The Nutrition of Mycorrhiza Plants: Calluna Vulgaris

In a paper published in 1915, the writer put on record the existence of an obligate relationship between the Ling (Calluna vulgaris) and its mycorrhizal fungus.

Recognition of the obligate nature of this association for one of the partners followed upon an experimental investigation into the behaviour of the two constituents when in “pure culture” and when brought together at a critical stage in the development of the Ling seedling. The results then recorded, and since that time often repeated and confirmed, provide conclusive evidence that, under the conditions described, the development of the seedling of Ling in all but the earliest stages is bound up with infection by the mycorrhizal fungus specific to the plant.

From observations on germinating seeds it is believed that a similar relationship exists in nature, a view not inconsistent with the belief that, under experimental conditions, it may prove possible to replace the stimulus to growth which normally follows infection by the addition of an appropriate organic substance to the rooting medium. That such requirement on the part of seedlings may be fulfilled exceptionally in nature is not impossible ; but no satisfactory evidence is at present known to the writer that plants of Ling ever occur under natural conditions or under cultivation entirely free from specific fungal infection (see p. 279 for paper by Christoph).

The unique feature of the association is not its obligate character, a feature shared by a majority of Orchids, but the wide distribution of the mycelium of the endophyte throughout the sub-aerial parts of the Calluna plant and its eventual appearance in the chambers of the fruit and upon the testas of ripening seeds. The provision of a mechanism for reinfection of the next generation of seedlings, and the method by which this is accomplished, are features peculiar to Ericaceæ and have been fully elucidated only in Calluna (cf. Lolium, McLennan, 1919).

Ignorance of the precise physiological relations between the constituents in mycorrhiza plants is still profound. In current botanical literature the latter are commonly described as examples of symbiosis between vascular plants and fungi, some degree of mutual benefit being assumed. The use of the term symbiosis in this sense to describe cases of association which have not been investigated experimentally is obviously unsatisfactory, and the whole subject needs revision in the light of modern experimental research. For this reason and also because the relations between vascular plant and fungus in Ericales are in themselves of singular interest, it has seemed to the writer worth continuing the work on Calluna in the hope of obtaining a working knowledge of the physiological basis of the relation in a case for which the main facts are already experimentally established. Unfortunately, investigations of this kind are difficult and laborious; the isolation of the endophyte, its maintenance in pure culture, and the production of synthetic plants all make demands on a complicated and specialised technique; while the difficulties of cytological investigation are increased in Ericaceæ by the fineness of the roots in which mycorrhiza are formed and the abundance of material of a fatty nature in the tissues.

During the course of the work, experimental confirmation of the wide distribution of mycelium in the shoot was obtained by studying the behaviour of shoot cuttings “struck” under controlled conditions. Publication of these observations is called for in view of a paper by Christoph (1921) published before that worker had access to the present writer’s earlier 266 paper on Calluna (1915). The conclusions reached by Christoph are greatly at variance with those expressed in the latter, and also with the experiments and observations on cuttings now recorded. They are discussed in a later section of the present paper.

The bionomics of mycorrhiza in Calluna are rendered difficult of investigation by the fact that it is impossible to cultivate the Ling plant in the absence of its fungal associate. The nutritive relations of the two plants can be studied therefore only by indirect methods supplemented by cytological observations. An intensive study of the (root) mycorrhiza was undertaken, because it seemed an essential step towards a better understanding of the physiological relations of which it is evidently an important expression.

The use of the term mycorrhiza to describe an association which extends to all organs of the vascular plant raises difficulties. “Root-mycorrhiza” is tautological. In what follows the term is used only in relation to the typical composite root structure. The facts already known may be briefly summarised as follows:—

The endophyte has been isolated and identified; it belongs to the genus Phoma, various species of which are parasites upon seed plants. Some indirect evidence has been provided by the writer (1922) that this fungus species can utilise atmospheric nitrogen when growing in association with the Calluna plant ; moreover, the work of Ternetz (1907) has provided direct experimental evidence that the biologic forms of Phoma associated with Calluna and a number of other ericaceous species all fix atmospheric nitrogen in greater or less degree. The reaction to certain media is also known, as is the fact that fruit bodies are not formed when the fungus is associated with healthy roots, while in pure culture pycnidia formation (sporing) is controlled by the substratum on which the fungus is grown, the age and history of the culture used for the inoculum, and other conditions. Further data are required as to the length of time that the endophyte can be maintained outside the plant, its reaction to temperature and to various nutritive substances, in particular those known to be present in the tissues of Calluna. For example, in view of the recent observations of Pearsall (1922), Priestley and Hinchliff (1922), and Priestley (1924), it would be of interest to learn the behaviour of the endophyte towards fats.

Turning now to the Calluna plant. Failing infection soon after germination, the seedlings are apparently incapable of further growth. The effect of infection is immediate and remarkable ; roots are quickly formed, the shoot develops, and production of chlorophyll is stimulated. Growth of mycelium reaches a maximum in the roots but extends throughout all organs of the seedling, the degree of development being closely correlated with the tissues concerned. Under experimental conditions the co-ordination normally established is easily disturbed, e.g. by altering the substratum, and may swing over towards parasitism on the part of the fungus. In all living tissues except those of the roots (and in air spaces elsewhere), mycelium is present in an attenuated condition; for this reason its presence is extremely difficult to establish with certainty in the shoot tissues. In young roots typical endotrophic mycorrhiza is formed, and it is only in this region that cytological evidence of metabolic exchange can be sought with any hope of success.

For reasons probably not unconnected with difficulties of technique, the roots of Ericaceæ have received less attention than has been the case with other mycorrhiza plants. In the paper already referred to (Rayner, 1915), previously recorded observations as to the distribution and appearance of the mycorrhiza were confirmed, but no special attention was given to cytological details. It was not until the autumn of 1923 that the writer, examining carefully stained material, noted that digestion of fungus mycelium was a common phenomenon in Calluna mycorrhiza.

The present communication is concerned with a seasonal study of the mycorrhiza and an account of the extensive digestion of mycelium now recorded for the first time. With this is included an account of experimental work on “cuttings” and a summarised statement of the available evidence for exchange of nutritive material between vascular plant and fungus.

The most striking result has been the observation of regular and well-marked digestion of the mycelial constituents of the root mycorrhiza. This phenomenon, common in mycorrhiza 368 plants generally and so well known in the roots of Orchids, has not been recorded previously for Ericaceæ either by the writer or by earlier observers. It was at first thought that a seasonal periodicity in the cytological relations might exist corresponding to periods of greater or less activity on the part of root-cells or mycelium (Rayner, 1923). This has not proved to be the case and, except for quantitative differences easily accounted for by seasonal variation, the “balance of power” between endophyte and root-cell, as deduced from cytological observations, appears to be uniform throughout the growing season. Active digestion takes place throughout the vegetative period; cells showing various stages are relatively abundant in young roots, and the process results in complete breakdown of the mycelial complex within the cortical (epidermal) cells.^* Digestion is not a phenomenon of senescence. It begins soon after the production of young roots in spring, is carried on throughout the growing season, during which mycelial activity also reaches a maximum, and continues until growth ceases in late autumn.

The amount of typical mycorrhiza formed is variable. Under certain conditions, especially in early spring, a majority of young roots may be clean, transparent, and apparently entirely or relatively free from fungal infection. Microscopic examination of such roots after special treatment almost invariably yields evidence of infection, though typical intracellular mycorrhiza is absent. Such uninfected or lightly infected roots are common in early spring at the start of root activity, but become relatively infrequent as the season of maximum vegetative growth approaches.

Clean roots are also characteristic of plants growing in sand or potted in sterilised (autoclaved) peaty soil. Much attention has been devoted to observations on this matter in view of their bearing on cases of supposed non-infection of roots, more especially those of cuttings. No evidence has been obtained that more than one fungus is concerned in the formation of mycorrhiza.

Material has been examined at intervals of a few weeks throughout the year over a period of two years in view of the possibility of seasonal changes, and it is believed that the condensed account of these observations now offered conveys a correct impression of the general sequence of events in Ling. It is based mainly on observations upon plants from the S.E. of England. Since these fit into a consistent scheme, it is believed that divergences, if observed elsewhere, can probably be related to local differences of climate and soil.

The work entailed in these observations has been considerable, but has been essential in order to understand the possibilities of nutritive exchange between the two plants.

The experimental work on cuttings has provided ample confirmation of the distribution of the endophyte throughout the shoot tissues. The liability to infection of adventitious roots as they emerge has been confirmed. Observations on the lack of typical mycorrhiza in older roots of cuttings grown in sterilised soil have been interpreted in the light of those on roots generally, and have provided a reasonable explanation of the discordant conclusions reached by Christoph.

Material for observation on roots was obtained from as many stations as possible. The most convenient method of procuring clean roots is to grow plants in pots of peaty compost plunged in ashes or other suitable material, removing roots as required from the outside of the ball of soil. This material has been supplemented by roots from plants cultivated in the open and from others growing naturally in typical calluneta. In my experience, the observations recorded are common to all healthy roots, the exact conditions varying with age, the season, and soil conditions.

Roots have been examined fresh, in lactic-phenol, and after fixation by 1 per cent osmic acid, Carnoy’s fluid and Flemming’s fluid. Fixed material has been stained in many different ways : lactic cotton blue, carbol thionin blue, etc., etc. In general, for cytological examination the most satisfactory results have been obtained by fixation of young roots in Carnoy’s fluid, staining in Heidenhain’s hæmatoxylin, and examination in cedar wood oil: except for observation of special details, whole roots have been found more satisfactory than microtome sections.

For observations on mycelium in the shoot, material has 270 been fixed in Carnoy’s fluid. Many methods have been tried in the hope of obtaining good differential staining of hyphæ in these tissues. Some success has been obtained with the following: cotton blue, picro-nigrosin, picric-aniline blue, Bismarck brown and methyl-violet (Huber method), hæmatoxylineosin (Durand’s method), aniline blue and vesuvin, or aniline blue and orseillin (Strasburger method), and a modification of the last method by substitution of cotton blue for aniline’ blue. After staining, sections were mounted in Canada balsam or (preferably) in Euparol.

Wratten colour screens in various combinations were used for stained sections. A pointolite lamp was employed as a source of light, care being taken that the illumination was critical. Without the two latter precautions, the identification of mycelium in the tissues of the stem becomes a matter of extreme difficulty.

The absorptive root system of Ling consists of a profusely branched mass of threadlike roots, formed annually by the outgrowth of laterals from the older roots; any or all of these young roots become mycorrhiza. In early spring, microscopic examination shows shrinkage of the tips and general exfoliation of the outer tissues, and it is evident that roots of the previous year do not function as absorbing organs during a second season. Hence, mycorrhiza is reformed annually by infection of young lateral roots developed in large part from the youngerregions of those of the preceding year. Actual dates have little significance, varying doubtless with the climate and season, but the general sequence of events is believed to be as follows.^*

As already noted, spring roots may appear to be uninfected, the duration of this phase varying with the season and possibly with variation in other conditions also. As the season advances, the typical conditionis re-established in which a majority of the outer cortical cells of all young roots become filled each with a densely branched “witches broom” of mycelium (Rayner, 1915, fig. 11). As judged by the frequency of cells showing stages in the process, digestion of mycelium continues actively throughout the vegetative season.

As might be expected, there appears to be a reciprocal relation between rate of growth of roots and intensity of infection. Thus, the clean and comparatively uninfected condition of roots in spring is held to be directly related to their active growth at a soil temperature unfavourable to activity of the mycelium, i.e. the temperature at which active growth is initiated is lower for the vascular plant than for the endophyte. Experimental evidence for this view was obtained by noting the earliest signs of root activity in Calluna plants growing in the open, and comparing with this the rate of growth of the endophyte in pure culture in various media, including Calluna root extract, when the culture vessels were buried to a depth of 6 inches in peaty soil adjacent to the plant roots. The photograph (fig. 1) indicates the relatively feeble growth of the latter cultures during the period 18th March to 1st April 1924, at a time when roots in the immediate neighbourhood were already making active growth. The daily soil temperatures for the district during this period were as follows: 36°, 36.3°, 36.7°, 37.7°, 39.5°, 43°, 45°, 44°, 44.7°, 42°, 40.7°, 38.7°, 39°, 38°, 38.7°. The controls were kept at a temperature of 6o° F.

Despite the absence of typical mycorrhiza, careful staining and examination under high magnification of these comparatively uninfected spring roots usually reveals the presence upon the surface and within their cells of fine hyphæ, the feeble staining capacity of which is regarded as further evidence of the relative inactivity of the mycelium at this stage. In addition to lack of vigour of the mycelium, there must also be taken into account a possible decreased “attractiveness” of the root-cells at this season, since rapid growth, especially in early spring, is incompatible with accumulation of reserves in the cortical cells.

Another set of factors is believed to be of importance in maintaining the partial and temporary immunity of roots under certain conditions. Based on numerous observations on roots in various stages of growth and at all seasons of the year, the view has been reached that although hyphæ may enter, mycorrhiza can only be established under certain conditions of the root-cells. There is, in short, a reciprocal relation of a very intimate kind between root-cells and hyphæ, the formation of a hyphal complex within the cell being the morphological expression of an equilibrium not always present, one which does not commonly exist, for example, in the actively growing roots of early spring. Thus it is possible at this season, and in early summer, to find cases showing the invasion of root-cells by hyphæ that undergo digestion before branching can take place within the cells. Similarly, digestion of hyphæ in the immediate neighbourhood of the apical meristem, presumably a region of great vegetative activity, may be so prompt that stages showing “clumping” of mycelium are not commonly found.

In general, careful study of early spring roots gives the impression of clean, rapid growth with much lighter infection than is the case later in the season. This relative immunity is believed to depend on the interaction of two sets of factors:—

• (1) The differential effect on growth of low soil temperature in early spring.

• (2) Internal factors regulating the reaction of the cortical cells to infection, related in turn with external conditions and the rate of growth.

Factors of the latter group are believed to determine not only the entry of the hyphæ, but also the formation of a mycelial complex within the cell. Assuming the correctness of this view, the formation of root mycorrhiza is a reciprocal phenomenon involving “co-operation” on the part of the rootcells. In other words, it represents a temporary phase of toleration on the part of the plant-cell interposed between one involving immediate destruction of an entering hypha, and the wholesale digestion of mycelium which eventually takes place.

Hyphæ may invade the root-cells at any point: they most frequently enter at the corners by solution of the middle lamella of the common wall (Plate III., fig. 18). In favourable preparations, the middle lamellæ of the root-cells may exhibit in surface view a network of fine hyphæ (Plate II, fig. 12). The endophyte is evidently well equipped for the solution of this part of the wall, and can maintain an existence there if unable to establish itself within the cells.

It is frequently stated that the root meristem of mycorrhiza plants is exempt from fungus attack. This is certainly not the case in Ling. It is the rule rather than the exception to find hyphæ associated with the root-cap cells ; invasion of cells of the apical meristem by branches from such hyphæ can be recognised, and typical digestion stages identified a short distance behind the apex (see Frank, loc. cit.). It appears that, in general, the nearer to the apical meristem, the more promptly doe’s digestion of the mycelial complex within the cell take place. The number of young roots infected in this way increases as the summer advances and may possibly vary with the season and with soil conditions. Absolute confirmation of this would require observations over a number of years.

Invasion of the apical meristem checks growth, and is probably responsible for two features characteristic of the rootsystem of the plant: (1) the profusion of branching, also doubtless affected by the general intensity of infection, and (2) the occasional development of intercalary growing regions in young roots.

As in Orchids, digestion appears to be directly related to increased activity of the nucleus, and the onset and completion of the process correspond with characteristic changes in its appearance. Cell activity in this respect is of an independent nature and adjoining root-cells often exhibit markedly different stages.

After fixation in Carnoy’s fluid and partial removal of the oily contents, a young root-cell, as yet uninfected, shows a delicate radiating cytoplasmic network surrounding a spherical nucleus. The nuclei are remarkably homogeneous after staining, with one or more deeply stained nucleoli. Except in spring or early summer, or under certain special rooting conditions, it is not easy to find such uninfected cells, or to observe the earlier stages of infection; nor is it easy to see the nuclei in the mycorrhiza cells until after the onset of digestion.

When hyphæ enter the cell they appear to be “attracted” by, and are usually found in close proximity to the nucleus : branching follows and the cell becomes filled with a closely branched mass of hyphæ of rather large diameter, the contents of which stain with 1 per cent, osmic acid and also with cotton blue, hæmatoxylin, and other stains. At this stage the staining reactions suggest a transference of stainable material from root-cell to hyphæ. On the other hand, it can frequently be observed that “entering” hyphæ are depleted of their contents, which have passed into the intracellular branches. From these and similar observations it is inferred that hyphæ absorb material from the cells and can also draw upon that in the mycelium outside the root. In this, regarded as the active mycorrhizal condition on the part of the fungus, the complex of hyphæ fills the cell lumen, and very beautiful preparations showing clearly the structure of the mycorrhiza can be prepared in early summer (May and June) by use of the lactic acid cotton blue and other staining methods. Later in the season the amount of mycelium outside the roots increases and, with increasing intensity of infection, the proportion of cells showing digestion stages becomes greater.

The onset of digestion is marked by “clumping” or contraction of the hyphæ about the nucleus. In Plate III., fig. 22, a cell can be seen in this early stage showing otherwise little sign of change except slight increase in size of the nucleus. Later stages of the process appear in Plate I., figs, 1, 2, 8, from which it is clear that the breaking down of hyphæ is initiated and is specially active in the immediate neighbourhood of the nucleus. The latter increases in size, becomes finely granular, and stains more deeply with hæmatoxylin. Slight deformation also occurs, so that it may be difficult to recognise the exact limits of nuclear process and hyphal remains (Plate I., fig. 7). Deformation of the nucleus is, however, relatively slight in Calluna; it is evidently similar in kind but less in degree than that described by Bernard for’the “verdauungzellen” of Orchids.

As digestion proceeds the hyphæ lose their identity, the change proceeding from the nucleus outwards, until the intracellular complex is represented only by a hyphal “stalk,” to which is attached a structureless mass of stainable material (Plate 1., figs. 6, 8a). Later, these cells become practically depleted of stainable contents, containing only the remains of the hyphal walls (Plate I., figs. 5, 8c). During the final stages of digestion the nucleus shrinks as evidenced by its angular outline (Plate I., fig. 5). It stains deeply with hæmatoxylin, and as more than one nucleus has been noted in exfoliated cells, it is believed that fragmentation may occur. Since old cells may be observed showing typical mycorrhiza and others whose structure suggests reinfection by fine hyphæ after the digestion cycle, it is probable that their fate varies with their position upon the root and the season of the year. Moreover, since the cortex is exfoliated and mycorrhiza reformed annually, the matter is not regarded as having much significance.

It is of some interest to note that Frank (1892) figures rootcells of Andromeda polifolia, apparently in the “clumped” stage of digestion, without evident appreciation of their significance.

Turning to the distribution of mycelium in the shoot, the relations of hyphæ with plant tissue become more difficult to observe. It is easy to learn that infection is not confined to the roots but extends to hypocotyl, seed leaves, and epicotyl; it is extremely difficult to obtain satisfactory ocular proof of the exact distribution of mycelium throughout the stem tissue of the mature shoot. Owing to the fineness and weak stainability of the mycelium and its distribution mainly in the middle lamellæ of the cell walls, convincing evidence of the presence of hyphæ throughout the tissues of the stem can be obtained only by extremely careful examination under high magnification. In thin sections only fragmentary portions of mycelium are exposed, and these are liable to be washed away during manipulation of the sections, thus rendering negative evidence of little value. Particularly favourable for examination are torn spaces on the edges of sections, where fragments of mycelium may often be found attached to the walls. Slow maceration of thick sections can also be made to yield positive evidence of the wide distribution of hyphæ. In cells of the pith an intracellular complex of mycelium showing evidence of digestion has been occasionally observed, otherwise the distribution of hyphæ in the living stem tissues appears to be mainly in the walls. The conclusion has been reached that, whereas very fine hyphæ may traverse cells of the stem as of the leaf tissues, formation of an intracellular complex is very rarely accomplished and is evidently subject to digestion as soon as formed. On the other hand, the corky tissues bounding the stem contain abundant mycelium, much of which is identical in character with the large hyphæ of clear brown colour formed by the endophyte upon the roots and in pure culture.

Satisfactory differential staining of the mycelium in the shoot tissues is difficult to secure owing doubtless to its attenuated character. Some success has been obtained with the stains mentioned on p. 271. The photographs in Plate II. show fine mycelium associated with the wood, on the borders of phloem and cortex, and in cells of the pith (Plate II., figs. 9, 14, 15).

The distribution and development of mycelium in the leaves has been described (Rayner, 1915). Hyphæ of normal diameter are found within the air spaces, continuous with others of an attenuated type growing in the walls or occasionally within the mesophyll cells. Rarely a mesophyll cell can be found containing an intracellular complex of hyphæ in a digested or partly digested condition. Fragments of hyphæ may be associated also with the trabeculæ of cell remains which cross the large air spaces in the leaf. These bridging strands in the leaf mesophyll consist mainly of cell walls, and are extremely difficult to differentiate clearly from fine hyphæ. The presence of the latter in this position has become intelligible now that the close association of mycelium with the cell wall has been realised.

This wide distribution of the mycelium in the shoot is so surprising, and the mycelium is so difficult to identify by ordinary methods, that much time has been devoted to further study of its exact distribution.

There can be no doubt whatever that all tissues of the healthy stem are penetrated by a network of extremely fine hyphæ. Although these grow mainly in the walls, branches may traverse the lumena both of living cells and of wood vessels. In the pith hyphæ often become rather more conspicuous, a condition doubtless associated with downgrade changes in cells of this region. The distribution of mycelium seems to be entirely casual, the, hyphæ are excessively fine and their stainability low. As might be expected, evidence of digestion is rare, but has been observed in cells of the pith and in those of the leaf mesophyll.

In general, growth of mycelium in the shoot appears to be restricted by factors present in the living tissues; when these are removed, as in air spaces of the leaves or in the dead tissues of the bark, the development becomes normal and resembles that of hyphæ associated with the root

It was thought that interesting confirmation of the distribution of mycelium in the shoot might be provided by shoot cuttings struck in sterilised media under controlled conditions. Plants propagated in this way should also provide evidence as to the relation of rooting to fungal infection, since the latter can evidently take place as the root emerges from the shoot tissues. Before the experiments bearing on this matter were completed there was published an account of similar experiments carried out by Christoph (1921). The conclusions of this worker on seed germination of Calluna have been criticised elsewhere and do not concern the matter in hand (Rayner, 1922). Before giving an account of the present investigation on cuttings of Ling and Erica carnea, the conclusions arrived at by Christoph on the same matter may be briefly summarised as follows.

The cultures were carried out by striking cuttings of young lateral shoots of Calluna and Erica carnea in sterilised and unsterilised peaty soils. Sterilisation of the soils was effected by heating dry at 120° C. for six hours, and again, after moistening, for thirty minutes at the same temperature. Cuttings rooted both in the sterilised and unsterilised soils. The former gave roots described as entirely free from infection and remained in this uninfected condition during two and a half years.

The ease with which ericaceous plants can be propagated by shoot cuttings varies with the species. In the case of Calluna and Erica camea small twigs of “half-ripened” wood from flowering branches root with fair readiness in moist sand during the summer. These cuttings do not form “callus” on the cut surface, but develop adventitious roots from any part of the stem at or below the surface of the sand.

Owing to the crowding of the small leaves and the roughness of the bark, it is not possible to ensure that the exterior is absolutely free from micro-organisms. Approximation to this ideal was obtained by selecting clean shoots, washing them in fast running water for some hours, making the cuttings, and then rinsing repeatedly in sterilised water. In some cases shoots were immersed in 0. 1 per cent mercuric chloride for one or two minutes and the cuttings made after several rinsings in sterilised water.

So treated, cuttings gave cultures showing cultures growth after many months in closed tubes in a moist atmosphere. Following several unsuccessful attempts, they were found to root well in $112$-inch U-tubes prepared as shown in the diagram (fig. 2). The use of U-tubes ensures an adequate water supply over long. periods without opening the cultures and permits of regulation of the supply by tilting the tube. They also allow free diffusion of gases; the addition of a small tube containing KOH facilitates removal of CO₂ but is not essential for success. Before planting, the whole apparatus was autoclaved at 120° C. on two successive days. If a potash tube is used, the small tubes can be placed in position for sterilisation and the potash added when the cuttings are inserted. After planting, the cultures were kept in a small cool greenhouse, precautions being taken to preserve a moist atmosphere around the tubes and adequate shading in hot weather. Cuttings of both species so treated rooted freely during July and August, often producing an abundance of clean white roots above the sand (figs. 3 and 4).

The roots are produced endogenously and grow out through the cortex and bark in the ordinary way. Microscopic examination, especially after staining, shows that they are infected by fine hyphæ in a manner quite indistinguishable from that associated with a similar stage of root development in seedlings (Plate II., fig. 10). In favourable preparations fine hyphæ may be seen invading the young roots as they emerge from the stem. The mycelium, identical in appearance with that found upon seedling roots, enters the cells freely and may show the early stages of mycorrhiza formation (Plate II, fig. 11). In the cases figured there was nothing to suggest that the mycelium was other than that of the endophyte.

After rooting, cuttings can be grown on in (U-tubes by replacing the rain-water with a complete nutrient solution, or may be transferred to autoclaved peat watered with sterilised rain-water; the latter method involves the possibility of soil infection from the air. In a study of infection by a specific endophyte this may be ignored, and the latter method was adopted as a means of obtaining older roots for comparison with those described by Christoph.

The compost used for plotting was a mixture of peat and sand in which Calluna and Erica camea were known to flourish. Pots and potting material were sterilised by autoclaving at 120° C. for thirty minutes on three successive days. The pots, after planting, were stood upon glass plates under a bell-glass in a small experimental greenhouse and watered with sterilised rain-water as required.

It is well known that the complete sterilisation of soil by heat causes the production of substances inimical to plant growth. Great difficulty has been experienced in cultivating rooted cuttings during the first few weeks after planting in peaty soil sterilised as described. Plants are slow in resuming growth and apt to show symptoms of toxic poisoning by discoloration of the shoot from below upwards. Both Calluna and Erica carnea suffer more severely in this way than do species of Vaccinium. After some weeks the toxic properties disappear, presumably as a result of leaching, and the plants grow in a normal manner. In the case of cuttings the difficulty is increased by the fact that these strike most readily during the summer, and consequently must be transferred to sterilised peat in late autumn when growth conditions generally are unfavourable. Many well-rooted cuttings of Ling died after removal to sterilised peat and some difficulty was experienced with older plants if repotting became necessary.

For critical study, the roots of a Calluna cutting, struck in sterilised sand as described, and subsequently grown from October to May in sterilised peat watered with sterilised rain-water, have been available. The history of the plant used is as follows:—

Well-rooted in sand in September, it was potted in sterilised peat early in October and survived the winter, growing vigorously until May, when the roots were examined. The roots showed the clean transparent growth characteristic of spring roots in the open (p. 269). Examined casually, they appeared to be free from mycelium. Borne by a plant the roots of which were known to be infected when transferred to peat, this seemed unlikely and after staining they were examined more carefully under high magnification. Signs of abortive infection were found to be abundant—fragments of mycelium attached to the corners of cells, coils and threads of fine hyphse upon the surface of the root (Plate III., figs. 17, 18). More detailed examination showed that in many places these roots exhibit in superficial view a continuous network of mycelium in the middle lamellæ of the walls, from which hyphL branches invade the cells in the typical manner. In short, the initial stages of mycorrhiza are present but the condition is not established, and it is evident from the staining reactions of the hyphæ that their granular condition of active growth in the middle lamell æ of the walls is quickly impaired when they enter the cells (Plate III., fig. 18). These roots confirm the view already expressed (p. 272), that mycorrhiza can only be formed under certain conditions of the root-cells. The roots of cuttings struck in the ordinary way in clean sand in a cold frame show a precisely similar condition, i.e. they provide ample evidence of infection, of the initial stages of mycorrhiza formation, and of destruction of intracellular mycelium before the typical condition can be established. If the objection is raised that a similar condition would be likely to arise if the mycelium present were that of a casual invader, it can be replied that there is no evidence that this is the case. The early stages of infection in the roots of cuttings struck under rigidly controlled conditions are in every way identical with those observed in roots of seedlings; the character of the mycelium, its mode of entry and appearance in the cells is similar, while the roots themselves show no signs of attack by an ordinary parasitic fungus.

Conclusive proof of the presence of the endophyte in an active condition in the shoot tissues was obtained by itsextraction from twigs of the plant in the following manner.Unsuccessful attempts had been made to root cuttings in tubes of nutrient agar as used for work on pure-culture seedlings. After some weeks in the tubes without rooting, these twigs developed a superficial growth of mycelium which closely resembled that formed on pureculture seedlings when inoculated by a too vigorous strain of the endophyte. By sub-culturing from this growth to suitable media, the endophyte was obtained and identified in the pycnidia stage. The culture was contaminated by Cladosporium sp., known to be a common “impurity” on the roots and shoots of the plant (fig. 5).

The imperfect development of mycorrhiza in roots of cuttings is regarded as extremely significant. It has enabled the writer to link up and interpret a number of puzzling facts noted at various stages of the investigation on Calluna, and has provided a means of explaining the observations on Vaccinium published by Stahl in 1901, and criticised by the present writer in an earlier paper (Rayner, 1915). It is now offered as an explanation of the discordant results obtained by Christoph (loc. cit.) on seedlings and cuttings of Calluna and Erica camea grown in sterilised peat. It is suggested that reexamination of these roots, using a technique similar to that described in this paper, might yield evidence of a similar type of root infection. When first observed, the older roots of cuttings just described (p. 282) appeared to be quite free from infection, and an elaborate technique and careful comparative observations have been necessary in order to provide material for the interpretation now offered.

The abortive stages of mycorrhiza formation described and figured (Plate III.) correspond exactly with those noted in the early spring roots of plants growing under field conditions, and also with those observable in seedlings normally infected but potted in or transferred to sterilised soil. They agree also with those found in the roots of pure culture seedlings grown in watery extract of a favourable soil (Rayner, 1921). The latter exhibit identical phenomena in the mycorrhiza cells, and a similar (apparent) freedom from infection when examined unstained without critical illumination.

The writer’s present conclusions may be summarised briefly as follows:—

The development of the endophyte in the mycorrhiza cells of Calluna is markedly inhibited by certain conditions in the rooting medium, and roots exposed to such conditions may appear to be uninfected.

The conditions at present known to be responsible in greater or less degree for this condition are as follows:—

• (a) Under experimental conditions.—A favourable soil or watery extract of such soil sterilised by auto-claving at 120° C. ; sterilised sand cultures irrigated with rain-water or a weak solution of mineral salts.

• (b) Under natural conditions.—Low soil temperature in early spring.

The disappearance of mycelium from the roots of plants in dry places and in pot cultures allowed to become dry recorded by Christoph (loc. cit.) is doubtless susceptible of a similar explanation. It may, indeed, be due to a similar physiological cause, eg. physical drought as compared with the physiological drought associated with low soil temperature.

As at present observed, restriction of growth of the endophyte in the root-cells leading to apparent suppression of mycorrhiza—or indeed to its actual suppression, if the use of the term is restricted to the more typical condition in which the root-cells are filled with mycelium—may be directly associated with conditions in the rooting medium unfavourable to vigorous though not to healthy growth, i.e. low temperature, drought, and a limited supply of nutrient salts, and is induced also by a sterile rooting medium.

It is reasonable to infer from this that the formation of active mycorrhiza is closely correlated with the activities of the root-cells as regards absorption of water and solutes ; and that the intracellular development of the endophyte varies with external factors in the rooting medium directly related with the nutrition of the vascular plant.

The exact ecological significance of these facts in Nature and their bearing on the edaphic relations and distribution of the species is not yet clear. Experimental work now in progress upon the nutrition of the endophyte in pure culture may throw light upon them.

As a consequence of the observations on roots, and the experimental work on cuttings, the view has been reached that fungal infection and the stimulus to development associated with it on the one hand, and the formation of root mycorrhiza on the other hand, must be regarded as distinct phenomena. The obligate relation is associated with the former: there is no experimental evidence, nor has a claim been put forward by the writer, that the formation of mycorrhiza is in any sense obligate, although under ordinary soil conditions it is the natural sequence to infection of the Ling seedling, and may be closely bound up with the soil relations of the roots.

The seasonal study of roots, together with the observations on cuttings now described, emphasise the importance of this distinction and justify the conclusion stated, namely, that mycorrhiza formation in Ling is a reciprocal phenomenon, conditioned, not only by the activity of the fungus, but by the reaction of the root-cells and the nature of the rooting medium.

The above considerations have strengthened the view expressed by the writer that in Calluna, as in the Orchids, the obligate nature of the relation between vascular plant and fungus concerns a particular stage of development only, i.e. could the stimulus to development ordinarily supplied by infection be replaced by the addition of a suitable substance to the rooting medium, plants raised in this way would continue to grow without formation of mycorrhiza. Experimental proof of the correctness of this view is lacking for Ericaceæ, inasmuch as the seed cultures recorded by Christoph were not carried out under strictly controlled conditions; moreover, no experimental proof has been offered that the seeds used were adequately sterilised. Indeed, such proof is obviously impossible in soil cultures in which an outgrowth of micro-organisms does not reveal itself as is the case in gelatine or agar. Moreover, assuming imperfect sterilisation of seed, the roots of seedlings grown in sterilised soil might be expected to show the condition of “suppressed mycorrhiza” now described, and unless examined at germination, the stage of active infection might easily be overlooked. Lacking convincing experimental proof, the conclusions of Christoph on this matter appear to the writer unjustified.

Assuming the practicability of raising rooted plants artificially without infection, the question arises, will they be helped or hindered in the “struggle for existence” by absence of the endophyte? Evidence bearing upon this could, be sought by growing plants, rooted without infection and normally infected respectively, upon sterilised substrata, but caution would have to be exercised in applying the results to those growing under natural conditions. Comparative data of this kind are not yet available for the Orchids, although seedlings rooted in media containing relatively high percentages of sugar have been found by Knudson (1924) to grow satisfactorily when transferred to a full nutrient without sugar.

The endophyte of Calluna has the general equipment of a parasite, and by suitable cultivation in pure culture can be so altered as to behave as a typical parasite when reinoculated into experimental plants. The conclusion seems irresistible that the present relation has evolved from one of parasitic attack by a soil fungus, aided possibly by the appearance of races or strains of the parasite of a more or less benign type.

It has long been clear to the writer that the relation of vascular plant and fungus in Ling cannot be explained as a case of parasitic attack, countered and held in check by the defence mechanism of the plant. The observed facts in this species are not explained, either by assuming with Christoph that the fungus is a “harmless parasite,” or believing, with Bernard,—on the analogy of the Orchids—that the condition is one of parasitic attack adequately met by an efficient mechanism of resistance in the vascular plant, Orchid or Heath as the case may be.

Co-ordination of growth has reached a relatively advanced stage in Calluna, and it is believed that similar methods of experiment on other members of the Ericales will yield evidence of a variable degree of specialisation within the group.

The varying degree of development of the endophyte observable under natural conditions at different seasons, and under different soil conditions (see present paper, p. 272, and Christoph, loc. cit.), and the definite correlation now recorded between this and the nature of the rooting medium in experimental cultures, suggest that the production of mycorrhiza by the Calluna plant is bound up with particular soil conditions. In view of the new facts recorded in this paper, the hypothesis may be put forward that the endophyte operates as an internal factor of a special kind directly related to the metabolism of the root cells in certain kinds of soil.

Throughout the work the writer has kept in view the original problem attacked, viz., the physiological significance of the calcifuge habit in Ling and allied species (Rayner, 1911). The observations on pure culture experiments with seedlings in favourable and unfavourable soil solutions are regarded as significant (Rayner, 1916): the facts now recorded may help to elucidate these results, inasmuch as they have provided a series of pictures of the cytology of the mycorrhiza cells under field conditions, and also when the rooting medium is strictly controlled in cultures.

The problem is obviously extremely intricate ; the prolonged attempt to unravel it in the case of Ling is justified not only by the practical interest of any contribution towards an explanation of the calcifuge habit, but also by that accruing to a better understanding of the significance of mycorrhiza in general Its widespread occurrence in timber trees, for example, arid the distribution of many such on soils akin to those favoured by Ling and Heather raise points of much practical interest.

The case for exchange of nutrient material, with a balance of profit on the side of the Ling plant, may be set out as follows:—

In the Orchid seedling, the stimulus of infection can be replaced artificially by supplying a nutrient substance, e.g. sugar. The great stimulus to growth, formation of chlorophyll and production of roots and leaves, which follows infection in Calluna, is similar in kind and suggests that the fungus brings in something directly or indirectly concerned in metabolic activity.

Whether this is of the nature of a vitamine essential to development and not to adult growth, or an enzyme capable of altering the osmotic pressure of the cells, or releasing some substance otherwise inert, is at present a matter for surmise. That it is of the nature of an ordinary food constituent is also possible, but is regarded as improbable in view of the ample reserves contained in the seeds.

Intracellular digestion of mycelium takes place throughout the plant tissues, regular and extensive in the young roots, sporadic elsewhere. In root-cells which have undergone digestion, granular and oily material undergoes change and disappears: whether such removal of contents is to be accounted for by leakage into the soil or by regular translocation of metabolic material to growing regions is unfortunately not susceptible of experimental proof. Assuming that the passage of soluble products to the vascular strands is inhibited by the endodermis, as stated by Priestley (1922), there seems to be no reason why such material should not be used either at the growing points of the roots concerned, or in the formation of the lateral roots that are freely produced during the growing season and again during the development of a new root-system in the following spring.

Fixation of atmospheric nitrogen by the endophyte is a specially tempting hypothesis in Ericaceæ, in view of their usual habitat on poor soils deficient in nitrates and suitable inorganic compounds of nitrogen. There is direct evidence that the endophytes of a number of ericaceous species can utilise atmospheric nitrogen in greater or less degree, supplemented in the case of Calluna by indirect evidence that “pure culture” seedlings grow with marked vigour on media free from combined nitrogen (Ternetz, 1907 ; Rayner, 1922). Very rigid proof is rightly demanded, but it is held that there is already a strong case for fixation of nitrogen by the endophyte of Calluna. If it occurs, digestion of mycelium by the root-cells in the wholesale manner described assumes an increased importance in the metabolism of the plant.

As already noted the endophyte is a facultative parasite, and at a certain stage undoubtedly absorbs material from the root-cells. Its exact behaviour towards various food-substances when growing in “pure culture” is at present under investigation. The results of this work, together with those derived from microchemical tests upon the mycorrhiza cells of the root, will be published shortly. The facts already ascertained have influenced the point of view adopted in the present paper.

To the writer, it seems important that the nutrition of mycorrhiza plants should be studied in relation to the physiology of parasitism in general. It is believed that experimental investigation of individual cases will provide evidence of evolutionary stages in the mycorrhiza habit, ranging from more or less effective resistance to parasitic attack to the stabilised relation displayed by Calluna.

Individual cases require investigation each on its merits; in those where the relations with the endophyte are habitual, specific, and very specialised, the “symbiosis” appears to be essentially of the nature of a “balanced parasitism,” in which the apparent stability may be dynamic rather than static, i.e. now one and then the other partner is more active. The reciprocity involved in the formation of active mycorrhiza as described in Calluna marks a relatively advanced stage, involving extremely intimate nutritive relations and resulting in a “balance of profit” for the vascular plant. The last term of the series—marked by complete dependence of higher plant upon fungus—will be represented in Ericales should, e.g. Monotropa— now awaiting experimental investigation—exhibit a relation comparable with that shown by Gastrodia among the Orchids (Kusano, 1911).

In conclusion, a plea is put forward for a revised usage of the term symbiosis. As at present employed in botanical literature, it implies a reciprocal relation involving mutual benefit to the participants, and is frequently applied to cases for which no experimental data are available. As descriptive of mycorrhiza plants, for example, it conveys the impression that the relation between endophyte and vascular plant is of similar nature throughout the group, and differs in some fundamental way from that which exists in parasitism generally.

At first defined by de Bary—the living together of dissimilar organisms— the term covers a vast biological field, and used in this broader sense the cases included would require further classification, eg. in the manner recently suggested by MacDougall (1918). In which category any particular case should be included depends upon the existence of adequate experimental data, and until these are provided, must remain, to some extent, a matter of opinion.

The experimental work described in this paper has been carried out in the Pilcher Research Laboratories, Bedford College, University of London, with the aid of a grant from the Dixon Fund of the University.

Figs, 1 to 18 are from untouched photomicrographs, taken with critical illuminetion and Wratten colour screens. Zeiss apo. horn. imm. Obj. 2 m.m. Compens. oculars 4 and 6 (except figs. 10 and 11). Root-cells figured are from whole roots.

Figs. 19 to 22 are camera lucida drawings of cells from whole roots.