Vegetative Anatomy in Donatia, Phyllachne, Forstera and Oreostylidium and its Taxonomic Significance
[Read before the Otago Branch, January 10, 1952; received by the Editor, March 6, 1952]
Anatomically New Zealand species of Phyllachne, Forstera and Oreostylidium are shown to be scarcely distinguishable. Features which set Donatia apart from them are: replacement of stem epidermis by hypodermis, presence of hairs and tannin cells, general lignification of the endodermis, a stele organised in distinct bundles, scalariform perforation plates in vessels, lignification of stelar parenchyma, heptarch roots, sclerenchymatous veins, subsidiary cells accompanying stomata. In conjunction with the floral features it is considered that these differences warrant recognition of a separate family Donatiaceae, but the Campanales is still considered to be the most appropriate order.
The Forsters6 classified their genus Donatia according to the Linnean system. Hooker8 in referring it to the Saxifragaceae states that he is accepting a view previously advanced by St. Hilaire and draws attention to several points of resemblance with the Stylidiaceae. The outstanding difficulty in regarding the genus as Saxifragaceous was the reduced number of stamens (2–3) and their grouping with the styles at the centre of the disc. Von Mueller16 was impressed by the general resemblance between Phyllachne and Donatia and accordingly transferred Donatia to the Stylidiaceae, pointing out that the approximation of stamens and styles should be regarded as a stage in their fusion into a column. In monographing the Stylidiaceae, Mildbraed12 considered that Von Mueller had undoubtedly taken the correct course and divided the family into two subfamilies, Stylidioideae with a sympetalous corolla and column and Donatioideae containing Donatia only and characterized by free petals and free stamens. This extension of the limits of the Stylidiaceae is followed in Cheeseman's flora5, but general works on classification still adhere to the older circumscription and ignore the problem of locating Donatia7,9,13. Chandler4 and Skottsberg14 have preferred to erect a special family for it, the former being satisfied, however, that the place of the Donatiaceae was close to the Stylidiaceae in the order Campanales.
Donatia novaezelandiae in habit and in ecological requirements is almost identical with Phyllachne colensoi. Though the latter appears to prefer better drainage, the two plants grow within a few yards of one another on the summit of Maungatua. Both have assumed a bryoid habit. If they are truly allied, their striking superficial resemblance should be accompanied by many correspondences in anatomy, and the purpose of this investigation was to ascertain how far this is the case. When differences were observed, the other species of Donatia (D. fascicularis Forst., of Chile) and other local members of the Stylidiaceae (P. colensoi and its var. haastii, P. clavigera, P. rubra, Forstera tenella, F. sedifolia, Oreostylidium subulatum) were also examined so as to ensure that impor-
[Footnote] * Now at Southland Technical College, Invercargill.
tance was not attached to features that are inconstant among species or genera whose interrelationships are not in doubt.
The author is indebted to Dr. G. T. S. Baylis for assistance and advice and to Dr. E. H. Walker, of the Smithsonian Institution and the Director, Royal Botanic Gardens, Kew, for material of D. fascicularis.
Certain features were common to all the material. Vascular and cork cambium were lacking and leaf traces were single strands. The angle between successive traces in the stem tip was approximately 135°, indicating a phyllotaxis of three-eighths. This was confirmed in Forstera by leaf counts and in Donatia by mapping the bundle system over several internodes. Lignification of the cortex was negligible, being restricted to a hypodermal layer (Donatia) or to a few outermost (Oreostylidium) or innermost (P. clavigera) layers. Phloem or pericyclic fibre was always absent and conducting cells of the phloem were ill-developed. Leaves were not reticulately veined, but provided only with a midrib from which arose 2 or 4 laterals which turned to run parallel with it. Despite much variation in the degree of cutinisation of the epidermis (between base and apex of the same leaf on the bryoid forms and between the two species of Forstera), in no case was a hypodermis developed, nor did the mesophyll show any marked production of palisade tissue. Since Mildbraed (l.c., p. 15) attached decisive significance to occurrence of inulin in Donatia, microchemical tests to confirm this10 were made with D. novaezelandiae and P. colensoi, in both cases with positive results.
There were several further details of anatomy in which Phyllachne, Forstera and Oreostylidium maintained similarity while Donatia diverged. These are summarized in the following table.
|Stem||Donatia||Phyllachne, Forstera, Oreostylidium|
|Epidermis||Replaced by lignified hypodermis||Persistent|
|Multicellular hairs||On leaf bases and stem tips||Absent|
|Tannin cells (figs. 13b, 13c)||* In stem, root and leaf||Absent|
|Endodermis (figs. 1–4)||*Lignified with passage cells||Caspary's band only|
|Vascular system (figs. 1–4, 13c, 14c)||8 stem bundles||No distinct bundles|
|Vessel elements (fig 5)||Oblique scalariform perforation plates||Almost transverse ends, simple perforations|
|Lignified ground tissue (figs 1–4, 13c, 14c)||Developed in pith and leaf gaps||Absent (except in stolons of Oreostylidium)|
|Stele (figs. 13b, 14b)||Mainly heptarch||Tetrarch or less|
|Sclerenchyma (figs 9–12)||Strand above main veins||Absent|
|Stomata (figs 6–8)||2 subsidiary cells parallel with pore||No subsidiary cells|
[Footnote] * D. novaezelandiae only, this feature being obscure in herbarium material of D. fascicularis.
1—Donatia novaezelandiae T. S. stem stele a—immature b—fully lignified × 180
2—T. S. portion of mature stem stele of Forstera sedifolia × 380
3—Ditto Phyllachne rubra × 380.
4—Ditto, Oreostylidium subulatum a—leafy crown, b—leafless stolon. × 180
5—Vessel elements a—D. novaezelandiae b and c—P. colensoi, d—O. subulatum. × 380 e = endodermis.
Some features mentioned in this table require explanation: The tannin cells of Donatia were very clearly demonstrated by microtoming material killed and fixed in formalin and ferrous sulphate10 and form a more or less continuous system surrounding the endodermis and leaf traces. Unfortunately, dried material of D. fascicularis did not show whether or not this is the same feature as the resin-canal system recorded in this species by Solereder15. The stele of Donatia, whether young (Fig. 1a) or mature (Figs. 1b, 13c), is quite dissimilar from that of the other genera. In young stems there are eight distinct protoxylem groups, while in mature ones the lignified pith and leaf gap parenchyma unite the wood into an unbroken lignified column. In Phyllachne, Forstera and Oreostylidium distinct bundles are never discernible, vessels always being scattered in a cylinder of thin-walled cells except in stolons of Oreostylidium, in which ground tissue of the stele lignifies. Vessel elements in Phyllachne and Forstera differ from those of Donatia not only in a sharply contrasted type of perforation plate but in being exclusively spirally thickened. Bordered pits are, however, plentiful in vessels of Oreostylidium.
Metcalfe and Chalk11 doubted whether anatomical evidence would assist in resolving the taxonomic problem presented by Donatia owing to the diverse and specialized habits of the Stylidiaceae. However, the New Zealand members of the Stylidioideae, despite the fact that they include cushion plants (Phyllachne), a rosette plant (Oreostylidium) and decumbent herbs (Forstera), prove anatomically to be remarkably uniform. The inclusion of Donatia, despite its close external resemblance to Phyllachne, destroys this unity in respect of such important features as vessel, endodermis, and stoma type and general vascular plan. Plainly, importance can no longer be attached to the similarity in habit between these two genera, since anatomically Oreostylidium is closer to Donatia, in addition to having a less specialized anther (Mildbraed, l.c., p. 18). There is justification for Chandler's suggestion that inclusion of Donatia in the Stylidiaceae makes the family an unnatural one. However, anatomical differences do not appear11 to exceed the limits of the Campanales even in its narrowest interpretation9, and because of the occurrence of inulin and the plausibility of interpreting androecium and gynoecium as part of the Campanalean sequence it seems best to continue to regard the Donatiaceae as allied to the Stylidiaceae. Presence of inulin has not, however, decisive taxonomic significance among Dicotyledons. There are records of its occurrence in families far removed from the Campanales, viz. Aristolochiaceae1, Marcgraviaceae3 and Gentianaceae2.
1. Aberhalden, E., 1914. Biochemisches Handlexicon, 8 Band.
2. — 1923. Ibid., 10 Band.
3. — 1931. Ibid., 13 Band.
4. Chandler, B., 1911. Note on Donatia novaezelandiae Hook. f. Notes Roy. Bot. Gard. Edin., 22, 43–48.
5. Cheeseman, T. F., 1925. Manual of the N.Z. Flora.
6. Forster, J. R. and G., 1776. Characteres Generum Plantarum.
7. Gundersen, A., 1950. Families of Dicotyledons.
8. Hooker, J. D., 1847. Flora Antarctica, vol. 2.
9. Hutchinson, J., 1926. The Families of Flowering Plants. 1—Dicotyledons.
10. Johansen, D. A., 1940. Plant Microtechnique.
11. Metcalfe, C. R., and Chalk, L., 1950. Anatomy of the Dicotyledons, vol. 2.
12. Mildbraed, J., 1908. Stylidiaceae. Das Pflanzenreich, 35, 4, 278.
13. Rendle, A. B., 1925. The Classification of Flowering Plants, vol. 2.
14. Skottsberg, C., 1915. Notes on the Relations between the Floras of Subantarctic South America and New Zealand. Plant World, 18, 129.
15. Solereder, H., 1908. Systematic Anatomy of the Dicotyledons. Translated by Boodle and Fritsch.
16. Thomson, G. M., 1880. Note on Donatia novae-zelandiae. Trans. Roy. Soc. N.Z., 13, 289–290.