The Mechanism of Leaf-Fall in Certain New Zealand Trees

By R. S. Russell, M.Sc., University of Otago.

From the phylogenetic viewpoint Bews (1927) has considered the origin of the deciduous habit. He suggests that this habit with its specialized leaf-abscission mechanism evolved from the evergreen habit as a physiological response to seasonal periodicity.

Lloyd (1916 b) and Sampson (1918) have discussed the external factors influencing leaf-fall.

Little has been written on the leaf-shedding of New Zealand trees. Rutland (1888) published a paper based on limited field observations, while Cockayne (1904 and 1906) touched briefly on the subject. More recently Cockayne (1928, pp. 143–6) supplied a list of deciduous species which, though not complete, contains the most outstanding New Zealand examples.

Aristotelia serrata and Fuchsia excorticata grow abundantly in the Dunedin “Town Belt,” where the material used in this research was procured. Material of Hoheria Lyallii, which is not found in the native state near Dunedin, was obtained from the Dunedin Botanical Gardens. In each case the trees were kept under observation for twelve months, and material was collected at regular intervals. The investigation of leaf-fall in Nothofagus has been less extensive.

Formalin-acetic-alcohol was the most frequently used killing and fixing reagent. Microtomed sections, prepared by the usual paraffin method, were employed throughout. Considerable difficulty in obtaining good paraffin infiltration was remedied by a method of gradual infiltration in which dehydrated material was kept in xylol for a week at room temperature under a copper gauze table supporting a block of slowly dissolving paraffin. In Fuchsia complications arose due to air held in the dead cortical cells, but were largely overcome by frequent evacuation. The “Cambridge Rocker” microtome was used extensively, but the more rigid “sliding” microtome was preferable for large woody objects. The majority of preparations were stained with safranin and light-green, several other well-known stains being also used. Numerous microchemical tests were employed, especially in the examination of the Protective-layer and the disorganization of the Separation-layer.

All the drawings illustrating this paper were made with the aid of either an Abbé camera lucida or a Leitz reflex drawing apparatus, while the photomicrographs were made using Imperial “Fine-grain” plates and appropriate “Wratten” filters.

The research on which this paper is based was carried out in the Department of Botany, University of Otago. To Dr. J. E. Holloway, who suggested this investigation and has given helpful encouragement and advice throughout, the writer owes a deep debt of gratitude. For the use of photomicrographic apparatus he desires to record his thanks to Professor Bevan Dodds, Dean of the Otago Dental School.

The Leaf-Abscission Mechanism

It is obvious that whatever may be the cause of “dipping periderm,” it will assist in the separation of the leaf by diminishing the crosssectional area of the plane of abscission.

The Scar Protective Mechanism. No protective mechanism has developed at the time of leaf-fall, but very soon after changes may be detected in the cell-walls of the uninjured cells immediately beneath the scar. The zone of cells whose walls are transformed is usually 3–5 cells deep and is known as the Protective-layer (Fig. 7). A microchemical examination has proved the walls to be lignosuberized. Tests for lignin and suberin show the entire cell-wall (including the middle lamella) to be lignified with a film of suberin deposited as an inner film on the wall. In some cases the middle lamella has been seen to be stained more deeply by phloroglucin than the lamellae on either side. The suberin forms a distinct superficial layer. When lignosuberization is complete no cell-contents remain, but a few granules which take the suberin stains adhere to the wall. It was possible to remove the lignin and cellulose by mounting in concentrated sulphuric acid a section which had been treated previously with Eau de Javelle. After this treatment the suberin lamellae alone remained. At least some of the tylosis walls in the region of the Protective-layer also become lignosuberized.

After the Protective-layer has developed, cells immediately beneath it become meristematic and divide by walls parallel to the surface of the scar to produce the scar Protective-periderm which is continuous with that of the stem (Fig. 12). Across the xylem of the leaf-trace the periderm is continued by a division of tyloses (Fig. 8). In the figure it will be seen that many segments have been cut off from the tyloses, and the xylem vessels are consequently ruptured. By this time the phloem elements are so compressed that they can scarcely be recognised in the neighbourhood of the scar-periderm. Within three weeks of leaf-fall the periderm initials have appeared, and scar protective mechanism is complete. Several abscissed axillary buds have been observed, the sequence of changes being similar to that of leaf-falls. The leaf and bud scars frequently appear quite continuous.

(b) Aristotelia serrata, J. R. et G. Forst.

Trees of Aristotelia serrata growing near Dunedin are not entirely without leaves at any season of the year. However, there is a genuine though most indefinite period of leaf-fall which may be extended over four months or even longer. Aristotelia serrata therefore has a low degree of deciduousness. Field-observations show that under local conditions the leaf-fall is more marked and occurs earlier in exposed situations than in shaded ones, but in all cases apical growth is more or less suspended from June to September.

Early in June leaf-fall commences in exposed situations when only a few leaves are affected, but the leaf-fall continues steadily, so that there are always a few yellow leaves on the tree. In the middle of September, when many trees show but few old leaves remaining, the resting buds resume vigorous growth. In shaded positions the leaf-fall is more gradual, and commences later, while the previous season's leaves linger on until December or even longer,

by which time the new season's leaves have fully developed. Cockayne (1928, p. 146) states that this tree varies from evergreen to deciduous, and its deciduous habit is stronger in more frosty localities.

Anatomical. Since the leaf abscission mechanism of Aristotelia serrata is closely similar to that of Hoheria Lyallii described above, it will be necessary merely to mention the points of difference between the two species. The triple leaf-trace of Aristotelia is in arrangement similar to that of Hoheria, but it leaves the cauline stele more sharply. Slightly more sclerenchyma is present, and occasionally it continues up into the petiole, while calcium oxalate crystals are not as abundant as in Hoheria. In its origin and development the Separation-layer (Figs. 13–16) is exactly similar to that of Hoheria, but its extent is considerably less, being only 5–8 cells deep, while in Hoheria the depth is 12–20 cells. Tylose development (Figs. 17–19) is not so extensive, but the Lignified-layer is more strongly developed. The dissolution of the cell-walls (Fig. 16) is a slower process, though the changes in the two species are identical in nature. Following the disorganization of the parenchymatous cells leaf-separation is effected by the rupture of the leaf-trace.

After leaf-fall the Protective-layer and periderm develop similarly to those in Hoheria, but in Aristotelia the cauline periderm has not developed at the time of the leaf-fall. The Protective-periderm is thus at first unconnected with any other development in the stem. Eventually the cauline periderm develops and becomes continuous with the Protective-periderm. This process is not complete till five or more years after leaf-fall.

(c) Fuchsia excorticata, J. R. et G. Forst.

Fuchsia excorticata is undoubtedly the best-known New Zealand deciduous tree. Accurate field-observations over a large area are wanting, but it is probably more consistently leafless in winter than any other tree. In Dunedin, during the winter of 1934, it was the only native tree entirely leafless. Leaf-fall commenced in trees in all situations during the early part of May, and was complete within three or four weeks. The trees were without leaves during the greater part of June, throughout July, and in early August, the new leaves making their appearance before the middle of that month. Exotic deciduous trees cultivated in Dunedin show a much longer leafless period. Cockayne (1928, p. 146) states that F. excorticata “may fluctuate, according to circumstances, from truly deciduous to evergreen,” but no case of this tree being leafy in winter has been met with during the present study.

Forced Leaf-fall. A series of experiments, similar to those of Sampson (1918) have been performed and have yielded corresponding results. Two series of experiments were carried out. Series I was carried out about three months before leaf-shedding would normally commence. The lamina of every second leaf was removed from a number of healthy twigs, and the time of shedding of normal and mutilated leaves was noted. After fifteen days 97% of the

mutilated and 2% of the normal leaves had fallen. Series II was conducted nearer the time of leaf-shedding and yielded similar results.

Anatomical. As in the description of Aristotelia serrata, only points of difference from the mechanism already described will be discussed. In Fuchsia the cauline periderm is deep-seated and develops before leaf-shedding (Fig. 20). This is responsible for several modifications in the abscission-mechanism. The leaf-trace consists of a single vascular strand unaccompanied by sclerenchyma. The Separation-layer is similar in extent to that of Aristotelia (Fig. 21). Cell-division is somewhat more regular, the majority of new walls being perpendicular to the axis of the petiole, while the Lignified-layer is developed only within that portion of the leaf-base which is inside the cauline periderm. Rows of mucilage cells develop in the parenchyma of the vascular strand, and tylose development is greater than that observed in any of the other species investigated. Several tyloses were seen which appeared to be cut off from their parent cells by cross septa, a clear example of this being illustrated in Fig. 22. There seems no doubt that this tylosis is cut off from its parent cell by the septum, as two facts preclude the explanation that there are two distinct tyloses in contact. In the first place, the use of the 1–12 inch oil-immersion objective failed to show the septum to be double, and its thickness appeared to be the same as that of the remainder of the wall surrounding the tylosis. In the second place, a slight triangular thickening is seen at either side of the septum where it joins the outer wall of the tylosis. Such thickenings occur where three walls join, but would not be present if the internal wall were merely the result of two tyloses being in contact.

By the time the leaf falls the cortical cells outside the cauline periderm are in most cases dead. Death of the remaining cortical cells soon follows, and is preceded by the lignosuberization of the cell-walls. The Protective-layer and the Protective-periderm develop only in the tissue within the cauline periderm, and, in nature, are identical to the corresponding mechanisms of the other species.

(d) Nothofagus Species.

The genus Nothofagus is widely distributed in the Southern Hemisphere, five species being in New Zealand, three in southeast Australia, and eight in southern South America. An evergreen section is recognised with five New Zealand, two Australian, and three South American species. The remainder of the genus is deciduous.

Field observations have shown an interesting series of conditions, intermediate between the evergreen and deciduous habits, to occur in the New Zealand Nothofagus species. Periodic leaf-shedding is most pronounced in N. fusca, which on occasions show a much more marked leaf-shedding than does Aristotelia serrata (see above) in the same localities. In all species the leaves are always abscissed before the end of the summer following that in which they develop.

The Changes in the Nature of the Cell-walls in the Separation-layer during Abscission.

The changes which occur in the nature of the cell-walls during the develoment of the separation-layer have been investigated in equal detail in Hoheria Lyallii and Aristotelia serrata. No differences have been found between these species either in the original nature of the cell-walls or in the changes which take place in them. In Fuchsia excorticata a less detailed examination has been made, but so far as it has been studied the sequence of changes is similar to that in the other two species. An account of these changes will now be given:—

In material stained with safranin and light green, changes in the appearance of the cell-walls are first observed near the time when the leaf turns yellow. The regular outline of the walls is lost, their staining becomes indefinite, and finally at one level the tissue is seen to disorganise. Closer examination reveals that each cell-cavity in the disorganized region is surrounded by a thin more or less distorted cellulose membrane which takes the green stain. The membranes of individual cells are independent of each other, and by their separation and rupture the severance of the petiole from the twig is effected. Special staining methods shed light on this process.

If sections showing swollen cell-walls are stained with ruthenium red and counter-stained for a very short period with light green, it is found that a central region of the walls has stained strongly with ruthenium red, indicating the presence of pectic substances, while a thin layer on either side takes the green stain, suggesting the presence of cellulose. Occasionally a darker red is taken by a band in the centre of the cell-wall—the middle lamella—than by the remainder of the red-staining region. As leaf-fall is approached the red-staining zone is found to increase, but no alteration is observed in the green-staining membranes. Normal cell-walls stained by the same method show a central region deeply stained with ruthenium red on either side of which is a region stained less strongly, and outside this again is a green-staining layer of the same thickness and appearance as that found in the swollen cells of the Separation-layer. In some cases it was impossible to distinguish between the two red-staining regions. If sections are treated with cuprammonia

In an attempt to determine the nature of the pectic substances present in the walls of the cells of the Separation-layer the following experiments were carried out on material of Aristotelia serrata. Sections from a leaf-base shortly before leaf-fall were treated with 2.5 per cent. ammonium oxalate solution at 50° C. for 24 hours, and at 100° C. for a short period. The sections thus treated showed the same staining reactions to ruthenium red as they had previously. Similar sections were treated with one of the following reagents:—5 per cent. hydrochloric acid, 5 per cent. sodium bicarbonate, and 3 per cent. caustic potash. In each case the sections were heated in the reagent for an hour and a-half on top of a hot-water bath. The material treated with acid was stained with ruthenium red after thorough washing, and showed no alteration in the amount of pectic substance present. When the potash-treated material was similarly stained it was found that little or no ruthenium red was held by the Separation-layer, though outside the Separation-layer the staining was normal. In cases where a certain amount of ruthenium red was held by the Separation-layer the part stained was invariably a thin band in the centre of the wall—the middle lamella. The material treated with sodium bicarbonate behaved similarly.

It was not expected that the material treated with ammonium, oxalate would stain deeply with ruthenium red, showing the pectic substances present to be insoluble in that reagent. It was found that the same result was obtained if the material was treated with acid after the oxalate. Failure to obtain the removal of the middle lamella under these conditions has also been experienced by Davison and Willaman (1927). Treatment with caustic alkalis, carbonates and certain other reagents is a standard method for removing pectic acids. It appears probable from the above evidence that, shortly before leaf-fall, the pectic substance in the cell wall is pectic acid and that the middle lamella, which is at first not composed of an alkali-soluble substance, is finally converted into pectic acid also.

The facts established here are in agreement, so far as they go, with the findings of Sampson (1918) quoted previously. However, it must be remembered that Sampson assumed the middle lamella to be composed of calcium pectate—a view which has now been largely abandoned (Nangi, 1925; Norris, 1925). No light has been shed by the present investigation on either this question or the mode of conversion of cellulose into pectic acid, but it has been shown that, during the development of the Separation-layer in the species here investigated, the secondary membranes of the cell-walls are converted directly or indirectly into pectic acid, and that finally the middle lamella is also converted into pectic acid. Separation of the cells is effected by the pectic acid becomng mucilaginous and thus freeing the individual cells surrounded by their tertiary membranes.

The Deciduous Habit in New Zealand.

The climate of New Zealand is of an equable rather than a truly alternating type, and these islands are thus able to support an evergreen flora. For this reason Rutland (1888) and Cockayne (1928, p. 143) have assumed that the leaf-shedding of the indigenous

that several of the New Zealand genera which show the deciduous habit belong to the South American “element” or the old Antarctic element in the flora, e.g., Fuchsia, Discaria and Muehlenbeckia, and also Nothofagus with its slight degree of deciduousness. These genera may be considered to have occurred in Tertiary times on the Antarctic continent and adjacent lands. If the deciduous habit had been possessed by the flora of that time we might reasonably expect it to be surviving here and there in the floras of southern South America and New Zealand. Little data on the South American flora are available to the writer, but it is known that of the seventeen species of Nothofagus six are truly deciduous, and of these five are in South America and one in Australia, while the New Zealand species (see above) show a mild degree of deciduousness. If the origin of the deciduous habit in this genus were of later date than the Tertiary it must have evolved independently in two distinct regions, and a marked tendency to it must have developed in a third—developments which would be unlikely in largely evergreen floras. The same is the case in Fuchsia, as some of the South American Fuchsias show leaf-fall. It must be remembered that, if the climates of the three regions were of strongly alternating type, this parallel development of the deciduous habit would be expected.

The fact that the South American genera in New Zealand and vice versa are not all deciduous could be explained by the assumption that the deciduous habit has been gradually abandoned in a more or less equable climate. That this is possible has been shown experimentally by Flammarion (quoted by Hicks, 1928). On account of the greater climatic alternation in the situations in New Zealand in which Hoheria Lyallii and certain other deciduous species grow, other things being equal, they would be likely to retain the deciduous habit longer than lowland forest-plants.

This discussion suggests the following tentative deductions, which, it must be remembered, are highly speculative:—

• (i) That the present climate of New Zealand is in certain habitats favourable to the deciduous habit.

• (ii) That the deciduous habit is probably not of recent origin in New Zealand.

• (iii) That, although there is at present little information to suggest the past history of the deciduous element, it might conceivably have originated in Tertiary times on the Great Antarctic Continent, and since then been gradually suppressed (so far as New Zealand is concerned) by the more equable climate.

Discussion

When the leaf-fall mechanisms of the New Zealand trees which are here described are compared with those found in Northern Hemisphere deciduous trees it will be seen that in all cases the mechanism is fundamentally of the same type. No useful purpose would be served by making a detailed comparison with individual types described by Lee (1911) and other workers, as the details of abscission mechanism are largely determined by the anatomical