Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 66, 1937
This text is also available in PDF
(540 KB) Opens in new window
– 71 –

Critical Notes on Celmisia Sinclairii Hook f.

[Read before the Otago Institute, September 10, 1935; received by the Editor, September 11, 1935; issued separately, June, 1936.]


Celmisia Sinclairii was first described in the Handbook of the Flora of New Zealand by Sir J. D. Hooker, the diagnosis being based on material collected and forwarded to Kew by Dr. Andrew Sinclair. This material consisted of three plants gathered in two widely separated localities—Tarndale, in the North-western Botanical District of Cockayne, and Dun Mt., in the Sounds-Nelson Botanical District. Notwithstanding some obvious differences in the characters of the plants forwarded, Hooker regarded them as two forms of the same species, mounted them on the same herbarium sheet, and described them conjointly under the name of Celmisia Sinclairii.

The absence from the Dominion not only of Hooker's type specimens, but of authentic co-types or paratypes of Celmisia Sinclairii and other allied species of Celmisia, notably of C. incana and of C. discolor, has been responsible for the prevalence of some confusion among field workers in New Zealand, more particularly because of the protean character of the genus.

In the case of Celmisia Sinclairii and allied forms New Zealand botanists have never been sure of their identifications mainly for the following reasons:—


As indicated above, no co-type or paratype material has been available for reference in the Dominion.


Until recently no plants matching Hooker's Tarndale specimens had been discovered, or at any rate recognised as such, by any Dominion botanist since Sinclair, and none existed in any New Zealand herbarium.


No recent field-worker has been able on Dun Mt. to find any plants matching or attributable to the Kew specimens of C. Sinclairii described by Hooker from that locality. The fact that plants from Dun Mt. labelled Celmisia Sinclairii in the handwriting of the late T. Kirk are preserved in the herbarium of the Dominion Museum has not hitherto been known.


Plants closely matching the Dun Mt. material at Kew have been obtained elsewhere both in the Sounds-Nelson and in the North-western Botanical Districts, but hitherto this fact was not certainly known to the collectors.


There exist on the mountains of the South Island of New Zealand several allied species comprising numerous jordanons, some of which answer to Hooker's description of Celmisia Sinclairii as set out in the Manual.

– 72 –

Now, although it has long been known that two distinct, forms existed in Hooker's type material, if, indeed, it can be so designated at all, nothing has been known as to their status. The solution of the problem as to whether they were in reality individuals of the same jordanon, or whether they represented two distinct jordanons either of the same or of an allied species demanded their re-discovery in the field, a task that happily has now been accomplished.

Re-discovery of Plants Matching Hooker's Type Specimens.

In January, 1933, whilst engaged in examining the vegetation of the rugged country near the source of the Waihopai River at the eastern margin of the North-western Botanical District in Marl-borough, the writer discovered at an elevation of 1350 metres on Mt. Schiza in the Bounds Range, growing in peaty soil at the base of some cliffs, a Celmisia that answered to Hooker's description of the Tarndale form of C. Sinclairii. On subsequently comparing this with a scrap of the original plant recently forwarded from Kew to the herbarium of the Canterbury Museum at the request of Prof. A. Wall, I was able to verify the identification. Dr. H. H. Allan, who recently had had an opportunity of examining the type material at Kew, has also confirmed the identification.

Subsequent study of the Mt. Schiza plants, both in many localities in nature and in the garden, has convinced me that the original plants sent to Kew from Tarndale were shade-epharmones (Martin, 1935, p. 174), and to this fact in part may have been due the difficulty experienced in re-locating the plant in nature.

Not having had an opportunity of examining the original Dun Mt. material preserved at Kew, I also submitted to Dr. Allan some plants gathered by me on Mt. Z, some miles to the north of Mt. Patriarch, in the Wairau Mountains, which, in view of their accordance with Hooker's description, I had provisionally labelled C. Sinclairii (Dun Mt. form). Of these Dr. Allan in his reply wrote: “These nearly match Sinclair's Dun Mt. plant,” and he has been good enough to forward for examination not only his tracings of the original plants, but also material collected by himself on Mt. Trovatore and on Mt. Mantell, which, he states, match it even more closely. Indeed, there exist on the mountains of the Sounds-Nelson and North-western Botanical Districts a considerable number of closely allied forms of a compound species to which the Dun Mt. plants must be referred, and which I have designated Celmisia Allanii (Martin, 1935, p. 181–2).

Specific Distinctness of the Tarndale and Dun Mt. Plants.

Field-study and garden-culture both confirm the conviction that each of the two forms of Hooker's Celmisia Sinclairii is specifically distinct from the other. The living plants have a very distinct appearance, and each has a distinct though overlapping area of distribution. One is invariably glabrous, the other tomentose on both surfaces of the leaf, a fact not indicated by Hooker, but one a knowledge of which would have prevented much of the present confusion. Dr. H. H. Allan and the late Dr. L. Cockayne have both

– 73 –

concurred with me as to the specific distinctness of the two forms.

It thus becomes necessary to restrict Hooker's name to one or other of the two plants involved. His use of the term “glabrous” before “hoary” or “white below” in the published description suggests my selection of the Tarndale plant as the lectotype of the species, notwithstanding that in citing the localities Hooker refers to Dun Mt. before Tarndale. I understand that Dr. Summerhayes, in a letter to Dr. Cockayne, had expressed similar views. This restriction has the effect of making the Canterbury Museum specimen a co-type of the species. Curiously enough, there is not a single specimen of Celmisia Sinclairii as defined below in any herbarium or garden in New Zealand, albeit it is quite a common plant within its circle of distribution.

Amended Description.

Celmisia Sinclairii Hook. f. (Handbk. N.Z. Flora, 1864, p. 132) sensu restricto.

Suffrutex perennis; caules a basi ramosi; rami breves, primo prostrati demum erecti vel ascendentes, reliquiis vaginarum et foliorum demortuorum persistentibus vestiti. Folia ± 10 a summis ramorum condensata, vaginam includens ± 7 cm. longa, ± 15 mm. lata; lamina obovato-spathulata vel elliptico-spathulata, obtusa vel subacuta, valde tenuia, glabra, viridis; margina integerrima vel remote brevissima dentata; sensim in partem petiolarem angustam contracta, denique in vaginam latam, striatam, glabram ampliata; subtus venae manifestae, costa media conspicua, excurrens; vagina membranacea, fulgida, saepe rubro-tincta base, lanata superna praesertim marginibus, ± 25 mm. longa, ± 6 mm. lata.

Scapus gracilis vix glaber, ± 15 cm. longus; bractae lineari-subulatae, circ. 8–9. Capitulum 3–4 cm. diam. involucri bractae lineares, subulatae, squarrosae, apice recurvatae; interiores bractae involucrae ad apicum ciliolatae, radii ligulae angustae, albae, circ. 8 mm. longae, obtusae, circ. 40; pappi setae graciles, barbellatae, nec apice incrassatae; achenia sericea.

South Island. Tarndale, Sinclair, in Herb. Roy. Bot. Gard. Kew, lectotype (co-type in Herb. Cant. Museum); Mt. Schiza, Raglan Mts., Mt. Bounds, St. Arnaud Mts., Mt. Fishtail, Mt. Richmond, Mt. Z.

Plants of the above species have grass-green leaves not unlike those of Celmisia prorepens, and this green colour is retained by the dried specimens if the process is carried out rapidly, but slow drying produces the brown colour seen in the Kew plant. The leaves of this species are said by Hooker to be more membraneous than those of other species of the genus, but in nature only the leaves of shade-epharmones are noticeably membraneous, and these invariably turn brown in drying.

In the light of my present knowledge of this species, it seems probable that there are not fewer than two distinct jordanons, one east and the other west of the Wairau River, but further research is necessary to establish this as a fact.

– 74 –

Species of Celmisia Hitherto confused with Celmisia Sinclairii.

The numerous forms of Celmisia commonly regarded by New Zealand botanists as C. Sinclairii are without exception most distinct from that species as now restricted; and, indeed, most are distinct from either form preserved at Kew. Numerous jordanons thus excluded have a glabrous or glabrate, green upper leaf-surface, and closely appressed dull or silky tomentum on the lower surface bisected by a prominent, usually glabrous midrib. Many of these belong to the compound species described below under the name Celmisia Du Rietzii.

A study of the various herbaria and plant collections in New Zealand makes it clear that jordanons of the following species have commonly been confused with Celmisia Sinclairii:—


Celmisia Allanii Martin.


Celmisia Cockayniana Petrie.


Celmisia discolor Hook. f.


Celmisia Du Rietzii Ckn. et Allan (spec. nov.)


Celmisia incana Hook. f.


Celmisia spp. (Undescribed).


Celmisia Allanii.

Hooker's Dun Mt. plant which I am excluding from the conception of C. Sinclairii is, in the opinion of Dr. Allan, who has examined it, conspecific with plants collected by himself, by W. Townson, by F. G. Gibbs, and by myself in the North-western and Sounds-Nelson Botanical Districts, but which have hitherto been catalogued as forms of Celmisia incana. These I have recently (Martin, 1935, p. 181–2) described as jordanons of Celmisia Allanii. In selecting as the type of this species a plant collected on Mt. Lockett by Mr F. G. Gibbs, I was guided by the opinion of Dr. Allan that of all specimens he had seen it most nearly matched the Dun Mt. plant discovered by Sinclair. The subsequent discovery by me in the Dominion Museum herbarium of a plant from Dun Mt. itself which closely matches Dr. Allan's tracing of the Kew specimen from the same locality, forces me to refer this latter plant to Celmisia Allanii var. canescens rather than to C. Allanii itself. All forms of Celmisia Allanii are tomentose on both surfaces, but on being dried the leaves in the case of var. canescens acquire a marked glabrous appearance on the upper surface. From C. incana this species is distinguished by the fewer, relatively narrower, longer leaves. The typical form has its involucral bracts more conspicuously reflexed than the typical form of C. incana, and altogether it possesses a very distinct appearance.


Celmisia Cockayniana.

This is a valid species of easy discrimination, distinguished from all forms of C. Du Rietzii by the foliaceous bracts. Its position may be regarded as intermediate between Celmisia hieracifolia and C. Du Rietzii. Hitherto this plant has been known only from Dr Cockayne's original plants gathered on Mt. Fyffe; nevertheless, it is a common

– 75 –

endemic of the eastern and central portions of the North-eastern Botanical District. For further notes on this common rupestral species see Martin (1935, pp. 176–7).


Celmisia discolor.

The typical form of this compound species is that which occurs commonly on the St. Arnaud Mts. A better understanding of this and all related forms is now possible consequent on recent field-work and an examination of the types preserved at Kew, and it is improbable that C. discolor will again be confused with C. Sinclairii or C. Allanii; but it remains a matter of some difficulty to say to which species some jordanons belong, which seem to link C. discolor on the one hand with C. Du Rietzii, C. angustifolia, and allied species on the other. From C. Du Rietzii it differs mainly in the smaller size of the leaves which grow at intervals along the stems, whereas in C. Du Rietzii, C. Sinclairii, and C. Allanii the living leaves are normally tufted at the ends of the stems. The length-to-breadth ratio rarely exceeds 2 in C. discolor, but in the case of the others it is rarely less than 3.


Celmisia Du Rietzii.

This is a compound species comprising numerous jordanons, and ranges throughout the full length of the South Island. It seems certain that in any subsequent revision of the genus this species will be sub-divided into a number of smaller groups each with specific or varietal rank. At present it includes by far the greater number of forms previously listed as C. Sinclairii. All have this in common that they are glabrous or glabrate above and white below. The midrib is usually glabrous and distinct and the tomentum thin and silky, and closely appressed to the leaf. As their specimens clearly show, what both Kirk and Cheeseman had in mind in their published descriptions of C. Sinclairii was undoubtedly that group of plants referred to by Dr. L. Cockayne in the 2nd Edition of his Vegetation of New Zealand under the nomen nudum of Celmisia Du Rietzii (1928, pp. 287–8), which name I propose to adopt for an extensive series of forms incorporated by both Kirk and Cheeseman in their conception of C. Sinclairii. A description of Celmisia Du Rietzii Ckn. et Allan is given below:—

Suffrutex perennis. Caules ad basem ramosi, primo plerumque prostrati demum erecti vel ascendentes, relinquiis vaginarum foliorumque demortuorum persistentibus vestitae. Folia prope apicum ramorum conferta, submembranacea, in petiolum angustam sensim attenuata, denique in vaginam tenuem membranaceam ampliata. Lamina ± 5 cm. longa, ± 1 cm. lata, anguste obovato-spathulata vel elliptico-spathulata, subacuta vel obtusa, minute et obscure dentata, superne glabra vel glabrata, subtus dense et appressa albotomentosa; costa media glabra, conspicua.

Scapi singulares vel plures, plerumque glabri ± 20 cm. longi; bractae circ. 8—9, inferior tomentosae, superior glabratae. Capitulum circ. 3—4 cm. diam. Involucrae bractae numerosae, recurvae. Achenia sericea.

– 76 –

South Island.—Type from Arthur's Pass; collected by me and deposited in the Plant Research Station, Palmerston North. In a broad sense this species is met with on the mountains of the western half of the South Island throughout most of its length.


Celmisia incana.

It has been customary to include in this species all Celmisias of fell-field having both leaf-surfaces clothed with soft, snow-white or silvery tomentum. The North Island plants to which the name was originally given have furrowed, serrulate leaves with a length-to-breadth ratio of from 2–2 ½, but many of the South Island jordanons have a ratio of from 4–6. Hence it has been found advisable to relegate the long-leafed forms to a distinct group to which the name C. Allanii has been given (Martin, 1935, p. 181), and to classify the remaining jordanons into distinct varieties of C. incana. In both species some forms have the tomentum of the upper leaf-surface very thin. When these leaves are-wetted or when they are pressed they not infrequently appear to be glabrous on the upper surface. In this condition varieties of C. Allanii might easily be confused with C. Du Rietzii.


Celmisia spp.

A study of plants collected on the mountains of western Otago points to the existence of several Celmisias specifically distinct from Celmisia Sinclairii, to which they have been referred, and from C. Du Rietzii, to which they must at present be referred. Much more field-work in this area would be necessary, however, before the writer could attempt their separation and description.


To summarise, all forms of C. Sinclairii are glabrous on both leaf-surfaces; all forms of C. incana and of C. Allanii are tomentose on both surfaces; C. Du Rietzii has leaves glabrous or glabrate above and tomentose below; C. Cockayniana differs from C. Du Rietzii in being entirely rupestral and in the possession of foliaceous bracts; while C. discolor differs from each of the above in the possession of small leaves not normally tufted at the ends of the stems.

The writer desires in conclusion to acknowledge the assistance he has received in this research from the curators of the museums at Auckland, Wellington, Christchurch, and Dunedin, who readily granted permission to study the herbarium material housed in these institutions, and from Miss L. Cranwell, M.A., Miss M. Sutherland, B.Sc. (For.), and Rev. Dr. J. E. Holloway, who facilitated that study. The late Dr. L. Cockayne was particularly interested in this investigation and greatly encouraged us in our work by ready assistance and helpful advice. Most of all, however, my thanks are due to Dr. H. H. Allan, without whose collaboration this paper would have been impossible in its present form. He willingly placed at my disposal all the information he had gleaned from a careful study

– 77 –

of the type-material at Kew. To my friend and field companion, Mr J. H. Hadfield, I am grateful for assistance in the cultivation of the various species of Celmisia involved in this research.

Literature Consulted or Cited in this Paper.

Cheeseman, T. F., 1881. Contributions to a Flora of the Nelson Provincial District, Trans. N.Z. Inst., vol. xiv, pp. 301–329.

——1928. Manual of the N.Z. Flora, 2nd Edition.

Cockayne, L., 1919. N.Z. Plants and their Story, 2nd Edition.

——Cockayne, L., 1928. The Vegetation of New Zealand, 2nd Edition.

Hooker, J. D., 1867. Handbook of the N.Z. Flora.

Kirk, T., 1899. Students' Flora of N.Z.

Martin, W., 1935. Species and Varieties of Celmisia Indigenous to Marlborough with Descriptions of New Species and Varieties. Trans. Roy. Soc. N.Z., vol. 65, pp. 169–185.