THE STATE OF TAXONOMY OF THE
GENUS ARMILLARIA
Harold H. Burdsall, Jr. and Thomas J. Volk
Center for Forest Mycology Research
Forest Service, U. S. D. A.
One Gifford Pinchot Drive
Madison, WI 53705
The genus Armillaria has been the bane of mycologists ever since its establishment by Fries as a tribe of Agaricus (1821), a genus to which he ascribed
nearly every gilled fungus, regardless of spore color, nutritional situation, or
other characteristics we use in taxonomy today. Even Fries was uncertain
as to how he wanted to recognize the tribe Armillaria; four years later he
placed the species in the genus Agaricus, tribe Lepiota (1825), then again recognized the tribe Armillaria in 1838 with twice as many species as it had
originally. Finally, in 1857 Staude raised Armillaria to generic rank, and
although there has been great discussion regarding the person “legally” responsible for this move, we agree with the arguments that support the authority
being Staude (Watling and Kile, 1982), with the genus formally designated
Armillaria (Fr.:Fr.) Staude.
Then there is the question of which generic name to apply to this group.
Controversy over which author was the first to raise Fries’ tribe Armillaria to
generic rank have led to the widespread use of Armillariella in the last 15
years or so. Fortunately this dilemma has been resolved, and the generic
name Armillariella has been gently laid to rest (stomped to death, actually.)
Roy Watling, Glen Kile, and Norma Gregory provided a beautiful eulogy in
1982. We can now ignore the name Armillariella as an obligate synonym of
Armillaria and get comfortable with Armillaria for these fungi.
When one considers the broad distribution of its species, it is very surprising that the genus Armillaria has gone so long with so little attention to its taxonomy. For example, at least until the early 1980’s here in North America, no
one had even heard of the widely distributed and very common A. ostoyae.
On the other hand, there have been over 250 species referred to the genus
Armillaria, although most of these are only distantly related to the group
now called by that name. Included among them are species of many descriptions and functions that can now be placed in over 25 modern genera. The
genus Armillaria in the modern and biologically functional sense is reserved
for facultatively parasitic root- and butt-rot fungi that produce rhizomorphs.
Common species such as Armillaria ponderosa (= Tricholoma magnivelare,
the American matsutake) are now found in the genus Ticholoma with other
mycorrhizal species. Conversely, the fungus formerly called Clitocybe tabescens
belongs in Armillaria with its root-rot relatives.
This paper will concentrate on the taxonomy of Armillaria in the Northern
Hemisphere. Most of our comments will also apply to the Southern
Hemisphere, but we are not as familiar with those species, and there will be
enough names and species concepts without introducing others that may only
confuse the issue. Photos of many of the species from both hemispheres can
be found in Watling et al. (1991).
4
Back in the early sixties the literature on A. mellea was extremely confusing. It was considered by different researchers to be either a virulent pathogen
(in the west) or an opportunistic pathogen (and then not very virulent) in the
eastern United States. Its host range was phenomenal, one of the broadest
known for fungi, and its morphology was extremely variable.
In the early seventies Hintikka (1973) developed a technique that allowed
the recognition of several European species in this group. Hintikka’s technique demonstrated the existence of five Armillaria species in the European
species complex. The technique depended on growing single spore isolates
together in a petri dish and observing the change or lack of change in colony
morphology. Single spore isolates of Armillaria species are generally white and
quite fluffy They need to fuse with a single spore isolate of opposite mating
type (not male with female, but similar in effect) in order to complete the life
cycle. When fusion of compatible mating types occurs, the coalesced colonies
become dark brown, appressed (flattened), crustose, and sometimes (depending on the species) produce rhizomorphs. If the single spore isolates are from
different species, the colonies will not grow together and will remain white
and fluffy. Using this method, Korhonen (1978) was able to distinguish five
Biological Species (BS). These are called biological species rather than morphological species because isolates within a species are able to interact in a biologically meaningful (i.e., sexual) way. In addition, the species were diflicult
to distinguish using morphological characteristics, which is still the case
today. Appropriate names have been applied to each of these European BS as
follows: BS A = A. borealis, BS B = A. cepistipes, BS C = A. ostoyae, BS D = A.
mellea, BS E = A. gallica (a.k.a. A. bulbosa and A. lutea ). Unfortunately, in
spite of the ability to distinguish the five European biological species by using
mating compatibility, the problem of distinguishing the morphological species
that we are more used to dealing with has not been entirely resolved.
Morphological distinctions were (are) still diftlcult to recognize. Added confusion resulted from the application of different names to these various BS. For
example, BS C was called A. ostoyae or A. obscura, depending on who was to
be followed and BS E was known as A. bulbosa or A. lutes at this particular
time. Later the name A. gallica was to be applied to this BS.
In the mid-seventies Anderson and Ullrich (1979) applied the techniques
developed by Hintikka and expanded by Korhonen to isolates collected from
widely distributed locations in North America. This work demonstrated that
what had been considered as Armillaria mellea in North America was actually 10 genetically isolated biological species (North American Biological
species or NABS). This led to substantial confusion in both mycology and
forest pathology circles; the more conservative element, who liked the simple concept of Armillaria mellea referring to all the taxa, felt that “those damn
splitters” were playing some more nomenclatural games (as taxonomists have
been known to do). In 1988, Anderson determined that two of the NABS
delimited in 1980 were superfluous. However, by this time another biological
species, NABS XI, had been recognized by Morrison et al. (as species F, 1985),
and with A. tabescens now considered a member of the genus, the species
count was once again ten. Anderson et al. (1980) demonstrated the genetic
compatibility of the NABS with all but one of the European BS; only A. borealis was found to be genetically isolated (incompatible) from all the NABS.
5
As a result the NABS have been designated by the names of their European
counterparts with the exception of NABS V and X which are incompletely
and inconsistently compatible with A. cepistipes of Europe.
With this new technology available for distinguishing species of Armillaria,
mycologists both in North America and in Europe began trying to distinguish
the biological species using morphological characters. This desire led to the
accumulation of as many characters as can be brought to bear to make these
distinctions. As always we like the easy characters, but easily used characters are in short supply in this group. Some of the more familiar characters,
including spore size and shape, are not useful for this endeavor, since they
overlap or are very similar between taxa.
Nevertheless, with practice and experience, there are morphological characteristics that can be used on fresh specimens. Color of the basidiome, presence and type of squamules, shape and color of the stipe, habit, host
associations, and geographical distribution can all be used in the effort to distinguish the species. However, applying the characters to dried specimens
is very difficult at best and providing a written description from which a dried
specimen can be identified, at least with some of these species, is practically
impossible. Distribution is an important character for some species, but more
as a negative character, and the possibility that an exception is in hand is
always a problem. For example, after three years of heavy collecting of
Armillaria in Wisconsin and northern Michigan we were “absolutely convinced” that A. mellea sensu stricto was restricted to the southern half of
Wisconsin, but in the fourth year we found A. mellea S.S. in the Upper
Peninsula of Michigan! The point is, since the concept of having many
Armillaria species is relatively new, the exact distribution of many of the
species is incomplete because the sometimes complicated determination of
the correct species is being practiced by only relatively few laboratories.
One of the primary questions that is asked is why does anyone care about
distinguishing the species of Armillaria? From the mycophagist’s view they are
all similarly edible and delicious, although there are unconfirmed reports
from the Pacific Northwest about slight toxicity of Armillaria ostoyae collected on hemlock. But from the view of the forester, it becomes very important to distinguish between species that vary greatly in their pathogenicity. For
example, A. mellea is known to kill trees, especially oaks, that have been
weakened by drought or by other pathogens. A. ostoyae is known as a virulent
pathogen of conifers. On the other hand, A. gallica is usually an innocuous
saprophyte, living on organic matter in the soil and not harming trees to any
great extent. A forester finding an Armillaria fruiting in the woods would
like to be able to tell whether or not there is a potential problem with
Armillaria root disease so that mitigative procedures can be taken if necessary.
Very little is known about the pathogenicity of other species of Armillaria at
present.
In spite of the difficulty in describing the species characteristics sufficiently
to allow easy identification of Armillaria species, three of the NABS not found
in Europe have been described as A. gemina (NABS II), A. calvescens (NABS
III) and A. sinapina (NABS V). The last mentioned is partially compatible
with A. cepistipes from Europe as is the unnamed NABS X. NABS IX, X, and
Xl are still to be described and named. These species are rarely found fruit-
6
7
ing and are thus delimited km the other species based on their incompatibility
reactions. It is possible that additional biological species will be found with
further collecting and testing against cultures of known species. It is also
possible that additional species will be “demoted” to subspecies.
Since the delimitation of the segregates of the Armillaria mellea complex
began, numerous nomenclatural problems and inconsistencies have been
encountered. An example is that of deciding whether the name A. obscura
or A. ostoyae should be used. It was finally determined that the concept of
A. obscura was not firm enough to be used and that because there was an
available type specimen for the name A. ostoyae, that name was the best to be
used (Marxmüller, 1992). The situation with the name A. bulbosa created
more confusion, because among the several different groups working on this
most widely distributed species, the names A. bulbosa, A. lutes, and A. gallica
were all being used. Finally with the publication of Marxmüller’s article
(1992) the nomenclature of the species was settled. Armillaria gallica is the
name to be used. Even though the other two names have priority (i.e., they
were published earlier), they were rejected as ambiguous names that could
refer to a number of species.
Using a broad spectrum of characteristics it is often possible to “provisionally” identify the more common Armillaria species of North America. The
following key may be of use in such a venture. We are currently developing
a key using both macroscopic and microscopic characters, once we figure out
what those are.
8
5. Annulus cortinaceous, very thin and collapsing quickly on the stipe as a
fibrous zone, usually with bright yellow tints, but sometimes pale cream; cap
brown to reddish brown, covered with rather dense brown fibrils, sometimes with small scales; stipe 5-8 cm long, 0.5-1 cm broad at apex, slightly
wider at base, shades of brown toward base but covered with a thin layer
of appressed yellow fibrils or more distinct annulations of bright yellow
fibrillar material (remains of the yellow fibrous universal veil); occurring
mainly on deciduous trees in the northern tier of the eastern United States
and throughout forested areas of southern Canada, but frequent on conifers
in the Pacific Northwest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.
. sinapina
5. Annulus not yellow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6.
6. Cap light to medium brown, covered with reddish brown to dark brown
fibrils sometimes clustered into hair-like tufts, occasionally scaly, annulus membranous to thick and cottony, with brown pigmented margin,
stipe nearly concolorous with cap, tapered to base, occurring in small
to large clusters, most frequently found at base of dead or dying conifers,
distributed throughout the northern United States and southern
Canada . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .A. ostoyae
6. Nearly identical to A. ostoyae but associated only with deciduous trees,
especially maple; distributed from southern Quebec and Ontario
through northeastern USA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .A. gemina
Morphological descriptions of species.
Armillaria mellea (Vahl:Fr.) Kummer. Although this name was widely used
for nearly all members of the genus until the 1970’s, true Armillaria mellea
has a smaller distribution, occurring mainly in the southeastern United
States up into the northeast into Quebec and in the midwest in the hardwood
forests, although it is occasionally found on conifers in mixed forests. It has
been reported from northern California but is not common in the west.
Although A. mellea is the type species of the genus, it is rather different than
the other species, and is relatively easy to identify with practice. It is characterized by its clustered caespitose habit, usually 8-10 but sometimes in
groups of 30 or more, which results in a tapered to pointed base on each individual stipe. The pileus is honey colored and the surface is smooth, with no
squamules. However, the most consistent character of this species is microscopic; this is the only species of Armillaria that lacks clamp connections at
the base of the basidia. Spores 7.2-8.9 X 5.6-6.7 µm.
Armillaria ostoyae (Romag.) Herink. This species is mostly restricted to
conifers and is a serious pathogen of many species in the northern conifer
zone. It is also found in caespitose clusters, sometimes pointed at the bases
of the individual stipes, with a brownish annulus. The pileus is brown, covered by dark scales, and can be very large, up to 1 foot in diameter, especially
in the Pacific Northwest. This is the species that has been reported to cause
some people some gastrointestinal upset if collected from hemlock, but we
believe that most of the reported upset is due to undercooking (and overeating) of some of these larger collections. Spore size 8-11 X 5.5-7 µm.
Armillaria sinapina Bérubé & Dessureault. This species is similar in color
9
to A. ostoyue, with similar, although somewhat smaller dark scales on the
pileus. It normally occurs singly or sometimes in clusters of two or three. In
the Northeast it can be found mainly on hardwoods and occasionally conifers,
but in the Pacific Northwest (Idaho, Washington, Alaska) it appears to occur
only on conifers. The major morphological character that distinguishes this
species from the others is the “presence of a golden yellow universal veil that
covers young fruiting bodies and later leaves remnants consisting of yellow
warts or lumps of tissue on the cap, a yellow fibrous annulus, and many
patches of fibers on the stipe” (Bérubé & Dessureault, 1988). Spore size 8.210 X 5.9-8 µm.
Armillaria gallica Marxmüller & Romagnesi. This is by far the most widely
distributed species east of the Rocky Mountains, including the northeast,
midwest, and Gulf coast, but is very rare in the west. It is almost always
found on hardwoods, but occasionally on conifers. It can be distinguished by
its solitary to gregarious habit, usually on soil, but occasionally “climbing”
onto logs or stumps to fruit. The fruit bodies, which tend to be smaller than
the other species, are directly attached to rhizomorphs (Darmono et al., 1992).
The partial veil is cortinaceous, i.e. very similar to the cobwebby cortina of a
Cortinarius species, leaving white “arachnoid” remnants on the stipe. The
pileus is “tan to pinkish brown” and “distinctly hairy” (Bérubé & Dessureault,
1988). The base is commonly swollen and sometimes stains yellow where
bruised. Spore size 7.2-9.5 X 4.8-6 µm.
Armillaria calvescens Bérubé & Dessureault. This species is very similar
to A. gallica. It is also found solitary on the soil, but is usually found in large
clusters, sometimes up to several hundred fruiting bodies, although the stipes
do not radiate from a common point as in A. mellea. Pileus tan to brown,
with no scales, but a finely fibrillous surface that can be easily rubbed off,
leaving a “mealy” texture between your fingers. Veil golden yellow, sometimes not easily observed, sometimes leaving fibrils on the stipe, which is
often swollen at the base. Usually found on maple in southeastern Canada and
New England, but also found on a variety of other hardwoods west to Michigan
and Wisconsin. The mycelium of this species has also been found in the
Canadian prairie provinces, although it has not been observed fruiting there
(Mallett, 1990). Spore size 8.5-10 X 5-7 µm.
Armillaria gemina Bérubé & Dessureault. This is a rare species, reported
only from southern Quebec, southern Ontario, upper New York state and
Vermont, down to West Virginia. It is identical to A. ostoyae in terms of habi
tats, pathogenicity, and morphology (Bérubé & Dessureault, 1989), but known
only from hardwoods. It can usually be distinguished by geography, but in
areas where the ranges of the two species overlap, they can be distinguished
only by cultural studies. Spore size 8.2-10 X 5.2-7 µm.
Armillaria tabescens (Scop.:Fr.) Dennis, Orton, & Hora. This is the only
species of Armillaria in North America without an annulus of any type. It
is very common in the southeastern United States, but can be found into Ohio,
Michigan, and Illinois, especially in moderated climates near the Great Lakes.
It is always found in caespitose clusters, usually 8-10 mushrooms, but sometimes of 50 or more. Spore size 8-10 X 5-7 µm.
10
Literature Cited
11
North American Biological Species (NABS) of Armillaria
some synonyms, and known distributions
NABS Species & Synonyms
Distribution
I Armillaria ostoyae (Romagn.) Herink
= Armillaria obscura (Schaeff.) Herink
[nomen ambiguum]
northern
conifer zone
II Armillaria gemina Bérubé & Dessureault
northeast USA,
Quebec, Ontario
III Armillaria calvescens Bérubé & Dessureault
Quebec to Michigan
and Wisconsin,
Canadian prairie
V (IV) Armillaria sinapina Bérubé & Dessureault,
but partially interfertile with A. cepistipes
northern
conifer zone
VI (VIII) Armillaria mellea (Vahl:Fr.) Kummer
hardwood zone,
N. California
to Wisconsin to the
Appalachians to
Quebec
VII Armillaria gallica Marxmüller & Romagn.
=Armillaria lutes Gillet [nomen ambiguum]
=Armillaria bulbosa (Barla)
Kile and Watling [misapplied name]
hardwoods, deep
South to Northeast
to Midwest,
rare in Pacific
northwest
IX unnamed
known from Idaho,
Connecticut, Alaska,
and British Columbia
X unnamed, but partially interfertile with
A. cepistipes
known only from
British Columbia
and Idaho
XI Interfertile with A.cepistipes Velenovsky
= Species F, Morrison et al. 1985
known only from
British Columbia
Armillaria tabescens (Scop.:Fr.)
Dennis, Orton, & Hora
Southeastern USA
into the Northeast,
west to Ohio, S coasts
of Great Lakes
Journal of American Amateur Mycology
Vol. 11 No. 1 1993
Publisbed by the North American Mycological Association
12
Printed on recycled paper