Other formats

    TEI XML file   ePub eBook file  

Connect

    mail icontwitter iconBlogspot iconrss icon

Tuatara: Volume 13, Issue 1, April 1965

Lunularia in New Zealand *

page 31

Lunularia in New Zealand *

* In this and a following article the New Zealand species of two liverwort genera of importance in elementary teaching are described. The present article considers Lunularia cruciata, and it is suggested that as this species is much more common than the standard textbook example Marchantia, greater use could be made of it in teaching.

In the second article the three species of Marchantia occurring in New Zealand will be described and details given of their occurrence and reproductive cycles.

Introduction

The Liverwort GenusLunularia has only one species, L. cruciata (L.) Dum. (Stephani, 1900). It belongs to the Marchantiales and, although showing an intriguing resemblance to Marchantia in its type of gemma, it is in many respects less specialised and is sometimes placed in a family of its own (Hassel de Menendez, 1962).

Although Lunularia was first described from Europe by Micheli in 1729, it is now known to be widely distributed (Frye and Clark, 1937). In many countries it is adventive, for it occurs in greenhouses and in the shaded parts of gardens. Watson (1959) remarks that it is perhaps not a true native of Britain, as it is rare away from habitations. Smith (1955) considers that it was introduced from Europe to U.S.A. with nursery stock and states that, although widely spread in greenhouses throughout the country, only in areas with a mild climate such as California has it become established outdoors.

Lunularia as described in the literature is morphologically uniform throughout its range except in South America. Here, along with forms differing in no respect from those in Europe, there occurs in Chile, Peru and Argentina a type in which the walls of the dorsal epidermal cells are greatly thickened and the ventral surface of the thallus is coloured dark purple. Although this type has been described as a separate species (Herzog, 1938), Hassel de Menendez by maintaining plants in cultivation has shown that the distinctive features become less pronounced. Also at the margin of its range she found plants of intermediate character (Hassel de Menendez, 1962). She considers there is justification for recognition of a form only and names it forma thaxteri (Evans and Herzog) Hassel de Menendez. Writers frequently remark on the fact that sporophytes are very rare. They have been found once at Cape Town (Saxton, 1930), at San Diego, California (Frye and Clark, 1937), at a few places in southwest England

* In this and a following article the New Zealand species of two liverwort genera of importance in elementary teaching are described. The present article considers Lunularia cruciata, and it is suggested that as this species is much more common than the standard textbook example Marchantia, greater use could be made of it in teaching.

In the second article the three species of Marchantia occurring in New Zealand will be described and details given of their occurrence and reproductive cycles.

page 32 (Saxton, 1930), and in southern Europe (Reader, 1884). None have been found in Argentina (Hassel de Menendez, 1962), in eastern and central U.S.A. (Frye and Clark, 1937), or in northern Europe (Reader, 1884).

Occurrence in New Zealand

In New Zealand Lunularia is abundant at the present time. However, Hooker (1867) makes no mention of its occurrence and, as it seems unlikely that a plant so well-known in Europe would have been overlooked by early collectors, it may be assumed that it has been introduced to this country. It occurs as a troublesome weed of greenhouses and shaded gardens but is also well established as a wild plant amongst native shrub vegetation, in some cases in isolated areas at a distance of up to 20 miles from the nearest farm homestead. Most flourishing colonies were found to form archegonia or antheridia. Sporophytes, however, are not often seen,* although they have been collected by the writer from plants growing in the wild in the vicinity of Dunedin, Palmerston North and Taihape. In one area near Palmerston North they were found to be plentiful during May and June. In all instances, the plants bearing sporophytes were growing at the base of damp, almost perpendicular, rock walls facing south or southeast in deep, winding, canyon-like gorges where, although well-illuminated because of a nearby stream and access roadway, they were protected to some extent from frost, wind and heavy rain and at no time were exposed to direct sunlight, to excessive heat or to air of high saturation deficit. These plants were found to be firmer in texture than plants grown in greenhouses, to have a high lipid content, and in the character of the epidermal cells and occasionally even in the purple colouration of the ventral surface to approach forma thaxteri.

Description of the New Zealand Plants

The thallus of Lunularia is normally green to yellowish-green in colour, but with age it turns brown, either at the edges or all over. Up to 4 cm, or rarely to 7 cm long, and up to 1 cm, or rarely to 1.8 cm wide, it grows flat on the ground or over existing thalli, often forming extensive colonies as it spreads.

Branching occurs by bifurcation of the apex and, when the latter ceases activity, regenerative growth occurs from adventitious shoots. Above, the thallus often appears somewhat glossy and under a lens can be seen to be marked out into polygonal areas each with a

* A further finding of both well-developed and mature sporophytes was made at Rangitoto Island in mid-February 1965, and of sporophytes with dehisced capsules by G. A. M. Scott at Dunedin in early February, 1965.

page 33 central, whitish air-pore. On the under surface there is a row of delicate, hyaline, or occasionally partially purplish scales on each side of the median line, each scale being 2 mm long and broadly lunate in general shape. At the middle of the scale there is a hyaline, or rarely partly purplish, appendage of approximately circular to reniform shape and of diameter 0.4 mm. Both in the scale and in the appendage there occur scattered cells containing brown oil-bodies, and on the margins there are club-shaped mucilage-cells. Where these latter wither off, the cells beneath them, more particularly on the appendage, often remain as a projection so making the edge irregularly toothed. Between the scales there arise two types of rhizoids, namely wide, anchoring rhizoids 30-40 u in diameter which grow vertically downwards, and narrower, tuberculate ones, 7-24 u in diameter, which near the thallus lie horizontally in bundles and distribute the available water evenly over the lower surface.

On the upper surface of some thalli are groups of disc-shaped gemmae lying in a cupule on the mid-line and protected on the posterior side by a crescent-shaped ridge with a crenate, or in old thalli an almost entire margin (Fig. 1). Each gemma has 2 opposite, lateral growing-points (Fig. 5) and is at first attached by a short stalk, but when mature it becomes detached, floats away in water and under favourable conditions grows into a new plant (Fig. 6). The crescentic cupules containing the gemmae are distinctive of the genus and are responsible for its name.

In structure the thallus shows considerable organisation. In the central portion it is 0.5-1.0 mm in depth and gradually becomes thinner towards the margin. The dorsal epidermis consists of colourless or almost colourless cells; sometimes the walls are thin but often those of centrally placed cells have triangular thickenings known as trigones at the angles and in plants from the open all the walls may be evenly thickened, this being particularly noticeable when the thallus is fresh (Fig. 4). Below the epidermis is the photosynthetic tissue consisting of a single layer of air-chambers separated by green, uniseriate partitions and occupied by numerous, erect, green filaments 3-5 cells high. The air-pores leading into the chambers are elevated above the surface of the thallus (Fig. 2) and are simple in structure, not barrel-shaped as in Marchantia. The compact, ventral tissue is 18-35 cells deep in the midrib region but gradually becomes thinner towards the wings; it is a storage region composed mainly of colourless cells with pitted walls, but scattered cells contain brown oil bodies, and septate fungal hyphae similar to those reported from Lunularia in South Africa (Auret, 1930) may be present in a zone of the midrib region. Fungal infection is sporadic and plays an insignificant role in the life of the plant as has been noted in other countries (Ridler, 1923; Auret, 1930; Nicolas, 1924, 1932). In plants growing in page 34 the open the lowermost 1-5 layers of cells contributing to the ventral tissue have thickened walls and brown, or occasionally purplish, colouration. Growth of the thallus originates from Marchantia-type initial cells situated at the base of the apical cavity. Although in vertical sections of the vegetative thallus a single wedge-shaped initial cell is apparent, in horizontal sections two similar rectangular cells are usually recognisable, as recorded also for Marchantia planiloba (Burgeff, 1943).

In regard to sexual reproduction Lunularia is dioecious. The antheridial receptacle is a slightly elevated, flat disc 3-4 mm in diameter surrounded by a circular, membranous, cupule-like sheath with a crenate edge (Fig. 1). Originally terminal on one branch of a dichotomy and alternating as to the side of occurrence on the plant, it soon becomes left behind by the onward growth of the thallus and appears to be situated in a lateral position. The antheridia (Fig. 9) are individually sunken in flask-shaped cavities opening to the surface by a simple pore at the end of a canal. In old receptacles the cells adjoining the cavity and the canal tend to turn reddish-purple. Young receptacles were forming in sheltered situations from May until the end of September, at which time most male thalli undergo regenerative growth and produce cupules. Development of the antheridium (Figs. 7 and 8) was found to follow the pattern shown by other members of the Marchantiales, as indeed Saxton (1930) implies, although other interpretations have been given (Chalaud, 1931). The sperm cells are exuded from the mature ovate antheridium in an opaque mucilaginous mass and lie on top of the disc until wetted, when the spermatozoids become free-swimming. Such masses of sperm cells are readily found on plants in the greenhouse and out-of-doors near Palmerston North from mid-July until September. Under conditions of water deficit, however, the antheridia are halted at various stages in development and remain in situ for many months.

The archegoniophore also develops in a terminal position on one branch of a dichotomy but sometimes owing to continued growth of the vegetative branch it becomes left behind in an apparently lateral position. It alternates as to the side of occurrence on the thallus and makes its appearance from May until September with different populations varying as to the starting time. The archegoniophore lies centrally on the floor of a shallow, circular cupule with a crenate rim. For a long time it remains as a page 35
Fig. 1: Male thallus. x 2. a. antheridial receptacle, c. cupule with gemmae. Fig. 2: Upper part of the thallus in section showing photosynthetic tissue and an air pore. x 230. Fig. 3: Female thallus with mature sporophyte. x 3. c. open capsule, d. disc, s. seta. Fig. 4: Dorsal epidermis in surface view from 3 plants. a. with thin walls, b. with trigone thickenings, c. with thickened walls. x 225. Fig. 5: Gemma. x 35. Fig. 6: Gemmaling at 12 weeks. x 2. Figs. 7 and 8: Vertical sections of young antheridia. x 380. Fig. 9: Vertical section of a mature antheridium. x 75. Fig. 10: Vertical section of an archegonium. x 400.

Fig. 1: Male thallus. x 2. a. antheridial receptacle, c. cupule with gemmae. Fig. 2: Upper part of the thallus in section showing photosynthetic tissue and an air pore. x 230. Fig. 3: Female thallus with mature sporophyte. x 3. c. open capsule, d. disc, s. seta. Fig. 4: Dorsal epidermis in surface view from 3 plants. a. with thin walls, b. with trigone thickenings, c. with thickened walls. x 225. Fig. 5: Gemma. x 35. Fig. 6: Gemmaling at 12 weeks. x 2. Figs. 7 and 8: Vertical sections of young antheridia. x 380. Fig. 9: Vertical section of a mature antheridium. x 75. Fig. 10: Vertical section of an archegonium. x 400.

page 36 slightly greenish, 4-lobed, domed disc attached by an extremely short stalk and hidden from view by a ring of overlapping, white scales up to 4 mm in length which arise near its base on the morphologically ventral side. The scales are rather firm due to thickened cell walls and the inner ones are fringed with white thread-like hairs up to 4 mm long. Similar uniseriate, multicellular hairs occur also within the sheath of scales. Archegonia arise in 4 radiating rows on the disc when it is merely a minute protuberance, each group being enclosed after fertilisation by a tubular involucre and carried to the underside by the expansion of the central dome. Rays such as occur in many Marchantia species are not present. Development of the archegonium occurs as in other members of the Marchantiales and the mature archegonium is typical of this group (Fig. 10).

Fertilisation takes place readily when spermatozoids are transferred in water by means of a pipette, and on occasions so abundantly that some developing embryos are suppressed. It occurs less readily by natural means both in the greenhouse and out-of-doors near Palmerston North.

The general development of the sporophytes has been described by Saxton (1930). However, from the abundant material available it is possible to add some supplementary notes regarding the early stages. The fertilised egg remains dormant for a period of 4 to 6 weeks though recognisable by its slight increase in size and its denser cytoplasm. Meanwhile both the enclosing calyptra and the archegoniophore are enlarging and accumulating much lipid and protein material. Then suddenly the embryo commences active growth while both the calyptra and the archegoniophore continue their enlargement. The first division in the embryo is transverse (Fig. 11), and this is followed by another transverse division in the upper cell so giving a file of 3 cells (Fig. 12), which by further development produce the foot, the seta and the capsule respectively (Fig. 16). In the foot region a vertical division is followed by further growth and by divisions in various planes until there is produced a massive structure which penetrates through the base of the archegonium into the tissue of the archegoniophore. The seta initial cell divides first by two intersecting vertical walls and then a few times in other planes but active growth and division in this region occurs only when the sporophyte is almost mature. In the capsule region the course of development, as pointed out by Saxton (1930), is of an unusual type. Two intersecting vertical walls give 4 cells each of which now divides transversely (Figs. 13 and 14). The formation of periclinal walls cuts off jacket cells from central cells (Fig. 15). Most of the capsule wall is derived from the jacket cells but the extra layers of the cap region are derived from the upper tier of the central cells. The lower tier of central cells cuts off page 37 a tier of sterile cells at the base and then becomes the archesporium composed of deeply-staining cells which at first divide by longitudinal walls only (Fig. 16).

When the sporophytes are nearly mature, the stalk of the archegoniophore elongates in just a fortnight to a height of up to 22 mm, so lifting the disc well above the ring of scales (Figs. 3 and 21). At this stage the whole archegoniophore is conspicuous by its general appearance of whiteness, although on close inspection the green foot of the sporophytes is visible within its tissue. The stalk has no rhizoid-furrow and no photosynthetic tissue. It is 1.0 mm in diameter and is more or less shaggy with scattered,
Fig. 11: A 2-celled embryo. x 260. Fig. 12: A 3-celled embryo. x 260. Fig. 13: Embryo showing a vertical wall in the capsule region. x 260. Fig. 14: Embryo showing vertical walls in all the cells and a transverse wall in the capsule region. x 260. Fig. 15: Embryo after the formation of periclinal walls in the capsule region. x 260. Fig. 16: Older embryo showing longitudinal divisions in the fertile tissue. x 260. c. capsule region, f. foot region, s. seta region. Fig. 17: Top of a ripe capsule in vertical section. x 240. Fig. 18: Very young sporeling showing the first vertical wall. x 380. Fig. 19: Young sporeling viewed from above. x 380. Apical cell is stippled. Fig. 20: Young sporeling in optical section. x 380. Apical cell is stippled.

Fig. 11: A 2-celled embryo. x 260. Fig. 12: A 3-celled embryo. x 260. Fig. 13: Embryo showing a vertical wall in the capsule region. x 260. Fig. 14: Embryo showing vertical walls in all the cells and a transverse wall in the capsule region. x 260. Fig. 15: Embryo after the formation of periclinal walls in the capsule region. x 260. Fig. 16: Older embryo showing longitudinal divisions in the fertile tissue. x 260. c. capsule region, f. foot region, s. seta region. Fig. 17: Top of a ripe capsule in vertical section. x 240. Fig. 18: Very young sporeling showing the first vertical wall. x 380. Fig. 19: Young sporeling viewed from above. x 380. Apical cell is stippled. Fig. 20: Young sporeling in optical section. x 380. Apical cell is stippled.

page 38 white, thread-like filaments similar to those on the scales at its base. The disc also lacks photosynthetic tissue; it is 4 mm in diameter and delicate in texture. From its shape as it expands is derived the specific name of the plant, for the 4 tubular drooping involucres spread outwards from the central portion giving the shape of a cross.

The mature-sporophyte consists of a small green foot, a colourless seta (stalk), and a dark-brown capsule which contains spores together with elaters. It develops within a calyptra and until almost mature is enclosed by the thin involucre. Then on a dry day the involucre unfolds distally to become bilabiate and the seta, which recently has been dividing actively, elongates rapidly to a length of 3 mm, so pushing the capsule beyond the opening (Fig. 3). In several instances there were 2 sporophytes within the one involucre, as was noted also in material from southern England (Saxton, 1930). The capsule is oval in shape and 1 mm long; its wall lacks annular thickenings and is made up of one layer of brown-walled cells except at the top where there is a minute cap 2-3 cells deep of a darker-brown colour (Fig. 17). The ripe capsule opens lengthwise to the base into 4 valves which sometimes begin to divide again lengthwise, but the tiny cap is shed intact. The spores are tetrahedral with a faint tri-radiate marking on the otherwise smooth wall; in colour they are a pale greenish-yellow and in size rather variable with a diameter ranging from 9 to 16 u. The elaters are bispiral, of width up to 10 u and of length 300-430 u, tapering at the ends to a long point. Once the spores are shed the fragile archegoniophore soon collapses.

In order to investigate the germination of the spores these were scattered either on inverted flower-pots filled with sphagnum moss and watered from below with Knop's solution, or in petri dishes on Knop's solution in agar surface-wetted with distilled water. Germination occurs in 10 to 20 days, this being longer than the time given by Chalaud (1932) perhaps due to the fact that the spores dissected from the capsules were not quite mature.

When the spores are well spaced out and the dishes are kept in the light, the course of germination was found to be as follows. The spore enlarges to twice its original diameter and opens in the region of the tri-radiate marking. A colourless rhizoid emerges and is cut off by a wall. The green cell grows rapidly and breaks out of the fragmenting spore coat, dividing first by a more or less vertical wall (Fig. 18), and then each resulting cell dividing both transversely and vertically in a plane at right angles to that of the first division, so giving an octant stage. The sporeling is now approximately spherical in shape but soon becomes irregular when an apical cell arises in one or other of the octants and begins to divide actively (Fig. 20). The apical cell has four faces in contact with neighbouring cells (Fig. page 39 19) and, although for a short time segments are cut off to right and left, the thallus is always more than one cell in thickness. An apical depression soon forms, converted later to an apical cavity, and the typical meristem becomes established. Chalaud (1932) has stated that there is a short filamentous stage but such was found only when sporelings are crowded on the pots, for then the sequence of the early divisions is irregular. Occasionally the first green cell divides transversely, or, more often, after a vertical division, one or both of the resulting cells again divides vertically, either once or a few times in the same plane, so producing a short filament before divisions occur in other planes and an apical cell becomes established.

Discussion

Lunularia as found in New Zealand corresponds with plants described from other countries. Transplant experiments indicate that in New Zealand the form with thin walls in the dorsal epidermis is a shade form developing under glass in greenhouses and in continuous shade outdoors. The form with trigones on the epidermal cells develops under better lighting, either in the open or in hardening-off frames. The form with much-thickened walls is found less commonly, for its requirements of comparatively high light-intensity together with coolness and high humidity are rarely
Fig. 21: Female plant with ripe capsules. Photograph by J. P. Skipworth.

Fig. 21: Female plant with ripe capsules. Photograph by J. P. Skipworth.

page 40 met; even the new thalli arising on such plants in winter, when frequent overcast days give subdued lighting, have thickening in the form of trigones only, and on plants transferred to greenhouses the new thalli formed have thin cell-walls.

Whereas some colonies near Palmerston North produce sporophytes freely, not all do so. Several reasons can be given to account for their absence. Firstly, some colonies consist of sexually immature or juvenile plants with a thin thallus and scanty food-reserves, for in many instances Lunularia behaves as an opportunist spreading rapidly by vegetative means on disturbed soil, on rock or on brickwork before the arrival of other species. Gemmalings are in this juvenile state for over a year, for even under favourable conditions those formed in autumn and winter do not become sexually mature until the winter or spring of the following year. When plants are damaged by dryness of the air or by extremes of temperature or are disturbed by management practices, re-establishment may occur from gemmae previously lodged amongst the rhizoids or in folds of the thallus, or from adventitious juvenile shoots developing on any still living remnants of the original thallus. Under natural conditions near Palmerston North these latter develop abundantly in late spring and to some extent in autumn. Continued repetition of the regeneration process may produce a colony in which the juvenile state is perpetually maintained.

Secondly, the development of antheridia and archegonia is not always synchronised. Female thalli in late autumn, even when green and fleshy and bearing sporophytes, do not continue apical growth but regenerate from approximately 3 strongly-growing adventitious thalli which, after producing cupules, being to produce archegoniophores at various times from May until September. Even when plants are grown alongside one another in the greenhouse, different populations vary as to when they start to form archegoniophores; and in the open the growth in almost all populations is liable to be checked at any time by unfavourable weather. Male plants behave differently as, even when dried out, they often grow again from the resistant apex once the rains commence in late autumn. New antheridiophores arise and old ones revive and continue development but as with the female plants growth may be halted at any time by unfavourable weather. Where, however, the male plants die off in autumn and regenerative growth occurs, the new thalli at first resemble gemmalings and grow very slowly; only in the greenhouse were they sufficiently advanced to form antheridiophores late in the season. In the open they develop so slowly as compared with female plants that the population under these conditions appears to consist solely of female plants. Populations entirely or predominantly of female plants have been noted also in other page 41 countries (Benson-Evans and Hughes, 1955; Campbell, 1918; Goebel, 1905; Goodman, 1956; Müller, 1906-11).

Provided sexually mature male and female plants are growing intermixed, fertilisation was found to occur quite readily. However, the sporophyte takes several months to mature and, since its delicate coverings are inadequate for protection from drying winds, few survive to maturity. In an unheated propagating pit in the greenhouse the sporophytes developed satisfactorily.

Recent experimental work with the Israeli strain of Lunularia has demonstrated a response to photoperiod (Nachmony-Bascomb and Schwabe, 1963; Schwabe and Nachmony-Bascomb, 1963; Wilson and Schwabe, 1964). However, any interpretation of the behaviour of the New Zealand plants in the light of these findings is complicated by the fact that winter temperatures in New Zealand are much lower than any used in the experimental work.

Literature Cited

Auret, T. B., 1930. Observations on the Reproduction and Fungal Endophytism of Lunularia cruciata (L.) Dum. Trans. Brit. Myc. Soc. 15: 163-176.

Benson-Evans, K., and Hughes, J. G., 1955. The Physiology of Sexual Reproduction in Lunularia cruciata (L.) Dum. Trans. Brit. Bryol. Soc. 2: 513-522.

Burgeff, H., 1943. Genefische Studien an Marchantia. Jena.

Campbell, D. H., 1918. The Structure and Development of Mosses and Ferns. New York.

Chalaud, G., 1931. La spermatogenese chez Lunularia cruciata (L.) Dum. Trav. Crypt. dedies a Louis Mangin: 113-126. Paris.

Chalaud, G., 1932. Germination des spores et phase protonemique, in Verdoorn Manual of Bryology: 89-108. The Hague.

Frye, T. C., and Clark, L., 1937. Hepaticae of North America, Univ. Wash. Pub. Biol. 6: 1-162. Seattle.

Goebel, K., 1905. Organography of Plants 2: 80. Oxford.

Goodman, G. T., 1956. Sexual Lunularia cruciata (L.) Dum. in South Wales. Trans. Brit. Bryol. Soc. 3: 98-102.

Hassel de Menendez, G. G., 1962. Estudio de las Anthocerotales y Marchantiales de la Argentina. Op. Lill. 7: 1-297.

Herzog, T. 1938. Contribucion al conocimiento de la Flora Briofita del Sur de Chile a) Parte sistematica. Arch. Esc. Farm. Cordoba 7: 1-56.

Hooker, J. D., 1867. Handbook of the New Zealand Flora. London.

Muller, K., 1906-11. Lebermoose Deutschlands Osterreichs und der Schweiz, in Rabenhorst's Kryptogamenflora, Band 6, Erganzungsbandlieferung 2: 352. Leipzig.

Nachmony-Bascomb, S., and Schwabe, W. W., 1963. Growth and Dormancy in Lunularia cruciata (L.) Dum. I. Journ. Exp. Bot. 14: 153-171.

Nicolas, G., 1924. The formation of mycorrhiza in a thallose hepatic Lunularia cruciata. C. R. Acad. Sci. Paris, 178-228.

Nicholas, G., 1932. Associations des Bryophytes avec d-autre organisms, in Verdoorn Manual of Bryology: 109-128. The Hague.

Reader, H. P., 1884. Lunularia vulgaris. Mich. Midl. Nat. 7: 277-279.

Ridler, W. F. F., 1923. The fungus present in Lunularia cruciata. Trans. Brit. Myc. Soc. 9: 82.

Sexton, W. T., 1930. The life-history of Lunularia cruciata (L.) Dum. with special reference to the archegoniophore and sporophyte. Trans. Roy. Soc. S. Afr. 19: 259-268.

Schwabe, W. W., and Nachmony-Bascomb, S., 1963. Growth and Dormancy In Lunularia cruciate (L.) Dum. II. Journ. Exp. Bot. 14: 353-378.

Smith, G. M., 1955. Cryptogamic Botany. Vol. 2. New York.

Stephani, F., 1900. Species Hepaticarum. Vol. 1. Geneva.

Watson, E. V., 1959. British Mosses and Liverworts. Cambridge.

Wilson, J. R., and Schwabe, W. W., 1964. Growth and Dormancy in Lunularia cruciata (L.) Dum. III. Journ. Exp. Bof. 15:368-380.