south, which is a mountainous region traversed by a single track along the summits between Table Hill and Port Pegasus, and densely forested on the lower slopes and valleys. In January of this year (1949), through the courtesy of Dr G. J. Williams, Dean of the Otago School of Mines, I was invited to accompany a geological expedition organized by himself to examine the country bordering Port Pegasus in the south of Stewart Island, and the Tin Range somewhat to the north. Three weeks of intensive study and collecting has resulted in a considerable addition to the list of bryophytes indigenous to the island, as well as a considerable amount of new data which forms the basis of this paper. As in the case of Part I, all hepaties were submitted to Mrs E. A. Hodgson for determination and many mosses to Mr G. O. K. Sainsbury, in order to insure accurate identification. For the accuracy of field notes I alone am responsible. It is obvious that in an area of approximately 200 square miles a thorough investigation would require months or even years of study; but by selecting areas typical of the dominant formations and habitats, it was hoped to secure the maximum knowledge possible in the time available. The accompanying map shows the areas investigated by me in this and former expeditions. The vegetation cover of the area comprises forest on the lower slopes merging into scrub on the upper levels and finally into meadow and bog on the summits of the higher tops, though many granite knobs are quite bare of vegetation and form a conspicuous feature of the landscape. Forests of the Pegasus area are distinguished from those in the north by the inclusion of much “yellow-pine” (Dacrydium intermedium) and by the much smaller quantity of rimu (D. cupressinum) and of kamahi (Weinmannia racemos). Full information regarding the forest and other formations is contained in Cockayne's Report (1909). Vandalism in the south has resulted in the destruction by fire of many square miles of forest which to-day is replaced by manuka and other serub formations; but a large area remains almost in its pristine state, modified in some measure by the foolish introduction of deer. The valley between the head of Crooked Reach and the Frazer Peaks is occupied by an area of bog meadow.

Climate of Soouthern Stewart Island No climatic statistics of any kind are available for the south of Stewart Island, and the only semi-permanent inhabitants are the men who operate the fish-freezing plant which formed the base camp for the present expedition. From these men particulars of a general nature were obtained fully in accord with the evidence obtainable from the vegetation itself. Rain is said to fall at frequent intervals, and the number of wholly rainless days probably does not exceed sixty or seventy per annum. The driest period is the first quarter of the year. The annual rainfall is reported to be much greater than at Half-moon Bay in the north and therefore may be placed at possibly 100 inches per annum. On the tops of the hills it doubtless exceeds 150 inches on an average. Overcast skies are the rule over much of the year and fog shrouds the summits for lengthy intervals. Snow, however, rarely falls, only two occasions being recalled in the past five years. Even on the high levels (1,500 ft. and over) snow is rarely seen, and when it does fall, it disappears in a day or two.

Frosts, too, are rare and, according to my informants, are never severe. This is borne out by the growth of frost-tender perennials in the cottage gardens. Winds are frequent and commonly are of gale force, resulting in much wind-shorn vegetation, and prostrate growth in the shrubs, particularly in the sub-alpine belt. The general opinion was that the climate even in winter could be considered as mild, though the summers were never hot. The average noon-day temperature during our stay (January 5 to January 26) was 61° Fah. with a range of only 6° Fah. Soils The whole of the area investigated is composed of granite. Soils derived from this rock are shallow and clay is rarely seen, save at the head of Crooked Reach, where a podsol has developed in limited areas. Most of the soils are composed of peat or organic matter. The average depth of forest soils as disclosed on stream banks or on coastal banks varies from three to six feet. Soils in bog and meadow are mainly peaty also. Comparison of the Bryophytic Floras of Northern and Southern Areas of Stewart Island Mosses, both in number of species and of individuals, exceed liverworts in the north of Stewart Island, whereas the reverse is the case in the south. A study of the following data, though based on figures necessarily incomplete, bears out the impression one gains from a superficial examination, which is in accord with the general rule that with increasing rainfall a point is reached where mosses decrease and liverworts increase in dominance. (a) Total moss species collected in Southern Area 143 (b) Total moss species collected in Northern Area 185 (c) Total hepatic species collected in Southern Area 165 (d) Total hepatic species collected in Northern Area 138 (e) Moss species noted only in Southern Area 12 (f) Moss species noted only in Northern Area 56 (g) Hepatic species noted only in Southern Area 80 (h) Hepatic species noted only in Northern Area 54 (i) Total moss species (+ named varieties) 199 (j) Total hepatic species (+ named varieties) 215 (k) Moss species common to both areas 131 (l) Hepatic species common to both areas 78 From the above figures it will be seen that not only the relative proportions, but also the species content, differs in the two areas to a rather surprising extent, for little more than a third of the hepaties have been collected from both areas. If one ignores the ubiquitus Dicranoloma billardieri, everywhere abundant, it will be found that in point of individuals, hepatics far outnumber mosses in southern Stewart Island. Ferns. too, and particularly Treeferns, are much less numerous in southern fortests than in those north of Paterson's Inlet.

Again, numerous plants normally subalpine are met with at sealevel in the Pegasus area. Cockayne (1909) has already drawn attention to this in respect of the flowering plants, but the same observation holds for the bryophytes also. Examples among the mosses include Andreaea subulata, A. acutifolia (?), Blindia tenuifolia, Breutelia elongata. Oligotrichum tenuirostre, and Rhacocarpus australis. It is worthy of remark that the first-named moss is here commonly an aquatic plant, and that at Cedric Creek in the North Arm of Paterson's Inlet it occurs on the lip of a waterfall in dense forest, stations quite unusual for this species in the South Island. It is not uncommon at Pegasus to find plants which normally grow as epiphytes, thriving on the peaty soil of the forest floor. Among flowering plants this was noted in the case of Dendrobium cunninghamii and Earina autumnalis: amongst ferns in the case of Asplenium flaccidum. Hymenophyllum revolutum, and Rumohra adiantiformis; amongst mosses in Dicnemon calycinum, Holomitrium perichaetiale, Isopterygium limatum, Leptostomum inclinans, Rhizogonium novae hollandiae, and Stereodon chrysogaster. Much the most outstanding difference between the bryophytes of southern and northern areas, however, is the development in the south of the mound or cushion form in both hepatics and mosses. Over large areas these mounds cover the forest floor and impart a unique character to the forest interior, making progress both slow and difficult. They are rarely if ever met with in forests north of Paterson's Inlet. They vary in size from one to three feet in width and are often as much as 18–24 inches high—in individual cases they may be three feet high. These cushions are not confined to the earthen floor of the forest, but grow also on stumps, logs, and low-growing branches which are frequently quite concealed by them. Contiguous mounds commonly merge, making it difficult to count the precise number in any given area; but two counts in areas twenty feet square gave 110 and 86 respectively. At the low estimate of 50 for this area, the number of mounds would work out at 5,400 per acre. Many of the mounds are composite in character, but in general the mosses and hepatics that develop the mound form are identical with those enumerated in Part I of this paper (p. 260) and need not be repeated. Some moss species common in the north which have not yet been detected in southern areas are Bellia nervosa, Cryphaea tenella. Cyrtopus setosus, Camptochaete arbuscula. Drepanocladus aduncus, Eurhynchium austrinum, E. prolongum, Hypopterygium setigerum, Tetraphidopsis pusilla, and Trachyloma planifolium; while others abundant in the north are conspicuously rarer in the south. This is particularly so of Dicranoloma menziesii, Breutelia pendula, Orthorrhynchium elegans, and Eriopus cristatus. On the other hand, several mosses appear to be more common in the southern area. This is true of Holomitrium perichaetiale, Macromitrium longirostre, and Schlotheimia campbelliana; but conspicuously so in the case

of Blindiopsis immersa, of which I have numerous gatherings from southern stream-beds, but none from north of Paterson's Inlet, save some stunted specimens on the rocky face of the waterfalls near the mouth of Cedric Creek and Martin's Creek arising in the Thomson Range and entering the North Arm of Paterson's Inlet. It will require much further research, however, to confirm the absence of a moss from any of the areas, and my conclusions must of necessity be tentative only. The hepatic Monoclea forsteri, common in the north, was nowhere observed in the south. Bryophyte Associations A. Forest (a) Forest Streams Moss species normally growing on submerged rocks or in the silty bed of forest streams are identical with those in the north, viz., Blindiopsis immersa, Fissidens rigidulus, and Thamnium pandum. The first-named moss is very common in southern streams, especially in Pegasus Creek and its tributaries, where it frequently acquires a surprising luxuriance. Plants fourteen inches (35 cm.) long were obtained, with as many as six capsules in a single perichaetium. The discovery of operculate capsules enabled me to confirm its gymnostomous character and the correctness of its generic determination. Hepatics that can thrive while more or less perpetually submerged appear to be few, and include Jamesoniella sonderi, Lophocolea notophylla, and two undetermined species of Riccardia. Not a few bryophytes, however, grow on rocks in the stream-bed or on the silty margin of streams, in positions liable to frequent submergence. Most unexpected among such mosses was Andreaea subulata, which normally grows neither in forest, at sea-level, nor in aquatic stations. Most of the others were identical with those enumerated for northern stream-beds, but Eriopus cristatus was seen but once, and Camptochacte gracilis and Thuidium laeviusculum very rarely. The hepatic flora of emergent rocks, however, shows considerable divergence from those enumerated for similar stations to the north. All these species must be capable of enduring frequent submergence. My list includes Acromastigum colensoanum, Bazzania novae-zelandiae, Balantiopsis aequiloba, Hymenophytum leptopodum (form approaching H. flabellatum), H. phyllanthus, Jamesoniella sonderi, Lepicolea scolopendra, L. ochroleuca, Lepidozia kirkii, L. praenitens, L. spinosissima, Lophocolea notophylla, L. leucophylla, Plagiochila gigantea, Riccardia spp., and Tylimanthus saccatus. The banks of forest streams in the Pegasus area are peopled by numerous bryophytes, but none were observed to be restricted to this substratum. (b) Forest Floor The outstanding feature of the forest floor in the wetter areas is the great development of mounds and hummocks both of mosses and hepatics, which, in many areas, are so numerous and close, that progress is rendered very difficult. The species involved are numerous (see list of mound-forming species in Part I, p. 260), but much the

commonest of all is Dicranoloma billardieri. In general the cushions are more or less composite. The moss D. menziesii in areas further north frequently forms balls, 2–4 inches in diameter and wholly detached from any substratum. Indeed, the initial point of attachment becomes central and cannot be seen from the outside at all. How these moss balls originate is difficult to determine, but they possibly form brackets on tree trunks which on becoming detached continue to thrive and take a spherical form. Such brackets have frequently been observed by me in the case both of D. menziesii and Isopterygium limatum. In many areas fully 50% of the forest floor is carpeted with bryophytes, hepatics exceeding mosses in number of species and of individuals. The most conspicuous hepatics in areas where mound formation is rare or absent are the numerous species of Schistochila, most of them with white, cream, or yellow branches, of which the commonest species are S. appendiculata, S. glaucescens, S. ciliata, and S. nobilis. The usual species of Plagiochila found on the forest floor in this region are P. ramosissima, P. lyallii, and in local areas P. sinclairii. Several undetermined species of Riccardia and a number of species of Lepidozia though perhaps commoner on logs, are common also on the earthen floor. Such include L. ulothrix (= L. albula of Part I), L. kirkii, and locally L. pulcherrima. A number of additional hepatics belong to the forest floor, but occur only in well-lighted stations such as on the forest margin or on bush tracks. Such include Acromastigum anisostomum, Anastrophyllum schismoides, Balantiopsis aequiloba, B. rosea, B. tumida, Isotachis montana, Isotachis lyallii, and Lepicolea ochroleuca. Chiloscyphus billardieri and in lesser amount C. coalitus are also to be noted both on logs and on the forest floor. Lembidium nutans, L. tenax, and the new species L. insulanum are likewise found equally on earth and on logs and even on rocks over which water is flowing. The commonest mosses are Dicranoloma billardieri, D. integrifolium?, D. platycaulon, and D. menziesii, Mniodendron comosum, Leucobryum candidum, and in wetter areas Pterygophyllum quadrifarium, Hypopterygium novae-seelandiae, Sciadocladus menziesii, Tayloria purpurascens, and Hypnodendron arcuatum. Acanthocladium extenuatum, Acrocladium auriculatum, Thuidium spp., Hypopterygium setigerum, all common in the north, are either rare or absent in the south. Attention has already been drawn to the presence on the forest floor of not a few mosses normally epiphytic. (c) Stumps and Logs Almost all the bryophytes of the forest floor are to be found on logs and stumps, as well as many plants normally epiphytic. On the other hand, many bryophytes rarely found either on the ground or on tree-trunks are to be found on these logs, and not infrequently on horizontal branches as well. Typical hepatics of this station include Balantiopsis tumida, Bazzania adnexa, B. reflexistipula, B. tayloriana, Chiloscyphus billardieri, C. allodontus, Cuspidatula monodon, Goebeliella cornigera, Lepidozia kirkii, L. praenitens, Lophocolea lenta,

Plagiochila conjugata, P. pleurota, P. lyallii, P. ramosissima, Radula uvifera, Riccardia alterniloba, R. eriocaulis, Schistochila balfouriana, S. ciliata, and S. glaucescens. The rare Dendroceros allisonii and Metzgeria violacea also belong here. Amongst the mosses Dicranoloma billardieri stands first in abundance, but very common associates are Stereodon chrysogaster, Hypopterygium novae-seelandiae, Rhizogonium spp., Pterygophyllum spp., Ptychomnion aciculare, Mniodendron comosum, Sematophyllum amoenum. Much less common in the southern forests than in the north are Eurhynchium muriculatum, Lembophyllum clandestinum, Leptotheca gaudichaudii, Rhacopilum strumiferum, Rhynchostegium tenuifolium, Sciadocladus menziesii, and Thuidium furfurosum. (d) Tree-ferns Epiphytes on the caudices of tree-ferns (mainly Dicksonia squarrosa) are usually very different from those near Half-moon Bay in the north. Thus Polyphlebium venosum is quite uncommon at Pegasus and Sphaerocionium ferrugineum, so abundant north of Paterson's Inlet, was nowhere seen by me in the south. Mecodium rarum and Tmesipteris tannenis are also seen much more rarely. Mosses of this substratum at Pegasus include Calomnion laetum and Rhizogonium pennatum var. aristatum, neither of them anything like so common in the north. Other mosses noted included Cyathophorum bulbosum var. minus, Dicranoloma menziesii, and its variety rigidum, likewise Goniobryum subbasilare. Hepatics included some not found elsewhere in Stewart Island. e.g. Lepidozia roseana, L. lindenbergii, Telaranea bisetula, and Zoopsis argentea. The more common hepatic epiphytes of tree-fern caudices comprise Bazzania monilinerve, B. reflexistipula, B. taylorianum, Chiloscyphus echinellus, Hymenophytum phyllanthus, Lembidium tenax, Plagiochila deltoidea, Schistochila appendiculata, and S. ciliata. More than one species of Metzgeria is common on this substratum. (e) Bark of Forest Trees Epiphytic bryophytes in the southern forests include many not seen in the north and omit others which are there common. Of mosses not observed in the north of the island, Sauloma tenella and a species of Orthotrichum new to New Zealand and possibly to science, found on rimu on Rosa Island in Pegasus Harbour may be mentioned. Epiphytic hepatics seldom seen elsewhere on Stewart Island include Chiloscyphus circumdentata (a new species), several species of Frullania (F. aterrima on rimu, F. setchellii on yellow pine, F. subdeplanata on Fuchsia excorticata, and F. rostellata on a forest log), Marsupidium epiphytum, Metzgeria nitida, Lepicolea ochroleuca (more common on logs and on the earthen forest floor), two species of Lophocolea (L. notophylla and L. subporosa), and the recently discovered Plagiochila simpsonii herein described. On horizontal branches the rare hepatics Diplophyllum densifolium and Lepidozia pulcherrima were obtained, the latter more usually on the ground. In the Pegasus forests the commonest epiphytic mosses are Leptostomum inclinans, Holomitrium perichaetiale, Macromitrium erosulum, M. gracilie vars. proboscideum and retusum, M. longirostre, Dicnemon calycinum, Schlotheimia campbelliana, and Stereodon chrysogaster,

Common mosses in forests on the north coast nowhere seen in the south include Bellia nervosa, Cryphaea tenella, Cyrtopus setosus, Neckera hymenodonta, Tetraphidopsis pusilla, and Trachyloma planifolium. Other common mosses very much rarer in the Pegasus area are Cladomnion ericoides, Calyptopogon mnioides, Macromitrium caducipilum, Orthorrhynchium elegans, Ptychomnion ericoides, Thuidium furfurosum, and Zygodon intermedius. The more common hepaties observed growing as epiphytes were: Bazzania novae zelandiae, Chiloscyphus menziesii, C. laxus, Frullania aterrima, F. ptychantha, F. rostrata, F. squarrosula, Lepidolaena clavigera and its coloured variety var. taylori, Plagiochila annotina, P. deltoidea, P. lyallii, P. pleurota, P. ramosissima, Radula uvifera, and Riccardia eriocaulis. The numerous species of Frullania with few exceptions prefer well-lighted stations. F. rostrata and F. squarrosula are often found on stems exposed to full daylight in bleak, wind-swept situations. Schistochila tuloides and S. pinnatifolia are epiphytic species of a genus most commonly found on the forest floor or on prostrate logs. Trichocolea lanata and T. tomentella are other common hepaties of the forest floor which are sometimes met with also as epiphytes, while the moss Pterygophyllum quadrifarium. normally a ground moss, has occasionally been found on tree trunks as much as ten feet above the ground. Additional epiphytic hepatics include: Adelanthus magellanicus, Drepanolejeunea latitans, Frullania setchellii, Lejeunea tumida, Lophocolea muricata, L. subporosa, Plagiochila radiculosa, P. retrospectans var. novae-zelandiae, P. simpsonii, Porella elegantula, and P. stangeri. Siphonolejeunea nudipes was recently detected as an epiphyte on low trees near Half-moon Bay in the north, by Mr Cedric Smith. (B) Rocks (a) Lowland Rocks True lithophytes are very rare on the granite rocks of the Pegasus region, but are rather more common on the schistose rocks on the summit of the Tin Range. Other than strictly coastal rocks, few low-altitude outcrops were observed. Macromitrium longirostre and Rhacomitrium crispulum were in fact the mosses most commonly observed by me on coastal rocks, though Lembophyllum clandestinum, Eriopus apiculata, and Zygodon menziesii were also found on sand-covered rocks in coastal stations. The Campbell Island moss Muelleriella crassifolium was obtained on rock at Wilson Bay and on the margin of a shallow rock pool on the flat between Crooked Reach and the Frazer Peaks, where, though my notes do not say so, it was doubtless attached to the granite rather than to the peat veneer. This is the second finding of this moss in New Zealand, Mr K. W. Allison having first obtained it at Akatore on the Otago Coast two years ago. On the rocks forming the lip and margin of the Pegasus Creek Falls. Blindia tenuifolia, Andreaea subulata, and Sematophyllum tenuirostre grow in quantity as well as the following hepatics: Jamesoniella sonderi, Balantiopsis aequiloba, Hymenophytum leptopodum, Lepidozia praenitens var. minor, Pachyglossa sp., and Riccardia sp. At Lord's River Frullania deplanata, and on an island

in Sawmillers' Arm Lophocolea novae-zelandiae were growing as lithophytes. Ditrichum brevirostrum grows on rocks at the terminus of the tramline on the Tin Range. (b) Subalpine Rocks On the summit rocks of the Tin Range, the following mosses were growing on a rock substratum: Andreaea acuminata (?), A. subulata, Blindia tenuifolia, Dicranoloma billardieri form integra (= D. pungentella), D. billardieri, D. integerrimum, D. sp., Dicranum trichopodum (?), Holomitrium perichaetiale, Leptostomum inclinans, Rhizogonium novae-hollandiae, Sematophyllum amoenum, and S. tenuirostre. The hepatics gathered on these summit rocks were: Herberta alpina, Jamesoniella pseudo-occlusa, J. sonderi, Lepicolea ochroleuca, L. scolopendra, Lepidozia patentissima, L. pulcherrima, L. ulothrix, Metzgeria nitida, Plagiochila deltoidea, P. ramosissima, and Pycnolejeunea zotovii. On granite rocks near Magog, one of the Frazer Peaks two species of Andreaea [A. rupestris (?) and A. acutifolia (?)] were the only lithophytes observed. Both, however, require further study. (C) Open Country There are few soils in the Pegasus area not peaty in character, and the majority of non-forested soils may be classed as bog. It is difficult to draw the dividing line between true bog and other soils; but, in general, the hilly slopes bordering the valley at the head of Crooked Reach are scarcely such as could be termed bog. All soils are, however, almost perpetually moist. On such soils in this area, a number of rare and interesting hepatics were discovered which may be enumerated as follows: Acromastigum anisostomum, A. colensoanum, A. martinii, Acrobolbus cinerascens, Lepidozia gottscheana, L. pulcherrima, L. setigera, L. herzogii, and Radula dentata. Of these, L. pulcherrima has not previously been collected save at Okarito, while A. martinii Hodgs. is a new discovery. One of the commonest hepatics here is Lepicolea ochroleuca. No mosses were detected as peculiar to this substratum, though Dicranoloma spp. were present, some of unusual form, and Sphagnum antarcticum and S. novo-zelandicum are everywhere common. In clearings near the “Freezer,” the ubiquitous mosses Funaria hygrometrica and Ceratodon purpureus were plentiful and likewise Bryum truncorum, B. dichotomum, as well as Marchantia tabularis. Campylopus introflexus is abundant in clearings elsewhere. Oligotrichum tenuirostre and Campylopodium euphorocladum grow abundantly on a slip south of the settlement, in association with Balantiopsis rosea. (D) Banks Certain bryophytes may almost be regarded as obligate to vertical banks, as Ditrichum flexifolium; but the majority of bryophytes found there at Port Pegasus are by no means restricted thereto. (a) Coastal Banks Omitting such as consist of bare rock, the only areas investigated were peaty banks on Rosa Island, at the “Freezer,” and at Fright

Cove in the South Arm. The only bryophytes found in this station only at Pegasus were Eriopus apiculatus, Orthodontium sulcatum, and Symphyomitra drummondii, but Ditrichum flexifolium, D. punctulatum, Rhacopilum strumiferum, and Rhizogonium novae-hollandiae were found there most commonly. Other mosses of this substratum were Campylopus clavatus and C. introflexus, while the following hepatics were listed: Balantiopsis convexiuscula, Bazzania involuta, Chandonanthus squarrosus, Chiloscyphus beckettianus, Cuspidatula monodon, Frullania rostrata, Isotachis lyallii, I. montana, Lophocolea subporosa, Plagiochila conjugata, P. radiculosa, and P. gigantea (one gathering). (b) Inland and Upland Banks This substratum on the Tin Range is usually weathered rock almost converted to clay. As a result, the vegetative cover is distinct from that on the peaty coastal banks. Here the lichens Baeomyces, Cladonia, and Stereocaulon are well represented and numerous seedlings of the nearby shrubs find anchorage. The commonest mosses on the bank margining the tramline on the Tin Range were Polytrichadelphus magellanicus and Ditrichum punctulatum with Campylopus clavatus common in places. Breutelia pendula, an undetermined Bryoid moss, Bryum obconicum, and a single clump of Bartramia norvegica were also listed. Hepatics from this station were Anastrophyllum schismoides, Bazzania novae-zelandiae, Balantiopsis tumida, Chiloscyphus billardieri, Isotachis lyallii, Lepidozia pendulina, Lembidium tenax, Radula dentata, Riccardia spp., one bronze and the other deep green, and Schistochila pinnatifolia, a plant more usually epiphytic. (E) Bog The only extensive areas of bog in the Pegasus area occur on the summit of the Tin Range and in the valley between Crooked Reach and the Frazer Peaks. The former are subalpine, the latter lowland; but there is little difference in either the phanerogamic or cryptogamic floras save that the upland bogs are somewhat richer in species. Species of moss not noted on bogs in the Frazer Peak area include Blindia tenuifolia, Holomitrium perichaetiale, and Breutelia elongata, and hepatics missing from the same area included Jamesoniella pseudoocclusa. Acromastigum colensoanum, A. martinii, and Balantiopsis rosea on the other hand were not observed on the Tin Range bogs, but the search was not extensive or intensive enough to dogmatise on these differences. The commonest bog mosses are Dicranoloma billardieri in a variety of forms, some quite remarkable, D. integrifolium, D. sp., Campylopus bicolor, C. introflexus, Sphagnum antarcticum, and S. novo-zelandicum. Rather less common were Breutelia elongata, Blindia tenuifolia, Holomitrium perichaetiale, Polytrichum commune, Rhacocarpus australis, and Rhacomitrium lanuginosum var. pruinosum. The following list comprises some of the commonest hepatics: Acromastigum colensoanum, Balantiopsis rosea, Isotachis montana, Lepidozia sp., L. ulothrix, Lepicolea ochroleuca, Cuspidatula monodon,

Jamesoniella pseudo-occlusa, J. sonderi, and Radula dentata. Notes on Individual Species (Contributed by E. Amy Hodgson) 445. Lepidolaena stangeri (G.) Martin and Hodgson. As far as can be ascertained, it is this species that in the Synopsis Hepaticarum is listed as var. stangeri of Polyotus clavigera (Hook.) G., which is a synonym for L. clavigera (Hook.) Dum. It differs from L. clavigera in its smaller size, glaucous colour, dentate-serrate branch leaves, and papillose lobules. If the original var. stangeri ever proves to be something different, then this species must be given a new name. 461. Bazzania quadrata Col. This is one of the many forms of an aggregate species with longish narrowed leaves and long apical teeth sometimes reduced to only 2. The present plant has leaves flatter than in the type, which appears to come very close to the Australian B. accreta, and may be a form of that species. 425. Lepidolaena clavigera (Hook.) Dum., var. taylori (G.) Mitt. In Taylor's herbarium this was labelled merely Jungermannia clavigera Hook. Gottsche made of it a separate species, Polyotus taylori G., which Mitten reduced to a variety of P. clavigera (Hook.) G. I agree with Mitten that it is a variety of L. clavigera (Hook.) Dum., somewhere intermediate between L. clavigera and L. palpebrifolia (Hook.) Dum. 479. Hymenophytum flabellatum (Hook.) St. (approaching). Stephani defines the species as with more than 2 primary divisions to the fan, stalk not at all winged, midrib lost below the apex and the involucre spinous. 420. Plagiochila annotina (Menz.) Ldbg. With sub-entire to entire margins, and strongly pigmented, this form appears to bear relationship to P. circinalis, or perhaps more likely it represents a separate species. 467, etc. Lepidozia pulcherrima St. Previously reported only from Okarito. 536. Lophocolea notophylla (Tayl.) comb. nov., Syn. Lophocolea okaritana St. Correctly identified in Kirk's collection No. 4863 as Jung. notophylla Tayl. 391. Radula unifera (Tayl.) Syn. Hep. Perianths in this specimen, though decurved (in some cases), are not ribbed as is usual in this species. 513. Lepidolaena clavigera (Hook.) Dum. With its purplish-brown colouring, elongate stems, shortly and regularly tri-pinnate, this specimen must come very close to L. brachyclada (L. and L.) St., which is recorded from New Zealand, but in all the descriptions is said to have ciliate stem leaves. 569 and 597. Metzgeria nitida Mitt. Evans (1923) reduced M. nitida Mitt. to M. hamata Lindb., into which it certainly merges, but with its flat margins, few, straight cilia, and sometimes very large cells, it looks very different; so different, in fact, that Stephani considered reducing M. decipiens (Massal.) Schiff. to its synonymy. 495. Metzgeria violacea (Ach.) Dum. In a recent publication,* Svensk Botanisk Tidskrift, Bd. 42, H. 3, 1948. we

reduced this to a variety of M. decipiens, but the name of M. violacca antedates that of M. decipiens by 85 years, and therefore must be retained. 623. Bazzania monilinerve (L. and L.) Syn. Hep. This specimen is very distinct in its well-defined vitta of large cells, as in Gottsche and Lindenberg's plate, also in the narrowed, 3-spinous (spines diverging) apex, which is characteristic of B. accreta, as also figured in the author's plate. A note by the authors which may be significant is that both B. accreta and B. monilinerve were in the same clump. They are Australian species. The Handbook accredits B. monilinerve to New Zealand, but Stephani does not. 718. Radula sp. cum fr. This fruiting specimen which seems to be a bigger form of the one thought to be R. sainsburiana (Martin in Trans. Roy. Soc., 1949, 273) is not paroicous, and must be considered an undescribed species. 720. Adelanthus falcatus (Hook.) Mitt. forma non-falcata. A. magellanicus (Lindenb.) Spr. given as from Campbell Is. and New Zealand by Stephani, has entire leaves according to the Synopsis description, though Stephani does not mention this. A. falcata has toothed leaves as in this specimen. 721. Porella stangeri (Syn. Hep.) Hodgson. If Mitten and Stephani are correct in considering Madotheca partita (Tayl.) Syn. Hep. as conspecific with M. stangeri Syn. Hep., then the latter name must give way, as Taylor's name was published first. 663, etc. Frullania setchellii Pearson = F. falciloba var. setchellii (Pears.) Hodgson. On further consideration, I would like to take this opportunity of restoring F. setchelli Pears. to specific rank. 457. Frullania scandens Mont. (Syn. F. subdeplanata St.). Stephani's description of F. scandens is at variance with Montagne's. 546. Probably the most interesting association in this remarkable collection is this combination, with others, of the North Auckland Radula dentata, a 3-fid-leaved specimen of Lepidozia gottscheana, Telaranea biselula, and a new species of Acromastigum, to which I have given the MS, name of Acromastigum martinii sp. nov. Descriptions of New Species Chiloscyphus circumdentatus Martin and Hodgson sp. nov. Planta sterilis, pulvinata cum aliis hepaticis, pallide brunnescente-viridis. Caulis ea. 1·5 cm., irregulariter ramosus, parum flexuosus. Folia caulinis opposita, imbricata, secunda, quadrato-rotundata vel latissime ovata, marginibus spinoso-dentatis, apicibus duobus spinis longioribus. Amphigastria imbricata, convexa, reniformia, dentata, utrinque cum foliis connata. Cellulae porosae, inaequales, ca. 20–35μ, parietibus interdum nodulosis, trigonis magnis. This species has all the appearance of a spinose C. menziesii (Syn. C. bidentatus), but the stipules, instead of being small, ovate and ± bidentate, are reniform and toothed as in C. billardieri. Although the latter is very variable, this new species is of smaller size, and has leaves too rounded and flat-margined to belong there.

Moss mounds in forest near mouth of Pegasus Creek, Stewart Island. Photo, Wm. Martin. Mounds of mosses and hepatics at 1,000 ft. alt., Pryse's Peak, Stewart Island. Photo, Wm. Martin.

On trees and logs in forest, Port Pegasus, 9/1/49, W. Martin 453; forming part of small cushion on limbs of trees and on logs, 1,000 ft., Pryse's Peak, 15/2/47, W. Martin A.460. Chiloscyphus erraticus Martin and Hodgson sp. nov. Planta sterilis, flavo- bis brunneo-viridis, vage pulvinata. Caulines ad 6 cm., simplices vel parum ramosi, laxe intricati, cum aliis hepaticis mixti. Folia imbricata, opposita, 1–1·5 mm., ovata, rotudato- vel oblongo-ovata, margine ventrale arcuatiore quam dorsale, plus minus undulato (siccato), apicibus multo variabilus, obtusis, cuspidatis, bi- vel unidentato-truncatis. Amphigastria parva, variabilia, ovato-quadrata, 2–4 denticulato-truncata. Cellulae rotundatae, ca. 30μ, trigonis parvis. This species resembles C. beckettianus St. in its quadrate and connate stipules and opposite leaves, but the leaf-apices are very variable as in C. compactus Col. (Syn. Leioscyphus irregularis St.). It may be a cross between the 2 species, but its discovery in 2 widely separated localities, together with its robust size, point to its being a good species. Part of a composite bryophyte cushion 1,000 ft., Pryse's Peak, 15/2/47. W. Martin A 462; near forest margin, 2,000 ft., Table Hill, W. Martin, 5/2/47; boggy ground, Table Hill top, Feb. 1947; bed of tributary of Pegasus Creek in dense forest, 5/1/49, W. Martin 481. Lembidium insulanum Martin and Hodgson sp. nov. Planta dioica, mediocre robusta, dilute virens, dendroidea, caespitosa. Caulis 3 cm., erectus, ramosus, e basi flagellis numerosis, primariis ramis caulis similibus, superne pinnatim ramulosis, incurvatis et apice convolutis. Folia caulinis et ramorum principalium ca. 1·2 mm., laxe imbricata, verticalia, ovata, hyalina et fragilia, appressa, apice brevissime 2–4 dentato. Amphigastria similia. Cellulae medianae ovaliae vel rectangulariae, ca. 45 x 25μ, basales elongatae. Folia ramulorum ovata, ca. 0·5 mm., imbricata, concava, patentia, apicibus breve 2–4–dentatis. Cellulae rotundatae, marginales ca. 15μ, medianae ca. 25μ. Amphigastria minora, simillia. Androecia numerosa, in ramula parva, curvatim patula, vermiculata, pallentia, bractiis imbricatis, concavis, breve bidentatis, monandris. Cellulae laxae, hyalinae. This notable new plant differs from L. tenax (Grev.) St. in the absence of cilia at the bases of leaves and stipules, and in the shortly dentate leaf apices. It is possible that these 2 with their dendroid habit may constitute a new genus. On logs and on ground near waterfall, Pegasus Creek, 9/1/49, 568, type; on ground, forest floor, Tin Range, 14/1/49, 543; subalpine forest, Table Hill, 5/2/47, A 512; on log in forested area, head of Crooked Reach, Port Pegasus, 21/1/49, 657, all coll. W. Martin. Plagiochila simpsonii Martin and Hodgson (sp. nov.) Planta parva dioica. Caulis flexuosus, 1–1·5 cm., simplex vel parum ramosus, apice parum recurvo. Rami interdum flagelliformes. Folia canlina oblique obovato-quadrata, patula vel appressa et secunda, normaliter biloba; lobo dorsali angusto, curvato, interdum acuminato,

sinu oblique lunato; margine ventrali areuato, lobato-serrato; margine dorsali decurrente, curvato, plano. Folia ramulina, patula, biloba, lobis sub-aequalibus, margine dorsali sub-stricto, ventrali parum arcuato. Cellulae formis variis, marginales ca. 18μ, medianae majores, ea. 35μ, trigonis confluentibus. Folia floralia caulinus, majora, simillima. Perianthium oblongum, ca. 2 mm. x 1 mm., ore dentato. Plants small, olive to dark green. Stem flexuous, 1–1·5 cm., or with few branches, curved backwards a little at the apex. Branches varying in size to flagelliform. Cauline leaves spreading, or imbricate and dorsally second, light green at the extremity of the stem, gradually increasing in size upwards, obliquely obovate-quadrate, normally bilobed; the dorsal or antical lobe narrow, curved, in a continuous line with the dorsal margin of the leaf; the ventral or postical lobe made up of several secondary lobes, or with marginal teeth or sub-entire; dorsal margin decurrent, flat, curved; ventral margin arched (in outline). The branch leaves and those on the lesser stems, contiguous to remote, bi-lobed to nearly ½, strongly resembling those of Sphenolobus perigonalis in both shape and direction; dorsal margin nearly straight, ventral a little arched. Areolation distinct with cells all separate, cell-cavities clear or more usually with a characteristically pitted appearance. Basal cells clongate with sinuous walls. Invol. leaves resembling the cauline, increasing in size to the uppermost pair. Perianth narrow-oblong to oblong. Head of L. Manapouri to Willmott Pass, G. Simpson, Dec.-Jan. 1945–46, H 841 Herb. K. W. Allison, type, Preservation Inlet (Long Sound), H 156 Fiordland Excursion collection. Epiphytic on forest trees, Pegasus, Stewart Is., 443, 454, 678, 696, 9/1/49; epiphyte, Pryse's Peak, Stewart Is., Feb. 1947, all coll. W. Martin. The flat appressed leaves receding from the dorsal side of the stem, with the curved leaf-lobes, give this species a very distinctive look. The above description is from a manuscript list of hepatics collected by members of the Fiordland Excursion, and is here published with the kind permission of Dr. H. H. Allan. Stewart Island Mosses Not Hitherto Recorded 1. Andreaea subulata Harv. 2. Bartramia norvegica (Gunn.) Lindb. 3. Bryum obconicum Hornsch. 4. Bryum pachytheca C. M. 5. Calomnion leatum H. f. & W. 6. Camptochactc ramulosa (Mitt.) Jaeg. 7. Campylopodium euphorocladum (C. M.) Besch. 8. Ditrichum brevirostrum (R. Br. ter.) Broth. 9. Drepanocladus aduncus (Hedw.) Moenk. 10. Fissidens campyloncurus (Hedw.) Moenk. 11. Fissidens tenellus H. f. & W. 12. Homalia pulchella H. f. & W. 13. Muelleriella crassifolia (H. f. & W.) Dus. 14. Oligotrichum tenuirostre (Hk.) Jaeg. 15. Orthotrichum beckettii C. M. 16. — sp. nov. (?) 17. Sauloma tenella (H. f. & W.) Mitt. Previous records now confirmed are Andreaea rupestris, Echinodium hispidum, and Rhizogonium pennatum var. aristatum. The moss Orthotrichum beckettii was first detected by Mr. Cedric Smith, also the hepatic Strepsilejeunea nudipes.

Stewart Island Hepatics Not Hitherto Recorded 1. Acrobolbus cinerascens (L. & L.) St. 2. Acromastigum martinii Hodgs. nom. nud. 3. Adelanthus magellanicus (Lindenb.) Spr. (probably) 4. Balantiopsis aequiloba f. subintegra 5. — convexiuscula Berggr. 6. — rosea Berggr. 7. Bazzania elegans (Col.) Mart. et Hodgs. 8. — quadrata (Col.) Mart. et Hodgs. 9. — reflexistipula. (L. & G.) 10. Blepharostoma pulchella (Hook.) St. 11. Chiloscyphus allodontus (Tayl.) Hodgs. 12. — circumdentata Mart, et Hodgs., sp. nov. 14. — triacanthus (Tayl.) St. 15. Diplophyllum densifolium (Hook.) St. 16. Dendroceros allisonii Herz. 17. Drepanolejeunea latitans (Tayl.) St. 18. Frullania aterrima. (Tayl.) Sp. Hep. 19. — fugax var. media Hodgs. 20. — reptans Matt. 21. — rostellata Matt. 22. — scandens Mont. 23. — setchellii Pears. 25. Geocalyx novae-zelandiae Herz. 26. Isotachis lyallii Mitt. 27. Lejeunea implexicaulis (Tayl.) Syn. Hep. 28. — tumida Mitt. (probably) 29. Lembidium insulanum Mart. et Hodgs., sp. nov. 30. — nutans (Tayl.) Mitt. 31. Lepicolea ochroleuca (Spreng.) Spr. 32. Lepidolaena clavigera var. taylori 33. — stangeri (G.) Mart. et Hodgs. 34. Lepidozia allisonii Herz. 35. — compacta St. 36. — dispar Mont. 37. — gottscheana. Ldbg. 38. — herzogii nom. nov. 39. — hirta St. 40. — kirkii St. 41. — procera Mitt. (?) 42. — pulcherrima St. 43. — pendulina (Hk.) Sp. Hep. 44. — roseana St. 45. — setigera, St. 46. — ulothrix (Schwaegr.) Syn. Hep. 47. — sp. nov. (?) 48. Lophocolea aculeata (Mitt.) comb. nov. 49. — dalliana St. 50. — lenta (Tayl.) Syn. Hep. 51. — notophylla (Tayl.) Mart. et Hodgs. 52. — sp. nov. 53. Marchantia berteroana L. & L. 54. Marsupidium abbreviatum (Tayl.) St 55. — epiphytum Col. 56. — piliferum St. 57. Megaceros sp. 58. Melzgeria decipiens (Massal.) Schiff. 59. — nitida Mitt. 60. — violacea (Ach.) Dum. 61. — sp. 62. Plagiochila radiculosa Mitt. 63. — retrospectans var. novae-zelandiae 64. — simpsonii Mart. et. Hodgs., 65. — sp. nov. (?) [n.s. 66. Porella elegantula (Mont.) Hodgs. 67. Psiloclada clandestina. Mitt. 68. Pycnolejeunea glauca St. 69. — zotovii Herz. (probably) 70. Radula dentata Mitt. 71. — sp. nov. (near R. sainsburiana) 72. Riccardia colensoi (St.) Mart. 73. — minuta (St.) Mart. 74. — sp. 75. Schistochila tuloides (Mitt.) St. 76. Siphonolejeunea nudipes (Tayl.) Herz. 77. Strepsilejeunea comitans (Tayl.) St. 78. Symphyomitra drummondii (Mitt.) St. 79. Treubia insignis Goeb. (probably) 80. Zoopsis argentea (Tayl.) Hook. Stewart Island Hepatics Previously Listed and Now Confirmed by Me Bazzania monilinerve, Chiloscyphus decipiens, C. echinellus, C. fissistipus, C. laxus, C. normalis, Lepidolaena magellanica, L. stangeri, Lophocolea muricata, Plagiochila deltoidea, P. pleurota, P. sinclairii. Additional Hepatics since Recognised Plagiochila prolifer, Radula allisonii, and a species of an undescribed genus.

(Note—On rocks on the lip of a waterfall in Pegasus Creek, I collected a hepatic which Mrs. Hodgson recognised to be new and forwarded to Prof. Herzog. He replied that it was a new genus and suggested the name Pachyglossa tristicha. If not a member of the Haplomitraceae, it probably represents a new family. However, the perianths were immature and better material may be necessary to establish its relationships.) Errata in Pt. I (Trans. Roy. Soc. N.Z., 77, 271–2) The following six, according to Mrs. E. A. Hodgson, are now regarded as synonyms: For No. 7. Bazzania involuta (Lindb.) read B. involuta (Mont.) Martin — No. 17. Chiloscyphus bidentatus read C. menziesii — No. 46. Hymenophytum leptodon read H. leptopodum — No. 61. Lepidolaena taylori read L. clavigera var. taylori — No. 63. Lepidozia albula read L. ulothrix — No. 64. Lepidozia bisetula Herz. read L. herzogii nom. nov. — No. 74. Lepidozia tenax read Lembidium tenax — No. 90. Mastigobryum novae-hollandiae read Bazzania adnexa Pegasus Expedition, 1949—Summary of Results 1. Eight bryophytes new to science or to New Zealand have been detected, four being described in this paper. 2. Seventeen mosses and eighty-three hepatics have been detected as indigenous to Stewart Island, in addition to the 182 mosses and 132 hepatics already detected, bringing the total known bryophytic flora from 314 to 414. 3. The hepatic element in the bryophytes increases steadily from north to south, or from 41% to 53% approximately. 4. The mound or ball form, adopted by numerous bryophytes, forms a unique physiognomic feature of the forests. 5. Strong climatic and edaphic contrasts between the northern and southern areas are doubtless responsible for the observed restriction to one or other area of many species of bryophyte. 6. Many orchids, ferns, mosses, and hepatics normally epiphytic, grow commonly on the forest floor. 7. Many bryophytes which on the main islands of New Zealand are alpine or subalpine here grow at sea-level. 8. Blindiopsis immersa, not seen by any other observer since 1874, is here abundant in forest streams. The Campbell Island moss Muelleriella crassifolia, first detected on the mainland in 1947 by K. W. Allison, was twice collected at Port Pegasus; while the moss Schlotheimia campbelliana, first recorded from Stewart Island in Part I of this paper, is now known to be abundant. Half-a-dozen hepatics hitherto known in New Zealand only from North Island, have been discovered in Stewart Island. Acknowledgments My thanks are extended to Dr G. J. Williams for his invitation to accompany the Geological Expedition to Port Pegasus, and for providing boat transport to various localities within the area, also

to Mr Alister Thomson for piloting me safely over many deep waters in the course of my studies. To Mrs E. A. Hodgson, who so willingly and generously undertook the determination of some 300 or more specimens of the hepatics, I owe a special debt of gratitude. Mr G. O. K. Sainsbury again assisted me with the moss determinations, and I cannot speak too highly of the value of his ever-ready assistance. Literature Consulted Cockayne, L., 1909. Report on a Botanical Survey of Stewart Island. Dept. of Lands, N.Z. Government, Wellington. Dixon, H. N., 1913–28. Studies in the Bryology of New Zealand. Bulletin N.Z. Institute, no. 3, parts 1–6. Hertzog, Th., 1936. Descriptions of New Species of New Zealand Hepatics. Trans. Roy. Soc. N.Z., 65, 350–356. — 1939. Ditto, Part II. Ibid., 68, 40–46. Hodgson, E. Amy, 1942. A Review of the N.Z. Species of Schistochila. Ibid., 71, 181–194. — 1944. A Review of the N.Z. Species of Chiloscyphus. Ibid., 73, 27–52. — 1944. A Review of the N.Z. Species of Plagiochila. Ibid., 73, 270–296. — 1944. A Review of the N.Z. Species of the Genus Radula. Ibid., 74, 273–287. — 1946. New Zealand Hepatieae: the Family Jungermanniaceae. Ibid., 76, 68–86. — 1949. A Review of the New Zealand Species of the Genus Frullania. Ibid., 77, 361–389. Martin, W., 1949. The Bryophytes of Stewart Island, Part I. Ibid., 77, 257–277. Sainsbury, G. O. K. New and Critical Species of N.Z. Mosses. Ibid., 75, 169–186. Williams, Gordon. The Geomorphology of Stewart Island. Geogr. Jnl., lxxxvii, 328–37.