Notes on New Zealand Floristic Botany, including Descriptions of New Species, &c. (No. 5)

Transactions and Proceedings of the Royal Society of New Zealand, Volume 57, 1927, Page 48

Notes on New Zealand Floristic Botany, including Descriptions of New Species, &c. (No. 5) By L. CockaynePh.D., F.R.S. and H. H. AllanD.Sc., F.L.S. [Read before the Philosophical Institute of Canterbury, 2nd December, 1925; received by Editor, 31st December, 1925; issued separately, 9th October, 1926.] 1. Taxonomic. The taxonomic details in this paper are based upon the principles outlined by L. Cockayne in No. 3 of this series (Trans. N.Z. Inst. 50 (1918) 161–2) together with the conclusions reached by the same author in his paper on New Zealand hybrids (New Phytologist 22 (1923) 105–127), and the terms defined by us in our paper on Hebe in this volume are used. Though Celmisia is badly in need of revision from the standpoint of the “natural method” in taxonomy the genus demands most critical treatment; but we propose later to devote a special paper to the subject. We take this opportunity of recording our most grateful thanks to the following gentlemen for assistance in various ways: Messrs. B. C. Aston, F.N.Z. Inst., F.I.C.; H. Carse; K. W. Dalrymple; W. R. B. Oliver, F.L.S., F.Z.S.; G. Simpson; J. Scott Thomson, F.C.S.; W. A. Thomson; E. Phillips Turner, F.R.G.S.; and A. W. Wastney. 67.* The numbers follow on consecutively in this series of papers. The previous paper appeared in vol. 56, pp. 21–33. Aciphylla Scott-Thomsonii Ckn. et Allan sp. nov. Species valde robustus, A. Colensoi Hook. f. affinis, a qua tenuioribus anguste linearo-lanceolatis, glaucis foliis, costis prominentibus differt. Caulis robustus c radice crassissimo ortus, usque ad 2 m. altus, 6 cm. diam. Folia radicalia glauca, numerosa, congesta, usque ad 1 m. longa, inferne bipinnata, superne pinnata, pinnis ± 4-jugis, ± 25 cm. longis; ultima segmenta 13–25 cm. longa, usque ad 25 mm. lata, anguste linearo-lanceolata, acuminata, pungentia, marginibus haud incrassatis, serrulatis; costa supra impressa, infra sub-carinata; lamina utrinque pagina nervis numerosissimis parallelis notata, nervis transversis brevibus irregulariter despositis, distinctis; petioli et petioluli crassi, plano-convexi vel bi-convex, supra vel infra interdum sub-carinati, basi vaginantes, serrulati. Bracteæ numerosæ. basi late vaginantes, apice pinnatæ, superne gradatim minoræ, vaginis usque ad 12 cm. longis, ± 2 cm. latis, segmentis anguste linearibus, pungentibus, striatis, marginibus serrulatis; segmentum centrale usque ad 30 cm. longum, pinnis plerumque 2-jugis multo brevioribus. Umbellæ numerosæ in paniculis magnis oblongis dispositæ. Inflorescentia ♂ in umbellis compositis disposita, peduncalis ± 12 cm. longis, bracteolis ovato-lanceolatis, ± acuminatis; umbellulæ numerosæ, satis patentes, pedicellis gracilibus; flores parvi, albi, calycis lobis acutis, antheris ± 1 mm. longis, late oblongis, filamentis gracilibus. Inflorescentia ♀ similes, umbellularum flores pauciores; fructus oblongus, ± 4 mm. longus.

South Island: South Otago Botanical District—Mount Maungatua, J. Scott Thomson; Powder Hill, Mount Flagstaff, J. Scott Thomson and L. C.; Swampy Hill, G. Simpson! This species is closely allied to A. Colensoi, but is distinguished by its constant glaucous colour, thinner linear-laceolata leaves, with well-marked mid-ribs. A. Colensoi is often epharmonically glaucous when growing in damp ground, but the glaucous hue of this species is independent of its growing-place. We are indebted to Messrs. Thomson and Simpson for living material and field notes. The description has been drawn up from material from Mount Maungatua. 68. Anisotome Enysii (T. Kirk) R. M. Laing var. tennysonianum R. M. Laing. Laing gave specimens of this to the herbarium of the Museum of Canterbury College. These exactly match Angelica trifoliolata (Hook. f.) Ckn., a fact first pointed out to the senior author by Professor A. Wall. The variety, then, must be removed from the flora. Up to Laing's discovery of the above species at Lake Tennyson (North-western Botanical District, at its junction with the North eastern Botanical District) it was only known from Mount Torlesse, where it grows in sphagnum bogs. In any case, it appears to be a very rare species. 69. Cassinia hybrids. While this paper is being revised prior to sending it for publication a series of specimens of Cassinia collected by Mr. George Simpson on Mount Maungatua have reached us. So interesting are they that a brief preliminary note seems necessary. Out of many plants differing from one another (“a tremendous mixture” according to Simpson's letter) we have received thirteen. Of these there is one we refer to C. albida (T. Kirk) Ckn.—a species previously unrecorded south of the extreme north of the Eastern Botanical District—, and another to typical C. Vauvilliersii Hook. f. The remainder we consider either hybrids between the above, or, as Simpson suggests, with C. fulvida also playing a part. No two plants are alike, the fundamental differences being size and shape of leaves, colour of tomentum—white to fulvous, with all intermediates—, and degree of greenness of the upper surface of the leaf. We defer giving a name to the group until further details are available; e.g., is the plant we refer to C. albida identical with the type? One interesting point stands out clearly, namely that C. fulvida var. linearis T. Kirk (Students' Flora (1899) 316—“leaves… distant… very narrow … clothed with white tomentum beneath”—italics ours), which was collected by Aston near Dunedin, and later by H. J. Matthews, is certainly one of this hybrid group. Cheeseman's remarks (Manual ed. 2 (1925) 986) that “the species [of Cassinia] are very closely allied, and are by no means easy to discriminate” are readily explained by the fact of there being swarms of hybrids: but we would say that the species themselves are quite distinct, their only polymorphy being that of slight fluctuating “variability,” or in some cases the presence of more than one jordanon. Information such as that supplied by Mr. Simpson is, in our opinion, far more valuable than the discovery of a new species.

70. Clematis australis T. Kirk. The original place of publication of this species was not, as always cited, The Students' Flora (1899) 3, but a sample page of what that work would be, published in 1891 at the time of the third meeting of the Australasian Association for the Advancement of Science. This description was widely circulated and was certainly “effective publication” according to the Rules of Botanical Nomenclature. Unfortunately we are not now in possession of a copy of the above publication and cannot give the actual citation. By Kirk, followed by Cheeseman in both editions of the Manual, the flowers are described as white, and this is used as a distinctive mark putting the species in the same group as C. indivisa and C. hexasepala; but, in point of fact, the flowers are invariably yellow. 71. Clematis Colensoi Hook. f. The plant of the Eastern Botanical District (Manual ed. 2 (1925) 431) referred to C. Colensoi can only be accepted with considerable hesitation, especially in the light of the statement that C. Colensoi is “a variable plant not always readily distinguishable from states of C. hexasepala and C. australis”—a statement which implies that its author did not know the true limits of these species. Petrie, in a letter to the senior author, was at one time inclined to separate the plant in question and a similar form from the North-eastern District as a new species. It is well that he refrained, as he had only a few dried plants for his material; but it is necessary that these eastern South Island forms of Clematis be examined carefully in the field. 72. Corallospartium racemosum (T. Kirk) Ckn. et Allan sp. nov. = C. crassicaule (Hook. f.) J. B. Armstg. var. racemosum T. Kirk in The Students' Flora (1899) 107.* No actual date of publication is given for The Students' Flora, so we take that of the preface. A fairly long description of this species is given by Kirk (loc. cit.) so that but little need be added here. The species is at once separated from C. crassicaule by its stems—less stout, almost flat, green (not yellow), more finely grooved; and by the flowers—in few-flowered racemes, the flowers with a deep purple blotch at base of standard, from which radiate light-purple lines to the margin, the wings with light purple lines but not fully covered with such, the keel with a purple blotch, and the calyx-teeth touched with purple—(information re colour supplied by J. S. Thomson). South Island: (1) North Otago Botanical District—near Lindis Pass, J. Buchanan ex Kirk. (2) South Otago Botanical District—Mount Roy (near Lake Wanaka), L.C.; Ben Lomond, on shady rocks, J. S. Thomson!, L.C.; Mount Dick, G. Simpson and J. S. Thomson! The Hector Mountains plant cited by Cheeseman as collected by W. Willcox most probably comes here. The species is evidently wide-spread, but local, in the South Otago District, and, unlike C. crassicaule, grows under fairly moist conditions.

73. Coriaria lurida T. Kirk. Kirk (The Students' Flora (1899) 98) separated his C. lurida from the group hitherto known as C. thymifolia, giving as its distinctive features “its lurid hue and strict habit, but presenting no structural points of difference.” Oliver (Trans. N.Z. Inst. 53 (1921) 364) enlarges the conception of C. lurida to include all the New Zealand forms up till then referred to C. thymifolia, stating “This New Zealand species, hitherto referred to C. thymifolia can easily be distinguished from all the American forms by the habit and shape of the leaves. C. thymifolia occurs from Mexico to Peru, and is a quite distinct species with small closely-set ovate acute leaves, which, though varying in size, are nearly constant in shape.” We consider the group to be a linneon, composed of several distinct jordanons, which hybridize when in company. A jordanon known to us from the bed of the Ashburton River answers quite well the brief description of C. thymifolia given by Oliver, as cited above. But Mr. Oliver, after examining specimens, has kindly informed us that it is distinct from the true C. thymifolia and has smaller leaves. Possibly Kirk's C. lurida is also a jordanon, but we are insufficiently acquainted with it at present to decide. In the meantime we follow Oliver in using C. lurida for the name of the compound species. The species of Coriaria hybridize most freely when they meet, and produce polymorphic groups, some forms of which are barely distinguishable from one or other of the parents, others come about midway between, and others again link up the whole swarm into a completely intergrading series. In all probability the North Island plants referred by Cheeseman (Manual ed. 2 (1925) 547) to C. angustissima belong to the group C. lurida×sarmentosa, and we unhesitatingly refer the so-called C. angustissima of Mount Egmont to that group, for which we propose the following name:×Coriaria sarlurida Ckn. et Allan. Plantae parentium signa diverse interligentes. Habitus sub-her-baceus vel suffruticosus. Folia inflorescentiaeque individuorum illas parentium interjacentes. We give this description, such as it is, perforce, since the present Rules demand one, but obviously nothing but a detailed analysis of the forms can give any satisfactory information as regards any hybrid swarm. Such group-names as we propose have hardly before been contemplated for wild hybrids, but we consider them of great convenience and importance, telling as they do at a glance the parentage involved, whereas an arbitrary name, after some collector, or what not, tells nothing of value. It is greatly to be hoped that when an International Conference is again held, some simple, commonsense rule for the making valid of such group-names will be drawn up. We also take out from the compound species the following varieties: (a.) Coriaria lurida T. Kirk var. parvifolia Ckn. et Allan var. nov. Planta humilis, sub-herbacea; rhizomata late vagantia; caules usque ad 1 dm. longi; folia ovata, acuta, vel sub-acuminata, plerumque ± 5 mm. longa, ± 2 mm. lata; racemi graciles ± 5cm. longi; flores parvi.

North Island: Egmont-Wanganui Botanical District—Mount Egmont, at c. 1200 m., on scoria-slopes, and in tussock-grassland: H.H.A. (b.) Coriaria lurida T. Kirk var. acuminata Ckn. et Allan var. nov. Suffrutex; rhizomata aliquanto condensata, caules usque ad 2 dm. longi; folia anguste ovato-lanceolata, valde acuminata, plerumque ± 25 mm. longa, ± 5 mm. lata; racemi et flori paulo majores. North Island: Egmont-Wanganui Botanical District—Mount Egmont, at c. 1150 m., subalpine shrubland, margins of streams in subalpine forest and scrub: H.H.A. These distinct jordanons have, on Mount Egmont, their own characteristic habitats, but where they meet hybrids are abundantly produced. Both probably have a fairly wide distribution, but in the absence of definite information we cite only the locality from which the type specimens were described. The Ashburton plant above-mentioned closely resembles var. parvifolia, and probably should be placed with it, while we have seen very similar specimens from Lochy River (near Lake Wakatipu) collected by Messrs. J. S. Thomson and G. Simpson. 74. Cotula Dendyi Ckn. in Trans. N.Z. Inst. 47 (1915) 118. This is rightly reduced by Cheeseman (Manual ed. 2 (1925), 993) to a variety of Cotula atrata Hook. f., but Cockayne is incorrectly cited as the author responsible for the change—for instance, in The Vegetation of New Zealand, (1921) 228, 344, the species is upheld by its author. Though the species was first published in 1915 the name had been in use for a good many years. The flowers, as given in the description, “vary from quite pale yellow to brown,” and we strongly support L. Cockayne's original suggestion, made in 1915, that plants with flowers other than yellow belong to C. atrata×var. Dendyi, for the species and its variety frequently occur in close proximity. Cotula atrata is thus increased in its content, and becomes a compound species made up of var. Dendyi Ckn. ex Cheesem. (but excluding the dark-flowered hybrids), and the type, for which we propose the name var. typica Ckn. et Allan var. nov., and which is identical with C. atrata Hook. f. in Handbk. N.Z. Flora (1864) 142. 75. Cotula sericea (T. Kirk) Ckn. et Allan sp. nov. = C. pectinata Hook. f. var. sericea T. Kirk in The Students' Flora (1899) 326. This species is readily distinguished from C. pectinata by its smaller dimensions, including the head only about 8 mm. diam., and the covering of dense silky hairs of nearly every part. It crosses with C. pectinata, and we have seen many “intermediates” of this class. South Island: probably on all the higher mountains of the South Otago Botanical District. 76. Danthonia setifolia (Hook. f.) Ckn. Hooker's variety setifolia of D. semiannularis R. Br. was created a species by L. C. in N.Z. Journ. of Agric. 23 (1921) 146 footnote. This change has not been cited by Cheeseman in the Manual ed. 2, but the species is most distinct, occurring unchanged throughout most of mountainous New Zealand. It is in fact far more distinct from

D. semiannularis than is that species from D. pilosa. As it is a somewhat important agricultural grass of the mountain sheep-runs it is desirable the name should receive recognition. 77. Discaria toumatou Raoul. The discovery of a semi-spineless form of the above (L. Cockayne, N.Z. Journ. of Sc. and Tech. 5 (1922) 206–8), which is a matter of distinct biological moment, is not mentioned in the Manual ed. 2. The shrub grows alongside the ordinary spinous form in identically the same habitat. It is of drooping habit; some of the twigs are spineless, while on others there are spines more or less stunted or up to 18 mm. long. The senior author has a vigorous plant grown from a cutting of the wild plant in his garden at Wellington that still possesses the same characters. South Island: Fiord Botanical District—Valley of the Dart near Kinloch, L. C., J. S. Thomson and G. Simpson (who have sent us fresh seed for genetic experiments). 78. Fuchsia perscandens Ckn. et Allan sp. nov. (Fig. 1). Caules scandentes, graciles, cortice pallide brunneo, tenui sed non chartaces, obtecti, parce ramosi, ramis divergentibus; ramuli hornotini gracillimi, pubescentes, nebulosi; ramuli annotini minute sulcati. Folia satis distantes, late ovata vel sub-orbiculata, sub-acuta, marginibus sinuatis, remote obtuseque dentati; tenuia, decidua, primo parce pubescentia, supra pallide viridia, infra glaucescentia, nervis reticulatis evidentis, primo ciliatis; petioli plerumque ± 2 cm. longi, graciles, nonnunquam multo longiores, laminis plerumque ± 3 cm. longi, ± 2 cm. lati. Flores solitarii, vel rarius 2–3 in axillis foliorum aggregati, penduli, pedunculis gracillimis ± 1 cm. longis; calycis tubus ± 9 mm. longus, basi inflatus, deinde abrupte angustatus, superne hypocrateriformis; calycis lobi ± 7 mm. longi, anguste ovati, acuminati, patentes; petala atro-purpurea, anguste oblonga, ± 2.5 mm. longa; stamina 8, filamentis purpureis ± 5 mm. longis; stylus gracilis, capitatus, ± 12 mm. longus, interdum longior; bacca atropurpurea, sub-cylindrica lateribus obscure tetragoniis, ± 1 cm. longa, ± 5 mm. diam. (a.) North Island: Ruahine-Cook Botanical District—forests near Feilding, abundant, H.H.A. (b.) South Island: (1) Sounds-Nelson Botanical District—E. C Jeffrey and L.C.; (2) North-western Botanical District—near Murchison, H.H.A.; (3) Eastern Botanical District—Malvern Hills, L. C. —said by G. Simpson and J. S. Thomson to be absent near Dunedin and by J. Crosby-Smith to be unknown in Southland. Fuchsia Colensoi Hook. f. was very briefly described in the Handbk. N.Z. Flora (1867) 728 as “intermediate in size [between F. excorticata and F. procumbens]. Leaves very variable, ovate orbicular or cordate. Flowers as large as in F. excorticata—Northern Island, Colenso. Middle Island, common, Canterbury Plains, Travers; Otago, Lindsay, Hector.” T. Kirk's description in The Students' Flora (1899) 181 includes the statement “A small erect or prostrate shrub with slender branchlets.” Cheeseman in Manual ed. 1 (1906) 187, describes the species as “A small shrub with long straggling

branches, sometimes producing slender flexuous unbranched shoots several feet in length,” and adds “a very variable plant, large forms of which almost pass into F. excorticata.” In the Manual ed. 2 (1925) 618 his original description is repeated, and the statement made, “This is either a very variable plant, or two very distinct forms are included in the present conception of the species.” Our opinion, based on field observations, so strongly supports Cheeseman's supposition of there being two distinct species included in Hooker's species that we restrict Fuchsia Colensoi Hook. f. to some jordanon or group of such, of shrubby habit, and define the true liane, which evidently is absent in many localities, as the new species F. perscandens—the name indicating a far more lianoid form than is seen in F. Colensoi, which according to Hooker and Cheeseman is a linneon. The linneon includes at the ends of the series F. excorticata and F. perscandens, while the problematical F. Colensoi comes more or less mid-way between the two, and all are connected by an almost bewildering group of hybrids, the parentage of which it is impossible to define unless one or other of the parents is far-distant from the locality where they occur. Fig. 1. In the foreground the lianoid stems of Fuschia perscandens Ckn. et Allan, which have fallen from their original support; Kitchener Park near Feilding. In foreground the lianoid stems of Fuchsia perscandens Ckn et Allan, which have fallen from their original support; Kitchener Park near Feilding. The photograph (Fig. 1) shows the cable-like stems of the liane. When growing in the open this species forms tangled rounded masses, similar to those formed by species of Rubi under similar conditions. Such rounded masses are easily distinguished from the hybrid forms by the absence of a stout branching, erect base. The description was drawn up from living plants growing in Kitchener Park, Feilding. 79. Gentiana serotina Ckn. In the Manual ed. 2, both in the body of that work and in the index, the specific name is written “scrotina.” By Cheeseman the

species was regarded as “a reduced state of G. corymbifera,” but as it grows in practically the same habitat as that species it is impossible to guess how such “reduction” could come about. To us it is one of the most distinct gentians in the flora. The species consists of two jordanons, the one with the corolla-lobes close together and the other with them further apart, the flower of the latter having a somewhat starry appearance. These forms can be readily observed in the field, and one or the other be gathered for certainty. 80. Helichrysum Purdiei Petrie. This is described by Cheeseman as a “very puzzling species not very closely allied to any other.” How it could be so designated, and not considered “a well-marked species,” as it was free from the troublesome “intermediates,” we do not know, but suspect that Cheeseman had more than a suspicion of its hybrid origin, since he speaks of the Hanmer specimens as growing “in company with H. bellidioides and H. glomeratum.” The truth is that, so far, wherever it has been found—Clarence Valley, Aston., to Arrowtown, L. C.— the above two species are always in the neighbourhood. The inflorescence is far from uniform in structure as defined in the Manual, and many distinct plants differing in various characteristics can be seen where the hybrid is in quantity, as near Hanmer. To Mr. C. E. Christensen is due the credit of first bringing proof that H. bellidioides and H. glomeratum crossed. The cross is the more interesting since it is between different sections of the genus—Xerochlaena and Ozothamnus. We do not propose a name and define such for the group at present since our material is not sufficient for a proper analysis, but the formula of course will be Helichrysum bellidioides×glomeratum. 81. Helichrysum Selago (Hook. f.) Benth. et Hook. f. var. tuberculata Cheesem. This, as collected by L. Cockayne, includes one of the plants upon which the variety is based. But it grew side by side with H. coralloides and H. Selago, and coming as an intermediate between the two is doubtless one of their hybrids. 82. Hoheria populnea A. Cunn. Cheeseman states (Manual ed. 2 (1925) 564) that H. populnea passes through a juvenile stage “markedly different to the adult,” and he bases this statement not on observations from the living plant during its development, but on specimens given to him by Mr. Carse. On the other hand, we are well acquainted with the life-history of the species in question, and can state without fear of contradiction that it is virtually the same in the juvenile as in the adult, and that it never possesses a juvenile divaricating form, as does H. angustifolia, or a twiggy, more or less weeping form like that of H. sexstylosa. There are certainly in the juvenile a few reduced leaves here and there, but mixed with far more leaves of the adult type. Further, we are greatly indebted to Mr. H. Carse for sending us a collection of specimens similar to those on which Cheeseman based his opinions, but they show only a small degree of leaf-reduction, and give no

idea of the growth-form. We have gone into this matter at some length for the different behaviour as a juvenile of H. populnea from that of its allies is of great importance from the standpoint of evolution. Our observations include the study of seedlings in gardens up to their reaching maturity, as well as field observations. 83. Lepidium tenuicaule T. Kirk. In Trans. N.Z. Inst. 14 (1882) 381 Kirk described his Lepidium tenuicaule and L. australe, from plants collected at Cape Wanbrow, Oamaru. In The Students' Flora (1899) 38 he reduced the latter to a variety of the former. Cheeseman in both editions of the Manual has treated L. australe as an epharmone of L. tenuicaule. After a study in the field of the Cape Wanbrow plants we consider this treatment to be correct. Robust forms growing through grass often develop more cauline leaves and are caused to assume a more erect habit of growth. In Trans. N.Z. Inst. 43 (1911) 175 Cheeseman created a variety minor for plants collected at Titahi Bay. This variety is retained in the Manual ed. 2 (1925) 473, where the statement is made “Both the type and var. minor on Kapiti Island, off Cook Strait.” An examination of the plants on Kapiti Island, and at Titahi Bay (where the type also occurs) led us to suspect that var. minor was also an epharmone, due to its special habitat conditions. To test this, plants of the different forms were collected and grown inland. These have become indistinguishable from one another, and accordingly we regard var. minor as an epharmone merely, and as a taxonomic variety it should be abandoned. The species occurs coastally at Timaru and Ashburton in addition to the localities cited in the Manual. In both these places it has been carried inland by gravel used for road—and railway—metal, as we have actually witnessed occurring at Oamaru. Probably most or all of the inland localities recorded are due to such carriage. 84. Libertia peregrinans Ckn. et Allan sp. nov. Species L. ixioides (Forst. f.) Spreng. affinis, sed rhizomatis late vagantibus facile distinguitur. Rhizoma horizontale, crassum, usque ad 8 mm. diam. Folia plura, ad nodos rhizomatorum conferta, angustissime linearia, apice attenuata, rigide crassa, usque ad 70 cm. longa, sub-nitentia, basi vaginantes, marginibus nonnunquam tenuiter scabridiusculis exceptis glabra; nervis lateralibus utrinque circiter 4, pagina utraque distincti, costis prominentibus subcomplanatis. Inflorescentia paniculata ramis alternatis, ramulis ultimis subumbellatis, pedicellis gracilibus; bracteae inferiorae usque ad 10 cm. longae, anguste linearolanceolatae, acuminatae, conduplicatae, superiorae gradatim minorae, chartaceae; tepala 3 exteriora patentia, supra albida, infra margine excepto cervina, ± 8 mm. longa, elliptico-oblonga, obtusa; tepala 3 interiora candida. ± 11 mm. longa, ± 9 mm. lata, late ellipticooblonga, retusa. in unguem brevem abrupte angustata; stamina 3, filamentis erectis, basim versus connatis, usque ad 6 mm. longis, antheris oblongis versatilibus, ± 3 mm. longis; styli 3-ramosi, ± 5 mm. longi, sub-patentes; capsula matura ± 1 cm. longa, oblongo-obovoidea.

Northern, Central, and Southern Botanical Provinces; Chatham Islands: forming extensive colonies in sand-hollows. The description is drawn up from plants collected in sand-hollows near Foxton (Ruahine-Cook Botanical District). Hitherto this species has been referred to L. ixioides, but it is at once distinguished by the far-spreading rhizomes, emitting dense tufts of leaves at intervals. It is confined to sand-hollows, and when growing wild is of a more or less bronze colouring in all the vegetative parts, but in cultivation inland the colour becomes distinctly more green. As a garden plant its stoloniferous habit makes it too aggressive for small gardens, or for alpine gardens, but where there is plenty of room it is well worth cultivating. 85. Macropiper excelsum (Forst. f.) Miq. var. major Cheesem. in Manual, N.Z. Flora ed. 1 (1906), 595. In the Manual ed. 2, Cheeseman follows R. M. Laing in changing the name major to psittacorum. Laing, however, was wrong in making the change, for though Piper psittacorum is a synonym for var. major, there is no obligation on the part of an author to follow the general rule of priority when transferring a species to a subdivision of such (See Art. 49, p. 47, Internat. Rules of Bot. Nomencl.), though he is rightly recommended to do so. But when transferring var. major of Piper excelsum to a var. of Macropiper excelsum, the original varietal name—major—must stand, it cannot be altered to psittacorum without ignoring Art. 48 loc. cit., therefore the valid name is Macropiper excelsum var. major as above. 86. Mida A. Cunn. This genus was published by Allan Cunningham in the Annals of Natural History 1 (1838) 376. He distinguished three species as follows: (1) M. salicifolia, foliis angusto-lanceolatis acuminatis, (2) M. eucalyptoides, foliis lanceolatis acuminatis, (3) M. myrtifolia, foliis ovalibus ovato-lanceolatis obtusiusculis acuminatisve. We consider Cunningham's genus amply distinct from Fusanus R. Br., under which it was treated as a section by Bentham and Hooker in Genera Plantarum, and accordingly restore it. Moreover, we accept Mida salicifolia and M. myrtifolia as valid species (M. eucalyptoides being perhaps a hybrid between these), as against W. Hooker's merging them into one in Icones Plantarum—a treatment followed by subsequent botanists, including Cheeseman. The species hybridize where they meet, and produce “intermediates.” Where only one species is present, as in the Ruahine-Cook District, it is constant, though each has its distinct juvenile form. Were the merging accepted in this case, it should logically be followed in other hybridizing groups, e.g. the species of Parsonsia. 87. Olearia Cheesemanii Ckn. et Allan sp. nov. = O. arborescens (Forst. f.) Ckn. et Laing var. angustifolia Cheesem. in Manual N.Z. Flora ed. 1 (1906) 285. As we reject all merging of extremely distinct true-breeding groups upon the unproved and highly improbable theory that they “vary” and “run into one another,” so must we separate the linneon O. arborescens into those simple or compound species, hybrids

and epharmones which form its content as defined in the Manual ed. 2. First of all comes out O. capillaris Buch., and its hybrids. Then probably var. cordatifolia is simply a series of hygrophytic epharmones, and it must be removed from the flora. Finally there is the very distinct var. angustifolia Cheesem., which is undoubtedly a valid species distinguished from the two other species by its thin lanceolate leaves with sinuate margins, lax, much-branched corymbs, but smaller than those of O. arborescens but much larger than those of O. capillaris. Another proof of its validity as a species is the record by Cheeseman in Trans. N.Z. Inst. 46 (1914) 4, of “a form almost intermediate between var. angustifolia and the type “Where the species occurs by itself, as in the Otaki Gorge, there are no “intermediates” i.e. no hybrids. So far Olearia Cheesemanii has been recorded from Ohinemuri Gorge, Lake Waikaremoana, Ngakawau near Westport, and K. W. Dalrymple and H. H. A. have recently found it in quantity in the Otaki Gorge, absolutely constant in form, growing on rock in the shade of Carmichaelia odorata and Hebe salicifolia var. angustissima. As a garden plant it would be of high excellence, and it will probably be hardy in many parts of Great Britain and Ireland. 88. Ranunculus Simpsonii Ckn. et Allan sp. nov. Herba R. sericophyllii Hook. f. affinis, foliis glabris, sub-carnosis, bracteis sub-foliaceis, achaeniis minute pubescentibus, stylis gracilioribus differt. Rhizomata crassa, ramosa, ramis ascendentibus. Folia radicalia in rosellas ad apices rhizomatorum ramorum conferta; petioli usque ad 10 cm. longi, plerumque breviores, sparsis gracilibus patentibus pilis induti, supra 1- sulcati, sulcis hirsutis, virides, late vaginantes; laminae ± 2 cm. longae, ± 2.5 cm. latae, 3–5- partitae fere ad basim, latissime ovatae, petiolulis inferne sulcatis, lobis tripinnatisectis, ultimis segmentis linearibus vel laneolato-oblongis, subacutis vel obtusis, paulo carnosis, glabris, vel primo parcissime pilosis, venis distinctis supra satis impressis. Scapi uniflori vel 2-flori, usque ad 8 cm. longi, satis crassi, ± 3 mm. diam., pilis gracilibus sub-patentibus, obtecti. Bracteae inferiorae foliis sub-similes, petiolatae; superiorae laminis 1–5 partitis, segmentis oblongis, obtusis, petiolis ± 1 cm. longis, supra glabris, infra pilis longis indutis. Sepala 5, pallide viride flava, obovato-oblonga, retusa vel 2-crenata, concava, ± 12 mm. longa, ± 6 mm. lata, infra sparsis pilis basi versus induta. Petala plerumque 5, late obovata vel oblongo-obovata, cuneata, sinuato-obtusa, in ungues breves angustata, ± 1.5 cm. longa, ± 1 cm. lata. Nectaria 3, depressa, apice unguis posita. Stamina numerosae, fila-mentis ± 6 mm. longis, antheris oblongis ± 1.5 mm. longis. Achaenia matura numerosissima in capitula subglobosa ± 1 cm. diam. aggregata, paulo turgida, minute pubescentia, ± 2 mm. longa, leviter carinata, in styla ± 2 mm. longa, satis gracilia, parum torta, angustata, ambitu dimidiato-turbinata. South Island: South Otago Botanical District—Bold Peak (near Lake Wakatipu) c. 1800 m., abundant, G. Simpson and J. Scott Thomson!

This handsome little species is allied to R. sericophyllus Hook. f., but is amply distinct, and is at once separated by its glabrous leaves. According to the discoverers it favours finely disintegrated rock as its habitat. It will make a good addition to the rock-garden Ranunculi. The descrption has been drawn up from living plants and from spirit-preserved material sent by J. Scott Thomson. 89. Rubus cissoides A. Cunn. var. subpauperatus Ckn. Cheeseman (Manual ed. 2 (1925) 500), cites Cockayne as the author of the above variety, which it is stated Cockayne published in his Report on a Botanical Survey of Stewart Island. This is not correct;—what was published, in that report was a species—certainly valid—named Rubus subpauperatus, which is only in part the var. subpauperatus of the Manual. Cheeseman further states that the variety produces flowers in the shade, implying that it does not do so when growing in the open. On the contrary, R. subpauperatus flowers freely when growing in full sunshine in contradistinction to R. cissoides var. pauperatus—an epharmone, not a jordanon and so as a variety should be removed from the flora—which rarely, if ever, blooms except in the shade. Finally R. subpauperatus Ckn. is absolutely distinct from any form of R. cissoides, and comes much closer to R. schmidelioides var. coloratus; in fact this group of plants might be joined to R. subpauperatus, thus making the latter a compound species. 90. Senecio rotundifolius Hook. f. var. ambiguus Cheesem. This variety was designed by Cheeseman to include the forms of shrubby Senecio growing on the west coast of the North-western Botanical District concerning which there has been great difference of opinion. The matter was gone into at considerable length by L. Cockayne (Trans. N.Z. Inst. 50 (1918) 183–4) who, however, had only a living specimen not in flower, collected by Mr. B. C. Aston at West Wanganui Inlet. The conclusion reached was that the taxonomic position of the plant in question was “quite uncertain,” that it was neither typical S. rotundifolius nor typical S. elaeagnifolius, but that it might be a variety of either, or a distinct, undescribed species. Cheeseman (Manual ed. 2 (1925) 1026) virtually follows Cockayne in founding the var. ambiguus. On the other hand, he and T. Kirk had earlier referred the group to typical S. rotundifolius, while Petrie (Trans. N.Z. Inst. 46 (1914) 30) had “without hesitation” referred the Cape Foulwind plant to S. elaeagnifolius. Upon examining the specimens in Petrie's herbarium we found they were by no means uniform, some resembling S. rotundifolius, some S. elaeagnifolius, and some more or less intermediate. Probably both species are present on the coast, and var. ambiguus is a hybrid between them, but we hope, later, to go fully into the question. 91. Teucridium parvifoliun Hook. f. var. luxurians Cheesem. This variety, described in the Manual ed. 2 (1925) 764, was based by Cheeseman on specimens collected from a single plant growing in forest shade on the banks of the Mangles River (North-western Botanical District). We have since studied the species as it grows in forests between Pohangina and Feilding (Ruahine-Cook Botanical

District). There it is abundant, and agrees exactly with Cheeseman's variety. A point not referred to by him, but quite marked in the original specimens, is that the leaves are frequently irregularly lobed. The corolla is invariably white. The production of 2–3 flowered cymes differs considerably in the individual plants, many being almost entirely with solitary flowers, many again with a majority of flowers in cymes. Moreover, an examination of individual plants has shown that the degree of production of cymose inflorescences may vary from season to season. At first sight the variety appears abundantly distinct from the very small-leaved form, but in this latter there is also an occasional development of cymose inflorescences, and it is desirable that further field study should be made to decide whether the variety is valid, or is a shade-epharmone. The plant in the localities here cited is plastic, getting smaller-leaved in the more open situations, and becoming a semi-liane in the denser shade. 92. Trisetum saxeticolum Ckn. et Allan sp. nov. Gramen perenne, caespitosum, ad 2 dm. altum, T. subspicati Beauv. affinis sed habitu rubustiore, culmis glabrescentibus distinguitur. Culmi erecti vel sub-geniculati in rhizomate brevi oblique arti conferti, ± 15 mm. longi, paulo incrassati, 2- vel 3-nodi, nodis plerumque in foliorum vaginis inclusis, teretes, glabrescentes, sulcati, sulcis minute pubescentibus. Foliorum vaginae striatae, sub-chartaceae, minute pubescentes, basales ± 2 cm. longae, persistentes, pallidac; ligulae breves, ± 1 mm. longae, laceratae; laminae lineares, ± 9 cm. longae, erectae, planae vel conduplicatae, in acumen filiforme attenuata, molliter pubescentes. Panicula angusta, spiciformia, multi-flora, ± 4 cm. longa; axis primarius leviter pilosus; rami graciles, erecti, leviter scabridiusculi; spiculae 2–3 florae, ± 6 mm. longae, arista excepta, in maturitate straminellae, sub-nitentia; gluma inferior ± 4 mm. longa, uninervia, acuminata, molliter pubescens, marginibus scariosis, carina scabrida; gluma superior similiter, ± 5 mm. longa, fere aristata, 3-nervia; rhachilla pilis albis longis gracilibus ornata; valva paulo incrassata, pubescens, ± 5 mm. longa, carinis scaberulis, bifida; arista ± 6 mm. longa, erecta vel aliquanto inclinata, minute scaberula; palea scariosa tenuiter bicarinata, apice bifida, carinis minute ciliatis; caryopsis ± 3 mm. longa, fusiformia. North Island: Ruahine-Cook Botanical District—in crevices of coastal rocks, northern shores of Cook Strait. The occurrence of this species on coastal rocks near Wellington has long been known; but if dealt with at all, it has previously been referred to T. subspicatum Beauv. It is certainly related to that species, but differs, among other things, in the robuster habit, the finer pubescence (T. subspicatum is rather tomentose than pubescent), the practically glabrous culms, much larger panicles, with the rhachillas pilose, and the glumes hardly shining. The type specimens were gathered by H. H. A. on rocks near Island Bay, Wellington. The species probably has a wider distribution than is at present known. Buchanan in Trans. N.Z. Inst. 6 (1874) 226, includes T. subspicatum in his list of the species occurring in the Wellington Province,

but gives no localities. Aston in Trans. N.Z. Inst. 43 (1911) 247 also includes in his list of plants of the Wellington Province “T. subspicatum Beauv. ascends to 5,000 ft.” We are not acquainted with any localities for T. subspicatum in North Island, nor does Cheeseman in either edition of the Manual cite any North Island localities. T. subspicatum is, of course, a compound species, but it would seem most probable that in part at least both Aston's and Buchanan's gatherings are attributable to T. saxeticolum. 2. Exotic Species Wrongly Included In The Flora. We give here a list of exotic species that should be excluded, in whole or in part, from the flora, but which are accepted in the Manual ed. 2, together with species as to the origin of which there is grave doubt. Cheeseman in the Manual ed. 2 (1925) 532, makes the following statement under his description of Geranium molle: “There can be little doubt that this is introduced, but as it has had a place given to it in previous works on New Zealand plants, and as it is now found in all soils and situations, and would certainly be considered indigenous by a stranger unacquainted with its history, it appears best to retain it in the Flora.” None of these arguments seem to us to have any cogency whatever—it does not seem well to repeat the mistakes of the past, the description of this particular plant's habitats is altogether too sweeping, and were the last reason valid then exotic species after species would have to be admitted. That it is desirable that a flora of the exotic species occurring in New Zealand should be prepared we do not gainsay, but in view of the important principles involved, not least from the phytogeographic standpoint, and the misinterpretations likely to be made were Cheeseman's procedure followed, we strongly deprecate any such admissions to floras explicitly claiming to deal with the indigenous plants of a country. It is indeed fortunate as far as the New Zealand flora is concerned that there are so few really doubtful cases. With the flora of such a country as Great Britain, of course, the matter is in quite other condition. 1. Ageratum conyzoides L.—This was recorded by Cheeseman in Trans. N.Z. Inst., 20 (1888) 169 for the Kermadec Islands. He remarks “Perhaps introduced, but it is truly native in many parts of Polynesia.” Oliver, rightly we think, placed it in his list of introduced species in Trans. N.Z. Inst. 42 (1910) 174. 2. Aleurites moluccana Willd.—This was also recorded by Cheeseman (loc. cit., p. 172). Oliver (loc. cit., p. 174) in treating of it as an introduced species says “The candle-nut (Aleurites moluccana), like the Polynesian ti, appears to be the survivor of the abandoned cultivations of the native race.” His action seems to be justified. 3. Alopecurus geniculatus L.—This was apparently first collected by Colenso in North Island, and by Lyall in South Island. It is much more common in North Island than the citation of special localities in the Manual would suggest. The species is a compound one, and whether there are indigenous forms or not we do not know, but in the great bulk of its present stations it is almost certainly introduced.

4. Atriplex patula L.—This was accepted by Hooker in Flora Nov.-Zel. 1 (1853) 215, for specimens sent by Colenso from the east coast of North Island. Cheeseman remarks (Manual ed. 2 (1925) 409) “How far it is indigenous in Australia and New Zealand is now very difficult to determine.” We regard it as undoubtedly introduced, as it could hardly have been missed by the earlier collectors had it been present then on the coasts. 5. Bidens pilosa L.—Hooker, Handbk. N.Z. Flora (1864) 139, records this for “Northern Island,” and for “Kermadec Island,” but of the former station says “a weed of cultivation, not indigenous.” Cheeseman accepts the species without comment. For the mainland, at any rate, it should be excluded from the flora. Oliver (loc. cit., p. 172) accepts it for the Kermadees. 6. Callitriche verna L.—Possibly there are indigenous varicties of this linneon, but there are almost certainly introduced forms also. The “species” in New Zealand need closer investigation. According to several informants it has of recent years shown a marked tendency to increase in slow-flowing waters in North Island. 7. Calystegia sepium R. Br.—This was collected by Banks and Solander, but as the “species” is a linneon the question should be taken up whether there exists an indigenous variety or varieties. It is certain that in many stations, cultivated land etc., it is introduced only. 8. Capsella procumbens Fries.—Cheeseman remarks (Manual ed. 2 (1925) 470) “I doubt this plant being a true native of New Zealand, and Sir J. D. Hooker has expressed the same opinion with regard to its supposed nativity in Australia and Tasmania.” In the stations in which we have seen it—clay cliffs at Cape Wanbrow, Oamaru; coast of Otago Peninsula (G. Simpson and J. S. Thomson!)—an exotic origin seems very probable. We think it should be excluded from the flora. 9. Carex leporina L.—This was first recorded by Cheeseman as occurring in various places “in the western portion of the Nelson district” (Trans. N.Z. Inst. 14 (1882) 301). In the various stations in which we know it we regard it as bearing all the signs of an introduced species. Cheeseman remarks in the Manual ed. 2, “It is probably introduced into New Zealand.” It should be excluded from the flora. 10. Centella asiatica (L.) Urban.—Nannfeldt in his “Revision des Verwandtschaftskreises von Centella asiatica (L.) Urb.” (in Svensk. Bot. Tidskr., 18 (1924) 397), removes the New Zealand plant from Centella asiatica and revives Colenso's name. The citation for the New Zealand species thus becomes Centella uniflora (Col.) Nannfeldt. This action is probably correct, but it still remains to make a proper analysis of the New Zealand forms. Colenso's description in Trans. N.Z. Inst. 17 (1885) 239 is based on plants collected on “Wet sides of slopes, gullies near Norsewood,” and evidently refers to one of the low-growing forms found in damp pastures. Plants taken from damp pastures near Feilding, which were then prostrate, with very small, thick, leaves on short petioles, have in cultivation in the junior

author's garden become quite luxuriant and show little if any difference from the form found in forest, or in shade in boggy places in the same locality. The forms would thus appear to be epharmones of a single jordanon. 11. Chenopodium ambrosioides L.—Hooker in the Handbk. N.Z. Flora (1864) 230 says “Northern Island: cultivated ground, Colenso, perhaps introduced.” He also cites Hector and Buchanan as having collected it in the “Waitaki Valley and lake district.” Cheeseman in the Manual ed. 2 (1925) 407, gives only North Island stations, and remarks “Has appeared as a naturalized plant near Wellington.” We remove it from the flora without hesitation. 12. Chenopodium carinatum R. Br.—The acceptance of this rests purely on its collection at the Bay of Islands by R. Cunningham in 1834. The station given by A. Cunningham (Mag. Zoo. and Bot. 1 (1838) 456) is “sandy shores of the Bay of Islands.” Cheeseman in the Manual ed. 2 (1925) 407 says “May be truly native in the North Auckland Peninsula. Elsewhere it is doubtless naturalized.” Now, by 1834, there had been ample time and opportunity for the plant to have been brought to the Bay of Islands, and for it to have spread considerably. We consider it should be removed from the flora. 13. Chenopodium urbicum L.—Hooker in the Handbk. N.Z. Flora (1864) 230 cites Colenso as having collected this in North Island, and Haast from “Ashburton River and Rangitata range,” and also “New River, Hb. A. Richard.” Cheeseman in the Manual ed. 2 (1925) 407, states it to be not uncommon in South Canterbury and Otago, and cites Colenso as having gathered it on the east coast of North Island. He adds, “Has also appeared as a naturalized plant near Wellington.” The species should certainly be excluded from the flora. 14. Cordyline terminalis Kunth.—Cheeseman accepts this for the Kermadec Islands in the Manual ed. 2 (1925) 310, but Oliver (loc. cit., p. 173), after careful study of the plants now occurring there, states “I have no hesitation in including, it among the list of plants introduced by man.” We consider this the correct course to take. 15. Cotula australis Hook. f.—According to Hooker in Handbk. N.Z. Flora (1846) 141 this is the Soliva tenella of A. Cunn. in the “Precursor” (Mag. Zoo. and Bot. 2 (1839) 128). Cunningham's plant was collected by R. Cunningham in 1834 “on the margins of fresh-water streams.” We have not seen it in any station where its nativity appeared at all certain, and we leave it in the flora with considerable doubt. 16. Cuscuta densiflora Hook. f.—As Cheeseman remarks, this is “very imperfectly known, and may not be truly indigenous.” The specimens on which Hooker founded his species were said to have been collected by Lyall at Port Underwood, Marlborough. It is desirable that further search be made. 17. Cyperus tenellus Linn. f.—Hooker in the Handbk. N.Z. Flora (1867) 745 gives the record “abundant in the Newton and Dedwood districts, Auckland, Mr. Kirk.” and adds “probably introduced both into Swan River and New Zealand.” Cheeseman in Manual ed. 2

(1925) 215, says “doubtfully indigenous in temperate Australia and New Zealand.” We do not hesitate to remove it from the flora. 18. Cyperus vegetus Willd.—Cheeseman states in the Manual ed. 2 (1925) 215, “there can be no doubt that it exists only as an introduced species in New Zealand.” With this we fully agree, and accordingly remove it from the flora. He further states that he retains it in the flora “because it has been twice described as an indigenous species, and on account of the remarkable fact that wherever found it presents all the apperance of a true native, and would certainly be taken as such by any one unacquainted with its origin.” To us these reasons do not seem valid; moreover, we have seen the plant associated with exotic species in sandy river-beds, ditches, damp pastures, and by the roadside. 19. Elatine americana Arn. var. australiensis Benth.—A: Cunningham (Mag. Zoo. and Bot. 4 (1840) 26) created his species E.? gratioloides for a plant collected by R. Cunningham “In a bog at Tauraki, Hokianga River.” Hooker, both in Flora Nov.-Zel. 1 (1853) 27, and in the Handbk. N.Z. Flora (1864) 28 gives E. gratioloides as a synonym for the American species E. americana. Bentham in Flora Austral. 1 (1863) 178 had given the varietal name australiensis for the Australian and Tasmanian plants. T. Kirk, and Cheeseman in both editions of the Manual, considered Bentham's variety well founded, and adopted it for the New Zealand plant. Niedenzu, however, in Die Natuerlichen Pflanzenfamilien 21 (1925) 276, restores Cunningham's name for the New Zealand, Australian, and Tasmanian species. From the differentiae given by him we consider this the best course to take. Hence the name E. americana disappears from the flora, if this view be upheld. 20. Festuca rubra L.—Cheeseman in both editions of the Manual has these remarks, “According to Professor Hackel, this constitutes the greater part of the F. duriuscula of the ‘Flora Novae-Zelandiae’ and the Handbook, the true F. duriuscula probably not existing in an indigenous state in New Zealand… It has considerable value as a sheep-grass, and is often sown on sheep-runs.” In ed. 2 he adds (p. 206) “From that fact it is doubtful whether some of the European forms that can now be readily collected even in remote districts may not have been introduced. Of some varieties, however, there is no reason to doubt their nativity.” We are not, however, given any diagnoses by which we may recognize them. Now to Hackel Festuca rubra L. was a huge compound species. Already in 1882 (Monographia Festucarum europaearum, p. 128) he had described six subspecies, with many varieties and subvarieties, and to us it seems to serve no useful purpose to include F. rubra L. in the New Zealand flora as a sort of dumping-ground for forms of festucae for which we cannot find any more suitable resting-place. Moreover it is highly probable that the now far-spread introduced species and varieties have hybridized with indigenous ones, so that our knowledge of this group is chaotic. The whole matter will have to be investigated de novo in the field; it cannot be done abroad by the herbarium method: and to us the only profitable course seems to be to admit to the flora only those “varieties” which have shown themselves by genetic

research to be jordanons, and which cannot be matched with exotic varieties. That this will be a task of extreme complexity, is obvious. That it will ever be done we are not so sure! But better by far left undone than have names based on a few scraps collected anyhow. 21. Geranium molle L.—As already explained, this species should undoubtedly be removed from the flora, where it never should have appeared. 22. Gypsophila tubulosa Boiss.—L. Cockayne in Trans. N.Z. Inst. 45 (1913) 255 gave as his reason for doubting that this species is indigenous the fact that “it has never been found in really virgin country, but always in places where sheep graze.” Cheeseman in the Manual ed. 1 (1906) 62 had stated that “no evidence has ever been obtained in support of such a view.” This statement is retained in ed. 2 (1925) 420, and he further argues that it is highly improbable that a plant found only in South Europe should reach New Zealand sufficiently early to be well established in 1834, when Colenso gathered it. He continues “On the assumption that it is a naturalized plant, it is also difficult to account for the fact that it has not further increased its range during the last fifty years.” Hooker in the Handbk. N.Z. Flora (1864) 22 says “it is worth observing whether it is rapidly increasing for if so it is probably an importation.” In the appendix (1867) 725 he remarks “appears to be rapidly spreading over New Zealand, and is, doubtless, an imported plant.” We may point out that such increase has been enormous in Central Otago, and further field observations have given us no reason to modify the statement of 1913. 23. Imperata arundinacea Cyr. var. Koenigii Benth—In the Manual ed. 2 (1925) 136 Cheeseman says “Perhaps introduced only, but it is one of those species which might be expected to be indigenous in the extreme north of the Dominion,* Of course “the extreme north of the Dominion” is in Polynesia and not in the New Zealand Region. and I have consequently given it the benefit of the doubt.” It was recorded in ed. 1 with the same remarks, as collected by R. H. Matthews near Kaitaia. No further collections are given in ed. 2. Where there is grave doubt our opinion is that it is better to exclude than include, especially in view of the great importance now apt to be attached to the study of those nonendemic supposedly indigenous species called by Willis “Wides” in his “Age and Area theory.” 24. Juncus bufonius L.—This was admitted to the flora by Hooker, without any expression of doubt, and this course was followed by Cheeseman in ed. 1 of the Manual. In ed. 2 p. 296, however, he remarks “Doubtless introduced with grass-seed in the early days of colonization.” Our field observations strongly support the view that it is introduced only, and we remove it from the flora. 25. Juncus lampocarpus Ehr.—This was first recorded in Trans. N.Z. Inst. 7 (1875) 378 by Kirk, who gathered it “about Karori and other places in the vicinity of Wellington,” and in “the southern part of the province of Otago.” He points out as features distinguishing the New Zealand plant from British specimens, the internal divisions of the leaf and stem more prominent, while both the inner and

outer segments of the perianth are acute and much shorter in relation to the capsule.” Cheeseman admits it to the flora with some doubt. Possibly there is both an indigenous and an introduced variety. Fresh study of the species in New Zealand is desirable. 26. Juncus tenuis Willd.—This was first recorded by Cheeseman in Trans. N.Z. Inst. 11 (1879) 433. His specimens were collected in 1875 “near Omano” on the Northern Wairoa River. He remarks “So many European plants are becoming naturalized in the colony that additional evidence will be required before Juncus tenuis can be included in the list of our indigenous species.” In both editions of the Manual he refers to its increase in range, and considers it “a very doubtful native.” From field observations we are satisfied it should be removed from the flora. 27. Kyllinga brevifolia Rottb.—This was collected first by Ball in ‘marshy ground on the north side of Mongonui harbour” and recorded by Cheeseman as K. monocephala Rottb. in Trans. N.Z. Inst. 11 (1879) 434. He remarks “So far as I can judge from the information supplied to me the plant appears to be truly indigenous; indeed it is precisely one of those species which might naturally be expected to occur in the northern extremity of the island.” In the Manual ed. 1 (1906) 765 Cheeseman says “possibly only naturalized in New Zealand,” and its range is given as from Mongonui and Ahipara to the North Cape. In Trans. N.Z. Inst. 51 (1919) 90 Cheeseman records it from near the Manakau Harbour, remarking “In this locality it is certainly a recent introduction … But its nativity in any part of New Zealand must be regarded as exceedingly doubtful.” In the Manual ed. 2 (1925) 214, however, the less emphatic expression of ed. 1 is retained. We think the species should be removed from the flora. 28. Mesembryanthemum aequilaterale Haw.—This was included in the flora by Kirk in The Students' Flora (1899) 184. The station given is “North Island: littoral. Napier (identified in the absence of flowers): Castle Point.” There is no other record. In both editions of the Manual Kirk's specimens cited as having been seen by Cheeseman. But Cheeseman remarks in ed. 2, p. 416, “I have seen no indigenous specimens.” It seems best to exclude the species from the flora. 29. Paspalum Digitaria Poir.—Apparently this was first collected in New Zealand by Cheeseman, near Ahipara. In Trans. N.Z. Inst. 46 (1914) 6 Cheeseman remarks, “Of late years this has increased enormously in the Lower Waikato.” In both editions of the Manual it is stated, “probably introduced.” The species should be removed from the flora. 30. Polygonum aviculare L.—Cheeseman, in both editions of the Manual, states that he retains the species in the flora only because of “the positive opinion expressed in favour of its nativity by the late Mr. Kirk.” It seems unnecessary to review the arguments of Kirk and Travers in various papers in Trans. N.Z. Inst. 4 and 5, but in view of all the evidence, and our own field observations we unhesitatingly exclude the species from the flora. As for the closely related P. plebeium we are pretty certain that stunted forms of P. aviculare

have been so referred, but even if it be in New Zealand it has probably been introduced. 31. Ranunculus sessiliflorus R. Br.—Hooker in the Handbk. N.Z. Flora (1864) 9 cites Colenso as the original collector, and remarks “perhaps introduced.” Cheeseman in both editions of the Manual says of it “once very plentiful, but now becoming rare.” We think the species were better excluded. 32. Salsola Kali L.—According to Hooker in the Handbk. N.Z. Flora (1864) 233, Colenso, who first collected the plant in New Zealand, observed that the plant was perhaps introduced. Cheeserman in the Manual ed. 2 (1925) 412, speaks of it as “of very doubtful nativity in New Zealand.” We consider it better to exclude the species from the flora. 34. Solanum nigrum L.—Banks and Solander collected what is believed to be this on Cook's first voyage, so a form of the compound species is undoubtedly indigenous. Whether this form truly matches any form, of other regions we cannot possibly say, but it is certain that more than one jordanon now occur in New Zealand. Dr. J. P. Lotsy felt sure what he saw was distinct from what he knew in Europe. Almost certainly many of the individuals met with in the settled country are of exotic origin. The epharmones of plants growing in shade and in the open respectively are very distinct in appearance. 34. Sonchus oleraceus L.—In the Handbk. N.Z. Flora (1864) 166 S. oleraceus L. var. a is stated to be “perhaps only introduced.” Kirk in The Students' Flora (1899) 362 rightly treats it as a naturalized species only. Cheeseman, however, in both editions of the Manual accepts it with some doubt. We think it should be removed from the flora. 35. Sonchus asper Hill.—This is the S. oleraceus var. γ of the Handbook, p. 166. Hooker says of it, “Certainly indigenous, being found by Banks and Solander and Forster, and at Chalky Bay by Lyall, and in the interior of the Northern Island by Colenso.” Of these collections that of Banks and Solander is the one of most moment in the argument. T. Kirk in Trans. N.Z. Inst. 26 (1894) 265, described the plant of maritime cliffs as S oleraceus L. γ littoralis. He makes the significant remarks “I have never seen this form on cultivated land, and, as far as I am aware, it is absolutely restricted to maritime localities. The singular absence of variation is a remarkable feature when this plant is compared with the typical form and varity γ. It seems not unlikely that the plant observed by Banks and Solander is identical with var. littoralis, the fruits of which resemble those of var. γ, but are slightly larger. This point could doubtless be settled by an examination of the specimens in the Banksian Herbarium.” This opinion is reinforced by the fact that the collections of Banks and Solander were made on the coast, and there is hardly a doubt also that the Chalky Bay plant collected by Lyall was var. littoralis. L. Cockayne in Rep. Bot. Survey of Kapiti Id. (1907) 21 raised the coastal plant to the rank of a species as S. littoralis (T. Kirk) Ckn. This action we consider amply justified and

it has been supported by Carse and others. It is highly desirable that the specimens of Banks and Solander be re-examined, but we have little or no hesitation in removing S. asper from the list of indigenous species. The two introduced species hybridize freely, thus accounting, in part, for the great “variability,” but there are doubtless several jordanons in each. 36. Sporobolus indicus R. Br.—Cheeseman in both editions of the Manual states of this “it is certainly introduced,” and cites Bishop Williams's statements as to its first appearance at the Bay of Islands. The species should be excluded from the flora. 37. Stipa setacea R. Br.—This is first recorded by Petrie in Trans. N.Z. Inst. 19 (1887) 326. He says “Several years ago I found this grass at Firewood Creek, Cromwell, and the Nevis Bluff, Kawarau River. As the plants were confined in these localities to small areas, and these were near an important highway, it was uncertain whether the species had not been accidentally introduced. In March of the present year [1886] I found it on the banks of the Waitaki River at Wharekuri. The discovery of this widely-distant habitat is sufficient to remove all reasonable doubt as to its being a genuine native of New Zealand.” His reason for accepting the species does not to us seem cogent. The plant is caten by sheep, and the fruits might easily be carried by them to runs at considerable distances, just as species of Danthonia are so carried. Cheeseman in the Manual ed. 2 (1925), p. 150, says “It is probably naturalized in New Zealand.” We would remove the species from the flora. 38. Veronica anagallis-aquatica L.— This was collected by Colenso in “watery places” at Hawke's Bay, and was accepted by Hooker without expression of doubt. Cheeseman in the Manual ed. 2 (1925) 832 says “Possibly Mr. Colenso's specimens were introduced as well [as in South Africa]; but if so, it is remarkable that the plant should have apparently disappeared.” We have no hesitation in removing it from the flora. It is not likely that a close search would secure a rediscovery of it. A closely allied species that has as yet not been certainly identified, occurs as a naturalized plant by the banks of the Waitaki River, near Glenavy. V. scutellata, also, is so far known as naturalized in New Zealand only in one small patch of Typha swamp near Feilding. All three records are thus from places subject to great modification during settlement, and might easily disappear before they had had a chance of spreading far. 3. Phytogeographic. The following does not by any means exhaust the list of unpublished localities of which we are cognizant. But as the second edition of the Manual must of necessity be the starting point, notwithstanding earlier publication of any locality, it is essential to study carefully the details in the Manual regarding the distribution of each species before adding such additional information as is required. Also the alterations in the conception of many species which are being made from time to time, will alter in no few instances the present dicta regarding the distribution of certain groups. We also deprecate the practice of publishing lists of “hitherto unrecorded habitats”

when the information given provides no real advance in our knowledge of the distribution of the species concerned. It is to be noted also in that in such lists there has generally been given no details of the habitat at all, but merely names of localities. 1. Agrostis muscosa T. Kirk. South Island: Eastern Botanical District—Coastal “loess” cliffs near Ashburton; H. H. A. 2. Anisotome diversifolia Cheesem. South Island: Sounds-Nelson Botanical District—Shingle-slips on Gordon's Knob, c. 1500 m.; H. H. A. 3. Aristotelia fruticosa Hook. f. North Island: Waikato Botanical Subdistrict—Kaikuri near Waitomo; E. Phillips Turner. 4. Arthropteris tenella (Forst. f.) J. Sm. South Island: Sounds-Nelson Botanical District—Various localities in subtropical rain-forest in the Marlborough section of the district; E. Phillips Turner, L. C.—Cheeseman had evidently had no specimens from South Island, though he cites the occurrence of the species from near Nelson, and on Banks Peninsula on the authority of T. Kirk and J. B. Armstrong respectively. R. M. Laing does not give it in his list, but W. Martin records it on the authority of D. G. Riches of Akaroa, who stated he had seen it but “exact locality forgotten.” 5. Celmisia coriacea Hook. f. var. stricta Ckn. South Island: South Otago Botanical District—On all the higher mountains. Cheeseman records it only from the Takitimu mountains where the “type” was collected. It apparently increases after the tussock-grassland is burnt. 6. Corokia buddleoides A. Cunn. var. linearis Cheesem. North Island: Thames Botanical Subdistrict—Mangarewa Gorge on the Mamaku Plateau, in open forest; L. C. 7. Dichelachne sciurea (R. Br.) Hook. f. South Island: (1) North-western Botanical District—near Grey-mouth, T. Kirk; in open manuka shrubland, H. H. A. Cheeseman had “seen no specimens” from that locality and merely quoted Kirk's record. (2) Western Botanical District—near Hokitika, in open manuka shrubland, H. H. A. 8. Epilobium rubromarginatum Ckn. South Island: Western and Eastern Botanical Districts at their junction—almost certainly the plant referred to E. tasmanicum by A. Wall in his Flora of Mount Cook from high altitudes is this species, which L. C. saw in abundance on the Sealey Range. 9. Gaultheria antipoda×oppositifolia. North Island: (1) Volcanic Plateau Botanical District—Near Lake Rotoma, on rock faces, common; J. P. Lotsy and L. C. (2) Egmont-Wanganui Botanical District—Mount Messenger on road-cutting, H. H. A. 10. Graultheria oppositifolia Hook. f. North Island: Egmont-Wanganui Botanical District—Pipiriki, E. Phillips Turner; Mount Messenger, E. P. Turner, H. H. A.

11. Gnaphalium Lyallii Hook. f. South Island: (1) Eastern Botanical District—Rakaia Gorge, on rocks by stream-side in forest, H. H. A. (2) North-eastern Botanical District—Shale Peak on rock in shade, L. C. in 1898. Cheeseman (Manual ed. 2 (1925) p. 963) only records this species from the western side of South Island, from “Collingwood to Milford Sound.” 12. Gnaphalium trinerve Forst. f. South Island: Eastern Botanical District—Rakaia Gorge, on rock by streamside in forest, H. H. A. Cheeseman, (Manual ed. 2, p. 964) has no record for the east of South Island, but he saw H. H. A.'s specimens, and quite concurred in the identification. 13. Helichrysum prostratum Hook, f. North Island: Volcanic Plateau, East Cape, Egmont-Wanganui, and Ruahine-Cook Botanical Districts—abundant on all the high mountains. H. bellidioides (Forst. f.) Willd., so far as we know, does not occur in North Island. Of course it is possible that the true H. prostratum is restricted to the Subantarctic Province, and if so No. 13 requires a name. Cheeseman says of H. prostratum “sparingly found in both North and South Islands,” but we do not know it from South Island. Regarding the matter of sessile to long-pedunculate heads either species may exhibit heads of this character. 14. Helichrysum bellidioides×Glomeratum Ckn. et Allan = in part H. Purdiei Petrie. South Island: (1) North-eastern Botanical District—Valley of the Clarence, Aston! (2) South Otago Botanical District—inland from Dunedin, G. Simpson and J. S. Thomson; near Arrowtown, L. C. 15. Hymenanthera crassifolia×obovata. (a) North Island: Ruahine-Cook Botanical District—Neighbourhood of Wellington, Aston! (b) South Island: Sounds-Nelson Botanical District—Coast near mouth of R. Whangamoa, L. C. 16. Lagenophora cuneata Petrie. South Island: Eastern Botanical District—Mount Peel, in tussock-grassland, H. H. A. We cannot discuss this species in this place, but suggest that it may be the true Lagenopora petiolata Hook. f., in that the heads are much smaller than in either L. pumila or Cheeseman's conception of L. petiolata; but we confess regarding the two last that we are never really sure of our identifications. 17. Leptospermum scoparium Forst. var. incanum Ckn. North Island: North Auckland Botanical District and probably Kaipara Botanical Subdistrict—The type came from near Rangaumu Harbour, but apparently the variety is of local distribution as indicated above. Cheeseman (Manual ed. 2 (1925) 588) gives no localities for the variety. 18. Muehlenbeckia complexa×ephedroides. South Island: North Otago Botanical District—near Kurow, on old flood-plain of R. Waitaki in scrub, L. C. Only one plant was seen but it showed most clearly the combination of parental characters.

19. × Muehlenbeckia muricatula (Col.) Ckn. et Allan as hyb. (M. axillaris×ephedroides). South Island: (1) North-eastern Botanical District—Awatere Valley, Aston! (2) Eastern Botanical District—Along the Ashburton R., coastal to well inland, H. H. A. The above name we only propose for the small group of hybrids intermediate between the two parents (= M. muricatella Col. in Trans. N.Z. Inst. 22 (1890) 482). For the whole hybrid group we intend to give the name×M. axephedra, but without a Latin diagnosis this is only a nomen nudum. 20. Myosurus novae-zelandiae W. R. B. Oliver. South Island: Eastern Botanical District—Near the Ashburton R. on the Canterbury Plain in bare places, H. H. A. This species has also increased to an almost unbelievable degree where the tussock grasses have been destroyed in semi-arid Central Otago (North Otago Botanical District). 21. Nertera Balfouriana Ckn. South Island: (1) North-western Botanical District—The St. Arnaud Mountains, by subalpine streams and tarns, H. H. A. (2) South Otago Botanical District—To west of the New River Estuary in low, flat, moist, sandy ground, L. C. in The Vegetation of New Zealand, 1921, p. 69. (This locality is omitted in the Manual ed. 2). 22. Nothopanax anomalum×simplex. [The hybrid midway between the two parents is×N. parvum (T. Kirk) Ckn. as hyb.] South Island: Fiord Botanical District—The Spey Valley in Nothofagus forest, extremely abundant and polymorphic; W. A. Thomson!, L. C. 23. Nothopanax Colensoi×simplex. South Island: Sounds-Nelson Botanical District—Near the track to the Dun Mountain; F. G. Gibbs and J. P. Lotsy! This is probably Nothopanax Macintyrei Cheesem. in Manual ed. 2, p. 636, which was described from a garden plant. We have seen similar forms, probably close to this hybrid in Nothofagus forest near Murchison, H. H. A., and in forest on the Pikikiruna Range, L. C. (North-western Botanical District.) 24. Olearia albida Hook, f. North Island: Egmont-Wanganui Botanical District—At mouth of R. Mokau; E. Phillips Turner, H. H. A. and K. W. Dalrymple. 25. Olearia arborescens×lacunosa. South Island: North-western, Botanical District—Near source of R. Inangahua, by side of stream, L. C. 26. Olearia Crosby-Smithiana Petrie. South Island: Fiord Botanical District—Mountains near Wilmot Saddle; W. A. Thomson, L. C. 27. Olearia divaricata Ckn. South Island: Fiord Botanical District—Near Wilmot Saddle in open, boggy ground, L. C. Not previously recorded from South

Island but supposed to be confined to Stewart Island. O. Macphersoni Ckn. (name only) in Flora, n.f. 18 (1925) 76 is a synonym. Mr. W. A. Thomson has sent us from near Lake Hauroko a series of forms which appear to be O, divaricata×lineata. One remarkable specimen carries leaves, some of which have the rusty tomentum of divaricata, others the white of lineata, and others again which are intermediate in colouring. 28. Ourisia glandulosa Hook. f. South Island: North-western Botanical District—Mount Mantell, in subalpine herb-field, H. H. A. Only recorded by Cheeseman (Manual ed. 2, p. 838 for certain localities in Otago, but Townson (Trans. N.Z. Inst., vol. 39, 1907, p. 419) recorded it from the Paparoa Mountains (North-western Botanical District). The same, or a closely similar species, occurs by tarns on the St. Arnaud Mountains. 29. Ourisia modesta Diels. South Island: Fiord Botanical District—Exact locality to be given in the next part of this series, W. A. Thomson! Previously only known from Stewart Island, and only from one locality and one plant! 30.×Plagianthus cymosus (T. Kirk) Ckn. et Allan as hyb. (= P. betulinus×divaricatus). North Island: Ruahine-Cook Botanical District—Miramar Peninsula, E. Phillips Turner! Plagianthus betulinus has not been recorded from that locality, but no one can say whether it has been there or the contrary, nor would its absence be evidence that the plant is not of hybrid origin. 31. Pomaderris phylicaefolia Lodd. South Island: Eastern Botanical District—Canterbury Plain, near Horrelville, in the extensive Leptospermum thicket association north of the R. Waimakariri, E. Phillips Turner. This is the first record for this species in South Island. In North Island it reaches from the far north to Otaki and Cape Palliser (Manual ed. 2, p. 555), but is very rare in the Ruahine-Cook Botanical District, and absent in its extreme southern part. To learn of its occurrence in Central Canterbury, while aware of its absence elsewhere in South Island, has occasioned us no small surprise and were the record not from one with such certain knowledge of the species as Mr. Turner possesses we should not have accepted the statement without further enquiry. It is, indeed, one of the most sensational records of distribution that have been made for some years in New Zealand, and it almost equals Mr. W. R. B. Oliver's discovery of Metrosideros Parkinsonii on Great Barrier Island. (Since the above was written Mr. Turner has sent us a specimen from the Canterbury Plain). 32. Suttonia chathamica (F. Muell.) Mez South Island: Sounds-Nelson Botanical District—(1) Red Hill, on the Mineral Belt, A. R. Wastney! (2) Whangamoa Valley, amongst small trees and shrubs on the bank of the river, E. C. Jeffrey and L. C. If this is not as referred above then it is an undescribed species. Only known previously from the Chatham Islands, where it is common, and from one or two localities in Stewart Island—close to the sea.