On the nutritional dependence of certain trees on root symbiosis with belowground fungi (an English translation of A.B. Frank's classic paper of 1885)

Full text

Mycorrhiza (2005) 15: 267–275
DOI 10.1007/s00572-004-0329-y
ORIGINAL PAPER
B. Frank
On the nutritional dependence of certain trees on root symbiosis
with belowground fungi (an English translation of
A.B. Frank’s classic paper of 1885)
Received: 1 June 2004 / Accepted: 14 September 2004 / Published online: 10 November 2004
# Springer-Verlag 2004
To promote the possibility of truffle cultivation in the
Kingdom of Prussia, His Excellency, the Minister of Agri-
culture, Domains and Forestry, commissioned me to ap-
proach the matter systematically. I was to begin with scientific
studies on the conditions of occurrence and development
of these fungi. Certain facts had already been established
through observations and experience. For example, true
truffles occur only with living trees, and in the Prussian
truffle districts these investigations had established a strong
relationship between truffle occurrence and particular tree
species: beech, hornbeam and oak. Above all else was the
union of Elaphomyces mycelium with pine roots, as rec-
ognized by Rees (Sitzungsber. D. physik.-med. Soc. zu
Erlangen, 10 May 1880).
From the outset these facts pointed the research towards
determining whether the true truffles also establish a
connection of the mycelium with living tree roots. As this
communication will show, the question must begin much
farther back, because it presupposes knowledge about the
nature and nutrition of plants not heretofore even slightly
suspected by science. This shall be nearly the only topic of
my present paper. It concerns the fact that certain tree
species, above all the Cupuliferae, do not nourish them-
selves independently in the soil but regularly establish a
symbiosis with fungal mycelium over their entire root
system. This mycelium performs a “wet nurse” function
and performs the entire nourishment of the tree from the
soil. Surprising though this may sound, it is solidly based
on the scope of my research.
When one examines feeder rootlets in the soil—the root
system’s end branches representing the actual organs of
nutrient uptake—of any of our native oaks, beech, horn-
beam, hazel or chestnut, it is evident they are generally
composed of two disparate components: a core, represent-
ing the actual tree root, and an organically united mantle of
fungal hyphae. This fungal mantle completely encloses the
rootlet, forming a continuous cover even over the growing
tip. It grows along with the root tip and behaves in every
respect as an organically united, peripheral tissue belonging
to the root. The entire structure is neither tree root nor
fungus alone but resembles the lichen thallus, a union of
two different organisms into a single, morphological organ.
It can be appropriately designated as a “fungus-root ” or
“mycorrhiza.”
Translation from German of “Über die auf Wurzelsymbiose
beruhende Ernährung gewisser Bäume durch unterirdische Pilze”
from Berichte der Deutschen Botanischen Gesellschaft (1885)
3:128–145, revised from an earlier translation in Molina R (ed)
Proceedings of the 6th North American Conference on Mycorrhizae
(1985)
Translator: James M. Trappe (e-mail: trappej@onid.orst.edu, Tel.:
+1-541-7378593, Fax: +1-541-7371393), Department of Forest
Science, Oregon State University, Corvallis, OR 97331–5752, USA
Translator’s note: This is not a literal translation, because the style of
nineteenth century German scientific writing often does not translate
comfortably into twenty-first century English. For example, the
original German title of Frank’s paper word-for-word in English
would read, “About the on-root-symbiosis-depending nutrition of
certain trees through underground fungi.” In adjusting syntax in the
translation, I have paid particular attention to preserving Frank’sorig-
inal intent and meaning. Frank used the family name “Cupuliferae,”
which was interpreted in the nineteenth century to include the present
Betulaceae and Fagaceae, or sometimes only the Fagaceae, but has
long since been discarded. Frank used it in this paper generally in
reference to Carpinus, Corylus, Castanea, Fagus and Quercus spp.).
In deriving the term “mycorrhiza” from the Greek μυκοριζα, Frank
transliterated it with a single “r”. Transliterating the Greek ρ (rho) into
Latin letters ideally requires that the “r” be doubled and followed by
an “h” in certain compound words to make the meaning and deriva-
tion clear, as discussed by W. T. Stearn [Botanical Latin, 4th Addition
(1992) Timber Press, Portland, Ore., p 261). However, many authors,
including Linnaeus and Frank, omitted the additional “r.” For
purposes of the science of mycorrhizae, it is trivial: whether with
one “r” or two, our communication is equally effective. Similarly,
whether one uses the Greek plural “mycorrhiza,” the Latin plural
“mycorrhizae,” the English plural “mycorrhizas,” or the plural end-
ings of other languages is of no consequence. “Rhiza” is feminine in
Greek; Frank retained that gender in German, using German feminine
case endings. To reflect his usage, I have used the English rather than
Latinized possessive and plural endings in this translation.
B. Frank (*)
Plant Physiology Institute,
Royal Agricultural University,
Berlin, Germany

Structure of the mycorrhiza
In surface view the mycorrhiza resembles many fungal
sclerotia in fine structure. It shows a pseudoparenchyma
constructed of irregularly and very tightly interwoven
hyphae, the cells ranging by diameter or length about 2.4–
100 µm (Fig. 3). The cell walls are relatively thin, some-
times nearly colorless, and other times bright or dark brown,
so that the mycorrhiza appears bright to brownish or nearly
black. The pseudoparenchyma is usually multi-layered,
forming a rather thick mantle as seen in cross- or longi-
tudinal section. This mantle does not simply lie intimately
on the root epidermis; hyphae from it penetrate between the
epidermal cells of the root itself (Fig. 6). The cortex, in-
cluding the epidermis, generally consists of about four cell
layers. The epidermis and the subepidermal layer, or only
the former, are composed of relatively broad, radially ex-
panded cells; as a rule only the epidermis is intergrown with
hyphae. The inner cortical cells always remain free. I could
never detect hyphae in the endodermis or fibrovascular
bundles. These endophytic hyphae always grow only at the
cell wall surface. They never enter the cell lumen but com-
pletely weave around all sides of most cells. The hyphae are
only 1.2–2.4 µm in diameter, considerably thinner than
those of the fungal mantle but clearly recognizable as a
continuation of the mantle. The thin walls of the peripheral
root cells restrict their growth to a small diameter. Accord-
ingly, one sees them most clearly in the surface view of the
cell walls. Because they are normally arranged in a tight,
almost pseudoparenchymatous layer, they seem at first to be
a delicate reticulation of the root cell wall. Every cross-
section through the cell wall, however, demonstrates that
this structure originates from the intercellular hyphae ap-
pressed to the cell wall.
The mycorrhiza surface is often perfectly smooth, so that
the fungal mantle is sharply differentiated from its sur-
roundings (Fig. 3). The tightly appressed fungal mantle
prevents formation of root hairs: I have never observed
them. However, root hairs are often replaced by similar
structures of the fungal mantle: its superficial cells extend as
filaments that elongate by tip growth to spread among the
surrounding soil particles. The manner in which this occurs
varies greatly. Sometimes the mycorrhiza is irregularly clad
with a thick, loose felt of bright or pale brown filaments
similar in diameter to those of the mantle but varying
between specimens from 1.2 to 3.6 µm in diameter. These
meander back and forth, generally spreading among the soil
particles (Fig. 4). Often they are enlarged where attached to
small soil particles in the manner of lichen rhizoids or plant
root hairs. Sometimes a multitude of long, straight, robust
hyphae, brown or concolorous with the fungal mantle or
blacker, extend into the soil, so that the mycorrhizas seem to
have a broom or tuft of hairs at their tips when suspended in
water. Sometimes the hyphae coalesce here and there into
strands ranging from a few hyphae to forming rhizomorphs
almost as thick as the mycorrhiza itself (Fig. 7). Without
careful examination, these can be easily confused with the
mycorrhizas themselves. The hyphae of these rhizomorphs,
commonly corresponding in color and size to those of the
mycorrhizal mantle, often emanate from the rhizomorph
surface into the surrounding soil as numerous, individual
filaments. In the truffle districts, they customarily abound in
the soil, especially near a fruiting truffle. They form a
system of innumerable branches and anastomoses spread-
ing through the soil. Their connection with the mycorrhizas
of any Cupuliferae present in the soil is easily confirmed.
Development of the mycorrhiza
Longitudinal sections show that the fungal mantle contin-
ues to the tip of the mycorrhiza and likewise completely
encloses it. The core of the tip elongates through a clearly
developed growing point, thus showing all characteristics
of a true root (Fig. 5). Therefore it must be assumed that
the fungal mantle can expand in pace with the elongation of
the root that it encloses. Indeed, it has its growing region
where the growing point and zone of elongation of the root
lie. The cells that comprise the fungal mantle over the root
tip are always much smaller than those farther back, where
the root is no longer elongating, i.e. 0.8–2.4 µm broad and
up to about 5 µm long. They gradually grade to the larger
cells farther back from the tip (Fig. 5). The fungal mantle
thus enlarges by the continuing insertion of new hyphae
between the existing ones at the tip of the mycorrhiza, and
the cells of the pseudoparenchyma thus constructed broad-
en to their final size. Both parts of the mycorrhiza keep up
with each other in this growth pattern, so that the fungal
mantle always compactly overlies the growing root tip.
Growth of hyphae into the epidermis first occurs where
length growth of the rootlet has stopped, not in its growing
region. Progressing from younger to older regions in lon-
gitudinal sections, one can clearly follow the gradual pen-
etration of the endophytic hyphae from the surface of the
epidermis.
In terms of growth phenomena, we differentiate the
tissues common to all root tips: apical meristem, procam-
bium, protoderm, and root cap. And, we see how these
meristems change to the usual permanent tissues of the
root: the central fibrovascular cylinder, the root cortex, and
the epidermis. The classification of these meristems follows
the types that apply to most dicotyledons, as discussed in de
Bary’s Comparative Anatomy, where the root apex with a
sharply differentiated apical meristem and procambium is
covered by a common layer of initials for the protoderm
and root cap (Fig. 5).
The weak development of the root cap is remarkable:
often at a given time only a single cell layer is present,
while as soon as the next one begins to split off from the
protoderm, the layer becomes disorganized. Remnants of
the older cap layers can often be recognized as thin, brown
masses compressed by the surrounding fungal mantle and
soon becoming unrecognizable. Reduced root cap forma-
tion may be understandable in this case, because the fungal
268

mantle occupies the space needed for root cap development
and necessarily replaces the root cap. Still, it is interesting
to see that the inherent histological differentiation of the
root remains intact despite the symbiosis.
How the mycorrhiza develops in the soil with the young
plant is another question. Naturally, the fungi do not
colonize the radicle of the embryo in the seed. The tap root
of the first stage of germination is also free of fungus. The
tap root soon develops lateral roots, which remain rather
slender and form numerous short, repeatedly branched,
feeder rootlets that appear almost coralloid. The fungal
colonization progresses on these lateral second and lower
order roots. At individual sites the hyphae at first appress to
the root epidermis for a short distance and, as they branch
and spread further over the root, they anastomose with
themselves and other hyphae. The fungal mantle is grad-
ually constructed from such starting points.
The colonization seems to progress most rapidly on
Carpinus; as a rule, the whole a bsorbing root system of
1-year-old plants is converted to mycorrhizas. It occurs
more slowly on Quercus, so one can easily follow the pro-
cess: 1- and 2-year-old plants or individual parts of the
root system of older plants are often only partially colo-
nized. These fungus-free feeder rootlets are clad with root
hairs, as are the feeder rootlets of plant species that never
form mycorrhizas. Still, such fungus-free roots of the
Cupuliferae are relatively infrequent. Moreover, at least the
tips of such rootlets are often colonized, while the my-
celium spreads mainly towards the younger part of the root
system. It soon is able to colon ize tips of the young root-
lets, because they grow slowly and always remain short.
Only the stronger and vigorously growing long roots, which
penetrate root-free parts of the soil and bear the actual feeder
rootlets, usually remain free of fungus. As with young
plants, the fungi colonize feeder rootlets on older parts of
the root system directly from the soil.
Fungus-induced changes in the root
The mycorrhiza differs in form from the noncolonized
rootlet. When we cultivate broad-leaved trees under fungus-
free conditions, as will be discussed below, the feeder roots
are relatively thin and long, their lateral branches emerge
monopodially at fairly distant intervals, and they resemble
the parent root in shape and branching. In contrast, the
mycorrhiza shows a very slow length growth but attains
greater thickness, because the cell layers of the apical mer-
istem and procambium proliferate somewhat more, and the
epidermal cells reach a greater width to form a short, rel-
atively thick structure. Moreover, a stronger tendency to
branching develops, so that the lateral rootlet branchlets
occur at short intervals close behind the tip. These branch-
lets behave similarly to the parent rootlet in growth, form
and branching. The mycorrhizas thereby form more or less
coralloid growths (Fig. 1) that often develop into large
clusters (Fig. 1).
The branching of the mycorrhiza occurs by the endog-
enous mode common for roots, and the new growing tip
that emerges from the parent rootlet thus is clad with the
parent rootlet’s fungal mantle from the beginning. From
then on the fungal mantle continues growth with the new
rootlet branch as described for the parent rootlet. The
branching is strongly monopodial: in spite of the coralloid
growth form, no dichotomy is to be seen.
The first branchlet always forms behind the tip where
length growth has ceased. Branching proceeds acropetally,
so the branchlet nearest the tip is always the youngest and
shortest. These branchlets form rather clearly in longitudi-
nal rows as is usual for roots: sometimes in two, sometimes
in three, sometimes in four rows, occasionally in only one
row, patterns that may partly depend on prevailing condi-
tions of space. In addition, gradations occur in the changes
of form produced by the fungus. Sometimes the feeder
rootlets approach the form of noncolonized roots (Fig. 1a).
In that case the fungal mantle still envelops them but is not
as thick as on roots with the most pronounced coralloid
form.
Subsequent fate of the mycorrhiza
The distinctive combined organ of fungus and root as de-
scribed above generally has a limited life span. Still, it
shares this property on the whole with feeder rootlets of
woody plants. As the trees expand their root systems with
age and invade new areas of soil, the feeder rootlets of the
aging parts are lost, while new ones develop on growing
parts of the system in other areas of the soil. The mycor-
rhizas, which have determinate growth, cease growing after
a while or regrow only on individual branchlets until they
stop altogether. Eventually they die off to shrivel and turn
dark-brown to black and brittle. Despite that, as already
mentioned, they can form anew at other places, often close
to a deceased mycorrhiza cluster.
How long mycorrhizas remain alive depends on a
number of circumstances and may vary considerably. They
may persist for many years: one often finds huge nests of
mycorrhizas which, considering their slow growth rate,
must have taken a long time to form. As with nonmycor-
rhizal tree roots in advanced age, the cortical cells turn
brown in the oldest parts of a mycorrhiza as the dying
process commences, while the fibrovascular strand con-
tinues to function within the protective endodermis. Death
of the fungal mantle goes hand in hand with that. The
stronger growing tips of some mycorrhizas may convert to
perennial, lignified branches of the root system through
further lengthening and thickening. They establish a cork
cambium beneath the endodermis and a vascular cambium
in the fibrovascular cylinder, in the process losing their
fungal mantle. The mantle is only for the younger rootlets,
particularly those involved with nutrient uptake.
269

270

Regular presence of the fungus in all ages
and on all roots of the tree
To study the roots of Cupuliferae at various ages, I acquired
1-, 2- and 3-year-old seedlings of oak, beech, hornbeam
and hazel from various localities, as well as root samples
with feeder rootlets of older trees: a 120-year-old beech, a
100-year -old hornbeam, and a 40-year-old hazel. The feeder
rootlets of all these trees at all ages were developed as
mycorrhizas, accompanied by the fungus through the entire
life of the root. It was interesting on the older trees how the
fungus behaves at the various soil depths in which the roots
occur. I pursued that especially with the beech and horn-
beam. In the uppermost soil layer, about 5 cm thick and
relatively humus-rich, the rootlets customarily form the
largest number of feeder rootlets, and these, as already
mentioned, are always developed as mycorrhizas. These
mycorrhizas are surprisingly abundant in this soil layer,
especially at truffle sites, so that the ripe truffles rest on, and
in, a thick weft of mycorrhizas. In deep soil, one can follow
how the frequency of feeder rootlets decreases with depth,
at first gradually but then increasingly so. The stronger
roots freely penetrate deep layers, but there they form
feeder rootlets only sparingly or do so mostly on branches
that have penetrated upward into more shallow soil layers.
In forest soil with the parent material at a depth of 50 cm,
roots could be traced in the parent material but formed
feeder rootlets only sparingly. Even there, however, the
feeder rootlets develop as mycorrhizas. One could explain
this by a distribution of the root-colonizing fungus in all
soil layers. More simply, however, it can be clarified by
penetration of the parasite into the soil along with the root,
which is always colonized as it elongates into the deeper
soil layers.
Occurrence of the root fungus on plant species
It is extremely interesting that this soil-inhabiting fungus
selects the rootlets that it colonizes strictly by species and
thus abides by a strong systematic constraint. For example,
in the soil of a beech stand, only the beech roots developed
as mycorrhizas. All the herbaceous vegetation that occurred
there, such as Oxalis acetosella, Mercurialis perennis,
Anemone nemorosa, Asperula odorata, Viola canina, Con-
vallaria multiflora, etc., as well as other woody plants, e.g.,
Hedera helix and Acer pseudoplatanus, have roots com-
pletely fungus-free and with root hairs, as is usual for plant
roots. This is even the case when those roots are near or
growing through a mycorrhizal cluster.
To accurately determine the spectrum of host plants used
by the root fungi, I examined most of our woody plant
genera and can specify those where the root fungi fail:
Betula alba, Alnus incana, Ulmus campestris, Morus alba,
Platanus occidentalis, Juglans regia, Pyrus malus, Sorbus
aucuparia, Crataegus oxyacantha, Prunus padus, Robinia
pseudoacacia, Tilia europeae, Acer platanoides and pseu-
doplatanus, Rhamnus cathartica, Cornus mas, Fraxinus
excelsior , Syringa vulgaris, Sambucus nigra. Thus the great
majority of plant families to which the native trees belong
are free of the root fungus, as judged by the representatives
here examined.
Because the limitation to a small spectrum of plants
always points to the Cupuliferae, I studied its most im-
portant representatives in this regard and verified the pres-
ence of root fungi without exception. The roots of these
trees show an essentially constant fungal colonization of the
kind described above: Carpinus betulus, Corylus avellana,
Fagus sylvatica, Quercus pedunculata and sessiliflora,
Castanea vesca, from samples from the Rheinland as well
as the Berlin Botanical Garden, and the American Quercus
rubra from the local botanical garden. From these results
one may assert that the root fungi are a special feature of all
the Cupuliferae. Indeed, this symbiosis is so constant for
this family that one could be tempted to regard it as a
systematic criterion. At any rate it is worth mentioning that
the inclusion of the Betulaceae in the Cupuliferae, as
accepted by the newer taxonomists, does not seem to be
supported as judged by the occurrence of the root fungi.
On the other hand, it also interesting to see a hint through
the occurrence of these fungi beyond the Cupuliferae, of a
certain kinship of some other families with the Cupuliferae:
the Salicaceae and the Coniferae. I have found mycorrhizas
3 Fig. 1 Root of Carpinus betulus laterally bearing mycorrhizas that
are mostly coralloid thickened. a A branch system that is thinner but
also colonized by fungi. ×1
Fig. 2 Cluster of mycorrhizas on a lateral root emerging from a
taproot (r) of a 1-year-old Carpinus betulus. ×1
Fig. 3 Part of a mycorrhiza of Carpinus betulus with a short lateral
branch; the surface view illustrates the structure of the outer, smooth
fungal mantle composed of a small-celled, pseudoparenchymatous
tissue, through which in places the contours of the underlying root
epidermis can be seen. ×145
Fig. 4 Tip of a mycorrhiza of Fagus sylvatica, showing the form of
the fungal mantle with loosely arranged, superficial hyphae that
grow out irregularly into the soil. ×145
Fig. 5 Longitudinal section through a growing mycorrhizal tip of a
1-year-old Carpinus betulus. p Procambium; r root cortex; e epi-
dermis; c ephemeral root cap layer tha covers the root apex; d general
layer of initials of the protoderm, root cap and epidermis and from
which forms the root cap layer through periodically repeated divisions
to the outside; m fungal mantle, comprised of pseudoparenchymatous
tissue that surrounds the entire growing tip of the root and the cells of
which are narrower at the root tip, the youngest part of the root, than
further back on the root. a Detached remnants of the once contiguous
cells of the now detached root cap layer surrounded by hyphae of the
fungal mantle
Fig. 6 Part of a longitudinal section through an older portion of a
mycorrhiza of Carpinus betulus, leading through the epidermis (e),
the cortex (r), and the cells that border the fibrovascular strand (f). a
The direction towards the root tip, b towards the root base, m the
fungal mantle, which forms the surface of the mycorrhiza and from
which a pseudoparenchymatic cell layer continues inward to weave
around the walls of the epidermal cells; in the lower part of the
figure one sees the same in cross section of the epidermal cell wall,
in the upper part a surface view, not unlike a fine, reticulate
thickening of the wall. ×480
Fig. 7 A mycorrhiza of Fagus sylvatica from Alefeld in the
vicinity of a truffle (Tuber aestivum), with a very strongly developed
fungal envelope, which extends through the soil partly as free my-
celial filaments and partly as almost rhizomorph-like mycelial strands
that in turn disperse as free mycelial filaments. ×40
271

also with them, but not so generally as with the Cupuliferae.
Salix viminalis, S. caprea, and S. aurita, as well as Populus
tremula, originating from many sites, were colonized by the
fungi in varying degrees, although no colonization was evi-
dent at other sites. I found roots of Taxus baccata, Juniperus
communis, and Larix europaea in the vicinity of Berlin to
be free of the fungi, but those of pine, spruce, and white fir
near Berlin to be mostly but not everywhere colonized in
the typical manner. Rees (loc. cit.) has similarly described
this colonization on pine roots in sites where Elaphomyces
occurs, but it is evident that the root fungus is much more
broadly distributed on conifers than Rees believed, includ-
ing sites where no sporocarps of Elaphomyces have been
found.
Geographic distribution of the root
fungi of the Cupuliferae
Once the fungi under consideration were found on the
Cupuliferae in sites that produce no truffles, so that the
occurrence of those fungi appeared to be more and more a
general phenomenon, it was appropriate to systematically
clarify their distribution. By arrangement with His Excel-
lency, the Minister, I have received for study roots of all
species of Cupuliferae occurring in a large number of forest
districts, representing as much as possible the variety of soil
conditions and geographical situations in the Kingdom of
Prussia. The plants were mostly 1– 3 years old, but samples
of older trees were also included. The primary result can be
revealed at the outset: the mycorrhizas occurred in all re-
gions, and no Cupuliferae were free of root fungi.
Beech, hornbeam and oak collected from various com-
partments of the forest district at our most southwestern
border near Saarbrücken were colonized by the fungi with-
out exception. So were beech from Rügen Island, hornbeam
from the Brödlauken Forest District in the Gumbinnen
Administrative District at our eastern border, and indeed
from all the regions in between from which Cupuliferae
were examined. The results show that differing elevations
and soil types do not limit the fungus: it occurs in the flood
plains, e.g., in the flood zone of the Elster of the Schkeuditz
Forest District and in the Elbe lowlands near Gräfenhain-
chen in the Merseburg Administrative District. It occurs
outside of floodplains as well as in completely flat areas or
gently rolling hills, e.g., it was constant on plants from the
forest complex of the Dübener Heath between the Elbe and
Mulde, south of Wittenberg from the Zöckeritz Forest Dis-
trict near Bitterfeld, from near Berlin as well as from the
Jülich flatlands.
In hill and mountain regions the fungus was found at all
elevations and exposures, on the plateaus as well as in the
valleys and hollows and on slopes, equally on north, south,
west and east aspects, and without difference between
gentle and steep slopes. The fungus also ascends with the
beech into the higher mountain regions.
Considering soil conditions, both shallow and deep soil
are suitable for the fungus; its behavior here has been
discussed above. Also, none of the geologic conditions of
the soil exclude the parasite. It is constant in diluvial soil
from various regions, indeed in strongly humic river loam
no less than in light, more or less humus-poor sand (e.g.,
from various places near Berlin), as well as in the inter-
mediate formations of sandy loam and loamy sand with
varying humus contents. Further, it is in the loamy sand soil
that is the weathering product both of carbonaceous sand-
stone (e.g., from Münster) and of the new red sandstone
(e.g., from Saarbrücken); in greywacke soil, e.g., from the
Eifel, Westerwald, etc.; then in the red loam soil weathered
from the red sedimentaries at Sangerhausen. Finally, it
occurs on all types of limestones, namely shell limestone
(e.g., Freiburg on the coast, Heldrungen, Wanfried on the
Werra, Friedland on the Leine), platey limestone (from
southern Hannover, e.g., Alefeld, etc.), and no less on Eifel
limestone (e.g. Schleiden on the Eifel) as well as chalk soils
of Rügen Island.
It merits mention that the fungus always seems to
develop most luxuriantly on limestone substrates. Finally, it
must be emphasized that no vegetation type hinders the
appearance of the fungus on the Cupuliferae present; it
occurs equally in high forests of trees grown from seed,
mixed high-coppice forests and coppice forests, in fields,
and no less outside the forest where Cupuliferae are raised,
such as parks, gardens, tree nurseries, etc. Indeed, I found
the fungus growing along with roots of plants I had potted
with soil in flower pots about 2 years before and then had
allowed to grow.
It might seem odd that botanists have missed the root
fungus of the Cupuliferae before now, as it has such a wide
distribution. Plant roots, especially the root tips, have been
studied botanically many times, but in general only seed-
ling radicles have been used. When that is done with the
Cupuliferae, the roots are at a developmental stage that
precedes appearance of the fungus. Researchers dealing
with root diseases of the Cupuliferae could hardly miss it,
however. Much contemporary plant pathology work is
conducted uncritically; this innocent fungus was found
accidently in pathology work, and, although no studies
were conducted on its significance, it was assumed and
declared to cause some anomalous growth phenomena.
However, it is an inalienable part of every beech and oak
tree, and, as we will see, serves as an important “wet nurse”
for their nutrition.
Misinterpretation happened in fact with the ink disease
of chestnut that occurs particularly in upper and central
Italy. It is called that because it commences as an initial
blackening and general dying and rotting of the roots. In
truth, the cause of this disease is not investigated at all.
Gibelli [Nuovi studi sulla malattia del castagno detta
dell’inchiostro (1883) Bologna], who has been much occu-
pied with this disease, believes it to lie in fungi that appear
on the rotting chestnut roots; he would characterize them as
272

Torula exitiosa de Seyn., Diplodia castaneae Sacc., and
Melanoma gibellianum Sacc. The evidence on the causal
relationship of these fungi to the disease is lacking, so much
the more so because these and similar fungus formations
occur as decomposers on plant parts rotting on and in the
soil after death by any cause.
Gibelli also noted the true mycorrhizal fungus on living
feeder rootlets of chestnut. His description leaves no doubt
about its identity with our fungus: the coralloid, tubercu-
late, swollen roots with tips capped by a pseudoparen-
chymatic mycelial tissue and entangled with branched
rhizomorphs, as illustrated in Plates IV and V of his report.
He was so biased towards the concept of a root-injuring
enemy, however, that he equated this root fungus with the
fungi on rotting roots listed above as a cause of the disease
under consideration. Gibelli did state that he had found the
characteristic root fungus to be general in Italy on roots of
healthy chestnut trees as well as oaks, beech, hazels, and
other Cupuliferae. But even these observations were insuf-
ficient to lead him to a different interpretation, that the
actual injurious parasite had already expanded to a wide
distribution in Italy and causes the disease, but the tree does
not suffer from the attack as long as it is growing vigor-
ously, rather only if it is weakened from other causes. No
further confirmation that Gibelli’s viewpoint is erroneous is
needed beyond what has been said above and what is to be
said below about the biological significance of the root
fungus. For us the interesting fact emerges that Italy is
included in the general geographical distribution of the
mycorrhizas of the Cupuliferae.
The biological discovery of mycorrhizas of the Cupu-
liferae could also have been possible earlier in Germany. R.
Hartig (Untersuchungen aus dem Forstbotanischen Institut
zu München (1880) I. Berlin, p 1 ff.) studied a root disease
of 1- to 3-year-old oaks in various sites, as reported under
the title, “The oak root killer, Rosellinia (Rhizoctonia)
quercina.” He regarded that fungus as the cause of the
disease, which produces a massive rotting of the taproot
and the lower parts of the stem in the field and also grows
on the soil surface. This fungus does not correspond with
our root fungus; R. Hartig did not confuse the two, because
he did not examine the oak roots well enough to discover
the mycorrhizal fungus.
Because the mycorrhizal fungus occurs so generally and
regularly that no Cupuliferae can be found without it, I
pondered on a way to artificially free the plant from its
“nurse” to force it to take up nutrients independently. I
succeeded in this by water culture: 1- and 2-year old seed-
lings were lifted from the soil in late winter and transferred
with intact root systems colonized by the fungus into a
nutrient solution composed of compounds common for
water culture.
After some weeks and before the buds burst, root for-
mation began. The mycorrhizas already present grew no
further, but very bright, new rootlets, easily differentiated
from the previously formed, darkly colored ones, were
produced laterally at various places. This is the common
phenomenon: roots formed in soil do not develop further
when placed in water. Rather, new roots initiate in the
water. The root fungus passed over to these new rootlets in
its characteristic form, partly by forming loose mantling
hyphae that assumed a kind of colorless water form. But it
unquestionably could no longer keep up with the rootlet
formation. The base of the new rootlets still showed the
extended fungal mantle, but it seemed less distinct, thinner,
and often so interrupted that broad stretches of the epider-
mis showed only a spotty fungal mantle. The remaining
surfaces were bare, thus prepared for root hair formation that
would otherwise be repressed by the fungal mantle. The tips
of the new rootlets were free of the fungus. A 3-year-old oak
cultivated from germination in water culture and never in
soil was in accord with this: its strongly developed root
system was completely fungus free. From the facts related
thus far, we must conclude that the fungi of the mycorrhiza
thrive best on roots in soil, are generally distributed in
vegetation-supporting soil, and from such soil colonize
rootlets of the Cupuliferae.
The question of the species of the root fungi
The systematic position of the fungi in question can be
determined only by identification of their fruiting bodies.
The occurrence of the mycelium on the roots necessarily
directs attention to the hypogeous fungi, above all the
Tuberaceae and many Gasteromycetes. It might appear
strange that the ubiquity of the root fungus mycelium is not
accompanied by a similarly general occurrence of the fun-
gal fruiting bodies. That can be for two reasons, however.
First, with extremely attentive searching, hypogeous fungi
can often be found even where one does not expect them.
Second and foremost, the presence of the mycelium of a
fungus is not necessarily accomp anied by the appearance
of its fruiting body at just any time. There are examples
enough that the mycelium of a fungus can grow year round
without forming fruiting bodies, and that the latter appear
only when certain environmental conditions are met. So,
we are presented with the question of whether or not it is
possible to determine the fungi by their mycelial character.
The many variations in form, thickness, color and con-
nections of the hyphae that branch off from the mycorrhiza
into the soil have been described above. More precise ob-
servations, however, quickly lead to the recognition that
these characteristics are not usable for specific differenti-
ation without elaboration, in that they may vary on the
same mycorrhiza; at least in part these variations represent
changes in form of a single mycelium.
If one compares the mycorrhizas of a truffle site with that
of a site not bearing truffles, no sharp differences are to be
seen, even in significant morphological characters. Often
the differences that do exist are primarily quantitative in the
biomass development of the mycorrhizas and the preva-
273

lence of the mycelium in the soil, which reaches its max-
imum in truffle sites. Accordingly, we can assume that the
fungi, which produce truffles in many regions, are much
more broadly distributed than are the truffles themselves.
Perhaps they are quite common fungi and, where truffles
are lacking, their fruiting body production is limited by lack
of the proper environmental conditions.
On the other hand, it is unwarranted to conclude without
further study that the similarity of the mycelia means the
fungi are everywhere the same. In keeping with the com-
mon rule for fungi that related species present no reliably
differentiating mycelial characters, various species of Tu-
beraceae probably cannot be differentiated by their myce-
lium alone. However, these questions are beyond the scope
of this paper and should be left for when results of pending
studies are available.
Biological and physiological
significance of the mycorrhiza
The roots of the Cupuliferae and the fungal mycelium
organically unite into a morphologically unique organ. The
intimate, reciprocal dependence that follows the growth of
both partners and the tight interrelationships of physiolog-
ical functions that must exist between the two appear to be a
new example of symbiosis in the plant kingdom. It goes
beyond the lower organisms to the most highly developed
plant form, the trees, and is incontestably both unexpected
and surprising. First of all, the fungus mycelium must be
regarded as an undoubted parasite on the living cupulifer
root, as is evident from the entire manner of its colonization
and penetration into the growing rootlet. As is the case for
all parasitic fungi, the basic nutritional needs of the fungus
are primarily the carbon compounds procured from the
photosynthesizing tree. In contrast, the fungus is evidently
independent in regard to uptake of soil minerals, in that it
alone contacts the soil by its peripheral position on the
mycorrhiza and the innumerable hyphae it extends into the
soil to grow around soil particles like root hairs.
Now, the question of great interest must be, is the tree
damaged by the fungal parasitism of its rootlets? We know
from a multitude of cases that parasitic fungi damage their
host plants. The morphological changes assumed by the
tree rootlets under influence of the parasite can be char-
acterized as hypertrophy or gall formation, albeit relatively
weak. This suggests an irritation by the fungus on root
growth. However, the root is in no way killed by the fun-
gus, and despite its change it does not lose the capacity to
function for the tree, as the prosperity of the latter ade-
quately demonstrates. For the same reason, the idea that the
fungus deprives the tree of mineral nutrients carries no
weight. Were this so, healthy beeches and oaks could not
exist, because each cupuliferous tree is accompanied by the
fungus from its first year to advanced age. We conclude
from all that, the root fungus, at least in the mycelial state,
can inflict absolutely no disadvantage on the tree.
This fact imposes the stamp of symbiosis on this rela-
tionship, because both of the united organisms live together
in reciprocal assistance without harm to each other. The
fungus fulfills a reciprocal service for what it receives from
the plant, a service of eminent significance, because it
represents the most important factor in the nutrition of the
tree. That soil water and nutrients needed by the tree are
supplied only through the mediation of the fungus cannot
be challenged: it envelops the entire surface of the feeder
rootlets, and its hyphae perform the role that root hairs do
for other plants in intimately contacting the soil particles.
The enlargement of the volume of the root epidermal cells
and their complete envelopment by the hyphae produce an
arrangement probably designed for nutrient uptake by the
tree. The fungus takes up soil minerals not only for its own
nutrition but also for that of the tree, so we must consider
that the root fungus is the sole organ for uptake of water and
soil nutrients by oaks, beech, etc. It functions as a wet nurse
of the tree in this respect. Thus, in contrast to autotrophic
plants and trees, the Cupuliferae show a relationship that
can be termed “heterotrophy,” i.e., nutrition from soil with
help of another organism on a truly splendid scale, known
before now only with lichen gonidia and some lower algae
living within higher plants.
The symbiosis of the Cupuliferae most closely parallels
that of the lichens, specifically in its biological character,
even allowing for the differences, i.e., the association meets
both the requirements and outputs for nourishment of both
partners. Indeed, the root fungus is analogous to the lichen
hyphae and the tree to the lichen alga; the comparison need
not be elaborated further. A complete analogy even appears
to prevail in reference to what degree these symbiotic re-
lationships are either necessary or dispensable for both
partners. The lichen algae are known to exist independently
of the lichen fungus and can develop as a free alga after
isolation from the lichen. Similarly, as previously men-
tioned, the Cupuliferae can be cultivated fungus-free in
water culture for years. Of course, the Cupuliferae do not
develop strongly when free of fungi in water culture. Still,
that is at least partly due to the unusual medium, for the
same thing shows up with other land plants grown in this
culture method. Whether the Cupuliferae can nourish
themselves better with their fungus nurse than without is
not known from these studies, because no adult Cupuliferae
seem to be fungus-free. On the other hand, as the lichen
hyphae do not develop strongly and, in any case, never
attain typical fruiting in the absence of the algae, so also the
mycorrhizal fungi seem to depend on the chlorophyllous
tree for their development.
So far I have not succeeded in growing the hyphae from
pieces of living mycorrhizas in water or fungal nutrient
solutions such as plum decoction. Moreover, the strong
dependence of the fruiting of truffles on the presence of
living trees is emphatically significant here. This would not
preclude the supposition that a weak, perhaps somewhat
saprobic, development of the fungus is possible in the soil
without the nourishing tree to explain the general distribu-
274

tion of the fungus in soil that supports plants. Finally, the
root fungi are also reminiscent of the dependence of lichens
on specific substrates. The occurrence and fruiting of many
lichens are restricted to specific types of rock. So also the
fruiting bodies of hypogeous fungi occur in striking rela-
tionship to soil properties aside from their dependence on
the nourishing trees. For example, the edible Tuber species
indicate with certainty an underlying limestone. The two
symbiotic relationships here compared thus differ only
morphologically through the differentiation and organi-
zation of the body of the tree as compared to the alga.
(received on 17 April 1885)
275